Page 7 :
Stanford Research Institute, , Mechanism and Structure, in Organic Chemistry, ), , HOLT, RINEHART AND WINSTON, New York - Chicago - San Francisco, Toronto - London
Page 8 :
lS'3,1, , £1^5") MEL, , NS‘j, , Copyright © 1959, by Holt, Rinehart and Winston, Inc., Library of Congress Catalog Card Number 59-8696, 10, , 11, , 12, , 13, , 14, , 15, , 16, , 17, , 18, , CFTRI-MYSO, 9537, Machan.srr, and st., , 2056109, , Printed in the United States of America, , 19
Page 9 :
Preface, , As a result of the very striking growth of physical organic chemistry during the, last thirty years, nearly all colleges and universities that grant advanced degrees, in chemistry offer at least one course in this field to graduates, and some offer, such a course to promising undergraduates as well. Moreover, many chemists, who obtained their degrees during the 1930’s or 1940’s have felt the need to, familiarize themselves with this subject, for much of the most interesting, .research being carried out today in organic chemistry is strongly influenced by, mechanistic thinking., , ,, , fr, , The bulk of this text is devoted to the consideration of the mechanisms of, homogeneous organic reactions, but the first three chapters are structural in, outlook. These deal with the structures-df atoVns and organic molecules and, with some of the physicochemical methods used to determine the positions of, atoms and to study the distribution of electronic charge. The treatments here,, although in accord with modern structural thought, tend to be brief; and, in, some cases, the approach is descriptive rather than rigorous. These chapters are, included because it is now clear that a satisfactory understanding of many, organic chemical phenomena cannot be obtained if molecules are visualized, merely as “ball-and-stick” models and atoms as miniature “solar systems.”, Much of the material in these chapters could be presented equally appropriately, in texts in inorganic and physical chemistry; and those who, in previous training, have become familiar with the fundamentals of structural chemistry may, wish to start directly with Chapter 4., A major difference between this text and others in the same field is the, incorporation of a rather large number of exercises at the end of each of the, sixteen chapters. It is intended that these develop the reader’s ability to suggest, mechanisms from experimental data, to relate structure to reactivity in many, different situations, to predict the course of reactions which are new to him to, explain features about reactions which might otherwise puzzle him, and’ if, possible, to propose experiments which serve to distinguish between a number, o possible mechanisms. Some of the exercises are relatively straightforward, but, many require considerable thought. Some, I hope, will challenge the efforts of, even the most capable readers., reader'll, °f, in the «** and in the exercises rises markedly as the, passes from the earlier to the later chapters. Such a trend is desirable, for, iii
Page 10 :
IV, , Preface, , a student should become more proficient in the subject at hand as the term, progresses. Nevertheless, a class for whom the material in the first six chapters, is new will not, except in rare instances, be able to master the entire remainder, of the text within a single term. It is suggested that in an introductory course in, theoretical organic chemistry for beginning graduates or undergraduates, the, first ten or eleven chapters be covered; for these allow consideration of many, of the more usual reactions upon which contemporary mechanistic thought is, based. A more sophisticated group of graduates, or one whose major interest, is organic chemistry, may, on the other hand, find it more profitable to begin, at Chapter 7 or 8, and cover as many of the subsequent chapters as time allows., The author of a textbook in physical organic chemistry must face the, problem of selection. No attempt has been made to consider, even briefly, two, broad classes of reactions, polymerization and heterogeneous reactions. Beyond, this limitation, I have chosen those of the more familiar organic reactions which,, mechanistically speaking, have been investigated most intensively. Less usual, reactions are included when they present points of fundamental interest that, enable us to put the more usual reactions into better perspective. Inevitably,, a number of reactions of some synthetic importance (including the Bucherer,, Leuckart, Reformatsky, Wolff-Kishner, Jacobsen, and Mannich reactions, and, the periodic acid and selenium dioxide oxidations) have been left out, largely, because they, at present, appear to contribute less to the overall mechanistic, picture than many others., For a text of this sort, complete documentation is not necessary, nor is it, desirable. Individual references to early work which has since been reviewed, many times are kept at a minimum, but references to papers published since, 1945, especially to those appearing in journals most readily available to readers, in western countries, are extensive. Altogether, about two thousand separate, references have been included., It is a pleasure to thank Professors Ronald Breslow and Murray Goodman,, who read and commented on major portions of the text; Professor George, Wheland, who made a number of valuable suggestions on the early chapters;, and Professor Harlan Goering, who read the final chapter. I am grateful also, for the help of Miss Elizabeth Brown, who prepared the entire manuscript; and, for the aid of my wife Marjorie, who made the preliminary sketches for the, many figures in the text and exercises and handled innumerable additional, details. Finally, my thanks goes to the National Science Foundation for a, Science Faculty Fellowship, during the tenure of which this book was completed., E.S.G., April, 1959, Brooklyn, New York
Page 11 :
Contents, , CHAPTER, , PAGE, , 1 Atomic and Molecular Structure. The Use of Resonance, Wave-mechanical Picture of the Electron, The Four Quantum Numbers. The Pauli Exclusion Principle, “Build-up” of the Periodic Table; s and p Electrons, Ionic and Other Electrostatic Bonds vs. Covalency, Formal Charge, Recent Interpretations of Covalent Bonding, Promotion, Hybridization, and the Tetrahedral Carbon Atom, Double and Triple Bonds, The Use of Resonance, The Transition between Ionic and Covalent Character, Limitations of the Use of Resonance, The Hydrogen Bond, ‘, , 2 Energies, Lengths, and Orders of Covalent Bonds, Bond Energies, Electronegativity, Bond Lengths and Covalent Radii, Contraction in Bond Lengths with Increasing Polarity, Bond Lengths in Resonance Hybrids, ^Hyperconjugation, van der Waals’ Radii, , 1, 2, 5, 9, 10, 13, 13, , 16, jg, 21, 25, 26, , 34, 34, 39, 42, 43, 45, 49, 50, , 3 Dipole Moments and Spectral Studies, Part I Dipole Moments, Definitions, Polarization, DWectric Constants and Dipole Moments, Bond Moments and Group Moments, Hydrocarbons, Halogen Derivatives, Monosubstituted Benzene Derivatives, Disubstituted Benzene Derivatives, Stene Inhibition of Resonance, Molecular Geometry, Intramolecular Rotation, , 57, 57, 58, 59, 61, 63, 65, 66, 68, 70, 71, 72, V
Page 12 :
Contents, , VI, CHAPTER, , 4, , Part II—Molecular Spectra, , 76, , Absorption of Energy by Molecules, , 77, , Rotational Spectra, , 78, , Vibration Spectra and Characteristic Frequencies, Additional Structural Information, , 79, , Electronic Spectra, , 85, , Acids and Bases. Nucleophiles and Electrophiles, , 93, , Hydrogen Ion Transfer. The Br0nsted-Lowry System, The Leveling Effect, , 93, 96, , Very Weak Bases. Studies in Concentrated Sulfuric Acid, , 5, , pAGE, , 82, , —-, , Quantitative Evaluation of Acidity. Concentrations vs. Activities, Dielectric Constants. Formation of Ion Pairs, Triplets, and Other Aggregates, , 100, 102, , The Use of Indicators in Media of High Acidity. Hammett's ho Function, The Grunwald Acidity Scale, Acid-base Catalysis, , 103, 107, 110, , The Br^nsted Catalysis Law, Electron-pair Transfer. The Lewis Acid-base System, , 113, 115, , Substitution Reaction, , 120, , Methods for Determining Reaction Mechanisms. Part I—, Nonkinetic Methods, , 127, , The Meaning of Reaction Mechanism, Energy Profile Diagrams. Intermediates vs. Transition States, , 127, 128, , Identification of Products, Testing Possible Intermediates, U‘Trapping” of Intermediates, Evidence from Reaction Catalysis, ^Crossover Experiments, /Isotopic Labeling, Stereochemical Studies, Limitations of Reactions, Physical Detection of Intermediates, , 6, , 97, , 131, 134, 137, 138, M3, 130, 152, , Methods for Determining Reaction Mechanisms. Part II—, Kinetic Methods, , M9, , Experimental Methods, First-, Second-, and Third-order Reactions, , M®, Ml, , Determination of the Order of Reactions, , M5, , Reversible Reactions, Consecutive Reactions; The Steady-state Approximation, , M8, , Parallel Reactions, Mechanistic Implications from Rate Laws, The Transition-state Theory. Energy of Activation, Entropy of Activation, Influence of Solvent, Influence of Ionic Strength. Salt Effects, , Ml, , M5
Page 15 :
Contents, CHAPTER, , 14, , 15, , 16, , -, , ix, PAOE, , Hydride-transfer Reactions, Stereochemistry of C—O Additions. Cram's Rule, , 544, 549, , Participation of Neighboring Groups in Nucleophilic Substi¬, tution Reactions and in Rearrangements, , 561, , Intramolecular Displacement by Oxygen, Neighboring Nitrogen, Sulfur, and Halogen, Aryl Participation. The Phenonium Ion, Intimate and Solvent-separated Ion Pairs, Alkyl and Cycloalkyl Participation, Neighboring Hydrogen, Bicyclic Systems, Transannular Rearrangements, The Pinacol and Related Rearrangements, Migratory Aptitude, Neighboring-group Participation in Elimination and Addition, , 563, 570, 575, 580, 584, 591, 594, 599, 601, 607, 610, , Further Molecular Rearrangements, , 618, , The Beckmann Rearrangement, The Hofmann Rearrangement, Reactions of Azides. The Curtius and Schmidt Rearrangements, Rearrangements of Peroxy Derivatives, The Benzilic Acid and Related Rearrangements, Rearrangements of Aldehydes and Ketones in Acid, Rearrangements Proceeding through Carbanions or Related Species, The Claisen Rearrangement and Related Reactions, The Rearrangements of N-Haloanilides and Related Aromatic “Rearrange, ments”, The Benzidine Rearrangement, , 618, 621, 623, 629, 635, 637, 640, 644, , Free-radical Reactions, Part I—Long-lived and Short-lived Free Radicals. Formation an, Detection of Free Radicals, Triarylmethyl Radicals, Further Types of Stable Free Radicals, Detection of Short-lived Free Radicals, The Configurations of Free Radicals, Formation of Free Radicals. Initiators, The Types of Free-radical Reactions and Some Common Characteristics, Part II—Homolyses and Free-radical Displacements, rree-radical Halogenations, Iodine Exchange Reactions, Autoxidations, Thermal Decompositions of Hydroperoxides and DialWl P, Decompositions of Diacyl Peroxides, Y Peroxides, Arylation of Aromatic Rings, Decompositions of Azo and Diazo Compounds, yi- he Sandmeyer Reaction, , 650, 656, , 672, 672, 672, 677, 682, 685, 687, 691, , 695, 695, 703, 705, 710, 714, 720, 724, 729
Page 16 :
X, , Contents, , CHAPTER, , PAGE, , Part III—Additions and Rearrangements of Free Radicals, , 730, , Hoinolytic Additions. Energetic Requirements, Addition of Hydrogen Halides, , 730, 732, , Hoinolytic Halogen Additions, Additions of Dinitrogen Tetroxide, Additions of Thiols, , 736, 739, 741, , Additions of Polyhalomethanes and Aldehydes, , 743, , Homolytic Cyclizations. Diradicals, , •, , 748, , Rearrangements of Free Radicals, , 755, , Author Index, , 771, , 780, , Subject Index, 9
Page 17 :
CHAPTER, , 1, , Atomic and Molecular Structure., The Use of Resonance, , In, , spite of the tremendous advances, , made in organic chemistry between, , 1860 and 1910, many workers in the inorganic and physical chemical fields, forty years ago regarded organic chemists not as scientists, but rather as very, clever artisans (although such opinions were rarely expressed in face-to-face, encounters). Perhaps the chief reason for this was the reluctance of the large, majority of organic chemists of that period to push their structural considera¬, tions beyond anything more fundamental than the carbon skeleton and the, functional groups bonded to it. Then, too, although the course of a new reaction, could presumably be predicted by consideration of the carbon skeletons and, the functional groups of the reagents, it often seemed to the observer that if the, inexplicably “correct” catalyst were omitted or if the correct reaction condi¬, tions were not rigorously maintained, the desired reaction would not occur., (In the latter event, the operator might obtain a product formed in a completely, unexpected molecular rearrangement or, more likely, a dark, intractable tar.), Today, although such intractable tars still arise in the laboratory, the, outlook toward organic chemistry has largely changed. During the last twentyfive years there has arisen an ever-increasing number of organic chemists who,, rather than search for new reactions, have devoted their attention to a close, examination of the older reactions in an attempt to obtain a more intimate, view of what happens between the time that the reagents are mixed and the, product (desired or undesired) is isolated., Such studies are said to have as their objective the determination of the, mechanisms of reactions-ideally, step-by-step descriptions of the paths of each, , of the atoms from start to finish. At present, all but a very small number of, 1
Page 18 :
2, , -, , Atomic and Molecular Structure. The Use of Resonance, , important organic type-reactions have been the subject of mechanistic investi¬, gations. Very often such investigations involve equilibrium or kinetic studies,, formerly considered to be in the province of the physical chemist. Hence, this, phase of chemistry is today known as physical organic chemistry., Since organic reactions are essentially a series of formations and breakings, of bonds between atoms, the development of the scientists’ picture of atoms and, chemical bonds during the last forty years has unavoidably affected the outlook, of the physical organic chemist; for any mechanistic pictures he proposes must, be in harmony with more general structural concepts., A simplified discussion of the modern view of atoms and of the bonds they, form comprises this first chapter., , The Wave-mechanical Picture of the Electron', The schematic sketch of the “solar-system” atom, with a nucleus at its center, and electrons revolving about it in circular or elliptical paths, has become, increasingly familiar during the last decade. Such sketches appear as insignia, for scientific organizations, as commercial trade-marks, and on the pages of, popular periodicals., This type of picture, which shows vividly the division of the atom into the, nucleus and orbital electrons, arose from studies of the spectra of atoms. Such, spectral studies indicated that electrons bound to an atom could have only a, discrete set of energies (that is, that atomic energy levels are quantized) and it, , was at the time natural to suppose that the change of an electron’s energy cor¬, responded to a shift from an outer “orbit” to an inner “orbit,” nearer the, positively charged nucleus, or perhaps in the opposite direction. (Why an elec¬, tron would be allowed in one orbit or another and yet be prohibited from taking, an intermediate position between the orbits was an open question.) It was even, possible to take such a picture and, with classical physical laws, to draw up, equations that described the supposed motion of the electrons in their circular, orbits. Using such simple assumptions, Bohr successfully explained the structure, of the spectrum of atomic hydrogen., Bohr’s treatment, however, could not be applied successfully to most other, atoms without drastic and somewhat unsatisfactory modifications. More im¬, portant, we are aware today that the laws of classical physics (which Bohr used, 1 For more detailed, but still largely qualitative, descriptions of the wave-mechanical, picture of bound electrons see: (a) Coulson, Valence, Oxford University Press, Oxford, 1952,, pp. 1-42; (b) Cartmell and Fowles, Valency and Molecular Structure, Butterworth’s Scientific, Publications, London, 1956, pp. 1-63; (c) Syrkin and Dyatkina, Structure of Molecules and, the Chemical Bond, Interscience Publishers, New York, 1950, pp. 1-41. More quantitative, treatments are given in a number of textbooks on quantum mechanics; see, for example,, (d) Pitzer, Quantum Chemistry, Prcntice-Hall, Inc., New York, 1953, pp. 127-188.
Page 19 :
The Wave-mechanical Picture of the Electron, , -, , 3, , to set up his equations) cannot be used to describe the behavior of very small, objects. The ordinary laws of mechanics and electrostatics are adequate for, phenomena in our everyday world, but they break down if they are applied, without substantial changes to nuclei and to electrons. This does not mean that, there are actually two sets of physical laws, one for the macro world and the, other for the ultramicro world. Rather this apparent “duality” arises because, the laws of classical physics turn out to be only approximations. For large ob¬, jects, however, they are excellent approximations; the inaccuracies in such, approximations become noticeable only when very small objects are being, investigated., For instance, the Heisenberg uncertainty principle gives us no trouble, in everyday physics. This principle says (among other things) that it is impos¬, sible to determine accurately both the position and the momentum (hence,, the kinetic energy) of a given object. It maintains further that the more accu¬, rately we know the energy, the less accurately we may know the position (and, vice versa). We do not concern ourselves with the principle when describing, an automobile, for we may specify both the position of the car at a particular, instant and its velocity (which allows us to calculate its kinetic energy). Nature,, however, opposes our making similarly specific statements about an electron;, for to determine the position of an electron, the electron must be observed., This means that it must be seen through a very powerful supermicroscope or it, must cause some variation in a test signal which we send out. Essentially all, that we see becomes visible by causing variation in a pattern of light rays, emerging from a source of illumination. Now, we cannot see an object that is, smaller than the wavelength of the light used for illumination, and to determine, accurately the position of an electron, we would have to use light having a very, short wavelength indeed. If radiation of too long a wavelength were used, the, image of the electron would become fuzzy, and the longer the wavelength,, the fuzzier would be the image., Very short wavelength radiation, however, consists of high-energy photons., Although it is generally assumed that we cannot move objects simply by illu¬, minating them, an electron can be moved by collision with a photon. Such a, photon may give up a little or much of its energy in such a collision; and the, higher the energy of the photon, the greater would be the uncertainty in the, energy of the electron after the impact. We see then that in trying to fix the, electron’s position very exactly by using photons of very short wavelength, we become very uncertain as to the energy of ,he electron since it would be, , knocked, , unknown direction with a very energetic photon. Philosophically,, , he Heisenberg prmc.ple anses because of the scientific necessity for the observathc character 'of pTnfcles 3 m'anS * °bSerVati°n radiation which, , displays
Page 20 :
4, , Atomic and Molecular Structure. The Use of Resonance, From what has been said, it follows that the classical picture of the electron, , as a compact mass in circular or elliptical motion about a heavy nucleus has, little significance since the exact fixing of the position of the electron would, give it a tremendous energy uncertainty. Because of the uncertainty principle,, the desciiption of the size and motion of electrons in terms of everyday physics, only cannot be given. Rather, a new system of mechanics, quantum mechanics,, yields a more satisfactory picture of the behavior of very small objects., In quantum-mechanical descriptions, the ordinary laws of motion are, replaced by equations representing probabilities. As a very crude example,, suppose we weie to know that if a particular electron had a certain energy—, a certain momentum—and was spinning in a certain direction, the probability, that it would be in a region a given distance from the nucleus would be “1 in 2”, or 0.5. The probability that the same electron would be in a region a little, further from the nucleus might be “1 in 5” or 0.2. Suppose, however, that a, second electron has a different energy or momentum; the probability that the, second electron might be in a region a given distance from the nucleus might, be “1 in 3,” or 0.333. The quantum-mechanical treatment of the electron seeks, to relate mathematically the energy, spin, momentum, and other character¬, istics of the electron to the probability of finding the electron at any distance, from the nucleus (or from any other point of reference). However, these char¬, acteristics will never determine with certainty just where the electron lies; all, that may be stated is the probability of finding the electron in a particular, region as compared to the probability for a different region. The equations, thus describe the way that the probability of finding an electron in a given region, varies as we look for it farther and farther from the nucleus. The electron is, more likely to be found in a “high-probability region” than in a “low-proba¬, bility region.” Indeed, it is possible to imagine the electron as a very small body, in rapid motion, darting back and forth over a large area, spending most of its, time in the high-probability regions. A more easily visualized picture of the, electron is that of a smear or cloud of negative charge, thickest in the regions of, high probability, more diffuse in the regions of low probability., We shall not try to describe rigorously the manner in which equations, dealing with such probabilities are set up but will indicate some of the results, of greatest interest to chemists. First, we may note that the most fundamental, equation used in describing bound electrons was set up because its originator,, E. Schroedinger, felt that the equation treating the “position probabilities ’ of, a bound electron was similar in nature to the type that describes the motion of a, point on a vibrating string. In order to get useful information from this very, general Schroedinger wave equation, it first becomes necessary to express, the energy of the electron in terms of its charge, its mass, and its other quan¬, tities. This general equation must then be simplified using whatever mathe-
Page 21 :
The Four Quantum Numbers. The Pauli Exclusion Principle, , 5, , matical devices are appropriate. Ultimately, one hopes to obtain a more specific, equation pertaining to the atom at hand, relating the “position probability, of an electron at a given point with the position of that point, and allowing us, to estimate the energy of the electron., Now the Schroedinger equation is a differential equation and, like other dif¬, ferential equations, it has a very large number of possible solutions. A number, of such solutions do not correspond to electrons as we know them. Some solutions, picture infinite electron densities at some positions, others allow two or more, position probabilities at a given position, still others allow sudden “blotches”, of electron density to appear in regions where the cloud is otherwise very thin., All such solutions must be described as being inconsistent with our experimental, picture of the electron. To obtain meaningful solutions of the Schroedinger, equation for a simple atom and, at the same time, a solution consistent with, observed atomic spectra, it is necessary to introduce two additional quantities, which, because of the wavelike nature of the electron, must be integers. (These, are the first two of the four quantum numbers described in the next section.), In one of the more useful solutions of the wave equation, a simpler equation, is obtained, expressing the energy of the electron in terms of these two “new”, quantities, along with the charge and mass of the electron, the charge on the, nucleus, and fundamental constants., , The Four Quantum Numbers. The Pauli Exclusion Principle, The two quantities thus arising are termed quantum numbers. Although their, values may vary, they must follow definite rules in such variation. The energy, of the electron is related to the values assumed by these quantum numbers, and, if we change the value of one or both of these quantum numbers, the energy of, the electron would be expected to change. Furthermore, the character of the, spectra of atoms tells us that a bound electron may have only certain energiesit then must follow that the numbers determining the energies may have only, limited combinations of possible values., A more complete description of an electron cloud involves two additional, quantum numbers which do not however generally affect the energy of an, , beTa™ clued!, n, , ^, , ma8"etiC fidds)' The f°Ur <lUantum, , the principal quantum number, , /-the subsidiary (or azimuthal) quantum number, the magnetic quantum number (sometimes abbreviated m), ., , the spin quantum number (sometimes abbreviated s), , From the way in which these numbers arise in ,k„ k, j,uicers arise in the handling of wave equations
Page 22 :
6, , •, , Atomic and Molecular Structure. The Use of Resonance, , and by analysis of atomic spectra, physicists are able to associate each of these, numbers with the characteristics of the electrons they describe., n, the principal quantum number is a rough measure of the size of the, , electron cloud. The larger the value of n, the greater is the volume of, the bulk of charge density., /, the azimuthal quantum number, is related to the shape of the electron, cloud. The value of l indicates whether the cloud is spherical, dumbbell¬, shaped, or perhaps more complicated., The magnetic quantum number, m is related to the orientation of the, electron cloud in space., Finally, the spin quantum number, j, indicates the direction of spin of the, electron. Experiments using an inhomogeneous magnetic field show that, electrons have a property that is roughly analogous to our everyday, concept of spin. (Although this language is inexact, it is quite suitable, for our present purposes.) Furthermore, this spinlike property is also, quantized; that is, a bound electron may “spin” only in two ways., Now the Schroedinger equation has not been solved for all 102 elements., Solving the equation for atoms or ions having more than one electron is at, present thought to be an impossible task, and even approximate treatments, are very difficult. Such approximate solutions as have been worked out, how¬, ever, indicate that we may describe the electrons in the heavier atoms in the, same language as that used to describe the energy levels in single-electron sys¬, tems. In fact, descriptions of the spectra of the elements are often given in the, same quantum terms that are used for the hydrogen atom., In the most common treatment of the wave equation, the restrictions on, the values of the four quantum numbers are as follows:, n—positive integers only; 1, 2, 3 . . . (but not 0), l—positive integers less than n; 0, 1, 2 . . . . (n — 1), m—0, +1, —1, +2, —2, etc.; up to +/ and —/, s—two values only;, , and —, , Note that n may assume any integral value. For electrons of atoms in unexcited, states, n takes values from 1 to 7, corresponding roughly to the seven horizontal, rows of the periodic table. For the atoms in which the organic chemist is most, interested, n is 1, 2, or 3., The values assumed by the azimuthal quantum number, /, depend on the, values of n. For elements up to calcium, / assumes only the values 0 (correspond¬, ing to a spherical electron cloud) and 1 (corresponding to a dumbbell-shaped, electron cloud). Certain electrons of the atoms of the heavier elements have
Page 23 :
The Four Quantum Numbers. The Pauli Exclusion Principle, / = 2; such electrons may be imagined as a quartet of sausages with ends, fastened to a point in space (Fig. 1-1)., The third quantum number, m, describes the orientation of the electron, cloud in space. For a spherical cloud, however, there is only one possible, orientation, for no matter how a sphere is twisted, it still looks the same to an, outsider. A dumbbell, on the other hand, may be lined up with one end toward, the observer or with its “handle” parallel to the height of the observer or with, its “handle” parallel to the line connecting the ears of the observer. There are,, in addition to these three orientations of a dumbbell, an infinite number of, intermediate positions with the “handle” at various inclinations. For an elec¬, tron, however, only three of these orientations are possible, meaning that in all, such experiments which attempt to investigate electron orientation, the orienta¬, tion of the electron cloud becomes quantized. Such experiments generally, , n—1, , 72 2, , 72=, , 3, , 1=1, , 1, , 2, , =, , /=0, , 1= 0, , =, , Fig. 1-1. Electron Clouds and Quantum Numbers, , make use of an externally applied magnetic field, and each of the possible, values of m relate to a given position of the electron cloud in space with respect, to such a magnetic field., The final quantum number, r, representing the “spin” direction, assumes, only two values. When this “spin” exhibits itself in the presence of an external, inhomogeneous magnetic field, it is found that electrons may “spin” either in, one direction or in the direction roughly opposite. The spin quantum number, may assume only two values, +y2 or ~y2, but to avoid concerning ourselves, either a" +or-“phT" fraCU°naI Va‘UeS> We sha11 refer to electrons as having, , ■■, , S"":;, , i! determine) by „ and t, and, g.nbrally ,,a,kiriy' ,t. ' ""crestcd' ,he e”«SV, , tu, , describe electrons of lower enertrv than H ,t,, qUantum numbers, electron with n = 1 and / - o, , ,, ° / lgher ‘>uantum numbers. (An, 1, 1 ana 1 — b would have less pnprmt, i, " = 3 and l = 1.) However this H, 8Y h, 3n electron with, , ,he 102 dem-ts in ^
Page 24 :
8, , -, , Atomic and Molecular Structure. The Use of Resonance, , values for these quantum numbers). Another factor, the Pauli exclusion, principle, must be considered when electrons are gathered around a nucleus., This principle applies to many problems, but for the purpose at hand we may, state the principle as,, , No two electrons bound to the same atom may have, , identical sets of four quantum numbers”; this statement is very roughly anal¬, ogous to the classical law, “Two bodies cannot be in the same place at the, same time.” Thus, two electrons of a given atom may have the same values for, three of their quantum numbers, but the values of the fourth quantum number, must be different., The one electron of the hydrogen atom in the unexcited state (or ground, state), , will have the lowest possible values for the quantum numbers n and l—, , that is, n = 1 and / = 0. Since m may not exceed /, m must also be 0. In helium, (atomic number 2) there are two electrons, one with the same set of quantum, numbers as the single electron of the hydrogen atom; the second helium elec¬, tron will have the same values for n, /, and m, but s will be different. If we, arbitrarily call ^ for the first electron —, then ^ for the second electron would, be +. Thus in helium there are two electrons situated around the nucleus,, appearing from the outside as a spherically symmetric electric cloud. The two, electrons are identical in all respects except that they are spinning in opposite, directions. Removal of an electron from the helium atom is much more difficult, than removal of an electron from the hydrogen atom, due chiefly to the added, positive nuclear charge (an extra proton). In helium there are thus two spheri¬, cally symmetric electrons with spins in opposite directions (that is, with spins, paired)., , With lithium, atomic number 3, there are two electrons in a spherical, cloud (as with helium) plus a third electron. Generally, in considering the elec¬, tronic configurations of an element, the “last” electron is the only one that, need be considered since each of the remaining electrons is present in the atom, preceding it in the periodic table. (The exceptions to this statement are not of, importance to the organic chemist.) For the third lithium electron there are no, more combinations of quantum numbers where n = 1, since neither / nor m, can exceed zero if n is unity. The last (outermost) electron in lithium therefore, has the quantum numbers:, n, , l, , m, , s, , 2, , 0, , 0, , -, , Here, again, the value of s is arbitrary since the two spin states are energetically, the same. Note that the additional electron has assumed a value of 2 for n., This last electron has the bulk of its charge density much farther from the nu¬, cleus than the first two and thus is held much more loosely than the other two., This electron is easily removed chemically, enabling lithium to assume a valence, of +1.
Page 25 :
"Build-up" of the Periodic Tables; s and p Electrons, , 9, , “Build-up” of the Periodic Tables; s and p Electrons, The last electrons in beryllium and in boron (atomic numbers 4 and 5, respec¬, tively) have quantum numbers:, , last electron in Be (Z = 4), B, , (Z = 5), , n, , l, , rn, , s, , 2, , 0, , 2, , 1-1-, , 0, , +, , Note that the fourth electron in beryllium “pairs up’’ with the third, just as the, second electron in helium “pairs up” with the first. However, the outer electron, in boron is the first unexcited electron in our “build-up” of the elements for, which / is unity. This final electron cloud is thus dumbbell-shaped, not spherical, as was the case for the preceding electrons. It is emphasized that the assignment, of the value of — 1 to m for this outer electron is completely arbitrary, just as, was the choice of — values of s before -f- values. The three orientations of the, dumbbell-shaped electron cloud are, in the absence of external magnetic fields,, energetically equivalent., Bearing in mind that the order of assignment of the values for m and s is, arbitrary, let us now summarize the four quantum numbers for the last electron, in each of the elements up to sodium (atomic number 11):, n, , /, , m, , 2, , 1, , -1, , —, , 1, , 0, , —, , 1, , +1, , —, , 1, , -1, , +, , 1, , 0, , +, , Ne (Z = 10), , 2, 2, 2, 2, 2, , 1, , +1, , +, , Na (Z = 11), , 3, , 0, , 0, , last electron in B (Z = 5), C (Z = 6), N (Z = 7), O (Z = 8), , F (Z = 9), , s, , —, , incoming electrons have taken all possible vali, pairing up, the same region in space and therefore repel each other more strongly than do, c ectrons of different m; therefore the “pairing up” process is to be postponed, as long as possible. The last three electrons in the nitrogen atom may be redireclns-th't, , !'lhaPed ^ ^ thd'', , Pointing in different, , elect,h, r™ m °XySen emerS °ne °f the clouds occupied by an, electron, whereupon the two electrons occupying the same cloud adopt different, spm directions. This “doubling up” or coupling raust obviously occur s nce Ze, are no further values that „ may assume if a and / remain the same, A group of bound electrons with the same value of « :
Page 26 :
10, , -, , Atomic and Molecular Structure. The Use of Resonance, , are said to occupy the same orbital. Thus in the neon atom, there are two filled, shells, the first with only two electrons (n = 1), the second with eight electrons, (n = 2). In the second shell of neon there are two subshells, the first having, two, the second having six electrons., After the tenth electron goes into neon (Z = 10), there are no more possi¬, ble combinations of quantum numbers having n = 2. The outermost electron, of the next element, sodium, must go into a higher energy state with n = 3,, much farther from the nucleus than the other ten. Thus, the outermost electron, of sodium, like that of lithium, is easily lost. More generally, since the properties, of elements are determined in a large measure by their outermost (valence), electrons, the periodicity of properties arises naturally from the quantum re¬, strictions. The final electrons for the elements in the second eight-membered, period (sodium to argon) have quantum numbers corresponding to those listed, for the preceding period with, however, the exception that n is 3 rather than 2., In describing the ten electrons of the neon atom, chemists often refer to, the two inner electrons (for which n = 1 and / = 0) as h electrons. The next, subshell (for which n = 2 and / = 0) is called the 2s subshell; and the next six, electrons (for which n = 2 and / = 1) comprise the 2p subshell. In this short¬, hand system, the initial number refers to the value of n and the small letter, refers to the value of /., Value of /, , Symbol, , 0, , s, , 1, , P, , 2, , d, , 3, , /, , (These letters are relics from the old designations that referred to certain spectral, lines as “sharp,” “principal,” “diffuse,” and “fundamental.”) The organic, chemist is interested chiefly in j- and />-type electrons. The d and / electrons, become very important when the transition metals and the rare earths are, being considered., , Ionic and Other Electrostatic Bonds versus Covalency, In attempting to answer the complex question, “What makes atoms come to¬, gether to form molecules?” the concept of ionic attraction is exceedingly useful., One has little hesitation in saying that positive and negative ions are held to¬, gether in pairs or groups because “unlike charges attract.” Thus, although the, LiCl molecule may not be dissociated into ions in the vapor state, it is still, reasonable to say that its stability results from interaction of the positive Li+, ion and the negative Cl~ ion. Electrostatic interpretations are often extended
Page 27 :
Ionic and Other Electrostatic Bonds versus Covalency, , -, , 11, , to account for other interactions, even if one or another of the reacting species, is not actually charged. We regard many molecules as having “positive ends”, and “negative ends” and refer to such molecules as dipoles. It is sometimes, convenient to designate positive or negative parts of molecules with the symbols, 5+ or 5 — , thus indicating positions of partial positive or negative charge; for, example:, 8+8-, , H3C, and, , 5+, , Li-Cl, , ^08—, H3C, , Dipoles may interact, sometimes with ions, sometimes with other dipoles. The, methylamine complex of silver ion is conveniently regarded as being held, together by attraction between positive Ag+ and the negative end of the amine, molecule:, , +, Ag, , .... 8, , /H, CH3 5+ forms, , /, Ag, , H, , —i +, , N —CH3, , H, , H, , Similarly, interaction of acetic acid with ammonia may be regarded as attraction, between the positive end of the water molecule and the negative end of the, ammonia molecule. Presumably, , 8-, , 5+, , ,0—H ., , o=c', , <5/H, N v H<5 +, \—, H, , /, , •, gives, , /O—H—N—H, _, /, \, 0=C., XH, , \, , CH-, , CH,, , bind )ab°Ve attracti°n’ as we Sha11 see’ is an “ample of the familiar hydrogen, Even though the simplicity of the electrostatic picture makes it attractive, it is obvtous that the p.cture cannot be extended to account for bonds between, , wo a,oms of he, , elemen( (for example> be(ween (he atQms ^, , «wee„, , or between the carbon atoms in C2H6). More generally it becomes difficu t to, assume tmpor,ant coulombic attraction between atoms of elements close to, each other in the periodic table., close to, When G. N. Lewis introduced the elert™«, , », , j, , ^ " *-—- ^S‘ZTsZ S5:f, Lewis, J. Am. Chem. Soc., 38, 762, , Atoms and Molecules, Chemical Catalog, , Co., New, , c, , York, , Valence and the Structure of
Page 28 :
12, , -, , Atomic and Molecular Structure. The Use of Resonance, , having eight valence electrons had already been recognized, it was reasonable, that the bond between atoms would be most stable if each atom was associated, with eight electrons, either by sole “ownership” or by sharing. Interpretations, using the electron-pair bond and the so-called rule of eight yielded a surprisingly, good correlation of the chemical properties of a great number of compounds., The majority of compounds of nonmetals in the first two (eight-membered), periods can be formulated using the Lewis rules (although not without some, serious objections). The arrangements of atoms in some polyatomic ions (such, as NH^, SOj, AlHj, and BF^) can also be pictured., In considering compounds containing carbon, oxygen, nitrogen, and (per¬, haps) sulfur, chemists had long been aware that more than one type of bond, could link two given atoms. The bond between the two carbons in ethane is, different from the bond between two carbons in ethylene; similarly, the bond, between carbon and nitrogen in methylamine is different from the bond be¬, tween carbon and nitrogen in acetonitrile. Such differences were readily ap¬, parent from the chemistries of the compounds and were later to be confirmed, by measurements of bond lengths and bond energies. Lewis represented multiple, {double and triple bonds) as quartets and sextets of electrons, arranged (perhaps, for typographical convenience) in adjacent groups of two., H, , H, , H, ,’C:, , :c;, , H: C: C, H, , H, H, ethylene, , acetonitrile, , Such pictures tend to suggest that the two bonds comprising a double bond are, equivalent and that the three bonds comprising a triple bond are also equivalent., Although the question cannot be regarded as settled, there is at present con¬, siderable evidence indicating that this is not so., In spite of the successes of the Lewis picture, it was obvious (even before, the development of the more recent pictures of covalent bonding) that import¬, ant exceptions existed. Stable compounds having odd numbers of electrons (for, example, NO, N02, and C102) had long been familiar. Certain compounds of, boron and aluminum had too few electrons to form the necessary number of, covalent bonds (for example, B4Hi0 and A12(CH3)g)- Other compounds arc, known which have more electrons than can be accommodated by octets around, each of the atoms (for example, SF6 and PCla)., Finally, there are an uncomfortably large number of cases where more than, one permissible Lewis structure can be drawn for a compound but where only, one compound exists (for example, N20,, , CH3N3,, , and o-xylene); in cases of this, , sort, the resonance concept (to be described shortly) becomes important.
Page 29 :
Recent Interpretations of Covalent Bonding, , -, , 13, , Formal Charge, This term is used less frequently but occasionally becomes important in discus¬, sions of the structures of a number of molecules having multiple bonds. To find, the formal charge of a given atom in a structure: (a) Start with the group num¬, ber of the atom in the periodic table, (b) subtract the number of unshared, electrons about the atom, and (c) subtract one half the number of shared elec¬, trons. In the structures for diazomethane (CH2N2) below, the formal charges, have been indicated:, H, , H, :N: :N:, , H: C: N:, .. +, , (A), Formal charge, , ~ (B), , may be loosely defined as the measure of excess charge about, , a bound atom. A bond such as the N—O bond in trimethylamine oxide,, -f", , • •, , —, , (CH3)3N:0: , having unlike formal charges on adjacent atoms is called by, some workers a coordinate link or semipolar double bond. The latter term, was introduced at a time when it was generally felt that the distinction between, ionic and covalent bonding was a sharp one. (Thus, it was believed that if, charges could be assigned to two atoms bound together by an electron pair,, then two bonds were present, one ionic and one covalent.) However, we now, recognize that almost all covalent linkages between different elements have, some ionic character (whether or not formal-charged structures can be drawn), and that therefore the “coordinate link” is not different in kind from many, other covalent bonds between unlike elements., , Recent Interpretations of Covalent Bonding3, fraction6" 77 7, C°Valen‘ b°"d k not due t0 °rdi"ary electrostatic atct.on, just what keeps the hound atoms together?” Such a question is some, , ;r”d, h t h, , h, , 7 descnl>ing the covalent b°nd as electrostatic in nature but, , SUUa,ed ^ 3 “neW” manner- °“ «-'d argue, for ins Lee, C°nSiStS f tW°, , near the middle, , 7th, , eharged protons held on, , n^T^ “77*'’ WhiCh " ~d, , arrangement should be more stable th,, , *° T, , Why SUch, , atoms, each with its own electron The shuat^8'6™ °f -W° IS°'ated M'Ogen, JF, ‘, station is certainly more complicated, , Modules, Prentice-Hall, Inc,, , New York, 1952 pp., , “79, , **■» and Dyatkina, gUS°n' EU“™ic, of Organic
Page 30 :
14, , •, , Atomic and Molecular Structure. The Use of Resonance, , than that in which a single electron is attracted to a singly charged, , nucleus., , In spite of the fact that the H2 molecule is one of the simplest molecules used in, the chemist’s laboratory, it nevertheless contains four bodies—two electrons, (diffuse and light), and two protons (tiny and massive). There are attractions, between each proton and each of the two electrons, a repulsion between the, two protons, and repulsion between the two electrons. Without knowing the, answer beforehand, it would be a very wise man who could guess the stable, layout of this four-bodied system and also predict how the probability densities, of each of the two electrons would depend on the geometry of the system., These are problems in quantum mechanics. As with single atoms, the ap¬, proach is made through the very general Schroedinger wave equation, which,, it will be remembered, treats the probability density functions of electrons in a, manner mathematically analogous to the motion of a point on a vibrating, string. The energy of the system may be expressed algebraically in terms of the, distances between bodies, and if the initial equation can be handled compe¬, tently, three desired items of information may be obtained. These items are:, (a) The relationship between electron density at a point and the position, of that point;, (b) The distance between the nuclei in the most stable layout of the system;, (c) An estimate of just how much more stable the bonded system is than, the isolated atoms (that is, an estimate of the bond energy)., For the isolated hydrogen atom, the energy depends on one coulombic, interaction between two bodies. For the hydrogen molecule, the energy depends, on six interactions among four bodies, a much more formidable system. In the, wave-mechanical treatment of molecules, more is involved than the ability to, select and use mathematical tricks. The problem is so complicated that it must, first be broken down into simpler problems, so that each of these may be, handled individually. Then the parts must be brought together in a rational, way, and, most important, the dimensions and energy of the molecule which, have been calculated must be checked against the dimensions and energy deter¬, mined, , by experiment. Such a “breakdown” of the problem requires both, , chemical and quantum-mechanical experience, and care must be taken that, in, the breakdown and subsequent recombination, an error does not slip in due to, some subtle limitation of the working method. There are indeed a number of, chemical physicists who have spent years devising methods of attack on molec¬, ular-wave equations. For many simple molecules, the quantum-mechanical, picture is not greatly different from the Lewis picture. For H2, theoretical, treatment predicts a cloud of two electrons (spins paired), relatively thick, between the two nuclei; the F, molecule, a much more complicated system,, can also be treated (making use of a number of simplifying approximations),
Page 31 :
Recent Interpretations of Covalent Bonding, , -, , 15, , arriving ultimately at a picture qualitatively similar to that of H2. For F2,, although there are fourteen valence electrons, only two have appreciable, probability densities on the line joining the two nuclei. Such a description, conforms to the Lewis single bond in elemental fluorine., The single bonds in H2 and in F2 are referred to as a bonds—that is,, electron clouds shaped much like elongated eggs with both nuclei inside, one, near each end (Fig. 1-2). The Greek designation a is chosen in analogy with the, Roman s, for, as may be visualized from the sketch, if the bound atoms were to, approach each other until the nuclei coalesced, the electron cloud would, assume the spherical symmetry typical of s electrons. Also shown in the sketch, is the so-called it bond, occurring in compounds whose Lewis structures include, double or triple bonds. If the nuclei bonded by a ir bond are considered to, come together, the electron cloud shrinks into the typical dumbbell-shaped, p electron., , Fig. 1-2. cr and, , tt, , Bonds, , However, whether we are considering a bonds or tt bonds, it can be said that, the wave nature of the electron is such that when two bound electron clouds, approach to form a bond, there will occur a thickening of the cloud in the, regions where overlap is greatest. What is not easy to visualize without some, physical insight is that wave interaction may allow a greater thickening than, can be accounted for by mere addition of the thicknesses of the separate clouds., Valence electrons that formerly were associated with just one of the atoms be¬, come associated with the entire system., Such a change in the configuration of an electron cloud may be called “delocalization.” Extensive delocalization is not favored for an electron bound to, a single atom because of the considerable coulombic energy required to move the, negative charge appreciable distances from the nucleus. Delocalization becomes, possible or polynuclear systems because there is more than one center of positive, covlnt bond, , “ 15 dd0Ca,lZati0n that, , *r the stability of the, , A covalent bond may be formed between two atoms only if the valence, , paired val, ' f, ‘W° a'°mS °Verlap' In Particular> an atom having un¬, paid valence electrons would be expected to form covalent linkages if overlap
Page 32 :
16, , -, , Atomic and Molecular Structure. The Use of Resonance, , with the orbital of an unpaired electron of another atom is possible. Such a, process would occur with the pairing of spins. Knowing the layout of the elec¬, tron clouds, it is sometimes possible to predict the type of bonds that will form, and the layout of the resulting molecule. In the simplest case, we have seen that, two hydrogen atoms will combine by overlap of their single, , s, , electrons if the, , electrons are spinning in opposite directions. Similarly, hydrogen can form a, single bond with chlorine by pairing its j electron with the one unpaired, , p, , electron of the chlorine atom., The oxygen atom, as we have pointed out, contains two unpaired, trons; the nitrogen atom has three unpaired, , p, , electrons. Since the, , p, , p, , elec¬, , electrons, , are dumbell-shaped and oriented at right angles, one might expect the bonds in, H20 and NH3 to be oriented at right angles as shown in Figure 1-3. Spectral, studies show that the actual H—O—H bond angle in water is 104°, whereas, H, , Fig., , H, , 1-3. Simplified Valence-bond Pictures of Water and, Ammonia, , the H—N—H bond angles in ammonia are 106V Although these values are, not far from the predicted 90°, they are also suspiciously close to 109 . The, latter figure is the bond angle in methane, CH4, for which the bonding must, be described differently. It is sometimes suggested that electrostatic iepulsions, between the positive hydrogen nuclei distort the bond angles in water and, ammonia from the “expected” 90V (However it appears at present that the, attractively simple pictures of these two compounds with bonds lying along the, , p, , orbitals may not be applicable without substantial modification.), , Promotion, Hybridization, and the Tetrahedral Carbon Atom, In almost all of its covalent compounds, the element carbon forms four bonds^, However, the carbon atom in its normal state (p. 9) has only, i Dennison, Revs. Mod. Phys., 12, 175 (1940), , 95°., , two, , unpaire
Page 33 :
Promotion, Hybridization, and the Tetrahedral Carbon Atom, , -, , 17, , electrons. By a process of reasoning similar to that in the preceding paragraph,, carbon would be expected to form only two covalent bonds. Confronted with, this inconsistency, one looks for ways in which the electrons around the carbon, atoms can be shuffled to yield four unpaired electrons. Suppose, for instance,, that the outside “j” subshell were broken up and one of the electrons were, “promoted” to the p level. The resulting atom, would have four outer electrons, with quantum numbers as tabulated below. However, this is an excited state,, and the energy required for the “promotion” of “electron 4” must be compen-, , Table 1-1. Quantum Numbers of the Valence Electrons in Carbon, “excited”, , “normal”, Electron Number, n, , l, , s, , n, , l, , 3, , 2, , 0, , 0, , —, , 2, , 0, , 0, , -, , 4 (to be “promoted”), 5, 6, , 2, 2, , 0, 1, , 0, , +, , 2, , 1, , -1, , —, , 2, , 1, , +1, 0, , -, , —, , 2, , 1, , 0, , 2, , 1, , -1, , -, , m, , m, , s, , sated for by the energy released when two extra bonds are formed. Often the, energy released in forming strong covalent bonds is sufficient to allow promo¬, tion, providing electrons are not promoted to levels beyond those characteristic, of the next rare gas., , Fig. 1-4. The Tetrahedral Carbon Atom, , Slnce '^e three /.-electron clouds of an unbound atom are dumbbell-shaped, " JT*, be oriented, , 17, , nfd ahS‘eSt,and SinCC ‘he '-e'eCtr0n d°Ud " ‘P^caUy ^ped,, H, * ^ f°Ur b°nd>, “-valent carbon would, , It is wetk‘ r‘Sh, , 7, , T, , th3t 'he f°Urth WOuld be oriented a. random., , corleTs lf a, ,, 7 1, thC f0Ur b°nds f0™ed ^ earbon point to the, ., rC*U ar tetrahedron (that is, a pyramid having four faces each of, which is an equilateral triangle Fig t 41 t„, ,u, S, ’, h °f, clouds of the Carbon at, ’ , g‘ _4^' ThuS’ the contours of the electron, carbon atom have been substantially changed by the approach
Page 34 :
18, , -, , Atomic and Molecular Structure. The Use of Resonance, , of four additional groups to form bonds. This change is not unexpected, for, there is no good reason to suppose that the electron clouds whose shapes are, the result of wave interaction in the field of a single nucleus should retain their, shapes when brought into the vicinity of additional positive centers. Wave, mechanics must again decide what new shapes the electron clouds may assume, to allow formation of the most stable bonds (that is, those for which overlap, with electron clouds from neighboring atoms is greatest). A rigorous treatment, of the problem is quite difficult, but several approximate treatments6 confirm, that the most stable “electronic layout” for four-bonded carbon is that of four, similar bonds pointing to the corners of a regular tetrahedron. The carbon atom, is at the center of gravity of the tetrahedron (three-fourths the distance from a, corner to the midpoint of the opposite triangular face). This corresponds (Ex. 7), to a H—C—H bond angle of 109.5°., By “reshuffling” the valence electrons in the lone carbon atom, we have, permitted the formation of four bonds, but the directions of the four bonds do, not correspond to the directions of the four original orbitals. The quantummechanical treatment is said to hybridize these electrons (that is, reorient the, shape of the electron-cloud picture without increasing the number of bonds)., The bonds formed by carbon are said to be spz hybrid bonds, the bonding pre¬, sumably made possible by vacancies in the s- and ^?-type subshells. Such hy¬, bridization is possible using other sets of orbitals; in particular, for compounds, of the transition metals, hybridization involving d orbitals becomes important., It has been suggested that the departures of the bond angles in water and, ammonia from the “expected” 90° are greater than can be accounted for by, mere electrostatic repulsion of the hydrogen atoms,5 and that the observed bond, angles in H20 and NH3 are better explained by the presence of hybridized, bonds than by simple p bonds that have been distorted. Ultimately, however,, hybridization, like any wave-mechanical treatment of electron systems, must, be based on an energy expression, itself electrostatic in nature, and this differ¬, ence in interpretation need not worry the organic chemist., , Double and Triple Bonds, Single covalent bonds (<r bonds) consist of electrons whose densities lie mainly, between one pair of atoms. Double bonds present a much different picture. In, addition to the <r bond (as is present in single-bonded compounds), there is, bonding interaction between the “dumbbell” electrons on adjacent atoms,, forming a rr bond, or, in the case of triple-bonded atoms, two it bonds. The bond' See, for example, (a, Pauling A,, 481 (1931). For a summary of such a treatment, see ( ), Bond, Cornell University Press, Ithaca, 1948, pp. 84-85., , 37
Page 35 :
Double and Triple Bonds, , 19, , ing in ethylene and in acetylene is often represented as in Figure 1-5. To sim¬, plify the picture, the a bonds are represented only as straight lines so that the, 7r, , bonds are more easily seen. Two drawings of ethylene appear in Figure 1-5., , The first attempts to show overlap between neighboring “dumbbell” electrons., The second picture is simply “stretched out” so that the portions of the molecule, may be more clearly labeled (although the interaction between the tt electrons, then becomes less clearly pictured)., , (A) 7r-bond picture, , H, H, , In acetylene, two pairs of 7r-electron, , (B) “streamer” picture, , c=c;, FT, , (C) benzene, , H, H, ;C=C:, H, , classical lormula, , Kelcule formulae, , of butadiene, , of benzene, , Fig. 1-6. Bonds in Butadiene and Benzene, , Jobes are shown at right angles by use of dotted lines. However, wave mechanics, , ^£=;, rie -, , as a group over a large section of the molecule. Figure 1-6 showsThel-Tect™
Page 36 :
20, , Atomic and Molecular Structure. The Use of Resonance, , structures for 1,3-butadiene and benzene. Again, the single (<r) bonds are, represented merely by lines., In this figure, the 7r-bond picture of butadiene (A) fails to show the delocalization of the 7r electrons. Picture B, the “streamer55 picture of butadiene,, shows the delocalization more clearly but is more difficult to draw (all C—H, bonds have been deliberately omitted). The “streamers” indicate that none of, the four electrons belongs to any particular carbon but that a composite cloud, is spread over the four-carbon chain. Benzene is shown as a 7r-bond picture, only, but a similar streamer picture may be drawn for it also, showing hexagonal, “streamers” of electric charge above and below the plane of the ring., The classical formulas of the two hydrocarbons are pictured below the, more “modern” formulas. In contrast to the Lewis picture, which puts a single, bond between the two center carbons in butadiene, the more recent pictures, show that the middle bond actually has “double-bond character,” as do the, end bonds. Moreover, it is to be emphasized that the 7r-bond picture of benzene, can correspond to either of the two “Kekule” forms but is “better” than either, alone because it does not convey the erroneous impression that three of the six, carbon-to-carbon bonds are single bonds., For delocalization of 7r-electron density along the length of a chain or the, circumference of a ring, it is necessary that each atom comprising the chain or, ring have associated with it at least one p-type electron that is not being em¬, ployed in formation of single bonds. (The classical chemist would say that each, atom in such a chain or ring should have some, , I, , ‘unsaturated character. ), , II, , III, , Fig. 1-7. Delocalized and Localized 7r-electron Systems, , Thus, Thus,, triene, triene, , the, the, (I), (I), , r-electron, r-electron, and, and over, over, , 7, 7, , cloud, cloud is, is delocalized over the entire chain in 1,3,5-hexathe, the ring, ring of the cyclic ether, furan (II), but the centra, , methylene, prevents interaction between the Tr-electron, methylene group, group in, in 1,4-pentadiene, 1,4-penta, clouds, As will be seen later, important differences in, clouds on, on the, the ends, ends of, of the, the chain., cha, chemical and physical properties stem from such differences in .-electron
Page 37 :
The Use of Resonance, , 21, , atoms in biphenyl lie in the same plane, and we may visualize single clouds of, r-electron density spread above and below the double-ring system. In contrast,, , 7, , the two rings in hydrocarbon IV lie in different planes. Although one might not, believe it from the structural formula of this compound, it may be calculated, from the van der Waals’ radius of the methyl group (Chap. 2) that the, methyl groups on the two rings would collide with each other if the rings were, H,C CHq, , \ ry/, H3C ch3, IV, to lie in the same plane. In IV, therefore, there is essentially no interac¬, tion between the 7r-electron systems of the two rings. The interaction or lack of, interaction may often be deduced from spectral studies. A number of examples, in which planarity (or departures from planarity) affects the properties, molecules will be described in later chapters., , of, , The Use of Resonance7, For most simple compounds having no more than one double bond, the “mod¬, ern picture” may be quite adequately represented by the Lewis structures, However, in cases where there is extensive delocalization of electron density,, the classical structures are not as suitable as either the 7r-electron structures or, the, 1, , streamer” pictures. Both of the latter type structures, however, are more, cult to draw and are far less convenient for the large number of chemists, , who were trained in the days before the concepts of cr and r electrons were, extensively used., Because of the convenience of thinking in terms of classical structures, t ere is in extensive use today a concept by which molecules that cannot be, a, , combination, , are represented instead by, of two or more Lewis structures. This concept is called, , resonance; we shall presently see why. When two or more legitimate Lewis, S, , ructures can be drawn for a compound, we say that the “true” structure, “ "'“"n'hc — -cep, has, , and Sons, Inc., New York, 1955; the final chapter of th', Chemistry, John Whey, naem, of some aspects of the resonance pfc,ure 1 For a Z,hr TZ *? quantitative treat,, the resonance concept (but not to its theoretical basis), l° CCrtain useso{, Huckd, Structural Chemistry of Inorganic Compounds (translated bv l, ', references, see, Co-, Inc., New York, 1950, pp. 434-437, d by, LonS)> Elsevier Publishing
Page 38 :
22, , -, , Atomic and Molecular Structure. The Use of Resonance, , (often called a resonance hybrid) is similar to each of the structures but, identical to neither. In short, it is intermediate in character. In the most familiar, case, the 7r-bond picture of benzene (Fig. 1-5) may be regarded as being a, combination, , of the two Kekule structures, both of which are wrong if con¬, , sidered alone., Similarly, resonance pictures may be used to show that there is double¬, bond character in certain of the bonds in 1,3,5-hexatriene (V), furan (VI), and, biphenyl (VII) which would be represented as single bonds in the classical, , CH2-CH=CH-CH=CH-CH7, , a O, +, V, , VI, , VII, , pictures. Thus, classical pictures represent the bonds labeled a and b in 1,3,5hexatriene, bond d in furan, and bond e in biphenyl as single bonds; but the, resonance forms above represent each of these as double bonds. (Other reso¬, nance pictures, in addition to those shown, may be drawn.), In recent years the dimensions of a large number of molecules have been, measured, and in many cases the resonance concept has been useful in rationaliz¬, ing observed dimensions that are significantly different from those predicted, from classical structures. The distance between the nuclei of two bonded atoms, is called a bond length. We shall see in the next chapter that a triple bond, between two atoms is shorter than a double bond between the same two atoms,, and that a single bond is longer than a double bond. Thus, the carbon-to-carbon, bond in ethane is longer than the carbon-to-carbon bond in ethylene, which is,, in turn, longer than the carbon-to-carbon bond in acetylene. In benzene, each, of the carbon-to-carbon bond lengths is the same, being greater than that cor¬, responding to a double bond but less than that corresponding to a single bond., The length of the bonds in benzene then confirms what the resonance picture, tries to show, that the carbon-carbon bonds in benzene are actually intermediate, in character between double and single. Likewise, bonds a and b in 1,3,5hexatriene, bond d in furan, and bond e in biphenyl are each intermediate in, length between a “normal” double and a “normal, , single bond., , Often resonance is indicated by a double-headed arrow,, , and the two, , or more contributing structures are called primary or canonical forms; for, example, for diazomethane:, , H, H:C::N::N:, , _, , H, <->, , H:C:N+:: :N:, • •, , (A), , ~, , (B)
Page 39 :
The Use of Resonance, , 23, , We have already seen that delocalization of electron clouds (that is,, spreading of electronic charge over the region of several nuclei) often leads to, structural stability. Since resonance is one way of picturing delocalization, one, might expect that if two or more permissible primary structures can be drawn, for a molecule, it should be more stable than a hypothetical molecule of the, same type with no delocalization. Stability of compounds is often determined, experimentally by measuring heats of chemical reactions; speaking very roughly,, a stable molecule will give off less heat in a given reaction than would a less, stable molecule of the same type. Now classical thermochemistry has been, developed to the point where heats of reactions can be broken down into a, series of makings and breakings of chemical bonds; and the heats of reactions, of the large majority of classical-structure compounds can be accurately pre¬, dicted by considering bond energies (Chap. 2). However, such predictions often, fail when applied to compounds for which two or more canonical forms may be, drawn. In such cases, measured stabilities are almost always greater than the, predicted stabilities of any of the canonical forms., Consider, for example, the combustion of benzene. The heat liberated in, burning a mole of benzene to carbon dioxide and water is sometimes estimated, by considering benzene as six C—H bonds, three C—C single bonds, and three, C=C double bonds. However the experimental heat of combustion of benzene, is 36 kilocalories per mole lower than the value predicted by using bond energies., Thus, it is often said that benzene is “stabilized by resonance” and has a reso¬, nance energy of 36 kcal per mole. It would seem that this choice of terms leaves, something to be desired. Actually the experimental energy of benzene is the, “correct” quantity. The resonance energy in a sense arises from the surprise of, the chemist when he realizes that the energy he incorrectly predicted using an, oversimplified picture was not what he found. The greater his surprise, the, greater would be the resonance energy. It should be noted, however,' that, resonance energies are somewhat more fundamental than mere “fudge-factors,”, used to adjust incorrect values to obtain better second guesses, for it is possible, to estimate relative resonance energies of a number of molecules by using one, of several quantum-mechanical treatments.5, Granted that the concept of resonance is useful in correlating the dimen¬, sions stability, and other properties of many compounds inadequately repre¬, sented by classical structures, one might well ask whether there is any funda, mental basis for the idea that a “true” structure can be represented at all by a, comb,natlon Qf pIctUres. We may ,nd.cate br.efly (he theore(, ;, , genw3"06 TeP' bymtr0dUCm* 3, -Hed a wave function", gen function, or (since it is generally abbreviated by the Greek letter ps(, , <*>, , **, , (London)’A,5S' 297 era*)!-
Page 40 :
24, , p),, , Atomic and Molecular Structure. The Use of Resonance, simply the, , p, , function. It will be recalled that quantum mechanics handles, , electron probabilities in space by using a mathematical analogy to the motion, of a point on a vibrating string. In the treatment of the string,, , p, , refers to the, , displacement of a given point from its position of rest; on the other hand, in the, treatment of electron probabilities, the square of the wave function,, , p~,, , deter¬, , mines the probability density at the position being considered. (Generally, in, a series of calculations, it is most convenient to work with, near the end of the calculations to square, , p, , p itself,, , waiting until, , to obtain a picture of probability, , density relationships.) This correspondence (although it probably would not be, predicted beforehand) is rational since probability density (or cloud thickness), can never be less than zero, whereas the displacement of a point can be + or, —, depending upon whether the point is above or below the position of rest., Now, for each Lewis picture, there exists, in principle at least, a particular, , p, , function. For simple structures such wave functions can be and have been, , developed, whereas for more complicated structures they presumably could be, developed if workers wished to expend the time and effort. These Lewis pictures,, and the, , p, , functions that correspond to them, represent “limiting” structures., , For molecules in which the resonance treatment is important, electron density, in these “limiting” structures is more localized than in the true structure, and, for a closer approximation to the truth, the mathematician must express delocalization of electronic charge over a greater area than indicated by the, limiting structures. Such delocalization may be handled in wave mechanics, by a linear combination of the wave functions of the individual limiting structures., (A linear combination of two quantities is obtained by multiplying each by a, constant and then adding.) More particularly, if we call the wave functions, associated with the two Lewis structures of diazomethane on page 22, , pB,, , it is possible to obtain a third function, , Pab, , pA, , and, , where, , Pab = C\pA + CiPb, and ci and c2 are the constants. From the wave-mechanical treatment of the, problem, it can be shown that PAB, corresponding to a structure over w ic, electron density is considerably delocalized, represents a, , better, , (that is, a structure of lower energy) than is represented either, , structure, , y Pa or pB-, , Thus, the mental combination of structures A and B of diazomethane corre¬, sponds, in a sense, to the mathematical combination of the wave functions, , pA, , and pb to obtain PAb, the wave function of a structure of lower energy,, , n, , carrying out the calculations expressing this delocalization, t e trea me^, becomes mathematically similar to the handling of certain vibrating ^tC"“, such as strings, tuning forks, or pendulums, in particular, , to t e mu, , fication in vibrations when two independently vibrating bodies are sudden y, attached to each other. We can see how the vibrations of one of such bodtes
Page 41 :
25, , The Transition Between Ionic and Covalent Character, , affects the other and often call this a “resonance” effect. However, the vibra¬, tions of the bodies are observable, whereas the contributing wave functions are, hypothetical. To say that an electron cloud “resonates” like a string or tuning, fork is clearly carrying the mathematical analogy too far., , The Transition Between Ionic and Covalent Character, Although we have devoted much space to discussion of the covalent bond, it, will be remembered that many bonds between metals and nonmetals are con¬, veniently considered as being formed from ions. In an extreme case such as, potassium chloride, the bond between a potassium ion and a neighboring, chloride ion in the KC1 lattice results almost exclusively from classical electro¬, static attraction. Although the positive and negative ions in the lattice are kept, from too close an approach to each other by repulsions of the electron clouds,, there is essentially no distortion of the clouds themselves as the ions approach, each other. Between two adjacent ions there is very little electron density that, can be thought of as belonging to the pair as a unit., Suppose we regard the hydrogen molecule as being formed from the coming, together of a hydrogen ion, H+, and a hydride ion, H:~ As these two ions, approach each other, the spherical electron cloud of the hydride ion becomes, very much distorted by the tiny positively charged proton. The electron cloud, spreads itself over the region of both nuclei with much of the cloud lying in the, space between the nuclei. The bond, of course, is the familiar covalent bond in, H2; in this interpretation, the covalent bond has resulted from extensive dis¬, tortion of the negative ion by the positive ion., Logically, between these two extremes in bond character there are cases, where there is an intermediate degree of distortion or deformation of the elec¬, tron clouds. Generally one regards the anions as being distorted by the cations, (a though the cations are also appreciably distorted in some cases). Such a, istortion is often called a polarization; a portion of the negative cloud of the, anion becomes “pulled” over toward the region of the positive cation. As the, distortion is increased, the bond between anion and cation takes on more and, more covalent character. Figure 1-8 crudely illustrates this trend. Such a pic, erallySSne°akeaf y, , m‘° <’uantitative language, for one does not gen-, , or d ” B, ‘°r T, S “35-percent distOT'«*” and “65-percent nondistorted. By means of the resonance concept, however, the character of a h !,, between two different atoms can be described in m, though it might well be argued that in tM c,, , 2, , f, , b, , d, , (al-, , glve the picture much additional physical £££%^ “* ^ *°, , completely ionic, , nor completely covalent, , but, , ££
Page 42 :
26, , Atomic and Molecular Structure. The Use of Resonance, , description is analogous to that used to describe compounds represented as, resonance hybrids. Accordingly, let us consider a completely ionic structure as, one extreme (having associated with it a wave function,, , \f/A+B-), , and a completely, , covalent structure as the second extreme (having associated with it a wave func¬, tion, ypA-.B). A better wave function, more nearly representing the “actual”, structure, is then obtained by a linear combination:, better, , =, , 'Pa+B*, , T, , C‘$Pa\B, , In such a linear combination, the relative importance of the ionic and covalent, structures is related to the ratio of the coefficients c\ and c2. For predominantly, covalent bonds c2 » C\., , Q©, practically no distortion, ionic bond, , some distortion, intermediate, , much distortion, covalent bond, , Fig. 1-8. Covalent Character and Distortion, , Now the energy of the form A+B~ may be estimated from electrostatics,, whereas the energies of both the actual molecule, AB, and its hypothetical com¬, pletely covalent structure, A:B, may be estimated from thermochemistry (as, described in Chap. 2). When comparisons are made between the experimental, stability of AB and the hypothetical stabilities of the two extreme structures, the, molecule itself proves more stable than either extreme form. This must be so if, the resonance treatment has been used correctly., , Limitations of the Use of Resonance, By shrewd and practised use of the resonance idea, one may make qualitative, and sometimes quantitative predictions about the structure and reactivity of, oceanic compounds. To the chemist, probably the greatest appeal of this treat¬, ment is its ability to correlate behavior previously regarded as anomalous in, terms of pictures that he already understands. However, it is obviously (and, perhaps fortunately) possible to draw and to manipulate resonance pictures, without understanding the nature of the physical wave picture that makes the, resonance concept a legitimate one. Given this human situation it is not su ¬, prising that the language of resonance has sometimes been taken over, , -, , critically, that meanings have been attached to pictures and operahons t, were never meant to apply, and that the ideas are sometimes modified to apply
Page 43 :
27, , Limitations of the Use of Resonance, , to situations for which they were not originally intended. Chemists who pre¬, sumably should know better are sometimes caught saying that a molecule, “stays in this resonance form only a small portion of the time,” that “one reso¬, nance form is more reactive than the other,” or that “resonance is decreased, since two types of resonance are going in opposite directions in the molecule.”, Aside from such misuses, however, there are aspects of the resonance treat¬, ment that are open to criticism. The selection of the admittedly arbitrary pri¬, mary (canonical) structures is the point at which the resonance treatment is, often attacked. There are three rules governing the selection of such canonical, structures; two of these are easy to apply, whereas the third sometimes causes, trouble. The three rules are these:, (1) Structures may contribute only if the relative positions of the atoms, are not changed. Thus, H, , H: C:: N:: N:, , and, , H: N:: C:: N: H, , cannot contribute to the same hybrid. Similarly,, H, , _, , H, , H: C:: C: O: H, , ••, , ••, , and, , H:C:C::0:, , H, , .. .., , H H, , are different compounds (tautomers), even though they are readily, interconvertible., (2) Only structures with the same number of unpaired electrons may con¬, tribute to a given resonance hybrid. (This restriction is important in, the consideration of organic biradicals, Chap. 16.), (3) Only those structures having similar energies (stabilities) can contribute, appreciably to a given hybrid. For example, there are three Lewis, structures of methyl azide, CH3N3:, CH3:N: :N: :N:, , CH3:N:N:: :N:, , (A), , <B>, , CH3:N:: :N:N:, (C), , ", , Of theSe, structures (A) and (B) are considered ,o contribute to the, hybrid form of methyl azide, but the contribution of form (Q is con, sidered negligible. Form (C) is sometimes termed a “h gh-energy, °rm> mean‘ng that “"Siderable electrostatic energy would be needed, to arrange the charges so that *, n, ., e eeded, one end of the structure and, e-negative charged atom is at, jacent atoms near the middle, , " Smgle'P°S“1Ve charS«, , on ad-
Page 44 :
28, , Atomic and Molecular Structure. The Use of Resonance, In many cases, it seems obvious that the primary structures chosen have, , the same energies since they are equivalent. The two Kekule structures for, benzene are, except for rotation, identical. The case of carbon dioxide, however,, is more troublesome. In C02, generally represented as 0=0=0, the carbonto-oxygen bond lengths are significantly less than the C-to-O bond lengths, in aldehydes and ketones. At an early point in the development of the resonance, treatment, it was suggested that this bond shortening indicated contribution, of triple-bonded structures, and C02 was accordingly represented as the hybrid:, 0=C=0, , 4—>, , (A), , ~0—C=0+, , 4—>, , (B), , +0=C—O(C), , To explain the bond shortening, Pauling estimated that the two triplebonded structures are about as important as the double-bonded one. Now it is, not easy to estimate the energy of a hypothetical structure such as (B) or (C),, but such estimates that have been made76 indicate that (B)—and hence (C)—, is far less stable (about 100 kcal per mole) than (A), chiefly because of the reluct¬, ance of the oxygen atom to assume a positive charge. Even if it is assumed that, the estimated value of the energy difference between (A) and (B) is consider¬, ably in error, the energy difference is almost certainly not small. Although it is, indisputable that there is a difference between the C=0 bonds in C02 and, those in aldehydes and ketones, it is doubtful that the difference is best expressed, by heavy contributions of high-energy triple-bonded structures., This is just a single example. At present a more satisfactory (and more, sophisticated) picture of the electron structure of C02 is available,9 and the, treatment indicated above may be said to be largely of historical importance., Nevertheless, it serves to show that application of the concept of resonance to, such species as carbon dioxide, ketene, the azide ion, or the cyanate ion is not, always as straightforward as we would like. On the other hand, the resonance, theory itself, for the most part, stands on firm ground. Moreover, the consider¬, able success of resonance in correlating the physical and chemical properties o, a host of multiple-bonded organic compounds makes the concept not on y, legitimate, but, as will be seen, exceedingly useful., , The Hydrogen Bond'", No discussion of chemical bonding should omit mention of the hydrogen bond, The organic chemist finds evidence of this type of bond in a large number o, compounds having O-H linkages and in a considerable number contammg, , 'Mlor mor * cTmpletc°discuss^ns otThe'hydrogen bond, see Pauling (Ret. 6c). pp. 284334; and Syrkin and Dyatkina (Ref. lc), pp. 273 286.
Page 45 :
The Hydrogen Bond, , -, , 29, , N—H linkages. Perhaps the most familiar evidences of the hydrogen bond are, the “abnormally high” boiling points of carboxylic acids, alcohols, and many, amines. For example, the two compounds, dimethyl ether, (CH3)20, and, ethanol, C2H5OH, have the same formula weight, but the latter (with an, O—H bond) boils over 100° higher than the former. Similarly, acetic acid, (bp 118°) and methyl formate (bp 31°) have the same formula weight, but dis¬, play a marked difference in volatility. The hydrogen bond is best considered, as a special example of dipole-dipole interaction in which the “negative”, oxygen atoms of one molecule are attracted electrostatically to the “positive”, hydrogen atoms of a second molecule., R, , R, I, , I, , ✓, , P\, , P, , H, , O, I, R, , H, V, I, R, , \X, o, I, R, , Such attraction must be stronger than the weak van der Waal's forces that hold, uncharged molecules together in the liquid and solid states. However, hydrogen, bonding cannot be as strong as ordinary ionic or covalent bonding, otherwise, the boiling points of alcohols and carboxylic acids would become comparable, to those of substances having multiple-bonded continuous networks such as, NaCl or diamond. Thus, the hydrogen bond is a moderately weak one, generally, about 5 to 10 percent as strong as ordinary covalent single bonds. The vast, majority of organic compounds that dissolve in water do so with attendant, formation of hydrogen bonds with the solvent., In cases where hydrogen bonding occurs, the two heavier atoms approach, eac, , ot er quite closely; in the solid state, this closeness of approach is detect¬, , able by x-ray methods and is an important indication of hydrogen bonding, (even though the hydrogen atoms themselves often cannot be located)^, sionally hydrogen bonding persists in the vapor state. Thus, gaseous HF is, associated at low temperatures, and formic add exists as a dimer in both the, vapor phase at low temperatures and in solutions in inert solvents. The structure, of this dimer is shown.”, structure, , H-, , C—H, , \, , O—H-O, , As indicated, the hydrogen bond between, , t, , metric; that is, the hydrogen is closer to, "° °XySen atoms is not sym, “Kapi, a, », ,, °Sen, ls, closer, to, one, oxygen, than to the other, Karle and Brockway,, Am. Cfem. ^ ^ „„ (, ^, °mer-
Page 46 :
30, , -, , Atomic and Molecular Structure. The Use of Resonance, , Important effects may be traced to the presence of hydrogen bonding, within a single molecule (intramolecular). In an extreme case, the hydrogen-, , bonded molecule, 2,6-dihydroxybenzoic acid (VIII), has an ionization constant, approximately 10 thousand times as great as that of its isomer, 3,5-dihydroxybenzoic acid (IX), in which intramolecular hydrogen bonding cannot occur., Another curious difference due to hydrogen-bonding effects is illustrated, by comparing o-nitrophenol (X) with its para isomer. Hydrogen bonding is, H, , H, , I, , I, , VIII, , IX, , X, , XI, , present in both isomers, but an intramolecular hydrogen-bonded structure, corresponding to XI is not possible for the para compound. Therefore, hydrogen, bonding in the latter must be between the hydroxyl group of one molecule, and the nitro group of another; that is, the para compound is associated whereas, the ortho compound is not. This structural difference is reflected in the much, greater volatility and solubility in nonpolar solvents exhibited by the ortho, compound., When the hydrogen atom of an —OH group attached to an aromatic ring, participates in intramolecular hydrogen-bond formation, the vibration between, the oxygen and hydrogen atom of the O—H bond (which may geneially be, observed in the infrared spectrum of the molecule) decreases in frequency., In effect, hydrogen bonding has “loosened” the O—H bond. In such cases, the, sharp infrared absorption peaks in the region of 7050 and 3525 cm ’, present in, “normal” phenols, are either absent or very much “smeared out” (Chap. 3)7*, In the majority of instances where intramolecular hydrogen bonding is, important, the hydrogen atom finds itself a member of a six-memberedplanar ring., In such cases, each of the remaining five atoms constituting the ring has a, T electron associated with it (essentially equivalent to the classical statement, , 7, , that such rings can be drawn with two double bonds). Thus, the intramolecular, hydrogen bonding present in salicylaldehyde (XII) is absent in its alicyclic analog, (XIII) since the H—O—C—C—C=0 system in the latter cannot be twisted, , » Wulf,, , at., J. Am. Chem. Soc., 58; 548, 1991, 2287 (1936): Errera, et at., Jour, de Physique, , et le Radium, 6; 154, 281 (1935).
Page 47 :
The Hydrogen Bond, , -, , 31, , into a planar configuration. Likewise, intramolecular hydrogen bonding present, in the enol form of acetylacetone (XIV) would not be expected in the enol form, of its homolog, acetonylacetone (XV), since there are too many carbon atoms, , H, , H3%Acr, CH3, i, ii, OH -O, , XII, , XIII, , XIV, , CH, CH3—C"" xCH2, OH ^C-CH3, o, XV, between the two oxygen atoms. When a compound containing an N—H or an, O—H linkage cannot undergo intramolecular hydrogen bonding, some degree, of association by intermolecular hydrogen bonding (for which steric require¬, ments are much less stringent) will almost always occur., , EXERCISES, , FOR, , CHAPTER, , 1, , 1. List all four quantum numbers for each of the 11 electrons in the sodium atom., 2. How many different electrons are possible with n = 4 and / = 3? How manv (total), with n = 4?, 3. Calculate the number of unpaired electrons in the atoms of Na, Al, and P., 4. (a) Write Lewis formulas for HN3, (CH3)3SBr, BH3CO, and CH3OCl., (b) Show that Lewis formulas cannot be written for B2H6, PF5, and N02., 5. Arrange the following bonds in order of increasing ionic character:, Be, , O, X, , O, O, , O, Si—O, Se-, , -yj, , u»> nr. -»*. tad :,i. £, , " “■, , *■w, uons is most like the carbon atoms in acetylene?
Page 48 :
32, , Atomic and Molecular Structure. The Use of Resonance, (b) Draw 7r-bond pictures of naphthalene and azulene (shown below). Show that, the Kekule forms of naphthalene are identical when represented in this way and, that the two forms of azulene are also identical., , naphthalene, , azulene, , 9. (a) Draw the possible Lewis structures for: COJ2 (3 forms); C302 (3 forms);, C6H5O- (5 forms)., (b) Sketch the canonical forms for each of the following, and label the formal, charges. Ignore structures in which more than one carbon bears a formal charge., , -C-CH,, »II, , o, , 2 Lewis structures, 5 “non-Lewis” structures, 6, , Lewis structures, , 5 “non-Lewis” structures, 10. The resonance theory stipulates that if two compounds have the same set of atoms, and bonds, then that compound for which the more canonical structures may be, drawn will be the more stable. On this basis, which would be the most stable,, anthracene, phenanthrene, or benzazulene?, , Q5> cA:, anthracene, , phenanthrene, , benzazulene, , 11. Explain why water has a higher boiling point than either hydrogen fluoride or, ethanol., 12. Certain of the compounds below participate in hydrogen bonding that may affect, their boiling points through association. Arrange them in the order of extent, of intermolecular hydrogen bonding (as would be evidenced by boiling-point, abnormalities):, acetic acid, acetaldehyde, ammonia, ethane, ethanol, fluoroform, hydro, cyanic acid, oxalic acid, salicylaldehyde, />-hydroxybenzaldehyde., 13, , Of the eight compounds shown below, four participate in intramolecular hydrogen, ' bonding (either in the indicated form or the enol form) and the remaining four do, not. Predict which four exhibit effects of intramolecular hydrogen bonding an, the reasons for your choices., , give
Page 50 :
CHAPTER 2, , The Energies, Lengths, and, Orders of Covalent Bonds, , Although the majority of organic chemists today picture the configurations, , of molecules in much the same way as did the chemists of 1910, our ideas con¬, cerning molecules have been brought into somewhat sharper focus by the, ability to assign energies to such molecules and to the bonds within them, and, further by the increased knowledge of atomic and molecular dimensions., , Bond Energies, For a diatomic molecule, the bond energy is defined as the energy needed to, split the molecule into atoms (both the molecule and component atoms being, in their ground states). Such a bond energy or “energy of dissociation” may, often be determined from the band spectrum of the molecule* or may be closely, approximated by measurements of the equilibrium constants for dissociation, of the molecule at high temperatures. The important van't Hof equation states, that a plot of the logarithms of such dissociation constants (values of In A') at, different temperatures against the reciprocals of the temperatures themselves, should yield a straight or nearly straight line whose slope is simply related to, the desired heat of dissociation,, d{In K) _ _H, d(\/T), , R, , where H is the heat of dissociation and R is the gas constant (1.99 calories pci, mole). Such thermodynamically determined heats of dissociation (w ic, , are,, , in actuality, enthalpy values) lie close to the energies of dissociation determine, - See, for example, G. Heraberg, Spectra of Diatonic Molecules. Van Nos.rand, Inc.,, Toronto, 1950, pp. 437-450., , 34
Page 51 :
Bond Energies, , -, , 35, , spectrally, but more detailed considerations indicate that the two types of, values should and do differ by small amounts. This distinction will be ignored, in the following discussions, and we shall conveniently (although somewhat, incorrectly) refer merely to bond “energies” whatever their source., For diatomic molecules, single-bond energies range from 135 kcal per mole, for HF to 36 kcal per mole for I2. By convention, bond energies are generally, given positive signs., For a molecule with three or more atoms (hence, two or more bonds),, the heat of formation from the isolated atoms may be regarded as the sum of, all the bond energies. The heat of formation of the Sg ring from eight sulfur, atoms should be eight times the average energy of formation of a single S—S, bond; similarly, the O—FI bond energy may be considered as being just half, of the heat of formation of a water molecule from individual hydrogen and, oxygen atoms. Now the energy of the reaction 2H + O —» H20 is not easily, measured, but the energy of the reaction 2H2 + 02 —>2H20 has been meas¬, ured calorimetrically many times. The latter value (116 kcal for 2 moles of, steam) may be combined with the known heats of dissociation of the 02 and, H2 molecules to find the O—H bond energy. For such combinations, we use, the familiar law of Hess that allows thermochemical equations to be handled, as algebraic equations:, , adding,, , 2H2 + 02 —> 2H20 (gas), , +116 kcal, , 4H, , -> 2H2, , + 207 kcal, , 20, , -*, , 02, , +118 kcal, , + 20 —-» 2H20 (4 O—H bonds) + 441 kcal, , The O—H bond energy thus calculated (110 kcal) is an average value, for it, should not be assumed that the formation of OH- from an oxygen and hydrogen, atom will release the same energy as the formation of H20 from OH- and HThe values of the C-H, C-C, C=C, and feC bond energies may be, determined by combining the readily measured heats of combustion of hydrocarmcthane"1, CH4, C (graphite), 2H2, 2H,0, , co2, adding,, , hCi“S °f C°mbuStion of carb™, + 202 -> 2H20 (gas), , hydrogen; thus, for, , + CO2 + 192 kcal, , sublime, , -* c (gas), ->4H->2H2, -* C (graphite), , — 170 kcal, — 207 kcal, C)2, + 02, , — 116 kcal, —, , 94, , kcal, , CI+, t4H- + C (gas), , - 395 kcal, , (breakup of 4 C—H bonds)
Page 52 :
36, , The Energies, Lengths, and Orders of Covalent Bonds, , The average value for the C—H bond energy, 99 kcal, is thus one fourth of the, energy released in the reaction C + 4H» —> CH4—that is, the formation of, methane from the free atoms in the gaseous state., It is important to emphasize the distinction between the ordinarily used, bond energies and bond-dissociation energies—the energies needed to rupture, specific bonds in specific gaseous molecules (for example, the C—H bond in, chloroform). The latter values are somewhat more useful and although they, are more difficult to obtain than “average bond energies,” a large number are, now available.8 However, we shall be mainly interested in carbon-hydrogen, and carbon-carbon bonds, and bond-dissociation energies have been determined, for only a few of these. Subsequent discussions will therefore be based on, “average” bond energies., Of the values necessary in calculating bond energies, the heat of atomization, of graphite to gaseous carbon atoms is subject to greatest doubt. The recent, value of 170 kcal per gram atoms(o) has been selected here, but a smaller, value, about 133 kcal,s(6) is used by a number of workers. As will be shown in, Exercise 4, the value chosen will affect the apparent values of bond energies, involving carbon, but many conclusions based upon comparisons of such bond, energies should not change. A similar uncertainty exists as to the true value of, the N=N triple-bond energy., Single-bond energies for the more familiar bonds appear in Table 2-1., A number of these have been compiled by Huggins.'* The energies of double and, triple bonds, often evaluated from heats of combustion or heats of hydrogenation, (Ex. 3), are found to be higher than the corresponding single-bond energies., C—C, , 80 kcal, , C—N, , 62 kcal, , C=C, , 142 kcal, , C=N, , 121 kcal, , C=C, , 186 kcal, , C=N, , 191 kcal, , Bond-energy values such as those given in Table 2-1 would be most useful, if we could assume that the energy of a bond between two given atoms in a, polyatomic molecule is independent of other atoms in the molecule, and that, therefore, for example, the heat of formation of ethanol, CHjCH.OH, from, the free atoms would be exactly equal to the bond energies for one C-C bond, , one c-O, , bond, one O-H bond, and five C-H bonds. There are indeed, , . For summaries of the methods by which bond-dissociation energies are evaluated see:, Cottrell, Thr Strengths cf Chernicnl Bonds,, , ^ntly determined bonT, " Free, , S°nS'i WC ChupWkaYOand l^m^L, Phys., 22, 1472, , "''THuggS75, 4125 (1953)., , ,««-• John Wiley and, , (1954,. (« Langer, Hippie, and
Page 53 :
Bond Energies, , -, , 37, , many cases where bond energies determined by investigation of one molecule, may be used to calculate the heats of formation and heats of reaction of other, molecules having similar bonds. If this were not the case Table 2-1 would be, of very little use. On the other hand, there are a considerable number of, organic compounds to which simple bond-energy calculations may not be, applied without substantial modification of treatment., The most important class of compounds for which bond energies alone, will not predict correct heats of formation or reaction are those compounds, having conjugated double bonds and which are commonly represented as, “resonance hybrids” of classical structures. The very familiar case of benzene, , Table 2-1. Energy Values for Some Single Bonds, , has been noted on page 23. As with benzene, the heats of formation of a num¬, ber of benzene derivatives (toluene, the xylenes, phenol, anisole, aniline, benzaldehyde, etc.) are each roughly 36 kcal per mole greater than would be, Pr!, T, , u, yrCOnSldenng ‘heSe comP°unds as constructions of ordinary single, , eaanch ofthe e d°ndS', ^ ^ °f C°mbuSti°" a"d hydrogenation of, of these derivatives are roughly 36 kcal per mole less than would be pre¬, dicted from single classical structures. The word “roughly” is not used her, , becau, , f the experimental errors that arise in measuring LI ofImbnstn, , of be,LrTn7o2ka“lUapery ^TuTeXT d, , °U‘ *° ““, , the variation in bond energies of “normal” C—C and O^Cbo ^, enough to allow a considerable latitude in guessing what the h , f, u, 86, or hydrogenation would be if o, ., f, e heat of combustion, by a single classical structure., , ^ d ^ represented adequately
Page 54 :
38, , The Energies, Lengths, and Orders of Covalent Bonds, The heats of hydrogenation and combustion of compounds having two or, , more benzene rings fall even further below the values that would be predicted, from bond energies alone (naphthalene—61 kcal per mole less; anthracene—, 84 kcal per mole less, etc.), whereas nonbenzenoid compounds containing con¬, jugated double bonds tend to show smaller departures (pyrrole—21 kcal per, mole less; 1,3-butadiene—roughly 3 kcal per mole less). Similarly, if the heats, of formation of esters or amides are estimated by considering the C=0 linkages, in these compounds similar to those in ketones, the values obtained are about, , 20 kcal per mole less than the true heats of formation., In each of the instances cited, the true heat of formation is more than the, value calculated from bond-energy values alone. The differences, the so-called, , resonance energies, arise because classical structures ignore the stability associated, with the spread of 7r-electron density over the conjugated systems. I.t is interesting, that certain contemporary workers have carried out theoretical evaluations of, resonance energies for a number of conjugated systems, using structural con¬, siderations alone. For descriptions of such calculations (an understanding of, which requires some familiarity with the fundamentals of wave mechanics) the, reader should consult more advanced works.5, Aside from conjugation effects, there are other cases in which strict addi¬, tivity of bond energies is not observed. Consider, for example, the heats of, combustion of isomeric paraffin hydrocarbons, , for example, n-pentane, iso¬, , pentane, and neopentane. Since each of these pentanes has the same number, of c—C bonds and C—H bonds, one might expect them to have the same heats, of combustion., , However, very accurate measurements by Rossini and co¬, , workers5 show that the heats of combustion of the three isomers are, respec¬, tively, 845.16, 843.24, and 840.49 kcal per mole. Similarly, the isomeric octanes, have different heats of combustion, ranging from 1317.45 kcal for the straightchain isomer, rc-octane, to 1313.27 kcal for the highly branched isomer, 2,2,3,3tetramethylbutane. If bond energies were constant, we should also expect that, the heats of hydrogenation of all olefins having just one double bond would be, the same since the hydrogenation process in each case involves the breakage, of an H—H bond, the creation of two new C—H bonds, and the conversion of a, C=C double bond to a single bond. However the heats of hydrogenation ol, monoolefins have been found to vary from 32.8 kcal per mole for ethylene to, 23.5 kcal per mole for cyclooctene.7, Such differences in heats of combustion and hydrogenation, as well as a, « See, for example, G. W. Wheland, Resonance in Organic Chemistry, John Wiley andSom,, Inc, Ne; York, ,955, pp. 629-64,, 665-672; H. Eyring, J.5 and G. E. K.mball,, Quantum Chemistry, John Wiley and Sons, Inc., New York, 1944, PP- 249 ^ •, . Rossini, et al., “Selected Values of the Properties of Hydrocarbons, U.S. Government, Office, Washington, D.C., 1952., Q_,, r Kistiakowski, et al., J. Am. Chem. Soc., 57, 65, 876 (1935),, , ('1936')., ,
Page 55 :
Electronegativity, , -, , 39, , number of additional discrepancies of similar magnitude, cannot easily be, ignored although, admittedly, they are small compared to the values of the, bond energies themselves. It is possible to refine work with bond energies, using,, in effect, a number of slightly different C—H, C—C, and C=C bond energies,, the choice of which depends upon groups adjacent to the bond being consid¬, ered.5 However, for many purposes it is acceptable (and certainly more con¬, venient) to use the value of 99 kcal per gram bond for C—H and to adopt, those single values for the C—C and C=C bonds that give the best approach, to additivity in the greatest number of cases. The values on pages 36 and 37, have been chosen on this basis., , Electronegativity, A variety of physical and chemical evidence indicates that the atoms of oxygen,, nitrogen, and the lighter halogens attract electrons more strongly than do the, atoms of carbon, hydrogen, or the metals. Consider, for example, the formation, of a C—Cl bond from the free atoms. As the atoms approach each other, there, occurs a redistribution of electron density with a marked thickening of the, valence electron cloud in the space between the two nuclei. After the bond is, formed, the bonding electrons are attracted simultaneously by both nuclei., These electrons, however, do not feel the pull of the entire nuclear charge., Situated between the valence electrons and the nuclei are the inner (non¬, valence) electrons, almost spherically symmetrical clouds of negative charge, which behave toward the outsider as if their charge were concentrated at their, centers—that is, in the nuclei. The effective positive charges on the nuclei are, thereby decreased by the nonvalence electrons, this being the so-called screening, effect. On this basis, the electrons constituting the C—Cl bond should be, attracted to the carbon nucleus by the action of four (6 - 2) protonic charges, and to the somewhat more distant chlorine nucleus by the action of seven (17*10) protomc charges. Although a more complete picture also requires that we, constder add.ttonal screening of each valence electron by those portions of the, remammg valence electrons lying close to the nuclei, the following qualitative, due to r‘,0n ,S 3 'I' °ne: The eIectron-attra«ing Power of an atom (which is, due to the portion of its nuclear charge that survives screening) dene d, (a) the difference between its atomic number and the number of its nonvakT", electrons and (b) the proximity of the bulk of the valence elec, on de, ,, the nucleus. In more familiar terms an atom will K, f ,, nS“y to, ifit has a large number of outer valence electron, , ,h ,, , ,, , r°n at'raCt°r, , periodic table) and if i, is small. Somewhat paradoxically”'0 ‘he, men, which attracts electrons strongly is hHffv electro ', 7, . "" ele’, ‘Klages, Ber., 82, 358 (1949)., ega ive, meaning that
Page 56 :
40, , The Energies, Lengths, and Orders of Covalent Bonds, , it will acquire the greater share of electron density when it competes with a, poorer electron attractor. Since the chlorine atom in the C—Cl bond is a better, electron attractor (that is, is more electronegative) than carbon, we frequently, regard this bond as polar with the chlorine at the negative end and the carbon, at the positive end. Fluorine is still more electronegative and a C—F bond is, even more polar., The concept of electronegativity is both important and useful. Differences, in electronegativity have been called upon to account for variations in molecular, dimensions, in dipole moments, and in molecular spectra as one atom is sub¬, stituted for another. Similarly, there are countless variations in reaction rates, and in positions of equilibria which are conveniently attributed to the electronattracting (or electron-repelling) abilities of atoms at or near the sites of reaction., However, the electronegativities of atoms cannot be measured directly and there, is some question as to which of a number of indirect methods9 of evaluation is, open to the least objection. Familiar electronegativity scales have been based, upon ionization potentials and electron affinities/0 on bond-stretching force, constants/1 and on atomic radii.1* A particularly convenient scale based on, bond energies, has been developed by Pauling.13, If we select any bond between unlike atoms from Table 2-1, designating, such a bond A—B, it appears that the energy of bond A—B is always greater, than the geometric mean of the energies of bonds A, , A and B, , B. Three ex¬, , amples are listed as follows:, , \/Ea-a, , X, , Ea-b, , Eb~b, , Differences, A, , AW, , V^si-si X E0-o = 38 kcal, , ESi-0, , =, , 89 kcal, , 51 kcal, , 7.2, , V^o-o X £f-f = 35 kcal, , Ep—o, , =, , 44 kcal, , 9 kcal, , 3.0, , V^si-si X £V-f = 40 kcal, , E9i-F, , = 128 kcal, , 88 kcal, , 9.4, , Note that the greater values of A are associated with bonds between atoms, lying far apart in the periodic table, that is, the value of A increases as the, “electronegativities” of the bound atoms diverge. Now Pauling suggests that, the geometric mean of the energies of the bonds A—A and B—B is the energy, that the bond A—B would have if it were completely covalent (as are the bonds, ^_A and B—B). The “extra” bond energy, A, is considered to be related to, • The construction of electronegativity scales has been recently reviewed, Skinner, Chem. Revs., 55, 745 (1955). This paper also makes clear some of the limitations, , 78Z, , Phys.,, , 2, 782 (1935): 3, 573 (1935)., , it Gordy, J. Chem. Phys., 14, 304 (1946)., , O—n “,y, , :;SpaanuH™^, pp. 58—69., , “**• 1940’
Page 57 :
Electronegativity, , -, , 41, , the ionic resonance picture of the pair of bonded atoms (page 26) and is thus, called the “extra ionic resonance energy” of the bond. If the bond energies of, A—A and B—B are not greatly different (Ex. 6), their arithmetic mean rather, , than their geometric mean may be used., From the figures on silicon, oxygen, and fluorine, we see that the values of, A are not additive (that is, ASi_0 + A0-F ^ ASi_F); however, the square roots, of these “ionic resonance energy” terms are much more nearly so. Thus, square, roots of the A values may be taken as electronegativity differences on which a, linear scale of electronegativities may be based. If the “extra ionic resonance, energies” are expressed in electron volts (1 ev per bond = 23 kcal per grambond), electronegativity differences lie in the convenient range between 3.3, units and zero. Electronegativity differences for the triad silicon-oxygen-fluorine, are then: 1.4 (ev)# for Si—O, 0.6 (ev)>* for F—O, and 1.9 (ev)^ for Si—F., Having thus obtained differences in electronegativity, Pauling fixes the elec¬, tronegative t) of the most negative element, fluorine, as 4.0 and assigns values, to the other elements on the bases of experimental (A)^ values. A typical set, of electronegativities appears in Table 2-2., , Table 2-2. Electronegativities (Pauling), H, 2.1, B, 2.0, Si, 1.8, , C, 2.5, P, 2.1, , As, 2.0, , S, 2.5, Se, 2.4, , O, , F, , 3.0, , 3.5, , 4.0, , 3.5, , 4.0, , Cl, 3.0, Br, 2.8, , Sb, , Te, , I, , 18, , 2.1, , 2.5, , 2.0, , N, , 2.5, , 3.0, , carbon,, , " d°, , ahnost ccnpletely covalent. Con^;, the eUonegaLities 0“^, , preaches thetlr (a" ^'T, , -----«, , ^+^, , --
Page 58 :
42, , The Energies, Lengths, and Orders of Covalent Bonds, , Bond Lengths and Covalent Radii, The mean distance between the nuclei of two bonded atoms is called a bond, length. Interatomic distances and bond lengths have been generally deter¬, mined either by electron diffraction, x-ray diffraction, spectral studies, or by a com¬, bination of these methods/4 A newer method, neutron diffraction, has been of, use only for a small number of substances.15, For uniformity within this chapter, interatomic distances are expressed to, o, , the “nearest” 0.01 A unit; however, the values obtained by electron diffraction, are often far less reliable, whereas some of the values obtained by x-ray diffrac¬, tion and by spectral studies are known to greater precision (a few, indeed,, to the nearest 0.001 A)., It has been found that the length of a covalent bond between two given, atoms is often (but not always) essentially independent of the nature of the, molecule or crystal network in which such a bond occurs. In most aliphatic, hydrocarbons, for example, the C—C single-bond length is very close to 1.54 A,, the same length as the C—C bond in diamond. Where shorter C, , C bonds, , occur, we shall see that the compounds concerned are often best repiesented, by combinations of structures., The single-bond covalent radius of carbon is then 0.77 A—just one half, Qf the q—0 bond length. Similarly, the covalent radii of chlorine and iodine, may be set at 0.99 A and 1.33 A, respectively—one half of the mternuclear, distances in the CI2 and I2 molecules., For a small number of compounds (which probably should be regarded as, exceptional) such radii are nearly additive. The C—Cl bond length in methyl, chloride is 1.76 A, almost exactly the sum of the covalent radii of carbon and, chlorine as determined from the elements. Similarly, the C—I bond in methyl, iodide has been found to be 2.10 A; this again is the sum of the covalent radii., Far more often, however, the length of a bond between two unlike atoms is, appreciably shorter than the sum of the covalent bond lengths as determined, from the elements., ., Covalent radii for the nonmetals are listed in Table 2-3. Note that for a gtven, , ..For discussions of electron diffraction, see, for, lilZnf/o^, Chen,., 5, 395 (1954); and Karle and Karle, in BraudeapplicaSlructurc, by Physical Methods, Academic Press, Inc New York 19 ’,(’^rtson in Organic Crystals, non of a-ray diffraction to organic compound,; is descrlpKL, methods, and Molecules, Cornell University Press, Ithaca, 1953. bate, , Nostrand> Inc, , Toronto, 1950;, , ^~John, and Sewlk ^rX^1947%20 (1950)?9’, , Wi,'y
Page 59 :
Contraction in Bond Lengths with Increasing Polarity, , -, , 43, , group in the periodic table, covalent radii generally decrease as one moves, from left to right. Students sometimes find this trend surprising, reasoning that, the addition of an electron could scarcely make an atom smaller. It should be, recalled, however, that the valence electrons of elements in the same period, generally have the same principal quantum number. Adding electrons tends, more to “thicken” the electron cloud about the outside of an atom than to, expand it. An element in a period also differs from the element preceding it, in its number of protons, and the addition of a proton to the nucleus tends to, draw the electron clouds inward. The largest element in each period thus has, the smallest atomic number., , Contraction in Bond Lengths with Increasing Polarity, As the difference in the electronegativity of two bound atoms increases, depar¬, tures from additivity of bond lengths become more and more marked. For, example, the covalent radii of carbon and nitrogen (half the C—C bond length, in diamond and half the N, , N bond length in hydrazine, respectively) add up, , to 1.51 A, whereas the “normal” C—N bond length is observed to be about, 1.47 A. Normal C, , O bonds tend to be even shorter (about 1.42 A), even, , though the covalent radii of nitrogen and oxygen are almost the same. Much, greater “shrinkages” are observed for bonds such as Si—Cl, Si_F, Si_O,, As—Cl, and P—F for which the more positive member is beyond the first, eight-membered period in the periodic table., The Schomaker-Stevenson relationship proposes that such bond shrinkage is, directly proportional to the electronegativity difference between the bound, atoms and that the interatomic distance, rA_B, for a single bond is thus linearly, related to the covalent radii, rA and r„:":, ta-b, , = rA -f- rB, , observed valuers obtained^:) C°Va'ejlt rad'*’ much be«<=r agreement with the, i, i, amed by use of the treatment proposed bv Hncrmnc n mi, latter shows that there k „, DY Muggins.17 The, bonds, having dimensions of length^rela”^11^, bond energy such that, °, th actual bond length and, ^, tab = ^ab, , |, , + — In, a, , EAB, , 16, 17, , Schomaker and Stevenson,, J-Am. Chem. Soc., 63, Huggins, J. Am. Chem. Soc.,, , 75, 4126 (1953)., , ,, , 37 (1941), , (2)
Page 60 :
44, , -, , The Energies, Lengths, and Orders of Covalent Bonds, , Now, a is a constant; when E is expressed in kcal per gram-bond, a is happily, close enough to the value 4.6 to permit equation (2) to be rewritten, , tab ~ rab, , + ^ l°gio Eab, , (3), , The r' values are called constant-energy distances; these are about 0.8 A greater, than ordinary covalent bond lengths and may be separated into contributions, of each of the two member atoms (just as may bond lengths in very nearly, covalent bonds). In analogy with atomic radii, the r' value for each atom is, called a constant-energy radius. Typical constant-energy radii are 1.22 A for, O, , °, , carbon (as compared to the covalent radius 0.77 A) and 1.73 A for iodine (coO, , valent radius 1.33 A). The Huggins relationship then becomes, , tab = ra + tb ~ 2, , l°gio, , (4), , Eab, , Only in the case of hydrogen, for which it is apparently not possible to assign, a good single constant-energy radius, does this treatment meet with appreciable, Table 2-3. Covalent Radii and "Constant-energy Radii”, , Covalent, Radius, , Constantenergy, , Covalent, , Radius, , Radius, , Constantenergy, Radius, , Ge, As, Se, Br, , 1.22 A, 1.21, 1.17, 1.14, , 1.61 A, 1.63, 1.58, 1.56, , 1.40, 1.41, , 1.80, 1.83, 1.79, 1.73, , H, , 0.28 A, , 0.82-0.88 A, , C, N, , 0.77, , 1.22, , 0.75, 0.74, , 1.12, , 0.72, , 1.11, , Sn, , 1.17, , 1.57, , Sb, Te, , 1.10, 1.04, , 1.53, 1.46, 1.44, , O, F, Si, P, S, Cl, , 0.99, , 1.12, , I, , 1.37, 1.33, , difficulty. The chief objection to this treatment would seem to be that the r, values are actually parameters whose values are fixed to give the best fit to, Equation (4) for the greatest number of bonds; thus they lack the physical, significance of covalent radii, which may be obtained by direct measuremen s,, Table 2-3 summarizes both the covalent radii and the “constant-energy radii, for a number of the nonmetallic elements.
Page 61 :
Bond Lengths in Resonance Hybrids, , -, , 45, , In Chapter 1 it was noted that a double bond (often referred to as a bond of, , bond order 2) between two atoms is shorter than a single bond (bond order 1), between the same two atoms; triple bonds are still shorter than double bonds., C—C, , 1.54 A, , C, , C, , 1.33 A, , C=C, , 1.19 A, , N—O, , 1.46 A, , N=0, , 1.14 A, , N=0, , 1.11 A18, , This shrinkage may be considered due to the attraction of the 7r-electron “dumb¬, bells” present on each atom forming a multiple bond. Recall that a single bond, is quite closely directed along the line of centers of the bonded atoms but that, the double bond has, in addition to this cylindrically symmetric bonding cloud,, extra attraction between the ir electrons, one on each atom. The atoms forming, a triple bond have two ir electrons each, resulting in two tt bonds in addition, to the a bond (which, when alone, comprises the ordinary single bond). We, might note that the decrease in bond length with bond order parallels an in¬, crease in bond energy; extra stability is often accompanied by bond shrinkage, (as is, of course, testified by both the Schomaker-Stevenson and Huggins equa¬, tions for single bonds)., , Bond Lengths in Resonance Hybrids, In discussing the resonance concept, we saw that a bond represented as a single, bond in one canonical form might well be represented as a double bond in, another form. In benzene, one of the simplest cases, the three double bonds in, one Kekule structure become single bonds in the other, and the same is true, lor the carbon-carbon bonds in the heterocyclic compound, pyrazine (I), , In, , I, this sense, the carbon-carbon bonds in benzene and in pyrazine are “midwav”, , may benaSscribednandordeer o" fsHimUarl’ ^, , V*, , °f reasonin«., , is that reported for nitrosyl perchlorate NO+ClO^^e ^AdH N==° “triPle-bond” length, (London), 9, 120, (1955)., ’, * See Addison and Lewis, Quart Revs, -iple, , C~C single bond and the C^C, , to-oxygen bonds. However, the “double bond’M, u" ^ lengths of the respective nitroeen, .O the "triple-bond” length in nitr^yl pe^Wor^On",?'^ Chl°dd' is, ^lyX, nitrosyl chloride molecule has considerable inn’, u ° thlS basis’ 11 may be argued that the, an important contributing form., lC C aracterJ that is, the structure N=0+C1~ is
Page 62 :
46, , The Energies, Lengths, and Orders of Covalent Bonds, , three surrounding carbons. Since carbon must form the equivalent of four, bonds, we may assign to each of the bonds in graphite an order of % or 1.33., Although the bond lengths in benzene, pyrazine, and graphite are, as expected,, , Fig. 2-1. The Structure of Graphite, , less than the C—C single-bond length and greater than the double-bond length,, the relationship between bond order and bond length is not a linear one, (to anticipate this, one would have to expect unusual cooperativeness on, Nature’s part)., However, Pauling has derived a relatively simple relationship between the, bond order y and the bond length r for bonds of order between 1.0 and 2.0.19, In this relationship,, (3y - 3) (rx - r2), r~ri~, , (2y —, , 1), , rx and r2 are the respective bond lengths for single and double bonds. The, observed bond lengths in benzene, pyrazine, and graphite fall very close to the, values predicted by equation (5) (1.39 A for benzene and pyrazine, 1.42 A, for graphite)., It is interesting that the lengths of the carbon-carbon bonds in the organoiron compound ferrocene*0 (II) may be estimated in a similar manner. This, , compound consists of an iron atom “sandwiched” between two cyclopentadienide pentagons, each of the ten carbons being the same distance from the, iron." It is possible to represent the molecule as a combination o struc ures,, it Pauling and Brock way, J. Am. Chem. Soc., 59, 1223 (1937)., " Wilkinson, et al., J. Am Chem. Soc., 74, 2125 (1952)., i/ (a) Eiland and Pepinsky, J. Am. Chem. Soc., 74, 49/1, Nature, 171, 121 (1953)., , DunitZ and Orgel,, K >
Page 63 :
Bond Lengths in Resonance Hybrids, , -, , 47, , each having one carbon per ring bonded to iron with the remaining carbons, forming double bonds. (That this is not the manner in which this molecule is, generally regarded need not concern us now.) Considering, for simplicity,, just one of the two rings, five structures similar to III may be drawn, each, having two double bonds. Each of the carbon-carbon linkages is represented, as a double bond in two of the five structures and as a single bond in three of, o, , the five. Setting the bond order then as 1%, a bond length of 1.41 A could be, predicted from equation (5); this is in excellent agreement with the observed, length. Similar treatment may be applied to a member of the polynuclear, benzenoid hydrocarbons (see Ex. 12) and the bond lengths predicted are some¬, times in surprisingly good agreement with those observed., Table 2-4 lists five of the many additional cases where bond shortening, is observed. For each of these compounds, the resonance interpretation predicts, , Table 2-4. Bond Lengths in Some Resonance Hybrids, Classical Structure, , Typical Additional Canonical Form, , 1.19 A, I, (A) H—C=C—C=C—H <-, , +, H—C=C=C=C—H, , T, 1.36 A, 1.16 A, , I, , (B) N=C—C=N, , : N=C=C=N:, , T, 1.37 A, 1.40 A, , (c) 0°r0, , 4-, , 1.19 A, , 1.38 A, , I, (D) H2C=CH—Cl, , «—, , T, 1.69, , H2C—C=C1:, , I, , A, , H, , 1.40 A, .0 4-, , KS
Page 64 :
48, , -, , The Energies, Lengths, and Orders of Covalent Bonds, , partial double-bond character in the “shortened” bonds. One typical canonical, form, in addition to the classical structure, is shown for each compound., As shown, the central bonds in diacetylene (A) and cyanogen (B), and the, exocyclic bonds in diphenylacetylene (C) are much shorter than the ordinary, C, , C single bonds (1.54 A). In the structures on the right each of these bonds, , is represented as a double bond, indicating that each has some double-bond, character. The same applies also to the C—Cl bond in vinyl chloride (D), o, , (“ordinary” C—Cl bond length, 1.76 A) and to the “single” bonds in furan, (E). For these five compounds, consideration of the structures on the right of, Table 2-4 might allow the anticipation of these bond shortenings22 but the, magnitudes of the shrinkages cannot be calculated. Even though we may draw, the important contributing structures for these compounds, the reasoning ap¬, plied to benzene, graphite, and ferrocene cannot be applied since the various, forms used to describe each of the compounds in Table 2-4 are not equivalent., The forms on the right, showing a “charge separation” are generally called, high-energy forms because electrical work has been presumably expended to, separate the charges to the positions shown; such forms are therefore “less, important” than the classical forms, although how much less important it is, difficult to say. It is, of course, possible to estimate the order of a particular, bond after its length has been determined, but at present there is no simple and, generally applicable way to confirm such an estimate independently. Con¬, siderable work has been carried out on theoretical methods for evaluating bond, orders from quantum-mechanical principles,25 and the bond orders so obtained, may be related to the observed bond lengths by equations somewhat different, from equation (5). These more complex treatments, which have been confined, almost exclusively to hydrocarbons, allow calculation of the bond orders in, compounds such as styrene and butadiene which cannot be directly treated by, inspection of resonance forms. However, for the benzenoid hydrocarbons, bond, lengths predicted by the more refined treatments are not, in general, signifi¬, cantly closer tff the observed values than those predicted by merely inspecting, «, Table, bonds, found, , In considering these structures, it might be noted that the canonical forms on the right of, 2-4 for diacetylene, cyanogen, and diphenylacetylene represent as, ^, which the classical structures represent as triple. Since the leng is, i 15 A', to be essentially identical with the lengths of “normal” triple bonds (C=N,1 5 A, P=C 1 19 A) it might well be asked why the contributions of the forms do not result in, , ££* ofte, due to a, ./bond order. This““l", too great a physica. s.gn.ficance to the individuaclose, indicate is that the carbons jo.ned by the single t.o, ^ ^ ^ ^ p,ace a,, trcanno, casi,y be madc by, o&tl/et/XpenneT, "ZmI ToX/T 158, 306 (1937); Coulson, 7W., Roy. Soc., (London) A 169, 413 (1939).
Page 65 :
Hyperconjugation, , -, , 49, , the bonds in the resonance structures in the manner described above for graphite, and illustrated in Exercise 12 for naphthalene., , Hyperconjugation, A carbon-carbon single bond, if adjacent to a carbon-carbon triple bond, is, almost invariably shorter than the “normal” length, 1.54 A. The C—C bond, lengths in methylacetylene and acetonitrile are, respectively, 1.46 and 1.49 A, —considerably longer than the “single” bonds in diacetylene and cyanogen, (for which the resonance treatment would predict shortening) but undeniably, shorter than the single bonds in ordinary paraffins and olefins. These “shrink¬, ages” are sometimes rationalized by assuming appreciable contribution of struc¬, tures such as IV and V; these are termed hyperconjugated structures in which a shift, , H+, , CH2=C=CH, IV, , H+, , CH2=C=N:, V, , of electron density toward the C—C bond has left the C—H bonds with partial, no-bond character/4 If these contractions were the only evidence of hyperconjugation in molecules, one might justifiably be extremely skeptical about, this effect, for the differences are slight and similar contractions are not observed, for C, , C bonds adjacent to double bonds. However, other evidence that this, , (or a related) effect operates comes from comparing the heats of combu^ion, of isomeric olefins or isomeric acetylenes. As an example, the heat of combustion, of isobutylene is 646.1 kcal per mole; this is significantly less than that for, 1-butene, 649.8 kcal per mole6, even though the two isomers have, of course,, the same number of C-C, C=C, and C-H bonds. To rationalize this and, similar differences using the hyperconjugation concept, we may note that, hyperconjugation of the type indicated involves a C—H bond one carbon, removed from a double or triple bond, that is, a-hydrogen atoms. It is therefore, reasonable that the effects of hyperconjugation become more pronounced as, the number of a-hydrogen atoms increases. Thus, hyperconjugation is more, CH,, important in isobutylene,, , CH2=C, , \, , , with, , six, , a-hydrogens,, , than in, , CH3, and Brown, J., °^:byTofth^^8by Mulliken> Ri<*e,, ent view, see Kreevoy and Eyri^g, iL 7^^, ?’ 2f4 <1957)‘ For a differ^, organic chemistry is discussed in detail by Baker in Hvb, ' ^ .rol® of hyperconjugation in, Oxford, 1952., 7, ker in HyPerconjugation, Oxford University Press,
Page 66 :
50, , -, , The Energies, Lengths, and Orders of Covalent Bonds, , 1-butene, CH2—CH, , CH2, , C’H3, with only two a-hydrogens; in the absence, , of any more important effects, isobutylene should be slightly more stable (hence, should have a lower heat of combustion). Likewise, tetramethylethylene, with, twelve a-hydrogens, should (and does) have a slightly lower heat of combustion, than, , 2-methyl-2-pentene,, , CH3CH2CH=C(CH3)2,, , with eight a-hvdrogens., , H3C+, H3C-C, , /, h3c, , There is evidence also for C—C hyperconjugation (for example, VI, , VI'), , which is, however, generally less important than C—H hyperconjugation (see,, for example, p. 87)., Many differences in the heats of hydrogenation of olefins can be explained,, at least in part, by hyperconjugative effects. As we have already noted, if, bond energies did not vary with environment, all monoolefins would have the, same heat of hydrogenation. However, the heat of hydrogenation of ethylene, (with no a-hydrogens) is 32.8 kcal per mole, that for 1-butene (with two, a-hydrogens) is 30.3 kcal per mole, and that for tetramethylethylene is 26.6, kcal per mole.7 Again, these small but real differences and others like them give, support to the concept of hyperconjugation. In later chapters the reader will, see how hyperconjugation may be called upon to “explain’ a number of trends, in reactivity., , Van der Waals’ Radii'3, When two molecules approach each other without forming a chemical bond,, there is a slight attraction between them due to a mutual distortion of theii, electron clouds. Such a force, called a van der Waals’ force or a dispersion force,36, exerts little effect on the molecules of a gaseous system, for their average separa¬, tion is too great. However, van der Waals’ forces draw together the molecules, of a liquid (although a great deal of “rolling over” obviously occurs); and these, same forces hold molecules in their places in a crystal at low temperatures, where kinetic energies have been cut down. For many polar molecules, van der, Waals’ forces are considerably augmented by dipole-dipole attractions. In the, condensed states, such forces tend to bring the molecules closer together; but, ts See also Pauling, Nature of the Chemical Bond, 2d ed„ Cornell University Press, Ithaca,, , 194°HP/or1fn'elementary and semiquantitative treatment of van der Waals* forces sec Rice,, Electronic*Structure and Chermcal Bonding, McGraw-Hill Book Co, New Yc, k, 1940,, pp. 354-358. For a more extended treatment, see Rowlmson, Quart. Revs., VIII, 168 QJ5 ).
Page 67 :
Van der Waals' Radii, , 51, , they must not approach each other too closely, for a repulsion sets in when the, respective electron clouds overlap appreciably. For each atom on the outside, of a molecule there is an optimum distance from the nucleus beyond which the, electron cloud of a nonbonded atom cannot easily advance; this distance is, called the van der Waals’ radius. From early crystallographic studies*7 the, van der Waals radii of chlorine and bromine were estimated to be, respectively,, o, , o, , 1.80 A and 2.02 A, these values being just half the shortest distance between, halogen atoms of adjacent molecules in solid /3-benzene hexachloride (VII) and in, solid hexabromobenzene. For organic chemistry, the van der Waals’ radii of, , the methyl group and methylene group are of considerable importance. These, two values, which should be roughly the same (Ex. lib), represent half of the, shortest distance between the carbon atoms of methyl or methylene groups in, adjacent molecules; they have been found by a large number of investigations, to approach 2.0 A.*s The van der Waals’ radii of oxygen, nitrogen, or sulfur, may be obtained by subtracting this figure from intermolecular carbon-to-oxygen,, carbon-to-mtrogen, or carbon-to-sulfur distances in crystals where there is, contact between the respective “hetero-atom” of one molecule and a methyl, or methylene group of another., Van der Waals’ radii of atoms of greatest interest to the organic chemist, are ltsted in Table 2-5 « Since there is a far wider variation in intermolecular, , Table 2-5. Van der Waals’ Radii, H, , 1.2 A, , F, , 1.4 A, , N, , 1.5, , Cl, , O, , 1 .4, , Br, , 1.8, 2.0, , S, , 1.9, , I, , 2.2, , *, , nonbonded than in bonded distances between two given atoms amnnv, , to ,hC nearest ai A’ PaulingTal, , -*7 Fnr o ___, , /., , %,, , - ., , Chtm 7-«>, , e™lec„u,ar, ■, , in octamethyl cycle,etrasiloxane, StejfinkhT'aT, tetramethyl pyraalne, Cromcr, ’ ^, , /, , methyl-,o-methyl distance, ^5
Page 68 :
52, , -, , The Energies, Lengths, and Orders of Covalent Bonds, , bond as the ion does in all directions.”*9 It should be noted that the van der, Waals radius is the maximum nonbonded radius of an atom—the distance between, the nucleus and the effective “outside” of the atom at a point directly opposite, the site of bonding. It is the radius of the atom as set off by a line forming an, angle of 180° with the bond direction. The atomic radius set off by a line, through the nucleus at any other angle must be greater than the covalent radius, but less than the van der Waals’ radius. This may be seen from Figure 2-2,, o, , rc = covalent radius of carbon (= 0.77 A), o, , rHr = covalent radius of bromine (= 1.14 A), r'Br = van der Waals’ radius of bromine, (= 2.0 A), , Fig. 2-2. Covalent and van der Waals’ Radii, , which shows two of the C—Br bonds in a compound such as CBr4. Often two, atoms bound to a common atom, but not to each other, are separated by a, distance significantly less than the sum of their van der Waals’ radii. For exam¬, ple, the nuclei of the bromine atoms in CBr4 (Br—C-—Br bond angle 109°), are separated by a distance of 3.1 A, almost 25 percent less than twice the van, der Waals’ radius of bromine. From Figure 2-2, we see also what is meant by, the phrase “pear-shaped atoms,” which is sometimes used to describe hydrogen, and halogen atoms in organic molecules., Along with the radii in Table 2-5, one additional value should be noted., t9 Some care, however, should be exercised in using ionic radii as van der Waals radii. To, cite a single example, the shortest intermolecular selenium-to-oxygen distance in transethanediseleninic anhydride (VIII) (Gould and Post, J. Am. Chem. Soc., 78, 5161 (1956)) is, , Se, , ?hAo, O, , /, , VIII, only 2 70 A, much less than 3.40 A, the sum of the ionic radii of selenium and oxygen. While, the radius of the selenide ion may be quite close to the van der Waals radius of selenium i, compound such as dimethyl selenide, it is apparently not appropriate for the selenium a om, ^, i vtti Thic discreDancv may be rationalized by supposing that the positive, m compound VIII I his discrepancy may, electron cloud in more tightly than, , is that the selenium atom i, ,, different van der Waals’ radii is less surpnsselenide. That the two types of selenium atom have diiterent van aci, ing than if they were found to have the same radii.
Page 69 :
Van der Waals' Radii, , 53, , This distance, about 1.85 A, is the “half thickness” of aromatic molecules—, that is, half the distance between parallel benzene rings of adjacent molecules, in the crystals of such compounds as naphthalene and anthracene. Note that, this value is much larger than the van der Waals’ radii of any of the nonmetals, of the first eight-membered period, indicating that the 7r electrons extending, above and below the planes of aromatic rings impart substantial thicknesses, to such rings. In graphite, however, the layers of rings (Fig. 2-2) are unexo, , pectedly close (3.4 A), suggesting that the layers are held together, not only by, van der Waals’ attraction, but by some auxiliary force, the nature of which, cannot at present be simply described., , EXERCISES, , FOR, , CHAPTER, , 2, , 1. The equilibrium constant for the reaction Cl2 —> 2C1 is 0.1 at 1800° K but only 10“10, at 800 K. Estimate the Cl, Table 2-1., , Cl bond energy and compare it with the value in, , 2. The formation of NH3 gas from the elements N2 and H2 liberates 11 kcal per mole of, NH3. The formation of hydrazine, NH2—NH2, from the elements is slightly endo¬, thermic, requiring 10 kcal per mole. The H—H bond energy is 104 kcal, whereas, the N=N triple bond energy is 170 kcal., (a) Calculate the N—H bond energy., (b) Calculate the N—N single-bond energy., (c) Using appropriate data from Table 2-1, calculate the heat released in the, reaction:, H2N-NH2 + 2H202 -> 4H20(gas) + N2, 3. The heats of hydrogenation of cyclohexene and benzene are, respectively, 28.6 and, 4y.« kcal per mole., (a) Calculate the C=C bond energy in cyclohexene., (b) Estimate the resonance energy of benzene., , 4' StphilrShTT^f °l alTt 4?Cal Per ?ram at°m in the heat, atomization of, the Wl, atomization is used in the calculation of bond, nergies, the following uncertainties should arise:, (a) 11 kcal for the C—H bond energy., (b) 22 kcal for the C—C bond energy., (c) 66 kcal for the C^C bond energy., (d) 33 kcal for the C=N bond energy., 5. Show that the uncertainty in the value for, not affect the following values:, , e, f atomiza'lon of graphite does, , (a) The difference between the heats of formation of C,H4 and C,H, (b) The resonance energy of benzene., 6', , 6- (a) Show that the arithmetic mean of two bond enemies F, , » a v, , greater than, or equal to, the geometric mean of teethe,, , *"
Page 70 : 54, , The Energies, Lengths, and Orders of Covalent Bonds, (b) What percent difference must there be between the values of two bond energies, before their arithmetic mean exceeds their geometric mean by more than, 10 percent of the latter?, (c) Show that by using the arithmetic mean, rather than the geometric mean, ol, bond energies Ea-a and Eb-b to estimate the energy of the hypothetical “com¬, pletely covalent bond" between A and B, it becomes possible to calculate the, “extra ionic resonance energy” of bond A—B merely by knowing the heat of, formation of the A—B molecule from the elements (that is, without knowing the, heats of atomization of A—A or B—B)., (d) Why is an estimate such as described in (c) possible for the “extra ionic resonance, energy” of the H—Cl bond but not for that of the N—H bond?, , 7. (a) From the data in Table 2-1, calculate the electronegativity difference between, H and Cl. Check with Table 2-2., (b) Using Table 2-3 and the Schomaker-Stevenson relationship, estimate the inter¬, atomic distance in HC1., 8. (a) Calculate the length of the Si—F bond in SiF4 using the Huggins relationship,, (b) Estimate the (nonbonded) fluorine-to-fluorine distance in the SiF4 molecule., (Hint: See Chap. 1, Ex. 7.), y, , Predict which member of each of the following pairs of compounds has the highci, resonance energy and justify your choice:, (a) Anthracene or phenanthrene?, (b) Ammonium acetate or acetamide?, (c) Cyclooctatetraene or styrene?, , (e) Benzene or hexamethylbenzene?, (f) /?-Benzoquinone or benzaldehyde?, , JO, or, , OH, , ,0. Arrange the following compounds in the order of .heir heats of hydrogenation:, , <a) Mrbon^n^^chjhwcan^,'carbon^^benzene^'carbon'hQ CF^emra! carbon in
Page 71 :
Exercises for Chapter 2, , 55, , CH2=C=CH2, central carbon in CH=C—CH2—C=CH, silicon in SiH4,, silicon in SiCl4, sulfur in SF6., (b) Explain why the van der Waals’ radius of the methylene group is about equal, to that for the methyl group even though the latter has one additional hydrogen, atom., 12. Draw pictures of the three Kekule forms of naphthalene. Assuming the three forms, are of equal importance, and neglecting high-energy forms, estimate the bond order, of the bonds designated as 1, 2, and 3. Estimate the lengths of each of the numbered, bonds and compare these lengths with the observed values (bond No. 1—1.43, bond No. 2—1.37 A; bond No. 3—1.40 A)., , 3, , A;
Page 72 :
CHAPTER 3, , Dipole Moments and Spectral, Studies, , Inorganic chemists sometimes refer, , to the “revolution" that has taken place, , in their field during the last forty years. These years have seen the development, of the modern methods of structural investigation; and today the structures of, inorganic compounds are the subject for systematic study, whereas before 1910, they were largely matters for speculation. Although the same methods have, inevitably affected organic chemistry, they have not revolutionized it in quite, the same way. Excellent structure determinations for thousands of organic, compounds have been based solely on the study of their chemistry, and physical, methods of investigation have often served to show that the older concepts of, molecular geometry of organic compounds were, in the large part, correct., However, two types of study, formerly carried out almost exclusively in physical, chemistry laboratories, have been particularly valuable to organic chemists., The first of these, the study of electric dipole moments, is occasionally, used for settling difficult questions concerning molecular geometry, but is more, generally used to provide hints as to the distribution of electron density in, molecules. Dipole moments have influenced organic thinking, but the measure¬, ment of such moments is still generally regarded as being in the province o, certain physical chemists who are especially interested in the subject., With molecular spectra, it is a different story. One of the most inking, differences between the organic publications of 1920 and th* of 1955 is the, very prominent role that spectral data have come to assume. While it .s still, possible for an organic research chemist to do good work without access to, spectral facilities, it is becoming increasingly rare. In many laboratories t, infrared spectrum of a new compound has come to be a routine item m its, characterization, just as are its melting point and chemical analysis., , 56
Page 73 :
Definitions, , 57, , A large area of information arising from spectral studies is of only passing, interest in the present text. We shall not be concerned here with the detection, of functional groups by infrared analysis, with the identification of compounds,, nor with the analysis of mixtures by infrared or ultraviolet spectroscopy. It is, also beyond the scope of this text to treat quantitatively the relationships be¬, tween the electronically excited states of molecules and their ultraviolet spectra., There are, however, a number of points of structural interest arising from, spectral studies with which we shall deal briefly after a somewhat more detailed, treatment of electric dipole moments. The reader, however, is cautioned against, evaluating the relative importance of these two types of investigations from the, lengths of the sections devoted to each., , Part I—DIPOLE MOMENTS'**, Definitions, \\c have already (p., , 11) referred to molecules with “positive ends” and, , negative ends” as dipoles. One of the simplest of these dipoles is the HC1, molecule, in which the hydrogen atom is the positive end and the chlorine atom, obviously the negative. Similarly, the region between the two hydrogen atoms, of the water molecule and the region between the three hydrogen atoms in, the ammonia molecule may be considered as the positive ends of these molecules, w ereas the respective heavier atoms are at the negative ends. It is sometimes, convenient to regard the positive or negative charges of a polyatomic molecule, as concentrated at single points in space. These points are then, in effect,, indT 7h°i f aVry °f P°SUiVe (°r neSative> charS<=;, indicated below for the H,G and CH2C12 molecules., , such centers of charge are, , centers of positive charge, , H, , H, sO", , C\'~ - —.cJ~^, ^, , V, , 1, , s, , A, , Cl, centers of negative charge, Physical 'C/Z^vTi,, , Longmans^, , A" Mmai, , Md«h- ^ £ Lo°n’d^31954’ PP- 287“541;, , £X—t, , Le-, , %, , crences are not given are taken from this worf "({Va'T'5' D,p°le m°mcnts, which, work. (6) Appllcati0ns of dipole moments to
Page 74 :
58, , -, , Dipole Moments and Spectral Studies, , The electric dipole moment is abbreviated ju. For a molecule (in which the, positive and negative charges must be equal in magnitude), /z is defined, (1), , M = zd, , where z is the value of the charge at either center and d is the distance between, centers., Suppose, for example, that the hydrogen chloride molecule were com¬, pletely ionic. It would then consist of a hydrogen ion (whose charge would be, the same magnitude as that of the electron, 4.8 X 10“10 esu) at one end of the, molecule and a chloride ion (with equal and opposite charge) at the other end,, 1.27 A away. The dipole moment would then be 6.1 X 10-10 A-esu units, more, conveniently represented 6.1 Debye units. Experimentally, the dipole moment, of HC1 has been found to be only 1.03 Debye, showing that the hydrogen, electron has not been transferred cleanly to the chlorine but has been smeared, between the atoms, perhaps a little thicker near the chlorine.3 Often, the pres¬, ence of a dipole is represented on the structural formula of a molecule by an, arrow, pointing from what is believed to be the positive end to the negative, end—for example,, , Cl, Compounds such as C02, CH4, BF3, benzene, and HgCl2, in which the centers, of positive and negative charge coincide, have no dipole moment; that is, they, are nonpolar., , Polarization, In determining a dipole moment, the sample is dissolved in a nonpolar solvent, or vaporized, then placed between two oppositely charged plates, becoming, in, effect, the dielectric material of a large electric condenser. As expected, the dipoles, structural organic chemistry are also discussed by Sutton in Braude and‘L 373-420, Hon of Organic Structures by Physical Methods, Academic Press, nc„ New York1955 pp. 373 420, • One is perhaps tempted to divide the observed dipole moment by that cdcubtedLtor a, completely, , ionic, , structure and, , done £ t”" ee, , 1940, p. 46.) Implicit in such an htterpretatton.s the a*umpUon, bond between two different atoms s ou, , res, , *', , tricPdis,ributio„ of, , neither verified nor disproved experi-, , ^/r^'Oxfo^^dniversiry, , p7e„ ;f u** by., bond, may have assented with them a substantia, dipole moment., , —
Page 75 :
Dielectric Constants and Dipole Moments, , 59, , tend to orient themselves so that the negative ends point toward the positive, plate and the positive ends toward the negative plate. (This orientation is, opposed by the random motion of the molecules, the “randomness” increasing, with temperature.) The substance, which is said to be thus polarized, increases, the capacitance of the condenser; that is, there is a slight increase of positive and, negative charge at the respective plates although the potential is kept the same., Although the circuits and techniques used in measuring the capacitances of, condensers will not be described here, we might note that in practically all, such measurements, alternating currents are used. The polarity of the con¬, denser plates reverses itself from 106 to 107 times each second, and any polar, molecules present between the plates rotate back and forth in an attempt to, keep up with the alternating polarity., In addition, distortions within the molecules themselves occur in the pres¬, ence of an electric field. Electron clouds, being negatively charged, are pulled, slightly toward the positive plate (electron polarization). If the polarity of the, plates is alternated, the electron clouds will shift, first in one direction, then in, the opposite direction. This polarization occurs both for polar and nonpolar, molecules. Less important, but sometimes appreciable, is atom polarization, in¬, volving slight twists or stretchings of bonds between unlike atoms such that, the more electronegative atoms suffer slight displacements in one direction,, the more electropositive atoms slight relative displacements in the other., Both electron and atom polarization effects may be grouped together as, induced polarizations, , that is, separations of charge that occur only under the, , influence of electric fields. Unlike the orientation of permanent dipoles in an, eectnc field (often called orientation polarization), induced polarization is senerally not affected by thermal agitation., , Dielectric Constants and Dipole Moments, A perfect vacuum between condenser plates yields the lowest possible capaci, tance. The dulectnc constant of a material is defined simply as the capacitance, of a condenser having the given material between chanted' nl„,, CapaCUanCe, , must be attributed to induced polarization., (i93;wat„„,,., , $, , n°nPolar, , (London), m> 569 (193]); i43> 55g (i934); [56 )3o
Page 76 :
60, , -, , Dipole Moments and Spectral Studies, , The relationship between the dipole moment of a dilute material and its, dielectric constant (the Clausius-Mosotti-Debye relationship) is5, (D - 1)3M, (D + 2)4 rd, , Ncc + N, , (2), , 3 kT, , where, D = the dielectric constant, , a = the polarizability, , M = molecular weight, , fi = the dipole moment, , d = density, , k = Boltzmann’s constant, , N = Avogadro’s number, , T = the absolute temperature, , Note that one of the two terms on the right side of Equation (2) is temperaturedependent whereas the other is not. The polarizability per molecule, a, which, includes effects of both electron polarization and atom polarization, is a con¬, stant for a given substance; this may sometimes be obtained from its refractive, index toward infrared radiation, or, more often, is estimated from its refractive, index toward visible radiation. In cases where the substance is stable in the, vapor state over a large range in temperature, both a and /jl2 can be estimated, from measurements of the dielectric constant at two (or more) temperatuies., The value of /jl is obviously zero if the quotient, , (D - 1 )M , „ , ,, ., ^, (cahed the molar, , polarization) is independent of temperature., When the material to be tested cannot conveniently be vaporized, it may, be dissolved in a nonpolar solvent, and the dielectric constants of the solutions, may be measured. Under these conditions, handling of the data is somewhat, more complicated, for there is often interaction between solvent and solute.5, Generally the dipole moment of a substance in the vapor state will differ, slightly from its dipole moment in solution. The dipole moments of simple, organic compounds lie in the range 0 to 6 D. The following values, which are, taken from Smythe2 and which refer to the vapor state, are representative., Toluene, Monomethylamine, Methyl chloride, Acetone, Nitrobenzene, , 0.37 D, 1.28, 1 .87, 2.85, 4.21, , . For the derivation of this equation, together with a number of relat.onsh.ps for con¬, centrated systems and their critical evaluation see Ref. 2(a , PR « com, ' Recently, the dipo e moments of “, obtained by the study of the vana ., , ds have been, , ("„ the microwave region), description of this method, which may, , moments'of organic compounds that had, been obtained in this way before 1953.
Page 77 :
Bond Moments and Group Moments, , 61, , Bond Moments and Group Moments, Dipole moments are generally handled as vectors,' and the dipole moment of a, molecule may be regarded as being the resultant of individual vectors associated, with each of the bonds within the molecule (and therefore called bond moments)., For a diatomic molecule such as HC1, the dipole moment is the same as the, bond moment, whereas molecules with three or more atoms must be treated, by vector addition, in which case it is necessary to know or to estimate, interbond angles in the molecule. Evaluations of the O—H and C—O bond, moments from the measured dipole moments of water and methanol (1.84, and 1.69 D, respectively) are illustrated in Figure 3-1. Note that in (A) the, m=1.84, , bond angle 104°, (A), 1.84 = 2/iq-h sin, , 180° - 104°, , 1.69 =, \/(1.50 — /iC_0 cos 70°)2 + (nc-o sin 70°)2, , Mo-u =, , '.50 D, , Mc-o, , = 1.44 D, , Fig. 3-1. Calculation of O—H and C—O Bond Moments, , observed dipole moment of the water molecule has been “resolved” into two, “’,heUb ln mff‘“de and maki"S “ angle with each other correspond., oLh bond, m, ang‘e- EaCH °f thCSe COmp°nent, - 'hus an, H bond moment (designated, After the value of Mo_„ has been deter, eXa’teTheT 7^7^ ““ °bSerVed dip°le, , of ethanol (B) to, , ■rated in Filnel ,, gure 3-1 assumes that in both the O—H and o, , °f me,hano1 “‘usn t, j, , atom is the more negative and that, therefore the bond, , 6^, , added in the same manner as those in water"b ^, .gnored the C-H bonds and moments associated with them The C, , H K, , d, , moment is known to be relatively small (about 0 3 to 0 4 m' nd th, siderable contention among workers in the fi m, ., there ls conwhether the carbon or the hydrogen at, ■ ., W aS t0 ltS ilrectlm (that is,, , For a short, g n atom is the more negative)/ Conveniently, Wiley and Sons,, Quart. Revs. (London), 2, 383 (1948)P, , ', , " **-* OWy, W», 3 more detaiIed treatment, see Gent,
Page 78 :
62, , -, , Dipole Moments and Spectral Studies, , conclusions based upon the discussions to follow will not be changed by assuming, blandly that this moment is vanishingly small. A representative set of bond, moments, obtained much in the same manner as Mc-o and mo-h? is, bond, moment (D), , H—N, 1.3, , H—O, 1.5, , H—S, 0.7, , C—N, 1.0, , C—O, 1.2, , C—Cl, 1.9, , C—Br, 1.8, , C=0, 2.7, , As was the case with bond energies (Chap. 2), bond moments obtained from, one compound may often be applied to other compounds, but, again, the, exceptional cases are sometimes the most interesting., Since the organic chemist tends to regard compounds as carbon chains or, aromatic rings to which one or more functional groups are attached, it is often, convenient to assign moments to groups as a whole. These group moments, are then the sum of the individual bond moments, possibly modified by any, Table 3-1. Group Moments9, Moment (D), , Moment (D), Group, , Group, Alkyl, , Aryl, , —och3, —nh2, , 1.22, , 1.35, , —COOH, , 1.2, , 1.48, , —CHO, , —Br, , 2.01, 2.05, 1.69, , 1.73, 1.70, 1.4, , —COCH3, , —Cl, —OH, , —NO 2, —C=N, , Alkyl, , Aryl, , 1.68, 2.73, 2.78, 3.68, 4.00, , 1.73, 2.76, 3.00, 4.21, 4.39, , . These values are taken from'Smyth(Ref. 2, p. 253). The aryl moments stem from measure¬, ments of benzene derivatives, the alkyl moments from ethane derivatives. The moments for the, _COOH and —CHO groups refer to solutions (in benzene), whereas the remainder, taken from studies in the vapor phase., ., ., Group moments for substituents on the higher aliphatic hydrocarbons lie very close, the “alkyl” values in the table, whereas group moments obtained from met y, , eriva iv es, , to be about 0.1 D lower than those given., , special interaction which is typical of the group. In general (for reasons that, will soon be evident) the moment of a group bound to an aliphatic carbon, is different from that of the same group bound to an aromatic ring., Table 3-1, two moments are given for each group. Group moments, , us, in, ave, , particularly helpful in considering aromatic compounds; here, the vector treat¬, ment is especially easy if the direction of the group moment “^sameas t a, of the bond joining the group to the ring (for example,, , or, , * •, , Figure 3-2 illustrates the use of group moments to est.mate the dipole moment, of rn-bromonitrobenzene.
Page 79 :
Hydrocarbons, , *, , 63, , n = y/(4.21 - 1.73 cos 60°)2 + (1.73 sin-60°)2, = 3.66 D, 1.73, , Fig. 3-2. Addition of Group Moments in m-BrCeH4N02, , The calculated moment, 3.66 D, is considerably greater than the observed mo¬, ment 3.4 D (for benzene solution). However, it must be remembered that the, bond moments used are those taken from gas-phase measurements. The dipole, moment of nitrobenzene in benzene solution is about 5 percent less than the, value for the vapor, whereas the moment for bromobenzene in benzene is about, 10 percent less than for the vapor. If we use instead the bond moments calculated, for solutions in benzene (1.54 and 3.98 D), the calculated dipole moment, becomes 3.48 D, in rather good agreement with the observed value., , Hydrocarbons, The zero (or almost zero) dipole moment of each of the paraffin hydrocarbons, suggests that, in the absence of complicating effects, the dipole moment of a, compound should not change if a methyl group is substituted for an aliphatic, hydrogen, that is, that the C, , H and C—CH3 moments are equivalent., , The small but definite dipole moment (0.4 D) of toluene then leads to, two questions, (a) which end of the toluene molecule is the “negative” end?, and (b) what is the nature of the “complicating effect” in this molecule? The, first question must be answered by comparing the dipole moments of two or, more compounds (stnce the study of a single compound may yield only the mag., ", ., “S m°ment’ never the direction). The reasoning is as follows: If two, negative (or two positive) groups are placed, say, at the para positions of a, benzene ring, the two group moments will tend to oppose each other just as, ::1 Tf°,hmeednirPeUc Un? “'T** CndS °f a r°Pe °r PU5hi"S a. opposite ends f a, box. If the directions of the group moments are the same as those of the bond,, , zz*;z°z, , loS^Ze!7“'may be °btained simp,y, , aPt0r:,hheer, , 1. b b^T; “fc', , T “ ~, , puU”, -V 1- obtained „
Page 80 :
64, , -, , Dipole Moments and Spectral Studies, , The positive character of the side chain in toluene may be cited as evidence, for the contribution of hyperconjugated structures (Chap. 2) of the type I' (eight, others may be drawn). A similar structure, IT, may be drawn for propylene to, rationalize its small but appreciable dipole moment (also 0.4 D). Since the, , ch3-ch=ch2, , > +H ch2=ch-ch2_, , H = 0.4 D, , r, , II, , II', , effect of the benzene ring is the same as that of the vinyl group, one would, expect the dipole moment of styrene to be zero or very close to zero (which is, the case). The dipole moment of methylacetylene is 0.74 D, suggesting that in, this molecule, hyperconjugation plays an even more important role—a con¬, clusion consistent with the marked shortening of the C—C single bond in this, compound, , (p. 49)., , Here, however, another factor probably should be con¬, , sidered, for it is now thought (Chap. 7) that, quite apart from hyperconjugation, effects, triple-bonded carbon atoms are more electronegative than single- or, double-bonded carbons., Two of the most interesting of the “polar hydrocarbons” which have been, the subject of recent theoretical treatments5 are 6,6-diphenylfulvene (III) and, azulene (IV). Substitution of chlorine atoms at the para positions of the benzene, , rings in diphenylfulvene lowers the dipole moment to 0.7 D, showing that the, cyclopentadiene “end” of the molecule is negative; and it is thought from, chemical evidence'0 that the same is true for azulene. This indicates a substantial drift of electron density into the five-membered rings of these two hydro¬, carbons from the remainder of the molecules, a charge distribution that may, 9 Sec, for example, Berthier J. ^ ^, * For example, treatment of azulene with, , ,, , UH&N_N, , resu, , in substitution on thc fivethat this end Gf, , membered ring (Treibs and Ziegenbein, Ann. 586, 149 (1W),, the molecule is more “electron-rich” than the seven-membered ring.
Page 81 :
Halogen Derivatives, , 65, , be indicated by contributions of forms such as III' and IV'. The cyclopentadiene, ring acquires part of an extra electron which brings its tc electron total nearer, to six. As we shall see in Chapter 11, conjugated cyclic systems containing six, r electrons have special stability. (In structure IV', the seven-membered ring, in azulene also has six 7r electrons.), , Halogen Derivatives, Dipole moments (vapor state) for some chloro compounds are listed in Table, 3-2. It is interesting that the moments of the corresponding bromo and iodo, compounds tend to lie within 0.1 D of the values for the chloro compounds,, the differences in electronegativity being offset partially by the increase in, molecular dimensions and partially perhaps by more subtle factors., , Table 3-2. Dipole Moments of Some Chloro Compounds, M, CH3C1, , 1.87 D, , C2H5C1, n-C3H7Cl, n-C5HnCl, , 2.05, 2.10, 2.12, , CH2=CHC1, HC=CC1, , 0.44, , C6H5C1, , 1.70, , o-C1C6H4C1, , 2.53, , 1.44 D, , The substitution of a methyl group for a hydrogen atom in alkyl halides, having three or more carbon atoms results (as we should expect) in practically no, change in dipole moment. However the ethyl halides have dipole moments 0.1, to 0.3 £1 greater than those of the respective methyl halides, whereas the n-propyl, halides have moments about 0.1 D greater than the respective ethyl halides., (Similar but less pronounced effects are observed for aliphatic nitro and cyano, compounds.) One current interpretation of this trend (which has recently, • ee" eX.P“ded lnto a factory quantitative treatment)" may be summarized, in the following argument: (a) At the carbon-halogen linkage, the halogen, a °m ‘S more neSative, the carbon therefore more positive, (b) Electrons from, atoms or groups bound to the halogena.ed carbon will be drawn slightly toward, he positive charge, resulting in a greater separation of positive and negmive, ,h, Kan Trease ” dip°le m°ment)- V The valence electrons assor¬, ted with a bound methyl group are more easily distorted by the nearby partial, , ated, , positive Charge than are the valence electrons of a bound hydrogen mom (tha, ,, , e methyl group ,s more polarizable than the hydrogen atom), (d) TlLre-
Page 82 :
66, , -, , Dipole Moments and Spectral Studies, , fore a methyl group attached to a halogenated carbon is more effective in, removing the positive charge from the halogenated carbon, and thus in increas¬, ing the dipole moment, (e) This effect, like a number of electrostatic effects, that we shall encounter (Chap. 7), falls off very rapidly as the distance from the, 4-, , —, , “primary dipole” (that is, C—Cl) increases, (f) Therefore, the change in dipole, moment in passing from ethyl to propyl halides is small and further changes on, increasing the chain length are negligible., The relatively low dipole moments of vinyl chloride, chloroacetylene, and, chlorobenzene (and the correspondingly low moments of the respective bromo, and iodo compounds) also indicate that another effect of importance has come, into play. In vinyl chloride, one of the p “dumbbells” of unshared electrons, on the chlorine atom is parallel to the pair of 7r-electron clouds of the double, bond, allowing delocalization of part of the excess negative charge over the, three-atom system and a corresponding decrease in dipole moment. (This, would mean also that the C—Cl bond had acquired partial double-bond, character—a situation we have already inferred from its short length, p. 47.), In chloroacetylene, both p orbitals of the chlorine atom are parallel to the ir, orbitals associated with the triple bond, and a further spreading of the excess, negative charge is possible. This lowering of moment is often termed a “reso¬, nance effect” because it may be described by the use of extra structures such as, V' for the halides involved. Because of the same effect, we would expect the, —, , -f", , HC=C—Cl <-> HC=C=C1:, , (V), , (vo, , H = 0.44 D, dipole moments of the aryl halides to be lower than those of the corresponding, alkyl compounds (indeed, that is why Table 3-1 gives two separate group, moments for halo substituents); however, it is somewhat surprising that this, effect for aryl compounds is much less marked than for vinyl compounds., , Monosubstituted Benzene Derivatives, The relatively low dipole moment of chlorobenzene (as well as the low values, for the other monohalogenated benzenes) may thus be explained by assuming, that forms such as VI' contribute to the structure of the molecule. Form, is similar to form I'—used to explain the dipole moment of toluene, , since in, , both the benzene ring has assumed a partial negative charge. However, it, should be noted from Table 3-1 that the moments of nitrobenzene and benzonitrile are considerably higher than those of their aliphatic analogs. Since there, is no doubt in our mind at present that both the nitro and the cyano groups are
Page 83 :
Monosubstituted Benzene Derivatives, , 67, , negative with respect to carbon, it appears then that the action of the benzene, rings in nitrobenzene and benzonitrile has made the substituents more negative;, that is, that electron density has been drawn out of the benzene ring and has, been spread somewhat more thickly over the substituents, leaving the ring with, a partial positive charge. Such an effect, which is represented by resonance, forms such as VII' and VIII', serves to emphasize what may be termed the, “electronic versatility” of the conjugated system of bonds of the benzene ring., , Electron density may be “pushed into” the ring from halogen atoms, from nitro¬, gen atoms of amino groups, or from oxygen atoms of —OH groups. On the, other hand, electron density may be “drawn out” of the ring by a —0=0,, , C, , N’ °r, , SrouP- (In Chap. 11 we shall see why these two classes of, , O, substituents correspond, as the reader must surely have noticed, to the «ortho, , groups”and the, , of, amines to about 1 6 D fin !, , ZZT X, , orgl, , of«, , lowniniof the dipoie —, } ^ the m°ment of ar°matic, , Wh™, , He in the range “ o to , 4 D), —nts, 3-3, the reJaot?:;^^;;"-of an.iine in Figure, to the N-G bond moment b Electron, , K a'm°St direC“y °PP°Sed, , »«»i,Z«L. Scy;
Page 84 :
68, , -, , Dipole Moments and Spectral Studies, , Fig. 3-3. The Dipole Moment of Aniline, , moments of nitrobenzene and chlorobenzene should be increased by introduction, of a —NH2 group para to these negative substituents; Table 3-3 shows this to, be the case. Similar arguments apply to dimethylaniline (/* = 1.6 D) and its, derivatives., , Disubstituted Benzene Derivatives, Further points of interest arise from consideration of the moments of disubsti¬, tuted benzenes. In Table 3-3, the observed values for a number of such moments, (in benzene solution) are compared with those calculated by vector addition,, using the moments of the monosubstituted benzenes on the left of the tabled', Note that in spite of the partial double character of the bond linking the nitrogen, atom in aniline to the benzene ring (see structure IX), the —NH2 group is not, quite coplanar with the ring (if it were, the moment of /;-phenylenediamine, would be zero instead of 1.5 D). The calculation of moments for benzene, derivatives having only halo, nitro, or cyano substituents is more straightforward, than for substituted anilines, phenols, and anisoles for which it is necessary to, know not only the group moments but also the moment angles, , that is, the angles, , between the bond moments and the bonds linking the —NH2, —OH, or, —OCH3 substituents to the ring. These angles have been evaluated/3 princi¬, pally by considering substituted toluenes, and are used in calculating the, moments of /?-nitrophenol and of jfr-nitro- and /?-chloroaniline in the, , table., , The treatment of ortho- and ^^-substituted phenols and anilines is even moie, complicated and the results subject to considerable doubt; the moments change, as the —OH or —NH2 groups rotate with respect to the ring, and it is necessary, to make somewhat arbitrary assumptions as to the ease of such rotation and the, possibility of favored configurations., Note that the agreement between calculated and observed moments is, quite good for the meta- and most of the para- disubstituted benzenes but much, poorer for the ortho compounds. In calculating moments for the ortho compounds,, it is generally assumed that the bonds holding the two substituents to the ben¬, zene ring lie in the plane of the ring and make an angle of 60° with each other., ii The values for monosubstituted benzenes refer to solutions in benzene and therefore, differ slightly from the group moments in Table 3-1 .which refer to the gas phase., 13 Marsdcn and Sutton, J. Chem. Soc., 1936, 59).
Page 85 :
69, , Disubstituted Benzene Derivatives, , This, however, is an oversimplification which ignores the possibility that ortho, substituents, because of their small separation in space, may interfere with, each other sterically. We now know, for example, that the halogen atoms in, o-dichloro and o-dibromobenzene are pushed out of the plane of the ring by, about 18°, one above the ring, one below.*-* It is probable that similar inter', ference occurs in other ortho compounds, increasing the angle between bond, moments and accounting, at least in part, for the differences. The situation is, further complicated by mutual distortion (polarization) of the electron clouds, associated with the two ortho substituents; although we cannot discuss the latter, , Table 3-3. Dipole Moments of Some Substituted Benzenes, MonoDisubstituted, , substituted, , 0, , m, , P, , Sub¬, stituent, , D, , Sub¬, , Ob-, , Calcu-, , Ob-, , Calcu-, , Ob-, , Calcu-, , stituents, , served, , lated, , served, , lated, , served, , lated, , Cl, , 1.58, , Cl, Cl, , 2.27, , 2.74, , 1.48, , 1.58, , 0, , 0, , Br, , 1.54, , Br, Br, , 2.1, , 2.67, , 1.46, , 1.54, , 0, , 0, , N02, , 3.98, , no2, no2, , 6.00, , 6.90, , 3.89, , 3.98, , NH2, , 1.53, , 0, , 0, , NHo, NH2, , 1.45, , 1.79, , 1.80, , OH, , 1.6, , 1.5, , no2, Cl, , 4.1, , 4.97, , 3.4, , 3.47, , 2.50, , 2.40, , no2, nh2, , 4.24, , 3.66, , 4.94, , 4.72, , Cl, nh2, , 6.2, , 5.17, , 1.77, , 1.71, , 2.27, , 2.30, , 5.04, , 4.34, , no2, OH, ., , effect m detail here,' it is clear that it too should reduce the reliability of dipole, moments that are calculated from simple vector addition., The two most interesting discrepancies in Table 3-3 are associated with, , —e, and these two compound; fJn^ £m^^, , sma„w.d ,d H_, ,, , ^
Page 86 :
70, , -, , Dipole Moments and Spectral Studies, , of such resonance effects on dipole moments. Unexpectedly large moments are, observed, , also for />-nitrophenol and />-aminobenzonitrile; for these, forms, , similar to X' and XI' may be drawn. The same effect is almost certainly im-, , X, , X', , XI, , XI', , portant for o-nitroaniline which, it will be observed, is the only ortho derivative, in Table 3-3 for which the observed dipole moment is significantly greater than, that calculated. On the other hand, the observed and calculated moments of, m-nitroaniline are in much better agreement, and a few attempts with pencil, and paper should convince the reader that quinonoid structures cannot be, drawn for this compound.16, , Steric Inhibition of Resonance, Structures such as X' take on added reality when moments of derivatives of, durene are compared with those of the corresponding benzene derivatives., Typically, the moment of nitroaminodurene (XII) is only 4.98 D in benzene,, about 1.2 D less than that of />-nitroaniline, although we might have expected, , h3c, XII, , XII', , ch3, XIII, , that substitution of four methyl groups, symmetrically arranged, would have, no effect upon the moment of the latter amine. A closer look at the problem,, however, can be taken by constructing scale models of XII. If the two oxygen, atoms and the ring methyl groups are made to lie in the same plane (as should, presumably be the case if form XII' is of considerable importance), one finds, » For a more extensive discussion of resonance effect, on dipole moments, « WfieUnd,, in Organic Chemistry, John Wiley and Sons, Inc., New York, 1955, pp., , 234.
Page 87 :
Molecular Geometry, , -, , 71, , that the oxygen atoms come uncomfortably close to the hydrogen atoms of, the methyl groups that are ortho to them. Apparently then, this steric inter¬, ference forces the oxygen of the nitro groups well out of the plane of the ring,, creating a sort of “bottleneck” between the nitro group and the carbon bonded, to it, and segregating the t electrons of the benzene ring from those of the nitro, group. Nitroaminodurene thus has little, if any, of the character of XII'., The same effect is even more pronounced in nitrodimethylaminodurene, (XIII), which has a dipole moment of only 4.11 D—almost 3 D less than that, of /?-nitrodimethylaniline (6.87 D). Here it seems, both the nitro oxygens and, the N-methyl groups have been pushed well out of the plane of the ring by, interference with the ortho-methyl groups. The bonds linking the —N02 and, —N(CH3)2 groups to the ring have lost their double-bond character and the, 7r-electron systems of the —N02 group, of the benzene ring, and of the amino, nitrogen have become independent of each other; this is in contrast to the, situation in />-nitrodimethylaniline, in which composite 7r-electron clouds are, , />-nitrodimethylaniline, , x-electron nodes, />-nitrodimethylaminodurene, , Fig. 3-4. tr-Electron Distribution in p-Nitrodimethyloniline and Nitrodimethylaminodurene, , delocalized along the entire length of the (almost) planar molecule, both above, and below the plane of the ring. The two molecules are compared schematically, in rigure 3-4., 7, f h, mu“ *, " decreases have been observed" in the dipole moments, benzaldehyde, acetophenone, and benzoyl chloride when methyl groups are, substituted into the 2 and 6 positions of these compounds. The differences are, of he order of 0.3 Z>, obviously far less spectacular than the effects we havl, noted for nitrated aromatic amines., , Molecular Geometry, , /"tv*, , 2, , as and trans configurations to the two 1 2, "Kadeschand VVeiier, v>euer,, , aSSlgnment of the, ,, c, 1’^hloroethvlenes having moments,, j. Am. Chem. Soc., 63, 1310 (1941).
Page 88 :
72, , -, , Dipole Moments and Spectral Studies, , respectively, of 1.8 D and zero. Similarly, the dipole moments of the two isomers, formed in the monochlorination of 1,4-endoxocyclohexane (XIV)18 support the, assignment of structures shown below:, , /hi ^, H = 1.70 D, XIV, The zero dipole moment of cyanogen (CN)2 indicates a symmetric linear, 4*, , configuration,, , 4-, , N=C—C=N (the configuration - : C=N—N=C : ~, which, , would also result in a zero moment, is ruled out by a variety of other considera¬, tions). The very small moment of 1,4-dioxane suggests that this molecule exists, , XV, , XV', , chiefly in the “chair” (XV) rather than the “boat” (XV') conformation,, whereas the moment of 3.5 D observed for gaseous diketene (C4H4O2) rules, out the structure XVI (formerly regarded as possible) although it does not allow, a choice between XVII, XVIII, and XIX, the last of which is now accepted, as correct.75 As a final example,, , c=o, , h2c—, I, , I, ch2, , o=c—, , XVI, , c=o, , h2c—, , I, , CHs—C=CH, , I, , HO—C=CH, XVII, , (U-O, XVIII, , ch2=c—ch2, , I, , I, , o—c=o, XIX, , Exercise 5 illustrates the use of the dipole moment of /,-nitrophenyl azide to, calculate the angle between the azido group and the bond linking this group to, the benzene ring., , Intramolecular Rotation, One of the basic tenets of elementary stereochemistry is that rotation of groups, around the axis of a normal double bond does not generally occur. Rotation, m Martin and Bartlett, J. Am. Chern. Soc., 79, 2533 (1957)., 19 Katz and Lipscomb, Acta Cryst., 5, 313 (1952).
Page 89 :
Intramolecular Rotation, , 73, , about single bonds, however, was assumed in the early days of chemistry to be, free. We now realize that even this freedom has important limitations. Let us, consider rotation in the molecule of ethylene chloride (1,2-dichloroethane)., Three possible conformations of this molecule are shown in Figure 3-5, both, in perspective and in a useful mode of stereochemical abbreviation50 that we, shall occasionally use in subsequent chapters. (In this abbreviation the bond, joining the two carbons is perpendicular to the paper and thus hidden from, view, the intersection of the three full bonds is the carbon nearest the eye, and, the circle represents the farther carbon.) From a glance, it seems apparent that, the two “staggered” conformations on the left should be preferred to that at, , H, , eclipsed, Fig. 3-5. Conformations (Rotational Forms) of 1,2-Dichloroethane, , “in 71’ 7 “, la"er thC at°mS b°nded '° the tw° carb°"s, very much, stable ^, S Way ” More°ver’ ,he '« conformation should be the most, each o heT T*, L ,b, t t,, , 7, ChWine atoms interfere the least with, y’ electro"-diff''action studies have shown that about three, , r°UleS °f, , 'T™ Chl°ride, , VaPOT ~, , c-W, , ad°PlinS thC, called, i, ., ^pole moment of this vapor is 1 12 D at 32° P, but nses w.th temperature, reaching the value 1 54 D at 27 ° Th, ,, suggests that as the temperature is raised, there is a shift in ! ' TK"S Y, more and more of the mnlpr„l~. -a, h ft ln equilibrium, and, practically no dipole moment) m ^ ^ ^ C°nf°rmatlon (which should have, * about 2.5 D). The moment at^ 2t77, , 1, , '^T, , Conformation (with, , 7. a..32, 3 4 1955, esponds quite clos^, A“, K-e, 7. iTk 425 (,,52)., , a
Page 90 :
74, , -, , Dipole Moments and Spectral Studies, , mixture having two skew molecules per trans molecule (if one chlorine is re¬, garded as fixed, there are two positions skew to it but only a single position, trans to it). Thus, rotation in the range of temperatures studied is not “free”, in the sense that the two halves of the molecule rotate without restriction;, they may indeed rotate in passing from one staggered conformation to another,, but they may not tarry on the way. This rotation is similar in principle to that, occurring in a number of additional aliphatic compounds, but derivatives of, the higher hydrocarbons generally have two or more axes about which di¬, pole rotation may occur and there are therefore a larger number of possible, conformations., A different situation prevails for most aliphatic carboxylic acids and their, esters. These show no significant variation of dipole moment with temperature., Moreover, the moments of most simple acids and esters lie in the range 1.71.9 D. Let us ignore possible rotation about all single bonds except the C—O, bond, for the dipole moment should depend almost solely on the configuration, , O, , h (r), , /, , \, , -O, , /, , o, , \, , o, , \, , /, , H (R), , R, , R, , cis, , trans, , Fig. 3-6. Rotational Forms of Carboxylic Acids (and Esters), , about this bond. The two extreme conformations of a carboxylic acid molecule, are shown in Figure 3-6. These may be designated as cis and trans conformations., Assuming the 0=C—O bond angle to be about 120° and the C—O—H bond, angle to be 110°, and neglecting (for simplicity) the R—C bond moment, the, moments of the cis and trans conformations may be shown (Ex. 7) to be 1.2 and, 3 7 D respectively. The apparent dipole moment of a 1:1 mixture of the two, conformations should be 2.8 D (which would also be the value corresponding, to completely free rotation around the C-O bond). The low moments for, acids therefore suggest that rotation about the C-O bond is definitely restricted, and the conformation of most, if not all, of the molecules is a, or a close ap¬, proximation to it" A similar argument applies to esters. For acids, it t us, appears that the electrostatic interaction between positive hydrogen and the, M The difference between the range of observed momenta j •, , J ^ molecules have, , lated for the cis form (1.2 D) does not necessarily suggest that, , °alculations> the bond, , adopted the ^ configuration. It should, angles, and the bond moments were relatively rough approxunat, , and the neglect of the, ,anarconf, ation, , R-C bond moment may result m an error, delations from planarwas assumed, whereas electron-diffraction studies (Ref. 14) show tna, ity as much as 30° may occur for the simple esters.
Page 91 :
Intramolecular Rotation, , 75, , partially negative oxygen atom of the carbonyl group favors the cis configura¬, tion, whereas the same conformation is assumed by esters because it permits, maximum separation of the large alkyl groups. It also seems that one 7r orbital, of the oxygen atom on the left has been lined up nearly parallel to the v orbital, of the carbonyl carbon. In the language of resonance, participation of structure, XX imparts significant double-bond character to the acyl-oxygen bond, and, rotation about this bond becomes almost completely restricted., , R, , XX, Note that in these arguments concerning rotation, the variation (or lack of, variation) of dipole moment with temperature has been a key point. When at¬, tempts are made to interpret data at only one temperature in terms of rotation,, ambiguities generally arise. For example, the moment of terephthalaldehyde, (XXI) is 2.35 D, whereas that of benzaldehyde is 2.76 D. We can then be sure, , trans, XXI, , dd° T h7 - *, , conformation (which, , r c7%urations 45, intermediate in structure to the two planar extremes3*"0", , , a*, , 1951. „ on thc, , COnfo™a“°n,, , and
Page 92 :
76, , Dipole Moments and Spectral Studies, , Part, , II—MOLECULAR SPECTRA54, , Ultraviolet, visible, and infrared spectra of molecules are almost always, absorption spectra rather than emission spectra. Polychromatic light from an outside, source is passed through (or reflected off of) the sample, the resulting beam is, separated into its component wavelengths using a prism or similar device, and, the extent of absorption of light at each wavelength is measured photoelectrically. A plot of the relative absorptions at the various wavelengths against, the values of the wavelengths themselves constitutes the spectrum of the sample., Often, particularly for infrared spectra, the wavelengths are replaced by their, reciprocals, the wave numbers. Maxima in such plots are absorption maxima; light, at such wavelengths is selectively absorbed and brings about some type of, excitation of the molecule. The energy, E, required for the excitation is related, by the Planck equation to the wavelength (X) of light absorbed by the sample:, , (3), , where h is Planck’s constant and c is the velocity of light., Raman spectra are obtained somewhat differently. Monochromatic light, is passed through the sample, and a small portion of the light is scattered,, emerging in directions other than that of the incident beam. Most of the scat¬, tered light has the same frequency as the incident light; but a small number of, additional frequencies appear. The differences between the frequencies of inci¬, dent and scattered light are the Raman frequencies. The Raman spectrum is, thus a measure of the alteration of the frequencies of photons, whereas absorption, spectra are records of photons that are “destroyed.’ In Raman scattering, only, a portion of the energy of each incident photon is required to excite the sample,, the remainder of the energy being emitted as a photon of lower frequency., (The frequency of an incident photon may also be raised by interaction with an, excited molecule.), In microwave and radio-frequency spectroscopy, the incident radiation is, a single wave of very precisely regulated frequency, generated electronically., The types of apparatus used for detecting and measuring the radiation are, of
Page 93 :
Absorption of Energy by Molecules, , 77, , corresponding, by the Planck equation, to transitions of very low energy, which, often arise from mere reorientations of atomic nuclei., , Absorption of Energy by Molecules, Photons of high energies (about 1.5 to 8.0 ev) can bring about electronic excita¬, tion of molecules, just as photons of similar energies can raise the energy levels, of the valence electrons of atoms. Such energy changes correspond to wave¬, lengths between 1500 and 8000 A—that is, the ultraviolet and visible regions of, the spectrum. Lower energy photons (0.05 to 1.2 ev) may increase the intensities, of vibrations within a molecule without causing electronic excitation; these, energies correspond to wavelengths between 10,000 and 250,000 A (or to wave, numbers between 400 and 10,000 cm-1)—that is, the “near” and “medium”, infrared. The vibrational energy levels between which transitions may occur are,, like the electronic energy levels of atoms and molecules, quantized rather than, continuous, although the energy differences between levels are obviously much, less. Even smaller energies (0.00025 to 0.0025 ev) bring about transitions be¬, tween rotational energy levels; the corresponding range of wave numbers (20 to, “00 cm ) ^ie in the, microwave region., , far, , infrared and near the low-wavelength end of the, , Complications in the study of molecular spectra arise from the fact that, the various types of excitation occur in combination rather than alone. Changes, in rotational levels generally accompany changes in vibrational levels, and, c anges both in rotational and vibrational levels often accompany electron, transitions. Figure 3-7 (a Grotrian energy-level d.agram, in which only the vertical, coordinate has significance) illustrates these transitions schematically. The two, ”, , “, , ', , .Hd‘agram (" = 1 and a = 2) represent electronic energy levels, , DllZ “Thtodvibrationt!levels which are’in turn’, , -‘°, , indication of the magnituTeTthe'rergy^iffr131 'T, * V“y r°Ugh, . ., 5, 1 me energy differences between the levels An, , Zr:Z:ibm ,the 7°, , " ^ * **"™, but it is obviou, , from, , -»the, , ::yeCi, , —r ---5, , lines. If the spectrum is poorly resolvld or”, " C‘OSdy SpaCed spectral, levels are further “smeared om” \, •, ’ m°re Partlcularly, if the energy, -te, the lines appear to coa.el ilT“t)“TT “ **“ ^, individual lines may appear as “vibr^f, i 1 ’, ^ g°°d resolution> the, , spectrum. Similarly, rotational effects areobserrable™fi"^, — spectra, and the nuclear, , elcCtronic
Page 94 :
78, , Dipole Moments and Spectral Studies, , vibrational, levels, , !, , rotational, , 1, , levels, , vibration-rotation, transition, Fig. 3-7. Energy Levels for a Simple Molecule (Schematic), , structure in rotational spectra may in turn be observed directly by measure, ments in the microwave region., , Rotational Spectra, Consider a molecule rotating in space about an axis passing through its center, of gravity. The moment of inertia of the molecule about this axis is defined as t e, sum of a number of terms of the type, , (one term for each atom), where m is, , the mass of an atom and r is the distance separating it from the axis of rotation.
Page 95 :
Vibration Spectra and Characteristic Frequencies, , -, , 79, , There are a great many ways in which a molecule can rotate, but rotational, spectral lines give us values for the moments of inertia only about certain of a, very small number of special axes (sometimes called the principal axes of rota¬, tion). For linear molecules, for example, we may observe only one set of spectral, , lines, corresponding to shifts between rotational levels about an axis perpen¬, dicular to the molecule. If the molecule has an axis of threefold or higher sym¬, metry (that is, if it is a “symmetric-top molecule”), we may obtain the moment, of inertia only about a second axis perpendicular to this symmetry axis. For less, symmetric molecules, three moments of inertia about mutually perpendicular, axes may sometimes be obtained, whereas if the molecule has no dipole moment,, it will exhibit no rotational spectrum at all., Whether such moments are evaluated from measurements in the far infra¬, red, from microwave spectra, or from the fine structure of vibration spectra,, their usefulness depends largely upon how readily they may be converted to, interatomic distances and bond angles. For diatomic molecules (in which, how¬, ever, the organic chemist has only minor interest), such a conversion is very easy., For polyatomic molecules, determination of dimensions from a single moment, of inertia amounts to solving two equations in three or more unknowns. In, many such cases, the data may be supplemented with the moment of inertia, of the same compound in which one or more isotopic substitutions have been carried out, (see Ex. 14). Such substitutions are assumed not to alter the geometry of the, molecule. Although the analysis of rotational spectra has been confined to rela¬, tively simple compounds (two of the most complex are CF3C=CH and tri¬, methylene oxide), the bond lengths and bond angles obtained may generally, be applied to more complicated molecules., , Vibration Spectra and Characteristic Frequencies, The problem of vibrations within polyatomic molecules would be appallingly, difficult if ,t were not possible to handle the equations of motion for even the, most complicated vibrations by combining the equations of motion of rather, simple vibrations. For example, all possible vibrations of the cyanogen chloride, molecule are said to be derived from “superposition" of the four mots of, ration indicated m Figure 3-8. These vibrations are the so-called normal, , Cl, , C---N->, , T, Cl—>C—N—>, , Cl - C1, , (1), , (2), , (3), , Pig. 3-8. Normal Modes of Vibrati<, , T, Cl—C—N, C in and out, Cl and N out and in, , (4)
Page 96 :
80, , -, , Dipole Moments and Spectral Studies, , modes of vibration; during each of them, all atoms pass through their equilibrium, positions at the same instant. All absorption in the infrared or Raman spectra, is associated with one of the normal modes of vibration, or with overtones or, combinations of these. This condition drastically limits the number of vibra¬, tion bands that may be observed, but it still may be shown that for a nonlinear, molecule of N atoms there are 2>N — 6 normal modes of vibration (and one, extra if the molecule happens to be linear). There are thus 30 normal modes for, the relatively simple molecule, benzene, and 216 for the more complicated, molecule,, , cholesterol., , Even, , though, , these numbers are sometimes further, , diminished by “selection rules” (which may be derived theoretically if the, structure of the compound is known and which are generally not the same for, infrared spectra as for Raman spectra), it should be quite clear that a complete, analysis of the vibration spectrum for any but the very simplest molecules is no, job for the chemist who is in a hurry. This does not mean, however, that vibra¬, tion spectra are useful only in the hands of a very competent mathematical, physicist. On the contrary, infrared examinations have aided in the identifica¬, tion and structure determination of thousands of very complicated organic, compounds, largely by workers whose analyses of the spectra amount to little, more than hurried scannings. Although the stretching or bending of a bond, between two atoms may be affected by other atoms in the same molecule, such, influence is often not sufficient to prevent the appearance of a characteristic fre¬, quency derived from such a stretching or bending. For example, almost all com¬, pounds having a carbonyl group (without a-/3 unsaturation) display a strong, “C=0 frequency” band between 1700 and 1800 cm-1; a compound having, one or more nonconjugated C==C bonds will generally absorb between 1640, and 1680 cm-1 (the region of the “C=C stretching frequency”). Likewise,, compounds having an aliphatic —OH group generally absorb near 3600 cm, , ,, , and various deformations of methyl and methylene groups (picturesquely called, “scissoring,” “wagging,” “twisting,” and “rocking”) may give rise to recog¬, nizable bands. Thus significant progress toward the structure proof of a new, compound may often result from even a superficial examination of its vibration, spectrum., The nature of functional groups themselves may occasionally be clarified, in much the same way. A class of compounds thought to be disulfoxides of the, type r_S—S—R' was found to absorb strongly at 1340 and 1150 cm-1, the, , I, , I, , O O, , derivatives, , regions characteristic of sulfones, , O, , ., , but not near 1040 cm \
Page 97 :
Vibration Spectra and Characteristic Frequencies, , -, , 81, , the region characteristic of sulfoxides.-5 On this basis the disulfoxide structure for, , O, this compound does not seems as likely as a thiosulfonate structure, R-, , S, , SRh, , O, , \, Similarly, N-nitroso compounds (generally represented with an, , _, , N—N—O, , linkage) exhibit bands near 1400 and 1200 cm-1, but none near 1520 cm-1, the, wave number to which the N=0 stretching frequency has been assigned;*5, this would suggest that the N-nitroso linkage might better be represented as, \+, N=N—O"., , /, As with rotational spectra, additional useful data may arise from isotopic, substitution in a molecule. Substitution of C14 for C12 or O18 for O16 may cause, small but noticeable shifts in the positions of spectral bands, and very sub¬, stantial “isotopic shifts” result when deuterium or tritium is substituted for, hydrogen. (It has been shown*7 that such a substitution may lower the C—H, stretching frequency by as much as 30 percent.) If it is possible to replace all, hydrogen atoms in a given position by deuterium, a band may seem to vanish, from the spectrum and reappear at a lower frequency region. As a single ex¬, ample of the use to which this effect may be put,*5 consider the identification, of bands associated with individual methyl groups in the spectrum of the steroid,, XXII. This compound has four narrowly separated maxima in the region, , 1357 cm-1, , XXII, " Cymerman and Willis, J. Chem., r ^uttl<e» J- Phys. Radium, 15, 633, (1947)., , f°r eXample-, , Soc., 1951 1332., (1954), , ref. 24. p. 334; also Halverson,, , which ^Su?lmat«op°e^tf;spec;ra htlidti“IT2*' F°r a, nngs, , Mod. Phys., I9. 87, -mple, in, , tr,terpenoids, see Barton, Page and Warnhoff, J. Chm sl^'isU^m5 CyCl°pr°panc
Page 98 :
82, , -, , Dipole Moments and Spectral Studies, , 1355-1385 cm-1, each presumed to be due to symmetrical bending vibration, of one or more of the methyl groups in the structure. The first of these, at, 1357 cm-1, disappears if the compound is treated with D20 and NaOD, sug¬, gesting that this maximum is associated with the enolizable hydrogen atoms,, alpha to the keto group, attached to the cyclopentane ring. Two additional, maxima, at 1365 and 1375 cm-1, disappear if CD3COOH rather than CH3COOH is used to esterify the parent alcohol, showing that both of these maxima, are associated in some way with the methyl of the acetate group. The remaining, maximum, at 1380 cm-1, is presumed to be associated with one or both of the, “angular” methyl groups as indicated., , Additional Structural Information, Sometimes an even more detailed picture may be obtained from the study of, vibration spectra. As an important example, it is largely through spectral evi¬, dence that we have become sure as to which of the two “strainless” conforma¬, tions of the cyclohexane ring is the correct one. Both the “chair” (XXIII) and, the “boat” (XXIV) forms of cyclohexane have 48 normal modes of vibration,, but selection rules stipulate that far fewer of these vibrations should appear in, , the infrared spectrum of the more symmetric chair form. It turns out, for ex¬, ample, that there are 18 “infrared active” C-C stretching and CH, rocking, and twisting vibrations possible for the boat form but only, , five, , of these for the, , chair form. Examination of the spectrum of cyclohexane in the region where, these vibrations should appear (700 to 1350 cm-) reveals the five bands expected for the chair form.*5, We have already seen (p. 73) how dipole moment studies of 1,2-dichloroethane indicate that the trans conformation predominates at low temperature, with the “skew” conformation becoming more important as the temperature, is raised. Spectral studies of this compound over a, 0, much the same story. The band at 1291 cm--attributed, “, vibrations” of the trans conformation-becomes less intense as the temperat, <» Rasmussen, J. Chem. Phys., 11, 249 O'143)., « Bernstein, J. Chem. Phys., 17, 262 (1949).
Page 99 :
Additional Structural Information, , 83, , is raised, whereas the band at 1235 cm-1—attributed to the same type of vibra¬, tions in the skew conformation—becomes more intense.30 Similar temperature, variations in the infrared spectra of alkyl nitrites suggest a shift in the equilibR, rium between the rotational forms, , OR, , \, , /, , O—N, , and, , \, , O—N, , %,o, , ,S1, , Quite commonly, significant structural information may be obtained from, relatively small shifts in the position of characteristic frequencies. Some of the, best evidence for hydrogen bonding (p. 28) has arisen this way. The “free”, O, , H vibration band at a frequency of about 3620 cm-1 generally appears, , strongly in the spectra of dilute solutions of alcohols and phenols in inert sol¬, vents. As the concentration of solute is increased, this band almost invariably, diminishes in intensity, with one or more new bands appearing at slightly, lower frequencies (3350-3500 cm-1). As a result of association, a number of, linkages of the type, , O, , H • • • O—H appear; the hydrogen bonded to, , one oxygen atom is pulled slightly away by its attraction to a second oxygen, atom; and the energy involved in the vibration of the somewhat weakened, O—H bond becomes less. As the strength of hydrogen bonding increases, the, position of the “associated O—H” band becomes farther and farther removed, from that of the “free O—H” band.32
Page 100 :
84, , -, , Dipole Moments and Spectral Studies, , axial (XXVII) positions.55 (Note that in the axial diol the hydroxyl groups are, much too far apart to allow intramolecular hydrogen bonding.), Shifts in the C=0 stretching frequencies have also proved instructive., The stretching frequency of any bond depends not only on the masses of the, bound atoms but also on a bond-force constant (actually a sort of Hooke’s Law, constant for the bond). The C=0 bond-stretching constant must be affected, when nearby substituents alter the charge distribution around the bond. More, particularly, substituents such as —Cl, —F, or —N02, which attract electron, density away from the carbon atom (leaving it with additional positive charge),, are found to tighten the C=0 bond—that is, increase its vibration frequency., This effect is very striking in the halogenated acetones., Compound, , C=0 Stretching Frequency, , CH3COCH3, CH3COCH2CI, CH3COCF3, , 1715 cm-1, 1724, 1769, , CF3COCF3, , 1801, , As the carbon atom is made more and more positive, the extra polarity aug¬, ments the already strong carbon-to-oxygen attraction. The C=0 stretching, frequencies of substituted acetic54 and benzoic acids55 show the same trend; those, substituents that boost the dissociation constants of these acids are found to, tighten the C=0 bond, whereas those substituents that lower the dissociation, constants loosen the C=0 bond., A C=C double bond in conjugation with the C=0 linkage generally, lowers the stretching frequency of the latter by about 30 cm 1 (for example,, cyclohexanone, 1715 cm-1; 2-cyclohexene-l-one, XXVIII, 1680 cm-1). Since, bond-force constants are known to decrease as bond order decreases, this shift, , XXVIII, , XXVIII, , suggests that the smearing of electron density over the conjugated system has left, the carbonyl linkage with a bond order slightly less than 2—a situation that, may be represented by contributions from additional resonance forms such as, XXVIII'. It is interesting that conjugation effects in the infrared are in the same, direction as conjugation effects in the visible and ultraviolet (that is, in both, cases conjugation shifts absorption to greater wavelengths). We should note,, « Kuhn, J. Am. Chem. Soc., 74, 2492 (1952)., Gillette, J. Am. Chem. Soc., 58, 1143 (1936)., « Flett, Trans. Faraday Soc., 44, 767 (1948).
Page 101 :
Electronic Spectra, , -, , 85, , however, that the two types of “bathochromic shifts” have wholly different, causes. Conjugation effects in infrared and Raman spectra arise because of, variations in bond-force constants, whereas conjugation effects in the visible, and ultraviolet stem from variations in the energy gaps between the ground, states and electronically excited states., , Electronic Spectra, Although in the past, visible and ultraviolet spectra were employed by organic, chemists largely for identification and analyses, electronic spectra are physically, important because of the information they should yield concerning the excited, states of molecules., At present, however, we cannot translate spectra of organic molecules into, quantitative descriptions of their energy levels (as is possible for atomic spectra)., The electronic energy levels for molecules are more numerous and the relation¬, ships between them more complicated than for atoms. Experimentally, rather, large uncertainties in the gaps between levels arise because of the smearing of, electronic lines by vibrational fine structure and because of modification of the, levels by solvation. Frequently, there is real difficulty in identifying each of the, observed absorption peaks with one of the theoretically possible transitions., The following section is only a very brief sampling of types of qualitative correla¬, tions that may be made between structural features and electronic bands., Interest in electronic spectra of organic molecules is largely centered, di¬, rectly or indirectly, about conjugated systems, for the characteristic spectra of, conjugated compounds lie within the range of wavelengths accessible to con¬, ventional instruments. Saturated aliphatic hydrocarbons absorb in the range, 1250 to 1750 A—that is, in the far (and relatively inaccessible) ultraviolet', electronic excitation of these compounds involves the boosting of an electron, already participating in a C-C or C-H bond to a higher level. Such a process, requires considerably more energy than does the excitation of T electrons in, S’ and the P^nce °f a C=C double bond will result in the appearance, one or more additional maxima at wavelengths above 2000 A ®> These, maxima arise from the excitation of the , electrons to higher energy orbitals, ctnushe, , rbbWike) CharaC‘er> bU‘ Call6d (f°r rCaSOns *«* need not com, , localized o„ ot amm, atoms connected by, , te’douhtbLd., , ' "T “ ^ a", , ZZtd Tl, , ^ugated system, the excited , electron, , - not, , *
Page 102 :
86, , -, , Dipole Moments and Spectral Studies, , conjugation; such delocalization further lowers the energy of the excited state., Since conjugation is found to shift absorption to longer wavelengths, it is temptmg to associate such a spectral shift directly with the drop in energy of the, excited state. It should be remembered, however, that conjugation lowers the, energy level of the ground state also (p. 38), and the observed bathochromic, shifts associated with conjugation must arise because the energy levels of the, excited states are somehow affected more than the energy levels of the ground, states. This situation is too complicated to be treated here, but the effect itself, is of extreme importance.®7, Conjugation shifts in the ultraviolet, in the absence of complicating factors,, are roughly “cumulative.” Polyenes with 2, 6, and 10 conjugated double bonds, o, , display maxima, respectively, near 2200, 3600, and 4500 A, the last being well, within the visible range; similarly, biphenyl shows a maximum near 2500 A,, whereas the polyphenyls XXIX and XXX absorb, as indicated, at somewhat, greater wavelengths., , \ r \ r\ r\j, XXIX Xmax 3000 A, , XXX Xmax 3175 A, , XXXII, , XXXIII, , Such conjugation effects tend to be most marked when the molecular, geometry allows maximum separation in space between the ends of the con«r for two qualitative explanations of conjugation effects *£TIectI,°|TC, rather Afferent points of view), see Wheland, (Ref. 16), PP. 257-278; and We1Ssberger, (Ref., 24), pp. 654-657.
Page 103 :
Electronic Spectra, , -, , 87, , jugated system. Of the four isomeric polycyclic aromatic hydrocarbons shown, below, naphthacene (XXXI) absorbs near 4600 A, whereas maxima for the, remaining three lie near 3500 A.ss Likewise, the trans form of a polyene (XXXV), will generally absorb at higher wavelengths than will the cis form (XXXVI).35, On the other hand, conjugation effects are diminished when the molecule, becomes unable to assume a planar configuration (Ex. 11)., , :c=c', , ;c=c^, , ;c=c;, , c=c, , c=c, , ,c=c:, xxxvi, , xxxv, , Hyperconjugation, which, as we have seen, affects heats of combustion, (p. 49,) and dipole moments (p. 64), also influences electronic spectra. Gen¬, erally the absorption maximum for an olefinic or aromatic compound is shifted, toward greater wavelengths when a methyl group is substituted for an a-hydroo, , gen. Typically, 1,3-butadiene absorbs at 2170 A, whereas both of the 1,3•, , ^, , •, , o, , pentadienes absorb near 2250 A; similarly, benzene absorbs at 2625 A, toluene, o, , at 2668, and mesitylene at 2700 A.40 Such shifts lend reality to forms such as, T and IT (p. 64) which attempt to show an effective extension of conjugation, by the methyl group., Unshared electrons are more easily excited than electrons involved in, ordinary <r bonding. Alkyl halides, alcohols, and amines thus absorb at greater, wavelengths than do their parent hydrocarbons. Moreover, if unshared electrons, are associated with double-bonded atoms, these electrons are very easily excited, to the antibonding tt level; most of the classical chromophoric groups (for, , O., example,, , C=6:,, , C=*S:, —N=N—, —N=6, , and —N, , are of, , 0:, , :, , * *, , /, , this type. Again, electronic excitation is further facilitated if these groups are, incorporated into conjugated systems. Thus nitromethane absorbs near 2700, nitrobenzene near 3500, and /i-nitroaniline near 3800 A., It would not be appropriate in a text of this kind to discuss at greater, ength the relat.onships between spectra (or color) and structure. For more informatton about this interesting topic, the student is referred to more specialized, ^Klcym5 and Platt, J. Chem. Phys., 17, 470 (1949)., oee, for example, Zechmeister and Pinckard / Am Ch, m c, nc n, stereoisomeric diphenyloctatetraenes., ’ ' A, °h, S°C’’ 76’ 4144 (1954), on the., ^The absorption maximum for <-butylbenzene, VE). The'fac^thaVthis^value, indicated, C-C hyperconjugation is less impLiant, , liec, , 1, , j, (p. lo,Tu«2, " ”*
Page 104 :
88, , -, , Dipole Moments and Spectral Studies, , works.4* It should be clear, however, that behind correlations such as we have, presented (which are little more than educated “rules of thumb”) lies an ex¬, tremely complicated picture of the electronically excited states of organic, molecules which today is just beginning to come into focus., , EXERCISES, , FOR CHAPTER, PART, , 3—, , I, , 1. Indicate which of the following molecules have dipole moments, representing the, dipole (if any) by an arrow pointing from the more positive end of the molecule to, the more negative end:, CH2CI2, naphthalene, OCS, HN3, (CH3)20,/>-fluoronitrobenzene, COCI2,, o-chlorotoluene, iodoacetylene, pyrrole, />-dimethoxybenzene, 2. From the Clausius-Mosotti-Debye relationship (Eq. 2) show that a plot of the, dielectric constant, D, for a gas vs. the reciprocal of the temperature should approach, a straight line (remember that D is quite close to unity), and that the slope, S, of the, line is related to the dipole moment by the relationship:, ,, M, , 3 MkS, ~ 4trNp, , where p is the density and the other symbols are defined on page 60., 3. (a) Use the bond moments on page 62 to calculate the dipole moment for phosgene,, C12C=0. The molecule is planar with a Cl—C—Cl bond angle of 113°., (b) The observed moment for phosgene is 1.18 D. Why is this value so much, different from the one calculated in (a)?, 4. Consider a benzene ring having two substituents with moments p\ and M2, and, assume that both group moments lie in the plane of the ring. Show that a simple, vector treatment predicts a resultant moment of \/p\ + p\, meta to each other, and a moment of VVi + M2 +, , ^2, , M1M2 if the groups are, , if the groups are ortho to, , each other., 5. Given the moments below, and assuming that the N3 group is linear, calculate the, angle between the N3 group and the C, , N bond in />-nitrophenyl azide., , p =2.96D, 4/ See for example, Mayer and Cook, The Chemistry of Natural Coloring Matters, Reinhola, Publishing Corp., New York J 943; Venkataraman, Chemistry of Synthetic Dyes Aca^nuc Pre^,, inc, New York, 1952; Grimmel in Gilman’s Organic Chemistry, an Advanced Treatise, Vol. 3,, , John Wiley and Sons, Inc., New York, 1953, p. 243.
Page 106 :
90, , -, , Dipole Moments and Spectral Studies, , 7. Using the sketches below and the bond moments on page 62, estimate the dipole, moments of the cis- and trans- rotational forms of acetic acid. Neglect the C—H bond, moments, assume the forms to be planar, and take the 0=C—O and C—O—H, bond angles as 120° and 110°, respectively., , cis, , trans, , 3. (a) Show that the apparent dipole moment of an equimolar mixture of two isomers, is the root mean square of the individual moments, that is, (m! + M2), (b) What would the dipole moment of acetic acid be if it consisted of an equimolar, mixture of the two forms in Exercise 7? Compare your value with that given on, page 74., , PART, , II, , 9. (a) The wavelength of visible light lies between 4000 and 7000 A. Show that a, molecule will be colored if the energy needed to excite it lies between 1.75 and, 3.0 electron volts (1.0 ev = 1.6 X 10 12 erg)., 0, (b) Explain why many compounds with absorption maxima well below 4000 A, and with no maxima above 4000 A (for example, w-nitroaniline Amai 3600 A), are colored to the eye., 10, , (a) The IR spectrum of an 0.01 molar solution of cyclohexanol in CC14 displays, only one O-H stretching frequency. An 0.03-molar solution shows two such, peaks, an 0.1-molar solution shows three such peaks, but a 1.0-molar solution, shows only two peaks. Explain., OH, (b) The IR spectrum of (CH3)3C—C—C(CH3)3 displays only one O, , H stretching, , CH, ch3, , ch3, , frequency, irrespective of dilution. Why is this?, (c) Suggest a reason why two O-H stretching frequencies appear in the p, very dilute solutions of o-chlorophenol, but only one each in the spectra, dilute solutions of phenol and, , 11., , t, , f, >, , 2,6-dichlorophenol., , Equal weights of toluene and 4.4'-bitolyl are diluted to ;equal volumes wi* tdcohol., The UV spectra of the two solutions are very much di eren .
Page 107 :
Exercises for Chapter 3—Part II, , 91, , ment is carried out with equal weights of mesitylene (XXXVII) and bimesityl, (XXXVIII), the two spectra are almost the same. Explain., , Me Me, , / \ / \, , Me, , Me, , Me Me, XXXVII, , XXXVIII, , 12. Cyclopropane derivatives of type (A) are known to rearrange under treatment with, HC1 to yield olefins of type (B)., , A compound of type (A) has a CH3 group attached either to position 2 or to position, 4 (but not to both). A sample of the compound is treated with HC1, and a second, sample with DC1. The IR spectra of the two products show the same intensities for, the peaks attributed to CH3 bending frequencies. Decide whether the CH3 group in, compound (A) is at position 2 or position 4., 13. Predict which compound in each of the following pairs displays an absorption maxi¬, mum nearer the red (high wavelength) end of the UV or visible spectrum. Justify, vour choice in each case., 7, , O, Ivl, , o, , G—CH;5 or, , —CH2—C—CH.i?, , (b) />-Nitrophenol or m-nitrophenol?, (c) />-Nitroaniline or its hydrochloride?, (d) ch2=ch, , ch=ch, , CH=CH, , or, , CH=CH., ch=ch2?, , ch2, , oh, , CH=CH, CH=CH, , (e) Biacetyl or acetonylacetone?
Page 108 :
92, , -, , Dipole Moments and Spectral Studies, , (§) (CeHj^C, , OH or (CeHs^C—O—SO3H?, , (h) Phenol or thiophenol?, , 14. Consider the rotation of the linear H—C=C—Cl35 molecule about the principal, axis A which passes through the center of gravity of the molecule at a distance x from, , ■ x-, , H, ^C-H, , drC-C, , C1 35, , drC-Cl, , the nucleus of the chlorine atom. Designate the (observed) moment of inertia of this, molecule as I., (a) Set up two independent equations in four unknowns, relating I, x, and the inteinuclear distances in the molecule., (b) Assume that the dimensions of the molecule will not change with isotopic sub¬, stitution. Show that if the moment of inertia,, , of D, , C=C, , Cl3j is known,, , we may obtain four equations in five unknowns, whereas if the moment of, inertia, I", for H—C=C—Cl37 is also known, we may obtain six equations in, six unknowns, thus allowing the calculation of the three bond lengths in the, iiiolecule.
Page 109 :
CHAPTER, , 4, , Acids and Bases. Nucleophiles, and Electrophiles, , By the classical or arrhenius definitions, bases are compounds that yield, , hydroxide ions when dissolved in water, whereas acids are compounds that, yield hydrogen ions. Such definitions are quite adequate if one considers reac¬, tions only in water; in particular, they serve nicely for the student of elementary, chemistry whose experiments are confined (for the sake of his own safety and the, peace of mind of his instructor) to aqueous solutions. However, the acid-base, concept is such a useful one that workers have redefined and generalized the, terms, , acid ’ and “base” in a number of ways; today many reactions are de¬, , scribed as acid-base reactions, even though neither the H30+ ion nor the OH', ion participates. In all such cases, however, there is some analogy to the classical, neutralization” reaction. Of the many generalized acid-base systems that, have been proposed,7 the two extensions most useful to organic chemists are, the proton-exchange approach, approach (Lewis system)., , (Br^nsted system), , the electron-pair transfer, , and, , Hydrogen Ion Transfer. The Br0nsted-Lowry System, In the BrjzJnsted system, an acid is defined as a species that can give up a proton, (an H, , ion) to another species. A Brfnsted base is a species that will accept a, , proton. Thus the classical acids are also Brjzfnsted acids., , John Wiley and Sons^Inc.^NtwTorrTS8nT30^335^, , b,y, , Moeller> Inorganic Chemistry,, , ™^,The Elects Theory of Acids 'and 42^^^, , 93
Page 110 :
94, , Acids and Bases. Nucleophiles and Electrophiles, HNOs, , -» H+ + NOj, , \, , CH3COOH —> H+ + CH3COO- > conjugate bases, C6H6OH, -> H+ + C6H50I, The ions on the right which remain after protons depart from each of the acids, are called the conjugate bases of the respective acids., Although the Br0nsted definition of an acid is very similar to the classical, definition, it includes a number of molecules or ions that classically might not, have been regarded as acids., , /, , O, , CH3NO0 -> H+ + ~CH2—N, O, (CH3)3NH+ -> H+ + (CH3)3N, A1(H20)+3, , H+ + A1(H20)50H+2, , Note that these reactions are not “spontaneous” ionizations. Rather, the H+ is, “pulled off” by an approaching base. Thus, the reactions above are actually, half reactions., Brjnsted bases, species that will take on protons, are generally negative, ions or neutral molecules, although there are a few of such bases (for example,, Zn(H20)30H+) that are positively charged. Basic half reactions of the hydroxide, ion, the carbonate ion, aniline, the methoxide ion,, , -pyrone, and acetic acid, , 7, , are listed below, OH", , +, , H+, , >, , h2o, , C072, , "h, , H+, , >, , HCO3-, , +, , -O, CH3COOH, , *T, , H+, , +, , H+, , V conjugate acids, , +, , +, , H+
Page 111 :
Hydrogen Ion Transfer. The Br0neted-Lowry System, , -, , 95, , As indicated, the species formed when a proton is added to a base is called the, conjugate acid of that base., A full acid-base reaction is a transfer of a hydrogen ion from an acid to, a base, yielding the respective conjugate base and conjugate acid (Table 4-1)., , Table 4-1, Acid, , 1. H30+, 2. H20, , Base, , Conjugate Base, , + OH+ H-, , 3. h2so4, + h2o, 4. CH3COOH + HCO^, 5. h2o, -f h2o, , Conjugate Acid, , h2o, , +, , h2o, , OH-, , +, , h2, , hso,-, , +, , h3o+, , CH3C00OH-, , +, , h2co, , +, , h3o+, , We see that water (in common with many hydroxylic solvents) may be either, an acid or a base, it may gain or lose a proton, depending upon the presence, of other acidic or basic species. Reaction 5 represents the so-called autopro¬, tolysis of water; similar “self-ionizations” occur to some extent in other polar, solvents, yielding equal quantities of lyonium (conjugate acid of the solvent), and lyate (conjugate base of the solvent) ions; for example, in methanol*, , +H, 2CH3OH -> CH3—O, , + CH3O-., , K = (CH3OH+XCH3O-) = 2 X 10-”, , H, and in formic acid, , OH, 2HCOOH, , H—C+, , + HCOO-., OH, K, , (HCOOH;f)(HCOO~) = 6 X 10~17, , From the manner in which the Boosted definitions are set up it follows, that very strong acids should have very weak cnni„ya,„ t, , j ., that i« if, *, uv., y weaK conjugate bases (and vice versa)tnat is, it the acid HX is stronger than HF then V— -n 1, A-. (Note, however, that it is possible fo HA ", 3 S‘r°nger baSe ‘ha", also a better base than H!- for examje, “, ^ HFand, ’ tor example, compare water and acetone.), -.l * Thc autoprotolysis constants refer to solutions at, Th, other soWents, are given by Hammett in Physical, xt0gether with values for, Inc., New York, 1940, p. 256, V, °rSamc Chemistry, McGraw-Hill Book Co
Page 112 :
96, , Acids and Bases. Nucleophiles and Electrophiles, , The Leveling Effect, When we say that an acid, hL4, is “fully ionized” in a solvent S, we mean that, its reaction with the solvent is virtually complete, being of the type, , HA + S -> HS+ + A~, We see then that the apparent strength of an acid will depend on the ability of, the solvent to accept protons—that is, on the basicity of the solvent. The, carboxylic acids, with very few exceptions, are “weak” in water, but they, become fully ionized when dissolved in liquid ammonia since their reactions, with this very basic solvent are almost complete., RCOOH -f NH3 -> NH+RCOOThe extreme basicity of liquid ammonia may be said to exert a leveling effect, on the apparent strengths of the acids that are dissolved in it. All carboxylic, acids have about the same degree of ionization in this solvent, and, moreover,, appear to be as strong as the common mineral acids. Experiments in the weaker, base, water, show us that the mineral acids are actually the stronger and also, allow us to arrange the weaker acids in the order of their strengths. To tell, further which of the mineral acids is the strongest, it is necessary to carry out, experiments in solvents of even lower basicity, or lower ionizing ability, or, both. In methanol, for example, nitric acid is partially ionized,3 but hydro¬, chloric acid is fully ionized. In anhydrous formic acid, however, HG1 becomes, partially ionized, whereas the first ionization of sulfuric acid is still virtually, complete/, Thus in comparing a number of different acids in the same solvent, that acid, is strongest which is most ionized. On the other hand, if we compare the acidities, of solutions of a single acid in a number of different solvents, we should remember, that the most strongly acid solution is that in which ionization is least. A com¬, parison between the strengths of aqueous HC1 and HC1 in benzene is often, made. In water, the solute is mainly in the form of H,0+ and Cl~ ions; that is,, the water has, in a sense, “neutralized” the HC1. In benzene, the HC1 is prac¬, tically unionized. If we were to test the two solutions with the same indicator,, the indicator would have to compete with the base HaO for protons in aqueous, solutions, but would compete with the much weaker base Cl, , for protons in, , benzene solution. More of the indicator would then be converted to its con¬, jugate acid in benzene than in water (where the competition is keener), thus, showing that the benzene solution is the better proton donor., * Deyrup, J. Am. Chem. Soc., 56, 60 (1934)., * Hammett and Deyrup, J. Am. Chem. Soc., 54, 4239 (, , )•
Page 113 :
Very Weak Acids, , 97, , Very Weak Acids, Among organic compounds there are a number of very weak acids which, even, in the basic solvent, ammonia, are not measurably ionized. These may be per¬, suaded to part with H+ ions only by the stronger negatively charged bases., The most common of such bases is, of course, the OH" ion, but this is too weak, a base to convert many of the very weak organic acids to their own conjugate, bases. For such purposes, even stronger bases (that is, conjugate bases of acids, weaker than water) are needed. However, the strongest base that can exist in, appreciable quantities in aqueous solutions is the OH- ion; stronger bases will, react with water to yield this ion. It then follows that the acidic properties of, very weak proton donors may not be observed by experiments in aqueous, media. A number of alcohols and amines, however, and some especially acidic, hydrocarbons may be converted to their conjugate bases by action of the amide, ion, NHj, in liquid ammonia. For acids still weaker than ammonia, proton, transfers may be carried out in ethers or hydrocarbons., The reactions given in Table 4-2 illustrate the conversion of some very weak, organic Br0nsted acids to their conjugate bases. In (1), acetoacetic ester is con¬, verted to its “enolate” ion by ethoxide ion in ethanol. “Deprotonations” of, £-butyl alcohol (2), fluorene (3), and diphenylamine (4) may be carried out in, liquid ammonia or ether,5 and the transfer of a proton between the methyl, , Table 4-2, Acid, , Base, , Conjugate Base, , O, , O—, , l S!?Tn~^2~C°OEt +, 2- (CH3)3C-OH, + NHy, , 3., , CH3-C=CH-COOEt +, -> (CH3)3C—Q~, +, , nh;, , 4. (C6Hs)2NH, , ,fV, , Conjugate Acid, , CH3, , +, , + (C6H5)3C:~—> (C6H5)2N:~, , + C6H5Na, , HOC0H5, nh3, , -, , 97™\?noraNnt and Wheland- J- Am. Chem. S, Z/3; (1925); McEwen, ibid., 58, 1124 (193f, , f~\— CH~Na+, , +, , NHc, (C6H8>3CH, , + CTT, , 6aa6, , 54, 1212 (1932); Kraus and Rosen, ibid., 47
Page 114 :
98, , -, , Acids and Bases. Nucleophiles and Electrophiles, , gi oup ot toluene and the sodium salt of the benzene anion may be carried out in, hydrocarbon solvents.5, , Very Weak Bases. Studies in Concentrated Sulfuric Acid, On the other side of the coin, we find a very large number of organic compounds, having very weak basicities. Among these are ethers, carboxylic acids, amides,, ketones, nitro compounds, and even some aromatic hydrocarbons. Although, certain of such compounds are appreciably soluble in water, additions of strong, acid to the dilute aqueous solutions result almost exclusively in the formation, of, , H30+,, , the conjugate acid of water. Since only tiny quantities of the con¬, , jugate acids of the weak organic bases are formed under such conditions, it is, necessary to observe the basicities in other ways., The most widely used solvent for studying the basicities of very weak bases, is 95 to 100 percent sulfuric acid. Particularly instructive data have arisen from, cryoscopic studies in this solvent.7 The freezing point, 10.36°, of the pure acid, lies conveniently between room temperature and “ice-bath” temperature and, its cryoscopic constant (over three times that of water) is relatively high. More, important, it is the most acidic solvent readily available, and cryoscopic meas¬, urements of electrolytes at convenient concentrations turn out to be less sub¬, ject to large “departures from ideality” than are measurements in aqueous, solutions.5 Somewhat paradoxically, data from such studies are most easily, interpreted if the sulfuric acid used is not 100 percent pure but contains about, 5 percent or more of water (Ex. 4c)., In such studies, the freezing-point lowering caused by a given solute is, compared with that caused by the same concentration of a nondissociated solute;, the ratio of the two depressions is the i factor or v factor of the solute, representing, the average number of separate dissolved particles produced by dissolving one, solute “molecule.”, Trifluoroacetic acid, sulfuryl chloride (SO2C.I2), and perchloric acid, which, ionize neither as acids nor bases in sulfuric acid, exhibit v factors very close to, unity. On the other hand, substances that behave as simple “monoacid bases, in this solvent show v factors near 2. Their reactions may be represented by, B +, , H2SO4 -> H£+ + HSO7, , (v = 2), , 6 Schorigin, Ber., 43, 1938 (1910)., ,, ,, 1on7, 1 The early work in this field was carried out by Hantzsch and co-workers betw, and 1930. This, and further contributions by Hammett and his co-workers, have been sum¬, marized (together with pertinent references) by Hammett, Ref. 2, pp., , •, more recent and detailed picture, see Gillespie and Leisten, Quart. Reus. VIII, 40 (, j, « The “pseudo-ideal” behavior of convenient concentrations of electrolytes in sultur, acid is discussed by Brand, James, and Rutherford, J. Chem. Soc., 1953, -447.
Page 115 :
Very Weak Bases. Studies in Concentrated Sulfuric Acid, , -, , 99, , In this category are: monobasic amines, simple ethers that form cations of the, R, , R, , ., , \+, , \+, , O—H, ketones that are converted to acids of the type, , type, , /, , R', , C—OH,, , R, , /, , simple carboxylic acids (but not substituted mesitoic acids) that are converted to, OH, , +/, , OH, , +,, , cations of the type R—C, , , esters that are protonated to R—C, OH, , OR', OH, , amides that form cations of the type R—C, , , and many additional com-, , nr;, pounds with less commonly occurring functional groups. Nitro compounds,, cyclic carboxylic anhydrides, and sulfonic acids display v factors lying between, 1 and 2, indicating that these compounds are incompletely converted to their, , conjugate acids.5, Of further interest are a number of compounds that exhibit v factors, greater than 2. In the three cases shown below, the cryoscopic data indicate, formation of carbonium ions (more will be said about these later)., (C6H5)3C—OH + 2H2SO,-^, , (c6H6)sGf, , +, , HsO+ +, , 2HSOr, , (v = 4) 10, , Me, Me-/, , VcOOEt, , + 3H2S04 •, Me \=/Ci° + Hs°+ + e‘OSO,H + 2HSO.(r=5)», , Me, , +, , 2H2S04, , + HaO+ + 2HSO;, , (y=4)12, , Although aromatic hydrocarbons havin- three or mnr. f, a •, soluble in concentrated sulfuric acid, the sLoler hvH, t, insoluble. However ben?enP, , ,, ^, }drocarbons are almost, owever, benzene and the methylated benzenes will dissolve in, Gillespie, J, Chem. Soc, , 1950 9^4? • xt, , (mthSee also Gillespie and L “in Re'f. 4, „ Ta,Jmett and Deyrup, J. Am. Chem. Soc, , “ a"d D'n°', 55 ionn ficrviN, , S 3\d Hammett’ ibid > ^ 1758 0937), , (, , 'man, kuivila, and Garrett, ibid., 67, 704 (1945)., , 33)-, , ^, , 73, 3561
Page 116 :
100, , -, , Acids and Bases. Nucleophiles and Electrophiles, , absolute hydrofluoric acid, yielding solutions of easily measurable conductivity/5, indicating that such hydrocarbons are acting as bases in the reaction,, ArH + 2HF —> ArH+ + HFy, The hydrocarbons may be recovered by prompt dilution of the hydrofluoric acid, solutions with water. The basicity of benzene may be demonstrated in another, way, if it is treated with D2SO4, deuterium atoms from the deuterosulfuric acid, are found to exchange with the hydrogen atoms on the ring. This exchange is, best interpreted in terms of the equilibria:14, , C6H6 + d2so4, , C6H5D + dhso4, , Note that the proposed cation intermediate I is the conjugate acid of benzene., , Quantitative Evaluation of Acidity. Concentrations vs. Activities, Comparison of the acidities of a number of solutions in a single solvent system, using indicators or potentiometric measurements is generally straightforward,, and it is a relatively simple matter for a worker to determine which of two acids, is the stronger (provided, of course, both are not weaker acids than the solvent, nor so strong that complications due to the leveling effect set in). The results, of such comparisons lead to an interesting and important question, “Why are, certain acids stronger than others?” Why, for example, should HI be stronger, than HF, phenol be stronger than ethanol, formic acid be stronger than acetic, acid, and jfr-nitrobenzoic acid be stronger than benzoic acid? Each of these, questions concerns the effect of structure on reactivity, a complex subject that, will be treated at some length in Chapter 7., Quite apart from this, however, is the perplexing question of the compari¬, son of acidities in different solvents. In deciding which acid catalyst to use for a, given reaction, the chemist frequently faces questions such as,, , Which of the, , three media is more acidic: dilute nitric acid in water, hydrogen chloride in, benzene, or anhydrous formic acid?” This type of question poses the double, problem of the methods of measurement and the interpretations of such meas¬, urements. To those whose early chemical training was based chiefly on the, ,s Kilpatrick and Luborski, J. Am. Chem. Soc., 75, 577 (1953)., u Ingold, Raisin, and Wilsin, J. Chem. Soc., 1936, 1637. It might be argued that the occur¬, rence of exchange does not, in itself, prove the intervention of cation I as an J^^n^The, that is, that the observed exchange could take place by direct disp acemen o, >, •, latter path is not, however, consistent with our present knowledge of the mechanisms y, similar aromatic substitution reactions occur (Chap. 11).
Page 117 :
Quantitative Evaluation of Acidity, , -, , 101, , study of dilute aqueous solutions, for which acidities could be compared merely, by comparing H30+ concentrations, it may seem perhaps a little surprising, that the quantitative comparison of acidities should become so much more, complicated when more than one solvent is considered. However, first let us, admit that the evaluation of acidities of aqueous solutions is not ideally simple,, for the more convenient methods of estimating hydrogen ions (potentiometric, measurements or experiments with indicators) generally yield the activity, a,, of hydrogen ions (a sort of “effective concentration”) rather than the true, concentration, c. The ratio of these two quantities, a/c, is the activity coefficient, y., , Activities, it will be recalled, are the quantities that must be used in equilibrium, constants for reactions in solutions if it is desired that such constants do not, vary when extra salts are added. Thus for the ionization of acetic acid in water,, HA + H20, K, , H30+ + A', , _ alliO+aA- _ CHjo, , 6q, , flH,, , ~, , +CU-. ., X, , CHA, , y H3Q+, ,+y' Aa A, , (The solvent generally does not appear in equilibrium constants.), In dilute aqueous solutions, activities lie very close to concentrations;, indeed, the activity coefficients of ions may be regarded as measures of depar¬, tures from the ideal behavior that would prevail at infinitely dilute solutions., In more concentrated solutions, electrostatic attraction between positive and, negauve ions cuts down the “freedom” of both, causing the activities of ions to, fall well below their concentrations. For very concentrated solutions, the activi¬, ties of ions sometimes rise above their concentrations; in such cases, a sizable, fraction of the water present has been incorporated into the hydration shells, commhtall8 ff, T’ ^ th"e, ^ ‘°, What We maV term "on* . .Y ( °r'herC 15 “ersy on this Poil«) “normal dilution effects ”, uv.ty coefficients for single ions ordinarily cannot be determined In, , dons.7,5 Such a set nf, , •, , evaluated indirectly (p., , ' 1 fimteIy dilute aqueous solu^, , irealmeniptcribed in, , S°lu'i0ns ** ,h' Debye-Huckel, , SSS sol f’, , holrl, , P, , efficients, must be, , degme,ate, , L., , example, Moore, P^Z, , i r u-U 10nS’ see ^arned and Owen The, , Phvri, , /, , . ', , or treatment, , PP- 40 Jr1, >, , • Am. Chem. Soc., 75, 565 (1953)., , of more con-, , Solutions,, , Rein.
Page 118 :
102, , Acids and Bases. Nucleophiles and Electrophiles, , Dielectric Constants. Formation of Ion Pairs, Triplets, and other, Aggregates, A dilute aqueous solution of a strong electrolyte (which we may arbitrarily, designate as A+B~) contains solute mainly in the form of ions of the type, /f(H20)+ and j5(H20)~. Each ion is thus surrounded by a “hydration shell”, of water molecules. Such shells may be several molecules thick; the number of, water molecules per ion and the tightness with which they are bound depend, upon the cation and anion involved. The ions in such solutions are quite “free”, in the classical sense and are often represented simply as A+ and Br ions. As we, increase the concentration of the electrolyte or, more particularly, if we sub¬, stitute an electrolyte containing polyvalent ions (such as d+2£-'-), the percentage, of solute existing as free ions falls and more and more of the ions come together,, forming relatively small aggregations, the simplest of which is the ion pair.17, Although it is sometimes convenient to represent such pairs as species of the, type A+B~y it should be remembered that such pairs have associated with them, a number of solvent molecules, one or more of which may separate A+ and B, within the pair itself. The ion pair will behave as a single particle in its effect, on the freezing point and vapor pressure of the solvent; moreover, a solution, consisting largely of ion pairs will be a much poorer conductor than would be, expected if the solution consisted chiefly of free ions. On the other hand, the, spectrum associated with a given ion will be much the same whether the ion, is free, is part of a pair, or part of a higher aggregate. In very concentrated, aqueous solutions appreciable portions of dissolved electrolytes exist as “ion, triplets” (of the type ABA+ and BAB~), as “quadruplets,” and even as higher, polymers. We see then why it is necessary to distinguish between ionization, and dissociation., Such a distinction is even more important for nonaqueous solutions, for, water is almost unique among the solvents commonly used by organic chemists., Table 4-3 lists the dielectric constants D (page 59) of a number of solvents., These are rough measures of the relative abilities of the solvents to fadelate the, separation oj positive and negative ions in solution. The dielectric constant of water is, , about 50 percent greater than that of formic acid, several times as large as, those of the lower alcohols and acetone, ten times as large as that of acetic acid,, and many times greater than the dielectric constants of ether and benzene., This means that ion association, such as occurs in concentrated aqueous solutions,, , 57,, , n A detailed picture of ion pairs and higher aggregates is «iven^”USj, 673, 954); and by Basolo and Pearson, Mahan,sms of Inargamc R'acUons, John, , (i, , .ley, , Sons, Inc., New York, 1958, pp. 376-385., R .., d Stokes, Electrolytic Solutions,, ,s These values are taken m large partRob^thc^ise stated, they refer, Butterworth’s Scientific Publications, London, 1955, p. 448. units, to liquids at 25 .
Page 119 :
The Use of Indicators in Media of High Acidity, , -, , 103, , may occur in vety dilute solutions of electrolytes in alcohols, ketones, carboxylic, acids, or in ethers. Indeed the concentrations of lone ions in such nonaqueous, solutions tend to be very small. A strong acid, HA, added to a solvent, S, of low, dielectric constant, may well undergo complete ionization to form the conjugate, acid of the solvent, but this conjugate acid will be tied up in ion pairs of the, type SH+ff- or in higher aggregates., , Table 4-3. Dielectric Constants of Some Liquids, Liquid, , Liquid, , D, , Hydrogen Cyanide, Formamide, Sulfuric acid, Water, , 123 (15.6°), , Ethanol, , 24.2, , 110, , Acetone, , 22, , Formic acid, Methanol, , (20°), , D, , Acetic acid, , 7.1, , 79, , Ether, , 4.5, , 50, , Benzene, , 2.3, , 31.5, , Pentane, , 1.8, , 110, , We now appreciate, at least partially, the difficulty encountered in con¬, structing an acidity scale that applies to a number of different solvents and that, will measure on a single yardstick the proton-donating abilities of various solu¬, tions in various of these solvents. Not only must we compare different acidic, species in different media, but we must also consider solutions under conditions, where the degree of association of the ions present may vary widely.19 A single,, thermodynamically rigorous acidity scale, applicable to solvents ranging in, basicity from sulfuric acid to ammonia (and covering the range of dielectric, constants lying between those of liquid hydrogen cyanide and pentane), seems, at present, to be completely out of the question, and even an approximate scale, covering a large number of solvent systems is probably too large an order, Nevertheless, more limited success has been achieved in setting up acidity scales, covering more modest, but still substantial, ranges in solvent character., , ThC Function!”iCa,0rS, , Medi° °f, , ^, , Acidi^ The Hammett A0, , best applicable^!!, ^ S°'Vent, “ that P™P°-d bY Hammett,, ^est applicable to acidlc media ranging in dielectric constant from about 50, of^ndiemors—that ^s'^ b ^ "°,(a.nhydr0US sulfuric acid)- This scale makes use, gators that is, bases (designated In) tha, are converted partially by, JohnWU^dWSohnshi '"v""]0" °f aciditY scales', , * to Hammett and Devruo ./ aI r,, , c, , lbtd-’ 56’ 827 (1934)- W paul and Long, Chem. ReZ'.’, 57,, , Inc‘> New York> 1954, pp. 122-155., W Hammett and Paul,
Page 120 :
104, , Acids and Bases. Nucleophiles and Electrophiles, , acidic solvents (designated SH) to their conjugate acids (designated InH+)., Success in working with indicators depends upon accurate estimate of the ratio, (InH+)/(In); this is generally done spectrophotometrically or colorometrically., Note that such measurements yield the ratio of the concentrations of the two, forms of the indicator, rather than the ratio of activities. For work in very, acidic media, very weakly basic indicators (for instance, substituted nitroanilines, and azobenzenes) must be used, otherwise the fraction of the indicator pres¬, ent in the basic form, In, becomes too small to measure. For the present dis¬, cussion, we shall compare indicators by considering the acidity constants of, their acid forms (which we shall designate as A’/„//+ values)., Although we recognize that a given acid ionizing in two different solvents, (and perhaps in two different mixtures of the same two solvents) is actually, undergoing two different reactions that should have different equilibrium, constants, thermodynamics prefers to regard the acidity constant of a given, acid at a given temperature as a fixed quantity, independent of solvent. Ap¬, parent variations of the acidity constant are then taken as reflections of varia¬, tions of the activity coefficients of the participating species. Letting a’s represent, activities, C’s represent concentration, and, , 7’s, , represent activity coefficients,, , we may write the acidity constant of indicator InH+ as, KlnH +, , Qn+Qln, Cjn, . . 7 In, - = 7;- X - X a H +, ajnll+, W„//+, 7 InH +, , (1), , or, — log aH+, , pKa = log, , (2), , Determination of the pKa value for an indicator such as/>-aminoazobenzene, _which is about as basic as the familiar indicator, methyl orange—is straight¬, forward, since the concentrations of both the acidic and basic forms are easily, measurable in solutions of dilute acid where the, , 7, , values approach 1. Assume, , that the pKa value for />-aminoazobenzene has thus been found to be 2.80., Now, consider an HC1 solution containing both this indicator (designated In), and the weaker base />-nitroaniline (designated In'). The concentration of acid, must be raised in order to convert measurable amounts of the latter amine, to the acid form; in which case almost all of the p-aminoazobenzene is converted, to its own acid form, leaving only a small (but measurable) amount in the basic, form. Expressions such as Equation (2) may be written for both indicators in, the solution, then subtracted., , PK., , ,, /ClnH+Cln'\ j_ lrt„, - PK-' - K«, + '°g, , (3), , Since the value of pK. is known and ,he value of «he first term - the right of, equation (3) may be estimated colorimetncally, we may calculate p .
Page 121 :
The Use of Indicators in Media of High Acidity, , -, , 105, , know the value of the logarithmic term involving the activity coefficients. It, would be particularly convenient if this term were zero; happily it turns out, to be negligibly small for the high dielectric-constant media under considera¬, tion. In other words, although the ratio yin/yinH+ varies if the medium is, changed, the ratio is nearly the same for all uncharged indicators in a given solution., Experimentally, this means that the concentration term in equation (3) may, be shown to be constant for a number of different acid solutions containing, the same two indicators, and from the value of this term we may calculate the, value of pKa' (for /i-nitroaniline). This turns out to be 1.11; and by carrying out, similar measurements with both this base and the still weaker base,, , 2,4-nitro-, , chloroaniline, we may show that the apparent difference in the pKa values for, these two very weak bases is 2.02 units in HC1 solutions, 2.08 in aqueous, HN03, 1.96 in aqueous H2S04, and 2.02 in aqueous HC104. To a rather good, approximation, therefore, the difference in the apparent pKa values of two, indicators is independent of the solvent. The procedure may be repeated by, comparing, in turn, weaker and weaker bases in solutions that are more and, more strongly acid, arriving ultimately at bases so weak that only concentrated, sulfuric acid will convert appreciable fractions to the respective conjugate, acids. The least basic of Hammett’s indicators is picramide (2,4,6-trinitroaniline) whose pKa value is —9.29, about 12 units less than pKa for jfr-aminoazobenzene., , Typical, , intermediate, , indicators, , are />-nitroazobenzene, , (pKa,, , — 3.3), benzalacetophenone (pK, —5.6), and anthraquinone (pK, —8.2) (note, that the very weak bases have very negative pKa values)., To understand the acidity scale that is constructed with these indicator, constants, let us rewrite equation (1):, , aa., n, , ., , I?, , Wn/7 +, , 1 InH*, , (4), , In a solution containing measurable amounts of the two forms of an indicator, whose constant ts known, we are prevented from calculating the activity of, V, , ogen ions by our ignorance of the value of 7,However since this, , ratio, to a very good approximation, is independent of the indicator used but, is characteristic of the medium, the quantity «.♦, , Moreover, , S°1U‘i°n, , 0 1S measurable whereas aH+ is not, , Wp, , defining h0,, ho = aa+, , y'In, 7 InH +, , (which may be abbre_, , K,Inll*, , «« as is, itself., cPP, r, ,, 'Ce a so ^rom the equation, , C,I nil+, CT, In, , (5)
Page 122 :
106, , -, , Acids and Bases. Nucleophiles and Electrophiles, , since ho is proportional to the ratio of concentrations of the acidic and basic, forms of the indicator. Note also that ho becomes identical to <2U+ in very dilute, aqueous solutions where the, , 7, , values approach unity. The properties of h0, , thus make it a rather useful quantitative measure of the relative proton-donating, abilities of solutions of high dielectric constant and high acidity. The quantity, Ho is defined by Hammett (in analogy to pH for hydrogen-ion activity) as, , — log Ao. Typical H0 values are: +7.0 for water, —1.5 for, — 2.0 for 6.9 molal HC1, —2.63 for, 8, , 8, , 6, , molal HNO3,, , molal HCIO4 (and about the same for, , molal H2SO4), —5.54 for 70 percent H2S04, and —10.60 for, , H2SO4. (On this basis, , 100, , percent sulfuric acid is over, , 100,000, , 100, , percent, , times as efficient, , a proton donor as 70 percent sulfuric acid.), In order that the same acidity scale measure also the ability of a solution, to donate a proton to a negatively charged base A~, it would be necessary that, the ratio ^ha/^a- for negative bases be equal to the ratio 7B//+/7B for neutral, bases in a given solution. Since these two ratios are practically never the same,, we might then ask, “Is there a different but analogous acidity function that, measures the ability of a solution to protonate negative ions?” Such a function,, designated by, , (in analogy to H0 for acidity toward neutral bases), was, , proposed by Hammett over twenty-five years ago, but it is still an open question, as to whether a useful //_ scale can be constructed. The ratio 7H4/7A- varies, greatly with the structure of the anion in ethanol-water solutions,^ and similar, variations occur for neutral acids in mixtures of isopropyl alcohol and ethanol'1, and almost certainly for acids in solvents of even lower dielectric constant. An, H_ scale then, if it can be constructed at all, must apply (as does the H0 scale), , to solutions of high dielectric constant. Indicators of the charge type In~ would, be converted very nearly completely to their conjugate acids by the more, acidic of such solvents (for example, H2S04, HF, and HGOOH), but measurable, amounts of both forms of the indicators should exist in hydrazine, formamide,, dimethyl sulfoxide, and in aqueous solutions of these. An //_ scale has been set, up by Deno** for mixtures of water and hydrazine, using as indicators such, uncharged weak acids as 3-nitrocarbazole (II),, (III),, , 4,4'-dinitrodiphenylmethane, , and jfr-nitrobenzeneazoresorcinol (IV), proceeding successively from, , indicator to indicator and from solution to solution in the same manner as was, described for the construction of Hammett’s H0 scale. The //_ value for pure, water is 7.0, whereas the values for 30 and 60 percent hydrazine have been, found to be 13.15 and 15.93, respectively. Thus, 30 percent hydrazine is almost, a thousand times as effective a proton donor (to negatively charged bases), as is 60 percent hydrazine. Deno’s scale, however, cannot be said to be genera ,, for it is based on measurements involving only a single pair of solvents.a Hine and Hine, J. Am. Chem. Soc., 74, 5266 (1952).
Page 123 :
The Grunwald Acidity Scale, , -, , 107, , The Grunwald Acidity Scale, The Grunwald scale*4 is one of the more interesting acidity scales proposed, during recent years, although it has been applied to organic systems in only a, very few cases (see, for example, p. 192) and although its theoretical basis, appears somewhat less firm than that of the Hammett scale. This scale corre¬, lates acidities in mixtures of ethanol and water, thus covering a range of di¬, electric constants between 24 and 79. The data on which this scale is based are, a collection of acidity constants for a number of carboxylic acids and for a num¬, ber of substituted ammonium and anilinium ions in ethanol, in water, and in, various mixtures of the two solvents., The carboxylic acids are more fully ionized in water than in ethanol- their, ionization involves separation of charge and this is greatly facilitated in the, substituted^ S, dleleCtriC COnStant' °" the °ther hand’ ,he acidities of the, ubstituted ammonium ions show considerably less variation with solvent The, acid reactions of these ions,, , BH+ + S-+SH+ + B, involves no marked separation of charge and is thus far less sensitive to variation, m the ionizing power of the solvent.**, variation, "Deno, J. Am. Chem. Soc., 74, 2039 (1952), for the, , -a4y5bean“, , Hammett’s, , T~TUur[C add and water- Moreover, Helv. Chm. Acta, 27, 348 (1944), to establish an H, , solutions is subject to the same objection., , ~, , lT, , C, , 0 scale,, , “ °nly °ne Pair, fcbwarzenbach and Sulzberger,, by measurements in aqueous, , and Berkowit2', , 73’, , oi, Whereas the reverse is true for others (for
Page 124 :
108, , -, , Acids and Bases. Nucleophiles and Electrophiles, , Each of the acidity constants, designated, , Ka,, , is compared with the constant, , for the same acid in pure water, designated KXJ, using for such comparisons the, logarithms of the constants. The constants for the carboxylic acids were found to, , fit nicely into a relationship, log K% - log, , Ka, , = log / + mAY, , or, using the familiar abbreviation whereby pA is defined as —log A,, pKA - pKJ = log / + mAY-, , Here,, , (6), , is a constant depending on the acid being considered, and / and Fl¬, , are constants depending upon the solvent being considered. The term, mA, is a, rough measure of the sensitivity of the acid to changes in the ionizing power of, the solvent mixture. For benzoic acid, whose pK value changes 2.59 units in, going from water to 80 percent ethanol, mA is 1.57; for formic acid, whose pK, value changes only 1.89 units as a result of a similar change in solvent, mA is, 0.77. The quantity, F_, called by Grunwald an activity function, is in essence a, measure of the relative ability of the solvent mixture to repress ionization of, —COOH groups; typically, F_ is 0.35 for 20 percent ethanol but 0.96 for 80, percent ethanol. (The significance of log / will become clear in a moment.), The acidity constants of the substituted ammonium and anilinium ions obey, a similar relationship; that is,, pKB h — pK%a = log / + tubbY o, , (7), , Again, mBn is a constant associated with a particular ammonium or amlinium, ion, and F0 and / depend upon the solvent. The Y terms are different for the, two types of Br^nsted acids, indicating, as expected, that the relative abilities, of the various solvent mixtures to repress ionization of —COOH groups are not, linearly related to the relative abilities to repress the ionization of the various, substituted ammonium ions (the subscripts indicate simply that the conjugate, bases of the carboxylic acids are negatively charged whereas those from the, ammonium salts are uncharged). It is important, however, that the / terms, which at first glance might appear to have been mcluded merely, , to make he, , answer come out right,” are the same for both series of acids,■ that is, / is a functio, only of the solvent system. Equations (6) and (7) are, ships; that is, thermodynamics does not predict them and they are almost, tainly approximate., , ___—, , example, N, N-dimethylanilinium and, , of a number, , of such, , ions are observed in, , have "not been simply explained. What, , these maxima, so far as the present author knows have no, should be emphasized is that there is My t^be «, , ^ variation in the, tQ water> in contrast to, , acidity constants of substituted ammo, ,., id in „0ing from ethanol to water,, the 100,000-fold increases in the constants of carboxylic acids in go g
Page 125 :
The Grunwald Acidity Scale, , 109, , The thermodynamic relationship proposed to relate “apparent ionization, constants” in these mixtures to the ionization constants in dilute aqueous solu¬, tions is, , KJ, , =, , (^sr) x (Xf)x (?& =, , x ft), , (8), , Note the two sets of activity coefficients; the 7 terms convert the concentrations, of the species (C values) to their activities (referred to very dilute solutions in, the solvent itself). These coefficients for the ions may be readily calculated by the, Debye-Huckel treatment, and the 7 value for HA is assumed to be unity; hence,, these terms need not concern us further. The 7' terms, which further link the, activities of the species in the solvent at hand to the activities of the correspond¬, ing species in very dilute aqueous solutions, are the designated “degenerate activity, coefficients.” If we assume that the latter have meaning (an opinion which is, by no means unanimous), it becomes interesting to rewrite equation (8) in the, logarithmic form,, pKA - pKJ, , log y'H+ + log, , (9), , Comparing Equations (9) and (6), we see that (log f + mAYJ) is equal to, (log Vn+ + log, , J. Moreover, the two expressions have the same form; one, , term of each (log } and log, , is independent of the acid under consideration, , but depends only on the solvent. The other term in each of the expressions, depends not only on the type of acid but also on the particular acid being treated., A completely analogous argument applied to the substituted ammonium and anilimum salts shows that for these (log / + Y0mBB) equals (log 7^+ + log, , V, , and again we may note the same correspondence in the nature of the terms*in, e two expressions. Since the properties of the series of empirical log / values, the na'tlTthe1105^"13"?, ^, ^ ^, (indeP<^ence of, the acid considered, approach to 0 for 100 percent water and, quantitative correlation of pK values), Grunwald’s scale makes use of the, , f, , then'are ‘the^T *7 “T” W°U'd, ^ °f, ^ffidents.*’ Thesi, then are the values by which hydrogen-ion activities in thn, , hydrion d~ ^
Page 126 :
110, , -, , Acids and Bases. Nucleophiles and Electrophiles, , found to be 3.6 and 14.2, respectively. This means that a solution in 65 percent, EtOH, having a total lyonium ion (H30+ + EtOH+) concentration of 0.01, molar, is approximately 3.6 times as efficient a proton donor as is aqueous, 0.01 molar HC1, whereas a similar solution in 80 percent ethanol is approxi¬, mately 14 times as efficient. These high activity coefficients are quantitative, reflections of a trend that has already been noted (p. 96); that is, a given acid, is strongest in the solvent of least basicity., , Acid-base Catalysis, To the organic chemist, it would be difficult to overemphasize the importance, of acid-catalyzed and base-catalyzed reactions. Among the many reactions, falling into these two very broad categories are the saponification, hydrolysis,, and synthesis of countless esters and amides, the hydrolysis of anhydrides and, alkyl and acyl halides, the bulk of carbonyl-addition reactions, and the aldol,, Claisen, Perkin, and Michael condensations. Of the 76 chapters describing re¬, actions of general utility in the eight volumes of Organic Reactions, almost half, are devoted to reactions that are catalyzed by acids, by bases, or by both/7, The strongest acid that can exist in large concentrations in a given solvent, is the conjugate acid derived from that solvent (the lyonium ion). This is gen¬, erally the most effective acid catalyst for reactions carried out in the solvent., Often it appears to be the only effective acid catalyst, and weak acids that, are present appear to accelerate the reaction only to the extent that they con¬, vert the solvent to the lyonium ion. Such reactions are said to be subject to, specific, , lyonium-ion, , catalysis, , (or,, , for, , aqueous solutions, in which the large, , majority of quantitative studies have been carried out, specific hydromum-ion, catalysis.) Among these are the acid hydrolyses of a number of esters and acetals., One of the earliest reactions of this type to be examined was the hydrolysis of, diethyl acetal, CH3CH(OC2H6)2, in formic acid-sodium formate buffer solu¬, tions/5 The formic acid concentration in a number of such solutions was varied, from 0.02 to 0.18 molar, but the ratio (HCOOH/HCOO~) was held at 2.96,, enough sodium chloride was added to each solution to bring the total concen¬, tration of positive ions to 0.100, keeping activity coefficient effects for all solu¬, tions the same. Each solution thus had the same H30+ concentration, and it was, further found that the rate of acetal hydrolysis was, within experimental error, the, same in each. (Note, however, that the concentration of formic acid was relative >, low even in the most concentrated of the solutions.), 47 For a detailed discussion of this subject, see Bell ^nrview '^e F^o^and" Pearson^, Press Oxford 1941. For a shorter but somewhat more recent vie ,, John Wiley and Sons,, ™, >>, v u Brdnsted and Wynne-Jones, Trans. Faraday Soc., 25,, )
Page 127 :
Acid-base Catalysis, , -, , 111, , Similarly, a number of base-catalyzed reactions appear to be catalyzed, specifically by the conjugate base of the solvent in which they are carried out., Among these are the ring closure of /3-styrene chlorohydrin, C6HB—CH—CH2C1,, OH, to the oxide, C6H5—CH-CH2, in water*5 and the formation of benzalaceto-, , \ /, O, , phenone, C6H5CH=CH—C—C6H5, from benzaldehvde and acetophenone in, , II, , O, ethanol.30 These reactions are said to be subject to specific lyate-ion (hydroxide- or, ethoxide-ion) catalysis., On the other hand, a large number of reactions suffer either general acid, or general base catalysis. Consider for example the rate of hydrolysis of ethyl, orthoacetate, CH3C(OC2HB)3, in a series of buffers derived from m-nitrophenol, and its sodium salt. This rate increases as the concentration of the phenol is, increased—even under conditions where both (H30+) and total ionic strength, are kept constant—showing that this hydrolysis, in contrast to the hydrolysis of, diethyl acetal described above, is catalyzed by 7/2-nitrophenol (and, presumably,, by other acids), as well as by H3O"1".*3 A somewhat analogous experiment in, basic catalysis is the bromination of bromoacetylacetone,, BrCH 2—C—CH 2—C—CH 3,, , in buffers of chloroacetic acid and its sodium salt.*' This reaction is catalyzed, not only by OH , but also by chloroacetate ion and by water. The halogenations of several additional ketones are also subject to general basic catalysis and, some are subject to general acid catalysis as well. Of particular interest is the, mutarotation of glucose, which is catalyzed by acids and by bases," but even, more effecttvely when both an acidic and basic center are appropriately located, in the same molecule33 (see p. 139)., y, Catalysis by more than one acid or base in a given solution complicates the, tnettc ptcture. For a reaction between species X and Y (first order in both), £££**, , **—, rate = k(X){Y){ H3@+), , 00), , t, , s-x""-, , ” Swain and Brown 1 A™ nu, n, I, ’, 2554 (1927)., rown’, Am. Chem. Soc., 74, 2538 (1952).
Page 128 :
112, , -, , Acids and Bases. Nucleophiles and Electrophiles, , whereas if general acid catalysis prevails, the observed rate will be the sum of, terms, each corresponding to one of the acids present in solution., , ra^e — (A)(l )[^'u,0 + (H30+) + *h20(H2O) + £ha(H/1) + knA'(\\A'), , •, , •, , • ], , (11), Here the k values are catalytic constants for the reaction at hand, one constant, being associated with each of the various acidic species present in solution., (Analogous expressions may be written for the rates of reactions subject to, general basic catalysis.) The stronger the acid, the higher is its catalytic constant, for a given acid-catalyzed reaction, but, as we shall see, the relationship between, ionization constants and catalytic potencies of acids is logarithmic rather than, linear. If the reaction is subject to both general acid and general base catalysis,, the kinetics may be even more complex., In attempting to distinguish general acid catalysis from specific hydroniumion catalysis, it is important that kinetic experiments be extended to solutions in, which the total concentration of the weaker acid or acids heavily exceeds the, concentration of hydronium ion. The opening of the epoxide ring of epichlorohydrin by iodide ion in acetic acid-acetate buffers, HA, , I- + CH2, , CH—CH2C1, , ICH2—CH—CH2C1 + A~, , (12), , OH, , O, , might appear to be catalyzed only by HsO+ if fairly rough measurements are, carried out on solutions in which the ratio (H0Ac)/(H30+) is less than 1000;, but if this ratio is boosted to about 100,000 by using higher concentrations of, buffer, catalysis by unionized HOAc becomes clearly evident.^ It is likewise, possible that other reactions that appear subject to specific catalysis would, if, examined at higher buffer concentrations, show evidence of general catalysis., The possible reaction mechanisms suggested by the various types of acid, and base catalysis will be considered after the fundamentals of kinetics have been, briefly reviewed in Chapter 6. At that point there will also be discussed the, closely related question as to why the rates of some acid-catalyzed reactions are, governed by the concentration of HsO+ ions whereas others depend upon the, Hammett H0 function., « Swain, 7., , A*. Chm. So*.,, , to its conjugate base A, , 74, 4108 (1952). Since there is a, , during this reaction, it cannot, , in the classical sense. However, the reaction ,s, , e sai, , of th,eadd hU, acid-catalyzed, l~on« and is con-, , reactions such as the hydrolysis of acetals an, tQ the “base-catalyzed”, veniently considered along with these. Similar y, w, reactions is not, halogenations of ketones even though the base consumed in one step, regenerated in subsequent steps.
Page 129 :
The Br0nsted Catalysis Law, , 113, , The Brtfnsted Catalysis Law, We have seen that for a reaction subject to general acid (or general base), catalysis, the strongest acids (or strongest bases) generally are the best catalysts., Typically, the dehydration of acetaldehyde hydrate,, HA, , CH3CH(OH)2—^ ch3cho + h2o, has been found to be catalyzed by 32 carboxylic acids.35 Catalytic constants, (in liters per mole-minute at 25°) associated with three typical acids are:, Acid, , Ka, , £cat. (liters per mole-min), , Propionic, , 1.35 X 10-6, , 18.0, , Phenoxyacetic, , 7.6, , X 10~4, , 92.0, , Dichloroacetic, , 5.0, , X 10-2, , 773, , If the logarithms of the catalytic constants for the 32 acids be plotted against the respective, pKa values, the points are found to lie very close to a straight line. Similar plots for, other reactions subject to general acid catalysis also yield points lying near, straight lines although the slopes of such lines vary from reaction to reaction., For general base catalysis, similar linear relationships exist between the loga¬, rithms of the catalytic constants for a given reaction and the pKb values of the, various basic catalysts. Again, the slope of the line representing such a relation¬, ship will depend upon the particular reaction under consideration., Mathematically, this type of relationship may be expressed, log £cat. = a log Ka +, where the constant a is the slope of the line and the constant b an axial intercept,, (actually the value of the catalytic constant, £cat, for a hypothetical acid whose, dissociation constant is unity). This, together with a similar expression for base, catalysis, constitutes the Br0nsted catalysis law. It is subject to two limitations., First, and most important, the relationships hold only if the catalytic action of, acids or bases of the “same type” are being compared. For example, aliphatic, mtro compounds and /3-diketones (weak acids with strengths comparable to, ‘h°Se °f‘hc leSS acidic Panels) will also catalyze the dehydration of acetalde¬, hyde hydrate, , but their catalytic constants are less than one twentieth those, , for phenols of comparable acidities. On the other hand, the catalytic constants, , Wasted c” H T*, ‘h°Se ** W°U‘d ^ Predicted from the, talysts law using a and b parameters derived from the data on caralTouvh, 8, , Tme U tme f°r r'aCti0nS SUbjeCt to, catalysieqUa“°n (13) ma>' aPP‘y t0 “‘alysis both by carboxylate ions and by, , Bell and Higginson, Proc. Roy. Soc. (London),, , A197,, , 141 (1949).
Page 130 :
114, , -, , Acids and Bases. Nucleophiles and Electrophiles, , substituted anilines, we should expect different a and b parameters for the two, types of bases., A second stipulation concerns statistical corrections that sometimes must, be applied to acidity or basicity constants. For example, even though the acid, constant of benzoic acid (6.3 X 10~5) is about twice that of the hydrogen fumarate anion (V) (Ka = 3 X 10~5), it might be argued that the proton-donating, abilities of the two acids are nearly equal. The acidity constant may be regarded, as the ratio of rate constants for the forward (proton-leaving) and reverse, (proton-returning) reactions. Now the conjugate base of the hydrogen fumarate, ion has four oxygens to which a proton may return, whereas the benzoate ion, , o, , O, C-CH, , X, , X-, , HO, , k, , O, , /, , CH-C, %, , kr, , /C-CH, Q/, , O, +, , \lH—, , H+, , o_, , O, , V, has only two such oxygens. This twofold statistical advantage for proton return, to the fumarate ion results in a twofold lowering of the acidity constant of its, conjugate acid. Since this effect is not, however, connected with the act of, proton leaving, a “fair” comparison between the proton-donating ability of the, hydrogen fumarate ion and that of any monocarboxylic acid requires that the, ionization constant of the former by multiplied by 2. By the same argument, the, basicity constant of an ion such as H2N—CH2—CH2—NH+ should be doubled, if a Brpsted law relationship being set up also involves monofunctional amines.35, At first glance, it might be suspected that a log-log relationship such as, the Br^nsted catalysis law was drawn up by a worker who felt that there must, surely be some type of quantitative relationship between the catalytic constants, of acids for a given reaction and their ionization constants, but who found that, simple linear, quadratic, cubic, and various other algebraic relationships did, not hold. If this were so, the Br^nsted relationship would be of very limited, interest. Let us recall, however, that the logarithm of a dissociation constant is, proportional to the free energy of dissociation. Furthermore, in Chapter 6 it will, be shown that the logarithm of the catalytic constant is proportional to the, free energy of the initial proton-transfer step in a reaction subject to general, acid or general base catalysis. The Br0nsted catalysis law is then a linear rela-, , *, , 9SS&SSSSSS&, .he subsided phenols are appreciably more effec-, , tive catalysts than are the fatty acids.
Page 131 :
Electron-pair Transfer. The Lewis Acid-base System, , -, , 115, , tionship between free energies—the first of such relationships that we have encoun¬, tered in this text, and also chronologically the first significant relationship of, this type proposed. Linear free-energy relationships will be further discussed in, Chapter 7., , Electron-pair Transfer. The Lewis Acid-base System, A more general interpretation of acid-base behavior was proposed by G. N., Lewis in 1923s7 and has become widely adopted since then, most especially, for descriptions of reactions in nonhydroxylic solvents. We have seen that a, proton is acidic because it may become bound to a basic species through an, unshared electron pair on the latter. However, a number of additional ions and, molecules (which we may call electron-pair acceptors) may likewise attach them¬, selves to the unshared electron pairs of bases. Lewis proposed that all such, acceptors be termed acids and that an acid-base reaction be simply the donation, of a pair of electrons from the base to the acid., Lewis acids are of several types. First, there are a number of compounds in, which one atom has less than a full octet of electrons. Typical members of this, class are trimethylboron, boron trifluoride, and sulfur trioxide. The reaction of, these acids with the Lewis bases, ammonia, ether, and pyridine are shown below:, Acid, , Base, , Me, , Adduct, Me, , H, , Me: B, , N:H, , Me, , \, >, , Et, , :o, :Q, , I, , Me—B:N—H, , H, , :F, •• • •, F: B, •• ••, :F, , 0:S, •• ••, , H, , O, , +, , >, , Chemical Catalog Co., New York,
Page 132 :
116, , -, , Acids and Bases. Nucleophiles and Electrophiles, , are often called nucleophilic reagents or nucleophiles since they generally, attack another molecule at a site where the atomic nucleus is poorly shielded, by the outer electrons., Positive ions may also be regarded as Lewis acids. Typically, the lithium, and silver ions will react with methanol and ammonia, respectively:, Acid, , Li +, , Base, , -f-, , ., , Me, , /, Li:Q, , H, , Ag+, , Me, , /, , H, , +, , Adduct, , :N:H, H, , XH, , 1), , ’, , H, , I, Ag:N-H, , * I, H, , (In both of the reactions above, more than one molecule of base may coordinate, with the acidic cation.) In aqueous solutions, many positive ions, particularly, polyvalent ions, are strongly hydrated and are thus Br^nsted acids. The salts, of tripositive iron, for example, form the ion Fe(H20)|3, an acid whose dis¬, sociation constant in water is close to that for H3PO4., Fe(H20)+3 + H20 -> Fe(H20)5(OH)+2 + HsO+, , K = 0.006, , Of particular importance in organic chemistry is a group of positive ions (not, derived from metals), the salts of which, although not unknown, are not gen¬, erally found in ordinary chemistry laboratories. Among these are themtromum, ion (NO£), the nitrosonium ion (NO+), the bromonium ion (:Br), and the, various acylium ions (R—£=0). All of these are thought to be active inter¬, mediates in aromatic substitution reactions (Chap. 11)., Certain halides in which the central atom may hold more than an octet of, electrons may show acidic properties. Tin (IV) chloride combines with ethanol, to form the addition compound SnCl4.2EtOHwhereas titanium (IV) chloride, forms adducts with ethers at low temperatures.” Although the structure, formulas of SnCl4 and TiCl4 might tend to put them in the same class as t e, tetrahalides of carbon (which are nonacidic), it should be remembered that, associated with both tin and titanium there are vacant d orb.tals of relative y, low energies. These may be used in bond formation with donor molecules., Similarly, the pentahalides of niobium and tantalum form addition compounds, ••Thiessen and Koerner, Z. mmg. Chm , 195, >8 <1931 >•, s» Stadnikov and Kaschanov, Ber., 61, 1389 (1 JZo).
Page 133 :
Electron-pair Transfer. The Lewis Acid-base System, , -, , 117, , with benzene and naphthalene, respectively.^0 Elemental iodine may also be, considered with such Lewis acids, for the I2 molecule coordinates with many, basic solvents, as evidenced by the familiar change in color when violet solutions, of iodine in CC14 or aliphatic hydrocarbons are diluted with alcohols, ethers,, or amines., In a similar sense, compounds having C=0, C=N, or N=0 double bonds, or C=N triple bonds are electrophilic. Although the Lewis classification in its, original form did not specifically regard compounds such as acetone or benzonitrile as acids, we see from reactions (14) and (15) that they accept electron, pairs from basic species in the process of coordination. The rate-determining, step in the conversion of acetone to its oxime in basic media is probably the, attack by the conjugate base of hydroxylamine, HO—NH~, on the 0=0, double bond, reaction (14),^ whereas the initial step in the basic hydrolysis, of nitriles is almost certainly the attack by OH~ on the C=N triple bond (15)., Base, , Acid, , H, , Me, , I, HO-N:~, , Adduct, , +, , H, ■>, , ^^C^O:, /, Me, , HO:', , H-, , Me, , I, , I .., HO-N-C-O:, I, *•, Me, OH, I, .., R—C=N, , R—C=N:, , (14), , (15), , Even the simple neutralization of aqueous C02 may be regarded in the same, way:, , H:6:', , +, , 0=C—6:, , ^ baSe aKacks the ™re positive member of the double-, , ror'triDleri, membe!, , o=c=o^, , ^ °f ^, , PUSh'S 3, , ^e more negative, , .t CrCud°Uble b°nd d°es n0t generaII>’ act as a" acidic site unless electron, , situated nelrbTThehdrTn fr°m, , ^, , dect,m'atlraamg group (Chap. 7), i“"a- ~, , anions; in the familiar Michael reaction a, , cart,, , of malonic ester is the basic reagent fin, , pi, , <»F„„k, , H m- a ,, , •, , ^, , 3, , nUmber, , of, , " SUCH 3S 'hC conJuSate base, , (16)l Flu°rmated olefins may also suffer, , ^Funk aiici Nicdedander, Ber. 61B, 1385 (1928), Olander, Z.physik Chem., 129, 1 (1927)., '
Page 134 :
118, , -, , Acids and Bases. Nucleophiles and Electrophiles, , attack by anions, such as occurs in the initial step of the base-catalyzed addition, of ethanol to tetrafluoroethylene (17)., Base, , Acid, , Adduct, , H, , CH(COOEt)2, -~~Jw, , (EtOOC)2Cr, , +, , R—CH=CH—C—R', , R—CH—CH=C—R', , IK, , 6, , -, , EtO', , +, , •'cf2=cf2, , EtO —CF2—CF2, , (16), , (17), , Although the Lewis system allows many species besides the hydrogen ion, to display acidic behavior, the species acting as bases under this system are, largely those that behave also as Br0nsted bases. The important nucleophilic, reagents fall into three broad classes:, (a) Molecules in which fifth- or sixth-group elements have unshared elec¬, tron pairs (such as, amines, alcohols, ethers, and mercaptans)., (b) Negative ions (for instance, halide, hydroxide, alkoxide, sulfide ions,, and carbanions)., (c) Olefins and aromatic hydrocarbons., The basic behavior of olefins and aromatic hydrocarbons assumes some¬, what more importance in the Lewis system than in the Brjzfnsted system. We, have seen (p. 100) that the basicity of aromatic hydrocarbons toward the H+, ion is indicated by evidence of an indirect sort, and the same might be said for, the Brjzfnsted basicity of olefins. On the other hand, a substantial number of, addition compounds of olefins or aromatic hydrocarbons with metal salts have, been isolated.** Typical adducts of this type are: (CH3)2C=CHCH3.ZnCl2,^, (C2H4)2-PtCl2/5 C6H6*AgC104,4S and, these, , C6H6—CH=CH—C6H5-2FeCl3/7 In, , the positive metal ion is presumably linked in some way to the unsatu¬, , rated site in the hydrocarbon molecule. In the benzene-silver perchlorate addi¬, tion compound, the silver ion is not bonded to a particular carbon but rather is, located equidistant from two carbon atoms, sitting, in effect, on t e o es o, two 7r electrons as shown in Figure 4-1., Griswold, J. Am. Chem. Soc., 70, 431 (1948)., , very broadest sense of the term., « Kondakov, et al., Chem. Zentr., 1930, I, 328 ., » Chatt and Wilkins, Nature, 165, 859 (1 'SO)., Rundle and Goring, J. dm. 6W .Sor., 72, ..., V Brass and Tcngler, Ber., 64B, 1650 (1931)., , ,, , ., , listed by Andrews,
Page 135 :
Electron-pair Transfer. The Lewis Acid-base System, , +, , -, , 119, , DfO, , acid, , base, Fig. 4-1. Formation of a, , tt, , Complex, , Thus, this adduct is called a it complex., At present we are not certain as to which of the addition compounds de¬, rived from olefins or from aromatic hydrocarbons are ir complexes and which, have the Lewis acid bound to a particular carbon atom by an ordinary single, (<r-type) bond. If the addition compound is isolable in the solid state and if the, acceptor species is not a hydrogen ion, it is possible (at least in principle) to, locate the position of the acid with respect to the remainder of the adduct, using, the methods of x-ray diffraction. However, a number of such adducts of con¬, siderable theoretical interest (including the intermediates in certain of the, aromatic substitution reactions) are formed only in relatively small concentra¬, tions in solution, and for these we may only speculate as to structural details., In this connection, let us compare the addition compounds formed by HC1 and, aromatic hydrocarbons in the absence of A1C1, to those formed in the presence, of Aid** The former are largely dissociated (even at -79° C) and are color¬, less nonconductors. Furthermore, if the aromatic hydrocarbon is recovered, from such a complex, in which DC1 is substituted for HC1, it may be shown that, Tneufr.‘,he deuterium has entered the benzene ring. (Were it not for the fact, t at HC1 is considerably more soluble in aromatic than in aliphatic hydro¬, carbons, we probably would not suspect that there was any interaction between, the components.) In contrast, the complexes formed in the presence of alumi¬, num halides are colored and conduct the electric current. With these, , if DC1, , is substituted for HC1, the deuterium atoms “exchange” with thehXl, atoms of the benzene ring. Clearly, the two types of complexes are very differ!”, in character, and it seems likely that complexes of the first class are , l, , ,, , (of the type V,), whereas complexes of thesecond Cass, , Aicr, VI, , 2SCS; IS.*—, for distinguishing, , '^nd^mpi;^’^prop^rf.1", , we cannot, , PaP'r' 3 nUmber of criteria
Page 136 :
120, , -, , Acids and Bases. Nucleophiles and Electrophiles, , Substitution Reactions49, The terms “electrophilic” and “nucleophilic” are often applied to substitution, reactions—that is, reactions in which a new covalent bond is formed and an, , old one is broken. Typical nucleophilic substitution reactions are, for example:, Attacking, reagent, , +, , c2h5ct, , HO-, , Product, , Substrate, , :NH3, , CH3CH2I, , nh3, 1, J, —-ch3ch2, , o2n—Cl, , +, , (CH3)3C, , Leaving, group, , Br, , +, , I’, , (18), , ->■ o2n^>-oc2h5 +, , cr (19), , (CH3)3C- OH, , Br" (20), , —>, , 4-, , In these, the attacking reagents and the leaving groups are Lewis bases or, nucleophiles. The electron pair,that bonds the attacking reagent to the product, is furnished by the former species. More specifically, reaction (18) may be, described as a nucleophilic substitution by nitrogen for iodine on carbon, whereas, (19) is a nucleophilic substitution by oxygen for chlorine on carbon. Although, reactions (18), (19), and (20) are classified together due to the similarity in the, net changes, it should not be assumed that the paths by which these three reac¬, tions occur are the same or closely analogous. As we shall presently see, there are, almost certainly fundamental differences in the mechanisms. Similarly, the, conversion of benzhydrol to benzhydryl chloride, H+C1- + (C6H5)2CH—OH —»(C6H5)2CH, , Cl + HOH, , (21), , is a nucleophilic substitution by chlorine for hydroxyl on carbon, .even though, the hydrogen ion is neeessary for the reaction (being, in fact, the species wh.ch, initially attacks the hydrol).50, 19 For * detailed discussion of the classification of organic reagents and ^nons, see, rjd, “mXL ,» ft** C^ry, Cornell University Pres,, I.haea, N.V.,, , ,95V,P„ cohering the mechanism of a reaction as will, , is broken by the following reaction:, H+ _p (C6H5)oCH—OH -» HOH + (CeH6)2CH+-, , which may be considered as electropl11he, on oxygen. Where possible, however, we s ic, , Lid base lur primary, , ccurs Subclasses will be introduced where, classification of reactions on the overall change that occurs, buocia, there is evidence of mechanistic differences.
Page 137 :
121, , Substitution Reactions, , In an electrophilic substitution, an electron-deficient species attacks the sub¬, strate, becoming bonded to it through a pair of electrons furnished by the latter,, and a second electron-deficient species departs from the reaction site. Of by far, the greatest importance in this category is the large group of aromatic substitu¬, tion reactions which includes alkylations, acylations, nitrations, sulfonations., halogenations (except for fluorinations), and diazonium coupling reactions. In, each of these types, the leaving group is simply the H+ ion. The attacking species, may be the incoming substituent in an electron-deficient form—for example, the, NO^ ion in nitra^n or the :Br ion in bromination:, • •, , Attacking, reagent, , Substrate, , [0=N—0]+ + Cl—C6H5, :Br+, , +C6H6, , Leaving, group, , Product, />-ClC6H4—N02 +, , H+, , C6H5Br, , H+, , +, , • •, , or a so-called ‘'carrier" of such a species (for example, CISO3H in sulfonation, or the RC1 + A1C13 reagent in alkylation)., Attacking, reagent, , Substrate, , Products, , O, , O, , r, , 11, , CISO3H, , + CH3C—NH—C6H5 —> />-CH3C—NH—G6H4, , RC1 : A1CL + C6H6, , c6H, , S03H -f H+, T Cl~, , R + H+ + AICI7, , As with nucleophilic substitutions, coclassification of these reactions on the, basts of similarity of overall changes does not imply that the attacking species, in the various reactions are necessarily analogous.57, The reactions that we have been discussing have one feature in common,, ectrons are transferred in pairs during the processes of bond formation and, rea age. When two species come together to form a bond, both bonding elecrons come from just one of the species; likewise, when a bond is broken, , both, , on tng electrons depart with one of the fragments. Because of the electric dis, symmetry of acid-base reactions, electrophilic and nucleophilic substitutions, classed m5e?a:t„rr'er°ePh;HC, , f ^ reaCti™S, , - - « «£ rc;ly’, , mediates tn such reactions have even numbers of electrons., It has already been shown that C=0 P—tsj, , ,, , susceptible to attack by nucleophiles, whereas C, vu nerable to electrophiles. In line w ,h such difference, between oUcrophUU and nncUopkilic addition L for ” ,, a ?r,r-,0nSJd° "0t seem “ useful as the corrSpoidin, and their adoption is, at present, far less widespread, , ', , ., ... bonds ,end to b', a, ° b°"ds ,end “ b' more, “T!"is “““ionally drawn., , H P ’ Rrf' 49' PP', These, deSCnpt.ons of substitution reactions,
Page 138 :
122, , Acids and Bases. Nucleophiles and Electrophiles, The reader is doubtless aware, however, that a number of important reac¬, , tions take place through intermediates with odd numbers of electrons. In these, reactions, covalent bonds may be broken homolytically—that is, with one bond¬, ing electron departing with each of the two fragments. Since species having odd, numbers of electrons are known as free radicals, such reactions are classified, together as free-radical reactions or homolytic reactions. Among these are, the halogenation and nitration of aliphatic hydrocarbons, the addition of HBr, to olefins in the presence of peroxides, the arylations of aromatic hydrocarbons, using diazonium compounds, and a large number (but not all) of the polymeri¬, zations of olefins. Dissociations, substitutions, and additions are observed among, free-radical reactions just as with polar reactions., , (22), (C6H5)3C—C(C6H5)3, , (23), , * 2(C6H5)3C*, , Cl. + CH3CH2COOH-4 HC1 + .ch2ch2cooh, Br- + CH2=CHBr, , » BrCH2—CHBr, , J dissociations, , (24), , substitution, , (25), , addition, , However, as we shall see in the chapter devoted to free-radical reactions, the, reactions in (22) through (26) are very different in character from their respec¬, tive counterparts among polar reactions—so much so, in fact, that correspond¬, ences between the two broad classes of reactions are little more than formal., , EXERCISES, , FOR CHAPTER 4, , 1. (a) Arrange the following in the order of decreasing acidity., , acetic acid, , nitromethane, , CH3C(OH)2, , H3SO|, , phenol, , h2so4, , cyclopentanol, , chloroacetic acid, , aniline, , benzene, (b) Arrange the following in the order of decreasing basicity:, (CH3)3C—oOH", H20, 2, , Br~, (C6H5)2CH-, , C2Hj, HOOC(CH2)2COO, , CelW>, ., Cl2CHC(OH)5, , NHJ, , Given a pH meter on which readings reliable to 0.02 pH unit may be made, what, is the minimum, 0.01, , K. value that a weak acid may have in order that the actdt y, , M solution in water may be detected by this meter?, , 3. Outline an experiment that would show which acid in each of the following pairs is, the stronger:
Page 139 :
123, , Exercises for Chapter 4, , (b) C6H5SO2H (contaminated with a trace, taminated with a trace of C6H5Se03H)?, (c) C12CH—C(OH)2 or Br2CH—C(OH)2?, (d) Benzene or cyclohexane?, (e), , or, , 4. (a) Predict approximate r factors for each of the following in concentrated sulfuric, acid:, water, , ethylenediamine, , sodium acetate, , chloroacetic acid, , sodium carbonate, , COOMe, , Me, , COOMe, , Mevv^\/Me, , ."aleS, , IfuTS “VT, , aT 1 ^ (C‘H‘°CH>> - -conHlSO‘Vnd ***, , freezing point by about, , . °, where!, , 0 8, , Ldtii, , grams of pure sulfuric acid lowers its freezing, acid to which a small amount of, , Pf“nVadd l0WerS itS, °k °'u, , “ °f Water to 1000, , JZ7 T' ^ ^ ^, , cryoscopic studies than pure sulfuric acid?, , -y the, , 3, , Why is sulfuric, , beUer solvent for, , (a) Why does the ratio *,/*, for ,o„g.chain dibasic acids approach, , ., , ?, , 4 00
Page 140 :
124, , Acids and Bases. Nucleophiles and Electrophiles, , and ChCHCOOH have about, , the same strength in water. Which is the stronger acid in ethanol? Explain., , 6. The following data, collected by Deno,** refer to the ratio of the concentrations of, the two forms (IrT and HIn) of three indicators in water-hydrazine mixtures., Percent, Hydrazine, , log {cjn-/Cj[jn), , 20.2, , —1.18, , II, , 20.2, , — 0.82, , II, , 33.2, , + 0.42, , III, , 33.2, , -1.18, , III, , 50.1, , +0.55, , Indicator, I, , Calculate the, , (pKa = 13.4), , pKa values for indicators II and III and the //_ value for 50.1 percent, , hydrazine., , 7. Given the followingpKa values in water, 35 percent ethanol, and absolute ethanol:, , pKa values, , (a) From, , show, , that, , the ratio, , Iha/ya- in water-alcohol solutions is not, , Percent, EtOH, , these data,, , HOBz, , HO Ac, , independent of the acid Hd., (b) T_ for absolute ethanol is set at 1.000 on, Grunwald's scale., , 0, , 4.76, , 4.20, , 35, , 5.43, , 5.24, , 100, , 10.32, , 10.25, , Calculate, , I _ for 35, , percent ethanol., (c) Calculate the difference in the, , values, , of HO Ac and HOBz., , 8. (a) From the data on page 113, estimate the value of ^at. associated with phenylacetic acid (Ka = 4.9 X 10~5) for the dehydration of acetaldehyde hydrat ., (The observed value is 33.0 liters per mole-min.), (b) For the same reaction, estimate Cat. associated with the hydrogen citrate ion,, OH, -o, , _C—CH2—C—CHo—C—O-,, , Ka = 4 X 10-6. For the statistical correc-, , O, COOH, O, tion, assume that the proton affinities of the three carboxyla.e groups of, citrate ion are essentially the same., , 9., , Two reactions are subject to general acid catalysis. For reaction (1), the Boosted, parameter, a. i, 0.1; for reaction (2). a is , .0. Both react.ons are canedoumbuder, made from an equimolal solution of a weak ac.d HA, , (K. - 10, , )
Page 141 :
Exercises for Chapter 4, , -, , 125, , Na'Ll-. Calculate, in each of the cases below, the fraction of the reaction due to, , HA, and H20. (Assume that the ionic, (a) Reaction (1); for the buffer (H.4) = (A~), (b) Reaction (2); for the buffer (Hd) = (A~), (c) Reaction (2); for the buffer (H.4) = (A~), H30+,, , strengths in all cases are the same.), = 0.01 molar, = 0.01 molar, , = 1.0 molar, , Show that if reaction (2) were studied only in dilute buffer solutions, it might appear, to be subject to specific hydronium-ion catalysis. (Note: Ka for H30+ is 55.5.), 10. Decide whether each of the following species may be Lewis acids, Lewis bases, both,, or neither:, h2o, , FeCL, , HgCl2, , CC14, , B(OC2H5)3, NO^, , naphthalene, , so2, , NO+, , (C6H5)3C+, , Cr(C104)3, , c6h5—c—ch2—c6h5, , SbFs, , —, , S03, , -, , (CH3)3C—O-, , O, , 11. Classify each of the following substitution reactions using the terms introduced in the, final section of this chapter; (for example, the first reaction is a nucleophilic sub¬, stitution by hydride ion for hydroxide on hydrogen)., Example: H~ + H20 —> H2 + OH", , rv*, , « \?-SOsH + OH, , (b) Cl- +, , OH + SO,H, , ■>, , CH., a, , (c) C6H6 +, , CH3C-CI, , A1C1;, , HC1, , +, , CHc, , rv., , \==/"(jfCH3 + h+aici4o, , (d) D2S04 +, , /, , o, , V-O- CH,, , och3 + dhso,, o, (e), , NaT, , +, , CH3 —O—S—^, , O, , Na, , o±Q, o, , + CH3I
Page 143 :
CHAPTER 5, , Methods for Determining, Reaction Mechanisms, I. Nonkinetic Studies, , The Meaning of “Reaction Mechanism”, In the ideal case, we may consider the mechanism of a chemical reaction as a, hypothetical motion picture of the behavior of the participating atoms/Such a, picture would presumably begin at some time before the reacting species ap¬, proach each other, then go on to record the continuous paths of the atoms (and, t eir electrons) during the reaction, and come to an end after the products, DktureThe, SmCeU iSrn°‘ genera“y P°SSib'e *° °btain SUch an intimate, don ,hV, T, f“°n, a mechanism has come to "lean obtaining informathat can furnish a picture of the participating species at one or more crucial, instants during the course of the reaction, , is, rtiiabiiuy a a,,, predictions as to how the, speed of ih,, •, me speea ot the reaction is, , '> "«■»")., f, d§ t0 <luantitative, u, anected by concentrations of, 127
Page 144 :
128, , Methods for Determining Reaction Mechanisms, , reactants, temperature, solvent, and the presence of catalysts. It is also desirable, that a proposed mechanism allow prediction of the manner in which the rate, of the reaction will change as the structures of one or more of the reactants are, subjected to a given change. Clearly, all of this constitutes a large order. In, actuality, experimental investigation of one or more of the above points leads, to a proposed mechanism that may then be confirmed (or excluded) by investi¬, gation of the remaining items., , Energy Profile Diagrams. Intermediates vs. Transition States, An intermediate in a chemical reaction may be defined as a species which is formed, (preferably in detectable amounts) from the reactants and which under the reac¬, tion conditions, is eventually converted to the reaction product or products., As an example, consider the conversions of N-bromoamides to isocyanates in, basic solution (a part of the well-known Hofmann rearrangement)., CH3CONHBr + OH- -> CH3N=C=0 + Br- + H20, In this case, we feel quite certain that the anion CH3CONBr- is an intermediate, in the reaction, for, with care, salts containing this anion may be isolated1, and such salts undergo the conversion to isocyanate at a rate no less than that, of the brominated amide., A large number of the known types of organic reactions proceed through, one or more intermediates, but there are some that do not. The basic hydrolysis, of methyl iodide is thought to be one of these., CH3I + OH- —> CH3OH + I", Although this type of reaction will be treated at some length in Chapter 7,, let us anticipate the discussion by picturing the course of such a reaction., Temporarily closing our eyes to the fact that both the hydroxide ion and the, iodide ion are “solvated” (hydrated, if the solvent is water), we may visualize, an energetic hydroxide ion approaching the methyl iodide molecule from a, direction opposite to that of the C-I bond. As the OH" ion gets nearer to the, carbon atom, the molecule “spreads out,” the C, , I bon, , wea, , “ h, , C-O bond begins to form. At some point in the reaction we may comider th, C-I bond as “half broken” and the C-O bond as “half formed., , At thi, , point (Fig. 5-1), the three nonparticipating hydrogen atom, w, , or near, , he, , rr ssssszxxs - - --£r, , C-O bond, tain, -bkh ,1™ to 11™ M.«V» —, / Mauguin,, , Ann. Chim. (8); 22, 297 (1911).
Page 145 :
Energy Profile Diagrams. Intermediates vs. Transition States, , 129, , hedral positions about the carbon. The situation represented in the center, section of Figure 5-1 may be regarded as the “halfway point” in the reaction—, that is, the point beyond which the system is more likely to progress to the, products than to retreat to the reactants. Such a configuration is generally, called the transition state or activated complex. Considerable energy is neces¬, sary to drive the reactants into this configuration; that is why the reaction pro¬, ceeds at only a moderate rate, for only the more energetic OH~ ions can push, the CH3I into the necessary transition state configuration. Furthermore, if the, reaction were to be carried out in reverse, it would require still more energy, to drive the system consisting of I" and CH3OH into the same transition state., , reactants, , products, , transition state, Fig. 5-1. The Basic Hydrolysis of Methyl Iodide, , Either the forward or the reverse reaction may then be imagined as going, up, then down, an “energy hill” with the transition state at the top. The, transition state is thus, in effect, defined as that point in the reaction where, the total energy of the system of atoms under consideration is maximum., Such a picture is sometimes extended by representing the course of reac¬, tions in energy profile diagrams., , CH3I, , Such representations, of which Figure 5-2, , + OH, , CHoOH, , energy, , + 1“, , reaction coordinate, Fig. 5-2. Energy Profile (Schematic) for CH3I -f- OH, , CH:iOH + I, , tictl diTP‘C’ WU1 be diSCUSSed here °nly very briefly- In such diagrams ver, , 4 SrelTaTeTr8' 4 ^ ^ ^ the, On this case an OH- il and aCH 1 m, Horizontal distances have no, , reaction. AH. The enerev diffe, , ’, , ^ * Particular molecular unit, ° CCU e) alonS the c°urse of the reaction., , ,, , N“ 'i» *•, , d fi, S rePresents the heat of, energy deference between the initial level and the top of
Page 146 :
130, , Methods for Determining Reaction Mechanisms, , the “hill'’ (the transition state) corresponds quite closely, but not exactly, to, Ht—the activation energy of the forward reaction. (The reason for this small, discrepancy need not concern us now.) This activation energy is the energy that, must be added to an “average” OH- ion and an “average” CH3I molecule to, drive them into the transition state. Often, activation energies are expressed in, kcal, under which circumstances they would refer to moles of “average” par¬, ticles, rather than individual units., For a reaction proceeding through true intermediate, there is a “dip” or, minimum in the energy profile diagram., , The deeper the dip, the more stable, , will be the intermediate and the more sure we may be of its existence. For the, anion intermediate in the Hofmann conversion (mentioned at the beginning, of the chapter), the dip is deep, perhaps reaching below the energy level of the, reactants. In contrast, the presence of the cation I in the nitration of benzene is, , (I), inferred only by indirect means (Chap. 11) and the energy profile, if it dips at, all, must do so very slightly. The two types of intermediate are compared sche¬, matically in Figure 5-3. In the extreme case the dip becomes so slight that the, intermediate a is experimentally indistinguishable from a transition state. If, the intermediate is isolable, we are in effect considering two reactions, each, with its own transition state. In such cases, a substance becomes designated as, an intermediate simply because the “reaction” originally chosen is a composite, one., a, metastable intermediate, b, stable intermediate, , products, reaction coordinate, Fig. 5-3. Energy Profiles for Reactions Proceeding through Intermediates, , In any event if the configuration and energy of each of the intermediates, and transition states through which a reacting system passes are known, it is, not too much to say that the mechanism of the reaction is understood., Since studies on the rates of reactions have become so important in giving, us information about reaction mechanisms, it is not surprising that in many
Page 147 :
Identification of Products, , 131, , minds “mechanistic studies” have become synonymous with “kinetic studies.”, Although an entire chapter is devoted to a discussion of kinetic evidence, let, us first examine a number of nonkinetic methods for obtaining mechanistic, information., , Identification of Products, The most fundamental basis for mechanistic speculation is the identification of, the reaction products (generally carried out whether or not a mechanistic ques¬, tion is at stake), for without such identification we cannot be sure which reaction, is actually under consideration. The student of elementary organic chemistry, learns that the chlorination of toluene yields benzyl chloride when carried out, in the vapor state with illumination, but that if the chlorination is carried out, in the liquid phase in the presence of aluminum chloride, o-chlorotoluene and, />-chlorotoluene result. On this basis, he becomes aware that a different chlorina¬, tion mechanism operates in each case., There might be some temptation to liken the replacement of chloride by, cyanide in benzyl chloride (II) to the replacement of chloride by cyanide in, furfuryl chloride (III) until the product of the latter reaction is identified as, , + CN', , O, , CHoCl, , ->, , NC-, , -CH,, , + cr, , III, IV, , 7, , ci—tt;;:'1 V)> shors that the position, * ** -oming, Af°Z " n0t the, S e as that occupied by the outgoing chloro group.*, fonic acid., , ^2™?;nt, , t may be, , mechanistically dissimilar when the product'f ^, the same carbon skeleton as the start', has undergone rearrangement., , ss, , rCplacements must be, ,cactlon v ls shown to have, , § matenal whereas <he product from VI, , 63’ 749 (193°)m, and Knell, J. Am. Chem. Soc., 73, 5004 (1951).
Page 148 :
132, , Methods for Determining Reaction Mechanisms, , (CH3)3C-CH2-0, , -S°2-CV, , CH3 + I~ —> (CH3)3CCH2I, , +TosO“, , neopentyl tosylate, V, , (c6h5)3 c—ch2—o—so2, , // \y-ch3 + och, , i, , tritylmethyl tosylate, VI, H, (C6H5)2CH-C-OCH3 + TosO", , Note that in the three examples cited, we have suggested no mechanistic, details. Nevertheless, the correct identifications of the products strongly suggest, dijferences in mechanism., Occasionally a reaction is used in the laboratory lor many years, fuinishing, satisfactory yields of a desired product and, at the same time, additional prod¬, ucts that may be discarded without identification. The present author has, converted aryl selenocyanates (VII) to diaryl diselenides many times using, methanolic KOH, yet knows essentially nothing of the path through which, the conversion occurs, mainly because the water-soluble pioducts, 2Ar—Se—CN + OH- -», , Ar—Se—Se—Ar, , + ???, , VII, are extracted and discarded. The identification of these products would seem, to be the first step in clarifying the mechanism of this reaction., Sometimes the absence of a particular product can be just as important, as the presence of another. The familiar Sommelet reaction converts halides, such as benzyl chloride to aldehydes by treatment with hexamethylene tetramine, (CH2)6N4. Since the latter compound is, in effect, a sort of “mixec, anhydride” of formaldehyde and ammonia, Sommelet' proposed that the reac¬, tion terminated in the three steps shown below., , ^, , y CH2—N=CH 2, , ♦ Sommelet,, , Compt. rend.,, , 157, 852 (1913).
Page 149 :
Identification of Products, , 133, , Today, however, although agreement is not complete as to what the mechanism, of this reaction is, we are quite certain that the mechanism proposed by the, discoverer is not the correct one since careful examination of the reaction mixture, failed to reveal appreciable quantities of methylamine.5, Often the most vexing question concerning the mechanism of a reaction is,, “Which of the bonds is broken during the reaction, and where is the new bond, (or new bonds) formed?” Consider, for example, the reaction between chloro¬, benzene and KNH2 to form aniline:, , Without knowing better, we might easily put this reaction in the same class as, the replacement of I, , by OH, , in methyl iodide (p. 128). It is certainly reason¬, , able that the NHy ion could approach the chlorinated carbon atom, that the, C, , Cl bond could weaken, and that the new N—C bond could form while the, , old C—Cl bond breaks (however, the geometry of the collision would have to, be slightly different from that in the case of methyl iodide because the geometry, of the two halides is not the same). One would have to admit, however, that if, the incoming —NH2 group were to occupy a position different from that occupied, by the leaving —Cl group (and if protons could move from one carbon to, another), such a path would not be detected merely by identification of the, reaction products. The fact that the action of NaNH2 on o-iodoanisole yields, rc-amsKhne* suggests, but of course does not prove, that in the “ammonolysis”, of chlorobenzene the carbon from which the chlorine departs is not the carbon, , +, , Nal, , gr0Up in the resulti"g molecule. As we shall see (page 144), tins doubt can be resolved by application of tracer methods, , by heaVm, , rearrangem“'’, , -o1, , Gil™1, , h2ch=ch., , OH, , and, F(assack’ J■ Chem- Soc.,, , Iman and Avakian,, , «■*» - converted, , 1949, 2700, , J. Am. Chem. Soc., 67, 349 (1945).
Page 150 :
134, , Methods for Determining Reaction Mechanisms, , Here again, one might ask, “Is the carbon atom that breaks away from the, ether linkage the same carbon atom that becomes attached to the benzene, ring?” Although in the example shown above, identification of products does, not furnish an answer, the same reaction, when carried out with the substituted, allyl ether VIII, yields the phenol IX, showing that the 7-carbon, rather, than the a-carbon, becomes attached to the ring during the rearrangement.7, , More will be said about this rearrangement in Chapter 15, but it is emphasized, that in this case, as in the five examples previously described, mechanistic, information is obtained by the simple process of identifying the reaction products., , Testing Possible Intermediates, A compound suspected of being an intermediate in a reaction may sometimes, be isolated, identified, and resubjected to the conditions that prevail during the, course of the reaction. If the appropriate products are formed at a rate no less, than that of the “uninterruptedr’ reaction, this is strong (although not unequivocal), evidence that the reaction proceeds through the intermediate that has been, isolated. For the Hoesch reaction of chloroacetonitrile with resorcinol and HC1, to form the hydrochloride X, the imino chloride XI is a likely intermediate, , OH, , since it has been shown to form from the nitrile and HCF and, upon treatment, with resorcinol under Hoesch conditions, readily yields, 7 Claisen and Tietze, Ber 58, 275 (1925)., « Troeger and Luning. J. prakl. Chem. (2) 6. ., » Stephen, J. Chem. Soc., 117, 1529 (1920)., , > roc
Page 151 :
135, , Testing Possible Intermediates, We have already mentioned anions of the type R, , C, , N, , Br, , as probable, , O, intermediates in the Hofmann rearrangement (p. 128). This reaction is often, regarded simply as a conversion of an amide to an amine having one less, carbon atom; that is, the reaction is written, RCONH2 + OBr- -> RNHo + C02 + Br~, Representing the reaction in this way is very much an oversimplification, for, it glosses over the fact that it is possible to isolate at least three intermediates,, an N-bromoamide, its anion, and an isocyanate., /RNH2, RCONH,, , RCONHBr, , RCNBr-, , R—N=C=0, , Jl, , |, , lco2, , In the Reissert reaction, compounds such as XII (N-benzoyl-dihydroquinaldonitrile) are converted by treatment with acid to quinaldamide (XIII), and aromatic aldehydes. If the original compound is treated with HC1 under, anhydrous conditions, a reddish-orange product forms; because this compound, , h, , cr, , XIV, , XV, , COPh, XVI, , vields the same amide and aldehyde when treated with, , ,, , • j, , consideration as an intermediate in the reaction even ,hT’ h, eaction even though its structure is
Page 152 :
136, , Methods for Determining Reaction Mechanisms, , not known with certainty (the structure XIV has been suggested).*0 On the, other hand, the substances XV and XVI cannot be intermediates in this reaction., The nitrile XV does not give quinaldamide, nor does the amide XVI yield, benzaldehyde when subjected to conditions under which the Reissert reaction, proceeds smoothly., Similarly, the reluctance of other presumed intermediates to give the, correct products under suitable reaction conditions has caused many a chemist, to revise his ideas. The Clemmensen reduction of 0=0 groups to CH2 groups, using zinc amalgam and HC1 cannot proceed through the corresponding carH, binol [, , that is, —C-, , O, , -C->—CH2-, , since the carbinols are them-, , OH, , selves not generally reduced with the same combination of reagents.** A somewhat, different argument may be used against a urethane intermediate in the forma¬, tion of allophanates (XVII) from alcohols and cyanic acid in ether. Although, it is true that urethanes, if treated with cyanic acid, form allophanates, the, reaction is much slower than the formation of the allophanates from the, alcohols.** Since the rate of a composite reaction must be less than or equal to, RO—C—NH—CNH2 (allophanate), , ROH + 2HN=C=0, , slow, , RO—C—NH2, , o, , o, XVII, , o, the rate of any step, the prevailing mode of formation of allophanate cannot be, through the urethane (although a minor amount could be formed in this way)., A reaction intermediate that can be formed in more than one way shou, exhibit a behavior independent of its mode of formation. Possible intermediates, may be eliminated from consideration on this basis even though they cannot, be isolated. Until recently, a number of workers felt that the Friede -Crafts type, methylation of aromatic hydrocarbons proceeded through a carbonmm, CH+, , ,, , However, the relative amounts of the three xylenes produced in, , methylation of toluene vary, depending upon whether methyl bro^de o, methyl iodide is used.'’ In a typical case, the para:meta ratio is 1.7 for methy, bromide but climbs to 3.3 when methyl iodide is used under the same, * McEwen and Cobb, J Am., £££ Inc., New York, 1942, p. 156., " Martin, in Organic Reaction, Vol. 1, Jo, » Close, and Spielman, J. Am Chm_ Soc 75,_4055 (1955)., » Brown and Jungk, J. Am. Chrn. Soc., 77, 5586 (1955).
Page 153 :
"Trapping” of Intermediates, , CH3X, , (V=Br or I), , \, , 137, , ALClf, , +, , />—CH3 —2^ o, m, or />-xylene + HC1, , \ A12C16, \, \, , A, , ~"his means that the CH+ ion is not an intermediate in both methylations and, probably not an intermediate in either case., , “Trapping” of Intermediates, Sometimes an intermediate may be detected, although it cannot be isolated,, by adding to the reaction a “trapping” reagent. The latter is added so that it, will combine with the intermediate to form a product that (the worker sup¬, poses) cannot be accounted for otherwise. The addition of bromine to many, olefins in polar solvents is thought to proceed through an intermediate, which, , /\ _ /V, we shall represent as ^/^\, , \) • (The reasons for assignment of a, , cyclic structure to this intermediate are discussed in Chap. 13.) In simple, additions, this cyclic “bromonium ion” is thought to react with Br~ present in, solution to form the observed dibromide. Strong evidence for such an inter¬, mediate is that it can be diverted from its ordinary reaction course by the pres¬, ence of basic reagents other than Br~. For example, when bromine is added to, stilbene (XVIII) in methanol, the bromoether XX may be isolated from the, reaction mixture. 4 This is presumably formed when the intermediate XIX, , Ph, Ph-CH=CH-Ph + Br2, , /, / \ / \, H, , XVIII, , \, , Br, , XIX, , Ph, , H, , PhCHBr— CHBrPh, , ^0^ PhCHBr — CHPh(OCH3), XX
Page 154 :
138, , ••, , Methods for Determining Reaction Mechanisms, , which itself reacts with the olefin to give the bromoether. We see, then, that, the trapping of intermediates by chemical means must be interpreted with, caution, for the reaction studied has been altered by addition of an extra, reagent. What the worker desires is that the alteration occurs after the formation, of the intermediate he seeks, rather than before., For a number of years it was an open question as to whether the reaction, of phenylmagnesium bromide with diphenylketene (XXI) involved addition, to the C=C bond or to the C=0 bond in the ketene. Hydrolysis of the reaction, mixture yields a mixture of the ketone XXII and its tautomer, the enol XXIII;, the products therefore do not indicate the direction of addition. However, if, the reaction mixture is treated with benzoyl chloride prior to addition of water,, the benzoate of the enol (XXIV) results.**7 This indicates that the C=0 bond,, rather than the C=C bond, in the ketene, has been subjected to attack, provided, O, , II, , Ph2CH—C-Ph, It XXII, Ph2C=C-Ph, OH, XXIII, , PhMgBr 4- Ph2C=C=0, XXI, , Ph2C=C—Ph, COn, , O, , XXIV, , c-c6h6, //, o, it is assumed (a) that there is no rearrangement of the addition product during, reaction with benzoyl chloride and (b) that the product formed in the reaction, with benzoyl chloride was isolated without suffering rearrangement., , Evidence from Reaction Catalysis, Often the manner in which a reaction may be accelerated or inhibited furnishes, a hint as to its mechanism. A large number of reactions are catalyzed by acid., In such cases it is reasonable to suppose either that the reaction intermediate, is a cation formed by removal of a basic fragment (OH , OAc , etc.) from a, reactant or else that the reaction intermediate is simply the conjugate add, (p, , 94) of one of the reactants. It is also possible that jhe reaction path may, , involve an intermediate of each type. Conversely, a reaction that is base cata¬, lyzed may be thought to proceed through an anion formed by removal of a, proton from one of the reactants (that is, through the conjugate base: of h, reactant), or else through an adduct of the added base with one of the, (page 190)., u, , Bed and Vejvoda,, , J. Am. Chem. Soc.,, , 76, 905 (1954).
Page 155 :
Evidence from Reaction Catalysis, , -, , 139, , The splitting of the hemiacetal linkage in the sugar derivative, a-D-tetramethylglucose (XXV) results in formation of the hydroxy aldehyde (XXVI),, which is in rapid equilibrium with both the starting material and the second, hemiacetal, /3-D-tetramethylglucose (XXVII). The conversion of either the, pure a or the pure /3 form in solution to the equilibrium mixture is accompanied, OMe, , OMe, , CHoOMe, , CHoOMe, , XXV, , OMe, CHoOMe, , by a drift in the value of the optical rotation of the solution until net reaction, ceases. This is the familiar phenomenon of mutarotation and offers a convenient, way for following the progress of the reaction. This reaction requires the presence, of both acid and base, for it is very slow in pyridine (which is basic but nonacidic), and very slow in cresol (which is acidic but only faintly basic). Mutarotation, proceeds far more rapidly in a mixture of pyridine and cresol/5 and rapidly, also in water (which is both acidic and basic). The reaction may be greatly, accelerated by using as a catalyst 2-hydroxypyridine, which (although a very, weak acid and a very weak base) has its acidic and basic sites held together, rigidly in the correct position for attack on the hemiacetal linkage (XXIX).", OMe, , OMe, , XXVIII, n o°W.ry an^ Faulkner, J. Chem. Soc., 127, 2883 (1925'), Swain and Brown I Am C'h*™ o, * /m--., ', , ’, , J- Am■ Lhem■ Soc-,, , 74, 2534, 2538 (1952).
Page 156 :
140, , -, , Methods for Determining Reaction Mechanisms, , The manner in which the ease of hydrolysis of salicyl phosphate (XXX), varies with acidity gives some hint about the hydrolysis intermediate. The rate, of hydrolysis is greatest at pH 5.3, diminishes as the pH is either raised or lowered, beyond this value, and becomes very small in either very acid or very basic, solutions.15 This trend suggests that the anion XXXI, possibly in its cyclic, form XXXII, is an intermediate in the hydrolysis and that the uncharged, , O, , o, , ^X>P(OH)2, , OH, COO', , 'COOH, , XXXII, , XXXI, , XXX, , molecule XXX (the predominant form in very acid solution) and the conjugate, base of XXXI (the predominant form in very basic'solution) are relatively, inert., One of the most convincing ways of showing that a reaction proceeds, through a path involving free radicals is the demonstration that the reaction is, accelerated by substances such as peroxides or azo compounds which readih, produce free radicals. If a reaction is photochemically induced, it is almost, certainly a free-radical reaction, and if it is inhibited by compounds such as, hydrogen iodide or hydroquinone, which are known to lower the concentration, of active free radicals, a similar conclusion may be drawn. A very familiar, example of a free-radical reaction is the “abnormal” addition of HBr to allyl, bromide. If the reagents are carefully purified, then mixed in the dark with care, taken to exclude oxidizing agents, the HBr molecule very slowly adds to the, double bond to form the 1,2-dibromide,15 that is, the addition is like that of, 3, , dark, no peroxides^ CH _CHBr_CH,Bl', , (slow), , /, CH2=CH—CH2Br + HBr, neroxides present, , HCl The addition can be accelerated by addition of benzoyl peroxide and, more, particularly, by illumination, but under these conditions the product is tie, 1,3-dibromide. The conditions favoring formation of the '>3- 1Jroml e, “, indicate that it proceeds through a free-radical mechanism, and the inh.b t.o, of the “abnormal” addition by hydroquinone or diphenylam.ne con . ms, .. Chanley, Gindler, and Sobolka J., » Kharasch and Mayo, J. Am. Chem., , Am. Chm., Soc., 55, 2568, , 7*, 5347 <1952)(1933).
Page 157 :
Crossover Experiments, , -, , 141, , Crossover Experiments, A number of molecular rearrangements may be regarded as processes in which, a fragment is broken off from its position in the reactant molecule and becomes, refastened to a different position in (a) the same or (b) a different molecule., In considering the mechanism of such reactions, it is important to know which, of these two possibilities is the correct one—that is, whether the rearrangement, is intramolecular or intermodular. This question may sometimes be answered by, carrying out the reaction with a mixture of two similar but nonidentical reac¬, tants and searching the product for compounds having fragments of both, reactants, thus seeing whether fragments from one reactant have “crossed over”, and have become attached to fragments of the other reactant., One of the most familiar examples of the use of this mode of investigation, is connected with the benzidine rearrangement, a reaction in which hydrazobenzenes are converted by acid to benzidines. If the reaction is carried out on a, , H', , H9N, H, , H, , {, , NH;, , mixture of 2,2/-dimethoxyhydrazobenzene (XXXIII) and its diethoxy analog, (XXXIV), ^only two benzidines can be isolated, both of them symmetrically, substituted.20 The nonformation of the unsymmetric benzidine XXXV (which, would result if fragments from two different hydrazobenzenes were to combine), is strong evidence that the reaction is intramolecular. For this reaction, it must, , r, , h9n, , H, ->, , to, , IngoM and Kidd,, , J. Chem. Soc.,, , 1933,
Page 158 :
142, , Methods for Determining Reaction Mechanisms, , be inferred either that the “new” bond (the C—C bond between benzene rings), is partially or wholly formed before the “old” bond (the N—N bond) is com¬, pletely broken, or, if this is not the case, that the two fragments from a given, molecule do not become free of each other’s influence long enough to allow, the fragment from another molecule to intercede. If, however, the reaction, turned out to be intermolecular, two types of mechanism could come to mind., The first would involve a transition state in which two molecules collide and, the, , rearranging, , fragments, , “exchange, , partners.”, , The, , second, , (and, , more, , likely) possibility would be that the dissociating fragments become essentially, independent of each other for at least a short time during the course of the, rearrangement., A somewhat different type of crossover experiment is possible if the species, thought to be one of the rearranging fragments is isolable. For example, the, rearrangements of 9-decalyl peroxybenzoate (XXXVI) and its peroxy-p-nitrobenzoate analog to the ester XXXVII are thought to proceed by migration of a, benzoate ion from an oxygen atom to a carbon atom, during which time the, remaining fragment of the molecule, the cation XXXVIII, presumably suffers, rearrangement (as shown below in the parenthesis). If the reaction does follow, such a path, the benzoate ion and the cation never become completely inde-, , “OCAr, , +, , II, O, XXXVIII, , pendent, for if the salt lithium /,-nitrobenzoate is added to the reaction mixture,, no 6-nitrobenzoate ester is found in the product.*' Likewise, if the experiment, is “reversed” and the peroxy-/>-nitrobenzoate ester is rearranged in the presence, of lithium benzoate (that is, if Ar = p-NO^Hj and Ar, , = CeHj), no un;, , substituted benzoate ester is found in the product. These experiments, if taken, alone, , do not rule out the possibility that the rearrangement occurs throng, , collision of two reactant molecules with a change of partners, a point that might, */ Goering and Olsen, J. Am. Chem. Soc., 75, 5853 (1953).
Page 159 :
Isotopic Labeling, , -, , 143, , best be settled by kinetic studies. It might also be argued that the reaction is, intermolecular but the fragments are radicals rather than anions; in this case,, added benzoate anions might not intercede in the reaction but added radicals, conceivably could. This point might be settled by showing that the rearrange¬, ment is not subject to free-radical catalysis., , Isotopic Labeling, We have already seen (p. 131) how, in favorable cases, mere identification of, the products of the reaction can indicate to a large degree which bonds are, broken and where new bonds are formed. At the same time it was pointed out, that more information was needed to answer such a question about the Claisen, rearrangement of allyl cresyl ethers or the ammonolysis of chlorobenzene. Both, of these reactions have, however, been studied using reactants especially pre¬, pared so that certain strategically located carbon atoms are the radioactive, isotope, C14. In such cases, only a small percentage of the molecules subjected, to the reaction are so labeled, but it is these that trace the reaction., The Claisen rearrangement has been carried out using allyl, , p-cresyl, , ether,, , labeled with C44 at the 7 position (XXXIX) " When the resulting allyl cresol, was subjected to oxidative degradation, the end carbon on the chain was, eliminated as formaldehyde. Since none of the radioactive carbon was lost, it, follows that the 7-carbon in the original ether does not become the end carbon, in the rearrangement product. Since it is very difficult to conceive of a reaction, , OCH2CH=C*IT, , OH, , rL i, , heat, ->■, , C*H2CH—CH2, , OsO«^, , hio4 ., , CH., OH, C*H2CHO, +, , CH3, , HCHO, , (*) designates labeled atom, , path in which the 0-carbon becomes attached to the ring durintr the, men., we may then conclude that the rearrangement, , to that of ether VIII (D, the benzene ring., "Schmid, , and, , 1 t.a\, , P', , Schmid., , ^, , . ., , ,, , ,, , ,S’ **“* the, , H,h. Chim. Ada, 35, 1879 (1952)., , einer AAXIX is similar, becomes attached to
Page 160 :
144, , Methods for Determining Reaction Mechanisms, The ammonolysis of chlorobenzene has been studied, using material in, , which the chlorine-bearing carbon is isotopically labeled. In this way, it is, possible to show that very nearly half of the incoming —NH2 groups become, bonded to the carbon from which the chlorine departed. The remaining —NH2, groups become bonded to carbons ortho to the labeled carbon, the other posi¬, tions in the ring not being attacked.23 This evidence points to a symmetric, intermediate, a “benzyne”, , (XLIII), which then suffers attack by NH3 or, , NHj at either of the triple-bonded carbons. The formation of a similar inter¬, mediate may be used to explain the rearrangement during the ammonolysis of, o-iodoanisole (p. 133)., « Roberts, Simmons, Carlsmith, and Vaughan, J. Am. Chem. Soc., 75, 3290 (1953). The, position (s) of the labeled carbon in the resulting aniline was determined by the following, reaction sequence:, OH, , NH,, b, , OH, , O, CrOg, , hno2, , ->, , b, b, a, H2N-CH2(CH2)3CH2COOH, , MnOr, , HNi, , h-», , H2N-CH2(CH2)3CH2COOH, XL, , b, b, HOOC(CH2)3COOH, XLI, , HOOC(CH.),COOH + 2HN, —* 2CO, + H,N(CH,),NH!, , If, in the original —U .he nitrogen*carbon, all of the labeled carbon s ion, in, eliminated as CO, by two successive treatwere the case, all of the labeled carbon wouW be, ^ L/of the radioactivity, , rdSr^:^r=ofdia.,ne(X^—, , the remaining half at positions b.
Page 161 :
Isotopic Labeling, , -, , 145, , When studies are made using C14, the route of the labeled atoms is naturally, traced by using a radiation-sensitive device to follow the, , radiation through a, , series of operations. Sometimes it is desired to carry out experiments with labeled, nitrogen or oxygen, but in these cases the procedure must be substantially, different. The greatest half life among the known radioisotopes of oxygen is, 126 seconds, and that among the radioisotopes of nitrogen is only 10 minutes., It is thus very difficult to obtain and handle appreciable quantities of these iso¬, topes and essentially impossible to perform a series of syntheses using them., Tracer work with oxygen or nitrogen must be carried out using samples of non¬, radioactive elements in which the atom percent of the rarer isotopes is sub¬, stantially greater than that found in nature. Samples of water “enriched”, with O18 and ammonium chloride “enriched” in N15 are available, and these, compounds may be used to synthesize oxygen-labeled and nitrogen-labeled, organic compounds. In such cases, the path of the “tagged” atoms is often, followed by taking samples of the products, converting them to gaseous com¬, pounds, and analyzing for the atoms of various masses in the mass spectrograph., Experiments with labeled hydrogen can be carried out in much the same way,, using hydrogen enriched with deuterium (H2); this isotope is much more readily, available than is the radioactive isotope tritium (H3), and its behavior is more, nearly the same as that of ordinary hydrogen., One of the most familiar studies using labeled isotopes is that of the, saponification of n-amyl acetate,^ using water enriched with H2018. Since the, , CH, , O, , I, , II, , c=o, , v, , C5HnOH, , acetate resulting from the reaction was enriched in O18, was enriched in O'8 whereas the amyl
Page 162 :
146, , Methods for Determining Reaction Mechanisms, H, OH -heat>, , R2C=0, , +, , R2C—H, , XLIV, , XLV, , (r=ch3o, , CCI3COOD, as a solvent, essentially no deuterium enters, , trichloracetic acid,, , the desoxy compound. This means that the hydrogen atom forming the new, C—H bond in XLV did not come from the solvent and therefore must have, come from the starting material. Moreover, it is well known that the hydrogens, of the —OH groups in alcohols are labile; that is, they easily come off and on, and may “exchange” with any acidic hydrogen or deuterium atoms in solution;, the hydrogen atoms joined to carbon are generally not labile. Very shortly, after the carbinol is mixed with the deuterated solvent, it becomes a mixture of, R2CHOH and R2CHOD. Since, however, no deuterium entered the desoxy, compound, XLV, it follows that none of the hydrogens bound to carbon in this, compound arose from O—H linkages. This study thus points to a mechanism, in which a hydrogen atom with its pair of electrons (that is, a hydride ion), is transferred from one carbinol molecule to the conjugate acid of another as, shown:, , H, , OH, , \ /, , r2ch +, 2, , c, / \, R, , H, , OH,, ->, , )H,, , R,C-H, , R, OH, R?C, , H, , + H-C-R, , +, , H20, , I, R, , In carrying out experiments involving isotopic, assumed, , that, , the, , tagging,, , it is generally, , tagged atom behaves essentially as would an untagge, , atom of the same atomic number. Although we recogntze today that the rat, of a chemical reaction and its position of equilibrium are affected slightly by, substitution of one isotope for another, this isotope effect h, nitrogen, and the heavier elements is almost always so small .ha, , » «, *< does not, , effect the interpretation of the results of isotopic mechanistic stud, , .
Page 163 :
147, , Stereochemical Studies, , deuterium or tritium more caution must be used, for it is not uncommon that, substitution of one hydrogen isotope for another in a reaction will change a rate, constant or an equilibrium constant by 50 percent or more. It should not, how¬, ever, be assumed that the “isotope effect” is merely a nuisance phenomenon, which the chemist must invariably avoid or correct for. As we shall see (p. 192), mechanistic evidence can sometimes be obtained by shrewdly putting this, effect to work., , Stereochemical Studies, Very often, organic reactions are applied to reactants capable of existing in, stereoisomeric forms and, , yield, , products, , that are also capable of existing, , in stereoisomeric forms. In many such cases, an examination of the stereochem¬, istry of the reaction may furnish important clues as to its mechanism. In two, such experiments, which are here selected from a vast number performed by, Hughes, Ingold, and co-workers/6 an optically active reactant yielded a product, which had practically no optical activity. Here, it must be assumed that both, , ^, , ^, , CH, , Cl (optically active) + H20, , —>-, , CH,, , /, , \—CH— OH, , ^', +, , \, , /, , irr, , ,, , ch3 (racemic), h+, , ^H~NH2 (optically active) + HN02, , cr, , +, , CH-OH, , CH3, , CHc, (racemic), +, , N,, , h9o, , /CH3, ->, , c, H, , XLVIII, , * ., 17ft (1950)., , ,, , £ °ne-inrediatC W, , «-* was, , '1S> suPerlmPosa^e on its mirror image. This and other, , ’ "8° ’ and c°-worl<erS; W J. Chm. S,c„ 1937, 1196-1243; (4) AW, ,66>
Page 164 :
148, , Methods for Determining Reaction Mechanisms, , evidence suggests that both of the reactions have the planar carbonium ion, XL VI as an intermediate. (In such a carbonium ion, the positive charge can, be spread over the benzene ring by partial extension of the x-electron density, to the carbon atom adjacent to the ring, a situation that is shown schematically, by including resonance pictures XLVII and XLVIII for the carbonium ion.), It is this carbonium ion that then reacts with the solvent, water, to form the, conjugate acid of the observed product., Far more commonly, reactions of optically active substances yield products, that display at least some optical activity. If this is so, more information must, be obtained before mechanistic conclusions can be drawn. Even if there is only, one asymmetric center in the reactant and one in the product, it is necessary, to know (a) whether the configuration of the product is the same as or different, from that of the reactant, and (b) whether the product is as optically pure as, the reactant or appreciably less so. (The answer to either of these questions is, obtained not by a single experiment but rather by a series of interconversions, and comparisons as will be described in Chap. 7.), If it can be determined that the configuration of the optically active, product is different from that of the optically active reactant, the reaction is said, to proceed with inversion of configuration. This occurs if the reaction path is, similar to the basic hydrolysis of methyl iodide described on page 129. The basic, hydrolysis of 2-rc-octyl bromide in aqueous alcohol is just such a case/60 Here,, the OH- ion hits the halide molecule on one side, the Br~ group leaves from the, opposite side, and the molecule in effect becomes “turned inside out.” This, , Me, , Me, , Me, (Hx=C6H13-), , XLIX, , [ CdLLam ai.'J, , -7, , *, , -, , ...., , 11,., , make the generalization that any reaction yielding an optically, , We may in fact, cannot proceed through an intermediate or transition state, active product cannot proceed through an, , that has lost its, , asymmetry. In particular, free carbonium ions,, , R
Page 165 :
149, , Stereochemical Studies, R', , and free radicals of the type R—C- are excluded as intermediates in such, R", reactions, since neither of these types of fragments can ordinarily support, asymmetry., In the Wolff rearrangement of ^r-butyl diazomethyl ketone (L) to 3-methylvaleric acid, the ^c-butyl group migrates from the carbonyl carbon to the ad¬, jacent carbon. If the diazoketone used is optically active, the resulting acid, , }, I, , I, , CHN,, , Et-CH, , Me/PC, /, a, , Et-CHMe, , 11, O, , *, , ch2, , +, , n2, , /, , G, , IP, , O, , OH, , L, also is optically active, with an optical purity essentially the same as that of, the starting material,57 showing that the migrating j^c-butyl group does not, escape from the remainder of the molecule for sufficient time to allow it to, racemize. More important, the configuration of the resulting acid was shown, to be the same as that of the diazoketone (L)—that is, the reaction occurred, with retention of configuration—indicating that the breaking of bond a and, the formation of bond b occurred on the same side of the ^c-butyl group. Less, frequently, a reaction occurs with retention of configuration because two steps, are ^involved, both with an inversion; the first step turns the molecule “inside, out ’ and the second turns it “outside out” again., We have already seen (p. 137) how the trapping of monobrominated intermediates indicates that the addition of bromine to olefins often takes place in, steps. This stepwise mechanism applies to the bromination of maleic and, fumanc acids, and stereochemical studies further indicate that the intermediate, in the bromination of maleic acid must be different from that in the bromination, o, , umartc acd. The chief dibromosuccinic acid obtained from maleic acid is, , Lined f, , f raCem‘C f°rm LI’ WhereaS the Chief dibromosuccinic acid ob-, , that themo„lranC “a .“ *he nonresolvab^, form, LII * Now suppose, the monobrominated intermediate in these reactions had bromine attached, <o only one carbon atom (that is, could be represented by struauPuP, , ■•mckIL-:,'1640 (1956)-
Page 166 :
150, , Methods for Determining Reaction Mechanisms, , Assuming the configuration about the positively charged carbon to be planar, and assuming free rotation about the C—C bond, the pair of enantiomorphic, intermediate ions from maleic acid would be the same as that from fumaric, acid, and the same mixture of dibromo acids should result in each case. Since, this is not what is observed, an intermediate other than LI 11 must be involved, , HOOG, , COOH, , HOOC, , ''c-0/, rt", , HOOC, , COOH, , HOOC, , nh, , H, , H, , >"C\ COOH, , COOH COOH, , H, , +, , H, , H, , Br, , H, , H, , COOH, , H, LIII, , LIV, , .Br', COOH, , HOOC,, , Br, , [-, , +, , H, , Br, , H, , HOOC, , H, , Br, , COOH, , acid, LI, , d-l, , in one or both cases. The cyclic bromonium ions LIV and LV are consistent, both with the stereochemistry" and the stepwise nature of the addition reaction, and are thus possible intermediates, whereas the cation LIII is not. The conversion of the intermediates LIV and LV to the acids LI and LII respectively,, is most easily visualized by imagining the bromide ion attacking the rig t- an, carbon atoms of the intermediates LIV and LV from the rear of the page., , Limitations of Reactions, Organic chemists are all too familiar with instances in which one compound, will undergo a given reaction successfully but another compound, apparent, , •milGoTe first, will stubbornly refuse to react under the same conditions., « Roberts and Kimball,, , J. Am. Chem. Soc.,, , 59, 947 (1937).
Page 167 :
151, , Limitations of Reactions, , Such differences are sometimes best explained by assuming a transition state, or intermediate into which the second compound, for some good reason, cannot, O, , II, be converted. As a very simple example, the conversion of an amide R, , C, , NHo, , to the amine RNH2 by the “Hofmann rearrangement” (p. 128) cannot be, used to convert the amide R—C—NHCH3 to RNH—CH3, simply because, , II, O, O, , II, species analogous to the necessary intermediates R—C—NBr and R—N—C—O, cannot form from the N-methyl amide., Similarly,, , most a,/?-unsaturated, , acids are readily decarboxylated, but, , the acid LVI is not.50 This difference may be explained by assuming that inter¬, mediates in the decarboxylations of a,/3-unsaturated acids are the isomeric, 0,7-unsaturated acids. The latter are known to suffer decarboxylation readily,, possibly through a cyclic intermediate or transition state such as LVII. On this, , H, , H, , /, (CH3)3C-CH=CH-COOH, LVI, , r-g=, H, , C, rch9-ch, , ^, , f CH,, , \"N, , +, , /, , o—c, , w, , CH., , o, , o, LVII, , o, , basis, acid LVI would be stable simply because it cannot be converted without, skeletal rearrangement to a 0,7-unsaturated acid., More subtly, dimethylaniline readily undergoes condensation with the, enzenediazomum ion to form />-dimethylaminoazobcnzene (LVIII), , whereas, , staetesucaheansLXmilar Xy'idene’ LIX’ d°CS n°‘-3' A" intermediate « transition, as LX .s consistent with this disparity if it is assumed that in such a, , CHS, Ph—N=N, , -N(CH3)2, , LVIII, , \—N(CH3)2, , PhN, , ux, S0‘” 72’ 4359 (1950)
Page 168 :
152, , Methods for Determining Reaction Mechanisms, , structure the substituents on the 2 and 6 positions of the benzene ring must lie, in or near the same plane as the two N-methyl groups. If the 2 and 6 substit¬, uents are hydrogen atoms (that is, R=H), such coplanarity is possible, but if, they are both methyl groups, it may be shown with atomic models that the, hydrogen atoms of the ring methyl groups interfere with those of the N-methyls,, preventing the necessary coplanarity. Thus, when the PhN£ ion attempts to, attack the 4 position in LIX, there is no way the p electrons of the amine nitrogen, atom can spread themselves over the ir orbitals of the ring to help accommodate, the added positive charge, and the new bond does not form., As a final example in which a reaction is rendered unsuccessful by a small, structural modification in the reactant, consider the rearrangement of the, bromomagnesium, , derivative, , of, , m-2-chloro-l-methyl-1 -indanol, , (LXI), , to, , 2-methyl-l -indanone (LXI I). If the trans isomer is used in place of the as, the, yield of indanone drops from 80 to 5 percent,5* and large amounts of tars appear, in the reaction mixture. This indicates that a successful reaction requires that, the C—Cl bond be broken by the magnesium atom on one side of the ring, while, at the same time and on the opposite side of the ring, a methyl group, migrates from one carbon atom to another. A transition state such as LXI 11, is thus consistent with the difference in reactivity of the two stereoisomers., , Br, LXI 11, , Physical Detection of Intermediates, [„ a few favorable cases the presence of an intermediate that may, , 3r impossible to isolate from the reaetton m.xture may be detected, .•Geissman, , and Akawie, J. Am. Chm. So'., 73, 1393 (1951)., , y
Page 169 :
153, , Physical Detection of Intermediates, , measurements. One of the most familiar examples is the nitronium ion, NO^,, the active nitrating agent in those mixtures of nitric and sulfuric acids used to, nitrate aromatic hydrocarbons. Examinations of the Raman spectra of such, nitrating mixtures revealed the existence of this ion35 many years before its, important role in nitration reactions became appreciated as a result of other, studies (Chap. 11)., In another case, cryoscopic (freezing-point) measurements have helped to, explain some puzzling differences in behavior., , Under ordinary conditions, , mesitoic acid (LXIV) is esterified only with difficulty. However, if this acid is, dissolved in concentrated sulfuric acid and the resulting mixture added to an, excess of alcohol, esterification is rapid and essentially complete. On the other, hand, this method for esterification is ineffective for benzoic acid. Now cryo¬, scopic measurements show that two moles of particles are produced for each, mole of benzoic acid dissolved in concentrated H2S04, but four moles of particles, arc produced for each mole of mesitoic acid dissolved.3-^ By assuming that, benzoic acid exhibits ordinary basic behavior in sulfuric acid, but that mesitoic, acid is converted to the carbonium ion, LXV, we may explain the differences, both in chemistry and cryoscopy. As is indicated, a carbonium ion of the type, CH3, , H3C-Q-COOH + 2 H2S04, CH,, LXIV, , 2HSO7, , COOMe, +, , h3g, MeOH, ->■, , H,0+ +, , .ch3, , CH,, -v'-, , 4 moles of particles, , %, , y—eooH, , + h2 so4, , —> HS07 +, \, , /, , C(OH)2, , no, , “V", , 2 moles of particles, 4*, , c—b; rrc, , we ^, , 'z
Page 170 :
154, , Methods for Determining Reaction Mechanisms, , species with unpaired electrons tend to move into an externally applied magnetic, field whereas essentially all other compounds and ions tend to move away, from such a field., , When the anion of phenanthraquinone-3-sulfonic acid, , (LXVI) is reduced in aqueous alkali with glucose, the solution turns brown, , -o*s, , 'o,s, , '03S, [H], , \ / \ /, o, , v r\ /, :o., , o, , [H], , •, , •, , oh, , HO, , OH, , LXVI 11, , LXVI I, , LXVI, , \ r\ /, , brown,, paramagnetic, and becomes paramagnetic; then, as the reaction proceeds, the color fades, and the paramagnetism disappears,55 suggesting that the brown intermediate, species is a free radical of the semiquinone type (LXVII). The final yellow, product is the corresponding hydroquinone, LXVIII., , EXERCISES FOR CHAPTER 5, 1. How can it be shown that the Wolff-Kishner reaction does not proceed by a freeradical mechanism?, 2. In the so-called Fries rearrangement, aryl esters such as phenyl acetate are con¬, verted to acyl phenols such as o-hydroxyacetophenone by action of Lewis acids., Devise an experiment to show whether this “rearrangement " is intramolecular or, intermolecular., 3, , Phenanthrene dibromide (I) is known to lose HBr on standing to yield9-bromophenanthrene (II). How would you show that Ms not an interme ,, , -HBr, , II, bromination of phenanthrene itself to form II?, 4. Why would you expect the hydrolysis of, , ££, , likely to go through a carbon.um ton, , of cLe halides does, , ethyl chloride (IV)? Devise an experiment to show that, not form a free earbonium ion during its hydrolysis., j( Michaelis, Bocker, and Reber, J. Am. Ch'm., , or.,, , ,
Page 171 :
Exercises for Chapter 5, , 155, , CHC1, , I, CHo, IV, 5. The decarboxylation of oxaloacetic has been found to be accelerated by Cu2+., HO—C—C—CH2COOH -^4 CO9 + HO—C—C—CH3, , II II, , ll ll, , o o, , o o, , It has been suggested that an intermediate is a chelate involving the anion of the, acid. Devise an experiment to show which of the chelate structures shown below is, more likely to be the active intermediate in this reaction., H, I, , S, , c, , -c-ch2cooh, , hooc-c', or, , cx, , P, , C, , I, o, , /, , Cu, \, , H, , .Cu, O, , 6. In the Curtins rearrangement, benzoyl azide, C6HS-C-N3, is converted to N, and, , .socyanate C,H.-N=C=0. How could you show that the four-membered, tng compound \ is not an intermediate in this reaction?, , phenyl, , c=o, N-H, V, 7. What evidence indicates that, the hydrolyses of the, two chlorides below proceed, through different mechanisms?, , (CH3)3CCH2C1, , 8. Diaryl sulfones, , are converted to diaryl selenides on heating with seleni, turn., O, Ar-S-Ar + Se -> Ar-S,Ar—Se—Ar + SO 2, , O
Page 172 :
156, , -, , Methods for Determining Reaction Mechanisms, , Devise experiments that would show:, (a) Whether the carbon atoms bound to selenium in the product are those bonded to, sulfur in the original sulfone., (b) Whether the two aryl groups in a molecule of the product came from a single, molecule or from two different molecules of sulfone., 9. In the Fischer indole synthesis, an example of which is shown below, a phenylhydrazone is converted to an indole by action of acid., , Outline two experiments, one using isotopic labeling, the other without, to determine, which of the two nitrogens of the phenylhydrazone is lost during the ieaction., 10. If the quaternary ammonium iodide VI is heated with aqueous base, the olefin VII, is slowly formed and trimethylamine is eliminated. Two mechanisms are proposed:, (a) a two-step process, and (b) a “concerted'’ process. Devise two separate experi¬, ments, one with isotopic labeling, the other without (neither involving kinetics) to, choose between (a) and (b)., , (a), , Me, , Me, , PhCH2-CH-CH2NMe3, , ^ °H, , I, , PhCH2-C-CH2NMe3, 2, (fast, equilibrium), , VI, ->, , PhCH2C— CH2NMe+3, , PhCH9—C=CH2 + NMe3, VII, , Me, , ', , (slow), , I, , Me, , (b), , Me, , PhCH2-CjCH2-NMe3, , ->, , PhCH2-C=CH2 + NMe3(slow), Me, , H, VII, , HO:^, Which, Wmcn method, memuu is the less equivocal?, -n—, , phenylpyruvic acid:, PhCH2—C—COOH, , O, , PhCH2COOH + CO
Page 173 :
157, , Exercises for Chapter 5, OH, , 12. Benzil, Ph—C—C—Ph, is converted by base to benzilic acid, Ph2C—COOH., , II II, , o o, Devise a nonkinetic experiment to choose between the stepwise mechanism (a) and, the “concerted” mechanism (b)., (a), , Ph, Ph, , ^ Ph, , "h, , OH, , '', , Ph, , C—C —O, , o o, , (fast, equilibrium), , O OH, , Ph2C—C=0, , I, , (slow), , I, , -O OH, , Ph2CI, -O, , -C-OH, , taut., , II, , ~>, , Ph2C—C—O', , o, , (very fast), , OH, , (b), , slow, , HO —C—CPh2, , II, , I, , o o~, 13', , fast, , ->, , 2, , 2-chiorrcIohexa„o„eis, , —, , O, II, , 0-C-C(Ph)2, OH, *, , proposed mechanisms (a) and (b)bdow Nme^ha" (bUnv'T' '°, be‘Ween, intermediate whereas (a) does not., ‘, V° VCS 3 cycIoProPanone, (a)
Page 174 :
158, , Methods for Determining Reaction Mechanisms, See how your method compares with that of Loftfield, J. Am. Chem. Soc., 73, 4707, (1951)., , 14. When the toluenesulfonic ester IX is dissolved in glacial acetic acid, it is converted, to the acetate XI. The cyclic ion X has been proposed as an intermediate. Devise an, experiment that would indicate that such an ion is an active intermediate., , XI
Page 175 :
CHAPTER, , 6, , Methods for Determining, Reaction Mechanisms, II. Kinetic Studies, , used for the investigation of reaction mechanisms, we, have postponed until last the method of most general applicability. For vir¬, tually all reactions there is (at least in principle) a quantitative relationship, between the reaction speed at a given temperature and the concentration of the, reagents. The chief object of a kinetic study is the determination of this rela■onship In addition, mechanistic hints may sometimes arise from measureIn discussing the tools, , sTrenvthof the'""I ” ** f***1 ^ with temperature, with the ionic, g, reaction medium, or with the solvent composition. The majority, usefu. organic reactions are complex, taking place in a series of steps o'e, these steps must be slower than the others, and this step often fh„t, f, .unate,y not always) determines ^ ^ „f ^, , i)ouicneckl’;The kinetic smdy °f *—, which species participl “, “, , Sha“ P°int °Ut> amhtguities may arise in intipremtiol^f, , 'mi S”’ 52s"araple’ W Mm!”, 1955, pp. 528-580. For more, , in, , physical them., , ed-> Prentice-Hall, Inc, , New Vrv.L, , Mechanism John Wiley'and Sons, Inc. N^YoT’, ^ ^ Pearson’ Kinetics and, ynng The Theory of Rate Processes, McGraw Hill Rr v n, ^ Glasstone> Laidler, and, P rticular interest to organic chemists is Friess and, ^ New York> 1941 • (<0 Of, Mechanisms” (Vol. VIII of Weissberger’s, W/^berger, “Investigations of Rates and, ers, Inc., New York, 1953, pp. M 535., ^ °/0^ Ch^ry), Interscience Publbh, , 159
Page 176 :
160, , Methods for Determining Reaction Mechanisms, This text is concerned mainly with reactions in solution. Gas-phase reactions, , are much less common in organic chemistry laboratories, and the treatment of, heterogeneous reactions is sufficiently complex so that it is best deferred to, more specialized works.2, , Experimental Methods, Success in a kinetic study depends largely on the selection of a suitable method, for following the course of the reaction. Chemical analyses are very frequently, used. Aliquots may be taken from a given reaction mixture at accurately deter¬, mined time intervals and then titrated for one or more of the reactants or prod¬, ucts. Alternately, a number of identical reaction mixtures may be made up,, the individual mixtures taken one by- one at fixed times during the course of, the reaction, and the total contents of each mixture titrated. The methods for, following the rates of a large number of organic reactions have been recently, summarized by Friess.;(<i), Unless the titration reaction is very much faster than the reaction being, studied, it is necessary to stop or slow down (to “quench ) the lattei by cooling,, by high dilution, or by chemical action before beginning the titration, other¬, wise the titration results obviously will not reflect the composition of the mixture, at the time the sample was removed., Where possible, it is generally preferable to follow the reaction by measur¬, ing some physical property of the solution that changes as the reaction pro¬, gresses. The need for quenching is eliminated in such cases, the system is not, disturbed by the sampling process, and more measurements may be carried, out within a given time. Spectrophotometric methods have proven particularly, valuable, especially in cases where one of the reactants absorbs strongly at, one wavelength with one of the products absorbing strongly at another. Meas¬, urements of changes in optical rotation, refractive index, conductivity and, dielectric constant have also been employed. Even the measurement"f, small volume changes accompanying a reaction in solution (d.latomet y), frequently used, but in such cases, the temperature of the reaction must be, keDt constant to within 0.002 or less., Reactions which are inconveniently slow at room temperatures may be, , ingeniously constructed to allow the composition o, , t e, , * See for example, Laidler, Chemical Kinetics, McGraw-Hill Book Co., Inc., New York,, 1950; See, for example, the study of aromatic nitration reported by Hughes, Ingold, and Reed,, J. Chem. Soc., 1950, 2438.
Page 177 :
161, , First-, Second-, and Third-order Reactions, , be determined after the reactants have been in contact with each other for, only a fraction of a second.4, , First-, Second-, and Third-order Reactions, It will be recalled that the rate of a first-order reaction is proportional to the con¬, centration of just one reactant. Representing the rate of decrease of the con¬, centration of reactant A in the usual manner as —d(A)/dt, the defining equation, for a first-order reaction is then, d(A), it, , (1), , =k^A), , or, in its integrated form,, In ^)° _, (A), , C, , 1, , ,, , (2), , (here, the concentration of reactant A at the beginning of the reaction is (4)0;, after time, t, the concentration drops to (A); and kx is the first-order rate, constant)., Equation (2) emphasizes an important feature of first-order reactions; the, time in which a given fraction of the reactant is consumed does not depend upon, concentration; that is, A/A„ depends only on the value of t. This characteristic, is exemplified by the most familiar type of first-order reaction, radioactive, decay, for which the half-life period (or, if you wish, the third-, fourth, or tenthlife period) of a decaying species is characteristic of the species but wholly, , A typical first-order reaction is the hydrolysis of /-butyl chloride in formic, acid containing small amounts of water. The rate depends only upon the con, centration of the alkyl chloride, being independent of the, , concLratuTof, , MejCCl + H20 -> Me.COI I + HC1., Jt- = *,(f-BuCl), Similarly, first-order kinetics have been observed for the loss of hal'd, , h", , (3), , f, , r “ 7™ akoho', , generally complex;, , ^ ", , mixture in spite of the apparent simplicity of the, Denbigh, W.tOFL^dlScU“e4d0^7l944)‘, , •"c,r.'ssr5 s, , uSnes> J. Chem. Soc., 1937, 1187., , kinetics, , ‘ ^, , reaCti°n, , * ■—*, , «, , Colloid Chm., 51, 505 (1947)- and by
Page 178 :
162, , Methods for Determining Reaction Mechanisms, A first-order reaction of a very different nature is the decomposition of, , di-/-butyl peroxide (CH3)3C—O—O—C(CH3)3.e Carried out in the vapor, phase in the presence of a large quantity of glass wool, the reaction is, 150°, , (CH3)3C—O—O—C(CH3)3-» 2(CH3)2C=0 + c2h6, — </(/-Bu202), , /iT) ^, -x—jt-- = kx(t-Bu202), , ,.N, (4), , As we shall see (Chap. 16), the yields and identities of the products may be, changed by removing the glass wool, or more particularly, by carrying out the, decomposition in a variety of solvents. The decomposition itself, however,, continues to obey first-order kinetics., For most second-order reactions, the rate is proportional to the product of, two concentrations. Calling these two concentrations (A) and (B), the defining, equation is then, zAAl = kM)(B), at, , (5), , which, when integrated, gives, 1, , ,, , (A)o(B), , (6), , = k2t, , (B)o — (^)o ° (B)o(A), where (^l)o and (B) 0 are the concentrations of the reagents at the beginning of, the reaction. Such a reaction, although “second-order overall,” is said to he, first-order with respect to (B) and first order with respect to (A)., Typical of second-order reactions are a tremendous number of nucleo¬, philic substitution reactions on primary alkyl halides-for example, the baste, hydrolysis of methyl bromide in dilute aqueous alcohol., , = *,(MeBr)(OH-), , CH3Br + OH- —> CH3OH + Br, , (7), , dt, , The bromination and iodination of acetone in dilute aqueous, , base are, to a, , good approximation, second-order reactions., , CH3—C—CH3 +, , CH3—C—CH2i, , I, , o, , O, — ^(Me2C=Q) =, , ^.2(OH-)(Me2C=0), , dT, 0 Raley, Rust, and Vaughan, J., , Am. Chem. Soc., 70, 1336 (1948), , (8)
Page 179 :
First-, Second-, and Third-order Reactions, , 163, , Note that the rate here is proportional to the concentrations of acetone and, hydroxide but independent of the concentration of iodine.7, Less often, the rate of a second-order reaction is proportional to the square, of the concentration of a single reagent. The “Diels-Alder type” dimerization, of cyclopentadiene is such a case.'8, , The integrated form of equation (9) is simply 1 /A — 1 /A0 = hot, an expression, that may be applied also to a second-order reaction between two different, reagents whose initial concentrations are the same., The rate of a third-order reaction may be proportional to the product of three, concentrations. The reaction of ethylene oxide, H>C-CH-, with HBr in, , \ X, o, , water is of this type.9, O, , HO, , / \, , Br, , ||, , H,C-CHo + H+ + Br-, , H2C—CH2, , —~jt-Z = *a(C2H40)(H+)(Br-), , (10), , The cyanide-catalyzed benzoin condensation is also a third-order reaction,, first order in cyanide and second order in benzaldehyde.70-^, , O, 2PhCHO XX ph—CH—C-Ph, OH, ■o'(PhCHO), -jt-= ij(PhCHO)2(C,N-), , (U), , Many reactions, a number of them treated in subsequent chapters present, me ac ptetures that are even more complicated. For example, the formation of, , in a very large excess is neveiSetaf not'fndudid'taThe, , “ ?•“' ''here 3 reaSent not present, As, , » Stern,V^r^50?5n(iC90’5f ^, , we, , sha"pres-, , 5'' 428 <1929>-, , cally. For a tabulation of integrated* ratVcaT™ (l0) and <"> are rather complex algebraikinetic situations, see, . numbt of
Page 180 :
164, , -, , Methods for Determining Reaction Mechanisms, , O, , I, , acetone semicarbazone, (Me2C=N—NH—C—NH2), from acetone and semicarbazide73-73 is subject to general acid catalysis. If the reaction be carried out, in the presence of a weak acid, H4, two terms will appear in the rate equation,, one involving the catalytic constant of the hydronium ion, £H+, the other involv¬, ing the catalytic constant associated with the weak acid, knA., Me2C=0 + NH2—NH—C—NHo, II, , H+, , Me,C=N—NH—C—NH2 + H,0, II, , o, , o, o, , Me2C=0) _ (Me2C=0) (“NH2—NH—C—NH2”), U'H/1 (Hd) +*u+(H+)], 1 + (H+)/A>, dt, , (12), , o, where (“NH2—NH—C—NH2”) refers to the total stoichiometric concentra¬, tion of added semicarbazide, including that which is converted to its conjugate, acid (acidity constant Ks+). The denominator in reaction (12) serves merely, to convert this total concentration to the concentration of unionized semi¬, carbazide (Ex. 2(b), p. 195), the reactive nucleophilic species. The rate law is, thus less formidable than it might seem at first glance., An equally complex rate law governs the reaction between H2S and cyanamide in slightly acid solution to form thiourea:7'', S, NH2—C=N + H2S —> NH2—C—nh2, -</(NH2CN), dt, , *2(“H2S”)(“NH2CN"), , (13), , [i + /cc/(h+)T[i + (h+)Z^3, , (Again, the quantities in quotation marks refer to “stoichiometric, , concentra¬, , tions, whereas Kc and tfH,s are the acidity constants of cyanamide and H2S,, , It is appropriate to emphasize at this poiat the d.stmct.on between the, order of a reaction and its modularity. We have seen that the order of a simple, reaction is merely a description of the algebraic form of its differential-rate, equation. On the other hand, we shall define the molecular,ty of a reaction as, the number of distinct chemical species (ions, molecules or free atoms), bonds or suffer the breakage of old bonds durtng the rate-determ,nmg step. The mole, larity of a reaction thus depends upon an assumed mechanism (which may, 11 Conant and Bartlett, J. Am. Chem. Soc., 54, 2881 (1932)., » Westheimer, ibid., 56, 1962 (1934)., (1930)- 53, 1270 (1931)., U Buchanan and Barsky, J. Am. Chem. Soc., 52, 195 (193U;,
Page 181 :
Determination of the Order of Reactions, , 165, , right or wrong) and depends further on the borderline one wishes to draw, between bond formation and electrostatic attraction. We see then how there, may be considerable controversy concerning the molecularity of a reaction,, whereas if adequate kinetic data are available there should be substantial, agreement as to its order., , Determination of the Order of Reactions, Kinetic data generally consist of reagent concentrations (or quantities that can, be converted to concentrations) at various times. For relatively simple reactions, of integral orders, such as reactions (7) through (10), the reaction order and, rate constant may be obtained from the available data in a number of different, ways. Often it is convenient to assume an order, to calculate the values of k, corresponding to each experimental point, and then to note whether the calcu¬, lated “k values” are substantially the same or whether they show a “drift”, as the reaction progresses. (Thus, all values of, , should be the same, , t, if the reaction is first order, whereas all values of V(^) ~ U(A)0 should bg the, same if the reaction is second order with the reactants present in equal con¬, centrations.) Alternatively, the values of the appropriate concentration func¬, tions, , ln(T),, , 1 /(A), , ln04)o(£) ~ ln(ff)0(T), , etc., , may be plotted against, , tune, noting which of these plots most closely approximates a straight line,, e rate constant may then be obtained by measuring the slope of the linear plot., i, . ,n,US?f. SUCh methods> two Provisions are important. First, the reaction, Should be followed until it is considerably more than half complete, for during, , he early stages of a reaction the observed rate often shows a satisfactory “fit”, to two or more of the common integrated rate expressions., tionsofCrfrf,h mmber °f mn! Sh0U‘d be made’ Var>'in« the ™tial concentrathe reactants and any suspected catalysts. Since the concentration, , sumed during the reaction’ i,, deceptively simple kineti, , ff, , fr?C“°n °f SUCh 3 reaSent wil1 be con-, , ’ • ** C ^ ^ ^ ^ may pass unnoticed, and a, , experimemoTthe baI“dP‘TUre T !* f0med' F°r «amP'c., , » single, , out with a large excess of blso^th, ^ br°midC (reaCtion 7), carried, concentration of the bromide’ tl( "j* '!°U.ld aPPear t0 depend only on the, kinetics”. Howevejf the sal, , ‘ p, , of hydroxide, the apparent “first ord'e" *, , Pparent, , S°-Calkd “P«udo-firs,-order, tW° dlffcrent concentrations, , first-order rate constant” will be found to be pro-
Page 182 :
166, , -, , Methods lor Determining Reaction Mechanisms, , portional to the concentration of base, showing that this concentration should, also be included in the rate law.15, Since the concentration of solvent remains very nearly constant as a reac¬, tion proceeds, conventional kinetic studies alone cannot detect the participation, of the solvent in a reaction. It is, of course, possible to select a reaction that is, generally carried out in, say, methanol and carry it out instead in benzene,, noting whether the reaction in benzene is accelerated by addition of small, amounts of methanol. By varying the concentration of methanol, the order of, the reaction with respect to methanol (in benzene) may be determined. The, transfer from methanol to benzene, however, involves a drastic change in the, reaction environment, and there is no guarantee that the nature of the reaction, in the two solvents is even approximately the same., Reactions having more complex rate laws, such as (12) and (13), cannot, truly be said to have a “reaction order.” For these, integrated rate expres¬, sions, if they can be derived at all, are apt to be extremely unwieldy. Rate, laws are obtained most conveniently by estimating the various values of the, rate itself (d(A)/dt) in a number of reaction mixtures having different con¬, centrations of reagents, catalysts, and possible inhibitors. The rate at a given, instant is the slope of a curve made by plotting concentration against time;, this slope (except for a zero-order reaction) must vary with the progress of the, reaction, but may be estimated at any desired moment by drawing a tangent, to the curve and measuring the slope of the tangent. This procedure, although, simple in principle, is somewhat unsatisfactory in practice unless carried out, with the aid of a prism,'*<“> a mirror, ,6m or other such device. The most sig¬, nificant datum is generally the initial rate of the reaction (that is, the slope, when t = 0) for at this point the concentrations of the various species are presumably known; complications due to reversibility or to the occurrence of side, reactions have not yet set in.1', , Reversible Reactions, Uthough in theory no reaction goes entirely “to completion,” the reader is, loubtless aware that a very large number proceed until the concent, a ion o, <■ Similarly, the nitration of toluene in aC'*“'wiTa, rder kinetics (rate independent of all concent: a, , ?, , of HNol, , u is attacked after the rate-determining, , rhe toluene does not appear in the, decrease in the concentration of the rate-determining, tep, whereas there is only a very' “ ^^tion. Under these conditions, the nitration proeagent, HN03, during the corns, , of the toluene is consumed., , ^, , -, , 56, 1696 (1944); French, ibid., 72, 48UO U, and 74, 1268 (1952)., , )>
Page 183 :
167, , Reversible Reactions, , one or more reactants is immeasurably small. However, for cases in which con¬, siderable quantities of reactants remain when equilibrium has been reached,, kinetics must account for the “slowing down” of the net reaction as the products, formed in the forward reaction become involved in the reverse reaction, regen¬, erating the reactants. We shall consider here only the simplest possible example, of a reversible reaction/5 the first-order conversion of reactant A to product B,, opposed by the first-order reconversion of B to A:, , A^B, , (14), , kr, , The rate constants for the forward and reverse reactions are designated as, , kf, , kr, respectively. For further simplicity, assume that initially no B is present., Reactant A is destroyed in the forward reaction but reformed in the reverse, and, , reaction, so the differential rate equation is, , - ^ = k,(A) - k,(B), or, (since every molecule of, , -, , (15), , B is formed at the expense of a molecule of A), = kf(A) -, , kr[(A)0 - (A)], , (16), , From this, we may obtain the integrated rate expression (Ex. 2a):, , t, , kf(A^ o, */(-4)o - [k,+ ir][(d)„ - (^)J = (*/ + k’)‘, , !f we let (A), be the concentration of, .17) may be simplified to, , In, , (17), , A when equilibrium is reached, equation, , (./i)0 — (A)e, (A) - (A), , — (kf + kr)t, , (is;, , /‘e th?‘ eqUa“°n (18) is similar in f°nn to equation (2), suggesting that th, “ toward equilibrium is a ,<first order, hav.ng 5 rat«cons(‘r, SUm may, 15 bC °b'ained fr°m the concentrations of A at th, the react,on, at time /, and at equilibrium. Furthermore, the quotier, ,/k, is simply the equilibrium constant for the reaction; this may be calculate, tart of h, , t, , rard7„“Ltrteconstandt * “, , 5n’ Ref- 1 (b), pp. 172-176., , allowi"S -para,ion of th, , PkX rcvcrsible tactions, see, for example, Frost and Peai
Page 184 :
168, , Methods for Determining Reaction Mechanisms, , Consecutive Reactions; The Steady-state Approximation, The ease in handling consecutive reactions (a set of reactions in which the, product from one step becomes the reactant in another) depends largely on the, relative magnitudes of the individual rate constants. Consider, for example, the, reaction sequence,, , k, k', T —> 5 —> C, , The mathematical treatment of such a sequence is most difficult when the rate, constants k and k' are of the same order of magnitude/9 but, happily, such, cases are quite rare. If k is much greater than k and if the conversion of A to, B is essentially irreversible, a large quantity of the product B will be formed in, a short time and will then be converted slowly to C. Thus, B will generally be, isolable, and if this is so, the “reaction sequence” is most simply studied as two, separate reactions., Often, however, the first step in such a sequence is rapid but reversible., A small equilibrium concentration of B is rapidly formed from A, and as B is, slowly consumed to form C, the A, , B equilibrium shifts, yielding more B., , The benzoin condensation may be regarded in this way:(m, /, , OH\, PhC, , OH, CN,, PhCHO + CN-, , Ph—C:~ (rapid, reversible), , Keq = (PhCHO) (CN“), , CN, , (19), Ph, , OH, Ph—C: ~ + PhCHO, , I, , CN, , Ph—CH—C—CN (slow), , I, O, , I, OH, - 1 (PhCHO), , (20), , Eliminating the concentration of the cyanohydrin anion from reactions (19), " yfelds a third-order rate expression identical to equat.on (11)., rate = *s/r„(PhCHO)*(CN-) = *a(PhCHO)*(CN-), « The treatment for consecutive ^t-order pactions, , (21), , (Xpp- 539 541., , iVer'ofLriiL (*. to amides, then to carhoxyiate sal.)., in excess base.
Page 186 :
170, , Methods for Determining Reaction Mechanisms, NO, +, NO+ + ArH, , Ar, , (fast, k'), H, , (29), , J, , NO, +, Ar, , + NOj -* ArN02 + HN03 (fast), H, , (30), , J, , The rate of the overall reaction should then be equal to the rate of reaction, (28), the dissociation of the conjugate acid of nitric into the nitronium ion and, water:, rate = ((H2NO+) = kK g(HNO,), , (31), , The rate is thus proportional only to (HNOa), and when this is in large excess,, the nitration rate is very nearly constant (“zero order”). The kinetic picture, changes, however, for less reactive aromatic compounds such as chlorobenzene, and ethyl benzoate. With these, step (29) becomes the slow, hence rate-deter¬, mining, step. Since this step involves (ArH), the nitration of these compounds, now becomes first order in the aromatic compound. Still another kinetic situa¬, tion arises when one or more fast steps follow a slow step that is, however,, reversible. This is the case, for example, in the hydrolysis of benzhydryl chloride, (Ph2CH—-Cl) in aqueous acetone:**, , Ph2CH—Cl, , Cl- + Ph2CH+, kr, , Ph2CH—OH, , (32), , *, , As shown, benzhydryl chloride ionizes slowly (rate constant kf) to the benzhydryl, ion, which may be converted back to the starting material (rate constant kr) or,, alternatively, may react with water to form benzhydrol (rate constant k')., The reaction is thus a competition between chloride and water for the benz¬, hydryl ion. The rate of the overall reaction depends not only on the rate of the, slow step but also upon the efficiency with which water competes against, chloride., The rigorous treatment of rate equations describing such a reaction se¬, quence, although possible, is difficult. The kinetics in this case may be handled, conveniently by calling upon the so-called steady-state approximation This, is used when all of the intermediates in a reaction are present m relatively, small quantities. In particular, considering the hydrolysis of benzhydryl chlo¬, ride, , we may assume that the concentration of the benzhydryl ion does, -■Hughes, Ingold, and Taher, 7. Om., , ibid., 966., , 1940, 949. Church, Hughes, and Ingold,
Page 187 :
Parallel Reactions, , 171, , vary appreciably during the course of the reaction/' The rate of formation of, this ion from benzhydryl chloride must then be equal to its rate of consumption, by the combined action of water and chloride., A,(Ph2CHCl) =, formation, , *r(Cl-)(Ph2CH+), , +, , consumption by Cl-, , *'(Ph2CH+), , (33), , consumption by water, , (Since water is a component of the solvent, its concentration, which remains, very nearly constant throughout the reaction, is incorporated into k!.) The, rate of the overall reaction is actually the rate of formation of benzhydrol from, the benzhydryl ion and water., rate = *'(Ph2CH+), , (34), , If we solve for the concentration of benzhydryl ion in reaction (33) and sub¬, stitute this expression in (34) we obtain:, £'£/(Ph2CHCl) _ £/(Ph2CHCl), k' + *r(Cl-), , 1 + Ay(Cl-), , From this, it may be seen that the rate of the reaction will be lowered by addi¬, tion of chloride ion. If no extra chloride is added, the reaction will appear to, follow first-order kinetics during the early stages since the concentration of, chloride ion will at first be very small/5, , Parallel Reactions, Quite often we are confronted with a system in which two or more reactions, are proceeding independently. (If this were not the case, the practicing organic, chemist would be pleased with quantitative yields of products from a large, percentage of his preparations.) Among the simplest systems involving parallel, reactions are substitution reactions in which more than one isomer may be, formed from the same starting material. In the nitration of chlorobenzene, for, xample, three mtrochlorobenzenes (abbreviated NCB) are formed, each at a, excesTof ah'^ T^0” ‘S, °Ut in nitrom«thane with a large, excess of nunc acid; under these conditions, the formation of each of the three, isomers is first order in chlorobenzene., , increase steadily, proc^ds untU £“ * “ ^ U?’ 'tS concentration should, by physical means. On the other hand if the •, Piesence could be eventually detected, formed ,he reaction could no, ge, under wly, Wm C°nSUmed W"*iably faster than i, is, for complex reaction sequences' nvoMngS'rnumberoflati°,nnaVe Pr°Ved ex,remely useful, hese, exact solutions become almost prohibitive^ L, mtermediat«; for many of, reatmg reaction sequences of many steps has he ^ A * ■ Cj V ^ generalized formulation for, 33B, 145 (1936) and 37B, 374Tl937) lZ I, £, deV1Sed ^ Chbstiansen, Z. physik. Chem, 1 Book Go-> Inc-, New York, 1940, p., ammett’ Physical Organic Chemistry, McGraw-
Page 188 :
172, , -, , Methods for Determining Reaction Mechanisms, , cf(o-NCB), dt, , = i„(PhCl); rf(,”-NCB) = *„(PhCl);, , ortho, , dSt^L>, , met a, , = i„(PhCl), , (36), , para, , The disappearance of chlorobenzene, —^(PhCl)^ -s aiso a ftrst_orcjer reaction, dt, having a rate constant (ka + km + kp). To evaluate the three individual rate, constants, we must know not only their sum (calculated by following the rate, of disappearance of chlorobenzene), but also their ratio. The ratio of any two, of the three rate constants may be obtained by dividing two of the three rate, equations in (36), then integrating. In this way, it may be shown that if the, reaction is stopped at any instant, the concentrations of the three isomers, present will stand in the same ratio as the respective rate constants; that is,, (o-NCB): (tw-NCB) : (/>-NCB) = k0:km:kp, , (37), , The reliability of the values of the individual rate constants depends largely, upon how accurately the composition of a mixture containing three closely, related compounds can be determined. Although ‘"analyses, , have been carried, , out in the past by fractional crystallization or distillation, such separation, methods are generally not as well suited for quantitative work as physico¬, chemical analyses that do not involve losses during separation. If three products, are involved, spectrophotometric analyses may often be used; with only two, products, measurements of refractive index, freezing point or other physical, properties may yield the necessary information. Recently, analysis by isotopic, dilution has become an effective analytical took4 (Ex. 3)., Similar considerations may be shown to apply to any system of parallel, reactions yielding two or more products if the order of each of the individual, reactions with respect to the given reactants is the same. Analyses of the prod¬, ucts will yield the relative rate constants even though it is not known whether, the reaction is first, second, or even zero order. There, is, however, one addi¬, tional proviso-no significant reversal of the individual reactions may occur, before the mixture is analyzed. If the individual reactions are substantia ly, reversible, the composition of the resulting mixture will drift toward that p, , -, , vailing at equilibrium, and the analysis of the product will reflect equ.libnu, condifions rather than the relative rates of the forward reactions. (In such a, case the composition of the product is said to be thermodynamually controlled,, rather than kinetically controlled.), Very frequently, two reactants will compete for, •, •, a*, tup kinetic, or for the same reaction intermediate., straightforward if the competing reactions are, , ., the same starting mater, treatment in such cases is, ^ order For example,, -, , ** See, for example, the work of Roberts and co-workers on the nitration, benzenes, J. Am. Chem. Soc., 76, 4525 (1954).
Page 189 :
Parallel Reactions, , 173, , when methyl iodide is treated with a solution containing both hydroxide and, phenoxide ions, both anions may bring about substitution reactions., , CH3I + OH-, , CH3OH, , CH3I + OPh-, , ► CH3OPh, , </(MeOH), dt, , = AOH(MeI)(OH-), , (38), , = ^orh(Mel) (OPh~), , (39), , (/(MeOPh), dt, , If both OH~ and OPh~ are in large excess, the disappearance of methyl iodide, becomes “pseudo first order” with a rate constant [£0h(OH~) + £0ph(OPh-)]., Dividing reaction (38) by (39), and integrating, we may obtain:, (MeOH) _, (MeOPh), , Aoh(OH-), (40), , ^Ph(OPh-), , By following the disappearance of methyl iodide, then analyzing the mixture, of products resulting from a solution in which (OH~) and (OPh~) are known,, we may evaluate the individual rate constants. If smaller amounts of OHand OPh~ are present initially so that the concentrations of these species show, substantial decrease during the reaction, division of reactions (38) by (39) and, integration gives:, , log, , (MeOH), 1, , (OH-)o, , *0H, ^OPh, , log, , 1, , -, , (MeOPh), , __ log of fraction OH, , remaining, , log of fraction OPh“ remaining, , (OPh-)o, , Similar treatments are appropriate for competitions between two or more species, or a reaction intermediate. For example, if a mixture having known quantities, o benzene and ^--xylene is nitrated and the reaction is quenched before either, o the aromatics is entirely consumed, analysis of the reaction mixture will, determine the relative rates at which the two hydrocarbons react with the NO+, wenhfversee„°ThVeer’, be *«“% evaluated since, as, e have seen, the overall nitration rate is simply the rate of formation of the, 02 ion and in this case is independent of the species being nitrated, , alky, bromides*5 may be regard d ', ", ^ °f 3, tion (independent of OH, l ! C°mpetItlon, second-order reac iol?of t^ “ ", „ See :, 10n (°f tHe, , °f S“°"d-y, a first-order reac-, , ^ hydr°lysis °f ^ b™»We) and a, the hydrolysis of methyl bromide)., , the basic hydrolysuVi^prop^l 'bromfdT sDc^aT^ '7', S°‘’’ ,937; 1177’ ”92' on, WHS1;*8 actually being consumed in two different ^ ° SltUatl°n may arise because the, The su ' oefrrgent U deStr°yed is kinetically comphcalT h°J “, sum of two simple reactions (see Chap. 8, Ex. 12), ’ °, , that the ProCess bV, approximated as, , COnven,entIy
Page 190 :
174, , -, , Methods for Determining Reaction Mechanisms, rate =, , — </(RBr), dt, , = Ay (RBr) + yt2(RBr)(OH-), , (42), , There are several ways for treating such “pairs” of reactions. Suppose, for, example,26 that the rate of appearance of bromide ion (disappearance of alkyl, bromide) is determined graphically for a number of mixtures containing known, concentrations of hydroxide and alkyl bromide. Since reaction (42) may be, rewritten,, rate, (RBr), , = Ay + *,(OH-), , (43), , we see that a new plot of the values of [rate/(RBr)] vs. (OH ) should yield a, straight line with Ay as a slope. The intercept—the specific rate when (OH~), is zero—should be Ay., , Mechanistic Implications from Rate Laws, There is no general way of proceeding from the empirical rate law for a reac¬, tion to its mechanism. Commonly, the worker lists the most plausible mecha¬, nisms, derives hypothetical rate laws for each, and then decides which of these, corresponds most closely to the observed rate law. If the reaction is relatively, simple, the experienced kineticist can often make a good guess (but still only a, guess) as to ifl^rtechanism merely by inspection of the rate law., The benzidine rearrangement, reaction (44), for example, has been founc, to show third-order kinetics, being second order in (H+):27, , — ^(PhNHNHPh) = £3(phNHNHPh)(H+)2, dt
Page 191 :
175, , Mechanistic Implications from Rate Laws, , 0.2 M) acid has been found to be second order in HN02, suggesting that two, molecules of nitrous acid somehow participate in the sequence leading to the, rate-determining step, even though the stoichiometry of the reaction demands, only one.**, PhNH+ + HN02, , PhN=N+ + H20, , rate = £3(HN02)2(PhNH2), , (45), , The kinetic data are consistent with a rate-determining step involving the, amine and a species containing two atoms of tripositive nitrogen, probably, N203:, 2HNO-, , H20, , +, , N203, , (fast, equilibrium), H, , Ph-N-H, , i, , H, , +, , \, , N=o] + +, , (b, , -> [Ph—N—, , N/ \, , 0=N, , N=0, , H, ■>, , [Ph-NH2-N=0]H, , PhN=N+, , 0-N=0“, (slow), , +, , H20, , (fast), , As indicated, the third step in the above sequence is composite; but since, kinetic data are singularly ineffective in furnishing information about the course, of a reaction sequence after the slow step, we can say little about the conversion, of the N-nitroso compound to the diazonium cation except that all the steps, involved are faster than the nitrosation step. Without the kinetic study, however,, we would be unaware of the participation of the second molecule of nitrous acid, earlier in the reaction., On the other hand, if a reagent is known to be consumed but does not, appear in the rate law, one may assume that its action is delayed until after the, rate-determining step. We have seen this to be the case for the nitration of, touene (p. 169), where the rate law does not include toluene. A similar conclusion may be drawn from the kinetics of the base-catalyzed iodination of, acetone reaction (8). The rate law, which includes the first powers of the con, a sulrr rre and hydrOXidC (bUt n0t iodi"e>’, be explained by, suming that before acetone may be iodinated, i, must be converted in a slow, step to its conjugate base., CTT, , /~'i, , rT, , OH ~, , -c—CH3-,, slow, , rapid, , o, 8 Schmid and Muhr, Ber, , i*, , ch3- -C—CHy, , o, 70, , 421, , CH;, , -c-, , -CHoI + I-, , I, , o, tvt *, , i, , wncentration of aniline rather than that of^lnihnium° onTh ^ ^ CXPression involves the, moderately acid solution is small and must be^^ caTe d« J r' conceun,rMion, the free base in, se and acid, using the known basicity constant or, ?"d ,fr°!n ,he concentrations of added, diaaotnations of aromatic amines may exhihh a °f a",lmC' " sh°nW also be emphasized thai, Hughe? "|’ ‘T, 3 de,ai,cd discussion of the kmctrss'arirf" laws’.dePending upon reaction, Hughes, Ingold, and Ridd, J. Chem. Sdc., 1958, 58 98, mechamsm °f diazotization, see
Page 192 :
176, , Methods for Determining Reaction Mechanisms, , From this mechanism, we would predict that any reagent that reacts rapidly, with the conjugate base of acetone should react with a given basic solution of, acetone at the same overall rate as is exhibited by iodine. Thus, the base-catalyzed, bromination of acetone has the same specific rate as the base-catalyzed iodination®9 since both reactions are governed by the same rate-determining step., A species appearing in the denominator of a rate law is generally formed in, a reversible reaction preceding the rate-determining step. An excess of this, species slows down the overall reaction by shifting the equilibrium so that a, smaller quantity of a reactive intermediate is available. The rate of basecatalyzed iodination of aniline in the presence of excess iodide is, for example,, inversely proportional to the square of the iodide concentration30, , + I3, , +, , + 5H, , B, , -</(PhNH2), , k(B~)( PhNH2)(Ij), , it, , (I")2, , +, , 21“, , (46), , where B~ represents the base. This suggests that an active intermediate is formed, along with two iodide ions in a reversible step early in the reaction. One of the, three possible mechanisms consistent wi h this rate law (Ex. 5) assumes the, attacking agent to be the iodonium ion, :I+ (or its hydrated form, H2OI+)., , -3, , ^, , :l:+, , +, , 21, , (rapid equilibrium, Kx), , _I/ ^, , -> BH, , (rapid equilibrium, K2), , ;NH;, , +, , I —)-NH,, , (slow, k), , The overall rate (.ha, is, the rate of the final step) is then in agreement with, reaction (46)., , -) = tlfs(PhNH2)(It)(B-) =, (PhNH2)(IJ)(^ ) (47), kK,K,, , *» Bartlett, J. Am., 30, , Berliner, J. Am, , Soc., 56, 967 (1934)., . Soc., 72, 4003 (1950).
Page 194 :
178, , Methods for Determining Reaction Mechanisms, , in the rate-determining step,3- and that this step is simply a direct displacement, of bromide by hydroxide., , +, , HO:', , H,C-Br, , ->, , HO-CH3, , +, , : Br:, , (51), , However, if the concentration of the entering group does not appear in the rate, law (indicating that this group enters the picture after the rate-determining, step), we may conclude that the slow step is the breakage of the carbon-halogen, bond in the alkyl halide (reaction 23). A similar mechanism is indicated by a, rate law such as given in reaction (35), but here the slow step is reversible., Much more will be said about the substitution reactions of alkyl halides in, Chapter, , ., , 8, , The Transition-state Theory. Energy of Activation, , 33, , One of the most familiar general phenomena of chemistry is the increase in the, rate of a reaction as the temperature is raised. This acceleration cannot be due, merely to the increased number of molecular collisions, for as a rule the reaction, rate increases with temperature far more sharply than does the collision fre¬, quency. It will be recalled that reactant molecules having “average” velocities, generally do not undergo chemical reactions since they have insufficient energy, to allow the formation of the necessary transition state. The transition state has, already been described (p. 129) as the summit of an “energy hill”: reactants, must have enough energy to “roll up” this hill before they may “roll down”, the other side, releasing energy and forming the products. The rate of the reac¬, tion is then related to the number of molecules that pass from the “reactant, side” to the “product side” in a given time. Raising the temperature increases, the concentration of molecules with sufficient energy to make the ascent, an, more conversions from reactants to products then occur., A molecule or group of molecules passing through the transition state is, said to be an activated complex. Considering a reaction between species d, and, , B, we may designate the activated complex as AB* and represent tie, , reaction as, , A + B —>, , ABX, , —> products, , (52), , activated, complex, , * however, methyl bromide were, consider the possibility that neither OH noi •, P, _ f, but rather that this step involves the conjugate ba^e ErGH^ , °, in the reaction, OH“ + CH3Br ^ H20 + BrCH* ‘, might then be similar to that for the ring c osure o, M For detailed discussions of'theLtransit‘°n^4_99, (1935); (6) Prosed, reaction-rate problem, the collision me y, topic, see Glasstone, Laidler, and Eyring, , d rapidly and reversibly, fsis Qf ^ reaction, , ’ lc„c 'hlorohydrin (reaction 22)., gee ca\ Eyring, Chem. Revs., 17, 65, ^’alternate method of handling the, , not be discussed here; for a treatment of this,, (Ref. lc), PP-
Page 195 :
179, , The Transition-state Theory. Energy of Activation, , The transition-state theory treats the activated complex as a chemical species, whose formation from the reactants is associated with an equilibrium constant,, K*, where, {AB*), , 7ABt, , (53), , (A)(B) yAy3, , We shall, for the time being, consider only dilute solutions in which the activity, coefficients approach unity, thus allowing us to ignore the y terms in (53)., Since the rate of reaction is assumed to be proportional to the number of, A-B couples passing through, , the transition state—that is, proportional to, , {ABl)—the specific rate [kr = rate/(T)(£)] for the reaction should be propor¬, tional to AT Furthermore, using the principles of statistical mechanics, it is, possible to show5^ that the constant of proportionality is very close to kT/h,, where k is Boltzmann’s constant, T is the absolute temperature, and h is Planck’s, constant. The transition-state theory further defines the following quantities,, analogous to the coi responding thermodynamic functions used in the descrip¬, tion of ordinary chemical changes:, The free energy of activation, AFb, , AFt = —RT In A* = —RT In, , Uf, (54), , kT, The heat of activation, AHb, AH* =, , —R, , Kt) - — R, , d{\/T), , ~, , tf(ln, , kT), , lorn, , + rj, , (55), , The entropy of activation, AAb, AS* =, , _ R, , rj-, ^(ln kf), , krh, , 1 ~dT~ +lnIr, , (56), , - 1, , (See Ex. 8a.), In addition to these quantities, workers often refer to the Arrhenius energy of, actwauon, E., obta.ned simply by plotting the logarithms of the rate constants, r a given reaction at a number of temperatures against the reciprocals of the, , »ul5C‘by' -T1™" meaSUrinS thC Sl°Pe °f thC “Hne” "**'***•” and, , „ XT' ,, , ■', , <f(ln k) _, , E„, , Wrn, , ~s, , (57), , ^ °n‘y ^ ab°Ut °'6 kCa‘ (‘he —, , ^ at ordi-, , 3' 107 0735)., , h, , reactions", ”,g«e“in* that «*>' activation, funcZ of, M “Ch T,S aic a«ua% curvedXha, is hXT''-Crai'mS °" very sensitive, ancnon of temperature. See, for examp£ LaMer Ind Miller', °f, “ a, UIcr’, Am- Chem. Soc., 57, 2674, , TaT'% 7*”*"
Page 196 :
180, , Methods for Determining Reaction Mechanisms, , nary temperatures). Since this difference is less than the uncertainty in many, measurements of activation energies, either Ea or AHx may be regarded as the, extra energy that must be imparted to an average molecular ensemble consisting, of discrete species A and B in order to allow formation of activated complex, AB*. The energy of activation is an important item in any mechanistic descrip¬, tion of a reaction, just as the free energy of reaction is an important item in its, thermodynamic description., In our earlier picture of the transition state in the reaction OH~ + CH3I, —» HOCH3 + I-, we saw that the new bond (C—O) was being formed at, the same time the old bond (C—I) was breaking. The activated complex in, such a reaction has a higher energy than either the reactants or the products,, largely because the energy released in the partial formation of the new bond is, considerably less than the energy needed to stretch the old bond well past its, most stable configuration. It has indeed been suggested that the activation, energy of a reaction is proportional to the strength of the bond that suffers, breakage.However, this is only a rough estimate that is approximately true, for a number of gas-phase reactions but completely inapplicable to reactions, in solution where solvation effects contribute to a much more complex picture., There appears to be considerable variation among reactions in the degree, of completeness with which the new bond is formed in the activated complex., Greater progress toward complete bond formation results in the availability of, more energy for the necessary distortion of the old bond, thus lowering the, activation energy. For example, the energy of activation for the reaction, (C2H6) 3N: + R—I -> [(C2H,)3N—R]+ + I-, , (58), , is only 9.7 kcal per mole when R is CH3— but is 16.0 kcal when R is, (CH3)2CH—.sr The attacking nitrogen atom may approach the carbon in, methyl iodide and, with little interference, begin the process of bond formation, that results ultimately in the displacement of the iodide io_ and the “Walden, inversion” (p. 148) of the methyl group. The reaction of the amine with iso¬, propyl iodide is subject to what is classically termed “steric hindrance —mean¬, ing that there is a great deal of interference between the bulky ethyl groups of, the amine and the a-methyl groups of the iodide when the amine draws near, to the iodinated carbon. The amine molecule must therefore show sufficient, energy to press on toward the reaction site while intermolecular interference is, forcing the a-methyl groups of the iodide out of the way., , This, , ,nutates, , both, , the, , inversion process and the breakage of the C-I bond well before substant, energy from the newly forming C-N bond has become available., We can also understand why the saponifications of esters have relatively, " Hirschfelder, J. Chem. Phys.,9, 645 (1941)*, ., 17 Brown and Eldrcd, J. Am. Chem. Soc., 71, 445 (1949).
Page 197 :
Entropy of Activation, , 181, , low activation energies (about 11 kcal)33 even though they involve the breakage, of the relatively strong C—O bond. The incoming OH~ group can bind itself, completely to the carbonyl carbon before the alkoxide group begins to break, off., , (P, , OH, ■>, , OH" + CHo-C, , \, OR, , OH, , /, , CHo-C-O, 3, I, OR, , ■> CHo-C, %, , + OR-(59), , O, , The energy barrier that hampers the reaction is lowered when the saponification, proceeds through anion V; for along such a route the system receives much of, its, , energy payment, , from the formation of the new C—OH bond before having, , to pay its “energy debt” for the breaking of the C—OR bond.33, , Entropy of Activation, Entropy is the measurement of the randomness of a system. If a reaction occurs, with an increase in entropy, there is more disorder possible among the products than, among the reactants—that is, there are more restrictions to the motion of the reac¬, tant molecules than to the motion of the product molecules. Analogously the, entropy of activation, which may be calculated from reaction (56), is a measure, of the freedom from restraint of motion among the reactants. For reactions in, solution, entropy effects also include changes in the randomness of the solvent, , the reactants"eW ^, few generalities1 First, , reqUiri"S differing, ^, , °f ’« formed from, ““"W ^ We "“Y n°te a, , rr’jsz-irr.* **•, , When two molecules come together Lt, , T, , " PeF deSIee-, , rs *•—--*?££ sasasassr, CyCllC products> a negative entropy of activation is to be ex, 99 The d, , Ohson, Z. physik. Chem., 118, 99 (1925)., , metastable intermediate, , compl«io„ Ofthc bond, , °VersimPlified since the ion V is actually a, , because the bond breaking £
Page 198 :
182, , Methods for Determining Reaction Mechanisms, , pected since free rotation about the single bonds becomes restricted during the, cyclization. Thus, the isomerization of allyl vinyl ether to (3-vinylpropionaldehyde (VII), thought to proceed through the cyclic intermediate VI, shows an, entropy of activation of — 7.7 calories per degree.40, , CHo=CH, ->, , CH?—CH, /, 2, %, H2C^, O, , ., ., (60), , ch=ch2, , CH —CH2, , VII, , VI, , A low entropy of activation may also reflect a crowded transition state in, which freedom of motion of the substituents is unduly hindered. The entropy, of activation for the reaction,, H, , \, ElO- + RCH,I, , ElO, , C., , EtO—CII2R + I-, , I, , (61), , •/ \, H, , R, , ., , is -9.5 calory per degree if R is H, but falls to -19.9 calory per degree if R, is, , t-Bu.4i, , The three extra methyl groups in the latter compound obviously, , add to the crowding in the transition state., Reactions of neutral molecules to yield ions invariably show negative, entropies of activation. For such cases, the charge separation begins in the, transition state and each end of the dipole becomes solvated with a sheath of, solvent molecules, which must however be suitably oriented. The increase in, orientation means, of course, a decrease in entropy. When a reaction in which, charge is being created is carried out in a number of solvents of varying po anty,, the greatest entropy decrease in the formation of the transition state will gen¬, erally be observed for the, , least, , polar solvent. For example, the reaction between, , aniline and phenacyl bromide,, + + Br~, C—CHoBr -> - Ph—C—CH>—NH,Ph“, O, , n, , O, , Stein and Murphy, J. Am Chem. Soc., 74^°41 (195“)Dostrovsky and Hughes, J. Chem. Soc., 1946,, , it Cox, J. Chem. Soc., 119, 142 (1921)., , (62)
Page 199 :
183, , Influence of Solvent, , bered that the entropy of activation represents a change in randomness and that, the molecules of polar solvents have a considerable measure of orientation,, even in the absence of solutes. The molecules of the less polar solvents therefore, suffer the greatest increase of orientation under the influence of a polar transition, state since they begin with so little orientation. As one would expect, a reaction, involving two ions of the same charge also exhibits a negative entropy of activa¬, tion (the transition state has a charge greater than either of the reactants),, whereas reactions in which two oppositely charged ions react to yield uncharged, products exhibit positive entropies of activation., For a reaction having a negative entropy of activation, we see that it is, not enough that the reactant molecules come together with the necessary, activation energy; they must also acquire an additional measure of orientation., Now, one of the fundamental characteristics of our physical world is that a, mixture in which the molecules are randomly oriented is a more probable sys¬, tem than the same mixture in which there is a high degree of order among the, molecules. It follows then that only a fraction of the collisions with the necessary, activation energy will actually result in reaction; the more highly negative, the entropy of activation, the less this fraction will be.* A reaction therefore, may be slowed down quite as effectively by a highly negative entropy of activation as by a large positive energy of activation., , Influence of Solvent, ^^eThWm 3ffeCt reaCti°n, rela, , d to, , he, , much «» *e same way as they, , ^, , and llion Cxhe, , ^ °f a action, , equilibrium between reactant, , honS bet":, uncharged molecules., , closely, , are “fw—, Reaction, , in polar solvents; typical of thi, , •, , , ., ™ *, , 1C, , lons are generated from, , ^rge.is CUated Proceed most rapidly, , amines by the “Menschutkin reacdon^(R3N+Tl, , seen, the extremely polar solvents are mLt capTbleof, solvation while themselves undergoing the least ren, , Crr ^ ^, , ^, , haVC, d‘SPerSmg CharSe, , -vents,, ^uPv°aStSible/° esti™te the number of, , -oleiuks
Page 200 :
184, , Methods for Determining Reaction Mechanisms, , such solvents will tend to facilitate the destruction of charge. Thus reactions, such as the Hofmann elimination (R3N+—C2H5 -f- OH- —» R3N + CH2=CH2, -f- H20), in which uncharged molecules are formed from ions, proceed more, rapidly in less polar solvents.44 When an ion reacts with an uncharged molecule, (for example, the reaction CH3I + OH~ —> CH3OH + I~), the charge, which, is localized to a single ion before reaction, becomes dispersed over a somewhat, larger area in the transition state. Since the charge density has been thus, (temporarily) diminished, the transition state requires less solvation than the, reactants. In this respect such a reaction is similar to that between ions of op¬, posite charge and should therefore be more rapid in poorly ionizing solvents., However, the effect of solvent upon velocity should be much less marked than, for reactions in which charge is destroyed. For reactions in which both the, reactants and products are uncharged, the effects of solvent on rate are slight, and need not be considered here., It should be noted that a variation in solvent may change not only the, speed of a reaction but also its apparent order. The ethanolysis of a-phenylethyl chloride in absolute ethanol is, like the hydrolysis of isopropyl bromide, (reaction 42), a competition between a first- and second-order reaction.4, Ph—CH—Cl + OEt~ —> Ph—CH—OEt + Cl-, , |, , I, , ch3, , /, , Me, , ch3, , \, , -APhCHCl/ =, d‘, , /PhCHCl\ + k2 /PhCHClX (OEr), (, , Me, , ), , \, , Me, , (63), , ), , However, if water is added to the mixture, increasing the polarity, the firstorder component of the reaction (which presumably involves an lomza, is accelerated, whereas the second-order component (which presuma, , ), , y >, , - volves the direct attack of an ion on a neutral alkyl halide molecule) is « ar e ■, In 80 percent alcohol, the latter component effectively vanishes, and, lion appears to obey an ordinary first-order rate law., « These arguments may be put on, to form such, distribution in the transition state, calc“la ? 1 differing dielectric constants, and equating, a transition state from the reactants in, a, example, Laidler and Eyeing, Ann., these energy values to the respective va iks, >, (1932). See also Ex. 6., nT.aJ&L 39, 303 (1940); ScatchardI, Cta.Sme, in essence, that a, Such treatments, however are °ve ‘“P, constant. It would then follow that a, solvent affects a reaction only through its dl, mixtures having the same dielectric conreaction should have identical rat<es inq, 80, with experience. See, for example, Brown, slant, a conclusionand Saville, W, 1955, 5114., and Hudson, J. Chan. Soc.,, >, 10117 1201, « Hughes, Ingold, and Scott, J.Chm. SoC„ 1937, 1201., it Ward, J. Chan. Soc., 1927, 445.
Page 201 :
Influence of Ionic Strength. Salt Effects, , 185, , Influence of Ionic Strength. Salt Effects, Let us reconsider a reaction between species A and B—first-order in both—, proceeding through the activated complex ABX., A + B —> (ABl) —> products, The expression for the rate constant, kr, for this reaction is given by the transi¬, tion state theory as, hr _, , (AB)* _ kT Rt yAyB, h, , (A) (B), , h, , yABt, , (64), , Since the activity coefficients (the y terms) of dissolved species vary with the, ionic strength^' of a solution, it follows that the rate constant for a reaction, should likewise depend upon the total concentration of ions in solution, whether, or not these ions participate in the reaction at hand. This is especially so if both, of the reactants are ions; for such cases, the effect of variations in ionic strength, may be predicted by the simple Debye-Huckel theory.* According to this, treatment, the logarithm of the activity coefficient y for ion A in a solution, having an ionic strength less than 0.01 is, to a fair approximation,, , equation (64), we obtain, , (68), 1, , where, valent
Page 202 :
186, , Methods for Determining Reaction Mechanisms, , Thus, a reaction between two positive or between two negative ions will be, accelerated by increases in ionic strength, whereas a reaction between a positive, and a negative ion will be slowed down. It is emphasized that equation (68), applies quantitatively only to dilute solutions in which the added salts are fully, dissociated and interionic attraction is at a minimum. For more concentrated, solutions, particularly in solvents of low dielectric constant, more complex, expressions must be used, but equation (68) still predicts the direction in which, ionic strength influences reaction rates., If one of the reactants (say A) is an ion, whereas B is a neutral molecule,, the activated complex, AB*, should have the same charge as A, and, accoiding, to equation (66), the same activity coefficient. If the activity coefficient of B, is also independent of ionic strength, equation (64) would predict that the rate, constant should not be subject to a salt effect (that is, the final term in equation, (68) should vanish). Although such reactions are considerably less sensitive to, salt effects than are reactions between ions, their rates are often measurably, affected by high concentrations of added salt. Not only is the activity coefficient, of a neutral solute generally raised by addition of salts (the “salting-out” effect),, but at high ionic strengths where equation (66) is no longer applicable, the, activity coefficient of an ion depends on factors other than its charge. There is,, at present, no simple way of predicting the magnitude, or even the direction,, of the salt effect on such reactions.^ Equation (68) predicts that the rate of a, reaction between neutral molecules to form ions should not be affected by ionic, strength. It is known, however, that such jeactions (for example, f-BuCl +, H.O —> t-BuOHf + Cl-) show positive salt ejects—that is, they go faster at hig, ionic strengths. Although the activated complex in such a reaction has no net, charge, it is a strong dipole since the separation of charges has already begun., Its activity coefficient Taat, which appears in the denominator of equation (64), decreases (and kr thus increases) with an increase in ionic strength., , We ca, , “Land why, in the absence of added electrolytes, *era. =, for a reaction that produces ions will show an upward drift as the reacuo, progr-s As more and more ions are formed, the magnitude of the positive, salt effect increases, and the specific rate rises., , i9 See, however, Amis and, , ^, , a reaction that produces ions,, , 10 A simplified, °, ^^940, 979), yields the relationship, given by Bateman, et al. fJ. Chem. i>oc.,, h y, ', , In kT = In k°r + aZW, where « is the rate constant at «, and the temperature, m the ionic stre g, , ,, , applicati0n of this equation to the quantitative, , and d the distance between the ends, prediction of salt effects is lirmtedlarge7, , the difficulty in estimating, a priori, the fract.ona, V, with an activated complex whose exist-
Page 203 :
Ambiguities in Interpreting Kinetic Data, , -, , 187, , In summary, then, reactions in which ionic charge is created are facilitated, by solvents having high dielectric constants and by high ionic strengths. Reac¬, tions in which ionic charge is destroyed are facilitated by solvents having low, dielectric constants and by low ionic strengths., The salt effects that we have been discussing are generally grouped to¬, gether as primary salt effects. They arise because of variations in the activity coeffi¬, cients of the species participating in the rate-determining step and the activity, coefficient, , of the activated complex. There is, however, another important, , way in which the rate of a reaction may be influenced by ionic strength. In, a number of reactions, the rate-determining step occurs after a rapid, reversible, , step involving ions, and the position of equilibrium in the preliminary step must, also be subject to salt effects. For example, an acid-catalyzed reaction occurring, in the presence of an uncharged weak acid, HA, may show a positive salt effect, simply because the degree of ionization of HA,, , HA, , A, , -j- H+ (solvated), , rises as the ionic strength is increased—that is, addition of salt results in a greater, conversion of HA into a more powerful catalyst, the lyonium ion. This so-called, , secondary salt effect would be far less important, perhaps negligible, if the catalyz¬, ing acid were of the type £H+. The equilibrium is subject only to minor salt, , B + H+ (solvated), effects since here ionic charge is neither created nor destroyed, In view of the complications that may be engendered by salt effects, conffol of, , ionic strength is essential in kinetic studies of reactions involving ions., , Often the ionic strength is kept essentially constant by the addition of f large, , I1“h"p'd’ d“"», V, , “, , desired the reaction rate is measured in a number, , r*, , 'ni *■, , »«...«, , :zz“, , of solutions of different, , Ambiguities in Interpreting Kinetic Data, , , *, , XSS t,T la7 f°r the *"* Bolyais of methy,, OHjBr and OH- o al7e™atel, ^‘“"g step involving, •hough the seconds, , “T ^, , HsO. A1, , of accord with a substantial body of chemical e'he kme'1C P‘CtUre’ “ is out, excluded. Similar ambiguities in Lterpretatlon of*, not so easiiy resoivcd-, , —* .he, , jUStifiably. *
Page 204 :
188, , -, , Methods for Determining Reaction Mechanisms, , with hypochlorous acid, momentarily ignoring all salt effects., , The reaction, , is first order in both reagents;5* the rate law may thus be written, RNH > + HOC1 —> RNHC.l + H20, , rate = £2(RNH2)(HOCl), , (70), , On the other hand, since the acid and the amine participate in the acid-base, equilibrium,, „, (RNH+XOC1-), A - ^7Nh2)(HOC1), , RNH2 + HOC1 ^ RNH+ + OC1, , nu, ^, , the product (RNH2)(HOCl) is proportional to the product (RNH+)(OCl ),, and the rate law has an alternate form,, rate = A'(RNH+)(OCl"), , O72), , which suggests that the chlorination reaction instead involves the RNH, an, OC1- ions" If we now bring the salt effect into the picture, a choice between, the two mechanisms might at first appear possible, since a reaction temeen, oppositely charged ions should show a pronounced negative salt effect wher, a reaction between uncharged species should show little or none. Howeve, a closer look shows us that an examination of the salt effect wil no giv, answer, , although, , reaction, , (70) would no, exhibit a negative prmary salt, , ‘effect, It is subject to a negative secondary salt effect. An increase in tome streng, eneci,, j, ions afe formed at the expense, , shifts the equ.hbnum (71) so that mo^ ^, , analysis of this problem, , (Exb^^show^tha^th^ primary salt effect ^nflimnringai^cfien^^ dilution, , patdhsdwould result in idemical rate, # choice must be, , laws even when activity coefficients.are, This type of, , another se’t that is in rapid equilibrium, , :adhethefiTlt ishideed one of the most important limitations on mechanistic, conclusions derived from kinetic studies., , Mechanisms, /vyecnamsma, , of -Acid and Base Catalysis, , i, . nf, Having examined the rate laws, position to understand why, catalysis, whereas others are su >j, , •, , number of reactions, we are now, s are subject to general acid or base, ific iyonium- or lyate-ion catalysis,, undergo a given reaction only, , Suppose, for example, that a substanC ', ‘, he formation of H.V+ is rapid, when converted to its conjugate acid, HA .111, Weil and Morris, 7., *0, 7., .664 (1949).
Page 205 :
Mechanisms of Acid and Base Catalysis, , -, , 189, , and reversible, the reaction sequence is, H+, , X, , Y, , HX+ —» products, rr, , k*, , r, , Keq, , and the rate law is, rate = *,(H X+)(Y) = k,K„(H+) (X)(Y), , (73), , The reaction is thus subject to specific lyomum-ion catalysis. However, if the forma¬, tion of HA+ is slower than its reaction with Y, the rate-determining step be¬, comes the transfer of a proton from an acid HA present in the mixture,, H.4, , „, , Y, , X ——> HA'+ —> products, slow, , fast, , If there are a number of acids (HA, HA', HA", etc.) present, each will transfer, protons to Ar at a different rate, and the rate of the overall reaction will be the, sum of a number of terms, each containing the concentration and rate constant, for one of the acids present., rate = k(X){HA) +, , k'(X)(HA'), , + k"(X)(HA") • • •, , (74), , This is a rate law for general acid catalysis., General acid catalysis is also observed if the rate-determining step involves, formed05'", , ”ded C°mP'eX °freagent X and acid HA’ reversibly and rapidly, HA, , y, , ^ ^=~ X HA —> products, *eq, , kl, , If there are several acids present, several complexes are possible each, one, reacting with reagent Y at its own rate., ’, rate - k2(X-HA)(Y) + k’2(XHA')(Y) + k'J(X-HA"){Y) • • ., (V(y)[X„k,(HA) + K’„k'2(HA') + K'^k”(HA'f) ■ ■ •], , (75), reaction, , Hi, , B, , X^==, , K,, , eq, , —> products, *2, , If the formation of ,YH+ is reversible and rapid, the rate law is, , The re, , •, , = ^.<>(B)(V)(H+) = k,_K„K%n,(BH+)(X), , a conclusion perhaps not hnmediamly, , (76), , ^ ^ add BH+’, , value of the product (B) X (H+), and both B
Page 206 :
190, , -, , Methods for Determining Reaction Mechanisms, , and H+ are present in the activated complex. Furthermore, if there are a num¬, ber of bases present, each may react with A'H+ at its own rate; the rate of, disappearance of X would be the sum of a number of terjns of the same type as, given in equation (76), each with the same value of Keq and (X), but with dif¬, ferent values of &2, 7C^H+, and (5H+). The reaction is therefore subject to general, catalysis by acids of the type BH+., By similar reasoning, it is easy to show that reactions which may be repre¬, sented by the following sequences are subject to general base catalysis,, B, , Y, , HX-> X~ —> products, slow, , fast, , Y, , fast, , and, , X + B, , X B-> products, eq, , slow, , whereas a reaction sequence of the following type:, Y, , fast, , HX + B ^ X~ + Z?H+-* products, eq, , slow, , should be subject to specific lyate-ion catalysis., Finally, if the reaction sequence is of the type:, B, , fast, , X + HA, , X HA-> products, eq, , slow, , the reaction is subject to general catalysis, both by acids and bases. (See, for example,, the mutarotation of tetramethyl glucose, p. 139.), , Reaction Rates and Acidity Scales52, It is now recognized that the rates of a number of acid-catalyzed reactions in, strongly, , acid, , media are proportional to Hammett’s h0 function (p., , whereas the rates of others are more nearly proportional to the concentration, of lyonium ion (in water, to (H,0+)). In the present text, we shall assume as ts, generally done, that reactions in the first class proceed through activated co, plexes that differ from the reactants only by addition of a proton. The justification, of such an inference is as follows. Suppose the reaction sequence ts, slow, , V _L H+ rfas^-" ATI+, A _r, *, , products, k, , The rate of this reaction at finite concentrations is given by the transition-state, « For more detailed discussions of this prohlem,, , see^, , (a) ZuckCT and H, , ^ Paul, , S:™ » “<Rtf-23,1, pp. 273—277.
Page 208 :
192, , -, , Methods for Determining Reaction Mechanisms, , this author, arc fully satisfactory. Thus, although proportionality between rate, and (H30+) is generally taken (probably, in most cases, correctly) to indicate, participation of one or more molecules of solvent in the rate-determining step,, the question must be considered open.5*(c,rf), It will be remembered that the ho function is a useful one in solutions of, high dielectric constant because the ratio Vv/7A-H+ in a given solvent mixture is, largely independent of the nature of X. This is not the case for solutions in, aqueous ethanol,52(d) for which, however, a treatment roughly analogous to, that given in equations (77)-(79) may be set up using the Grunwald acidity, scale (p. 107). This is most easily done by expressing 7CAh+ in the solvent at, hand in terms of the corresponding acidity constant in pure water, using the, additional parameters m0 (specific for reactant X), Y0 (specific for the solvent, mixture), and/H+ (the degenerate activity coefficient for the hydrogen ion in the, given mixture). It may thus be shown (Ex. 12) that if the log (kr/fR+) for a num¬, ber of mixtures be plotted against the F0 values of these mixtures, the resulting, curve should be very nearly a straight line of slope mo. The second-order rate, constants for the acid-catalyzed isomerization VIII —> IX in water-ethanol, mixtures containing from 35 to 100 percent ethanol have been found to fit, such a relationship.55, H, , / %, , C-CH=CH-CH:, OH, VIII, , Isotope Effects, In the discussion of isotopic labeling in the previous chapter, it was pointed out, that isotopic substitution may cause appreciable variation m the rate of a, reaction. This isotope effect is greatest when deuterium (H ) or tntium (H ), is substituted for ordinary hydrogen, particularly when the reaction involves, breaking the bond to the labeled atom. Since the chemistry of a species is ge e ally lugh. to depend upon its electronic configuration rather than on the, , - - “r• - -S r m, , -r ”,, , with this vibration, which persists down to abso, , ,, , t nu, 1948 1982* Gutbezahl and Grunwald, J. Am. Chem., a Braude and Stern, J. Chem. Soc., 1948, 1M, ou, Soc., 75, 572 (1953)., . ., effect see Bigeleisen, J. Chem. Phys., 17, 675, si (a) For a rigorous treatment o, e, P, isot0pe effect is given by YViberg,, (1949). (b) A review of chemical aspects of the deuterium, F, Chem. Revs., 55, 713 (1955).
Page 209 :
Isotope Effects, , -, , 193, , energy; it is greatest for the very light atoms. A C—H bond, for example, has a, greater (about 1.2 kcal per mole greater) zero-point energy than a C—D, bond of the same type; hence the C—H bond is a “looser bond” and is more, easily broken. Furthermore, if the formation of an activated complex requires, partial or total rupture of a carbon-hydrogen bond, the activated complex, will be formed more easily from a C—H bond than from a C—D bond, and the, C—H compound will react more rapidly than the C—D compound. Such an, isotope effect may be detected by actual comparison of the reaction rates of, the labeled and nonlabeled compounds. A more subtle method is to subject, a known isotopic mixture to an incomplete reaction, isolate the unreacted, starting material, and determine whether the isotopic composition has changed, (that is, whether the percentage of C, creased because the C, , D compound in the reactant has in¬, , H compound is consumed more rapidly)., , Since the chemical equilibria depends upon the rates of the forward and, reverse reactions, the position of an equilibria may also be slightly affected by, isotopic substitution. The direction of shift will depend upon whether the dif¬, ference between zero-point energies of labeled and unlabeled compound is, greater for the reactant or for the product. Specifically, substitution with a, , keaVm isot°Pe wil1 favor that bond for which the zero-point energy is less, sensitive to isotopic substitution., It follows then that a complex reaction may show an isotope effect for one, of two reasons. The bond to the labeled atom may be broken or stretched in, e rate-determining step or, alternatively, a rapid and reversible reaction prethis? ‘tk, ?P may mVOlVe breakage °f this bond> the ^tope effect on, qui 1 rium eing reflected in an observable kinetic isotope effect for the, overall reaction. If the bond to the labeled atom is broken a,tl the ra e deter, mtnmg step, little or no kinetic isotope effect is to be expected, Consider for example, the oxidation of benzaldehyde by permanganate, , Zr;, riX'Srtr r ~ d° is broken during the rate-rieie-, , ••• '•, , “h bm7“r ,7“"' k“‘, , bond m the aldeMe group, , /-H 7 is oxidized in this reaction Ts*,' ^ T* °rdinary benzaldehyde at, reaction 7.5 tunes as fast as is C,H6-C=0.« The, , equmbrium inv°iving ^age or the, deuterated aldehyde in ordinary water’TnVdete", °Xidation °f the, hydrogen has replaced the deuterium 1 ,u, ‘ermlninS whether any “light”, On the other hand the, •, ,n, & unreacted starting material., hand, there ,s » observable hydrogen isotope effect in the, 1 "8 and Stcwarb J■ Am- Ch,m. Soc, 77, 1786 (1955).
Page 210 :
194, , -, , Methods for Determining Reaction Mechanisms, , nitration of benzene and its derivatives. Deuterated nitrobenzene, C6D5N02,, for example, is nitrated at the same rate as ordinary nitrobenzene under the, same conditions.56 This means that the C—H (or C—D) bond is not broken in, any equilibrium prior to the rate-determining step, and it is often taken to, mean that the C—H bond is not broken during the rate-determining step either, —that is, that the breakage of the C—H bond occurs in a rapid step following, the slow formation of intermediate X. While the absence of a hydrogen, , H, X, , isotope effect is consistent with the existence of intermediate X, it does not, demand it. Strictly speaking, we may infer only that the stretching of the carbonhydrogen bond (which should be more difficult for C—D than for C—H) does, not play a significant role in the activation process for the slow step—that is,, that the stretching of this bond does not effect the energy of activation for this, step. Thus, although the reactants almost certainly pass through a configuration, corresponding to X, we cannot say, without further knowledge, whether this, represents an actual intermediate (a point of minimum energy in the progiess, of the reaction) or simply an activated complex (a point of maximum energy)., , EXERCISES, , FOR, , CHAPTER 6, , 1. Suggest a method for following the rate of each of the reactions below:, , O, (a) ph—N—C—Ph + C6H6 —> Ph—Ph + N2 + Ph, , COOH, , I, , N=0, (b) C2H4 (gas) + H2 -» C2H6 (gas), (c) PhCH2F + H20 -> PhCH2OH + HF ^, (d) CH3—O—SO7 + Br“ -> CH3Br + SO;, (e) PhN=0 + PhNH2 -» Ph—N—N—Ph + H20, , m, , O, w Gold and Hawes, J. Chem.Soc., 1951, 2102, Hammond, J. Am. Chem. Soc., 77, 334 (1955), n For further discussion of this question, see H
Page 211 :
Exercises for Chapter 6, , 195, , (g), , cr+ ch,-o-so;, , -> 02N—f, , \S-CH,, , S + CH,0S02Ph, , y, , / -OCH3 + SO4", , + PhSO?, , 2. (a) Show how the integrated expressions for the simple reversible reaction, kf, , A^B, kr, , (equations 17 and 18) are obtained from the differential rate equation (16)., (b) The reaction of semicarbazide with acetone, , o, , O, , Me2c, , O + NH2 NH—C—NH,, , MejC=N—NH—G—NH2 -f H20, , is subject to general acid catalysis and is first order both in acetone and in semi¬, carbazide. On this basis derive the rate law given for this reaction (equation 12)., , 3., , Chlorobenzene in which some of the chlorine atoms are radioactive Cl36 is subjected, to nitration in mtromethane. The resulting mixture is neutralized and the organic, ayer separated into three portions having the weight ratio 1:8:1. To the first, portion‘s added 10 0 grams of pure e-nitrochlorobenzene; the mixture is heated until, all of the solid dissolves, then cooled. The o-ni,rochlorobenzene which precipitates is, recrystallized repeatedly until its activity remains constant (after a small correction, for decay) at 1638 counts per minute per gram., The second and third portions of the neutralized nitration mixture are treated with, , m mi m ■— -—- 4. Show that a reaction following the sequence:, , HS +, S~ +, , X, Y, , -S, , + XH+ (fast, reversible), products, (slow), , should be subject to specific lyate-ion catalysis., , 176) for the iodination, (b) One of these mechanisms is extremely^unl Wy T'”" ^, formation of the observed maior isnm/ * • ^ Y’ b, g lnconsistent with the, (c) Propose an experiment to, Ex»lai"using the isotope effect., 1 ie iemainmg two mechanisms,
Page 214 :
198, , Methods for Determining Reaction Mechanisms, O, , O, , II, , II, , O, OE,-, , II, , (c) If the reaction: CH3—C—CH2—C—CH3-> CH3COOEt + CH3—C—CH3, EtOH, , is carried out in the presence of excess ethoxide, its rate is proportional to the, concentration of added acetylacetone. If it is carried out in the presence of excess, acetylacetone, the rate is proportional to the concentration of added ethoxide., Cl, I, , ^d) The reaction CeH5—N—COCH3 —»/>-C1C6H4NHAc is accelerated by HC1 but, not by HOAc or H0SO4., (e) The formation of semicarbazones from ketones and semicarbazide is accelerated, by small concentrations of acid but retarded by large concentrations of acid., Br, (f) The bromide C5H12—CH2—CH—CH3 reacts with water over 30 times as fast as, does the corresponding chloride. The products in both cases are the alcohol, C5H12—CHo—CHOH—CH3 and the olefin C5H12—CH=CH—CH3; and the, ratio olefin to alcohol in the mixture of products is the same for the bromide as, for the chloride, despite the difference in rates., (g) The, , decomposition, , of azomethane,, , CH3—N=N—CH3—>C2H6 + N2 has, , an entropy of activation of +17 calories per degree, whereas the decomposi¬, tion of acetic anhydride in the vapor phase, (CH3CO)oO —> CH3COOH +, CH2=C=0, has an entropy of activation of -4 calory per degree., , (h) The hydrolysis of benzhydryl chloride, Ph2CHCl, is more effectively accelerated, by addition of Li2S04 than by addition of LiN3. However, LiN3 is more effective, in diverting the benzhydryl group to benzhydryl azide than is Li,S04 in divert¬, ing the benzhydryl group to benzhydryl hydrogen sulfate or benzhydryl sulfate., (i) The benzidine rearrangement and the acid-catalyzed hydrolyses of esters and, acetals proceed more rapidly in D20 than in H20., (j) The oxidation of (CH3)2CDOH by chromate is subject to a much more marked, isotope effect than is the oxidation of (CH3)2CHOD., (k) The nitration of toluene in nitromethane with excess nitric acid is retarded by, addition of LiNQ3, but the reaction continues to obey zero-order kinetics., 12. Consider the reaction sequence,, , X, + H+ ^ XH+, (fast, eq), XH+ + Y —» products (slow, £2), carried out in a series of water-ethanol mixtures in which the Grunwald treatment, “applicable Set up the rate law in terms of*, and, , and express the apparent, , specific rate (*, = rate/W(H+)(K». ™, °( th' val ue ° +* “ f”!T, and the other Grunwald parameters. Show that a plot of the values of log (kJU), for the various solvent mixtures vs. the, line of slope, , Y. values for these mtxtures is nearly, , ma. What approximation is necessary?, , stra.g
Page 216 :
200, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , approach toward a simple quantitative treatment of the influence of structure, on reactivity will be based on just such a division/, It is to be emphasized that a given structural change may affect the rate, of a reaction and its position of equilibrium in opposite directions; that is, it may, slow down a reaction but allow it to go more nearly to completion. For example,, if a methyl group is substituted for each of the three a-hydrogens in acetalde¬, hyde, the rate constant for semicarbazone formation is lowered by a factor of, twenty at 25°, but the equilibrium constant for semicarbazone formation is, increased by about 60 percent/ The semicarbazone of trimethylacetaldehyde, is thus more stable to hydrolysis, for the semicarbazone of acetaldehyde, al¬, though formed more rapidly, is hydrolyzed 28 times as rapidly. Substitution, of methyl groups for the a-hydrogens in this reaction has raised the value of, AF* but lowered the value of AF. It appears that equilibrium is much more, quickly established in the system involving the light two-carbon aldehyde and, its semicarbazone than in the system involving the more massive and bulky, five-carbon aldehyde and its semicarbazone., , Inductive and Field Effects, The ionization of carboxylic acids has been more extensively studied than has, any other type of organic chemical equilibrium.5 For such reactions, any, change in structure that will facilitate the removal of the H+ ion from the acid, molecule or hinder its return to the carboxylate anion should increase the ioniza¬, tion constant of the acid—that is, lower the free energy of ionization. (Analogous, 1 Before the 1950’s, it was generally felt that attempts to link structure and reactivity, ex¬, cept for a number of very special cases, would meet one major obstacle. It seemed that al¬, though a structural formula, when interpreted in the light of modern concepts, would tell us, much about potential-energy effects (bond energies, resonance energies and dipole interac¬, tions) it would tell little about the kinetic energy of a reacting species (that is, about the oc¬, cupancy of the various possible rotational and vibrational energy levels) and about the degree, of its solvation. Since both potential and kinetic energies change during ordinary reactions, it, was felt that a structural change might result in predictable potent.ai-energy cffects wh.^, would be, however, offset by unpredictable kinetic energy or solvation effects (See, for exam, pie, Hammett, Physical Organic Chemistry, McGraw-Hill Book Co., Inc., New York, 1, pp. 69i^)ecent, , ,, , ^ however> evidence has been accumulating showing that quantitative, , correlations maybe made by considering independent contributions of inductive, resonance, , cannot generally be (and often need not be) dtsentangW., , Acadde^cafpX inc^Net Yori. 1955, ’pp-' 567-562 Dissociation constants quoted, without' reference within this chapter are taken largely from this source.
Page 217 :
Inductive and Field Effects, , 201, , statements apply, of course, to other Br^nsted acids.) Table 7-1 compares the, pK values (in water at 25°) of a number of aliphatic acids to that of acetic acid, (note once again that a large ionization constant corresponds to a small pK, value)., , Table 7-1. pK Values for Some Carboxylic Acids, pK, , Acid, , CH3COOH, , 4.80, , pK, , Acid, , Acid, , pK, , FCH2COOH, , 2.66, , HOCH.COOH, , 3.83, , (CH3)3N+—CH2—COOH 1.83, , cich2cooh, , 2.86, , N=C—CH2—COOH, , 2.43, , H3N+— (CH2)4—COOH, , 4.27, , CbCHCOOH, , 1.30, , HOOC—CH2—COOH 2.83, , -02C—CH,—COOH, , 5.69, , C13C—COOH, , 0.65, , CH3—CH2—COOH, , 4.88, , -02C(CH2)4C00H, , 5.41, , Cl—(CH2)2COOH 4.0, , (CH3)3C—COOH, , 5.05, , HCOOH, , 3.77, , From the first column, we see the effect of introducing a charged sub¬, stituent. As we would suspect merely from electrostatic considerations, a positive, center such as (CH3)3N+ or NH+ eases the departure of the positive hydrogen, ion from the —COOH group. The effect is large (almost a thousandfold, increase in Ka) for (CH3)3N+—CH2COOH, in which the positive center lies, close to the carboxyl group; it is much smaller (about a threefold increase in, Ka) for H3N+— (CH2)4—COOH, in which the positive center is five atoms re¬, moved from the carboxyl group. This is again in harmony with our knowledge, that electrostatic interactions become weaker as the distance between charges, is increased., Conversely, the introduction of a negative charge should increase the energy, needed to remove an H+ ton and should thus weaken the acid. Substitution, C02 group for an n-hydrogen in acetic acid yields the hydrogen, malonate ton -02C-CH2-C00H. We must, however, be a little carefuf in, , czz, , rrh ofthis acid wi*,hat°face,k’ f°r the ™i°™e, , has four equivalent basic sites whereas the acetate ion has only two., , 1, , o, \, rC-CHo-C, , /, , o, , /, , O', , \, o, , o, ch3-c, , \o, , ion would be half as strong an add as acetic adT Act^ U, , h>/r°§en malonate, , one eighth as strong, so the electrostatic effect here isM’the'', ^, ^, but its magnitude is quite small Th., H, C°rrCCt dlrection”, and the hydrogen malonate ion is, , add hsdf, ppropriate and much more striking.
Page 218 :
202, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , By a statistical argument similar to that above, the two acids should differ in, strength by a factor of 4 if other effects are absent. In truth, malonic acid is, over four hundred times as strong an acid as the hydrogen malonate ion—, that is, the ratio K\/Ki for malonic acid is 734. As would be expected, the ratio, K1/K2 decreases as the distance between the carboxyl groups is increased; this, ratio is 19.2 for succinic, 11.9 for glutaric, and 6.1 for azelaic acid (the ninecarbon dibasic acid)., Furthermore, the ratio Ki/4K2 (note the statistical factor 4) may be taken, as a measure of the electrostatic effect exerted by the negative —COy group, at one end of a molecule on the arrival or departure of a H+ ion at the other, end. In a simple treatment based upon this principle, Bjerrum4 equated the, free-energy difference corresponding to RT In K\/4K2 to the classical electro¬, static energy necessary for complete separation of a positive and a negative, charge that are separated by a distance r in the molecule. This energy is in¬, versely proportional to D, the dielectric constant of the medium lying between, the charges, leading to Bjerrum’s relationship,, , RT In, , K,, , Ne2, , 4 K2, , Dr, , (1), , where e is the charge on the electron and N is Avogadro’s number. There is, some difficulty in selecting a suitable value for D\ for if the dielectric constant, for the pure solvent is used, the values of r calculated by equation (1) are over, 30 percent less than may be accommodated by our present picture of molecu¬, lar dimensions. A similar but more sophisticated treatment by Kirkwood and, Westheimer51-1 recognizes that a sizable portion of the space between the, centers of charge is occupied, not by molecules of solvent, but by the carbon, chain of the acid. Insofar as a “microscopic dielectric constant” can be assigne, to this carbon chain, it should be much nearer to the very low values (about 2, of the paraffin hydrocarbons than to the high value of the ionizing solven, (80 for water). Assuming (for mathematical simplicity) that a mo ecu e a, as a homogeneous ellipsoidal cavity of low dielectric constant surrounded by a, homogeneous region of high dielectric constant, it is possible to estimate an, “effective dielectric constant” having a value that depends somewhat upon h, assumed shape of the molecular cavity but that, in any event, , « ™, lower than the dielectric constant of the solvent. The Kirkwood-Westheim, treatment leads to more satisfactory values for r-the distance between the ends, of the molecules—but still must be regarded as a relative y cru e aPPr“, tion since the molecules involved are obviously not true ellipsoids., :, , % 506 5,31<1938,. « For further, , merits in treatment, see Tanford, J. An,. Chm. Soc, 79, 5348 [1957,., , refine-
Page 219 :
Inductive and Field Effects, , 203, , less expresses the important point that electrostatic effects are transmitted, through the carbon chain (largely by electron polarization, p. 59) as well as, through the solvent (largely by orientation polarization in the case of polar, solvents)., This treatment also helps to account for the striking increase in the ratio, K1/K2 as bulky alkyl groups are introduced between the —COOH groups in, dicarboxylic acids., , malonic, , ~ = 734, , K2, , diethylmalonic, , succinic, , tetramethylsuccinic, , glutaric, , /3,/S-di-n-propylglutaric, , 121,000, , 19.2, , 6130, , 11.9, , 4180, , Two effects appear to operate here. The first, and probably the most important,, concerns dielectric constant. In malonic, succinic, and glutaric acids, much of, the space between the —COOH groups on the ends of the molecules is occupied, by sections of solvent molecules. In the alkylated acids, however, most of the, space between —COOH groups is occupied by the alkyl groups on the chain., Thus, for the alkylated acids, the “low-dielectric-constant cavity” is fat, whereas, for the nonalkylated acids it is thin. The —COOH groups in the alkylated acids, interact through a medium of low dielectric constant, whereas those in the non¬, alkylated acids interact largely through a medium of high dielectric constant., Since electrostatic interaction is inversely proportional to the dielectric constant, of the medium between interacting species, the acid-weakening action of a, —COO- group on a —COOH group at the other end of the molecule is more, strongly transmitted in the alkylated acids. A second effect concerns steric, crowding. Ordinarily a flexible molecule of a dicarboxylic acid in a solvent of, high dielectric constant will tend to adopt the conformation (p. 73) in which, the —COOH groups are as far from each other as possible. However, as, crowding in the molecule is increased by the introduction of alkyl substituents,, t e, , COOH groups become pushed somewhat closer together and the inter-, , action between themjs increased. There is, in fact, good evidence that hydrogen, bondmg is of some importance in the monoanions of the alkylated diacids (p., , A number of workers refer to electrostatic action transmitted through, chains of atoms as inductive effects and electrostatic action transmitted either, , -rr r- -
Page 220 :
204, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , because of their greater measure of unscreened nuclear charge, better electron, attractors (that is, are more electronegative) than the carbon and hydrogen, atoms. When a more electronegative atom is substituted for hydrogen in a, molecule, electron density is pulled toward this atom and away from the nearby, atoms; and those atoms nearest the position of substitution will suffer the, greatest induced polarization. If the molecule in question is an acid, this sub¬, stitution will make it a stronger acid, just as would the introduction of a positive, charge., For example, the second column in Table 7-1 shows that the substitution, of chlorine or fluorine for an a-hydrogen atom in acetic acid strengthens the, acid, and the further substitutions of two and three chlorine atoms cause addi¬, tional increases in acidity. Similarly, glycolic acid, HO, which the electronegative oxygen atom of the, , CH2, , -COOH, in, , OH group has replaced an, , a-hydrogen of acetic acid, is almost ten times as strong an acid as the latter., The substitution of chlorine for a /3-hydrogen in propionic acid also increases, the acidity, but, as might be expected, the increment is much less than for, « substitution since the electron attractor is farther away from the site of reac¬, tion. Electron shifts associated with inductive effects are sometimes indicated, by arrowheads, attached to the bonds in structural formulas, and pointing in, the direction toward which the electron density is shifted. The inductive effects, of the chlorine atoms in dichloroacetic and /3-chloropropionic acids may thus, be represented,, , Cl, t, , Cl-, , -c<, I, H, , 0«-H, , /, , H, , H, , I, , I, , I, H, , I, H, , ci-*— c-*-cO, , 0«-H, , /, , O, , showing schematically how the polarization is passed from one atom to another, along the length of the chain. However, such representat.ons fail to show (a), that the inductive effect falls off with distance from the primary electron, luractor and (b) that a portion of the observed effect (just how much, we, cannot be sure) is transmitted, not through the chain but throughi the space, outside the molecule, , generally through molecules of solvent. This, , cn if the molecule is kinked or curled., , _N “ n. -"Sot ttrt, , cspec, , • sr:, -, , or triply bonded to a more positive atom that is, m tu.n, s.ngl>
Page 221 :
205, , Inductive and Field Effects, remainder of the molecule. Considering, for example, the acyl group, O, , C, , ,, , R, it will be recalled that one of the electrons associated with the carbonyl carbon, is a 7r electron (p., , 19). This is more easily polarized than are the strongly, , localized a electrons comprising ordinary single bonds. The carbonyl oxygen, can thus withdraw electron density much more effectively from its neighboring, carbon (leaving the latter with a partial positive charge) than can the singlebonded oxygen atoms in ethers and alcohols. We should therefore expect a, keto group to be more effective in boosting the strength of an acid than a simi¬, larly located hydroxide or alkoxide group. Analogous arguments may be ap¬, plied to the additional functional groups listed above.7 To express these argu¬, ments more briefly in the language of resonance, we may say that the character, of each of the functional groups involved is, to some extent, represented by the, “polar forms” listed below on the right:, , O', , O, , ++/, —N, , \, , o_, , o_, , -> —C=N, , — C=N <O, , o, , —c, , <-, , \, , \, OH(, , +/, -> —C, , \, , OR), , />, —c, , ^-, , \ \, H( R), , \, , OH(, , OR), , +/°~, , -cx, , H(VR), , No,e that in each of these “polar forms” the atom bearing a positive formal, charge is (in the absence of molecular coiling) closest to the remainder of the, molecule. Since we may imagine each of these atoms as a positive center, withdrawing electron density from its neighbors, the resonance concept furl, , dot:;;:?: VIVld PiCtUre °f ,he, , character of these func-, , Turning now to the electron-attracting abilities of alkyl groups, , we find, , .hem to be very nearly the same as that of the hydrogen atlimZm I last, , second -COOH group into a carbut also because of a statistical effect • th t •, ^ becfuse of lts electron-attracting properties, as likely ,o lose a pro,on i U ihe SUI^C
Page 222 :
206, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , column in Table 7-1 it. can be seen that substituting methyl groups for all three, a-hydrogens in acetic acid decreases the acidity, but merely by a factor of two;, alkyl substitution in other acids generally results in similarly small decreases, in acidity. If we push the site of substitution still nearer to the reaction center, by replacing the nonacidic hydrogen of H—C—OH with a methyl group, we, O, lower the dissociation constant by a factor of about ten. It appears then that, the methyl group (as well as other alkyl groups) is a slightly poorer electron, withdrawer than the hydrogen atom. This “acid-weakening” character of the, methyl group (in comparison with the hydrogen atom) has been recognized for, many years and a number of structural explanations have been suggested., None of these, in the opinion of the author, is fully convincing., In contrast to alkyl substitution, replacement of an a-hydrogen in acetic, acid by a vinyl or phenyl group increases the acidity (by a factor of about two, in each case). It appears then that an unsaturated carbon atom is a better, electron attractor (that is, more electronegative) than a saturated carbon atom, —a conclusion that is in accord with considerable additional evidence. What, appears to be the most satisfactory explanation of this difference9 is based upon, difference in bond hybridizations. Let us recall that the bonds to a saturated, carbon atom are “hybrid bonds,” formed from the combination of a single, , 5 and three p orbitals of the carbon atom (p. 18). On the other hand, a doublebonded carbon will have one of its p electrons participating in 7r-bond formation,, with only two of its p electrons available for hybridization to form «r bonds., The bonds around the unsaturated carbon atom are consequently called, , sP, , hybrids.” Now a spherically symmetrical 2, electron has the bulk of its charge, density closer to the nucleus than does a dumbbell-shaped 2p electron. Similarly, an ./hybrid bond should have the bulk of its charge density closer to the center, Of coordination than does an, , hybrid bond (since the former rs more hke a, , pure , bond than the latter). This means that a carbon atom partictpating, ,, 8 Some workers (see, , r, „„i_ Tncrnld Structure and Mechanism in Organic Chemistry,, for example, I g M ^ ^ add.weakening properties of the, , Cornell University Press, Ithaca, 1953, p., , ), , h, , the hydrogen atom—that is, that, , methyl group arise because this group 1S, ^:ob than the lone valence electron of hydrogen, its seven valence electrons do a slightly better j, the —COOH group. This, in satisfying the demand of the 'J^^^^hydrogen, seems almost a paraphrase of the statement t, 1, electron attractor than the methyl gro p,, , &, , a slightly more powerful, a subtle difference. Ir contrast, , atom is, , nresumably have an “intrinsic polarity,, , to the chloro, cyano, and, jf°hen under the influence of other polar groups within, alkyl groups exhibit polar effects, y, is generally determined by a study of a, ,he molecule. However, the polar effect of a, “ 8™r more different groups for elecrron, by '"so,d is, meaningful., ingful., », , Walsh,, , Discussions Faraday Sac.,, , ,, ifl U947)- Bartlett,, 2, 18 (1M/;, oar, , J. Chem. Ed.,, , 30, 29 (1953).
Page 223 :
Inductive and Field Effects, , 207, , sp» bonding (an unsaturated carbon) is pulling in its electrons more efficiently, than a carbon atom participating in sps bonding (a saturated carbon). By an, extension of this argument, a carbon atom bearing a triple bond should be, an even stronger electron attractor., By convention, groups which are more powerful electron attractors than, the hydrogen atom are said to exhibit negative inductive ( — I) effects, whereas, those which are poorer electron attractors than hydrogen display positive in¬, ductive (+/) effects. Table 7-2 summarizes the inductive effects of the more, usual substituent groups., , Table 7-2. Inductive Effects of Groups, — I Groups, -NH+, , -NRJ, —no2, —C=N, —COOH, —COOR, , —CHO, —c=o, , +/ Groups, —OR, -SH, —SR, —ch=ch2, —cr=cr2, —C=C—H, , 1, R, —F, —Cl, —Br, —OH, , —ch3, —ch2r, —chr2, -cr3, —c—o||, , o, , Although we have chosen to illustrate inductive effects of groups by com¬, paring dissociation constants of acids, it must be understood that such effects, influence virtually all heterolytic organic reactions. However, some care should, be exercised in ascribing an observed change in reactivity to an inductive effect, since other effects frequently become important. As we shall soon see, the reac¬, tivities of aromatic compounds are often influenced by conjugation, whereas the, reactivities of many aliphatic compounds are influenced also by steric factors., ( ndeed, there are a number of instances in which variations in reactivity are, ascribed by one group of workers to inductive effects but by another group of, workers, equally authoritative, to steric effects.) It appears, however, that in¬, duct,ve effects are relatively clear-cut in the reactions of ^-substituted ben¬, zene denvanves and those of relatively rigid alicyclic systems such as the denvatives of bicyclo(2.2.2)-octane (I).ff, , I
Page 224 :
208, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, In spite of the difficulties in disentangling inductive effects from other in¬, , fluences, a number of relatively safe generalizations may be made. An electronattracting group that boosts the acidity of an acid will, if substituted for a, hydrogen atom of an amine molecule, withdraw electrons from the basic nitro¬, gen, thus lowering its basicity. Likewise, such a substitution should reduce the, formation constants of addition compounds formed by the amine with metal, ions. In the extreme case, the fully fluorinated amine (CF3)3N shows no basic, properties at all., By a similar argument, electronegative groups that ease the departure of, H+ from acids will make more difficult the departure of the negative CN- ion, from the cyanohydrin anion in the equilibrium,, R—CH—C=N, , RCHO + CN-, , (2), , o_, thus stabilizing the cyanohydrin anion with respect to the aldehyde.*0, The same holds true with reaction rates, for substitution of electronattracting or electron-repelling groups can strongly influence the equilibrium, between reactants and activated complex.** A reaction between an electro¬, philic and a nucleophilic reagent will be accelerated by electron-attracting, substituents on the electrophile (since these would tend to make it more electro¬, philic) but slowed down by electron-attracting substituents on the nucleophile, (since these would tend to drain electron density away from the nucleophilic, center). Thus, carbonyl-addition reactions are almost invariably accelerated, when electronegative groups are substituted on the carbonyl compound, whereas, nucleophilic displacements by amines are generally retarded by substitution of, electronegative groups in the amine molecule. Substitution reactions of tertiary, halides are almost always accelerated by electron-repelling substituents (whic, should ease the departure of the negative halide ion in the rate-determining, ionizing step) and, conversely, are retarded by electron-attracting groups., ,, A, 1r, Chpm Soc 1928, 2533. The more usual way of describing, 10 Lapworth and Manske, J.Chem., h , molecule increase the positive, this effect is that electron-attracting groups on the aide y, nucleophile, character of the carbonyl carbon, thus, , or even water,, , does not have to be negatively charged, u may, " A number of workers (see, for example, I g, , >, , •, , cfej. Soc., 1933, 1120) prefer to, from those electronic shifts that, , differentiate inductive influences in reMtan sdemands Gf one reagent on another,, occur in the activated complex as a result o, inductomeric effects. We shall not make this, The former are termed inductive effects, the a, ., h ffccts together simply as inducdistinction in subsequent discussions, but will group all such ellects tog, tiVC f ?he same cannot, however, be said for, these generally proceed through a, atom at the substitution site is partially, , -"S t", !0°both the incoming and the leaving group,, depending upon other circumstances,, , For these reactions,, cause either an increase or decrease in rate. bee to, 77, 3886 (1955) and Ballinger, et al., J. Chem. S c.,, , a, 2641., ,, , J. sta. C*». **
Page 225 : Hydrogen Bonding and Acid Strength, , -, , 209, , In a complex reaction, the equilibrium constant lor a rapid and reversible, step preceding the rate-determining step generally becomes incorporated into, the rate expression for the overall reaction. In such a case, the reaction may, appear subject to inductive effects largely because of the influence of the, inductive effect on a preliminary equilibrium. For example, the substitution of, electron-attracting groups in the acyl portion /R—C—\ of esters speeds up, , \, , O /, , their saponification;75 the withdrawal of electron density from the carbonyl, group facilitates attack by the OH- ion, thus favoring the formation of anion, III:, , O, , OH, , O, , OH, , R—C—OEt ^, , R—C—OEt, , R—C—OH + OEt~, , (3), , III, , On the other hand, substitution of electronegative groups generally slows, down the acid-catalyzed hydrolyses of such carbonyl derivatives as imines,, oximes, and hydrazones since, in these cases, the preliminary equilibrium in¬, volves attachment of a positive H+ ion to the C=N bond., As we shall presently see, the most reliable estimate of inductive effects, may be obtained by comparing the rates of acid- and base-catalyzed hydrolyses, of substituted esters (p. 228)., , Hydrogen Bonding and Acid Strength, The hydrogen bond (p. 28) may be regarded as a special type of short-range, e ectrostat.e .nteraction which becomes important only when the positive center, ( he hydrogen atom and the negative center (an oxygen, nitrogen, or fluX, atom) he wtthtn 2 A of each other. Although it is reasonable to, , uspectTa, , such tnteractton might affect the reactivities of, say, the nitro group in e nimo, , parkotTf, , H, , ^ " "'Mroxyacetophenone, we must turn to the com-, , stants of some substituted benzoic acids'* follows., , 69, 279S2“^°36)Xan,Ple' Kind"r', , *•* “, dlsso«at,on con-, , 45°' 1 «««>* «*. 90 (1927); 464, 278 (.928);, , at 25" Cf"' ValUCS’ <aken fr°m Dippy' a,m-, , 25, 151 (1939), refer, , to aqueous solutions
Page 226 :
210, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, COOH, , COOH, , COOH, , X)CH, , KX, , 105 6.3, , 8.1, , COOH, , COOH, , 3.4, COOH, , OH, , 2.9, , iCXlO5 105, , 5000, , Note that the substitution of an —OH group para to the, , COOH group, , in benzoic acid lowers the acidity by a factor of two, but that the substitution of, an ortho hydroxy group raises the acidity sixteenfold. Furthermore, the substi¬, tution of two ortho hydroxy groups raises the acidity almost a thousandfold., Now it is well known that the substitution of almost any neutral group ortho to, the _COOH group in benzoic acid causes an increase in acidity (p. 236),, but for no other group are the effects of ortho substitution and para substitution, so different as for the hydroxy group. Substitution of an, , -CH3O— group, , 0, , (which is a bit larger than a hydroxy group but otherwise very similar) results, in less than a 30 percent increase in acidity. The difference is obviously that, the methoxy group, unlike the hydroxy group, has no acidic hydrogens to par¬, ticipate in hydrogen bonding. In the hydrogen-bonded form of salicylic acid, (IV), the positive (hydrogen) end of the phenolic hydroxyl group lies very near, he —COOH group, pulling electron density away from the carboxyl hydrogen, 1mm and easing its departure. The hydrogen bonding occurs also in the anion, H, , —H+, , IV, , V, , V, the negative charge being spread over the six-m
Page 227 :
Hydrogen Bonding and Acid Strength, , -, , 211, , that includes the hydrogen bridge, rather than being localized merely to the, carboxylate group. In 2,6-dihydroxybenzoic acid (VI) both phenolic —OH, groups participate in hydrogen bonding; in the anion of this acid the negative, charge is spread over 'two chelate rings. Note that in these cases the hydrogen, bonding persists in the carboxylate anibn as well as in the acid itself. In con¬, trast, although intramolecular hydrogen bonding occurs in o-nitrophenol (as, m, , evidenced from its high volatility, p. 30), hydrogen bonding cannot occur in, the o-nitrophenolate anion, and the acidity of this phenol is nearly the same as, that of its para isomer.75, In a number of dicarboxylic acids such as maleic acid, VII, the —COOH, u, , o, , I, H, i, i, i, O, , H-C, H-C., 'C, I, O\, , - H+, , -ol, , H-C, , H, , H-C., ^c, II, , o, , H, , VII, , VIII, , groups are so situated that intramolecular hydrogen bonding may take place, between the hydrogen atom'of one -COOH group and the oxygen atom of the, seeond. The first ionization constants of such acids, like the ionization constant, of salicylic acid, should be “abnormally high”; bu, since the remaining acidic, hydrogen atom of the mono anion (for example, anion VIII) is incorporated, into a negatively charged cyclic system, the seeond ionization constant of such, an acid should be “abnormally low.” Thus, intramolecular hydrogen bonding, , zzzsz ssjrsr-, , -f*—, , its monoester should be very similar except that h -d ^ ', 'ir, d‘aC‘d 3nd, in the anion derived from the latter) Us', ,J r°fn bondlng cannot occur, , ahown that for those dibas” acids rih h?n, factor by which such bonding raises the valued, , S Crlteri°n’ “ can further, lmPortant, the, , ™“ation,, y g eater ,han um'y for diethylmalonic, , W**" This ratio is substanti^y gr a ° fh‘a‘°, , „ Rr,„ .. K „, ., , “Wes, , diadd-i “m1 ^ *“ £ diacidC to 'acfjictr56*' T”' T “ 'h' d'-mi„ato,, •he absence ofa„ other effec,s-w„uld bejufe
Page 228 :
212, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , (16.0), tetramethylsuccinic (13.5), and maleic (5.3) acids but very nearly unity, for malonic (1.8), succinic (1.06), and fumaric (1.0)* acids.'7 We can appreciate, that the trans configuration of fumaric acid does not allow the close approach, of the —COOH groups that is necessary for intramolecular hydrogen bonding, (as, for example, in structure VII). Such a close approach appears to be, relatively rare in aliphatic dicarboxylic acids, except, presumably, in very, “crowded” acids such as diethylmalonic, tetramethylsuccinic, and /3,/3-di-/zpropylglutaric acids. In these cases, the formation of hydrogen-bonded struc¬, tures is apparently favored because in this way the —COOH groups tend to be, “pulled out of the way” of the bulky alkyl groups, increasing the freedom of, motion (hence the entropy) in the remainder of the molecule., , Conjugation Effects, Important as the inductive effect is, it often fails badly if used to predict differ¬, ences in reactivity among aromatic compounds. Considering only the inductive, effect, for example, we might anticipate />-hydroxybenzoic acid to be a stronger, acid than benzoic, whereas the reverse is found to be true. We might predict,, using the inductive effect, that ra-nitrophenol is a stronger acid than the para, isomer, but, again, the opposite is true. Similarly, the inductive effect, by itself,, cannot account for the very great difference in the basicities of aniline and, benzylamine (benzylamine is the stronger base by a factor of 50,000)., It may be recalled, however, from the discussion of dipole moments (Chap., , 3) that the transmission of electrical effects along the chain of u bonds compris¬, ing an aliphatic molecule is very much different in character from the transmis¬, sion of effects involving tt electrons along the tt bonds that comprise a conjugated, system. In the first case, the inductive effect, a relay of electron polarization, occurs but decreases sharply as one moves along the aliphatic chain away from, the primary pole. In contrast, a disturbance of 7r-electron density at one atom, in a conjugated system may become distributed over the “x-electron streamer”, (p, , 20) associated with the entire system, with those atoms far from the souice, , of'the disturbance being just as much affected as are the atoms close to it, This mode of transmission may be termed a resonance or conjugation effect, since it is readily described using the language of resonance. For example, two, resonance structures for nitrobenzene (IX and IX') indicate that the substitu-, , ester. The above authors present a more, the hydrogen bonding in the monoanion oft e, , i, , ^, , aTso^hefidativelylmall) differences, ^, thp diacid and that of the, , , ,1“—, , n}, , forra of ,he diacid and tha‘ °''he, , 3, , respective ™o„oc,M esters.
Page 229 :
Conjugation Effects, , -, , 213, , tion of a nitro group on a benzene ring results in withdrawal of 7r-electron density, from the ring, especially from the ortho and para positions. In the same way,, structures X and X' for the phenoxide ion indicate that the negative charge is, not localized on the oxygen atom but also resides partially on the ring, again, at the ortho and para positions., , The resonance effects of the, represented, r,, a, ., N°2 and, substituents may also be, esented hv, by cu, d P , is shifted., i r;rWd arrOWS’ POiminS “ward the direction in which .-electron, density, , “shifts”.) B°«hntyhpe!, , from the indicate, , resonance effects; they are transmitted ,, , lmP°«ant characteristic <, , system. Resonance forms analogous to Jx'lX' x', vr T “ ^ COnjUSa“, °r native charges in the meta nositin ’ " ’ X’ °r X > whlch put the positiv, , -e unsatisfactory fa ure t L^r01 *, "A Centre such as X, 1 h, ’, , ,C, , ^, , Whh°Ut -'udin, , Whik the difficult y i,, the negative charge of the phenoxide on the met, , o
Page 230 :
214, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , drawing such forms is consistent with the “alternating character” of the atoms, forming a conjugated system, it cannot be considered an “explanation.” (In¬, deed a similar alternating character may be predicted without consideration, of resonance forms as such by using a different approach to the theoretical, treatment of conjugated systems—the so-called molecular orbital method.i9(o), A satisfactory explanation of this aspect of conjugation has not been given in, nonmathematical language, although there have been several attempts.19(b), Before considering the transmission of effects between substituents through, conjugated systems, we should recall that the reactivities of a number of sub¬, stituents are markedly altered when they themselves are attached to unsatur¬, ated carbon atoms. The —OH groups in phenols, for example, are more acidic, than the —OH groups in alcohols. The direction of this effect is not surprising,, for the phenyl group is known to be a slightly stronger electron attractor than, the ordinary alkyl groups (p. 206). However, the magnitude of this effect seems, far greater than would be expected on the basis of the inductive effect alone, (for example, pKa for phenol is about 10, pKa for methanol is about 18), and, it is currently felt that the acidity of phenol must also be boosted by a resonance, effect. As structures X and X' indicate, the negative charge on the phenolate, ion is distributed over the benzene ring, whereas the negative charge on an, alkoxide ion is localized at the oxygen atom. Since the phenolate ion is thus, stabilized in a manner not possible for alkoxide ions, we would expect the loss, of a hydrogen ion from phenol to occur more readily than the loss of a hydrogen, ion from an alcohol. One may well wonder how much of the difference in, acidity between phenols and alcohols is due to the inductive effect of the ben¬, zene ring and how much is due to the “resonance stabilization” of the phenoxide, position, is open to some objection because it depicts a bond between two nonadjacent carbons, fn a planar ring. The distance separating these two atoms is about 2.4 A, far greater than t, length of normal carbon-carbon bonds, and any “trans-annular” bonding action of this, would be expected to be extremely weak., m„mherpd rinir, We might also note that if the conjugated system is a five- or seven-membered, g, , XIII, , Wi“ WtteXTerW, , XIV, , £S£ "£ E, , 238-258; and KatellatChemical Cons^or^ f, f59; Dickens and Linnett, Quart. Revs.,, 273-288; (b) Longuet-Higgins, Proc. Chem. tsoc., 1^/,, , XI, 310 (1957).
Page 231 :
Conjugation Effects, , -, , 215, , anion. In this case there is at present no experimental method for making a, clear-cut separation of the two effects., Similarly, the basicity of aniline falls below the basicities of the ordinary, aliphatic amines by a factor of about 106. Here, the electron-attracting power, of the phenyl group would be expected to lower the basicity of the aromatic, amine, but, again, the effect is far too large to be attributed merely to an, inductive effect. The unshared electrons responsible for the basicity of aromatic, amines comprise part of the 7r-electron system that also includes the ir electrons, associated with the aromatic carbon atoms. Electron density, which, in ali¬, phatic amines, is localized on the nitrogen atom, becomes partially drained off, into the regions above and below the plane of the benzene ring in aniline. As a, result, the nitrogen atom in aniline assumes a partial positive charge (forms, XV and XV') and its basicity is greatly diminished. Saying this a little differ¬, , ently, the anilmium ion (pKa 4.58) is a stronger acid than the ammonium ions, derived from aliphatic amines (pKa values lying between 10 and 11) because of, delocalization of negative charge in the basic form in a manner prohibited for, the acidic form. This argument corresponds closely to that which we have, used to account for the acidity of phenois, but for anilinium ions we can say, with more confidence how much of the enhanced acidity is due to the inductive, effect and how much to the resonance effect. The pK. values of the conjugate, fXVim ^, (XVI)’ benzo<)uinuclidin<i (XVII), and quinuclidine, V111) are given below:*0, CH3, , CH,
Page 232 :
216, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , tropy effects), the inductive effect accounts for approximately 3 of the 5.5 pK, units by which the basicities of aliphatic amines differ from those of aromatic, amines., Perhaps the most striking difference between alkyl and aryl halides is the, extreme slowness with which the aryl derivatives (except for nitro compounds), undergo nucleophilic substitution reactions under conditions where most alkyl, halides react readily. Chlorobenzene, for example, survives for days in basic, solutions at room temperature although thermodynamics predicts its hydrolysis., Since this is a kinetic, rather than a thermodynamic effect, it must be explained, on the basis of the energy gap that separates the reactants (C6H5C1 + OH-), from the activated complex. In chlorobenzene, the lobes of two of the unshared, electrons on the chlorine atom lie perpendicular to the benzene ring; hence,, these two electrons have become part of the 7r-electron system lying above and, below the plane of the molecule, and the C—Cl bond (like the C—N bond in, aniline and the C—O bond in phenol) assumes some double-bond character, tl Similar arguments are sometimes invoked to explain why carboxylic acids are stronger, acids than alcohols and why amides are weaker bases than amines. Carboxylate anions are, stabilized by the distribution of negative charge over two oxygen atoms in the —GOO, , o-, , —c, , /, , \>, , group, , O, , /, , whereas a neutral amide molecule (but not its conjugate acid) is, , «-> —c, , oJ, o-, , o, stabilized by a similar delocalization, , r, , /, , «-+— c, , On the other hand, we might, , \+, N—I, , N—, , . t.t th_ differences between carboxylic acids and alcohols (and between amides and, amS are due largely to the strong inductive effect of the carbonyl group. It is pertinent, however^, , hat^limethyldxhydroresorcinol (“dimedon,” XIX) andmethylketene dimer, , XV, , -both of them enols with a" hydroxy group conjugated with a carbonyl group-show acuhtxes, , Me, , O, , \_f, , (.pKa, , (.pKa 2.8), , 5.2), , XIX, , /, , _, , HO, , XX, , \, Me, , comparable ,o those of carboxylic acids. XIX°n' ““e “is^Zg^dward a™3, Schilling, Am., 308, 193 (1900 ),.where-i,5® “ a^*k““ £ ,hL compound, are, Small, J. Am. Chem. Soc., 72, 1(, sufficiently far from the hydroxyl groups, , ■, , the inductive effects may be considered, hese cases are due largely to conjugation, , relatively small, it seems that the high acidi, respective anions by distribution, effects-ffiat is, stabilization of the, are of, over conjugated systems. By inference, then, it is reasonao, comparable importance in carboxylic acids and amides.
Page 233 : Conjugation Effects, , 217, , (XXI *-> XXIV* To convert the chlorobenzene molecule to the activated, complex for basic hydrolysis (XXII ^ XXII') requires a substantial expendi¬, ture of energy, for in such a process a 7r-electron system embracing seven atoms, is cut down to one involving only five. The difficulty with which nucleophilic, , XXI, , XXI', , XXII, , XXII', , substitution reactions on vinyl halides proceed may be similarly explained., It thus appears that the —NH2 and —Cl substituents, although capable of, withdrawing electron density from saturated carbon chains by induction, are, , capable also of supplying 7r-electron density to conjugated systems. The same, is true for the substituents —OH, —OR, —O—C—R, —F, and —Br; when, , O, any of these substituents is put on a conjugated system, resonance forms anal¬, ogous to XXI and XXI' may be drawn for the resulting compound. On the, other hand, the —NO,, —C=N, —COOH, —COOR, —C=0, —C=0, , substituents, each of which is associated with a negative inductive effect, also, withdraw 7r-electron density from conjugated systems. When any of ’these, groups is affixed to a conjugated system, structures analogous to IX and IX', may be drawn for the resulting compound., Let us then (in analogy with the classification of inductive effects) designate, nose groups as +R in character that, by resonance effects, supply electron, density to conjugated systems, and those groups as -R that withdraw electron, density from such systems. Resonance effects (R effects) of the more usual, groups are summarized in Table 7-3.25, , CrCI, , chlorobenzene is consent, , l“? P'T’ ^ “V «"”■ **■. 59,, , 2)81 (1937)), as comparedTillX'noLal”TST, <P, , If" XXr alS° h'lp *, , ^ relatively W dipolf, , 0^06^, , Particularly ‘hOSe °f *he “En*lish, , petr^
Page 234 :
218, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, Table 7-3. Resonance Effects of Groups, -{-R, —I Groups, , — R, —I Groups, , —F, , -\-R, +/ Groups, , o-, , -NO,, —C=N, , —Cl, —Br, —I, —OH, , —CHO, , —, —s—ch3, , —c=o, , —cr3, , 1, R, —COOH, , —OR, —O—C—R, , —COOR, —conh2, , II, O, , O, , —SH, —SR, —NHo, , 1, —S—R, |, , —nr2, —NH—C—R, , O, —cf3, , O, , Note that each of the +R groups (except, , CH3 and, , GR3) is attached, , to the remainder of the molecule by an atom having one or more unshared pairs of, electrons. The —CH3 and —CR3 groups exhibit very slight +R effects, pre¬, sumably as a result of hyperconjugation (XXIII ^ XXIII'). In each of the, , O, -R groups ( except —S—R and —CF3 j, the atom closest to the remainder, , O, of the molecule is also multiple bonded to a more electronegative element (for, , example, —N, , O,, , C, , =0, and — C=N). The -R character of the —S—R^, , O, and -CF,« groups has, , XXIV, , (Inwhieh the'sulfur atom has ten rather than eight valence electrons), XX V (in which t, the electron.attracting effect of the sulfone, group!whereas0form XXV (again a sort of “hyperconjugated" form) has been, suggested to explain the -R effect of the -CF, group., &&, , „, , into, , flOW), , .. Bordwell and Cooper -/. dm CW f.-^ ^o)’^, ts Roberts, Webb, and McElhUl, tbid., 72, 4U8
Page 235 :
Conjugation Effects, , 219, , Generally, the best indications of the nature of the resonance effect of a, substituent arise from the study of reactions in which that substituent is para, to the reaction center in an aromatic system. Inductive effects are also present, in such cases, but these are often relatively small since the substituent and the, center of reaction are four carbons distant. When the substituent and the reac¬, tion center lie meta to each other, inductive effects are more evident. Here, due, to the alternating nature of the atoms in a conjugated system, resonance effects, become less, although they do not, as might be supposed, entirely disappear., (When the substituent is ortho to the reaction center, both inductive and reso¬, nance effects may operate strongly, but superimposed on these are proximity, or steric effects that further complicate the interpretation of data.) We can now, appreciate why />-hydroxybenzoic acid (pK 4.58)—in which the —COOH, group feels the + /? effect of the —OH group more strongly than its —I effect, -is a weaker acid than benzoic acid (pK 4.20), whereas m-hydroxybenzoic, acid (pK 4.08)—in which the importance of the two effects is reversed—is a, slightly stronger acid than benzoic. Similarly, jft-nitrobenzoic acid (pH 3.43) is, a slightly stronger acid than its meta isomer (pK 3.45) since the -R effect of, 1 ', , ffN°? f °UP is m°re imP^tant in the para compound (even though the, e ect is less important). In ^-fluorobenzoic acid (pK 4.14) the +R and the, , acid744ot K, In general I, , S-°UP near‘y CanCd, °f ** +*, , °ther’ but i^-chlorobenzoic, <* the chloro group is much less., , (4 ToHTnh”i lSrmrPOrtant f0T SUbSti,UmtS, (—Cl _Br ’and, , Sm TK, , ^T"'5, , roughly the’same size as tie' °T T, , effective. (There are L fact, , 7, , the heavier elements, ^ ^, , ■ 7*™*, °f the, , atoms, but the ou r (3h aSZ, erably ,arge and 0 b 4nd P), , first-row events, , ", , ekments ««, carbon, elements are consid-, , f°r bondi"S, , b less, , having true multiple bonds invoking'eTemems^Iher!‘ CXampleS °f ““Pounds, nitrogen.), thc tllan carbon, oxygen, and, Since the />-CH3 group lowers the acidities of benzoic, , acid, phenol, and
Page 236 :
220, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , the anilinium ion appreciably more than does the W-CH3 group, we may con¬, clude that the electron-repelling action of this group operates through con¬, jugation effects as well as through induction, this being perhaps the chief, chemical justification for writing structures such as XXIII and XXIII'., As with inductive effects, our preliminary discussion of resonance effects, has been based upon the influence of substituents on acidities—mainly because, of the large quantity of available data. It should be re-emphasized, however,, that the resonance effects of substituents appear to influence practically all, types of reactions in which the reaction center is attached to an aromatic sys¬, tem. This is true both for equilibria and reaction rates. Furthermore, if a reac¬, tion is aided by electron withdrawal due to induction, it should be aided also, by electron withdrawal due to resonance., , The Hammett Equation26, Thus far, our considerations have been mainly qualitative. At present, however,, there exist a number of quantitative (or almost quantitative) relationships, between structure and reactivity. One of the oldest and most familiar of these is, the Hammett equation, which relates structure to both equilibrium constants and, rate constants for the reactions of meta- and jf?ara-substituted benzene deriva¬, tives/7 Consider a series of aromatic compounds, each with the same reaction, center present as a side chain, but each having, in addition, a different sub¬, stituent situated meta or para to that reaction center—for example, a group of, substituted benzyl chlorides. The Hammett relationship stipulates that the rate, or equilibrium constant associated with the reaction of any one of these com¬, pounds (say the specific rate of hydrolysis of /?-nitrobenzyl chloride) may be, determined from the corresponding constant from the “parent compound, (benzyl chloride itself) if two parameters are known. The first parameter (a), is characteristic only of the substituent (in this case the p-nitro group) and, represents the ability of the group to attract or repel electrons by a combina¬, tion of its / and R effects. The second parameter (p), characteristic of the reac¬, tion series at hand (in this case the hydrolyses of substituted benzyl chlorides),, is a measure of the sensitivity of this type of reaction series to ring substitution., If k and k0 are the rate constants for reaction of the substituted and unsubstituted compound, respectively, the Hammett equation may be written:, * (.) Hammett, Physical Organic Chemistry McGraw-Hill Book Co., Inc, New York, 1940,, , ^cyclic^Re^T) ^^d^heterocydic^(Elderfield and Siegel, ibid., 73, 5622 (1951)) derivatives.
Page 237 :
The Hammett Equation, , -, , 221, , l°g (g) =, , (4), , [If equilibrium, rather than rate, constants are being considered, the term on, the left is log (K/Ko).] Now the, , <7, , value for a substituent may be obtained most, , directly by measuring the effect of that substituent on the ionization constant, of benzoic acid in water at 25°. cr is defined, , o- = log, , (5), \, , nC8H6COOH, , /, , where /fx-c6H4-cooH and ^c6h6cooh are the ionization constants for the substi¬, tuted and unsubstituted benzoic acids, respectively. In effect, the ionization of, benzoic acid has been arbitrarily chosen as a standard reaction type (for which, P, , is fixed at unity), and a is defined on the basis of this standard. A positive, , o value for a substituent indicates that the substituent is a stronger electron, , attractor than hydrogen; substituents with negative a values are weaker electron, attractors than hydrogen. To obtain p for a given reaction series, it is necessary, to measure rate or equilibrium constants for a number of compounds, each, having the reaction center under consideration and each having a different, ring substituent with a known a value. The logarithms of the measured con¬, stants are plotted against the corresponding a values, and the slope of the best, straight line through the points on such a plot is the p value for the reaction, series at hand (Ex., , 6)., , Reactions with positive p values are aided by electron with¬, , drawal from the benzene ring, whereas those with negative p values are made, more difficult by electron withdrawal. The, , cr, , values for twenty of the more familiar, , substituents are listed in Table 7-4, and the p values for some typical reaction, senes are given in Table 7-5., , Table 7-4. Substituent Constants, Group, , ~ch3, , &m, , -0.07, , (Tp, , Group, , -0.17 —F, , &m, , TO. 34, , (d, , (T p, , values)*ff(6), Group, , TO. 06 —NH—C—Ph, , <rm, , TO. 22, , TO. 08, , TO. 36, TO.10, , TO. 27, TO.13, , II, -0.15 —Cl, TO.01 —Br, , TO. 37, TO. 39, , O, TO. 23 —COOH, TO. 23 —c—o-, , TO. 42, 0.00, , TO. 55 —I, -0.46 —no2, , TO. 35, TO.71, , TO. 28 —CHO, TO. 78 —c—ch3, , TO. 38, TO.31, , TO. 22, TO. 52, , —o-, , —0.71, , —OCHs, , TO. 12, , -0.52 -nh2, -0.27 —N(CH3)3+, , -0.16, , o, -0.66 —CN, TO. 86, , TO. 68, , TO.63, , 1, , 1, , -0.04, TO. 06, , OQ, , —c2h6, -c6h5, , <rp, , TO.91, , o
Page 238 :
222, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, Table 7-5. Reaction Constants (p-values)?<?(6), Reaction, , P, , 1. Ionization of benzoic acids, water, 25° (eq), 2. Ionization of benzoic acids, ethanol, 25° (eq), 3. Ionization of phenols, water, 25° (eq), 4. Acidities of anilinium ions, water, 25° (eq), 5. Addition of HCN to benzaldehydes, ethanol, 20° (eq), 6. Ionization of triarylmethyl chlorides, SO?, 0° (eq), 7. Saponification of methyl benzoates, 60 percent acetone, 0° (rate), 8. Acid hydrolysis of ethyl benzoates, 60 percent EtOH, 100° (rate), 9. Hydrolyses of benzyl chlorides, 50 percent acetone, 60° (rate), , 1.000, 1.957, 2.113, 2.767, 1.492, -3.974, 2.460, 0.144, , 1.688, , -, , 10. Benzoylation of aromatic amines, benzene, 25° (rate), 11. cis-trans isomerization of substituent azobenzenes, C6H6, 25°, , -2.781, -0.610, , (rate), Reduction of nitrobenzenes with SnCl2, water, 90° (rate), Addition of HCN to benzaldehydes, ethanol, 20° (rate), Side-chain bromination of acetophenones, water, 25° (rate), Decomposition of substituted benzoyl peroxides, acetophenone,, , 1.149, 2.329, 0.417, 0.374, , 12., 13., 14., 15., , 80° (rate), At the present time, substituent constants for over 110 different substituents, have been calculated, and reaction constants for almost 400 different reaction, series are known.** This means that about 44,000 rate and equilibrium con¬, stants of meta- and /^-substituted benzene derivatives may in principle be, calculated from only 510 parameters. Over 3000 of these constants have been, measured and have been found to agree with the values predicted by the, Hammett equation with a probable error of 15 percent (only 0.06 logarithmic, units)., . ., , ., Two questions come to mind. First, why should the Hammett relationship, hold at all? Second, why is it that this relationship does not also encompass the, reactions of orrte-substituted benzene derivatives and those of aliphatic com¬, pounds as well? (As we shall see, similar, but less general, relationships may, hold for these.) It will be recalled that log K„ for a reaction is proportional to, the standard free-energy change, AF°, and that log kMtor a reaction is pro, portional (from the arguments of the transition-state theory P; 1791 tothe, free energy of activation, A FT If we are considering the equilibrium consta, associated with a given reaction series, we may rewrite the Hammett equation, '• In addition, the Hammett equation has, , fTJrda“dC°r^“^io“C^frequencies, and, , half-wave potentials from polarographic reduc 1, ’associated with the fluorine nuclei in, nuclear magnetic resonance-absorption frequen, D 214, Substituted fluorobenzenes. These are summarized by Jaffe, Ref. 26b, p. 214.
Page 239 :
The Hammett Equation, , -, , 223, , as, , log K = pa + log Ko, , (6), , or, in terms of free-energy changes, -AF, , =, , RTp)a - AF0, , (, , (7), , (The AF values refer to standard free-energy changes, but the superscript zeros, have been omitted for simplicity.) Now, for a given reaction series at a given, temperature, T, p, and AF0 are constants; equation (7) is therefore of the form, y = ax + b. The free-energy changes associated with the reactions of the mem¬, , bers of a series are thus linearly related to the respective a values (and, from the, definition of a, linearly related to the standard free energies of ionization of the, correspondingly substituted benzoic acids). If we are dealing with reaction, rates, an analogous equation may be derived, this time expressing a linear rela¬, tionship between the values of AF* and the a values., In meta- and ^^-substituted benzene derivatives, the substituents are rela¬, tively rigid and lie far enough from the reaction center for us to feel safe in, assuming that steric interaction between the substituent and reaction center is, negligible. It appears that in most of such cases, substituents may affect reac¬, tivity, directly or indirectly, solely by their ability to withdraw electrons from, or, supply electrons to, the reaction site. It is tempting, and in many cases it is correct, , to assume that the change in AF° (or in AF*) resulting from the introduction, of a substituent merely represents an increase or decrease in the classical elec¬, trostatic energy gap that separates the reactants from the products (or from the, activated complex, if we are considering rate constants). Such an increase or, ™e 15 Presumably directly proportional to the electron-attracting or, electron-repelling power of the substituent, as represented by its „ value We, ZTJr? h0WT’ lhat 3 SUbstitUent’ due '° ^ electron-attracting, , Ity, may also introduce kinetic energy effects; it may affect the vitror of, , degree1of Z T" Tu r°tali°nS, ** ^ ^, a'<er the mode and, g, solvation of the species involved in the reaction. In order that the, effects1 be neT'hl ^ aPpUcable> “ is "'^ssary either that these kinetic energy, It sh, , 1H, , lgl ’ C,°r that they be pr°P°r'ional to potential-energy effects «, , no,genXpnp°Lt,eC :a;ry ^ "7™" «>'“*»’ “, Cb, Hem the subs i u, s e ,, °f ‘^bstituted benzene derivatives., , *- -* r, , srxsr i
Page 240 :
224, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , that a linear relationship between free energies and substituents apply to these, reactions, two conditions must hold. First, each ortho substituent must have, what may be termed a “steric interaction capacity” which, like its electronattracting ability, is independent of the reaction at hand (this is true in some, cases).50 Second, the sensitivity of reactions to steric effects must quantitatively, parallel their sensitivity to electron withdrawal or supply (generally this is not, true)., Similarly, it would be anticipated that the reactions of aliphatic com¬, pounds would not follow the Hammett equation, for even if the substituent is, several carbons removed from the reaction site, the bending and twisting of the, aliphatic chain might allow appreciable steric interaction between groups. On, the other hand, it might be suspected that the reactions of relatively rigid, alicyclic compounds may be correlated by a linear free-energy relationship, similar to the Hammett equation. This has been found to be the case for reac¬, tions, , of derivatives of 4-substituted bicyclo[2.2.2]octane-1-carboxylic acid, , (XXVI).6 For such compounds a new set of a values must be used since there is, , XXVI, no conjugation in the ring system, and all electron withdrawal or supply by the, substituent must therefore occur by the inductive (or a combination of the, inductive and field) effects. Moreover, the reactions of a number of aliphatic, compounds that lack the rigidity of the bicyclo[2.2.2]octane system also follow, linear free-energy relationships similar to the Hammett equation For such, compounds, the steric interactions present in the reactant molecules may, , e, , considerable, but such interactions also persist in the products and » the, activated complex. Thus, the free energies of reaction or of activation, which, rely wholly on fences, are not significantly affected. We see then that steric, interaction is not a sufficient condition for failure of Hammett-like relationships, there must be sizable variation of such interaction during the progress, “ould be well at this point to make a distinction between the two ways, in which the resonance effect of a substituent may influence reactnuty m^e, , r:i i., , J0Taft, J. Am. Chem. Soc., 75, 4538 (1953).
Page 241 : The Hammett Equation, , -, , 225, , acidic —1OH group is in direct conjugation with the nitro group.(XXVI11)., , °x, , N, , <p, , a, , r, , \J, , L\\, , H, XXVII, , XXVIII, , In both cases the nitro group is acid strengthening, but, as might be expected,, it is more effective in the second. The />-nitro group boosts the acidity of benzoic, acid by a factor of 6 and should, according to the Hammett treatment, increase, the acidity of phenol by a factor of about 36 (since p for this reaction is about 2)., However ^-nitrophenol is almost 600 times as strong an acid as phenol. Like¬, wise the />-nitroanilinium ion is 4000 times as strong an acid as the anilinium, ion although the simple Hammett treatment would predict it to be only about, 125 times as strong; again, the —NOz group is in direct conjugation with the, NH2 group in the basic form. Similar departures are generally observed for, rates and equilibria of reactions involving derivatives of anilines or phenols and, having, at the same time, -R groups (Table 7-3) para to the reaction centers., owever, a rather good fit to the Hammett equation is achieved by assigning to, each of these -R groups a second substituent constant, which may be designated, 'h,e suPerscriPt c designating direct conjugation. These at values (which, XitVd Td’ f°r eTmPle’ fr°m the addities of ,he appropriate parabst tuted ami,mum tons) are greater than the normal a values and are used, in Table'?"’'3''10"8 ^ Phe"°1S’ aniUneS’ ^ ,hdr derivatives. They are listed, , each'S Xpln .“pS, , sh°“ld b= -ociated, , with, , carbonyl compounds (where there is direct conhitrar1^ k l° reactlons of esters, acids, and, center), the second to reactions ofphenZsaXTJXh’, jugation ,s absent). The substituent constants nr,?, n, ^, repellmg groups apply to the first class of reart', , ,!, , 7" Subs,i>f« and reaction, ?rivatlves (where direct con-, , enved by Hammett for electron-, , «° reactions of anil" I and pLlu" ou d^H^ “ 'tT L, *, °f'* «*“». aPPhcable, tion for the latter class (see for examnln t?, somewhat better fit to the Hammett equamoment constants, m„rru3a^,«Z^Vc, errors; introduced by adopting a single value for tJP, spread adoption of a double scale. S, H, , ’ 579' 87 <*»»)>• Howc^Z, g gr°UpS tend t0 be small> a"d the, n0t Senous enough to warrant wide-, , choo1- parameters) ^^y^be ob^ained'by^assmnin^'a? ^r’ h<TVer’ the introduction ofaddi-, , tHC rangC ^^t^ee^'mc^st^appropriaprfo*%**** Substit-nts and, , PP. 576-58326<b>' PP- 228-22!»- For a third approach
Page 242 :
226, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , Table 7-6. Substituent Constants for Electron-attracting Groups to, be Used for Reactions of Anilines, Phenols, and, Their Derivatives~6(b), Group, —no2, , —COOH, —CHO, , 1.27, , Group, , *CP, , -C—ch3, II, , 0.87, , O, —CN, , 0.73, 1.13, , 1.00, , Recent work suggests that the Hammett equation may also apply to the, specific rates of electrophilic substitution reactions on the benzene ring itself, (aromatic chlorination, nitration, mercuration, etc.).s* It appears that the, specific rates of m^a-substitution reactions may be correlated rather well using, the ordinary <rm values, but an additional set of substituent constants (designated, ff+ values, since the attacking reagent is generally a cation) must be used for, para substitution. Conveniently (but as yet inexplicably), the ap values may be, , estimated simply by subtracting 0.13 from the corresponding <rp values (or, from the cc values when these are available). Since the reaction site is on the, benzene ring itself, rather than on a side chain, these substitution reactions are, very sensitive to the presence of electron-attracting or electron-repelling groups;, that is, the p values are unusually large (for example, p for aromatic chlorination, in acetic acid is —10.1, whereas p for aromatic nitration in acetic anhydride is, — 6.5). This treatment is particularly useful for estimating the ratio of meta, to para substitution in a given reaction; but in making such estimates it must, be remembered that meta substitution, aside from polar effects, has a 2 to 1, statistical advantage over para substitution since there are two positions on a, ring meta to a given substituent but only one position para to it., The Br^nsted catalysis law (p. 113),, log *eat., , =, , a, , l0g, , K*, , +, , b, , ^, , which expresses the relationship between the acidity constant A', for an acid and, its catalytic constant, , for a given acid-catalyzed react.on, ts also a hnea, , free-energy relationship. On the basis of the transition state theory, equat.on (8), may be rewritten:, = a\F> + [ RT (log ^ - »), , (9), , the expression in brackets being a constant at a given temperature. This is, i» McGary, Okamoto, and Brown, J. Am. Chem. Soc., 77, 3037 (1955).
Page 243 :
Polar Effects in Aliphatic Compounds. The Taft Treatment, , 227, , most easily explained if the mechanism of catalysis is, , A, , ArH+ r, , HA, ->, slow, , A~, , fast, , products, , for here, the catalytic constant is simply the rate constant for the slow step, involving proton transfer (p. 189). Since the free energy of activation for this, proton transfer is linearly related to the standard free energy of ionization of the, acid, the change in AF* caused by a structural alteration in HA is proportional, to the corresponding change in AF° of ionization. Moreover, the proportionality, constant, a, is found practically always to be less than unity. A structural change, that lowers the free energy needed for complete proton removal (ionization, of HA) will also lower the free energy needed for partial proton removal (forma¬, tion of the activated complex X • • • H • • • A); but for partial proton, removal, the effect will, as expected, be less.35 The fact that the same constant,, a, applies to all acids of a given series (for example, to all carboxylic acids or to, , all phenols) suggests that each of these acids forms an activated complex in, which the degree of proton transfer to the substrate X is the same. It is further, reasonable to suppose that as the Br^nsted parameter a approaches unity, the, proton transfer from HA to X in the transition state becomes more and more, nearly complete., , Polar Effects in Aliphatic Compounds. The Taft Treatment54, The quantitative correlation between structure and reactivity for aliphatic, compounds is simplest when (a) the reaction center is not part of a conjugated, system and (b) when the degree of steric interaction between substitut'd, “mi,1,, (and d h ,, , "0t, IT appreciab1^ as the —‘ion progresses. These are, neVerthdess’ a "umber of aliphatic reaction series are known, , condWo, F S T m°re eXiS° Where a‘ *eaSt S°me members conform to these, be, J -°r SUCh reactlons> as for the reactions of meta- and para-substituted, ^— derivatives, substituents influence rates and equilibria Tnly though, equilibria to such ra«-ablIltheS' ^, , ^ SUrprisinS> therefore, that rates and, , identical to the Hammett equadLT a“°nShlP WhlCh’ at first Slancc, appears, , log, , ©-, , a*p*, , (10), , for a particular member of theC°nStam), , (a), , Taft, , T, , op. 586-629. ' •, , A, , T, , °f ihis ,opk’sce BeU-, , ’, , ra University, , ^*729.3,20 0,52>; 75, 423, (1953). (4) sMatoRef. 29>
Page 244 :
228, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , for the “parent compound” in the series (generally the methyl compound, since, data are most readily available for these); p* is the reaction constant, analogous, to Hammett’s p; and a* is the polar substituent constant, again representing the, electron-attracting ability of the substituent being considered, this time as, transmitted through an aliphatic chain. Unlike the a constants, the values of, a* are not defined in terms of dissociation constants of carboxylic acids since, , not all aliphatic acids obey equation (10) (probably because the larger substitu¬, ents, irrespective of polarity, interfere with solvation of the carboxylate anion)., It appears, however, that a rather good <j* value for a given substituent, may be obtained by selecting an ester having that substituent alpha to the car¬, bonyl group, then comparing the specific rates for acidic and basic hydrolysis of that, ester (at the same temperature in the same solvent).55 Taft then defines a*:, , a* =, , 2.5, , log, , (ft 108 (ft, , (11), , ~, , where the subscripts B and A refer respectively to basic and acidic hydrolysis, and the k0 values refer to the specific rates of corresponding hydrolyses of the, acetate ester (the “parent compound”). The factor 2.5 is purely arbitrary,, being introduced to put the Taft a* values on about the same scale as the Ham¬, mett cr values. To understand the basis for this definition, let us compare reac¬, tion series 7 and 8 in Table 7-5. We see that the rates of basic hydrolysis (sa¬, ponification) of benzoic esters, of which series 7 is typical, are very sensitive, to electron attraction or repulsion by substituents (that is, the p value for the, series is high). However, the rates of acid hydrolysis, of which series 8 is typical,, are practically unaffected by electron attraction or repulsion. The p values are close to, zero; in series 8, for example, the rate of hydrolysis for the />-mtrobenzoic ester, exceeds that for the />-methoxy ester only by about 12 percent. On the other, hand, the rates of acid hydrolysis of aliphatic esters are strongly affected by, substituents. Since there is no reason why these esters should be more susceptible, to polar effects than are aromatic esters and since resonance effects cannot be, transmitted along aliphatic chains, we may conclude that steric effects are, intervening in the aliphatic series. With negligible polar and resonance effects, the ratio of kA (the specific rate of acid hydrolysis for the substituted ester), (koU (the specific rate for the parent ester) then becomes a measure of the, , te ic effect of the substituent, and log (*/*°)a becomes a measure of the merement in AF*, arising as a result of this steric effect. Taft, the activated complex in a normal acid-catalyzed hydrolysis (XXIX) d.Hcrs, f, , that in base catalyzed hydrolysis (XXX) only by the presence of two, , additional protons, steric effects in the two types of hydrolyses should *, *, "IT Since both polar and steric effects are important in base-catalyzed hy, u This was apparently first suggested by Ingold, J. Chem. Soc., 1930, 1032
Page 245 :
229, , Polar Effects in Aliphatic Compounds. The Taft Treatment, —, , O, /, , —c, , R—, , \, \, , -I-, , O, /, R-— C, \, , O, \, , -, , O, , O, / \, H H, , H, /, —O, \, R', , /, H, XXIX, , XXX, , drolysis, the term log (k/k^B is a measure of the increment in AF* arising from, a combination of both effects. The difference, [log (k/ko)B — log (A://:o)^], is, thus taken as a measure of the polar effect alone.se, A number of <r* values are listed in Table 7-7; substituents having positive, <r* values are, by Taft’s convention, stronger electron attractors than the methyl, group. The p* values for seven reaction series, which conform in part to the, Taft relationship, are listed in Table 7-8. (Unless otherwise stated, these reac¬, tions are carried out in aqueous solution at 25° C.) Also included (designated n), is the number of members of each series found to give reasonably good agree¬, ment to the Taft equation. As with the Hammett equation, a positive p*, value signifies that the reactions constituting a series are facilitated by electron, withdrawal., The reaction series in Table 7-8 show a wide variation in type, but only, a small number of compounds in each series are known to give good fits to the, Taft equation. We may not assume that because certain members of a series give, good agreement with equation (8), all members of that series will give good, agreement. In very general terms, we may say that quantitative predictions, ol rate or eqmhbrium constants from equation (8) become less reliable as the, Stenc requirements of the substituent increase and as the position of substitution, raws near to the site of reaction. Nevertheless, it is difficult to say a priori, U Ct, ‘, v, , °\ not the Taft cquation will hold in a particular case. Although Taft, , values for over fifty substituents, reaction series including more than, , of these are rare. About twenty of these substituents have not been included, , i:r;rtio; s;tas yet’and dght °th-—, is a, n, ’, c rdationshiP> though of considerable theoretical interest, ’, present, significantly less useful for predictive purposes than is the M, ', mett equation 37, Toft, •, ,, ^, purposes than is the Ham„^, ThC Taft eqUa,1°n has als° been extended to certain reactions, , Z\ki l° a g°°d approximationnipend'enTonemn''^ “, , SUbsti,u™, , O'" acyl, , S°'Ven'’ and 'h' —, sutailueM, , ta, , obtaincd, , <^>s ,he'*values ,o Hamme,t's ““-Lr7tssaa„ed so., , P™por,i„„al, , z, , thc, , ssi
Page 247 :
Steric Effects, , 231, , of aliphatic ketones are severely inhibited by branching at the a positions., Such effects are often coclassified as steric hindrance phenomena; that is, situa¬, tions where contact between reaction centers is made more difficult by inter¬, ference of groups that do not otherwise participate. Although “steric hindrance”, has, in the past, been used to account for a number of trends in reactivity that, are now far more convincingly explained electronically, there remains a sub¬, stantial body of chemical facts which are best explained with the aid of steric, arguments. We should note also that, given the correct conditions, steric effects, may accelerate, rather than retard, reactions. Moreover, although steric factors, are most often considered in relation to reaction rates, equilibria too may be, s-terically affected., To date, the most extensive investigations of steric effects on equilibria have, been carried out by H. C. Brown and his co-workers.88 Brown’s conclusions, are based largely on the dissociation constants of addition compounds formed, by Lewis acids, such as trialkylborons, with amines., K3JN, , :BR'3, , R3N, , -f, , BR', , Consider, for example, the adducts formed by (CH3)3B with ammonia, methylamine, dimethylamine, and trimethylamine. The dissociation constants for, these addition compounds (in the gaseous state at 100°) are as follows:, H3N : BMe3, , MeNH2 : BMe3, , Me2NH : BMe3, , Me3N : BMe, , 0.0350, , 0.0214, , 0.477, , K, , Recall,ng that methyl groups are slightly less effective electron attractors than, ,hyat7hnf,at°mS r"" arC *hUS, stre"Sthe"ing”), we are not surprised, t at substitution of a single methyl group for hydrogen in ammonia leads to a, cmases7latesdn7Ph X Z, , Substhution of a «°nd -ethyl group de-, , meZ 7“, l, ,, W°nderS’ h°WeVer> why subs<i<ution of a third, methyl reverses the trend (this is also the case if H+ is the reference acidl, From the corresponding data for the adducts of the ethylated amines, we see, , K, , H3N : BMe3, , EtNH2 : BMes, , 4-6, , 0.0705, , ducVvZZr;'1:;:of on,y, , Et2NH : BMe3, , 1.22, , ethyi ~, , Et3N : BMe3, very large, , ^ tre„d«bY the in., , m ammonia can loweZhe’ smbility'TZ, , gr°UP ^ 3 hydroSen, , nikyl group is sufficiently bulky (for example Tc H, , tni"e'hylb°r0n if the, , n Wiley and Sons, Inc., 1956, pp. 454-46(/, , °rgamc, , Chemi^ry (edited by
Page 248 :
232, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , group in the 3 or 4 position of the pyridine ring, but is decreased by methyl, substitution in the 2 position. Substitution of methyl groups in both the 2 and, 6 positions of pyridine prevents formation of a trimethylboron addition com¬, pound, even at —80°. It seems clear that the stabilities of these boron-nitrogen, adducts are decreased when the alkyl groups of the amine “get in the way”, of the alkyl groups of the boron compound; if there is too much interference,, an adduct will not form at all. Conversely, quinuclidine (XVIII), in which the, substituents on the nitrogen atom have been “tied back” so that they cannot, interfere with the front-side approach of the boron compound, forms an ex¬, tremely stable adduct with trimethylboron. Steric interaction between the, groups on two different atoms, which arises during the formation of a bond, between these atoms, is designated front strain (“F strain”)., The surprisingly low basicities exhibited by trialkylamines (even, pre¬, sumably, when the steric requirements of the acid are very slight) are more, difficult to explain. Brown originally felt that the coordination of such an amine, with an acidic species—that is, the formation of a fourth bond by nitrogen, requires that the alkyl groups of the amine be pushed back, close enough to¬, gether so that they interfere with each other. Such steric interaction, designated, back strain (“2? strain”), was presumed to destabilize the addition compound,, making the amine appear less basic than primary or secondary amines. Since,, however, the bond angles about the nitrogen atoms in amines are now known, to be nearly the same as the bond angles around the nitrogen atoms in trialkylammonium salts, the formation of a fourth bond by nitrogen should cause, practically no change in steric interaction between the alkyl groups bound to, nitrogen. Moreover, it has been found that the basicities of tertiary amines in, water may be correlated with the Taft cr* values of the alkyl groups,<»<•> which, presumably measure only inductive effects. Since we may thus conclude that, steric factors do not significantly affect the base strengths of aliphatic amines in, water, we may most logically attribute the low basicities of tertiary amines to, solvation effects.*5(6) More specifically, the conjugate acid of a tertiary amine,, R.NH+, , has only one N—H bond, whereas the conjugate acids of secon ary, , and primary amines have, respecdvely, two and three N-^H bonds^Th.s means, that a cation of the type R,NH+ is not as strongly solvated in a hyd . ,, R NH4- and RNHt for solvation of such ions occurs, solvent as are cations R2NH2 and Kixn3, iui, ily by hydrogen bonding through the N—H bonds:, mainl, H, , O, , +/, H—N—, , JJJ, , 39, , 1518., , _ ., n., ~, 79 5441 (1957). (b) Bell and Bayles, J. Chan. Soc., 1952,, (a) Hall, J. Am. Chem. Soc., 79, 5441, \ >
Page 249 :
233, , Steric Effects, , Weak solvation of the ions derived from tertiary amines thus tends to diminish, the basicity of these amines in water. On the other hand, the weak basicities of, tertiary amines toward acids of moderate or large steric requirements (for, example, toward Me3B and /-Pr3B) in nonaqueous media are probably due, at, least in part, to, , F, , strain. Back strain may well enter the picture in accounting, , for the acidities of boron trialkyls, for with these the bond angles around the, boron atom are near 120°. When the boron atom forms a fourth bond, these, angles are reduced to near 110°, thus increasing steric interaction between the, alkyl groups. However, front strain and the inductive effect also decrease the, acidity in such cases, and separation of the three effects is, at present, not, possible., In the dissociation of substitution hexaphenylethanes to triphenylmethyl, radicals (Ar3C:CAr3 ^ 2Ar3C-), the bond angles about the central carbons are, increased from about 110° in the ethane to 120° in the radical. This dissociation, relieves steric strain, not only because it eliminates interference between aryl, groups on different carbon atoms, but also because it lessens interaction be¬, tween aryl groups bound to the same carbon atom. We should then suspect, that substituents that increase the steric requirements of the aryl group should, increase the degree of dissociation of the ethane. Thus, at 25° in benzene, the, dissociation constant for diphenyltetra-o-tolylethane (XXXI) is 1.50, whereas, that for the tetra-/>-tolyl compound (XXXII) is only 0.0013; a difference that, is almost certainly due to the increased crowding in the ortho compound.-*0, , Me, , This reaction is further discussed in Chapter 16, , infcr-hat steric —■, 1° the transition state but will retard a reaction if the7 8°m® fr°m 'he reaCtants, crowded. Most of the reactions subject, * Marvel, Kaplan, and Himel, , J Am Ch, c, ^, c1’ J. Am. Chem. Soc., 63, 1892 (1941)., , lnC ‘anCe, , (for
Page 250 :
234, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , example, the more usual types of esterifications, saponifications, and carbonyladdition reactions) fall into the second category—that is, reactions in which the, density of material around the reaction site is increased in the activated com¬, plex. However, the solvolyses of tertiary halides in water-alcohol mixtures, generally proceed through preliminary loss of halide ion to form a carbonium, ion. The bond angles around the halogen-bearing carbon increase from about, 110° in the reactant to about 120° in the carbonium ion intermediate, and lie, somewhere between these two values in the activated complex. This then is a, case where the formation of the activated complex releases strain, and, as ex¬, pected, the presence of bulky groups near the reaction site leads to increases in, rate. The specific rate of solvolysis for the moderately crowded chloride XXXIII, (at 25° in 80 percent ethanol) is over 21 times that for /-butyl chloride, whereas, that for the very crowded chloride XXXIV is over 500 times as large as for, /-butyl chloride.41 These accelerations seem far greater than can be accounted, , H3Cx, , /CHa, , H3C-C-CH2—C-Cl, h3c, , ch3, , h3c, 3, , \, , h3c-g-ch2h3c, , /, , c, , /, , CH 3, , \, , 2, , XXXIV, , XXXIII, , for merely on the basis of the slight inductive effects of the alkyl groups., Although steric effects are easily visualized qualitatively, a fundamental, quantitative treatment is a task of considerable magnitude. De la Mare and, his co-workers4* have recently proposed such a treatment for a relatively simple, type of reaction—the exchange of Br“ for the bromine atom in alkyl bromides., At first glance, this may not seem to be a true chemical reaction at all, but its, rate may be followed by using radioactive bromide ion in solution and measur¬, ing the speed in which radioactivity is incorporated into the organic bromide., *Br~ + RBr, , *Br—R + Br", , rate = *2(*Br ) (RBr), , (12), , Theoretical calculations were carried out, comparing energies *"d “tr°Pies, of activation for exchange reactions in which the alkyl group was methyl, ethyl, n-propyl, i-propyl, f-butyl, 1-butyl, and neopentyl. In each case, a transit on, sate having the incoming and outgoing bromine atoms equidistant from the, rtion c/nter was assumed a d, , st:: him, , {z;:z:zx^ ^ ---“rr;, , previously been derived for bond-stretching, bond-bending, an, , H Brown and Berneis, J. Am Chem. Sac 75, 10 0953)., it De la Mare, Fowden, Hughes, Ingold and Mack e, J, detailed paper has been reviewed by Ingold, Quart. Rev,, XI,, , or non o, , 1955, 3196. This very, (1957).
Page 251 :
Steric Effects, , 235, , compression energies. Entropy changes in going from reactant to transition, state may be calculated using quantum mechanical treatments for entropies of, translation, rotation, and vibration. Such calculations are obviously complex,, and we cannot describe them here, but we may note two conclusions that, emerge. First, this work confirms the experimental observation that substituents, in the (3 position are considerably more effective “steric hinderers'” than a sub¬, stituents. The energy of activation for the bromide exchange reaction on neo¬, pentyl bromide, (CH3)3C—CH2Br, is estimated to be almost 5 kcal per mole, greater than that for /-butyl bromide, (CH3)3C—Br, this energy difference, corresponding to a difference of almost 4 powers of 10 in the specific rates of, exchange for these two bromides. It can, in fact, be shown, using a scale model, of the transition state, that hydrogen atoms belonging to /?-methyl groups are, better able to interfere with the bromine atoms than are hydrogen atoms, belonging to a-methyl groups^3 (p. 276). Secondly, the quantum-mechanical, relationships used to calculate changes in translational, rotational, and vibra¬, tional entropies (which are then combined to estimate entropies of activation), include the masses of the species involved, and any alteration of structure that, increases the mass of a reactant or alters the distribution of mass will change its, entropy. In particular, when the calculations for the transition state in halideexchange reactions are carried out, it turns out that substitution of a heavy, group for a hydrogen atom in the alkyl chain invariably results in a more nega¬, tive entropy of activation., Branching in the alkyl chain will decrease the rate of exchange, even if the, space-filling abilities of the substituent are ignored. That portion ’of the effect of added, , masses which depends upon their weight but not their volume has been called, by de la Mare and co-workers the ponderal effect. Since it is often in the same, direction as the steric effect, we can see why it is so often included with the latter, Nevertheless, de la Mare’s calculations show that insofar as entropies of activalaliri, , ,,e7 anse in the higher alk>'1 halides diff" from those for methyl, , ”r *“ —■», , ™ —, , factors generally effect the reaction rafJin the same^irect, Th11 ^ ^ 3 change in steric, outstanding exception is the case of bromide exchan^nTk /Vk”0' Tariably so‘ The one, bromide. Here, the substitution of methyl clouds fo^th, *‘butyl brortllde vs. that in methyl, activation (tending to decrease the rate) buf S atom f a-hydr°gens boost*, energy of, (which ,n itself, would tend to increase the rate) As inaM, P°STC entr°Py °f activation, assumed the structure that results in the smallest incre, ^ Cases’ the activated complex has, o stretch the C-Br bond more in the Thu^ T"' The CXtra ^ergy needed, compensated for by the extra freedom associated, the methyl reaction is partially, reaction. Nevertheless, the net steric effect of the substk t^ aCtlVated ComPlex m the former, « A typical case in which ponderal and tl Substltutlon “ strongly retarding., , 1, , which'7', , Am\ Chem' S°C-', , h u s,m,,ar in, , 75’ 4538 (1953)' see1aLCRefaT91UmPCrd,to°gether iS 3 treatment by, *° his—-
Page 252 :
236, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , Steric Inhibition of Resonance, Resonance effects associated with a conjugated system are most pronounced, when the atoms of such a system lie in a common plane; such effects fall off, rapidly as departures from planarity increase. We have seen how steric inhibi¬, tion of resonance affects dipole moments (p. 70), and we need not be surprised, that it may affect rates and equilibria also. It has already been noted (p. 210), that essentially all ortho substituents boost the acidity constant of benzoic acid,, a trend that cannot be due to ordinary steric effects since the steric require¬, ments of the H+ ion are negligible. However, it will be recalled that the, —COOH group is a “ — R group”; contributions of structures such as XXXV, and XXXV' result in transfer of 7r-electron density from the benzene ring to, the —COOH group, thus tending to weaken its, , acidity., , (Corresponding, , structures such as XXX\ I may be drawn for the benzoate ion, but these are, , XXXV, , XXXV', , XXXVI, , considerably less important because of the accumulation of negative charge on, the carboxylate grotip.) However, ortho substituents will “get in the way” of, the carboxyl oxygens unless the latter move out of the plane of the benzene, ring, and when this happens, the acid-weakening resonance effect is greatly, diminished., ,., On the other hand, since a nitro group substituted para to a phenohc, hydroxide group greatly increases its acidity, again by a conjugation effect, we, should expect that the acidity of the phenol may be lowered by pushing the, nitro group out of the plane of the benzene ring by ortho substitution Thus, the two nitro-m-xylenols, XXXVII and XXXVIII, the latter ts the weaker, acid« even though the methyl groups (whose inductive effect^m, , weaken the acidity) are closer, weaken, ”, nitro group in XXXV ill can, effect is cut down. Similarly,, , to the acidic -OH group in XXXVII. The, exhibit jts inductive effect, but its resonance, exniuu, fVXXIX') is a far, N,N-dimethylpicramide (XXXIX), , this treatment is largely confined to series, varied, and fails when electronegattve suhsntuent, “Steric substituent constants, , ten, , are assigned ty T I, , I'^Suced mSTthe reaction site., ^, a,kyl groups by noting the, (, , 229), but such con-, , effects of these group, on the rates°J‘^d ponderal effects., , *■*,nc> New York’, , 'p
Page 253 :
237, , Steric Inhibition of Resonance, OH, , OH, , XXXVIII, (pKa = 8.24), , stronger base than picramide itself (XL)^ since the o-nitro groups in the former, compound force the —NMe2 group out of the plane of the benzene ring, in¬, , sulating the amino nitrogen from the electron-withdrawing resonance effects, of all three nitro groups. In picramide, however, there is very little interference, between the small —NH2 group and the nitro groups ortho to it; departures, from coplanarity are slight, and considerable 7r-electron density is drained off, from the amino group to the nitro groups., Similar effects appear with respect to reaction rates. Aromatic substitution, reactions are generally accelerated by substituents which allow the charge, originally associated with the attacking reagent to be spread over a large area,, either in the activated complex or in a high-energy intermediate. Familiarly,, the bromine atom in />-nitrobromobenzene is far more easily displaced by, bases than is the bromine atom in bromobenzene, principally because the high, concentration of negative charge originally centered at the attacking position, in the base becomes redistributed in the intermediate XLI, not only over the, , R2NH + Br, , RoN-, , azonium ion C H N+'i f, .-e,ectro, , no2, , Convers;ly> the benzene ring in, P i ic reagents (such as the benzenedi-, , deS, Iart:°re, tha" is can, b“, ensuy ffrom, the ammo m(rogen, shift-If,, imo largely because, , Hammett and Paul. X Am. a,m. Soc., 56, 827 (,,34).
Page 254 :
238, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , of the benzene ring in the intermediate XLII to help accommodate the added, positive charge. Thus, the nitro group has activated the benzene ring toward, , PhN+ +, , nucleophilic substitution, whereas the dimethylamino group has activated it, toward electrophilic substitution. Both of these are resonance effects whose, magnitudes, , may, , be, , severely cut down, , by substitution of methyl groups, , ortho to the activating groups, since the latter, under such circumstances, would, be pushed far out of the plane of the benzene ring. On this basis, we can see, why displacements of bromide from the nitrobromo-zw-xylene (XLIII) are, far slower than from jb-nitrobromobenzene.47 A similar argument explains why, the hindered amine, XLIV, does not react with diazonium compounds under, conditions where dimethylaniline reacts readily.4® Nucleophilic substitution on, compound XLII I and electrophilic substitution on compound XLIV may be, , NO 2, XLIII, , XLIV, , said to be “hindered,” but in both cases the effect is obviously very much dif¬, ferent from classical “steric hindrance,” since here there is no direct interference, between the hindering groups (the methyls) and the attacking reagent., The question as to whether hyperconjugation is also subject to steric, hindrance is not easy to answer. Suppose, for example, that we substitute two, methyl groups ortho to a substituent participating in hyperconjugation. Since, hyperconjugation effects are relatively small, it should be diflicult to disentangle, effects supposedly due to inhibition of hyperconjugation from the inductive, effects of the two methyl groups themselves. It has been noted, however, that, the chloride XLV is hydrolyzed in aqueous acetone almost twice as rapidly, as is the chloride XLVII.49 Both hydrolyses are thought to proceed through, carbonium ions, the ion XLVI derived from XLV presumably being stabilized, V Spitzer and Wheland, J. Am. Chem. Soc., 62, 2995 (1940)., ls Friedlander, Monatsh., 19, 627 (1898)., ., «■, ••Baddeley and Gordon, J. Chen,. See., 1954. 2190. For a more defied discuss,on o,, , steric inhibition of hyperconjugation, see Jaffe and Roberts, J. Am. Chem. See., 79, 39, (1957,
Page 255 :
Interactions in Alicyclic Systems, , -, , 239, , by hyperconjugation. Using molecular models, it is possible to show that al-, , XLV, , XLVI, , XL VI', , a, , XLVII, , though both the five-membered ring in compound XLV and the seven-membered ring in compound XLVII are puckered, the latter ring is much more so., The bond from Ca to the benzene ring lies much more nearly in the plane of, the benzene ring in XLV than in XLVII. It might then be argued that hyper¬, conjugation in the carbonium ion derived from XLVII should be considerably, less important than in carbonium ion XLVI, and to the extent that the sta¬, bility of the intermediate carbonium ion governs the rate of solvolysis, XLV, should react more rapidly than XLVII., , Interactions in Alicyclic Systems, Polar and steric effects of substituents that influence the reactivities of aliphatic, compounds, will, of course, be present in alicyclic compounds as well. However,, t e operation of such effects, which depends largely upon how closely the sub¬, stituents approach the reaction center under consideration, may be modified, by the geometry of the cyclic structure. Of the more usual ring systems, inter¬, pretation of the action of substituents should be most straightforward for de¬, rivatives of cyclobutane and eye,open,ane. The cyclobutane ring appears to, ringPareasligh,, dePartUrCS, fr°m, * the cyclopen,ane, ■ng are slight, - bCe, being, at present, of little chemical significance., the, hC reaT°nS 0f cyclopropyl compounds are often strongly influenced bv, ^substantia, «*»/ ** present in thc three.membered ^, , value, , 109 5°" I ^, , “V °n'y 6°<’’ ''“le m0re than one half their “preferred”, , , for example, Pfizer, We, 101, 672 (,945); also Aston in Ref., , 3, pp. 546-548.
Page 256 :
240, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , tion state in which the bonds about the carbon atom are opened out well, beyond the normal tetrahedral angle,57 a transition state which, however, is, not possible for cyclopropyl compounds. Thus at 95° the hydrolysis of cyclo¬, propyl chloride is only 3^00 as fast As that of cyclobutyl chloride and Hoo, as fast as that of cyclopentyl chloride. Likewise, cyclopropyl bromide is unreac¬, tive towards iodide ion under conditions where the higher cycloalkyl bromides, react readily. Furthermore, since the C—N—C bond angles in amines are, ordinarily increased (although sometimes only slightly) when an acidic species, forms a fourth bond to nitrogen, it is not surprising that the highly strained base,, ethyleneimine (XLVIII), is a weaker base than alicyclic amines having the, nitrogen atom as part of a four, five, or six-membered ring. The derivatives of, h2c-ch2, , \ /, N, , I, , H, XLVIII, cyclohexane are of particular interest,5* for among alicyclic compounds, it is, the cyclohexyl system that is most frequently encountered. The cyclohexane, ring is not planar but is known to assume a “zig-zag” or “chairlike” conforma¬, tion (XLIX), in which each of the C—C—G bond angles is very close to, , equatorial bonds, the normal value for tetrahedral carbon, 109.5°. Through the center of the, cyclohexane ring a line may be drawn (AA' in structure L) which crystallographers would call a “six-fold axis of alternating symmetry.” Six of the twelve, bonds projecting from the cyclohexane ring are parallel to this axis and are, thus termed axial bonds (L), whereas the remaining six bonds, which extend, outward from the axis and make angles of 109.5° with that axis, are equatmal, , 7, , type (see Chap. )•, Brown and Gerstein, , P, (J. Am. Chem. Soc.,, , to account for trends in activity among the arger-nng, and, 1956., , sssrSiS'S, , 2926 (1950)) who use it, with less striking success,, 72,, 0 U, w, ,, n b, , (fl) Dauben, *44'
Page 257 :
Interactions in Alicyclic Systems, , -, , 241, , bonds (LI). (Axial and equatorial positions are often abbreviated a and e.), It is to be emphasized that for 1,3 substitution, a pair of cis substituents may, be aa or ee, whereas a pair of trans substituents may be ae or ea. For 1,2 and 1,4, substitution the reverse is true: trans substituents may be aa or ee, whereas cis, substituents must be ae or ea. The two conformations of a given cis- or transsubstituted cyclohexane do not represent different forms separable under ordi¬, nary conditions, for the cyclohexane ring is flexible enough to “turn itself inside, out,” whereupon all bonds that were originally axial become equatorial and, vice versa. This is best shown with a three-dimensional model., Examination of models will reveal an additional feature of the cyclohexane, ring that almost certainly would not be noticed from a planar projection. Any, substituent except hydrogen (and possibly fluorine) in an axial position inter¬, feres with other axial hydrogen atoms, not those bound to adjacent carbons, (for these are on the other side of the ring), but rather with those axial hydrogens, bound to carbons in the 3-positions. As the substituent becomes bulkier, its, preference for the equatorial position becomes stronger. Thus, for /-butylcyclohexane, it has been estimated that the conformation with the /-butyl group, in the equatorial position is more stable than that with the /-butyl group in the, axial position, the difference in energies being about 5.6 kcal per mole.53 This, means that at room temperature, only about 1 molecule in 10,000 will adopt, the latter conformation. The very strong preference of the /-butyl group for an, equatorial position results in “conformational purity,” not only for /-butylcyclohexane, but also for derivatives of it. The 4-substituents must occupy, equatorial positions when trans to the /-butyl group (LI I) and axial positions, when cis to this group (LIII). The reverse is true for 3-substituents (LIV and, LV). Remembering that the axial positions are considerably more hindered, R, , (LIUR-OH, , atT “*, , (LII R =~OH?o, , "1 7”' *hat eSt"S °f^^utylcyclohexanol, , ,rrV T y SaP°nified tha" ,h°Se °f the, , the trans isomer (LIV with th, , isomer, , ’ W“h ‘hC CSterS °f ^bw^yclohexanol,, , to the cyclohexane ring breaks the onn •,, connectlng substituent “R”, „w., ., g, CakSj thC °PP°Slte sltuation prevails. Since an axial, mstem and Holness, J. Am. Chem. Soc., 77, 5562 (1955).
Page 258 :
242, , -, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , substituent is more crowded than an equatorial one, the partial breakage of the, bond to a substituent in the transition state will release more strain if the sub¬, stituent is axial. We may then expect steric acceleration (p. 231) to assist breakage, of bonds to axial substituents.54, As a final example of conformational effects, let us consider the acidities, of alicyclic dicarboxylic acids, a number of which are listed in Table 7-9., , Table 7-9. Acidities of Alicyclic Dicarboxylic Acids3, pK", , PK \, , Ki/4Ki, , Cyclopropane-, , pK\, , pK\, , Ai/4A2, , 1,2-m, , 4.34, , 6.76, , 66.7, , 1,2-trans, , 4.18, , 5.93, , 14.0, , 1,3 -cis, , 4.10, , 5.46, , 5.8, , 1,3-trans, , 4.31, , 5.73, , 6.5, , 1,4-cis, , 4.44, , 5.79, , 5.5, , 1,4-trans, , 4.18, , 5.42, , 4.3, , Cyclohexane-, , 1,2 -cis, , 3.33, , 6.47, , 1,2-trans, , 3.65, , 5.13, , 345, 7.6, , Cyclobutane1,3-m, , 4.03, , 5.31, , 4.8, , 1,3-trans, , 3.81, , 5.28, , 7.6, , Cyclopentane1,2 -cis, , 4.37, , 6.51, , 34.5, , \,2-trans, , 3.89, , 5.91, , 24.2, , 1,3 -cis, , 4.23, , 5.53, , 5.0, , \,3-trans, , 4.40, , 5.45, , 2.8, , As we have already seen (p. 202), the ratio A1/4A2 may be taken as a, measure of the influence of the —COO- group in the monoanion on the acidity, of the remaining —COOH group. Since the carboxyl groups are generally, closer to each other in a m-dicarboxylic acid than in the corresponding trans, isomer, we may expect this ratio to be larger for the cis diacid. The 1,3-cyclohexanedicarboxylic acids are seen to be exceptional, for with these the Ah/4A2, ratio is the larger for the trans diacid. This is in agreement with measurements, made on three-dimensional molecular models which show that for 1,3-substituted cyclohexanes, the substituents are slightly closer to each other when in, the trans (ae) positions (LVI) than in the cis (ee) positions (LVII). However,, similar measurements of 1,2-substituted cyclohexanes show that substituents, in the cis (ae) positions (LVIII) ar* the same distance from each other as substituents, « For example, it has been found (Eliel and Ro,, the />-toluenesulfonate, , group, , m, , J. Am. Chem. Soc., 79,, , 59^, , ^almost 20 times ’as fast as, , SSSSassar
Page 259 :
Exercises for Chapter 7, , 243, , in the “preferred” trans positions—that is, both equatorial (LIX). Why, then,, should the ratio K1/AK2 for the 1,2-trans diacid be so much less than the ratio, for the cis isomer? It is likely that in this case, repulsion between carboxyl groups, forces the trans acid and the anions derived from it to assume the conformation, in which both —COOH groups are axial (LX) and as far from each other as pos¬, sible—this despite the increase in crowding associated with that conformation., COOH, , R, LVI, , LIX, , LX, , It will also be noted that the ratio K1/AK2 is greater for trans-\,?>-cyc\obutanecarboxylic acid than for its cis isomer. This is not a conformational effect,, for the distance between carboxyl groups is unquestionably greater in the trans, diacid. However, the space between carboxyl groups in the cis diacid (LXI), is occupied chiefly by molecules of solvent (that is, by material of high dielectric, constant), whereas the space between carboxyl groups in the trans diacid (LXII), is occupied largely by the cyclobutane ring (that is, a region of low dielectric, constant). Since polar effects are more powerfully transmitted across material, of low dielectric constant, the interaction between carboxyl groups is slightly, more effective for the trans diacid despite the greater separation between the, groups in this isomer. This argument is essentially the same as that used to, account for the high KX/AK, ratios for diethylmalonic and tetramethylsuccinic, (p ->03) C0"traSt ‘° thC COrreSp°ndinS ratios for clonic and succinic acids, , COOH, , HOOC, , COOH, , HOOC, LXI, exercises, , (b) —Si(CH3)3 or —Si(CH3)2?, , LXII, for, , chapter, , 7
Page 262 :
246, , Inductive, Resonance, Steric Effects Upon Reactivity of Molecules, , 6. The following are the acidity constants (in water at 25°) for some substituted benzeneseleninic acids, ArSe02H:, , Substituent, , K X 105, , Substituent, , K X 105, , none, , 1.6, , m-N02, , 8.5, , />-MeO—, , 0.89, , /^-Br, , 3.2, , m-MeO—, , 2.2, , J&-C6H5O—, , 1.3, , m-Cl, , 3.5, , (a) Show that equilibria for the reaction series, , ArSeO.H + H20 <=> ArSeO^ + HaO+, , are governed by the Hammett equation, and calculate the p value for this series., (b) Are the sign and magnitude of p about what you would expect? Explain., (c) Calculate the dissociation constant for /?-nitrobenzeneseleninic acid. (The, observed value is 1.0 X 10~4.) Should <rp or acp be used in this calculation?, (d) Calculate the a constant for the />-C6H50— substituent., 7. Predict which base in each of the following pairs is the stronger if the reference acid, is the proton, H+. Indicate in which cases a reversal of order would result if triisopropylboron, (t'-Pr)3B, is taken as the reference acid., (a) Trimethylamine or trimethylphosphine?, (b) NH2—CH2—CH2—NHt or NH2—CH2—CH2, , NH2?, , (c) Ethylenediamine or N,N'-dimethylethylenediamine?, , (d) Allylamine or trimethylamine?, (e) C6H50-or C6H5CH20-?, (f) CH3NHC1 or NH3?, NH2, , (h) Aniline or /rarzj-4-aminocyclohexanol?, (i) m-Aminobenzonitrile or />-aminobenzonitrile?, (j) /)-Nitroaniline or l-amino-4-nitronaphthalene?
Page 265 :
Exercises for Chapter 7, , 249, , (h) When the dinitro compound, LXIII, is treated with base, the nitro group in the, 1, , position, although more “hindered,” is more easily displaced than that at the, , 4 position., , LXIII, (i) Anthracene-9-carboxylic acid is six times as strong an acid in water as is anthracene-2-carboxylic acid., (j) The loss of Cl from C6H5 C(CH3) 2C1 in aqueous alcohol is retarded slightly by, a p-Cl substituent but strongly by an o-Cl substituent., (k) The dissociation constant of benzoic acid is increased by a factor of 100 by an, o-nitro substituent, but that of benzeneboronic acid, PhB(OH),, is increased, only by a factor of 3 by a similar substitution., (l) ^-2-Hydroxycyclohexanecarboxylic acid is a stronger acid than its cis isomer, in water, but in ethanol the reverse is true.
Page 266 :
CHAPTER 8, , Nucleophilic Substitution Reactions, in Aliphatic Systems, , Of, , the various classes, , of organic reactions (p. 120), nucleophilic substitution, , reactions on carbon have, to date, been studied most intensively.* The net, change in such reactions may be described simply as follows: A basic reagent,, using a pair of its electrons, forms a new bond to the carbon atom under, attack, and one of the substituents originally bound to this carbon atom is, freed, departing with the pair of electrons comprising the bond that has been, broken. This very broad class includes many familiar and useful organic reac¬, tions, among them the Williamson ether synthesis,, [r—6: ]~ + R'—Br-> R—O—R' -f Br~, the alkylations of amines and sulfides,, R3N: -f R'—BrR.S: + R'—I-, , -> [R3N—R']+ + Br-> [R,S-R']+ + I-, , the synthesis and hydrolysis of alkyl halides,, f, , H“i+, , /, Br- +, , Br—R T H2O, , R—O, , \, L, , hj, , i For detailed discussions of nucleophilic aliphatic substitution see: (a) ][ng°ld Structure, !, (b) Hine, Physical Organic Chemistry, McGraw-Hill Book Co., Inc., New or,, , pp., , and Mechanism in Organic Chemistry, Cornell University Press, Ithaca, 1‘>52, PP-, , (c) Streitwicser, Chem. Revs., 56, 573-735 (1956)., , 250
Page 267 :
Unimolecular and Bimolecular Substitution Reactions, H—O:, , 251, , H—O—R + Br¬, , + R—Br, , and the alkylations of sodiomalonic ester,, -» (EtOOC)2CHR + Br, , [(EtOOC)2CH:]- + R—Br —, , Mechanistic investigations have been centered largely about the reactions, of alkyl halides. Among the other types of substrate that have been studied are, trialkylsulfonium (R'—SRf) and tetraalkylammonium (Rr—NR+) salts,' and, /R2C-CR'\, substituted ethylene oxides (, \ /, )• Also of considerable interest are, , V, , O, , /, , />-toluenesulfonate and />-bromobenzenesulfonate esters (often referred to as, tosylates and brosylates); for the substitution reactions of these compounds, unlike, , those of the more usual esters, generally involve the breakage of an alkylO, oxygen bond, R —O—S—Ar., , O, In a great many nucleophilic substitutions, the attacking nucleophile is, either water or the hydroxide ion. These are, of course, the familiar hydrolysis, reactions. More generally, a substitution reaction in which the attacking nu¬, cleophile is the solvent is commonly known as a solvolysis., , Unimolecular and Bimolecular Substitution Reactions, By 1935 it had become evident, largely through the investigations of Hughes, ngold, and co-workers,8 that the more usual nucleophilic substitution reactions, cou d proceed by what appear to be two distinctly different paths. In the first, and eS\, , "eW b°nd ^ bei"S f°rmed at the same time the °ld bond is breaking, , ‘‘h in VT“,0n Sta‘e> the inC°minS SrouP and ‘he leaving group are both, half bonded to the carbon atom being attacked. This is the path by whkh, the basic hydrolysis of methyl iodide proceeds., H, HO- + CH3I, , HO-C-1, -, , reactants, , HO—CH3 + I-, , H/ Ni J, transition state, , products, , Cjp°5 dp‘128) tChPti°n °f the aCtiVated COmP‘eX in this reaction « given, ', , 11, , rCtl0nS are> “ effCCt’, , reoaJ., , Cleave, Hughes, and ’ingold,'“^'935'”HUgh'S and Ingold>, , iM- 15
Page 268 :
252, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , In the second path, the bond to the leaving group is broken before the new, bond is created, the reaction proceeding through an intermediate of somewhat, the same nature as a free carbonium ion. (As we shall see, just how “free” this, carbonium ion is constitutes an important question.) Reactions proceeding, by this mechanism may thus be represented in two steps—for instance, for, hydrolysis of /-butyl chloride,, _ Cl~, , (CH3)3C—Cl —4, reactant, , I, , (CH3)3C+, , IT Q, , —4 (CH3)C—OH + H+, , intermediate, , products, , This path is sometimes called the dissociation, heterolysis, or carbonium-ion mecha¬, nism, or the solvolytic displacement. Here, the leaving group is “pulled off” by, , the solvent, whereas in the direct displacement reaction, it is “pushed off” by, the incoming nucleophile but still with the help of pulling action by the solvent., In the “dissociation mechanism,” the initial step is generally the slow one.s, Since the molecularity of a reaction has been defined as the number of chemical, species that form new bonds or suffer the breakage of old bonds during the rate¬, determining step, we see that the direct displacement mechanism is bimolecular,, whereas the “dissociation” mechanism is unimolecular. (Molecules of solvent, doubtless participate in both types of reaction, but they do not suffer changes in, covalence.) The two types of mechanism are often abbreviated, respectively,, as, , the Sn2, , (substitution-nucleophilic-bimolecular), , and Sn 1, , (substitution-, , nucleophilic-unimolecular) mechanisms., Speaking very broadly, nucleophilic substitutions on primary carbon atoms, tend to proceed by the SN2 mechanism, those on tertiary carbon atoms tend to, proceed by the 6V1 mechanism, and those on secondary carbon atoms often con¬, stitute borderline cases. For the SN1 mechanism, the activated complex in the, slow (bond-breaking) step assumes some of the character of an ion pair; that is,, its formation requires considerable separation of charge. Alkyl groups, as we, have already observed (p. 205), are more effective electron releasers than, hydrogen. Hence, we would expect the leaving group \X (with its pair of bond¬, ing electrons) to depart more easily from Alk3C:T than from Alk-CH2:X, Aryl groups linked to the a-carbon are even more effective in facilitating um. Recently evidence has been obtained that in a few S* reactions, , proceeding, , through pre¬, , liminary dissociation, the initial step is more rapid than, for exampie, Gellcs, Hughes and Ingold, , . C, , ^, , ^ exceptionai cases., , J. Am. Chem. Soc., 77, 11ZZ fl JOO). i nesc, i, F, * Two additional effects may account, in part, for the ^, , unimolecular substitution on, lf ., crowded, , tertiary carbon atoms. Since the tertiary carbon, “Sfrom steric strain, surroundings than does a primary carbon atom, we may expec ^ &, ^ carbon. Such, on expulsion of the leaving group, , tom ^c-Lci are known to undergo unimolecular solvoly-
Page 269 :
Unimolecular and Bimolecular Substitution Reactions, , -, , 253, , molecular substitution than are alkyl groups, for such groups stabilize the, carbonium ion (and the carbonium-ion-like activated complex) by supplying, 7r-electron, , density from the ring (II <-> II')-, , The ease of both the 6V1- and ,5V2-type reactions is affected by the polarity, of the solvent. To predict the direction of such effects, we may use the prin¬, ciples given in Chapter 6 (p. 183). In doing so, however, it should be remem¬, bered that there are many cases in which a solvent may exert specific effects, on a reaction, exclusive of its polarity. In making the following generalizations,, we are therefore speaking broadly. Consider first the tSVl reactions of alkyl, halides and sulfonate esters. Activation in such reactions requires separation of, unlike charge, and we would suspect that such reactions are more rapid in, solvents of high dielectric constant (for instance, water and formic acid) than, in solvents of low dielectric constant (ethanol, ac tone, and dioxane). That,, in general, is what we find. Similar reasoning may be applied to other nucleo¬, philic substitution reactions, both of the 6*^1 and 6V2 types, varying the “charge, type, , of the reactants. The results are summarized in Table 8-1., One further reminder: For reactions in which ionic charge is created or, , destroyed during formation of the transition state, the effects of solvent polarity, are much more striking than for reactions in which charge is merely dispersed, during activation. Although quantitative correlations between solvent and, reactivity will be discussed later in this chapter, the following comparison indicates the extent of the difference. It appears that the specific rates of substitution, or" Zc‘,n, f t ‘he f0nTi0" °f the aCtiV3ted ComP'- requires creation, destruetton of charge will suffer changes of from 3 to 6 powers of 10 when, such as I', Such hyperconjugation would be expected to be more important if there are three, , H, , I, , H, , I, , -C—C+, , I, j? STfihJt ifthe'a-carbon, C°HHh >»»* are loosened, , oSewe’dt 285)rd,ng ““““, , • :A', , c=c, , <-> —, , —, , :x-, , V, , ST *, ^ --bon, , ^
Page 270 :
254, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , transferred from water to absolute ethanol. In contrast, SN reactions for which, activation requires only dispersal of charge are from 3 to 10 times as rapid in, ethanol as in water.5, , Table 8-1. Predicted Effects of Solvent Polarity on Rates, of Nucleophilic Substitution Reactions5(a), Effect upon Rate, of Increasing, Type of Reaction, , RX->R+X~ (SN\), , Requirements for, , Dielectric Con¬, , Activation Process, , stant of Medium, , Separation, , of unlike, , Increase, , charges, Y~ + RX -► YR + X- (Sn2), , Slight dispersal of, , Decrease, , negative charge, Y: + RA' —* Y:R+ + A'- (SN2), , Separation, , of unlike, , Increase, , charges, R3S+->R+ + R2S (M), , Slight dispersal of, , Decrease, , positive charge, R3S+ + Y-, , RF + R2S (Sn2), , Partial neutralization, , Decrease, , of charge, R3S+ + R3N -> R2S + R4N+ (Sn2), , Slight dispersal of, , Decrease, , positive charge, It is interesting that a change in solvent may bring about marked alter¬, ations in the mechanism of a reaction. As we shall see, a number of the, reactions of /-butyl halides that proceed by the ^1 mechanism in aqueous, alcohol will proceed by the Sn2 mechanism in acetone. Conversely, reactions, of methyl halides that proceed by the SN2 mechanism in aqueous alcohol may, adopt the SN1 mechanism in the more strongly ionizing solvent, formic acid., , Kinetics, Without giving the matter much thought, one might expect bimolecular sub¬, stitution reactions to display simple second-order kinetics,, rate = *2(R:A)(T:), , 0), , and unimolecular substitution reactions to display simple first-order kinetics., t (a) Examples of the solvent effects for Sn reactions of the various charge types are listed, by Ingold Ref. 1(a), pp. 347-350, and by Streitwieser, Ref. 1(c), p. 602. See also de la N_are,, Leffck, , and Salami, J. Chem. Soc., 1956, 3686. (b) For■eampta.ofspecific^solvent effects, , (exclusive of dielectric constant, see Grim, Ruf, and Wolff, Z. physik. Chem., B13, 3, and Norris and Haines, J. Am. Chem. Soc., 75, 1425 (1953)., , (
Page 271 :
Kinetics, rate = ^(R:^), , 255, (2), , Furthermore, one would anticipate that all SnI reactions of a given substrate, would proceed at the same specific rate (as long as solvent, temperature, and, ionic strength are kept constant), since all such substitutions should have the, same rate-determining step—the initial bond breaking. Such a correspondence, between order and molecularity is often, but by no means always, observed., Quite obviously, if the attacking reagent in an 6V2 substitution is a solvent, molecule, the reaction will appear to follow first-order kinetics since the con¬, centration of the solvent will not change significantly during the course of the, reaction., It is also important that a unimolecular substitution reaction will show, substantial departures from first-order kinetics if the initial ionization step is, significantly reversible—that is, if the reaction may be represented by the fol¬, lowing sequence:, , R:X^ :X+, , R+R:Y, , (3), , we have already considered a substitution reaction of this type—the hydrolysis, of benzhydryl chloride (CeHs^CHCl, , (p. 170). Here, the concentration of the, , attacking reagent, water, was considered constant, and application of the steady, state approximation gave the rate expression,, rf(RF) _, , dt, , kf(RX), <, kr(:X), ^ k', , ., , (4), , If the concentration of the attacking reagent (:F) varies during the reaction,, the rate expression becomes, , rate =, , i +, , kr(\X), k'(-.Y), , (5), , A rate expression such as equation (4) or (5) is very compelling evidence for a, preliminary bond-breaking step-much more so than is the rate expression, S'ven in equation (2), The overall rate of unimolecular substitution should,, , inolecular^ThcpreHminar^bond-br^ak^nit ueD^'j^*101'01! rePrese"ted b>’ sequence (3) as uniSlow step, its rate does not in itself determine the * uninJ°*ecular! but although this may be the, equation (4), the rale of tSe overall^ ™TonH, °T, reacti°"- As * «**«« from, speaking, for a complex reaction one is nn <= f, ds ^pon kr and k a* well as kf. Strictly, of individual steps only We shall hm, ^ .grOUnd when one refers to the molecularity, reactions involving preUnVry^, , ^ to, , “"^hUic substitut on, , terminology, although no, completely cormcf l , ,, T, rea<:,i°"s” *"« such, organic chemist’s point of view such reaction, I’, CXtreme,1y widespread, and since from the, , TTnlT'", , which the overall rate is ,he St 'he ^ of, *", *» th“' * reactions*, me rate oi the preliminary bond-breaking step.
Page 272 :
256, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , in the absence oi other effects, be decreased by addition of nucleophile :A'from, an outside source, for some of the added :A' should combine with R+ present, in solution, reversing the initial “dissociation” step. Since such an effect de¬, pends upon a simple shift in the equilibrium of the bond-breaking step, it is, often called a mass-law effect and will be operative only if the added nucleophile, is chemically the same as the fragment resulting from the initial bond breaking., Thus, we should expect that an alkyl bromide which hydrolyzes via the SN1, path should be hydrolyzed more slowly if an ionic bromide is added to the, reaction mixture, but not if an ionic chloride is added. (The reverse should, be true for the hydrolysis of an alkyl chloride, and neither hydrolysis should be, retarded by addition of an ionic nitrate or fluoride.) And that is not all; the, mass-law effect should come into play even though no common ion is added;, for as a substitution reaction proceeds, the concentration of the leaving group, must increase. This means that a SnI reaction, during its early stage, may, display first-order kinetics, but as the leaving group accumulates in solution,, reversal of the bond-breaking step may become important. The reaction rate, should lag farther and farther behind that predicted from the initial first-order, rate constant, but the rate will fit an expression such as equation (4) nicely., However, these mass-law effects are, , not, , observable for all iVl reactions., , For example, the hydrolyses of (C6H5)3C—Cl in “85 percent aqueous acetone,”7, of benzhydryl chloride and bromide in 80 percent aqueous acetone,5 and benzyl, chloride in 61 percent by weight aqueous dioxane0 are retarded by addition of ex¬, cess “common ion,” but the hydrolyses of ^-butyl bromide in “90 percent aqueous, acetone70” and .y-butyl bromide in “75 percent aqueous alcohol17" are not., The mass-law effect will not be observed for an 6V1 reaction if the ratio, kr^^l in equation (4) or the ratio, , in equation (5) is very much less, , than unity. This will be the case if the carbonium-ion intermediate, R+, is, much more likely, under the reaction conditions employed, to combine with, the attacking nucleophile than to revert back to the original reactant. In the, hydrolysis reactions just cited, water molecules are, in effect, competing with, halide ions for the carbonium ions formed in the initial ionization step.. A, carbonium ion, during its rather short lifetime, is surrounded by a “solvation, cage,” consisting largely of water molecules since water is a much more effective, solvating agent than acetone, dioxane, or alcohol. Such a solvated carbonium, r swain Scott, and Lohmann, J. Am. Chem. Soc., 75, 136 (1953). Compositions of mixed, , :, , 'A 5735 (1*53).
Page 273 :
Kinetics, , -, , 257, , ion will persist either until it pulls a water molecule out of its solvation cage, (forming a molecule of the carbinol and causing the remainder of the cage to, collapse), or until a halide ion pushes through the solvation cage to attack the, positively charged carbon. The relatively unstable /-butyl and y-butyl carbonium ions, in which positive charge is concentrated over a small volume, will, tend to destroy their solvation cages before an external halide ion can attack., As the stability, and hence the average lifetime, of the carbonium ion increases,, reaction with such a halide ion (hence reversal of the ionization step) becomes, more likely. Since such reversal is the basis of the mass-law effect, we see why, the importance of this effect increases as the intermediate carbonium ion becomes more stable., The competition between water molecules and halide ions for a carboniumion intermediate is somewhat one sided because of the favored position of water, molecules in the solvation cage of the carbonium ion. If, however, the attacking, reagent is not a component of the solvent, a closer competition (and, hence, a, mass-law effect of greater magnitude) should result. This is the case for sub¬, stitution reactions carried out in liquid SO2. Thus, the rate of conversion of, m-chlorobenzhydryl chloride to the corresponding fluoride, m-C\—C6H4—CHPh—Cl + F-, , m-C\—C6H4—CHPh—F + Cl¬, , ean be depressed by more than 99 percent by addition of excess chloride ion.i0, On the other hand, as the bond making and bond breaking become more, nearly contemporaneous—that is, as the substitution becomes more nearly, bimolecular in character, , the carbonium-ion intermediate may no longer be, , thought to have even a transitory existence, and the mass-law effect disappears., In addition to the common-ion effect (and independent of it), substitution, reactions are subject to the same salt effects that influence virtually all reactions, involving ions (p. 185). If a reaction requires separation of positive and negative, charge during formation of the transition state, it will be favored by high ionic, strengths. If charge is partially neutralized or dispersed during formation of, the transition state, the reaction will generally proceed more rapidly in solu¬, tions of low ionic strengths. Since these effects are in the same direction as the, etlects of dielectric constant, they need not be discussed here in detail. However, we should note that the magnitude of such salt effects depends, in part, upon, he dielectric constant of the medium; for charges operate more effectively, con°stam,mr, '“"r, CO"StantS than throuSh media °f high dielectric, constants. Consider, for example, the hydrolysis of 0.1 molar /-butyl bromide in, law effectCbutne' ,h, ions increases ,h, , h TT"’ “ * ^, Sh°WS "° “PP"^ mass7, ^ procceds’ the accumulation of H,0+ and Br", , charge b th: £jT, , ^ hydr°‘ySiS, , -P-™*- of, , magnitude of which, V"’ ^, rate wlU show an uPward drift, the, which must necessarily increase throughout the course of the
Page 274 :
258, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , reaction. Moreover, as expected, the salt effect in “90 percent aqueous acetone”, is significantly greater than that in “70 percent aqueous acetone.” In the first, (the less aqueous) of these solvent mixtures, the apparent first-order rate con¬, stant has risen by about 16 percent of its initial value when the reaction is, four fifths complete; in the second of these mixtures, the corresponding increase, is only about 7 percent., In describing what we have chosen to regard as two distinct mechanisms, for nucleophilic substitution, we have, quite naturally, tended to emphasize, the extreme cases. However, it was noted that a number of such substitutions,, particularly those on secondary carbon atoms, constitute what were noncommitally called “borderline cases.” Many of these follow, to a rather good ap¬, proximation, a rate law of the following type:, rate = ^(R: X) + *2(R: X) (: Y), , (6), , and it is perhaps natural to suppose that these borderline cases are simply, reactions in which some of the individual acts of substitution are unimolecular, and others are bimolecular. Yet, in a number of these cases, the rate data give, an equally good fit to more complicated expressions that are consistent with a, picture in which each of the individual acts of substitution in a given reaction is, very nearly the same.IJ(o) In such intermediate mechanisms, the attacking, reagent may be considered to approach the substrate and, either by ion-dipole, or dipole-dipole interaction, to supply the “push” necessary for bond breaking, in the substrate without actually colliding with it. Using such a picture, it may, be shown (Ex. 12) that the rate should depend upon the concentration of at¬, tacking reagent but not be proportional to it. The question as to which picture, of a borderline-type substitution reaction is more nearly correct is both im¬, portant and difficult, and strong arguments exist in support of both sides.**, , Attacking Reagents and Leaving Groups, That atom of a nucleophilic reagent which, during a substitution reaction on, carbon, becomes bonded to the carbon atom may be conveniently designated, as the “attacking atom.” Because a nucleophile attacks by using a pair of its, own electrons, we would suspect the most effective nucleophile to be the one, whose attacking atom has the valence electrons most available for coordination., Since this seems, broadly speaking, to be the criterion for base strength, it is, reasonable that the strongest bases should be the most effective reagents for, « For arguments that “borderline” Sv reactions proceed by a single mechanism, I, than a combination of unimolecular and bimolecular mechanisms see (a) Bird Hughes, a ^, Ingold, J. Chem. Soc., 1954, 634 and (b) Winstein, Grunwald, and Jones,,J., 73, 2300 (1951). For rather compelling arguments to the contrary, see (c), 1956, 4633., , ., ,, , • ^
Page 275 :
Attacking Reagents and Leaving Groups, nucleophilic substitution, , reactions., , But, , basicity, , involves, , 259, , thermodynamic, , measurements whereas nucleophilic “pushing power” (or, as it is often called,, nucleophilicity), , generally involves reaction rates. Again, basicity most often refers, , to coordination with H+, whereas nucleophilicity, as we use the term, involves, coordination to carbon. Therefore, a correspondence between the order of, basicities and the order of nucleophilicities among various reagents, although, reasonable, is by no means axiomatic.*5, Postponing for the present the discussion of the measurement of nucleo¬, philicity, we may note that such a correspondence does exist; that is, the strong¬, est bases generally, but not always, carry out displacement reactions most, rapidly. Yet even this rougn correspondence is subject to an important provision:, it will hold only as long as we limit our considerations to nucleophiles whose at¬, tacking atoms are in the same horizontal row in the periodic table. In particular, it is, most useful for groups of nucleophiles having the same attacking atom. Thus,, for nucleophiles in which the attacking atom is oxygen, we may list, in order of, decreasing nucleophilicity:, , OH">OPh" >COi">OAc" >H, , Br, , so7>H2o>cior, , and for nucleophiles in which the attacking atom is nitrogen:, , o, , nh^>nh3>n, , nh2>o2n, , Keeping the charge type the same but varying the attacking atom gives the, following sequence:, R3C:- > R2N- > RO- >FSuch lists must be interpreted with care, for there are a number of cases known, where the relative nucleophilicities of two reagents will be reversed as the sub¬, strate is changed. Occasionally such reversals may be rationalized by calling on, steric effects,, , 4, , but often there appears to be no satisfactory explanation. Next, , we may consider such series as F~, Cl“, Br~, I~ and OR- SR~, SeR-—that, the HlmZInTT, eXfeCt the same type of structure-reactivity equation (for instance,, constants h, \ aft e<luatl°n) to apply to a set of bases with respect not only to their basicity, corrdation bet, , ° l° *<75pccific rates of ^ir participation in ^ reactions. In such cases the, , SS"" baS'C,'y and nucle°Philici‘y is obviously no. only qualitative but also, doespy’ridine*but'“ IT"'"1,, mor' raPidly with methyl iodide than, with isopropyl iodide (Brown and Fid ’ ,c,ties °f th',wo amines is reversed in their reactions, between EtjN and, activated complex., , P, , h’, ^ CW ?*" 7‘- 445 <1949»- The faction, be SCVerely retarded bV “front strain” (p. 232) in the
Page 276 :
260, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , is. nucleophiles in which the attacking atoms are members of the same vertical, group in the periodic table. Here, the weaker bases are derived from the, , heavier, , member8 of the series; yet these bases prove to be the most effective nucleo¬, philes. Thus, we have, in the order of decreasing nucleophilicity,, I- > Br~ > Cl- > F~, SeR- > SR- > OR- and, R2Se > R2S > R.O, The reason for these trends is a matter that merits conjecture, especially since, the stronger nucleophiles in these series tend to form the weaker bonds with, aliphatic carbon atoms. It is sometimes suggested that the heavier members of, these groups are more nucleophilic largely because their atoms are more, polarizable; that is, the clouds of valence electrons about these atoms are most, easily distorted by electrophilic centers. In visualizing the formation of the, activated complex, we may imagine the electron cloud of the attacking atom, to be pulled in toward the carbon atom under attack. (This is during the early, stages of the reaction, before appreciable energy due to the formation of the, new bond is released.) The more easily this electron cloud is distorted, the lower, will be the activation energy required for substitution. Negative ions having the, very polarizable sulfur atom free to attack (thiocyanate, thiosulfate, sulfite,, and sulfide) would be expected to be (and are) especially effective nucleophiles., Similarly effective are the azide (Nj) and cyanide ions, in which the negative, charge is distributed over more than one polarizable atom., Yet polarizability can hardly be the entire story. If the attack is by an, anion, solvation must also be considered. During the activation process, the, solvation cage of the attacking nucleophile must be disrupted and a new “cage”, built up around the activated complex. Any difficulty in breaking up the solva¬, tion cage of the attacking species must be reflected in an increase in the activa¬, tion energy for substitution, hence in a decrease in reaction rate. Now, it is well, known that the solvation energies of ions increase as their charge to size ratios, increase; that is, it takes more energy to “desolvate” (partially or wholly) a, small ion than a large one having the same charge. Thus, the relati\el> low, nucleophilicities of reagents with small attacking atoms are due, at least in part,, to the difficulty in removing molecules of solvent from the attacking site., » Edwards (7. Am. Chem. Soc., 76, 1540 (1954)) has shown that the, number of reagents are closely related to their oxidation potent,als (that is, to the hatf-ceU p, . . f, o y-_. V, 2r~) This correlation suggests that the same lacto, , r0:?;g:(e^, , should govern both properties. It may be shown tsee, ior example, uuu ,, s, SS Henry Holland Co., New York, 1955 p. 207, .ha, the spines * », *, ,o Xt if: M .he X—X bond energy is high; <« .he ,on.Za,.on po.ent.al of A, s ow, O, mlvation energy of X~ is low; and (d) the solvation entropy of X is high. Of these lour i, , also govern the nucleophilicity of X~, it is not surprising that this correlation
Page 277 :
Attacking Reagents and Leaving Groups, , -, , 261, , It is to be emphasized that the specific rate of a nucleophilic substitution, , not, , will, , depend markedly on the nucleophilicity of the attacking reagent if the, , substitution proceeds by a carbonium-ion mechanism., , In fact, for those, , SN1, , reactions in which the dissociation step is essentially irreversible, the specific, rate should be independent of the attacking reagent. Thus, ^-butyl bromide in, nitromethane undergoes substitution by a number of reagents,, chloride ion,, , bromide ion,, , pyridine,, , among them, , and water. The specific rates for these, , reactions are very nearly the same although the products are, of course, dif¬, ferent.^ Even if the dissociation step in an, the, , initial, , Sn\, , reaction is significantly reversible,, , rate should be independent of the identity of the substituting agent., , During the later stages of the reaction, the substituting agent may influence, the overall, , reaction, , rate,, , depending, , upon how, , successfully, , it, , can, , compete, , with the leaving group for the carbonium ion formed in the initial step. A very, , k (*Ar), effective nucleophile will lower the value of the fraction A *, . in equation (5),, , /c, , i), , thus tending to eliminate the downward drift in the apparent first-order rate, constant as the reaction proceeds. But there is a limit; no matter how effective, a nucleophile the attacking reagent is, it obviously cannot make the overall, reaction faster than the initial bond breaking without changing the reaction, mechanism., Turning now to the leaving group,, , :X,, , we may note that the activation, , process, both in bimolecular and unimolecular substitution, requires a stretching, of the, , C:X, , bond. To be an effective leaving group, a substituent should be, , bound to the carbon atom with a relatively weak bond. Moreover, a rather, good, , negative correlation, , tendencies ’ of substituents., , exists, , between, , the, , basicities, , and, , the, , “departing, , The less basic the substituent, the more easily it is pulled, , off by solvent or pushed off by an attacking nucleophile., , The following groups are thus, , arranged in order of the ease with which they are displaced from aliphatic, carbon atoms:, -NR2 < -OH ~ -OR <, , -NR+ < — OAc «, , —F <, , -Cl<-Br<-I<-0S02-^~Y-CH3<— OSO, , (Again, there may be individual cases where the order of “departing abilities”, of two substituents may be reversed when attacking reagent, alkyl group, , or, , -—■»..*, 2, , I, , *, , n t^iese» t^ie added proton draws the, , H, ,2”8°ld', , J- a™- Soc-., , 1954, 647., , De la, , Mare, Hughes,
Page 278 :
262, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , electrons of the alkyl-oxygen bond toward the oxygen atom, greatly facilitat¬, ing departure of the substituent. Thus, alcohols and ethers may be converted, to alkyl bromides by HBr although they are inert to Br- in neutral solutions, Similarly,, , although, , displacement of the —N(CH,), group from aliphatic, , carbon is ordinarily not possible, the “leaving efficiency” of this group may be, very much increased when it is methylated to —N(CH3)+., It is interesting that iodide, besides being a very powerful nucleophile,, is a very effective leaving group. This at first seems to be an inconsistency, and, one may ask, “If a C—I bond is so easily broken, why is such a bond so readily, , formed in displacement reactions?” In answering this, we should note once more, that the activation process for breaking a bond involves stretching that bond;, it is here that the low C—I bond energy (57 kcal per gram bond) facilitates, reaction. On the other hand, for a reaction in which a new G—I bond is being, formed, bond formation has barely begun in the activated complex; and the, energy of the bond-to-be is not nearly so important in determining the activa¬, tion energy as are the ease of polarization and the ease of desolvation which, make the iodide ion such an effective attacking reagent. The dual ability of, iodide to attack and depart readily in nucleophilic displacements makes it an, effective catalyst for other such displacements. Thus, addition of small amounts, of iodide to an aqueous solution of methyl bromide accelerates the hydrolysis, of the latter/7 since, in addition to the ordinary slow reaction of water with, CH3Br, the following series of more rapid reactions may occur, _ CH*Br (CHsI, , IBr-, , h’°, , fH+I-, , CH*Br \CH3I, , * ICH3OH, , * IBr-, , etc., , ->, , The net result of this reaction series is the destruction of methyl bromide while, the iodide is continually being used, then regenerated., The identity of the leaving group in an Sn2 reaction should obviously, affect the reaction rate (that is, an alkyl chloride, bromide, and toluenesulfonate, should undergo substitution reactions at different rates), and if more than one, attacking nucleophile is present, the leaving group may also influence the rela¬, tive amounts of products formed. With an 6V1 reaction, the leaving group, will likewise determine the reaction rate, but it should not influence the ratio, of products formed, for the formation of the new bond must presumably wait, until the leaving group is removed from the reaction site. A mixture of products,, the composition of which is independent of the leaving group, is an excellent, indication that the unimolecular mechanism is operating, just as is a substitution, rate which is independent of the identity of the attacking group. For example,, let, , us, , compare, , the, , reactions, , of benzhydryl, , chloride,, , n Moelwyn-Hughes, J. Cfietn. Soc., 1938, 779., 18 Church, Hughes, and Ingold, J. Chem. Soc., 1940, 966, , ., , (C6H5)2CHC1,, , and
Page 279 :
Stereochemistry, , -, , 263, , benzhydryl bromide, (C6H6)2CHBr, with 90 percent aqueous acetone which, is also 0.10 molar in sodium azide, NaN3." The products in both cases are, benzhydrol, (C6H5)2CH—OH, formed in the attack by the water, and benz¬, hydryl azide, formed in the attack by the very nucleophilic azide ion. Although, the bromide is found to react 33 times as rapidly as the chloride (at 50°), the, mixtures of products resulting from the two reactions have, to within experi¬, mental error, the same composition (66 percent hydrol and 34 percent azide)., When the carbonium-ion intermediate in an 6V1 reaction has one or more, /3-hydrogen atoms, another type of reaction may become important. Instead of, undergoing substitution, the carbonium ion may undergo elimination of a hydro¬, gen ion, forming an olefin.19, , \, , /, , +/-h-\, , CH—C, , ->, , \, , /, , C=C, , /, , \, , Since the laws of chance governing the fate of a given carbonium ion do not, depend upon the mode of origin of the ion, the ratio of the olefin to the substi¬, tution product (but not the specific rate of formation of each) in an Sn\ reac¬, tion should be independent of the leaving group. This is often, but not invari¬, ably, the case. Typically, both /-butyl iodide and /-butyl bromide undergo, solvolysis in aqueous ethanol at 25°. Although the iodide is destroyed almost, three times as rapidly as the bromide, the mole fraction of f-butene in the mix¬, ture of products (0.13) is the same for both halides/0 With /-butyl chloride,, which is destroyed about 3^43 as rapidly as the bromide under the same condi¬, tions, the mole fraction of olefin in the product rises to 0.17. While this differ¬, ence is small, it seems larger than the experimental error, and similar small, differences have been observed for the /-amyl halides. Although the question, must be considered an open one, it appears at present that nonparticipating, ions may influence (albeit only slightly) the relative rates of the possible reactions, of a carbonium ion in solution., Furthermore, if more than one olefin is obtained from a given carbonium, ■on, the ratio of olefins in the product should also be independent of the source, Of the carbonium ion. We shall see in Chapter 12 that this is the case., , Stereochemistry, grotna hitsle ^, !T 3, disPlaceme"‘ action, the incomin,, group h ts the carbon atom under attack on one side while the leaving grout, upon the ratt^of'appearanc^of^he free^eavim?1'^ b^“°n’ “V ***> study base,, combined rate of substitution and elimination reactions The, ”, f°llowinS th, sis based upon the rate of appearance of H+., ', Same 1S trUC °f a study of solvoly, Cooper, Hughes, and Ingold, J. Chem. Soc., 1937, 1280; 1948, 2038.
Page 280 :
264, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , departs from the opposite side. This process, known as the Walden inversion,, is said to turn the molecule “inside out,” a description that is best understood, if we examine a case in which four different groups are attached to the carbon, atom under attack., , X, , Y, , T, , \, , Note that the section U—C— of the substrate is converted to its mirror image,, V, , —C—U, as a result of the reaction, and that if T, U, and V represent different, , \, V, , substituents, the mirror images cannot be superimposed. In short, the configura¬, tion about the carbon atom has become inverted during the course of the dis¬, placement reaction., Since we obviously cannot see the individual acts of molecular substitution,, how can we be sure that this inversion occurs? It might at first seem that if a, dextrorotatory reactant yields a levorotatory product (or vice versa),, , then, , inversion has taken place; that is, a change in direction of rotation indicates an, inversion of configuration about the asymmetric carbon atom. Although this, is very often true, there are a significant number of cases known where an, alteration of an optically active molecule may reverse the direction of rotation, even though the configuration about the asymmetric atom remains unchanged., •i For example, acetylating the amino group and ethylating the carboxyl group in dextro¬, rotatory alanine, CH3-CH(NH2)-COOH, yield a levorotatory ester, although neither of, these structural changes should affect the configuration about the gV™na«nccgto atom, Even more disturbingly, the mere conversion of lactic acid, CH3—CH(OH) COOH, to, anion reverses the direction of rotation., ,, The correlation between configuration and direction of rotation can be made somewha, more reliable by comparing not only the molecular rotations for the compounds under con, sideration but abo the changes in rotation when temperature, solvent, concentration, and w, S ZolarSd light are varied. A further degree of reliability is obtained by comparing, changes in rotation when the compounds under consideration are subjected to simil^ ahe tions in structure which must not, however, alter the configurations about the, ^, asymmetry (acylation of ammo groups, esterification (192£°°0 2447 0927); 61, 1083,, example, Freudenburg and co-workers, Ber., 56, 193, (1923), bU,, )>, (V^ ke^rtheiei’nL absence of other information, conclusions‘'h« ha«b«n b^dup®, such comparisons, however reliable they may be, cannot, ^ch comparisons leave unanswered the, a product is in comparison with the reactant that is, j, place during the reaction., , e regar, , optically pure
Page 281 :
Stereochemistry, , -, , 265, , Since direction of rotation is thus a useful, but by no means conclusive, indica¬, tion of relative configuration, we must turn to more rigorous methods to demon¬, strate that every act of direct displacement in a given reaction is accompanied, by inversion of configuration., Two important groups of experiments unequivocally established this for, a number of direct displacement reactions, and the assumption that all direct, displacement reactions invert, when further applied to many additional groups, of reactions, has led to no significant inconsistencies. Let us consider a single, example** from nearly a dozen reaction series investigated by Kenyon, Phillips,, and their co-workers. In these, an optically active alcohol (with the alphacarbon asymmetric) is converted to its enantiomorph. (In the following series,, the a values designate the specific rotation for each of the compounds at 5461 A, and 23° C.), Me, , Me, , PhCH2—CH—OH, , OAc, , PhCH2—CH—OTos ^ PhCH2—CH—Me, , a = +33.02°, , a = +31.11°, , a = —7.06°, OH, PhCH2—CH—Me, a = -32.18°, , SO 2—), , (Tos=H3C, , Since the resulting carbinol has a rotation opposite in sign but nearly equal in, magnitude to that of the original carbinol, an inversion of configuration at the, asymmetric carbon atom must have occurred in one of the three steps. In the, first step-the conversion of the alcohol to the tosylate-the reaction takes, th, , aV 6 °X^en, , t^ie, , group rather than at the asymmetric carbon;, , Suratlon about the carbon must therefore remain unchanged here!, ir, , step is the basic hydrolysis of an ester; since it is now known that such, , reactions ordinarily** involve breakage of the acyl-oxygen, rather than the, alkyl-oxygen bond /that is, —C—O-, , C— rather than —C, , b, , /, , J. Chem Soc., 1923, 64; 1925, 399, 2552; 1926 ''052 ^, “U, he supposed that Inclusionof.S «, argument given; for, we may ask, how do we kn„, , .1, , 3, , -O—C-, , the, , O, , + Vr™' "llllll>s and co-workers, , i, °rdl.na"l>', , 1072- 1663i 1936' 503, substantially weakens til., , rather than the “extraordinary” t p^ al ; ", ? ^ hyd'°'^s is of, “ordinarypresent to say that considerable confidence to those ,knOW 'n°u8l> ahout ester hydrolysis a, alky-oxygen breakage are of a few very spec alLed 7T 'rh,chr u"d'rK° saponification witl, O, r r ar CS"r- the Position of bond breaking ma n7- a, ChaP- 9)- Moreover, for an,, ut, saponification in water “labeled^ with"heavy^xygen6
Page 282 :
266, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , configuration about the asymmetric carbon atom remains unaltered in this step., The inversion must therefore have taken place in the second step, the displace¬, ment of tosylate by acetate. The same conclusion arises from a number of, similar Kenyon-Phillips cycles, and it can be likewise shown that the displace¬, ment of TosO~ by EtO- results in inversion of configuration. It is significant,, however, that if such cycles are applied to carbinols having a phenyl group, alpha to the asymmetric carbon, the resulting carbinol has a rotation opposite, in sign to that of the original carbinol, but the magnitude of its rotation has fallen, substantially; that is, considerable racemization has accompanied the substitu¬, tion of acetate for tosylate. Such substitutions, however, no longer display, second-order kinetics., A second set of experiments of considerable fundamental interest employs, optically active alkyl halides, again with the a-carbon atom asymmetric.^, , I, , I, When, for example, </-2-iodooctane, CH3—CH—C6Hu, is added to a solution, containing free iodide ion, the organic iodide undergoes racemization. Further¬, more, the rate of racemization is proportional to the concentration of I-., Iodide ions are quite obviously attacking the ^/-iodide molecules, forming new, C—I bonds and breaking old ones., </-C8H17I + I- —> /-c8h17i + IDextrorotatory molecules are being converted to levorotatory molecules iin, some or all of these displacements (without further evidence, we cannot decide, which). When half of the original d molecules have been so inverted, the solu¬, tion will have become optically inactive and any further acts of displacement, will bring about no change in specific rotation; for at that point in the reaction,, a d molecule will be just as likely to be attacked as will be an / molecule. But, this exchange of iodine between inorganic and organic iodide may be observed, in another way. If radioactive iodide is used in solution, the iodide exchange, will result in incorporation of radioactive iodine atoms in the organic iodide., The fraction of radioactive iodine in the iodooctane will increase until it be¬, comes equal to the fraction of radioactive iodine in the inorganic iodide. Now, if every displacement of iodide by iodide results in inversion, the rate of racemiza¬, tion should be just twice the rate of iodide exchange (as indicated by transfer, of radioactive iodine); this must be so, for a given group of molecules of d-iodooctane will be completely racemized when just half of them have been invertedthat is, when exchange has proceeded half way to compietion. The ratio of, racemization to exchange in this case has been found to be: 1.93 ± 0.16., similar halide exchange experiments with a-phenylethyl bromide and a-bromoHughes and co-workers, J. Chem. Soc., 1935, 1525; 1936, 1173; 1938, 209.
Page 283 :
Stereochemistry, , -, , 267, , propionic acid, using radioactive bromide, the corresponding ratios have been, found to be 1.82 + 0.19 and 2.03 ± 0.12.** In these displacements then, as, in the Kenyon-Phillips cycles, virtually every act of bimolecular substitution entails, inversion of configuration. Moreover, at present, this principle, is thought to apply, , to all Sn2 reactions on carbon., With the .SV1 mechanism, the story should be different. Suppose that a, substitution on an asymmetric carbon atom proceeds through a free carbonium, , in, , and suppose further that the bonds to the electron-deficient, , T—C, , ion,, , \, , VJ, , carbon atom lie in a single plane. We should then expect such a reaction to be, accompanied by complete racemization (unless, of course, there were an additional, asymmetric center not at the reaction site), for the carbonium-ion intermediate,, having a plane of symmetry, should not give rise to an optically active product, (p. 149)., Complete or almost complete racemization is sometimes observed in 6V1, reactions. For example, the hydrolysis of optically active cx-phenylethyl chloride,, CH3—CHPh—Cl, in 80 percent aqueous acetone proceeds with about 97 per¬, cent racemization.** />-Methoxybenzhydryl hydrogen phthalate (III) is one of, a special group of esters that suffer alkyl-oxygen bond breakage on hydrolysis, and solvolysis. If an optically active form of this ester is dissolved in methanol,, unimolecular solvolysis occurs and the methyl ether resulting (IV) is racemic.*7, , MeO, , CH'-o-C, , + MeOH, , III, active, , COOH, COOH, , deSircd *1™ ™r b' determined by diluting the, uring the radioactivity of the nonaoueous, experiments and the racemization exocrimr^t, displacement reaction, and the exchange, of the radioactive iodide. For a brief, H^ett, IW Organic, t7, see: Balfe, ^, , ^, , 3 nonpoIar solvent> then measancdin?, ^ata, both the exchange, Seated slightly by the reversibility of the, by the sl™ decaV, , McGr^^^, , anc* Scott, J. Chem. Soc., 1932, 1232, ^, , Fo^similarexamples,
Page 284 :
268, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , More commonly, however,, , 1 reactions are found to lead to a combina¬, , SN, , tion of in\ ersion and racemization; that is, the product has a configuration opposite to, that of the reactant, but its optical purity is substantially less. Typically, when optically, , active 2-bromooctane is dissolved in 60 percent aqueous ethanol, it is converted, to a mixture of 2-octanol and its ethyl ether. Some inversion has occurred in the, formation of both of these products, but both are about two thirds racemizedW, Yet we must be careful in interpreting such an experiment, for as the reaction, proceeds, Br~ is formed, and this can racemize the starting material (just as, I, , can bring about the racemization of 2-iodooctane by iodide exchange)., , Racemization due to bromide exchange is negligible near the beginning of the, reaction when (Br~) is very small, but it becomes important near the half-way, point in the solvolysis. It is possible (although the mathematical treatment is a, little complex) to estimate the specific rate of bromide exchange by following, both the appearance of H+ volumetrically and the disappearance of optical, activity polarimetrically during the course of the reaction. After corrections, are made for racemization due to bromide exchange, it turns out that, aside, from this effect, there is 32 percent racemization in the formation of the alcohol, and 26 percent racemization in the formation of the ether. Similarly, in the, , CH3, , I, , (CH3)2CH—(CH2)3 —c—c2h6, Cl, V, solvolysis of the tertiary chloride V in aqueous ethanol, 70 to 80 percent race¬, mization is observed/9 (Here, interpretation of the results is straightforward,, for this alkyl chloride is not appreciably racemized by 5V2-type chloride ex¬, change.) Likewise, when the optically active tosylates VI and VII are treated, with acetic acid, the corresponding inverted acetates are obtained, but 10, , ch3, C2H5—CH—OTos, VI, n Hughes, , /-Bu—CHPh—OTos, VII, , Ingold, and Masterman, J. Chem. Sac., 1937, 1196. The optical purity of the, , ether mayk determined by comparing the product from this reaction with the et er or, in the reaction of 2-bromooctane with OEt" in absolute ethanol. In the latter solutio ,, substitution is unmistakably bimolecular in character and may be assumed to ProoV, almost 100 percent inversion The optical purity of the alcohol is determined by comparing the, product obtained with 2-octanol that has beenresolvedbyconventional methods., P, , » Hughes, Ingold, Martin, and Meigh, Nature, 166, 679 (1950).
Page 285 :
Stereochemistry, , -, , 269, , percent racemization is observed in the first case50 and 90 percent racemization, in the second.5/, .The persistence of optical activity in the produc.ts from many reactions, that are unmistakably iSVl in character indicates that the intermediate, though, it may be electron deficient, is not satisfactorily described as a carbonium ion, unless definite restrictions as to its freedom are recognized. It is sometimes, said that the carbonium-ion intermediate may be shielded by the leaving group, so that the incoming reagent is more likely to attack the ion on the “unshielded”, side than on the “shielded” side. In effect, the carbonium ion, although it is, itself symmetric, may, at the moment of attack, be in an asymmetric environ¬, ment. Furthermore, the shield is not stationary; just after the bond breaking, has occurred, the leaving group lies close to the newly formed carbonium ion,, but as the carbonium ion grows older, the leaving group recedes and its shielding, action becomes less and less efficient. The longer the “lifetime” of the carbonium, ion, the more nearly free it can become before it is attacked by the entering, group. On this basis, we would expect those substrates that give the most, stable carbonium ions to undergo the greatest degree of racemization in their, substitution reactions. This is what we find), Although such a picture of the 6V1 reaction is easily visualized, we have, found it necessary to use in our description the terms “carbonium-ion lifetime”, and “shielding efficiency.” These terms represent quantities that are not easily, defined. An alternate description, which is most readily applied to solvolyses,, has recently been suggested.55 It is proposed that solvolytic substitution reac¬, tions proceed through an intermediate VIII that has much of the character of, an ion pair.55 The intermediate VIII may then react in one of two ways. The, , (inverted), S, , fC)-X, , s—CcY-.s, substrate, , VIII, , racemic, products, , Ix, , irot anTf' the Carb°nat°m in VIU, “-hybridize,” expelling the leaving, group and forming the .averted product; or an additional molecule of solvent, , j, , ,, , P aCG the leavin§ SrouP, converting VIII into IX, the solvated free, , ar onmm ion. Since IX has a plane of symmetry, it must yield racemic prod-, , (1952'd, , ', , ^(c)’ P- 641- See also Doering and Zeiss, J. Am. Cfum. Soc., 75, 4733, , negatively charged. If :.y iTunchametrh'ln!^™^3'' °nlV ‘f 'hC leavinS group, :X, is, leaving group, and solvem> ^ (og ^ by, carbonium ion.
Page 286 :
270, , Nucleophilic Substitution Reactions in Aliphatic Systems, , ucts. This picture allows the degree of racemization to be expressed in terms of, concepts that have been rigorously defined. The ratio of inversion to racemiza¬, tion should be equal to the ratio, , ki/kr,, , the relative rates of the competing reac¬, , tions by which intermediate VIII is consumed. This interpretation, however,, leaves unanswered the important question, “What is the nature of the process, by which VIII is converted to IX?”, In discussing the stereochemistry of nucleophilic substitution reactions, we, should not overlook a number of special cases in which net retention of configuration, is observed. In the predominant product resulting from such reactions, the, configuration about the reaction center is the same as that in the substrate. (In, such cases, some degree of racemization may also occur.) In order that a com¬, pound react in this manner, there must be, in addition to the leaving group,, a nearby substituent, often called a neighboring group, which has nucleo¬, philic character. This group may intervene in the substitution reaction, not, only affecting the stereochemical outcome, but also increasing the rate and, in, many cases, bringing about molecular rearrangements. Neighboring group, reactions are discussed at some length in Chapter 14; here we are interested, only in stereochemical matters. The classic example of neighboring-group par¬, ticipation is encountered in the basic hydrolysis of the a-bromopropionate ion,, CH3—CHBr—COO-.34 If an optically active form of this ion is hydrolyzed in, very concentrated base, the lactate ion, CH3—CHOH—COO , which is, formed has, as may be expected, an inverted configuration; here, a direct dis¬, placement reaction has occurred. However, if the hydrolysis is carried out in, dilute base, the lactate is formed with retention of configuration. In the latter, reaction, there are two acts of substitution at the asymmetric carbon, the first, bringing about an inversion, the second bringing about an inversion of the, original inversion. It appears that the slow step of the hydrolysis in dilute base, is the breaking of the C—Br bond, rapidly followed by formation of an a-lactone, X (first inversion), after which the lactone is broken by water (second, inversion).55, , •A, H, , Me, , H, , Me, first, , Br-, , ^C—Br, , HA, "O, , inversion, , Me, , Me, , > \/ H;°' *f°"d >\—oh2, inversion, , /, , H, _o—", , o, , O, , %o, , n Cowdrey, Hughes, and Ingold, J. Chem. Soc., 1937, 1208., « The rate law for the reaction will not tell whether the formation o, ring occurs at the same time as the breakage, broken first, since in either case the reaction should, shown, however; that the reaction shows a positive salt effect (see, , o, , _o, , 1, , m-mhered, e, , Winstein,
Page 287 :
Stereochemistry, , -, , 271, , A more recent example of neighboring-group participation is the reaction, of trans-2-acetoxycyclohexy 1 brosylate (XII) with acetate in glacial acetic acid., , As shown, a second acetoxyl group replaces the brosylate group. The product, is not the «V-diacetate (which we would expect if acetate ion attacked the ring, on one side while the brosylate ion departed from the other), nor is it a mixture, of the cis- and /ra/zj-diacetates with the former predominating (which we would, expect if the brosylate ion initially departed, leaving a “free” carbonium ion)., Instead, the chief product is the /ra/zj-diacetate (XIV), indicating that the, acetoxy group, like the carboxylate group, is capable of preserving the configuration, about a neighboring carbon atom while the latter is subject to nucleophilic attack.86 The, , proposed intermediate is the cyclic carbonium ion XIII, which may suffer, attack either at carbon-1 or -2. In either case, the W/zj-diacetate should result., (However, if the starting material is optically active, the resulting diacetate, should be, and is observed to be, racemic.), , J~Am. Chem. Soc., 70, 841 (1948)), suggesting that the formation of the activated complex, s, , ra,her ,han dispersal of char?e; that is-,hat, , particularly since* aS^nT0”*1 “ 1", has ”'ver b«=" plated and, more, Wold Rpf'i / \, r prepare “dactones have failed, some workers (see for example, Ingold, Ref. 1 („), p. 384) prefer to regard the intermediate in this reaction as a Type of, ~~o, , o=c, , \vH, — c', , XI, , XMe, in the correct configS’aaon'fOT'fubseauem™^ thc =arboxVlate group holds the carbonium ion, method for, by "aKr' At th' P^nt time, the lack of a, largely a semantic one., °ne and 2wlttenon intermediate makes this question, , retentionof configuration^its’hydrolysis must"involve“ *° account for ,h' overall, Ch” thofuyl'°XyBen bond- Sneh cleavage, fc™ rV ^ alkyl-oxygen bond rather, , •• wim,u'„s:mHtr, , rrrbP™, served for ordinary es,ers, Grunwald, and Ingraham, «*.,, S°‘ ’ 64' 2780. 27«7. ™ (1942). Winstein,
Page 288 :
272, , Nucleophilic Substitution Reactions in Aliphatic Systems, , Removal of the Leaving Group. Reactions in Nonpolar Solvents, A negatively charged leaving group, whether it arises from an .SV1- or .S^-type, reaction, is ordinarily solvated. Negative ions are most effectively solvated by, hydroxyl-containing compounds (which operate through the formation of, hydrogen bonds), and this is one of the chief reasons why the vast majority of, Sn reactions of halides and sulfonate esters are carried out in solvents consisting, , wholly or partially of water, alcohols, or carboxylic acids. When the hydroxylic, solvent is present in very large excess, ordinary kinetic studies cannot, of course,, detect its participation in a reaction. To obtain what is hoped will be a better, picture of the role played by such a solvent we may transfer the reaction to a, relatively inert solvent (such as benzene or CC14) and note the effect of adding, small measured amounts of a hydroxylic solvent in which the reaction is known, to occur readily. However, we should exercise considerable reserve in applying, conclusions based upon the study of a reaction in a nonpolar solvent to the, “same reaction” in a large excess of hydroxylic solvent; for the reaction environ¬, ments in the two cases must be very different. Any phenomenon involving ions, in a polar solvent may suffer modification when transferred to a nonpolar sol¬, vent, since single ions cannot exist in appreciable concentrations in the latter, but are associated to ion pairs, triplets, and higher aggregates., Consider the reaction between methyl bromide and pyridine in benzene., This is accelerated by adding small quantities of alcohols or phenols,sr although, the reaction will proceed at a measurable rate in their absence. Moreover, the, more acidic hydroxylic compounds (for example, jfr-nitrophenol) are the more, effective catalysts, presumably because they form the stronger hydrogen bonds, to the departing bromide ion. In a manner of speaking, the added molecules, of alcohol or phenol are “pulling” at the bromide while the nitrogen of the, pyridine molecule is “pushing” at the methyl carbon, and for this reason the, catalyzed reaction is sometimes said to proceed by a “push-pull” mechanism.5, The reactions of triphenylmethyl chloride in benzene are of consideiable, interest. This chloride undergoes chloride exchange with such benzene-soluble, ionic chlorides as Bu4N+C1- and (C18H37)2NMe+Cl- at a rate independent of, fhe concentration of chloride,59 indicating that, as in polar solvents, the rate« (a) Swain and Eddy, J. Am. Chem. Soc., 70, 2989 (1948). (b) Pocker, J. Chem. Soc., 1957,, 1279, , « This reaction and, by inference, a number of similar reactions were formerly sa^ a so to, be “termolecular,” largely because early data (Ref. 37(a)) indicated that they were third, order—that is, first order each in attacking reagent, substrate and pulling reagent, basis of more recent and more complete data by Pocker ( c ., ( ))> ‘ n°W,^y and hcnoi, reactions are, in general, of no particular integral order wit 1 respect o a e, fo seems, although they are indeed first order in both stacking reagent and subs rate. It here, adequate to describe these as S*2 reactions, proceeding with the help of pulling action y, alC°h«1(a°)r Sw'nan^Kreevoy, J. Am. Chem. So,, 77, 1122 (1955). (5) Hughes, Ingold, Mok,
Page 289 :
Removal of the Leaving Group. Reactions in Nonpolar Solvents, , 273, , determining step in such exchange is simply the ionization:, Ph3CCl -> Ph3C+ + ClOn the other hand, the methanolysis of Ph3CCl in benzene is first order in, methanol, provided the concentration of methanol is very low (below 0.001 M),, , and the methanolysis is considerably slower than chloride exchange/0 Here, then,, it appears that the reaction of the Ph3C+ ion with methanol is slow, compared, both to the ionization and to its reversal—that is, that the methanolysis sequence, is, fast, , MeOH, , Ph3CCl ^ Ph3C+Cl--> Ph3C—OMe + HC1, fast, , slow, , which would be consistent with the observed rate law., At higher concentrations of methanol, the kinetic picture seems to change,, for here the reaction order with respect to (methanol) rises from unity to well, over 2. Thus suggests the involvement of one or two additional molecules of, alcohol in the activated complex, and, once again, it may be supposed that these, serve to pull at the chloride while the first methanol molecule attacks the, phenylated carbon., f 3S t, , Ph3C—Cl + MeOH ^ Ph3C+ "Cl •, fast, , MeOH, , H—OMe->, slow, , Me, , +, , H, , \ /, O, , Ph3C, , + Cl- T HOMe, , There is another way in which the “pulling action” in halide substitutions, may be intensified. Dipositive mercury and unipositive silver are known to, e unusually electrophilic toward halide ions (aside from F~), and the substi¬, tution reactions of a number of alkyl halides are markedly accelerated by the, addition of silver or mercuric compounds-a “catalysis” quite obviously due, catal^^ 11C • PUl1 by thCSe heaV> mCtal ionS‘ However> the kinetics of such, _>ZCd reactlons maY become quite complex. In the course of a reaction, Patai, and Pocker, J. Chem Sor, , lQ<t7 ioas, , m-, , fr°m work described in four papers immediate^Ceding ™?marizCS the elusions arising, , ,aW, , authors is different from that proposed by Hughes el al (Ref, ProPosed bY these, this methanolysis becomes zero order rather thin, * (? f‘ .39£))» whose data indicate that, of the latter. This important discreoancv R I, f, $ °rfder’ 10 MeOH at low concentrations, studies in this system"The, carrying out kinetic, he reaction itself, but also by traces of moisture? « Catalyzed bY HCI produced (not only by, sW1 ^ ^ CarryinS out the methanolysis in the ZrZ a"d, attemPt to minimize this, showing also (in the opinion of the D™t“., PreSenfCe of a number of tertiary amines, significantly affect the methanolysis., h°r’ Satlsfactorily) that these amines do not
Page 290 :
274, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , catalyzed by Ag+, an insoluble silver halide will generally precipitate; and, since this halide is itself often a catalyst, the reaction will take on the complica¬, tions typical of heterogeneous reactions. With Hg2+, difficulties arise from the, multiplicity of species formed by dipositive mercury in the presence of halide, (Hg2+, HgA'+, HgA2, HgAj and HgA'f-). However, if the constants governing, the various mercuric-halide equilibria are known, kinetic treatments are possi¬, ble.Silver- or mercury-catalyzed substitution reactions, with but few exceptions,, have much of the character of >SV1 reactions (although since an additional species, is participating in the reaction, they are, in the strict sense, bimolecular)., Like Sjvl reactions, they occur most readily with tertiary halides and least, readily with primary halides; and like Sn^ reactions, substitution at an asym¬, metric carbon is attended by partial racemization, although, in the absence of, neighboring-group effects (p. 270), there is net inversion of configuration. Thus, such reactions may be considered to pass through an electron-deficient inter¬, mediate which, as in the case of Sn 1 reactions, we represent as a carbonium, ion., Ag +, , R—X, , slow, , »Agx+R+^Rr, , Hg.Y2, , and, , »HgXj + R+ ^ RK, , RA, slow, , In a similar manner, carbonium ions may be generated from alkyl halides by, the action of electrophilic metal chlorides in inert solvents. On this basis, we, can see why optically active a-phenylethyl chloride is racemized by action of, SnCl4 in CC14.", Ph‘ +, SnCU „, , ., , ^-MeCHPhCl-> SnClj +, , SnCU, CH3—C, , </,/-MeCHPhCl, , H J, , Steric Effects, We have already pointed out (and it will bear repetition) that atoms become, more crowded in the formation of the activated complex in an 5x2 rea, but become less crowiei during activation in an Sul reaction., reactions should be far more effectively hindered by the, , ere o >, rth, , groups. In fact, *,1 reactions would be expected to be subject, ", Ion (p 234) if they exhibit steric effects at all. The most straightforward co, l,usioPns concerning steric hindrance should arise from examining the influ, o Roberts and Hammell, J, (1948); Oae and VanderWerf, ibid., 75, 27, 4* Heald and, , Gwyn-Williams,, , j., '«« <1W);, (, )•, J. Chem. Soc., 1934, 3 ., , "" 7°’
Page 291 :
Steric Effects, , 275, , Df alkyl substitution on reaction rates, since the inductive effects of alkyl groups,, although by no means negligible, are generally much smaller than those associ¬, ated with other substituents (p. 205). For Sn reactions, we should be particuarly interested in the effects of successive substitution at positions a and 0 to, the reaction site; for it is now recognized that substitution at the y position and, fat positions still farther removed have, except for special cases, only slight, effects on reactivity., One of the simplest possible nucleophilic substitutions is the replacement, of one halogen by another in an alkyl halide (the so-called “Finkelstein, reaction”).43, RX + Y~ —* R Y + X-, , Such a reaction can be followed kinetically, even if the attacking and leaving, groups are chemically the same, since radioactive isotopes of chlorine, bromine,, and iodine have been available for some years. The reaction series of greatest, interest to us are, R. = Me, Et, f-Pr, /-Bu (successive substitution of CH3 groups for a-hydrogens);, and, R. = Et, rc-Pr, f-Bu, neo-Pe (Me3C—CH2—), (successive substitution of CH3 groups for /3-hydrogens), In comparing rates of reactions of halides in these series, we must be sure, hat each halide reacts by the same mechanism, frhe reactions are therefore, -arried out in the poorly ionizing solvent, dry acetone, in which even the f-butyl, aalides undergo substitution by the Sn2 mechanism. Relative specific rates, [with k2 for the methyl halide in each series set at unity) for two types of halide, -xchange are given in Table 8-2. We should remember that these trends are, lot due solely to steric effects. When the reactivities of molecules having different, masses are being compared, ponderal effects must enter the picture. These it will, remembered (p. 235), depend exclusively upon the distributions of masses in, the reactants and in the activated complex. These effects operate appreciably, mly on the entropies of activation, almost invariably decreasing the rates of, pactions of heavier compounds in comparison to lighter compounds. Ponderal, : ects that cannot yet be measured must be estimated theoretically 43«» Thus, -eactTonU!hrnlhaVef, , reP°rted, , ** bro“de-bromide exchange, , js re“ rd;‘ “ ,H, °rkeXaT ’ that the br°mide CXChange of '-butyl bromide, n era y ,y a factor of about 10 (as compared to the reaction of, sy ClifSTaW, 'o-workers in eight additional papers ./. Chem .W, «ary, see Ingold, Quart. Revs., XI, 1 (1957)., ’’, , lQ^^iVo, , hKaVC b“" summar'2ed, by de la Mare a^d, , ’ 3169 3236, ^ For a recent sum-
Page 292 :
276, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , Table 3-2. Relative Rate Constants for Sjv2 Halide Exchanges, in Acetone (25°), R =, Me, , Cl- + RI, , RC1 + I", , Br* + RBr —> RBr~ + Br~, , 1.00, , 1.00, , 0.089, , 0.013, , 0.0028, , 1.4 X 10-4, , /-Bu, , —, , 3.9 X 10-5, , zz-Pr, z-Bu, , 0.053, 0.0034, , neo-Pe, , 1.2 X 10-6, , Et, z’-Pr, , 0.0085, 4.4 X 10-4, 2 X 10~7, , methyl bromide) and that in neopentyl bromide the exchange is retarded, ponderally by a factor of about 100., , It is also likely that inductive effects play a small part in determining the:, relative rates in the “a series” above (Me, Et, z'-Pr, and /-Bu), but not in the, , “/3-series” (Et, zz-Pr, z-Bu, and zz^o-Pe). It is difficult to say a priori how large such i, , effects are, but de la Mare and co-workersW6) estimate that each a-methyll, group inductively retards bromide-bromide exchange by a factor of about 5., While the magnitude (and, perhaps, even the direction) of this effect may be, , open to question, it is obvious that inductive effects, like ponderal effects,,, account for only a small measure of the observed differences in rate (6 to 7, powers of 10) recorded in Table 8-2., Figure 8-1 is a sketch of the activated complex for the SN2 bromideexchange reaction of 2-bromobutane (chosen because it has both or- and /3-methyl, , groups). We see that the configuration about the carbon atom under attack is:, such that the a-methyl group tends to be held out of the way of the very, large bromide ion but that substantial overlap is possible with the /3-methyl, group. If there is only one /3-methyl group, as in zz-propyl bromide, this inter-, , Fig. 8-1. Activated Complex in SN2 Bromide Exchange.
Page 293 :
Steric Effects, , 277, , ference may be avoided by rotating the Ca—bond, thus pushing the /3-methyl, group “out of the way” of the bromide ions. If there are two /3-methyl groups,, both will try to avoid interfering with the bromide ions, and serious restrictions, will be placed upon the freedom of rotation of the carbon-carbon bonds in the, transition state. These restrictions in freedom correspond to a more negative, entropy of activation, which corresponds, in turn, to the low rate of substitution, that is observed.44 With three /3-methyl groups (neopentyl bromide) there is no, way in which such interference may be avoided; the reaction requires attack by, one of the small number of very energetic Br~ ions in solution and therefore it, will be very slow indeed. Neopentyl halides are, in fact, extremely inactive, toward nucleophilic reagents in bimolecular substitutions. (They will react at, moderate rates, albeit with extensive rearrangement, under conditions that, favor unimolecular substitution.) Broadly speaking, the trends which we have, discussed for halide exchanges are observed for other series of ^2 reactions., Bimolecular substitutions may also be retarded when the steric require¬, ments of the attacking reagent become excessive. We may show this by selecting, a series of amines or alkoxide bases having varying degrees of branching and, using these as attacking reagents toward a common substrate. Below, for ex¬, ample, are listed the relative specific rates45 (in nitrobenzene at 25°) for the, reaction of some substituted pyridines with methyl iodide:, , n, , N, , + CH3I, , N—CH<, =/, , +, , I, , Me, , X, ST, , Me^, , 1.00, , 2.27, , pyr, , Me", , Me, 0.042, , CHMe2, , 0.072, , "N', , "CMe3, , 0.0002, , slower than those ohtVnMUd^b^au^he^Mt" ^0' 0"3 °h "‘’?i0pyl halkl'-'S were slightly, -te for the propyl ootids, «*—*», (See, for example, Denbar and Hammett T A VA, c ~ those for eth>rl compounds,, calculations byP* ,a Mare and co w^ers (ReT 4UM ^’y'09 (195°>>. ‘he, activation enti npies between the ethyl and Dronvl ™, ’ Pj 3232 Suggest that dlfferences in, m origin., 6 Cmyl and Pr°Py1 impounds are ponderal, rather than steric,, “ Br°Wn' GiMis" and P°da". J- Am. Chem. Soc, 78, 5376 (1956).
Page 294 :
278, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , Substitution of a methyl group in the 4-position of the pyridine ring in¬, creases the rate, for the electron-repelling effect of the methyl group has made, , the nitrogen atom more nucleophilic. However, substitution of a methyl group, in the 2-position, or, more particularly, methyl groups in the 2- and 6-positions, retards the substitution. This effect is obviously steric in origin, pointing to, , interference between the hydrogen atoms of the ring methyl and those in the, substrate (Fig. 8-2). Moreover substitution of the very bulky f-butyl group in, the 2-position lowers the rate of substitution by a factor of 5000., , Fig. 8-2. Activated Complex in the Pyridine-Methyl Iodide Reaction., , Similarly, quinuclidine (XV) has been found to react with methyl iodide, over 50 times as rapidly as triethylamine.^ The two amines have similar, , XV, , basicities toward H+, but in quinuclidine the substituents on the nitrogen atom, are held back, allowing no steric interference with the substrate. The difference, is even more striking if a more hindered substrate is used. With isopropyl iodide,, quinuclidine reacts over 700 times as rapidly at Et3N., Let us turn now to unimolecular substitution. To begin with, we can do, little more than speculate as to the steric effects of a-alkyl substituents in the, substrate. We are, of course, very much aware that such substituents great y, accelerate Sivl-type reactions, for it is the presence or absence of alkyl groups, on the a-carbon that determines whether we are dealing with a reaction on a, primary, secondary, or tertiary carbon. Yet, we cannot say how much of this, accelerative effect is to be attributed to induction, how much to hyperconjug, tion, and how much, if any, to steric acceleration. With 0 substitution t le pic, Bi own and Eldred, J. Am. Chem. Soc., 71, 445 (1949).
Page 295 :
Steric Effects, , 279, , is a little clearer. Successive incorporation of alkyl groups at the f3 position, generally results in small, and apparently nonsystematic, accelerations of SN1, reactions until the substrate becomes very crowded indeed, whereupon reaction, rates increase sharply, possibly by several powers of 10. Thus substitution of, one, two, and three methyl groups for /^-hydrogens in /-butyl bromide increases, the solvolysis rate in 80 percent aqueous ethanol at 25, , by 78, 22, and 68 pei-, , cent, respectively.^7 However, the solvoly&es of such crowded chlorides as XVI, and XVII appear to be several thousand times as fast as that of /-BuCl under, the same conditions.^ Although the fact that rearrangements accompany the, solvolyses of the very crowded halides has led to some apprehension,^ it would, CH3v, , /CH3, , N /, , CH3, , CH, , ch3x, , Cl, , CH3, , I, , ch3— c—c-, , ch3—c—c—c—ch3, , -Cl, , ch3/, CH, CH 3, , ch3, , ch3, , ch3, , CH3, XVI, , XVII, , seem that we almost certainly have here an example of steric acceleration; that is,, a reaction in which formation of the transition state eases the crowding that, prevailed in the reactant and which therefore is assisted by a maximum of, crowding in the reactant., Substitution reactions on cyclic systems were discussed briefly in the pre¬, ceding chapter (p. 241), but there is one type of cyclic system that merits, special consideration. Compounds such as apocamphyl tosylate (XVIII) and, 1-iodotrypticine (XIX) (p. 280) with leaving groups on “bridgehead” carbon, atoms are especially inactive to nucleophilic substitution, both by the unimolecular, and bimolecular mechanisms. Tosylate XVIII is inert to lithium iodide in boiling, , acetone (a powerful ^2-type reagent), and the corresponding chloride is inert, both to silver nitrate (an 5*1-type reagent) at 160° and to concentrated alcoholic, *7 Ingold, Ref. 1(a),, , p., , 414., , !“1, formed by, , 'w'Th, , ““f"', , °T m°r' addi<‘onal eWorld!, , carried out on mixtures of chlorides and since each'chfon^eln the8!"*.'" ^, were, reacting in more than one, ^ the mixture may have been, although we may still be quite sure tlm eaSToAhe rh!^^ S.hould be accePtcd with reserve,, many times as rapidly as /-butyl chloride., 011C < s in each of the mixtures reacted, ** See, for example, Bartlett, J., ^treitwieser, Ref., , 1, , Chem Ed, , (c), pp. 709-710, also, , 30 77, , discusses^his probRm., , . r, , , ,, g<, , ’ *** 1 (a)’ P‘ 417‘
Page 296 :
280, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , KOH (an ^2-type reagent).50 Iodide XIX, and the corresponding bromide, also, are unreactive to concentrated alcoholic NaOH, Na2S, and silver nitrate, in boiling ethanol,'51 being thus inert both to SN1- and ^2-type reagents., , I, , XVIII, , XIX, , XX, , The striking lack of reactivity of compounds of this type shows how necessary, it is that the bonds to a carbon atom on which a nucleophilic substitution is, occurring have substantial freedom of motion. In XVIII and XIX the bonds, to the a-carbon atoms are held rigidly in place by the cyclic structures. Since, Walden inversions about such carbon atoms cannot occur, Sn2 reactions at, these carbons are virtually impossible., •SV1 reactions are presumed to proceed through carbonium ions. Such ions, are most stable when the three bonds to the positive carbon atom lie in the same, plane, and their stability falls off sharply as departures from coplanarity increase., The bonds to the bridgehead carbons in XVIII and XIX are obviously not, coplanar, and even an approach to coplanarity is prohibited by the rigidity of, the bicyclic systems. While it is not impossible to convert such compounds to, carbonium ions, such ions should be of very high energy and should form very, infrequently. One would then predict that, with persistence, it should be possible, to force such compounds as XVIII and l-bromobicyclo(2.2.2)-octane (XX), (in which the bicyclic system is slightly more flexible) to undergo SN\ reactions,, although extremely sluggishly. Actually XX reacts very slowly with aqueous, AgNOg to give the corresponding alcohol,5* but the more rigid trypticyl halides, (for example, XIX) will not react, even under the most severe forcing conditions., , Structure and Reactivity; Further Considerations, Thus far a number of relationships between structure and reactivity have, emerged in our discussion of nucleophilic subsdtutions. We have learned that, .Sjvl reactions are accelerated by a-alkyl substituents, and, more strikingly, by, 60, 51, , Bartlett and Knox, J. Am. Chem. Soc., 61, 3184 (1939)., 1005 (1950); Bartlett and Greene, ibid.,, Bartlett and Lewis, J. Am. Chem. Soc., 72,, , 1088 (1954)., 1008 (1953)., e! Doering, Levitz, Sayigh, Sprecher , and Whelan, J. Am. Chem. Soc., 75,
Page 297 :
Structure and Reactivity; Further Considerations, , 281, , a-aryl and a-vinyl substituents; but that Sn2 reactions are retarded by a;-alkyl, substituents and also by /3-alkyl substituents. We have learned that the rate of, an Sn2 reaction may generally be increased by increasing the basicity of the, attacking reagent, that the rate of an, , SN1, , reaction is necessarily independent of, , the attacking reagent, but that both types of reaction are accelerated if the, basicity of the leaving group is decreased. With respect to the reactions of, halides and sulfonate esters, we have learned that SnI reactions are favored by, increasing the ionizing power of the solvent, but that the solvent effects in an, Sn2 reaction will depend upon the charge on the attacking group. | Finally,, we have learned that Sn2 reactions are completely prohibited and SnI reactions, very seriously retarded if the leaving group is attached to a bridgehead carbon., The accelerative effect of an a-aryl group on an 6V1 reaction has been, attributed to the ability of the benzene ring to stabilize the partial positive, charge in the carbonium-ion-like transition state by delocalization (XXI'). We, should then expect, and we find, that substituents on the benzene ring which, further increase the effective length of the conjugated system, either by electron, donation (XXII) or by hyperconjugation, , (XXIII), , will further facilitate, , unimolecular substitution. Attempts have been made to put the accelerative, , CR,-X', , XXIII, (or retarding) effects of substituents on a more quantitative basis. The Hammett, equation (p. 220) has been applied to the alcoholyses of ring-substituted benzhydryl chlorides and triphenylmethyl chlorides,ss but with only moderate, success However, it has been shown that the rates of hydrolysis of ring-subsututed a-phenylethyl chlorides in 90 percent aqueous acetone can be correlated, , ya, , ammett-hke relationship, using, instead of the, , ordinary rvalues. Brown’s, , intended fo', T ^ TT*1"6 that this shou!d be *>. for the <r+ values are, r reactions in which the transition state is positively charged, he accelerating action of a-halogens and a-alkoxide groups on uni, tnolecular substitution may be similarly explained. Again, Jlectl den^, Forfurthe^referenMs.'^e'jaffc'am'SyM S"";.'7', “ Bro- “0 Okamolo,, , j. AlcZ'slt 79,, , Chtm' Soc-’ 7b 3649 (1952)., , ^ <1953>-
Page 298 :
282, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , drifts from the halogen or oxygen atom toward the partially positive carbon, atom in the activated complex, stabilizing the latter., \ :C1:CR+ —X~<—> :G1: :CR2 -Z-l and, , [R:O:CR£—<—>R:0: :CR2--X~\, ••, , ••, , J, , As with the reactions of substituted benzene derivatives (p. 218), the “+R”, effects of the halogen and alkoxide substituents are opposed by their respective, "—I” effects, (that is, electron withdrawal by induction), which, if present, , alone, would be expected to retard carbonium ion formation.55 If a halogen or, alkoxide substituent is attached to a /3-carbon, its conjugative effect may no, longer operate but its inductive effect may persist. Therefore such substituents,, except for cases involving neighboring-group participation (Chap. 14) will, retard unimolecular substitutions., , Perhaps the most useful generalization concerning the relationship of, structure to ease of unimolecular substitution reactions is this. In forming the, activated complex, the leaving group has begun to depart with its pair of bond¬, ing electrons; the reaction will therefore be facilitated by any feature that in¬, creases the electron density at the carbon constituting the reaction site. We, obviously cannot apply this rule to bimolecular substitutions, for electron den¬, sity is being supplied to the reaction site by the attacking reagent at the same, time that electron density is being removed by the leaving group. For any, particular reaction, we cannot compare the quantity of electron density that, has been supplied with that which has been removed at the instant the system, passes through the transition state. We therefore cannot say whether the carbon, at the reaction site is more positive or more negative in the activated complex, than in the reactant; consequently an Sn2 reaction might be accelerated or, retarded by electron-rich substituents on the substrate., However, this much is certain; if, bound to the a-carbon, there is a group, that can delocalize either positive or negative charge, such a group should, facilitate bimolecular substitution, irrespective of the extent of bond making, and bond breaking that has transpired in the transition state. Both the phenyl, and vinyl groups are of this type., « Although insufficient data have been obtained to allow a clear decision iit appea.rs that, unlike the heavier halogens, the -/effect of the a-fluoro group outweighs its +^eff^tinsofa, as these effects influence unimolecular substitution. It has been found, for examp e (, as these el, 74 3182, PhCFjBr loses Br" in 50 percent aqueous ace-, , (1952)) that, , £, , »d«iphCH,Br. This is ir.contrast to Che, , substituent on reactions of benzene derivatives, where,, , it w.U ItV tte substituent., , '.he'ouorc substituent is only ear — moved from the reaction site.
Page 299 :
283, , Structure and Reactivity; Further Considerations, , Hence,, , both, , benzyl, , and, , allyl derivatives, , undergo, , Sx2, , reactions, , more, , readily than saturated compounds of an otherwise similar nature. Roughly, speaking, the Sx2 reactions of allyl compounds are ten to one hundred times as, rapid as those of the corresponding ethyl compounds at or near room tempera¬, ture, whereas those of benzyl compounds are one to several hundred times as, rapid.55, By noting the effect of electron-attracting and electron-donating substit¬, uents on the rate of Sx2 reactions of benzyl derivatives, we should be able to, determine whether the carbon atom under attack in a particular type of reac¬, tion acquires a partial positive or partial negative charge in the activated com¬, plex. To date, too few reaction series have been studied in this way to allow, a definite conclusion, but for benzyl halides it appears that the a-carbon atom, will become more positive in the transition state (that is, the reaction will be, retarded by electron-withdrawing groups) if the attacking reagent is uncharged., If, however, the attacking reagent is negatively charged (OR~ OH~ I~, etc.),, the a-carbon becomes more negative in the transition state and the reaction, is facilitated by electron withdrawal. This is reasonable, for the activated com¬, plex in the attack by a neutral molecule has much of the character of an ion, pair;, Ar, Y:, , +, , ArCH9X-, , Y-C-—X, , -, , H, , /\, , [lr-CH2Ar]+, , + X, , H, , whereas if both the attacking and leaving group arc negatively charged it is, , hi™™, , of ,he excess negative char?e per“^, , The incorporation of an a-carbonyl, a-carbethoxy, or a-cyano group into, somedmesTuitrst'k''’1, SUbstitution' This rate enhancement is, q, mg’ as ln the halide exchange reaction of alkyl chlorides, or recent compilations, see Streitwieser, Ref. 1(e), pp, 584, 585, 591.
Page 300 :
284, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , with KI in acetone. The following relative rate constants (w-PrCl, T — 50°) have been observed:57, , 1.00,, , CICHoCOOEt, , ClCH2COPh, , ClCH2COCH3, , C1CH2CN, , 1600, , 100,000, , 33,000, , 2800, , This effect is surprisingly consistent, being independent of the charge on the, attacking reagent. Of a number of explanations that have been suggested, per¬, haps the most satisfactory is as follows:55 It is proposed that there is appreciable, attraction between the electron-rich attacking reagent and the electrondeficient carbon of the carbonyl (or cyano) group. Put a little differently,, there should be significant overlap between the bond-forming p orbital of the, attacking atom and the electron-deficient tt orbital of the carbonyl carbon., This interaction constitutes partial bond formation between the attacking atom, and the carbonyl carbon (XXIV7), and helps to stabilize the activated com¬, plex, thus speeding up the substitution. On the other hand, if there is a nonY, i, , \!, c-c=o, , /!, i, , X, XXIV, , <-, , \ /, , \, , c-c=o, , /!i, X, XXIV', , participating halogen atom bound to the a-carbon atom in the substrate, it, should repel the attacking reagent (since both species are electron rich), thus, lowering the stability of the activated complex and retarding the bimolecular, substitution.55 By the same argument, we would predict that an a-oxygen, atom, as in an alkoxy or acyloxy group, would also retard Sn2 reactions. Al¬, though less than a dozen cases have been reported, it appears that here our, predictions are seriously wrong, for these groups are found to increase the rates, of such substitutions, sometimes by several powers of ten.37,50 The following,, for example, are relative rate constants (n-BuCl = 1, T = 50°) for chlorideiodide exchange in acetone:, MeO—CH2C1, , AcO—CH2C1, , PhCOO—CH2C1, , 900, , 270, , 60, , CBuCl, , " Conant, Kirner, and Hussey, J. Am. Chem. Soc., 47 488 (1925)., « See (a) Bartlett, in Gilman’s Organic Chemistry, Vol. Ill, John Wiley and So, New York, 1953, p. 35; (b) Streitwieser, Ref. 1(c), pp. 597-601., 59 Hine, Thomas, and Ehrenson, J. Am. Chem. Soc., 77 3886 (, )., to Ballinger, de la Mare, Kohnstam, and Prestt, J. Chem. Soc., 1955, 3641, K, Am. Chem. Soc., 48, 2945 (1926)., , ,, y
Page 301 :
Structure and Reactivity; Further Considerations, , -, , 285, , The rate-enhancing action of a-oxygen atoms has been rationalized in several, ways, but none appears to be completely convincing.61, Somewhat surprisingly, 6V1 reactions may be substantially retarded by, substitution of deuterium atoms for hydrogen atoms at one or more /9-carbons., Typically, the following deuterated compounds undergo solvolysis more slowly, than the corresponding nonlabeled compounds by the factors indicated.6*, , Cl, , Cl, , Cl, , I, (CD3)2CCD2CH3, , Me2CCD2CH3, , Me2CD—C(CH3) 2, , 0.43, , 0.71, , 0.78, OTos, , Br, EtCD2—CH—CD3, , EtCD2—CH—CD3, , 0.72, , 0.64, , kH, , These isotope effects suggest that somehow the /9-hydrogen atoms are, involved in the rate-determining ionization step. Since the observed effects, are relatively small and since we know that the C—H bonds are not broken in, Sn, , reactions, we may infer that the effects arise from stretching of the 0-C—H, , bonds in the activated complex. This fits in with our feeling that an aliphatic, carbonium ion, and an activated complex having some of the character of such, an ion, may be stabilized by hyperconjugation in which the /9-C—H bonds, assume some degree of “no-bond” character (XXV—XXV'). Since /9-C—D, , r h, I, , h+, , -i, , I, , —C—C+-X- <-> — C=C-X-, , iXXVi, , J, , XXV', , bonds are somewhat “tighter” bonds than 0-C—H bonds (p. 192), we should, expect substitution of deuterium for d-hydrogen to lower the extent of stabiliza¬, tion by hyperconjugation in the activated complex, boost the energy of activa¬, tion, and thus retard the reaction. Similarly, the solvolysis of chloride XXVI, cause",L0ox7gTitombou^t„er,heRef' 'P’^, ^ ,Ws ^deration occurs be¬, ing .he parS, charce ™ ,tr;e, >0" t'lpS ‘° S‘abiUze 'he aCtivated “"V1'* bV dispersto that used to explain the accelerating effector, Th‘S rCason Seems closel>' similar, and should be applicable onlv if the ?, i, °i ^ a koxide SrouP on SV1 reactions (p. 282),, charge during activation In the view of the" ^ aCqmr.ed an additional measure of positive, ^particularly if the att^^ntt a^"^ ^ C°nditi°n ^, , ^ ^, , -srre,hano1 at 25°(see, , volyzed in formic acid at 98° (see Lewis and Rn, hi", aSt tW° cornPounds were solapproach to this problem,, ^ ™ (1954))-For a more detailed, ’ agow, rahey, and Suzuki, ibid., 80, 2326 (1958).
Page 302 :
286, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , in acetic acid at 50° occurs only nine tenths as rapidly as that of the correspond¬, ing nonlabeled chloride, suggesting that hyperconjugated structures of the, type XXVII stabilize the transition state in this solvolysis.®3 As expected, the, , XXVI, , XXVII, , corresponding deuterium isotope effect for kSat2 reactions is much smaller. The, specific rate of reaction of f-PrBr with NaOEt is only 13 percent larger than that, for the same reaction of (CD3)2CHBr.®'{ This difference may be due merely to a, ponderal effect., , Allylic Halides. Allylic Rearrangement and Internal Return, It has been remarked that a carbonium ion derived from ionization of an allylic, substrate may be represented by a pair of structures (XXVI <-> XXVI'),, , Rx, , R\, , c=c-c, H 7, , n H, , +/H, , h, XXVI, , “, , H, , ", , /H, , /c-?=cv, , H, , H, , H, , XXVI', , o, , 0/ H, , U^.C—c, u'O, , h, , XXVII, , indicating that the positive charge is distributed over a x-orbital embracing, three carbon atoms (XXVII), rather than being localized on the a-carbon., In a substitution reaction proceeding through such a carbonium ion, the in¬, coming nucleophile may attack either at the y- or the a-carbon. If attack is at, the a-carbon an “ordinary substitution product,” of course, results, but if the, attack is on the y-carbon, the reaction becomes an allylic rearrangement., In the latter case, the double bond shifts and the incoming group takes a posi¬, tion two carbons removed from that originally occupied by the leaving group. », Typically, the solvolysis of crotyl chloride (XXVIII) in “50 percent, aqueous acetone” (with a little CaCOi added to remove the HC1 formed) yiel s, a mixture of two isomeric alcohols., « Lewis and Coppinger, J. Am. Chem. Soc., 76, 4495 (1954)., Shiner, J. Am. Chem. Soc., 74, 5285 (1952)., Young Chem. Revs., 56,, 65 For a recent review of allylic rearrangemen s, see, unpublished experiments by, 784-801 (1956). Included in this are accounts of a number of unpublished, pe, Young and co-workers, to which reference is made in the present text.
Page 303 :
Allylic Halides. Allylic Rearrangement and Internal Return, , 287, , CH 3—CH=CH—CH 2OH (56 percent), CH3—CH=GH—CH2C1, , water-acetone, ->, , CH3—CH—CH=CH2, , (44 percent), , 47°, , OH, XXVIII, Similarly, if the same chloride is solvolyzed in boiling ethanol (again in the, presence of CaC03), a mixture of the two corresponding ethyl ethers is obtained,, 91 percent of which is the ether of crotyl alcohol and 9 percent the rearranged, ether, CH3—CH—CH=CH2.65 Now, if the solvolysis of a-methylallyl chloride,, OEt, CH3—CHC1—CH=CH2, proceeds through the same carbonium ion inter¬, mediate as does crotyl chloride, we would expect to get the same pair of products, (and in the same proportions) from both solvolyses, since the fate of an inter¬, mediate species should be independent of its mode of birth. In actuality, the, pairs of products obtained from the two isomeric chlorides are the same, but, the ratio of products varies. Of the mixture of alcohols from the hydrolysis of, , a-methylallyl chloride in water-acetone, 43 percent (rather than 56 percent), is crotyl alcohol (that is, the so-called product spread for the solvolysis of this pair, of halides is 13 percent). Likewise, of the mixture of ethers resulting from the, ethanolysis of this chloride, 53 percent (rather than 91 percent) is ethyl crotyl, ether; that is, the product spread is 38 percent. As a rule, similar variations, are also observed in the compositions of mixtures obtained from other pairs of, isomeric allylic halides. Moreover, such variations are consistent in character;, a mixture resulting from the solvolysis of a primary halide has a greater pro¬, portion of “primary product'5 and less “nonprimary product” than has the, mixture resulting from the solvolysis of the isomeric nonprimary halide. Thus,, in the solvolysis of such a pair of isomeric halides, it appears either that the, intermediates in the two cases differ in some respect or that the solvolysis of one, or both halides is proceeding partially by some other path that circumvents, the common intermediate. Significantly, the product spread in such a pair of, solvolyses may generally be reduced by increasing the ionizing power of the, medium, and further reduced by addition of Ag+., The interpretation of these facts depends upon how we wish to regard, borderline substitutions (p. 258). If we take the view that they are reactions, n which some of the individual acts of substitution are unimolecular and the, lions"h I, **a, ’ We may C°nsider the solvol>'ses of allV>ic halides as competi¬, tions between direct displacement by solvent (in which there should be no rengetnent) and substitution via preliminary ionization (in which rearrange~y rr) We may, thC «“* °f bi—ular substitution by, easing the ionizing power of the solvent or by adding an electrophilic
Page 304 :
288, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , catalyst. In doing so, the extent of solvolysis through preliminary ionization is, increased; in the extreme case, the solvolysis may become completely unimolecular in character, and the product spread should disappear.65 The ratio, of products then represents the relative rates of solvent attack at the a and, 7, , positions of the carbonium-ion intermediate., In referring to such a “carbonium ion” we must not forget that it is not a, , free cation,, , but rather that it constitutes part of an ion pair. At present,, , it appears that this type of ion pair is somewhat unusual, having the halide, ion located equidistant (or nearly equidistant) from the a- and, , 7-carbons, , (XXIX). The evidence for such an arrangement stems largely from a phenomeC1, \, / \\, , H3C, \ T /, , CH, , \, , \, , \, , \“, rCH,, , CHXXIX, , non, first reported in 1951, now called internal return.67 When a,a-dimethylallyl chloride (XXX) is subjected to solvolysis, either in acetic acid67 or in, aqueous ethanol,66 a portion of the halide becomes converted to its isomer,, 7,7-dimethylallyl, , chloride (XXXI)., HO Ac, , ch2=ch—c(ch3)2ci, , (or Ti2k^-£TUri], , (CH3)2C=CH—CH2C1, , XXX, , XXXI, -f- solvolysis products, , The rate of the isomerization is proportional only to the concentration of the, original allylic chloride. This at first suggests that the isomerization might be, due to a rapid attack of Cl~ on a carbonium ion that is formed in a slow step;, but if this were the case, the solvolyses reactions themselves would exhibit, mass-law effects (p. 256) when carried out in the presence of extra Cl". No, such effects have been observed. Moreover, when the isomerization is carried, out in the presence of excess radioactive Or, it is found that the isomerization, proceeds much more rapidly than the incorporation of labeled Cl, , into the, , organic molecule. Evidently, the Cl" that departs from the a-carbon atom is,, M For example, no significant product spread is observed in the, CHCH.C, and, NfcCCCH^CH, in-or£, atoms -ha. he,ps accommodate .he, positive charge in the carbonium ion is tertiary., n951\, «r Young, Winstein, and Goering, J. Am. Chem. Soc., 73, 1958 (1, )., cs de la Mare and Vernon, J. Chem. Soc., 1954, 2504.
Page 305 :
AUylic Halides. Allylic Rearrangement and Internal Return, more often than not, the Cl" that becomes bound to the, , 7-carbon, , -, , 289, , atom in the, , same molecule; that is, a chlorine is more likely to change its location laa given, molecule than to break free and become equilibrated with Cl, , in solution., , Intermediate XXXII is consistent with this evidence, for in this ion pair the, Cl- has taken a position favorable for the required migration., , /, , CH2, , \, , \, , /CH3, CH-C, a, , ^ch3, , XXXII, Granted that allylic isomerizations proceed through a “bridged” ion pair, such as XXIX, what is the evidence that solvolyses (with and without allylic, rearrangement) pass through the same intermediate? Let us, for an instant,, suppose the contrary—that solvolyses of allylic halides proceed through some, intermediate other than XXXII, perhaps a solvated carbonium ion. Whatever, the nature of this intermediate, its formation requires separation of unlike, charge, just as does the formation of XXIX and XXXII. It should therefore, be formed more readily in dry EtOH (dielectric constant 24) than in dry, HOAc (dielectric constant, , 6);, , as expected, the solvolyses of chlorides XXX, , and XXXI are somewhat faster in ethanol than in acetic acid. But, it has been, observed that the isomerization of XXX to XXXI does not take place in dry, ethanol.65 It is difficult to believe that this is because the “bridged” ion-pair, intermediate, XXXII, cannot be formed in dry ethanol even though it is formed, in dry acetic acid and in wet ethanol. Rather, it seems much more likely that, XXXII is also formed in dry ethanol, but that very shortly after its formation it is, converted by ethanol to one of the two possible isomeric ethyl ethers—that is, that sol¬, , volysis in dry ethanol has proceeded through XXXII. We may also then infer, that solvolysis in wet ethanol and in acetic acid proceeds through XXXII,, for there seems no reason why the bridged ion pair should be destroyed in dry, ethanol but not in other hydroxylic solvents. It is appreciated that this evidence, is indicative rather than conclusive. The question is by no means closed., Internal return has been observed in a number of additional systems, among them the derivatives of 4-methylcyclohexene.66 For example, if one of the, enantiomorphs of the />-nitrobenzoate XXXIII is dissolved in 80 percent, Both°of T°ne’ U Underg°eS " C°mbination of inte™al return and hydrolysis., XXXIV tT ?Ctl7 PrOCeed thrOUgh the ^^tric bridged ion pair,, XXXIV. Therefore, the products from both reactions must be racemic l’, "Cocrmg and co-workcrs;, , J. Am., , Chm. See.,, , 77,, , 1129, 5026, 6249 (1955).
Page 306 :
290, , Nucleophilic Substitution Reactions in Aliphatic Systems, , breakage at (1) or (3), , racemic, XXXIII, , CH3, XXXIII, , this case, the rate of racemization due to internal return is very nearly half the, rate of solvolysis. On the other hand, when the corresponding chloride (trails5~methyl-3-chlorocyclohexene) is solvolyzed in acetic acid, the rate of racemiza¬, tion is over four times the rate of solvolysis. As expected, internal return may, compete with solvolysis most effectively in solvents that are only weakly, nucleophilic., From the discussion of allylic rearrangements thus far, it might be sus¬, pected that such arrangements may occur only through ionization. This was, thought to be the case until 1949-1951, during which period it was shown that, there are a number of substitution reactions of allylic compounds, distinctly, bimolecular in character but nevertheless involving allylic rearrangement.70 One, , such reaction is that between diethylamine and a-methylallyl chloride, first, order in both reagents, yielding the crotylammonium derivative, XXXV., CH,, , CH3, , H, , I, , „—^ i, CH2=CH-CH-C1, , X, , Et2N-, , I, , cr, , Et9NH—CH,—CH=CH, , XXXV, , This rearrangement is not due to a preliminary isomerization of the staiting, chloride, for although the chloride is known to isomerize, it undergoes the, observed substitution reaction far more rapidly. Furthermore, the reaction is, not a normal displacement followed by a rearrangement of the ammonium, H, salt, CH2=CH—CHCH3—NEt£Cl“; for this salt is known to be stable under, the reaction conditions used. The most reasonable interpretation of the facts, is indicated in the equation above. The attacking reagent attacks the, pushing 7r-electron density from the /3—7 bond to the a, , -carbon,, , 7, , -ft bond as the chloride, , ion is pulled off by solvent. This mode of attack is often referred to as an ab» (a) Kepner, Winstein, and Young, J. Am. Chem. Soc, , 71 115 (1949); (b) Young Webb,, , and Goering, ibid., 73, 1076 (1951); (c) England and Hughes, Nature, 168, 1002 (195 ).
Page 307 : Epoxides, , -, , normal bimolecular displacement or SN2' reaction.71 It is generally o serve, , 291, w en, , bimolecular substitution at the a position of an allyl derivative (but not at its, T position) is seriously retarded by steric hindrance., By carrying out this type of reaction on cyclohexene derivatives, it is pos, sible to show that the entering group comes in and the leaving group departs, from the same side of the substrate molecule. (This is, of course, in contrast to, the stereochemical situation in an ordinary Sn2 reaction.) For example, ordi¬, nary bimolecular substitution on the 2,6-dichlorobenzoate of the hindered, cyclohexenol XXXVI is virtually prohibited sterically, but the SN2' reaction, (using piperidine or sodiomalonic ester as an attacking reagent) takes place, , easily.72 As shown, the attacking reagent comes in cis to the leaving group., On the basis of the limited data now available, it appears that the cis relation¬, ship between the entering and leaving groups in the Sn2' reaction is general., However, no completely satisfying explanation for this stereochemistry seems, to have appeared, although there have been a number of attempts.70®’72, , Epoxides 73, , Epoxides ^C———derivatives^ are of some interest as substrates in, , nucleophilic substitutions, for in these there are two possible reaction sites that,, for unsymmetrical epoxides, are nonequivalent. Ether linkages are ordinarily, not broken by the more usual nucleophilic reagents, but the highly strained, t ree-membered epoxide ring is exceptional in this respect. The most straight(1956)F°r 3, 7,, , °f thC ‘SV2, reaction> see de W°he and Young, Chem. Revs., 56, 769-784, , p°rk and White> J-Am. Chem. Soc., 78, 4609 (1956)., , field’s, (4) Eliel i„, 1956, pp. 106-114., , ,, , 1, , wTidCS !T 9<a) Yins"in and Hcnderson in Elder•’ o, W,'7 and Sons’ Inc - N'w York, 1950, pp. 22-46,, “, ga"“ Chemilry- John Wiley and Sons, Inc., New York,
Page 308 :
292, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , forward mechanism for the opening of the epoxide ring is closely analogous, to ordinary bimolecular substitution., CH30:-, , + H2C-CHMe, , -> CH3 O: CH2— CH — O- 74, Me, 75(a), , SCN, , +, , /, , H2C — GH-CH2C1, l\ /, V°, , c6h6ch2nh:, , O', , -> NCS: CH2—CH, , CH2C1, , h2c^ch-cgh5, , C6H5CH2NH2—CH2—CHPh-O" 75(6), , Like the direct displacement mechanism, this mode of ring opening is favored, by very nucleophilic attacking reagents and by the absence of alkyl substituents, on the carbon atom under attack. As indicated, if one of the carbons in the, ethylene oxide ring is less substituted than the other, attack will occur prefer¬, entially (sometimes exclusively) at the less substituted carbon., Openings of epoxide rings may be acid catalyzed. In these, the conjugate, \, \ +, C, O—H, , acid, , ], , is involved. The ring-opening step may be unimolec-, , ular, , \, , X, C, , /vx, , H+, =F-, , /, , \, , +_dow^, , (/, , /, , V x, , I \, , fast, , /I, Y, , HO, , I, H, , \, HO, , or it may be bimolecular., K, , /, , V, Y:, , +, , V x, , ->, , /, C—OH, , /, , \, , H, The ring opening of ethylene oxide itself with HC1 and HBr has been found to, exhibit a third-order rate law 76, n Reeve and Sadie, J Am. Ckm.S*,. 72, 1251 0 «0>., « (a) Nichols and Ingham, ibid., 77, 054/ (waa;. w u, , Lu, , j 0, , 1187.5Sted, Kilpatrick, and Kilpatrick, J. Am. a,m. S.51, 428 (1929)., , Cto... 17.
Page 309 :
Epoxides, , -, , 293, (7), , rate = £3(oxide)(H+) (halide ), , clearly consistent with the second of these mechanisms and inconsistent with, the first (in which the nucleophile attacks after the rate-determining step)., Moreover, a number of acid-catalyzed epoxide openings proceed with almost, complete inversion of configuration at the carbon atom under attack, , a stereo¬, , chemical situation generally considered typical of direct displacement reactions., In one of the simplest of such cases, the acid-catalyzed methanolysis of cyclopenteneoxide (XXXVII) yields only /rafl.r-2-methoxycyclopentanol (XXXVIII),, whereas if the reaction passed through a carbonium ion, both cis and trans, isomers would be formed.77, , OMe, MeOH, h2so4, , OH, XXXVII, , XXXVIII, , On the other hand, it now appears that the acid-catalyzed hydrolyses of, simpler epoxides in water almost certainly proceed through carbonium-ion, intermediates.75 Although we might suspect that the rate law for such reactions, could not tell us whether or not water is involved in the rate-determining step, (since water is present in large excess), careful study of the kinetics of these, hydrolyses in strong acid shows that their rates are proportional, not to (H30+),, but rather to Hammett’s ho function. We have already seen that such a depend¬, ence indicates that the activated complex for the rate-determining step consists, of the substrate, a proton, but nothing else (p. 190); this strongly suggests a unimolecular slow step., The entropies of activation for such hydrolyses point to the same conclusion, It is interesting that AS' values for acid-catalyzed breakage of C—O bonds, tend to fall rather sharply into two ranges: (a) from -20 to -25 cal per degree, and (b) from 0 to +10 cal per degree. From independent evidence it is believed, that the activated complexes for hydrolyses in the first group (hydrolyses of, or 'nary esters and 7-lactones) involve just one molecule of water. The transiton states for hydrolyses in the second group (which includes simple acetals, waterUCThU dT 'h°USht (ag3in fr°m independent evid<="«) to involve «,, waten Tins d.stmct.on is reasonable, for the mere loosening of a C-O bond, wate;dZJ?t y affr‘ the f-d-.°f.motion in a system, but removing a, strate will 'U 6, ' S°lvent and llnklnS 11 to a single molecule of the subimpose severe restrictions on its randomness, resulting in a sizable, stereospecificity inlc'id "^1™/epSidf”', " Pritchard and Long, X ^ CW, , 21% (1953>' For additional examples of, W
Page 310 :
294, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , decrease in entropy. Now, the AS* values for the hydrolyses of the simpler, epoxides have been found75 to lie between - 3 and - 6 cal per degree, suggesting, that these reactions proceed through transition states involving no water molecule., However, the fact that these A,S* values actually lie between the two ranges, given above would make us rather reluctant to accept this conclusion were it, not for other evidence., When the opening of an epoxide ring proceeds through a carbonium ion,, we should expect the ring to break preferentially at the most highly substituted, carbon atom; for, as we have repeatedly noted, secondary and tertiary carbons, are better able to tolerate a positive charge than are primary carbons. This, means that the reaction of an unsymmetric epoxide with a given nucleophile, may yield mainly one product in the absence of acid, but mainly another with, acid present. We have, for example, seen that the reaction of methoxide with, propylene oxide and that of benzylamine with styrene oxide yield secondary, alcohols (attack at primary carbon) under basic conditions. Yet, under acidic, conditions, these same pairs of reagents yield mainly primary alcohols, for the, ring opens more easily at the secondary carbon atoms., EtOH, , + H2C-CHMc —> HO—CH2—CHMe—OEt (mostly)74, , PhCHoNHo + H2C-CHPh ->, , \ /, O, , HO—CHo—CHPh—NHCH2Ph (mostly)75(6), , The Internal Nucleophilic Substitution (S^i) Mechanism, One of the more usual methods of converting alcohols to alkyl chlorides is, treatment with thionyl chloride, SOCl2. This reaction has been shown to pro¬, ceed through an alkyl chlorosulfite (XXXIX), an ester that, with care, can be, isolated and shown to decompose upon heating to the alkyl chloride and S02., , O, ROH + SOCl2 -> R-O-S-Cl —> RC1 + S02, XXXIX, This substitution is almost unique in that it sometimes (but not always) results, in retention of configuration about the a-carbon atom, even though no neig boring groups are involved. However, if the reaction is carried out in the pres¬, ence of amines (as is sometimes done in the laboratory in order to remove the, * (fl) Bartlett and Herbrandson,, , J. Am., , ibid., 74, 308 (1952); 75, 3182 (1953)., , Chem. So,, 74, 5971 (1952). (*) Lewis and Boozer,
Page 311 :
The Internal Nucleophilic Substitution (S^i) Mechanism, , 295, , HC1 formed), the substitution proceeds with inversion of configuration. These, facts suggest that the chlorosulfite (XXXIX) decomposes through a cyclic, intermediate, XL. As shown, the intermediate XL may break down in a sort, \, , V ., , \, , c-o-s-ci, /', o, , -I, , C-Cl + SO;, , ret, , o, , 7, , —G, , /, , "Cl, , Cl—G— + SO, + Cl", \, 2, , XL, of “internal return” reaction, forming the chloride with retention of configura¬, tion; or, if excess Cl- is present, an ordinary displacement reaction may take, place, forming the alkyl chloride with inversion. The latter reaction occurs, when the reaction is carried out in the presence of base; for the HC1 formed in, the initial condensation of alcohol and SOCl2 is converted to the nucleophilic, Cl~ ion. We strongly suspect that intermediate XL is an ion pair with a partial, positive charge on the a-carbon, for the reaction resulting in retention exhibits, certain of the characteristics of Sn 1 reactions. It is retarded when deuterium, atoms are substituted for /3-hydrogens,70(6) and it takes place most readily when, phenyl groups are bound to the a-carbon.50 Furthermore, the chlorosulfites, of primary alcohols5' or alcohols in which the a-carbon is at a bridgehead50 do, not decompose to yield alkyl chlorides at all. Since the substitution takes place, through a type of “internal return” reaction, we often refer to it as an SNi (sub¬, stitution-nucleophilic-internal) reaction.8* The same type of mechanism seems, to apply also to the decomposition of secondary chlorocarbonates.55 As with, , \, , O, , -c-o-c-ci, , ret^r 7C-CI + C02, , v UT, , c=cr, / \../J, M?;, •pi, inv\\, , —o', , ¥■, , ••, , _, , ,, , /, , Cl—c— + co2 4- cr, , chlorosulfites, inversion, rather than retention, prevails if the reaction is carried, out in an excess of amine., , n, cyclic carbiS"may°ftSe’plac^hrough anT, 540, 250 (1939); 543, , 198 11940, , ,, , ", , &, , low temperatures (L„enc and RotheT", R, , isxi, ^, , mechanism, , i937-,252reaC,ion °f PC'S wi,h certain, (Huckel and co-workers, Ann, , °f HBr with Phenylalkyl carbinoU a, , seem ,0 be exceptional cLes As yftonW,t, Z-237 <1939»- Al Pr«en, these, considered as generally useful fo/br’inging aboutTTre ' ',10ny ch'orlde and phosgene, may be, must-be carefully controlled., § g, S reactions, and even with these, conditions, Chem. Soc., 77, 2774 (195S)!PS* J, , ^ 1932’ 108’ 1232, ^ WibcrS and Shryne, J. Am
Page 312 :
296, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , With allylic systems, another mode of reaction is possible—internal sub¬, stitution with allylic rearrangement (which we may abbreviate SNi')., , ^CH, R-CII^XCH,, RCH= CH - CH 2— O - S - Cl, , t, , R—CH— CH= CH,, , O, , :C1, , O, , Cl, , S', , I, , + so,, , o, , This is, in fact, a much more favorable reaction path for allyl chlorosulfites, than the ordinary Sni reaction. The study of the reactions of substituted allyl, chlorosulfites is not easy, for although allyl chlorosulfite itself has been char¬, acterized, the substituted allyl chlorosulfites have not. They must, therefore,, be prepared on the spot from the alcohol and SOCI2, a reaction yielding HC1, along with the desired ester. In a polar solvent, HC1 will be partially ionized,, and in the presence of Cl~, the chlorosulfite may conceivably react by five, mechanisms: Sn 1, Sn2> Sn2', Sni, and Sm'. However, if the reaction is carried, out in dry ether, in which dry HC1 is known to be unionized,54 the first three, of these mechanisms may be assumed to be unimportant.55 Under these condi¬, tions it is found that, , 7-methylallyl, , alcohol (crotyl alcohol) is converted only to, , a-methylallyl chloride, whereas the a-substituted alcohol is converted only to, the Y-substituted chloride, that is, in both cases SNi' reaction predominates to, the virtual exclusion of the Sni reaction., Cl, SOCh, ether, , X, , ~T t, , t, , CH3—CH=CH—CH2—OH-» CH3—C—CH=CH2, H, CH3—CH—CH=CH2—► CH3—CH=CH—CH2, , Cl, , OH, , Substitution Reactions of Ambident Nucleophiles86, Just as allyl derivatives and epoxides are of special interest because they may, undergo nucleophilic substitution at more than one site, such nucleophiles as, the nitrite and cyanide ions and the anion of acetoacetic ester, Bushwell, Rodebush, and Roy, J. Am. Chem. Soc., ^0, 2528 (1938)., « Young and co-workers, J Am. Chem. Sac 77, 4182 (1955); ■“^3 (1953).^, (a) For a discussion of this subject, together wit re trenc, blum, Smiley, Blackwood, and Iffland, J. Am Chem. Sac., 77, 6269 (1955). (6) See, strom, Arkiv for Kemi; 6, 155 (1953); 7, 81 (1, , )•, , Brand-
Page 313 :
Substitution Reactions of Ambident Nucleophiles, , 297, , 'CH3—C=CH—COOEO, , o_, are of interest because they contain more than one atom that may function as, an attacking atom. These reagents are sometimes said to be ambident, signifying, that they may attack with “either tooth., , It has, for example, been recognized, , for many years that alkyl bromides ordinarily form nitriles (R—C=N) when, treated with NaCN but yield predominantly isonitriles, , (R—N=C), , when, , treated with AgCN. The reactions of alkyl iodides, either with NaN02 or, AgN02, yield mixtures of nitroalkane (RN02) and alkyl nitrite (R—O—N=0),, but the proportion of the latter is invariably greater from the reaction with, AgN02. In the past it was thought that these differences arose because the, silver ion somehow modified the nature of the attacking reagent, but it now, seems much more likely that the influence of the silver is due to its interaction, with the substrate., More particularly, it appears that as the substitution reaction acquires, an increasing degree of “iSVl character,” the incoming nucleophile tends to, attack with its most electronegative atom, but if the reaction acquires more of, the character of a bimolecular substitution, the incoming group attacks with a, less electronegative atom. Since the Ag+ ion, and even sparingly soluble silver, salts, aid in the formation of a carbonium ion by removing the halide from the, substrate (p. 274), the action of silver salts favors attack by the nitrogen atom, of the CN, , ion and by the oxygen atom of the NO^- ion. There is considerable, , evidence (much of it pertaining to the reactions of AgN02) to support this, conclusion. For example, those alkyl halides giving the highest ratios of nitro¬, alkane to alkyl nitrite on treatment with silver nitrite are primary halides (for, which substitution reactions have the least “SVl-like character”); those halides, giving the largest nitrite :nitro ratio are tertiary halides (for which substitutions, have the greatest degree of “.SVl-likc character”). Furthermore, the nitrite:, mtro ratio in a given reaction can be increased by transferring the reaction, from a less polar to a more polar solvent. When n-heptyl iodide is treated with, silver nitrite in ether, the nitrite to nitro ratio in the resulting product is 1:8,, ut when the reaction is carried out in acetonitrile, the ratio rises to 2 • 5 Treat¬, ment of benzyl bromide with silver nitrite yields a mixture of products in which, the mtro compound predominates over the nitrite ester by a ratio of about, rinj the, ameth°Xy gr°Up is introdu“d on the para position of the benzene, mg the positive charge in the carbonium ionlike transition state is stabilized, of about’2 1, , o'7 T br°r, , the Pred°mi"a"' pr°d-‘. again by a ratio, , “destabilized” be 7, ’, °harge in ,he transition state is, to about 10 1., 3 P'mtr° SrOUP’ the nitro:"itri“ ratio in the product rises
Page 314 :
298, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , The story is much the same for other ambident reagents. Treatment of the, silver salt of a-pyridone (XLI) with ethyl iodide yields an O-ethyl derivative, (XLII); but if the potassium salt is used, an N-ethyl derivative (XLIII) re¬, sults.57 Similar effects may be observed in the absence of univalent silver if the, Ag+, EtI, , XLII, N', , 'N"X>, , OEt, , K+ EtI, , XLIII, , Et, reagents are judiciously chosen. The anion of aceotoacetic ester (XLIV) attacks, ordinary primary halides (whose substitution reactions are generally bimolecular) yielding C-alkyl derivatives (XLV); but the same anion attacks the a-chloro, ether (XLVI) (whose substitution reactions in ethanol are unimolecular) to, \88, yield an O-alkyl derivative (XLVII)*, , CH3-C-CH- COOEt, II, I, (XLV), O Me, [CH3—C-CH—COOEt ^ CH3—C=CH —COOEt], CH3—C=CH-COOEt, , :Q:, , O, XLIV, , ;CI, , <j,, , \, , (XLVII), CH.—OCH,, , The rationalization of these contrasts in the behavior of ambident anions, (which the reader may or may not find fully convincing) is as follows. In an, SN\ reaction, the incoming nucleophile, if an anion, attacks the carbonium-ion, , intermediate largely because of the gain in electrostatic stability resulting from, the neutralization of charge. It will be to the advantage of the carbonium ion to, approach the anion near the spot where the latter has Us highest concentration of, negative charge. Since excess negative charge is heaviest on the most electronega¬, tive atom, the new bond should form at this atom. On the other hand, in a, direct displacement reaction, a section of the solvation shell about the attacking, reagent must be disrupted and the formation of the new bond initiated while, the leaving group is still partially bound to the carbon atom under attack., Since the solvation shell is thickest about the atom having the highest concentra¬, tion of negative charge, less reorganization of solvent molecules will be necessan, if the solvation shell in the region of the less electronegative atom is d.srupte ., Hence, bond formation at this atom is favored. In effect, then, t e most nuc eo, philic site in an ambident anion is the least electronegative atom having, unshared electron pair., , n ]t, r>, 04 3148 (1891); Rath, Ann., 489, 107 (1931)., 87 Von Pechmann and Baltzer, Ber., 24,, t, ', 88 Simonsen and Storey, J. Chew. Soc., 95, 210 (
Page 315 :
299, , Attempted Correlations of Substitution Rates, , Attempted Correlations of Substitution Rates. The Swain and Winstein-Grunwald Equations, The problem of quantitative correlation of nucleophilic substitution rates has, aroused the interest of a number of workers. What, in general, is desired is the, tabulation of a series of constants (“parameters”) associated with the various, substrates, attacking reagents, and solvents, which would, in principle, allow, the a priori calculation of the specific rate of attack by any reagent on any, substrate in any solvent. The most ambitious treatment of this sort seems to, be that proposed by Swain,59 who pictured all Sn reactions, irrespective of, mechanistic type, as resulting from the combination of a “pushing”, , and, , “pulling” action. The “pull” may be exerted by one or more solvent molecules, or by an added Lewis acid (for example, Ag+ or SnCL); in all cases the pulling, agent is acting in an electrophilic capacity. The “push” in an Sn2 reaction is, exerted by the entering group. In an Sn\ reaction it is exerted by the solvent,, but may become very feeble in cases where the carbonium ion intermediate is, unusually stable. In a “borderline substitution” the push may be exerted by the, incoming group, by the solvent, or by a combination of these. In all cases, the, pushing reagent is acting in a nucleophilic capacity. We may then represent an, SN reaction on substrate R:A' schematically, and very generally, as, ■—x, , N:, , ' + 'XR('-X, , +, , E, , ratedetermining, , [N:-~■ R, , +, , X-E\-, , _, , final, products, , where A : and E are the pushing and pulling reagent, respectively. (Note that, we have not specified the number of molecules of AT or E, nor the nature of the, bonding in AT —R and X: - E.), The Swain treatment stipulates, in effect, that the nucleophilic “push”, and the electrophilic “pull” associated with a substitution reaction make inde¬, pendent contributions to its free energy of activation, and that if we select a, substitution reaction of a given substrate in a given solvent, we may calculate, us specific rate, k, if we know the solvolysis rate,, , of the same substrate in a, , standard solvent. (The standard solvent chosen by Swain, in agreement with, a related treatment by YVinstein, p. 302, was 80 percent aqueous alcohol.), i ne owain relationship is, , l0g ko = nSn +, , (8), , where a is a nucleopkilicity parameter associated with reagent AT, , is an electro(1253); Swain,
Page 316 :
300, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , philicity parameter associated with reagent E, sn is the sensitivity of the reaction, under consideration to nucleophilic push, and se is the sensitivity of this reaction, to electrophilic pull., The Swain equation, like the Hammett equation (p. 220) and the Taft, equation (p. 227), is a linear free-energy relationship. Since four parameters, rather, than two, are involved, four points of reference must be arbitrarily chosen., By selecting solvolysis in 80 percent aqueous alcohol as a “standard reaction,”, we are automatically fixing two of these; n and e for this reagent are 0.00. The, remaining two points of reference become fixed upon choosing a “reference, substrate” for which both sn and se are given the value 1.00.30 /-Butyl chloride, was chosen as this standard.It then becomes necessary to obtain n values for, as many nucleophilic reagents as possible, e values for as many electrophilic, reagents as possible (polar solvents will have both an n and an e value), and, sn and se values for as many substrates as possible. From rate data on 152, solvolyses involving 25 substrates and 17 solvent systems, values for 78 Swain, parameters giving the “best fit” to equation (8) were calculated (a task better, carried out with an electronic calculator rather than with a slide rule). A num¬, ber of typical Swain parameters are listed in Tables 8-3 and 8-4. With these,, and others not included, rates of solvolyses within the scope of this treatment, may be estimated with a probable error of 0.12 in the logarithm, even though, the rate of solvolysis of a given substrate may change by several powers of 10, when transferred from one solvent to another., The orders of electrophilicities and nucleophilicities given in Table 8-3 are,, qualitatively at least, not greatly different from what we would expect. How90, , jn Swain’s 1955 papers, only solvolysis reactions were considered; that is, reactions in, , which Nl is the same as E. This limitation would ordinarily prevent evaluation of the various, parameters associated with the reagents and substrates. In fact, it may easily be shown that no, matter how many data were obtained, evaluation of the Swain parameters would squire solv¬, ing x simultaneous equations in * + 2 unknowns. For this reason, Swain made two additiomd, assumptions, fixing sjst for methyl bromide as 3.0 and sjs, PhsC-F as Vs (It, be, recalled that in substitution reactions of MeBr, the push tends to be considerablyjnore, important than the pull, whereas the reverse is true for such reactions of triphenylme hy, halides.) There is, of course, no guarantee that the two magnitudes have anything but qua tatlV- Swafn’s^umption that * = * for /-BuCl has subsequently provedl toJ, , ^latwely, , unwise one. Recent experiments by Doering and Finkelstein, 1 (c), p. 642) show that the observed variation in rates of solvolysis, closely parallels the observed variation is solvolysis rates of bromide XLV III which, , XLVIII
Page 317 :
Attempted Correlations of Substitution Rates, , 301, , ever some of the individual j„ and s0 values in Table 8-4 are not in accord with, chemical experience. Benzhydryl chloride is listed as being more sensitive to, electrophilic pull than is triphenylmethyl fluoride, whereas both benzhydryl, chloride and /-butyl chloride, are listed as being more sensitive to nucleophilic, push than are methyl, ethyl, and n-butyl bromides. Inconsistencies of this kind, lead one to doubt the theoretical validity of the Swain equation, despite its, usefulness for predictive purposes., , Table 8-3. Electrophilic and Nucleophilic Reagent (Solvent), Parameters in the Swain Treatment, Solvent, , e, , n, , HCOOH, 97.5 percent Ac20, 2.5 percent HOAc, , + 6.53, + 5.34, , -4.40, , h2o, HOAc, , +4.01, +3.12, , -0.44, -4.82, , -8.77, , 80 percent EtOH, MeOH, EtOH, , 0.00, , 0.00, , -0.73, , -0.05, , -1.03, , +0.53, , 90 percent acetone, , -1.52, , -0.53, , Table 8-4. Nucleophilic Susceptibility and Electrophilic Susceptibility, Substrate Parameters in the Swain Treatment, Substrate, , Sn, , ■Te, , $n/$e, , Triphenylmethyl fluoride, Benzhydryl chloride, /-Butyl chloride, f-Propyl bromide, , 0.37, , 0.33, , 1.24, , 1.12, 1.25, , 1.00, , 1.00, , 1.00, , 0.90, , 0.58, , 1.55, , Benzyl tosylate, Ethyl bromide, , 0.69, , 0.39, , 1.77, , 0.80, , 2.22, , 0.77, , 0.36, 0.34, , 0.81, , 0.27, , 3.00, , n-Butyl bromide, Methyl bromide, , 0.99, , 2.26, , The Swain treatment has also been applied to a series of SN2 reactions in, water (m which case the is, term in equation (8) remains a constant). The, tosvlX, ,COrrda,:°n’ aPP‘ied '° such subs‘rates as benzyl chloride and ethyl, tosylate and to such nucleophiles as CN~ I~ N~ and OHY, , .a™,.,™. m «_, ", , V., , ; r,r.'S“ ”, , ... dmvrd from subslitution rearrioa. oo saturatrd oarbom doer oof.
Page 318 :
302, , Nucleophilic Substitution Reactions in Aliphatic Systems, , in general, apply to substitutions on sulfur, mercury, or even on unsaturated, carbon.92, On the other hand, for a group of substitution reactions that are essentially, insensitive to nucleophilic push, the Swain equation may, in principle, be sim¬, plified to, (9), Actually, the relationship, , (10), which, except for symbols, is identical to (9), was proposed by Winstein and, Grunwald,53 five years before the presentation of the Swain equation. As in the, Swain treatment, 80 percent alcohol was chosen as the reference solvent and, ^-BuCl as the reference substrate (for which m was fixed as 1.00). The Y value, (which Winstein prefers to regard as a measure of ionizing ability) for any solvent, is thus obtained by comparing the specific rate of solvolysis of /-BuCl in this, solvent to that for its solvolysis in 80 percent ethanol. The m value for a given, substrate could be obtained by measuring the specific rates of solvolysis of that, substrate in a number of solvents of known Y values, plotting their logarithms,, and drawing the best straight line through the experimental points, , (Ex., , 10). The Winstein-Grunwald equation (10) gives a satisfactory correlation of, the rates of solvolysis of a few tertiary halides and secondary sulfonate esters,, and attempts have been made to apply it also to secondary halides. Here,, agreement is good for solvolyses in various mixtures of the same two solvents,, but is much poorer if reactions in a number of different solvents are being com¬, pared. In other words, if we take the Y values obtained from the study of, the solvolysis of ^-BuCl in various solvents and solvent mixtures and plot them, against the logarithms of the specific solvolysis rates of, say, benzhydryl chloride, in these same solvents, the points are badly scattered. Those for 20, 40, 60, and, 80 percent ethanol lie on one line; those for the various water-dioxane mixtures, lie on a different line; and those for water-acetic acid mixtures lie on still a, third line. If equation (10) were to apply rigorously to benzhydryl chloride, all, points should lie on the same line., It has recently been shown5* that the Winstein-Grunwald equation may be, applied to a large number of substrates if, to each substrate there is assigned, »* For a discussion of this phase of the problem, see Edwards, J. Am. Chem. Soc., 76, 1540, 095^Winstein and Grunwald, J. Am. Chem. Soc., 70, 846 (1948); Winstein, Grunwald, and, 1602,4151,5937, , Jones, J. Am. Chem. Soc., 73, 2700 (1951)., , n Winstein, Fainberg, and Grunwald, J. Am. Chem. Soc., 79,, , (1957)., , ,
Page 319 :
Long-lived-Carbonium Ions, , -, , 303, , not a single m value, but a series of m values, one for each solvent pair. Typically,, the m value for benzhydryl bromide has been fixed at 1.69 for HOAc, , H20, , mixtures, 0.91 for water-acetone mixtures, and 0.95 for water-dioxane mixtures., While such an expedient improves the correlation, it obviously makes the treat¬, ment less useful, for it greatly increases the number of parameters that must be, determined experimentally., The partial failure of equation (10) when applied to secondary halides is, to be expected, for the applicability of this relationship should be limited to, substitutions in which nucleophilic push contributes little or none of the driving, force of the reaction. While this may be true of the solvolysis of tertiary halides,, it is not, as we have seen, true for secondary halides. Moreover, the WinsteinGrunwald relationship assumes that every act of ionization of the substrate, results in an act of solvolysis—that is, that there is no “internal return” (p., 288). This is much more likely to be the case in nucleophilic solvents such as, alcohols than in carboxylic acids. What is perhaps surprising about the WinsteinGrunwald relationship when applied to a large range of solvents is not that its, ; success is limited, but rather that it yields any measure of correlation at all., , Long-lived Carbonium Ions, The carbonium ions considered thus far have short lives; evidence for their, intervention in reactions is largely kinetic and stereochemical. However, a, number of carbonium ions are known which may exist for extended periods of, time in favorable surroundings. Triarylmethyl carbonium ions (Ar3C+) are, especially stable, and may be formed simply by the ionization of triarylmethyl, halides in polar solvents which should not, however, react with the ions. It has, long been known, for example, that solutions of such halides in liquid S02, conduct electric current,*95<‘> and that the observed relationships between con¬, ductivity and concentration are typical of a weak electrolyte.S5<6> These halides, are, as a rule, partially ionized in liquid S02 to ion pairs (which do not contribute, to the conductivity), and the ion pairs are, at moderate concentrations, par¬, tially dissociated to ions (which do contribute)., liquid SO2, , Ar3C—Cl ^-=± Ar3C+Cl~ ^ Ar3C+ + Cl~, ^r;hy?rrato'on the °ther hand> appear to be ^p^iy io^ed, , dis'ocLed, , ’ °nCe agai"’ thC reSU‘ting i0" Pair$ are, , Part‘ally, , AriC-OC,°* Si AraOCOr ^ Ar3C+ + CIO,, ■«. 35, 2403 (1902). (4) Ziegkr,
Page 320 :
304, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , Now the dissociation constants of the perchlorates have been found to be inde¬, pendent of the nature of the aryl groups,'95<6) and it may be supposed that this, is also true for solutions of the triarylmethyl halides. On this basis, the different, conductivities exhibited by solutions of the various triarylmethyl chlorides, merely reflect differences in their degrees of ionization. As may be expected,, ionization is favored by />-alkyl and /?-phenyl groups on the benzene rings and, is hampered, , by the />-chloro groups.95 Indeed, the jb-methoxy derivative,, , XLIX, appears to be very nearly completely ionized in S02 at 0°. On the other, , hand,, , benzhydryl halides (diarylmethyl halides) do not conduct measurably, , in S02., Triarylmethyl derivatives may also be ionized in less polar solvents, but, if dissociation of the ion pairs and higher aggregates is slight, ionization will not, be readily detectable by conductimetric studies. It may, however, be detectable, spectrophotometrically, for the spectrum associated with a carbonium ion is, modified to only a slight degree when that ion is incorporated into an ion pair., Thus, it has been shown that carbonium salts are formed when triarylmethyl, chlorides are treated with a number of Lewis acids in aromatic solvents; for, example,, Ph3C—Cl -f- SnCl4 (in benzene) —> Ph3C+SnCli-'9r(a), , )8C—Cl + HgCh (in chlorobenzene), , )8G+HgCir»7^, , The solutions formed in these reactions exhibit spectra closely akin to those of, the corresponding triarylmethyl perchlorates in SO,; moreover, evaporation of, these solutions yields salts that retain the spectra typical of the carbonium ions,, indicating that carbonium ions may also exist in the solid state. The spectrum, of the triphenylcarbonium ion is also exhibited by solutions of triphenyl carbmol, in concentrated sulfuric acid;!"’<“> the observed “four-fold freezing-point depres-, , 1957» w Hanlasch, Z. physik. Chm.. 61, 257 (1908). (4) Hammett and Deyrup, J. An. CHm., Soc., 55, 1900 (1933).
Page 321 :
Long-lived Carbonium Ions, , -, , 305, , sion” (p. 99) observed for such solutions*™’ indicates that conversioh to the, carbonium ion is very nearly complete., Ph3C—OH + 2H2S04-*, , Ph3C+ + H3O+ + 2HSO7, (4 moles of particles per mole PhaCOH), , Maximum resonance stabilization of the triphenylcarbonium ion would, presumably be achieved if all three benzene rings and the central carbon atom, were coplanar. However, it may readily be shown by use of scale models that, such a structure would involve considerable steric interference between the, ortho hydrogens of one benzene ring and those of another. Thus, one or more of, , the benzene rings must be twisted out of the plane containing the three bonds, to the central carbon atom. It is currently believed, largely on the basis of, spectral evidence,5*(o), , that all three rings are so twisted. If this is so, the, , ion may exist in the symmetric “propellerlike” conformation, L, or in the unsymmetric conformation, LI, or may be an equilibrium mixture of both.99(6), , In addition to triarylcarbonium ions, a number of additional long-lived, carbonium ions are known. Three of the more interesting of these are shown, below., , Lipoo(o), , LHpoo(6), , For each of these, a number of resonance structures may be drawn, showing, that m all three cases the positive charge is spread over a number of atoms, rather than being confined to a single atom., 99, , (a) Deno, Jaruzelski, and Schriescheim, J. Ore. Chem 19 1SS, lk\ c, 1, evidence that certain triarylcarbonium ions may conJst of a mixture oflothfi }, SPreCtral, Magel, and Lipkin, J. Am. Chem. Soc., 64, 1774 (1942), orms, see Lewis,, , M 3SS? a"d K"°-, , ^ O"., , 76, 3203
Page 327 : 311, , Exercises for Chapter 8, , Me, SOC1;, , ether, , OH, 9. Aside from its acid- and base-catalyzed hydrolyses, ethylene oxide undergoes slow, hydrolysis in aqueous solutions buffered at pH 6 to 8 with a rate independent oi pH., Two mechanisms have been suggested for this reaction:, slow, , "f*, , fast, , (1) H->0 + H2C-CH >-> H20—CH2—CH20--> HOCH2CH2OH, and, , \ /, O, , eq, , (2) H+, , + H2C-, , -ch2, , -CH2 + -^Uhoch2ch2oh, , H,C-, , o, , slow, , O, H, , (a) Show that both proposed mechanisms should lead to the same rate law—that is,, that both are consistent with a rate independent of pH., (b) It has been found (Pritchard and Long, J. Am. Chem. Soc.,, , 78, 6008 (1956)),, , that the hydrolysis is about 10 percent faster in H20 than in D20. Heavy, water is known to be about half as strong an acid as ordinary water and about, one third as strong a base. Which of the proposed mechanisms is in better agree¬, ment with the observed (small) isotope effect? Explain., 10. Given the following rate constants for the solvolyses of T-BuCl and /-BuBr in the, solvents listed (25°):, , Solvent, , ^BuCl X 106, ^BuBr X 104, , 80%, EtOH, , 100%, EtOH, , 40%, EtOH, , 9.2, 3.4, , 0.097, 0.057, , 1300, 350, , MeOH, , 70%, MeOH, , HOAc, , 0.75, 0.34, , 98, 31, , 0.21, 0.10, , 80%, Me2C=0, , 2 0, 1.1, , (a) Calculate Y values for each of the solvents., (b) Estimate m for the solvolyses of /-BuBr., 11. Explain each of the following:, <a), , molar>in 0.01 N NaOH are in the, , ;b;, , -, , (C) Sls'fTatrS t^d7sTM™ aC" bV addition of stnal. quan., PhjCHOEt in’the product, S'gn,ficant lncrease in the ratio Ph,CHOH/, , lej An aqueous solution of the optically active salt TV, “Kr, , LVI'in, , • u, , M Of the /groups have^been'racemized. S‘anding’ ‘',e
Page 328 :
312, , -, , Nucleophilic Substitution Reactions in Aliphatic Systems, , COOR, , ^^/COOR, , ^wCOO Na+, , +, COCTNa4-, , k"^^COOR, , LV, , 'COO-Na+, , LVI, (R= PhO, , r\, , CHPh—), , (f) The relative specific rates of unimolecular solvolysis of the following chlorides in, ethanol are, HC=C—CMe2—Cl = 1.0, MeC=C—CMe2Cl = 2000, H2C=CH—CMe2Cl = 10,000, (g) Treatment of substituted anisoles with HI yields CH3I and a substituted phenol,, rather than CH3OH and a substituted iodobenzene., (h) Br~ is displaced by the Sv2 mechanism more rapidly from CH2BrCl than from, CH2Br2., (i) The second-order reaction between />-nitrobenzyl bromide and LiCl in 10 per¬, cent aqueous dioxane is accelerated when the water content is raised to 30 per¬, cent, but is retarded if the water content is increased further to 50 percent., (j) The rate of racemization of chloride LVII exceeds its rate of solvolysis in acetic, acid, but the rate of racemization of brosylate LVIII equals its rate of solvolysis., , OAc, LVIII, , LVII, , OBs, (k) A /?-CH30 group generally retards the hydrolysis of an alkyl chloride, but with a, j8-CH3S group its acceleration is sometimes spectacular., 12, , Consider a description proposed by Bird, Hughes, and Ingold,'«<•> intended to apply, to both types of limiting substitution mechanisms and to borderline cases as well., Suppose that for any substitution reaction there is a volume a (which we may ca, -critical-reaction volume”) that surrounds each molecule of ^ate^Supf», further that if one or more molecules of attacking reagent are with.it this volum ,, there is a significant probability of reaction, but, if not, there ts a neghgtb.e prota, hilitv of reaction For Ss2 reactions this volume is very small, for n, this volume is much larger, for borderline cases this volume is of an tntermed.a, (a^ShovMhat if there are a reagent particles in a solution of volume F, the probability of finding at least one particle in the critical reaction volume, , - (’ - O', (This calculation is very much like the calculation of ',W, a. least one queen from a full deck of playing cards in a tndependen., , mpts.), P
Page 329 :
Exercises for Chapter 8, , 313, , (b) Note that v « V and that n is very large. Show that the rate of reaction may be, represented by the expression,, rate = K(RX)(\ - e~Nvc), where (RX) is the concentration of the substrate, K is a proportionality constant,, c is the concentration of the attacking reagent, and N is Avogadro’s number., ^Hint: log* x = (x - 1) - ^ (x - l)2 + j (x - l)3 * ‘, , j, , (c) Show that the reaction will appear to be first order (independent of c) if v is, large, but second order if v is small., ^•2, , Note: ex = 1 + x -j- — + — + ' * ', (d) Assume that a reaction may be regarded as first order if the term Ke~Nvc (which, represents the departure of the rate from that corresponding to first-order kinet¬, ics) is less than 1 percent of K. Show that on this basis, a reaction will appear to, be first order when the critical reaction volume is more than Jive times the average, volume of solution per molecule of attacking reagent., (e) Assume that a reaction may be regarded as second order if KNvc (the specific, rate corresponding to second order kinetics) is within 1 percent of K{\ — e~Nvc)., Show that on this basis, the reaction will appear to be second order when the, critical reaction volume, v, is less than J-fso of the average volume of solution per, molecule of attacking reagent.
Page 330 :
CHAPTER 9, , Reactions of Carboxylic Acids, and Esters, , Formation and Hydrolysis of Esters. Plurality of Mechanism7, There are a number, , of paths by which esterifications and ester hydrolyses may, , proceed. The hydrolysis of esters, like ring opening in epoxides (p. 291), may, involve bond breakage at one of two possible sites. In many instances, the, acyl-to-oxygen bond breaks /R—C —O—R\; but sometimes it is the alkyl-to-, , O, O, oxygen bond that breaks \R—C—O—|R/. In the first case, ester hydrolysis is a, nucleophilic substitution (—OH for —OR) at the acyl carbon; in the second,, the reaction, , O, , may be regarded as a nucleophilic substitution, , (, , OH for, , \, , _o_C—R/ at an alkyl carbon. It should therefore not surprise us that either, of these reactions may, like the reactions considered in the preceding chapter,, proceed either by a unimolecular or a bimolecular mechanism. Furthermore, as, with the opening of epoxide rings, ester hydrolysis may occur through attack on, the ester itself (in basic or neutral solution) or, alternatively, through attack on its, R, conjugate acid, R-C-6X, , (In add solution). In view of these three mecha-, , r \H, nistic dichotomies, we can visualize eight (2>) possible mechanisms for ester, , For a review of .his topic, see Day and Ingold, Trans. Faraday Sac., 37, 686 (1941)., , 314
Page 331 :
Ester Saponification. The BAC2 Mechanism, , -, , 315, , hydrolysis; to date six of these have been observed. A corresponding set of eight, mechanisms may be proposed for the reverse reaction, esterification, but since, esterifications are almost invariably carried out in acid solutions, we need con¬, sider only four of these. To date, three different esterification mechanisms have, been observed. We may describe the various paths for esterification and hy¬, drolysis, using a shorthand notation proposed by Ingold,2 in which the letters, B or A stipulate whether the substrate itself or its conjugate acid is attacked,, the symbols AC and AL indicate whether acyl-oxygen or alkyl-oxygen bond, breakage occurs, and the numbers “1” and “2” refer to the molecularity of the, rate-determining step. Thus, a BAc2 hydrolysis is a bimolecular reaction in¬, volving acyl-oxygen bond breakage, carried out in a basic (or possibly in a, neutral) medium. The possible mechanisms for esterification and ester hydrolysis, are summarized in Table 9-1. The reader is reminded that a carboxylic acid in, a basic medium is converted to a carboxylate ion, RCOO-, which, because of, its negative charge, is not subject to nucleophilic attack by alcohols or their, conjugate bases. Esterifications in basic media are therefore not observed., , Ester Saponification. The Bac2 Mechanism, It has long been known that the basic hydrolyses (saponifications) of ordinary, esters are second-order reactions.3 The transition states in these reactions may, then be assumed to contain one molecule of ester and one OH~ ion; that is,, the reactions are bimolecular. The site at which the ester molecule is broken is,, as we have seen (p. 145), established by experiments using water enriched in, O18. When an alkali metal hydroxide is dissolved in “labeled” water, the OH~, ions quickly become labeled also, because of the very rapid transfer of protons, between H2018 and OH~. The hydrolysis of rc-amyl acetate (and, presumably,, of other simple esters as well) in such 018-enriched solutions yields an “un¬, labeled, , alcohol but a “labeled” carboxylate anion., , o, o18-h, , +, , O, , %, , /O—Am —^ HO — C, , c, , /, , \, Me, , Me, , + OAm', , O, O18—G, , \, , + AmOH, , (Am=«-G5HU—), , Me, 1953^, , 754.’, , and Muhanism » °'S°™ Clumistry, Cornell University Press, Ithaca,, , in one^nheVefy ear^, , *” Sap°nifica,i°" °f «hy> acetate
Page 332 :
316, , Reactions of Carboxylic Acids and Esters, Table 9-1. Mechanisms for Esterification and Ester Hydrolysis, , Type, , Remarks, , Bac%, , Very common; includes almost all basic ester hydrolyses (saponifications), , Aac2, , Very common; includes acid hydrolyses of esters of primary and most, secondary alcohols, and most ordinary esterification reactions, , Bac1, , Not observed to date, , Aac^, , Rare; some esterifications and hydrolyses in concentrated solutions of, very strong acids are of this type, Extremely rare; observed only for hydrolyses of /3-lactones in the, , Bai2, , absence both of strong base and strong acid, Not observed to date, , Aal2, , Bal\\ Apparently quite general for the hydrolysis of esters of tertiary alcohols, and those secondary alcohols that yield the most stable carbonium ions, (for example, benzhydrol and a-methylallyl alcohol). Not generally, AaiA, , ), , observed for hydrolysis in concentrated bases, , clearly pointing to acyl-oxygen breakage.A mechanism consistent both with, the rate law and the observed site of bond-breakage is, , O, , O, , O, , HO' + C-OR'^, , HO-C-OR', , fast, , R, , R, , II, , HO-C + OR'', I, R, , fast^, , jRCOcr, \r'oh, , I, The overall reaction is irreversible. Although the first, and very probably the, second, steps are reversible, the final step is not., Now a question concerning anion I comes to mind, “Is this an intermediate, or merely an activated complex?” To answer this, let us consider another group, of experiments involving labeled oxygen,5 this time present in the acyl group, < Experiments intended to establish acyl-oxygen fission have bee:a carried out °n the, esters of such alcohols as a-methylallyl alcohol and neopentyl alcohol—alcohols which y, carbonium ions known to undergo rearrangement. (See, for example, N°rt°n an Ringed, J. Am. Chem. Soc., 62 1170 (1940).) Since saponification of such esten. yields ", products, it is assumed that the free carbonium ions are not ^termed, Q18, ■ h^d water., of such experiments is not as clear cut as is that of the, °, ‘occurs,, What they show is that either acyl-oxygen fission occurs, or that if alkyl-oxygen, , ,, , it is bimolecular., 1 Bender, J. Am. Chem. Soc., 73, 1626 (1951).
Page 333 :
Ester Saponification. The Bac2 Mechanism, , 317, , O18, of the ester, R—C—O—R'. Suppose ethyl, isopropyl, or /-butyl benzoate,, labeled at the benzoyl oxygen, is subjected to saponification in ordinary water,, but the reaction is stopped before completion and the unreacted ester ieisolated. It is then found that the O18 content of the unreacted ester has decreased., Depending upon the ester, the rate of O18 loss is from one tenth to one third the, rate of saponification. The sequence below suggests how this oxygen exchangecan occur:, , O*, II, HO"-f C-OR', I, R, I, R, , Since, in the exchange reaction, the labeled oxygen must depart as 0*H~,, the lifetime of anion I must be sufficient to allow a significant chance for its, isomerization to anion II, which may break up in the required manner. Since, the existence of an activated complex is only momentary, anion I must instead, be an intermediate., Two further consequences arise from the presently accepted mechanism, for saponification. In forming the intermediate I we are increasing the density, of negative charge at the reaction center. Therefore, saponification should be, facilitated by groups that will withdraw some of the excess negative charge, (that is, by electron-attracting substituents). This should be true whether such, substituents are located on the alkyl or acyl portion of the ester. Moreover, in, forming intermediate I, we are increasing the extent of crowding; therefore, we, should expect saponification to be subject to steric retardation. Both steric and, polar effects are in the expected direction. The following relative saponification, rates6 are illustrative of polar effects:, CH3COOMe, , UH2UlCOOMe, , unui2c<JOMe, , 1.0, , 761, , 16,000, , ffMeOAo, , (COOMe)2 *, , CHsCOOEt, , 170,000, , 0.60, , k, MeO Ac, , CH3—C—COOEt, , o, , 10,000, , McGraw-HurBook1^^ ^e,^e^'^0^,e”^®^P>pe211-2?2.rTheyrefcrlo<ireactU)ns<in-watcrait
Page 334 :
318, , Reactions of Carboxylic Acids and Esters, , Similarly, the three reaction series—the saponification (a) of methyl esters of, substituted benzoic acids, (b) of the ethyl esters of these acids, and (c) of the, benzoates of substituted phenols7—exhibit Hammett p values that are positive., Thus, as predicted, these saponifications are accelerated by such substituents, as —Cl, —Br, and —N02 (both on the alkyl and acyl sections of the molecule),, and are retarded by such substituents as —CH3, —NH2, and />-CH30—., Benzoates are, in general, saponified more slowly than acetates or even pro¬, pionates. This observation may cause some surprise since the phenyl group is, generally considered a more effective electron attractor than the methyl and ethyl, groups. But it should be remembered that the benzene ring, despite its — I, effect, is conjugated with a carbonyl group in a benzoate, and may donate, 7r-electron density to it through conjugation;, , +, , -*■, , c-cr, i, , OR, whereas conjugate effects of this sort are absent in aliphatic esters., With respect to steric effects, we may consider the following two series:5, CH3COOEt, , C2H5COOEt, , (CH3)2CHCOOEt, , (CH3)3CCOOEt, , 1.0, , 0.47, , 0.10, , 0.011, , k, ^EtOAc, , and, , CH3CH2OAc (CH3)2CHCH2OAc (CH3)3CCH2OAc (C2H6)3CCH2OAc, , 1.0, , 0.18, , 0.70, , 0.031, , f EtO Ac, , Although it might be argued that the progressive drop in saponification rate, with successive methyl substitution, in the first series, is due, at least in part, to, the electron-repelling action of the methyl groups, this cannot be said about, the second series because here the structural alterations are too far (four atoms, removed) from the reaction site. The decrease in rate in this case must be largely, steric in origin., , Esterification and Acid-catalyzed Hydrolysis of the Most Usual, Type. The, , Aac2, , Mechanism, , The most obvious difference between the basic hydrolysis of estersandtor, acid-catalyzed hydrolysis is that the latter is expenmentally reverse, its reversa, > See, for example, Tommila and KLmhelwood, . The data for the ethyl esters, m», , T, , 1938^'801 •, , ,hey, , on the' aeeiates are reported by New, , aTc^, John Wiley and Son,, Inc., 1956, „ 220; they refer to, , reactions in 70 per cent dioxane at 20 .
Page 335 :
Esterification and Acid-catalyzed Hydrolysis, , -, , 319, , being the acid-catalyzed esterification reaction. Some of the earliest experiments, demonstrating the phenomena of equilibrium and reversibility in chemistry, were, in fact, concerned with esterification and ester hydrolysis.5 Now an im¬, portant feature of reversible reactions is that the mechanism of the reverse, reaction is known with as much certainty as is the mechanism of the forward, reaction. This is a consequence of the so-called principle of microscopic reversibility,, which stipulates that if a system is at equilibrium and if there are a number of, (necessarily reversible) steps occurring, then the total number of molecular, systems participating in a given step per unit time must be the same as the total, number of systems participating in the reverse of this step. More particularly,, if a given sequence of steps constitutes the favored mechanism for the forward reaction, the, reverse sequence of these steps constitutes the favored mechanism for the reverse reaction.10, Thus, if we determine the mechanism for acid hydrolysis, we automatically, establish the mechanism for acid-catalyzed esterification., Let us first consider evidence similar to that upon which our proposed, mechanism for saponification is based. Hydrolyses of ethyl hydrogen succinate, (III)11 and the cyclic ester, y-butyrolactone (IV),le in H2O18 establish acylH+, , HOOC— CH2CH2—C —OEt + H2018->, o, III, HOOC—CH2—CH2—C—Ox8H + EtOH, , II, , o, h2o18, , ho-ch2— ch2—ch2— c-o18h, , II, , o, oxygen cleavage; whereas experiments with Ph—C—OEt in acid suggest that, , 018, acid-catalyzed hydrolysis (and, consequently, esterification) passes through an, intermediate having a lifetime long enough to allow appreciable oxygen exWaag,Xtfeal®^9.G69 0879):, , 385 <1862); 68’ 225 <I863>- G“Mberg and, trese„a,emd° here" f“, , IJ^ivmitVpr^^oSord, , Th^, , reversible^reaction, since *an^activated^rn Pf, , °f, , W°UW, , ,T°lman S Statist“al MecfuOd^Srfb£d, °f C°UrSC’ °bvi°US for a °ne-step, , favored, both for the forward and reversr.^actionr'Forie'T"1, be m°St, ever, jhe argument becomes more complex., reactions having several steps, how-, , „ ?atta’ DaY. fnd Ingold, J. Chem. Soc., 1939, 838., ong and Friedman, J. Am. Chem. Soc., 72, 3692 (1950).
Page 336 :
320, , Reactions of Carboxylic Acids and Esters, , change with the solvent.73 That the rates of acid-catalyzed hydrolyses are pro¬, portional both to (H+) and (ester) has long been recognized, but the kinetic, participation of water is more difficult to demonstrate when the latter is a, major component of the system. Rates of hydrolyses carried out in acetone with, measured amounts of water have been shown74 to be very nearly proportional, to (HoO), but we must use reserve in interpreting experiments of this sort,, since what is observed may conceivably be a rate increase due to an increase, in solvent polarity as water is added. However, it is possible, by a close examina¬, tion of the dependence of rate on acidity, to decide for this reaction whether, water is involved in the transition state for the slow step; for, it will be recalled,, if the activated complex consists merely of a molecule of ester and a hydrogen, ion, the reaction rate should be proportional to Hammett’s ho function;, rate - £(RCOOR')/*o, , (1), , whereas if a water molecule is also involved, the rate should be proportional to, the hydronium-ion concentration (p. 190)., , (2), , rate = £(RC00R')(H30+), , A distinction cannot be made in dilute aqueous solutions where h0 approaches, (H30+), but the two quantities are measurably different at acid concentrations, greater than 1.0 molar. In such acidic solutions, the rates of hydrolyses of, the more usual esters75<‘> and of, , -butyrolactone, (IV),7^ have been found, , 7, , to be very nearly proportional to (H30+). On theother hand, the rates of hydroly0pj2, 0=o, Me—CH—C O, ses of /3-propiolactone,, , |, ’, I, ch2—O, , , and, , -butyrolactone,, , 0, , ^ ^, , >, , 2, , are proportional to h0;'6 but it may be assumed that the /3-lactones, because, of strain associated with the four-membered rings, behave abnormally., The mechanism for ordinary acid-catalyzed hydrolyses (and esterifications), must then meet the following requirements: (1) all steps must be significantly, reversible; (2) acyl-oxygen bond breakage must occur; (3) the transition state, in the hydrolysis must consist of a molecule of ester, a molecule of water, and, an H+ ion- and (4) the reaction must pass through an intermediate t at ca, urvle long enough to allow oxygen exchange with the solvent. The Mowing, mechanism^ involving the conjugate acids of both the ester and the carboxylic, acid, fits these conditions:77, J. Am. Chem. Soc., 75, 5986 (1953h, ,ml), Friedmann and Elmore, J. Am., em', ’1301 (1940)- Bell, Dowding, and Noble,, « (a) Duboux and de Sousa, Helv. Chim. Acta 23, «« l1™”’ “ 3’326 (19586). {b) Long,, J. cJn.Soc., 1955, 3106; Chtniel and Long Td- <*.., 78,, Bender,, , Dunkle, and McDevit, J. Phys Chem., 5\829 ( 3267 (1950)<, i. Long and Purchase, J. Am. Chem. Soc., 72, 326/
Page 337 :
321, , Esterification and Acid-catalyzed Hydrolysis, , R', , ?, , + H2Q, slow, , R-C-O, , II, , "-H20, fast, , \, , H, , H, -R/ OH, fast ^, , R-C-O, \, , H, , O, , /, , H, , AR', , +X, R—C— O., , II, , + R' OH, slow, , o, , H, , H, , V, As required, the active intermediate, V, is a complex of ester, water, and H+., Assuming that shifts of protons from one oxygen atom to another may be very, rapid, we may represent the oxygen exchange accompanying acid-catalyzed, hydrolyses as, H, , *, , ?, , /H, , R-C-Ox, /°\, H, , , H, R', , H, R', , very rapid, , s, -, , R-C-O, I, , O, , /°x, H, , +/, R-C-O, , +, , h9o*, , H, , R', , If the ester is dissolved in an alcohol, R"OH, and acid is added, the original, ester, RCOOR', is converted to a second, RCOOR", with an alkyl group, derived from the alcohol. This is the familiar ester interchange reaction; it almost, certainly proceeds through intermediate VI, which is identical to V except that, an R"OH group has replaced an HOH group., , R', , O, , R', , O-C-O, /, , H, , I, , R, , \, , H, , VI, In acid-catalyzed hydrolysis, ester interchange, and esterification, the bond, angle about the acyl carbon is reduced from near 120° to near 109° during the, rate-determining step, and steric crowding is thereby increased. We should then, expect these reactions to be subject to steric retardation, just as is saponification., Moreover intermediate V differs from the corresponding saponification inter¬, mediate I only by the presence of two protons; and this is true also for the, respective transition states leading to these intermediates. Since the steric, rcw"cramimSH0frd P,r°t0nS may be C°nsidered negligible, it is very likely that, virtually3 the^, ^^ ffeSterification- and saponification are subject to, rrtually the same steric effects. Polar influences, however, should be much, ZZnTlZ rf add'catfyzed —‘ions than for saponification, for the, sc, eaction is jointly controlled by two factors that respond differently
Page 338 : 322, , Reactions of Carboxylic Acids and Esters, , to polarity. Such a reaction will be accelerated if the conversion of the substrate, to its conjugate acid is made more complete (this will be facilitated by electronrepelling groups), but it will also be accelerated if the protonated substrate, coordinates more readily with the displacing reagent (this will be facilitated by, electron-attracting groups). If we are considering reactions in very acidic media, where virtually all of the substrate will be converted to its conjugate acid, we, may be reasonably sure that the Aac2 reaction will be accelerated by electronattracting groups, but we are far more often interested in reactions in only, moderately acidic solutions, and here we cannot safely predict the direction of, polar effects (although we may expect their magnitudes to be small). The, reasonably successful Taft treatment of polar and steric effects (p. 227) is based, upon the assumption that polar effects in the acid hydrolyses of the usual esters, are negligible. Typically, the specific rate of hydrolysis of ethyl ^-nitrobenzoate, (in “60 percent alcohol” at 100°) exceeds that of ethyl /?-methoxybenzoate by, only 12 percent/7 whereas the specific rate of esterification of />-methoxybenzoic, acid (in MeOH at 25°) exceeds that for jfr-nitrobenzoic acid by about 70, percent.15, As an example of steric effects in Aac2 reactions, let us consider the esteri¬, fication rates of a number of aliphatic acids. The following are a few of the many, relative specific rates that have been determined, , _L, , (MeOH, 40°) :19, , CH3COOH, , n-C3H7COOH, , (CH3)3CCOOH, , 1.0, , 0.51, , 0.037, , ^HOAo, , ch3, , k, 110, , (CH3)3CCH2COOH, , (CH3)3C—ch—cooh, , 0.023, , 0.00062, , (CH3) 3C—C(CH3) 2—COOH, , Et3C—COOH, , 0.00013, , 0.00016, , Ao, Et, , h, , (f-Pr) 2CHCOOH, , (CH3) 3c—ch—cooh, , <10~4, , <10-4, , ^HOAc, , The sharp decrease in rate with the increasing chain branching suggests the, combination of steric and ponderal effects which we have already d.scusse, in regard to the halide-exchange reaction <p. 275). As wtth the latter, /3-alkyl, 11 Timm and Hinshelwood, J. Chem.Soc., 1932, 55., Hartman and Gassmann, J. Am. C em. oc.,, *, (1952)., i. Loening, Garrett, and Newman, J. Am. Chem. Soc., 74, 392 J (1M).
Page 339 :
Esterification and Acid-catalyzed Hydrolysis, , 323, , substituents are significantly more effective “steric hinderers” than a-alkyl, substituents.*0 Thus, incorporation of three a-methyl groups into acetic acid, "he esterification rate by a factor of 27, but incorporation of three, 0-methyl groups into propionic acid lowers the rate by a factor of 36. More, spectacularly, if the methyl groups in trimethylacetic acid are replaced with, ethyl groups (thus effectively adding three 0-carbons), the esterification rate is, lowered by a factor of 230. In extreme cases, di-isopropylacetic acid and a-tbutylbutyric acid, both of which have four 0-methyl groups, are esterified too, slowly to measure., Very nearly the same idea may be expressed somewhat differently using, the Newman rule of six,*1 which says, in effect, that those atoms which are most, effective in providing steric hindrance to addition are separated from the at¬, tacking atom in the transition state by a chain of four atoms. This means that, if either the attacking atom or the carbonyl oxygen is designated “1,” the, “blocking atom” will be in the “6” position (Fig. 9-1). If the acid is assumed to, have a coiled structure with normal bond lengths and bond angles, it is possible, to show, using molecular models, that an atom in the 6-position is much more, likely to draw close to the path of the attacking reagent than is an atom in, say,, the 5- or 7-position. If there are only a few (3 or 6) atoms in the 6-positions, they, may be moved out of the way of the attacking nucleophile by rotation about, the C3—C4 bond; but if there are 9 or 12 atoms in the 6-positions, it becomes, increasingly difficult to twist the acid chain into a permissible conformation, while avoiding close interaction with the 1-position. Of the acids whose esterifi¬, cation rates are listed on page 322, the final two have 12 hydrogen atoms in the, 6-positions, and the four preceding these have 9 each. Although f-butylacetic, acid is esterified more rapidly than we would predict, the trend is otherwise, quite definite., 10 It might appear that the effectiveness of /3-alkyl substituents in steric hindrance may be, rationalized in the same way for esterifications as for halide-exchange reactions, that is, ^ar ?X^C ac*d molecules, like alkyl halide molecules, are chains which are bent or coiled so, that hydrogens on 0-alkyl groups block the path of the attacking reagent more effectively than, hose on a-alkyl groups (see Fig. 8-1). While such an explanation may not actually be wrong, nfCmde!iCrffPtl0n °f theSC tW° reactions in similar language masks important differences. Because, tion7c?h5erCnt “nVeuntl°f in nomenclature, the atom under attack in a halide-exchange reaca hvdr, a"Carbo"> but that under attack in esterification is adjacent to the a-carbon. Therefore, atom "nS ethe°arf, Y ST? l* a halide-exchange reaction is separated from the attacking, tr atom, J, complex by three atoms, but in an esterification, it is separated by, different for theT^T’ * ^ configuration of bonds about the carbon atom under attack is, chemi0/ >, ,, tyPCS °f transition state. Thus, except in a rather broad sense the stereo, chemical situations in the two types of reactions are dissimilar., ’, , ^ph^Te'mtrlLeoTthti'^72! 178\(I95K0)', °riSinal form> ,his rul<=, to, oxygen although mL n, ? ’ relaUonship between the blocking atom and the carbonyl, It should also be”ottd that tho rnlo r7, , a ““‘“"‘V»" and the incoming nucleophile,, , figuration about the atom under attack & teUhedmUn thetamidonsmte.1" WhiCh ,h'
Page 340 :
324, , -, , Reactions of Carboxylic Acids and Esters, , The esterification of benzoic acid is retarded by introducing a substituent, ortho to the, , COOH group, and (unless unusually severe conditions are em¬, , ployed) virtually prevented by a pair of substituents at the 2- and 6-positions., This is doubtless a steric effect, for it is observed whether the substituents are, electron attracting or electron repelling. Typically, an o-C2H5 group lowers the, rate of esterification (methanol, 15°) of benzoic acid by a factor of 5, whereas, an 0-NO2 group lowers it by a factor of 32." Both 2,4,6-tribromobenzoic and, 2,4,6-trimethylbenzoic acids are resistant to esterification under the usual con¬, ditions (although, as we shall soon see, the latter may be easily esterified by, employing the correct tactics). The nature of these “ortho effects” is readily, understood if we recall (p. 236) that substituents ortho to the —COOH group, , force the latter well out of the plane of the benzene ring. Although we know, somewhat less about the geometric details of esterification than those of nucleo¬, philic displacement, we feel reasonably certain that the attack on an unsaturated, OH, function, , and this includes the —C, , group, , occurs from a direction, , OH, perpendicular to the plane defined by that group. As shown in Figure 9-2, if there is one, substituent, X, ortho to the protonated carboxyl group, the incoming alcoho, molecule can attack only by route A; if two ortho substituents are present, this, route is also closed., ., ., Since the reactants in acid-catalyzed ester hydrolyses are very similar, nature to those in esterifieation, and sinee the transition states for the two reac¬, tions are almost identical, we would expect steric effects in ester hydro VJ, be very nearly the same as those in esterification. A close parallelism has indee, « Sudborough and Turner, J. Chem. Soc., 101, 237 (1912).
Page 341 :
Reactions Proceeding Through Acylium Ions, , -, , 325, , been observed/5 Furthermore, since the specific rates of esterification and, hydrolysis vary in almost exactly the same manner as the structure of the sub¬, strate is modified, the ratio of these specific rates—the equilibrium constant for, esterification—should undergo very little variation with structure. This is also, known to be the case/'* Structural modifications that cause decreases of several, powers of ten in the esterification rate change the equilibrium constant by less, than 20 percent., , Fig. 9-2. Steric, Benzoic Acids., , Hindrance, , in, , Esterification of, , ortho-substituted, , Reactions Proceeding Through Acylium Ions. The, , AacI, , Mechanism, , Mesitoic acid (2,4,6-trimethylbenzoic acid) is stubbornly resistant to attempted, esterifications under ordinary conditions, but it may be easily esterified by dis¬, solving it in concentrated sulfuric acid, then pouring the resulting solution into, cold methanol/5 Conversely, the methyl ester of this acid, which can be saponi¬, fied only with extreme difficulty, may be hydrolyzed simply by dissolving in, concentrated H2S04 and diluting this solution with ice water/5 Apparently,, concentrated sulfuric acid has converted mesitoic acid and its ester to inter¬, mediates that are far more readily attacked than are the respective conjugate, acids, , intermediates in which the “blocking effectiveness” of the ortho-methyl, , groups has been greatly decreased. The mesitoyl cation, VII, should be such an, intermediate, for the —C=0 group is linear and lies in the plane of the benzene, nng; attacking reagents may approach the carbonyl carbon perpendicular to, the plane of the ring, essentially without hindrance. Indeed, we have already, d" J\Amr> ^ S°c” 63’ 1556 (194°); 63> 3463 (1941)., ’ f, example, Branch and McKittrick, J. Am. Chem Soc 45 321 <4 923V, "Newman, J. Am. Chem. Soc., 63, 2431 (1941)., *, ’, (1923)<, , „J, , 1 reffers and Hammett, J. Am. Chem. Soc., 59, 1708 (1937).
Page 342 :
326, , Reactions of Carboxylic Acids and Esters, , seen (p. 99) that cryoscopic determinations of the number of particles per, “molecule” of mesitoic acid or ester in concentrated sulfuric acid indicate, that both solutes yield just this cation—an ion that, upon dilution with, water or alcohol, should once more form the parent acid or ester. The nearly, cone, , Ms-COOH or Ms-COOMe, , h2so4 >, Ms-COOMe, (Ms=H3C, , ), , Ms-COOH, VII, complete conversion of an acid or ester to an acylium ion is only rarely ob¬, served, for such an ionization apparently requires both steric assistance (that, is, a very crowded —COOH or —COOR group) and stabilization of the posi¬, tive charge by electron-donating groups. (Yet, such groups should not be too, effective as electron donors lest the molecule suffer sulfonation.) Thus, 2,4dimethylbenzoic acid and, , 2,4,6-tribromobenzoic, , verted to their respective acylium ions in, , acid are not measurably con¬, percent H2SO4, whereas, , 100, , 2,6-, , dimethylbenzoic acid is only partially converted/7 However, it should not be, assumed that because conversion of an acid or an ester to its acylium ion is not, detectable cryoscopically, none occurs. A reaction involving the slow and in¬, complete formation of an acylium ion is, in fact, far more suited to kinetic study, than are the interconversions of mesitoic acid and its esters in the manner just, described (for in these, the rate of conversion is generally governed by the delay, in adding a third reagent to a mixture of the remaining two). A kinetic study, of the hydrolyses of methyl benzoate and methyl />-toluate in sulfuric acid con¬, taining small amounts of water suggests that these reactions proceed through, acylium ions/* These hydrolyses are first order in ester but independent o, , furic acid. Like ntesitoic este*. i., converted to its half ester in 65 percent yie, y, ,he solution to cold water. (See c°rey X da., , !, , 5902 (1952).), (, >, « (a) Experiments by Graham and Hughes, reported by ing, ,, , also Leisten, J. Chem. Soc., 1956, 1572., , S, , . 5, , 771. (« See
Page 343 :
327, , Further Substitution Reactions on Acyl Carbon Atoms, , H, , Me, H2SO4, , PhCOOMe, , Ph—C—O, , h2o, , Ph-C=0, , slow, , fast, , -)-, , Ph-C-O, faSt, , H, , o, , H, , o, , the concentration of added water (when the latter is kept below 1.0 M). The, half lives for these reactions at 20° are 7.7 hours (for PhCOOMe) and 2.0, hours (for/>-CH3C6H4COOMe); hence the concentration of ac.ylium ion at any, point in the reaction is far too small to be detected cryoscopically., Evidence of a different sort indicates that the Aac 1 mechanism operates, in the hydrolyses of /3-lactones in strongly acid solutions (but not, as we shall, presently see, in weakly acid solutions). The hydrolyses of /3-propiolactone, and its homolog, /3-butyrolactone, in 2 to 5N H2SO4 or HCIO4 occur with, acyl-oxygen cleavage; for if H2018 is used, almost no O18 enters the /3-OH, group of the resulting acid.*9 Unlike, , 7-lactones, , (and presumably ordinary, , esters also), the rates of hydrolysis of /3-lactones are proportional to Hammett’s, h0 function/5 a proportionality that, we have learned, indicates that the acti¬, vated complex in the slow step consists only of a molecule of substrate and a, proton. The following mechanism is thus consistent with the known facts about, this hydrolysis:, ch2ch2, , CH0CH0, , fast) eq, , I, , K, , o—c=o, , I, , o—c=o, , slow, , _, , +, , H2O, , -» HO—CH2CH2—C=0 —>, fast, , ki, , ' +, , *H, , H, HO—CH2CH2—C—O, +v, o, rate = k1(LH+) = *i/C(L)(H+), , (3), H, , = k'(L)au+ — = k'(L)h0, Tiu+, , 7zh +, , (L = lactone), As with hydrolysis of methyl benzoate in concentrated sulfuric acid, the rate, of reaction is simply the rate of unimolecular breakage of the acyl-oxygen bond., , Further Substitution Reactions on Acyl Carbon Atoms. The Formation, and Hydrolysis of Amides, We have now examined just half of the known mechanisms for ester hydrolysist ose involving substitution on the acyl carbon. Before taking up the remaining, rr?’, atoms"/, so, , far, , H, , lnv0lve Substitution, , the alkyl carbon, let us consider a, , ad lUOnal reactlons that are. in essence, attacks on acyl-carbon, , he-le 77 ln SOme resPects- surprisingly similar to the ester hydrolyses, , •• o:c;;b:d„F:, much, Ison and Hyde, e;a;p'e>, J. Am. Chem., Soc., °f, 63,what, 2459 has, (1941)., , con^^ L jj
Page 344 :
328, , -, , Reactions of Carboxylic Acids and Esters, , modes of ester hydrolysis (by the BAC2 and Aac2 mechanisms) may also be, applied to the hydrolysis of amides. The following similarities stand out:, (a) Rates of basic and acid hydrolysis of many amides at moderate concentra¬, tions of base and acid have been found to be proportional, respectively, to, (amide) X (OH~) and (amide) X (H+).so, (b) The hydrolyses of N-substituted amides necessarily involve acylnitrogen, rather than alkyl-nitrogen, bond breakage. If this were not so, alco¬, hols, rather than amines, would be formed in such solvolyses., (c) The basic hydrolysis of Ph—C—NH2 in ordinary water is accompanied, , II, Ol8, , by the loss of O18 to the solvent,57 just as are the saponifications of acyl-labeled, benzoates (p. 317), demonstrating the existence of an intermediate (IX) having, a sufficiently long lifetime to allow proton exchange between oxygen atoms., (d) Both acidic55(o) and basic55(6) hydrolyses of amides are subject to steric, retardation by bulky groups, just as are the hydrolyses of esters., (e) The basic hydrolysis of amides is accelerated by electron-attracting, groups, as is the saponification of esters, but polar effects in the acid hydrolyses, of amides are slight.50,55(o), In line with these similarities, it seems likely that the basic hydrolysis of, amides proceeds as follows:, , o, HO- +, , O, , C—NHR, , ,, , fast, , slow, , <=>, , HO—C—NHR', R, , R, , IX, , o, , RCOO, fast, , HO—C + NHR'-, , I, , R'NH2, , R, and that acid hydrolysis takes the path:, "H, O, , H, R—ON—IV, O, , H, , H, , h2o, -), slow, , O, , H, , ', fast, , RCOOH, , R'NH2, , R'NH+, , R—C—N—R', O, , + fast, , R—C—OH 2, , H, X, , 30 See, for example, Reid, Am. Chem. J., 21, 284 (1899)., , __ 34g n955)., si Bender, Ginger, and Kemp, J. Am. Chem. Soc., 76, 3350 (, ),, ,, 0 953)« (a) Reid, Am. Chem. J., 24, 397 (1900). (b) Cason, et at., J. Org. Chem., 18,, (
Page 345 : Further Substitution Reactions on Acyl Carbon Atoms, , 329, , The overall reactions in both cases are virtually irreversible: the first because of, the conversion of the acid to its anion, the second because of the conversion of, the amine (or ammonia) to its conjugate acid., In concentrated acid solutions, rates of amide hydrolysis are found to, pass through a maximum and the value of the hydrogen ion concentration cor¬, responding to this peak is, in general, different for each amide.ss(o) This maxi¬, mum appears to correspond to the point where virtually all of the amide is, converted to its conjugate acid, and further addition of strong acid merely, serves to “tie up” the water present in the solvation shells of H+ and the anion,, decreasing the quantity of “free” water available for attack at the carbonyl, carbon., The observed rate law for the basic hydrolyses of N-methylanilides,S5(6), rate — ^(OH ) (anilide) + £3(OH )2(anilide), suggests that these hydrolyses are occurring by two paths. The second-order, term corresponds to the ordinary hydrolysis through an intermediate analogous, to IX, whereas the third-order term probably results from hydrolysis proceeding, through conjugate base XI, the concentration of which is proportional to, (OH-)2., Ph, R—C—, , \, , ', , Me, , Ph, , HO, , r, , -, , o, \, , ., /, 1_, , -, , third-order, , 1, , 1, , 2, , OH-, , ph n 2, , —U, , \, , 1, , 1, , —U-, , 04, , o-, , ", , term, , K, n>, , H, , O, , -> RCOO- + PhNHMe, , \e_, , O, , XI, second-order term, One additional feature deserves mention. During the acid hvdrol , •, , situat^n^differenT^rom^that^obscrved1^ be'WCen S°IVen‘ ^, , r, , <a, , This means that the breakup of cadol X “o formlmTne R'NH^ ^, rapid than the transfer nf nmr, c, amine R NH, is much more, hs breakup to form water. In fact TsTd'T, evidence at present that cation X is, an activated complex, , " <«) Krieble, J- C/um. Soc., 1957, 2000. (6) Beichler, , °X)'gen, , al°m to another in X and, , * C ^ .analogy’ there IS no significant, ° ^ lnte™diate. It may merely be, , 7 7, ^, , », , ^ ^ ” SO'Uti°n “ f°™, Am’CAem, ,Edw"d “d Meacock., Am- Lhem■ Soc., 79, 4927 (1957).
Page 346 :
330, , Reactions of Carboxylic Acids and Esters, , (although such reactions are known at elevated temperatures). Instead, the, preparation of amides is most often carried out by treating an ester, an acid, anhydride, or an acyl halide with an amine or with ammonia. The formation, of amides from acyl halides is often very rapid and in other ways unsuited for, convenient kinetic studies, and the reactions of anhydrides with amines (per¬, haps with less justification) have, until very recently, also been slighted. The, conversions of esters to amides have been investigated by a number of workers, and it is now recognized that these are very similar in character to ester saponifications and basic hydrolyses of amides. They are accelerated by electronattracting groups,34 retarded by bulky groups,35 necessarily involve acyl-oxygen, cleavage, and their kinetics (with suitable interpretation) suggest bimolecular, rate-determining steps. Unlike basic hydrolyses of amides and esters, however,, the conversions of esters to amides are reversible, and may thus be represented, as follows:, , O, , HOI ■+■, , slow, , RNH2 + C—OR", I, , O-, , I, , fast, , R—N—C—OR", fast, , R', , I, , I, , I, , "f", , ^, , fas*, , RNH—C—O—R" *==*, fast, , HR', XII, , |, , |, , slow, , R' H, XIII, , o, RNH—C + R"OH, , I, , R', , The “suitable interpretation,” mentioned parenthetically above, may become, necessary because the ester is attacked, not only by the amine, but also by the, conjugate base of the amine, RNH". Although this anion should be present in, low concentration, it is far more nucleophilic than the amine itself and, molecule, for molecule, it should be a much more efficient attacking reagent. For example,, the rate law for the conversion of PhCH2COOMe to PhCH2CONH2, using, ammonia in methanol,35 suggests that ammonia and the amide ion are compet¬, ing for the ester molecules in the initial step. It may be shown (Ex. 3) that sue, a competition leads to a rate expression having two terms, the first proportional, to (NHj), the second proportional to (NH3)94., ., When an ester is converted to an amide in an aqueous solvent, apprec.abl, hydrolysis may accompany amidization, for although water is less nucleoph., 1946 (.948); 7t, 1245 (.949); 72, 5635 (1950)1 75,, , "50”Betts and, , Hammett,V., , Am. Chm. Soc., 59,, , 1568 (1937).
Page 347 :
Further Substitution Reactions on Acyl Carbon Atoms, , 331, , than the vast majority of amines, it is present in large excess. Of particular, interest are the reactions of esters with tertiary amines; here the intermediate, XIV (analogous to XII) cannot be converted to an amide, for this requires, loss of a proton from the nitrogen atom and a tertiary amine has no proton, to lose. In this case, intermediate XIV may decompose into its parent ester, and amine, or it may lose OR"~ to form an N-acylammonium ion:, , o+, , R3N + R'COOR", , O, , r3n, , I, , R3N—C—OR", , I, , R', , R3N—C—R', N-acylammonium, ion, , R'COOH, R"OH, , XIV, , Now, N-acylammonium ions derived from tertiary amines are very readily, hydrolyzed. Hence it is quite possible that if the departure of OR"- is suffi¬, ciently rapid compared to the reversal of the original association step, the overall, reaction (which is simply an ester hydrolysis) will be faster than hydrolysis of, the ester through direct attack by a water molecule. As pointed out in the pre¬, ceding chapter, an effective leaving group must be only weakly basic—that is,, it should be the conjugate base of a relatively strong acid. We would thus ex¬, pect the amine-catalyzed ester hydrolysis to take place most readily if R"OH,, the alcohol or phenol from which the ester is derived, is unusually acidic. It has’, ■n fact, been shown that the hydrolyses of the esters of /.-nitrophenol and 2 4 ’, din.trophenol are strikingly catalyzed by such amines as trimethylamine, hvdnr0o,ne’HPyKidine’ and the piCOlineS" Typically- ^-nitrophenyl acetate b, as fast ini m M, , ^ faS* In °'01 M a£Iueous Me*N a"d four times, , the presence 0^“, 1, ^ ^, are *°‘ due *°, affected bvadd-t, /V, S°lutl°ns’ f°r the observed Elyses are not, (which, m f, ?, “"Jugate acids of the respective tertiary amines, centra,, f OWCT 6 COncentration °f OH- in solution but leave the con, entratton of amine unchanged). The heterocyclic base imidazole XV •, Particularly effective catalyst; for the hydrolysis in 0 01 Af, ’ *S 3, almost 100 times as fast, *, y, y S in 0.01 M aqueous imidazole is, , J, , may lose a pToton f, N-acetylimidazok, , XVI, , V, , ^ Wa'er', , whT, , thi$ C3Se> the N-acetyl derivative, , °f **“ imida2ole ri"S> yWding, , 245 nZ The lane, ‘, ”ay be detected spectroscopically X, a®, nM). the latter is considerably more stable th.n XT, ,, \ V max 1, a"d It is possible to alter the kinetic characteristics of ^, arge excess of imidazole; for then intermedilte XV., r "T'0” ^ ^ 3, hydrolysis becomes the rate-determining step The, raP‘dly a"d "S, ■’ Bender and TurnquK, y ^, ^ TheSe ami"~atalyzed hydrolyses, quest, J. Am. Chem. Soc., 79, 1652, 1656 (1957).
Page 348 :
332, , -, , Reactions of Carboxylic Acids and Esters, , \, N, N-, , /, H, , _, , °, -f C-OAr, Ale, , O, , \, , II, , O, _TT +, , N—C, N -J+, , /, , H,0, , —-—>■, , N-C, \, Ale, , N=/, , \, , Ale, , H, XV, , XVI, , A, N—H + AleCOOH, N-, , are important because of the possibility that they may constitute relatively, simple models for hydrolysis reactions catalyzed by such enzymes as a-chymotrypsin and papain. Indeed, the imidazole ring (a section of the amino acid,, histidine) has been proposed as a vital part of the active catalytic sites in these, enzymes.57,55, , Hydrolysis and Alcoholysis of Acyl Chlorides, Along with the substitution reactions on the acyl-carbon atoms of acids, esters,, and amides, we should consider such reactions on the acyl carbons of acid, halides. Our glimpse will be brief, however, for these reactions, in many ways,, resemble the substitution reactions of alkyl halides described in the preceding, chapter.55 As with alkyl halides, a duality of mechanism and a range of border¬, line cases are recognized. In the alcoholyses of substituted benzoyl chlorides,, for example, both extreme mechanisms appear to operate within the same, reaction series/0 Reaction rates for monosubstituted benzoyl chlorides follow a, trend typical of those bimolecular nucleophilic substitutions in which the bond¬, making, , process governs the activation energy; specifically, the reaction is, , accelerated by substitution of halo and nitro groups in the benzene ring but, retarded by substitution of a p-CH3 group. In line with this trend, we would, expect the alcoholysis of mesitoyl chloride to be extremely slow, both because, of electron repulsion by the three methyl groups, and because of steric hindrance, by those methyls in the 2- and 6-positions. Since this alcoholysis is, on the con, trary, “immeasurably fast,” we may safely assume that the Avl mechanism, 38 Doherty and Vaslow, J. Am. Chem. Soc., 74, 931 (1952)., 88 The reactions of acid anhydrides are, in a number of respects, similar in character ro, those of acid chlorides. However, considerably less quantitative data exists on the former group, of reactions and they will not be discussed here. For recent work on acid anhydrides see paper, by Gold and co-workers, Trans. Faraday Soc., 44, 506 (1948), J.C/iem. Soc., 1953,, 1416, by Berliner and Altschul, J. Am. Chem. Soc., 74, 4110 (1952), and by Denney anc, baUr*o Norris^ancfco-workersj J. Am. Chem. Soc., 57,1415, 1420 (1935); 61,1418 (1939),, , ,, , >
Page 349 :
333, , Hydrolysis and Alcoholysis of Acyl Chlorides, , has taken over; for reactions proceeding by this mechanism are subject to steric, assistance and to acceleration by electron donation'. Similarly, the hydrolysis, of mesitoyl chloride in 95 percent aqueous acetone is also unimolecular, for its, rate is not affected by addition of hydroxide ions.^, It further appears that, as with some alkyl halides, the solvolyses of acyl, chlorides may undergo a change in mechanism as the polarity of the solvent is, changed. The relative rates of solvolysis of the following /-substituted benzoyl, chlorides, for example, suffer an almost complete reversion in order as the, reaction is transferred from 40/60 ethanol-ether (dielectric constant 13) to, 50/50 water-acetone (dielectric constant 53).** Note that in the alcohol-ether, , Table 9-2. Effect of Solvent Changes on Relative Solvolysis Rates, of /-Substituted Benzoyl Chlorides, Relative Solvolysis Rates, Medium, , D, p-N02, , p-Br, , H, , />-ch3 />-ch3o, , 40% EtOH + 60% Et20, , 13, , 32, , 2.5, , 1, , 0.47, , 0.25, , Pure EtOH, , 22, , 22, , 2.1, , 1, , 0.70, , 0.81, , 50% HoO + 50% Me2C=0, , 52, , 12, , 0.92, , 1, , 2.9, , 30, , mixture, the trend in reactivities is much the same as that observed for the, saponification of esters and the basic hydrolysis of amides—typical bimolecular, processes. When the reactions are transferred to 100 percent ethanol,, , the, , p-methoxy compound moves into the borderline region, since it now reacts, more rapidly than the /-methyl compound. In 50/50 water-acetone, the solvolysis of only the /-nitro compound appears to have retained its bimolecular, Trt h):dr0lrses of ,he lining acyl chlorides have become disgroups, , ’, , arC n°W SUbjeC‘ t0 aCceleration by electron-donating, , as st^TiC ry °f uhe hydr°lyses and alcoholyses of acyl halides is not, , c r ul v, , zz, T, , 7 m‘S, , bC deSired’ f°r °ne mUSt Sdec‘ a sokent rather, , ::::L b0out in media, 7 containing, e thewater, roie °forthe, almhnl, , «a, , -action „, oo ~, rate u,ir, L, °, OI alc°ool as a major component the, te law may be noncommittal on this point On the other hnnH, fu, is carried out ;n * ™, i, ,, ther hand’ lf the reaction, water, the kinetic orde, ,;, e,IC °rder IS, 4t Rrown and Hudson, J., rown and Hudson, J., , 7, “l’ addmg measured quantities of alcohol or, °ften masked by the medium effects arising as a result, Chem. Soc., 1953 3352., Chem. Soc., 1953* 883.
Page 350 :
334, , -, , Reactions of Carboxylic Acids and Esters, , of adding a polar liquid to a nonpolar one.45 To date, the most informative, kinetic results have been obtained from experiments in ether. The ethanolyses, of acetyl chloride45(o) and ^-nitrobenzoyl chloride44 in ether have been found to, be second order in ethanol over a large concentration range. Thus, the activated, complex in these reactions may be presumed to consist of a molecule of acyl, chloride, two molecules of alcohol, and an unknown number of molecules of, ether. Recalling Swain’s proposal that both “push” and “pull” are necessary, in many (if not all) nucleophilic substitutions (p. 299), we are tempted to picture, one of the alcohol molecules as attacking the carbonyl carbon while the other, pulls at the chlorine atom through hydrogen bonding., , H, , \ -—-3a /-*, +, C—Cl-H-OEt--*, O, I, /, , Et, , O, , H, , I, /, V, o--c, , /, Et, , + [Cl-HOEt], , o, , Before assuming such a mechanism to be general, however, one would like, to know whether similar rate laws govern the reactions between other acyl, chlorides and other alcohols and whether, in particular, a molecule of phenol, (which should form stronger hydrogen bonds to chlorine than does alcohol), can replace a molecule of alcohol in the transition state when the reaction is, carried out in the presence of both phenol and alcohol., , The Claisen Condensation, The Claisen condensation is related to the saponification and “amidization, of esters. In saponification, a hydroxide ion attacks the acyl carbon of the ester,, in “amidization,” an amine or its conjugate base attacks. In the Claisen con¬, densation, the attacking reagent is a carbanion, obtained by removal of a slightly, acidic hydrogen from a ketone, a nitrile, or, most often, from another molecu e, of ester In the cyclic counterpart of the Claisen condensation, the Dieckmann, reaction, the acyl carbon of a -COOR group is attacked by a negatively charged, carbon somewhere else in the same molecule, forming a new C-C bond an, a new ring., * Typically, the rate of ethanolysis of acetyl chloride in CC1;, tion of ethanol, but the reaction cannot be snd to, orders” with respect to ethanol may be obtained y p, rates against the logarithms of ethanol concentrations, , gmeasuri, g, , the slope of the resulting, P, solutions to, , curve, but such “apparent orders” are found, J, &*., 1955, 4121., unity in more concentrated solutions. See ( ), hvdrolvsis of benzoyl chloride in, Similarly, .he “apparent order” with respect to, 1 Vfltreen acetone l near seven, water-acetone mixtures has been found to vary from, in 50 percent acetone; see « H^dsonand Archer J. Om. Soc., .950,, u Ashdown, J. Am. Chem. Soc., 52, 269 (193 ).
Page 351 :
The Claisen Condensation, , +, , —C:, , |-, , /, , -c-c—c, II I II, o, o, , C—OR', , o11 i, , 335, , R, , R, -C—C“, , -, , A, o, , G—OR->- —C-C, , I, , I, , O, , +, , OR, , II, , +, , OR', , (Claisen), , (Dieckmann), , V, , O, , An important requirement for the Claisen condensation is that carbanions, be available, at least in small concentration. Although the factors favoring the, stability of carbanions are not to be discussed until the following chapter, we, might note here the evidence that carbanions are present in significant quan¬, tities in solutions in which the Claisen condensation is occurring. It is now well, recognized that esters, ketones, and nitriles having one or more a-hydrogen, atoms undergo hydrogen exchange at the a-position when dissolved in “labeled”, ethanol, C2H5OD, containing NaOEt/5 indicating the existence of the following, mobile equilibria:, , H—C—C— + OEt~, , I, , II, O, , I, ~:C—C— + EtOD v, , and, , I, , : C—C-b EtOH,, , I, , II, O, , I, D—C—C— + OEt-, , II, , II, , O, , O, R, , Furthermore, optically active esters of the type, , ^CH—COOEt are racemized, , by ethoxide ion," the asymmetry of the a-carbon atom presumably being lost in, , the carbanion,, , ^C'—COOEt, R', , Having established the existence of carbanions in the reaction mixture, , ::zzz;, --srr, - * -»-, , cr, , ^:trisionaiiy) for the cuisen «-*■««« a, , CH*, , COOEt + OEt- ^ “:CH2, , COOEt + EtOH, , ^ «•, , *>3 (19401.
Page 352 :
336, , Reactions of Carboxylic Acids and Esters, , o, , O, , II, EtO—C, , +-:CHoCOOEt, , EtO—C—CH2COOEt, , I, ch3, , I, ch3, XVII, , o, EtO- + C—CH2COOEt, , ch3, CH3—C—CH2—COOEt + OEr, , CH3—C—CH—COOEt', , o, , + EtOH, , O, , Here, we cannot say whether anion XVII is a true intermediate or merely an, activated, , complex., , (Moreover,, , experiments in which the acyl oxygen is, , “labeled” cannot answer this question, for even if XVII were long lived, there, is no path by which the acyl oxygen can exchange with solvent oxygen.) Note, that in the final step, the product, acetoacetic ester, is converted by base to its, anion. Since this /3-ketoester is almost as strong an acid as phenol, this conversion, should be nearly complete if an excess of ethoxide (which is a somewhat stronger, base than OH~) is used. The success of the Claisen condensation, as it is usually, carried out, depends upon this final step which removes the desired keto ester, from the reaction site; for this condensation is known to be reversible,^ and if, the final proton transfer could not occur, the equilibrium involved would, allow only a small degree of conversion. Ethyl isobutyrate is not appreciably, converted to keto ester XVIII by sodium ethoxide since the latter has no, OEt", , O, , "V, , -^V, , 2(CH3)2GH—COOEt', , ;(CH3)2CH-C, , C(CH3)2, , COOEt + EtOH, , \ MsMgBr orPh3CNa, , x-—', , XVIII, , a-hydrogen atoms and therefore cannot be converted to an anion by NaOEt., However, as indicated, the preparation of XVIII can be carried out ustng a, stronger base than ethoxide, such as sodium triphenylmethtde, Na+CPh„ or, mesitylmagnesium bromide, XIX (which, although a Grignard reagent, is too, MgBr, , X, , (MsMgBr, XIX), , Me, See, for example, Kutz and Adkins,, , J., , Am. Chem. Soc., 52, 4391 (1930).
Page 353 :
337, , The Cleavage of (5-Diketones, , hindered to undergo the typical Grignard addition reactions).^ The use of, these very strong bases assists the reaction in two ways. The removal of the, very weakly acidic a-hydrogen from isobutyric ester is accelerated, and, what, is more important, the resulting keto ester is partially converted to the anion, ~:CMe2—C—CMe2—COOEt, by removal of a y-hydrogen, with the resulting, , II, , O, favorable effect on the equilibrium., , The Cleavage of ^-Diketones, The basic cleavage of /?-diketones is essentially a reversal of the Claisen con¬, densation, for here an OH- or OR- ion attacks an acyl carbon, displacing a, carbanion. Let us compare the following cleavages, for which rate laws have, been determined.^9, , 9, , Me O, , I, HO, , +, , O, , II, , Me O, , II, , C—C—C—Me —HO—C, Me Me, , Me, , I, , II, , |CH3COO_, , + “:C-C-Me, , Me2CH— C—Me, , Me, , o, , XX, , 9., EtO_+, , o, , o, , o, , c—CH2— C-Me ==—- EtO—C + ~:CH2 —C—Me, Me, , r;Ie, , The first of these reactions, rendered irreversible because of the formation of, , the acetate ion in the final step, takes place readily, and its rate is simply pro¬, portional to (OH-) X (diketone).^) The second cleavage of acetylacetone is, addid’ V OUf,We Sh°Uld PerhapS eXpeCt the °PP°site. reasoning that the, ster a lvlnT d 8"T w ^ &St diket°ne Sh°U‘d retard the a«ack, both, tencally and inductively. Moreover, the kinetics of the second reaction appear, ^ lh‘ dik“0ne °r ,hl eth0Xii° i0n " in “*«• W <he diketone is in, S *h?.,reactl0n [ate 15 ProPortional to added ethoxide but independent, o added diketone, whereas if e.hoxide is in excess, the reaction is first orderTn, diketone but zero order in ethoxide ww tk;, i •, •, , of cleavage is oronnrtin, , ,, , relationship suggests that the rate, , ^, , nearly compL, , «■. ^ mTSS “d MayerlC', , ■4m-, , 73,, , 926, , (,951). (6) Pearson and Sandy>
Page 354 :
338, , Reactions of Carboxylic Acids and Esters, , the number of moles of this species would be determined by whichever of the, two reagents is not in excess. Immediately the conjugate base of acetylacetone, CH3—C—CH—C—CH3<->CH3—C=CH—C—CH3\ comes to mind, and, , 1, , -, , II, , 1, , 00, , 11, , o_, , o, , the kinetic picture is consistent with a mechanism in which the rate-determining, step is the reaction of the anion with a molecule of solvent. Yet this cannot be,, for dimethylacetylacetone, XX, which cannot lose a proton in the same way,, reacts more rapidly than acetylacetone (both with EtO~ and OH-). Let us, how¬, ever, recall a fundamental limitation of kinetic studies: they cannot tell us, whether the rate-determining step in a reaction sequence involves one group, of reactants or a second group in rapid equilibrium with the first. In particular,, the anion XXI is in equilibrium with the diketone and ethoxide, and a reaction, rate proportional to the concentration of XXI is necessarily proportional to, the product (OEt-) X (diketone):, O, , II, CH;, , C—CH2—C—CH3 + OEt- ^ CH3—C=CH—C—CH3 + EtOH, M, , II, , o, , o, , I, , o_, XXI, _ _anion_, eq, , (diketone) (OEt-), , and,, , (4), , rate = A(anion) = k'(diketone) (OEt ), , The rate law is thus equally consistent with a rate-determining attack by, ethoxide on the diketone., O, , O, ., , Et0- + C—CH2—C—Me, Me, , O, , O, , ||, , II, , EtO-C, , EtOH, , + -;CH2-C-Me-., fast, , Me, EtOAc + Me2C—O + EtO, , The kinetic complications arise because these two species may react in an alter¬, nate way—that is, by simple proton transfer. This is why acetylacetone is, cleaved much more slowly than dimethylacetylacetone, XX. In the, compound, the alternate reaction path, which greatly lowers the concentrations, of the reacting species in the cleavage reaction ^s closed.^^, For an unsymmetrical /?-diketone, R, , C, O, , 2, , n, O, , ^ ^ of, ’
Page 355 :
Esterifications and Hydrolyses, , -, , 339, , cleavage, leading to two different acids (or esters), are possible. If the substrate, exists very largely in the keto (rather than in one or both of the enol) forms,, the preferred position of attack is easily predicted; for a nucleophilic OH, or OR- ion will be most likely to attack the more electrophilic of the two carbonyl, groups. The acid RCOOH should be formed in preference to R'COOH if, R— is a stronger electron attractor than R'—; this is equivalent to saying that, the cleavage should yield, as the principal product, the stronger of the two possible, , acids (or its ester).50(o) When considerable enolization of the diketone occurs,, the situation becomes more complicated, for, although there is only one pos¬, sible enolate anion, there are two possible enols: R—C=CH—C—R' and, OH, , O, , R—C—CH=C—R'. Since, it appears, attack takes place on a keto (rather, O, , OH, , than on an enol) carbon atom, the first of these two forms shown is protected, from attack at the carbon adjacent to R— but not adjacent to R'—, whereas, the reverse is true for the second form. Thus, the mode of attack will depend, not only on the electrophilicities of the carbonyl groups in the keto form, but, also on the relative amounts of the two enol forms present in the reaction mix¬, ture. If substituent R—favors enolization more than does R'—, then R'—COOH, will tend to be formed rather than R—COOH. Thus the cleavage of benzoylacetone, PhCOCH2COCH3, yields mainly acetic acid,50(6) rather than the, stronger acid, benzoic acid. This is in line with our knowledge that a phenyl, group promotes enolization more effectively than a methyl group (Chap. 10)., , TLten!‘Catl0ns and Hydro,yses Proceeding Through Carbonium Ions., Tne Aal] Mechanism5*, Returning now to ester hydrolysis and esterification, let us consider the acidcatalyzed hydrolyses of the esters of tertiary alcohols. If we apply to such esters, the same types of tests that, for the esters of primary and most secondary alcoos,, , indicate or suggest acyl-oxygen cleavage, we obtain different answers, , es,aehrCh hydr0‘r °f ''bUtyl aCCta‘e in H2°18 yidds labekd '-butyl alcohol,, establishing, in this case at least, alkyl-oxygen cleavage. 62, 50, , ^.,®r2a,d^36an09M)b'rtS°n’, , ^^, , Davi‘?tdaK^n^,°QfSriX,2oT(^56t, , I926> 2356; <*> Kut* and Adki-,, , a"d reU“d, , resul,i„gUraho°iXled wht, O^bumhuta’ T h’50' ’l"* hydr°l>'sis' >he, , -ha„ge does occur, it U . much siower reaction, , J- An,, «, acid, , an
Page 356 :
340, , Reactions of Carboxylic Acids and Esters, , (CH3)3C, , -o—C—CH3, , H+, , + HoO18, , (CH 3) 3C—018—H + CH3COOH, , O, A similar, but more definite, conclusion arises from the acid-catalyzed hydroly¬, , sis of the optically active acetate, XXII. The resulting carbinol, XXIII, is?, almost completely racemized.53 This shows not only that alkyl-oxygen cleavageEt, , O, , I, , I, , (CH 3) oCHCH ,CH 2CH2—C—O—C—CH 3-*, water dioxane, , Me, , Et, , I, , XXII (active), , i-C6H13—C—OH + CH3COOH, , I, , Me, 90 percent (racemized), XXIII, , has occurred, but also that the reaction is unimolecular—that is, that it proceeds;, Et, , through the carbonium ion, i'-C6Hi,—C+. If it were bimolecular—that is, at, Me, , direct displacement of acetate by water—inversion of configuration about the-, , asymmetric carbon, rather than racemization, should occur. Assuming that:, acetate XXII and /-butyl acetate are representative esters of tertiary alcohols,, we may represent the acid-catalyzed hydrolysis of esters in this class as follows:, H, , H, —R'COOH, slow, , R'—C—O—CR3, +, , + HjO, fast, , —- R3C+ s + R'COOH, fast, , +', , - RaC—O, , -HsO, slow, , N, , H, , o, , Although the reaction is represented as being reversible, esterifications of tertiary;, alcohols are seldom carried out in the manner indicated. However, the esteri¬, , fication of the optically active secondary alcohol, 2-octanol, using sulfuric act -, , in a large excess of acetic acid, appears to take place, at least in part,Jby t e<, , AAl\, , mechanism, for extensive racemization accompanies this reaction., Both the, , Aac1, , and the, , Aal\, , mechanisms are, quite naturally, facilita, , by strongly acidic solvents, but the structural features within the ester whi, favor the two mechanisms are different. The breakage of the acyl-oxygen, « Bunton, Hughes, Ingold, and Meigh, Nature 166,,679 (1950)., , it Hughes, Ingold, and Masterman, J. Chem. Soc., 1939, », , •, , ■
Page 357 :
Esterifications and Hydrolyses, , -, , 341, , O, to form an acylium ion [ R—C : 6: R' 1 is favored by the presence of electron-, , I, , H, attracting substituents in R' and electron-repelling substituents in R, whereas, the reverse is true if the alkyl-oxygen bond is to be broken, forming a carbonium, , R'+ 1. Bearing this in mind, we can rationalize (although we, , probably would not have predicted) the order of the hydrolysis rates for alkyl, benzoates in concentrated sulfuric acid.55, (CH3)3C—OBz > (CH3)2CH—OBz > CH3—OBz > CH3CH2—OBz, The hydrolysis of methyl benzoate in concentrated sulfuric acid proceeds, as, we have noted (p. 326), by the AAc\ mechanism. Ethyl benzoate is apparently, hydrolyzed via the same route, since the substitution of an electron-repelling, methyl group for an a-hydrogen in the alkyl section of the molecule decreases, the hydrolysis rate by about 50 percent. But the substitution of two, and,, more particularly, three methyl groups for a-hydrogens in the alkyl section, , boosts the reaction rate again. The trend is, of course, very much like that for, the basic hydrolyses of the corresponding alkyl bromides (p. 311, Ex. 11a), and, strongly suggests a change in mechanism; that is, f-propyl and /-butyl benzoates, are hydrolyzed via the Aal\ mechanism (where electron-repelling substituents, in the alkyl group aid reaction), rather than via the Aac 1 mechanism (where, such substituents inhibit the reaction). In line with this supposition we are not, surprised to learn that the electron-attracting />-nitro group retards the hydrolysis, of ethyl benzoate by a factor of 63, whereas the same group accelerates the hy¬, drolysis of z-propyl benzoate by a factor of 200.55^, The intervention of alkyl-oxygen cleavage in the reactions between esters, and alcohols may be detected without difficulty by simply identifying the products Acyl-oxygen cleavage in such reactions would result in ester interchange, , Z,, , °Ut lf ^‘oxygen cleavage occurs, the ester is converted instead to an, ether and a carboxylic acid., xv, , R'+, , +, R—C—O-, , A i, O, , ^, , H, , un, , * R'—O—R" + H+, , R', R—COOH, , Tc! brate and triphenylmethV> -etate, when treated, , ' 'J. Am. Chem. Soc., 71, 1575 ('1940') • on T, ♦, r, ’, ’, Leisten, J. Chan. Soc., 1956, 1572.
Page 358 :
342, , Reactions of Carboxylic Acids and Esters, , with methanolic HC1, have been found to yield, respectively, i-BuOMe and:, Ph3COMe, together with the respective parent acids.56, , Alkyl-Oxygen Cleavage in Neutral and Basic Media. The BAL1 and, BAL2 Mechanisms, , The unimolecular cleavage of esters at the alkyl-oxygen bond is, in essence, ar, , Sn\ reaction in which the carboxylate ion is the leaving group. On this basis, , we may predict a number of characteristics of this mechanism: (a) It should be, , observed only for those esters in which the alkyl group forms a relatively stabl., , carbonium ion—that is, esters of tertiary alcohols and certain special secondary, , alcohols such as benzhydrols and a-substituted allyl alcohols, (b) It should be, , facilitated by electron-repelling substituents on the alkyl group (which woulc, stabilize the carbonium ion intermediate), and by electron-attracting substitu, ents on the acyloxy group (which would tend to ease the departure of the latter), , (c) Aside from an expected positive salt effect, the rates of reactions proceeding, , by the BAL1 mechanism should not be increased by addition of OH-, (d) Thu, rate constant should be subject to decrease by the mass-law effect (p. 256) a:, the reaction progresses, (e) This mechanism should be favored by media o, high dielectric constant; in particular, the rates of such reactions carried ou, , in aqueous acetone, aqueous alcohol, or aqueous dioxane should depend upon, , the concentration of water but should not be proportional to it. (f) This mecha¬, nism should not be subject to steric hindrance, (g) When such reactions occur, at an asymmetric carbon atom, extensive racemization should occur, (hi, , Finally, when an ester of a substituted allyl alcohol is hydrolyzed by this, mechanism, partial allylic rearrangement (p. 286) should be observed., , The stereochemical studies of reactions proceeding by the BAL1 mechanism, , have, to date, been concerned largely with the half esters of phthalic acid (these,, , besides being esters, are acids and may generally be resolved directly, using the, conventional basic resolving agents). For example, the hydrolyses of the opticall>, , active half esters, XXIV, XXV, and XXVI, in dilute aqueous NaOH yield the, , corresponding racemic alcohols, and XXIV also yields a racemic alcohol m, , ION NaOH.57 Again, the stereochemical results point both to ummolecu an >, , and alkyl-oxygen cleavage. If, however, half esters XXV or XXVI are hydro¬, , lyzed in concentrated base, the resulting alcohols are formed with retention, of configuration, indicating that the more usual BAC2 mechanism has taken ove, Indeed, , it is quite likely that hydrolysis by both mechanisms is occurring |, , dilute base, but that hydrolysis by the BAL\ mechanism is so fast that the alterna, « W Cohen and Schneider, 7., , Gum. Soc, 63, 3382 (1941). (A) Bunton and Wood, , * ^Kcny’olTalT&m. Soc, 1936, 85, 576; (b) 1942, 605; « 1946, 803, 807.
Page 359 :
343, , Alkyl-Oxygen Cleavage in Neutral and Basic Media, , mechanism is difficult or impossible to detect. As the concentration of base is, progressively increased,, , hydrolysis, , by, , the BAC2, , mechanism, , is, , necessarily, , accelerated, but that by the BAL\ mechanism is very nearly unaffected, and the, rate of the former may overtake, then pass, the rate of the latter.55 (With half, ester XXIV, which forms an exceptionally stable carboniuin ion, BAL1 hy-, , MeO, , COO, , XXV, COO, , CH3, , I, , ch3ch= ch- ch—o—c, *, , o, XXVI, drolysis is presumably so rapid that it predominates even in concentrated base.), We may likewise understand why the half ester XXVII, when hydrolyzed in, concentrated base in aqueous methanol, yielded its parent alcohol, but why,, when treated with dilute base in the more polar solvent water, it yielded both, its parent alcohol and the isomer formed by allylic rearrangement.59, Me, PhCH= CH-CH-OH, , {fiAC2), , O, PhCH=CH-CH-0-C—/—\, , ', , Me, COOH, XXVI1, , fPhCH-CH=CHMe, [phCH=CH-^CHCH3]—<, , (b, i), AL. *, , OH, , PhCH=CH-CH-OH, , I, , Me, M, , product is racemic, , the rattTof jlc '‘n l'rmer ground if it were shown that when the hydrolysis, , hydrolysis o^S?- (OH"); tha,, »ncen,?tion range where raccmi2a^nl^aS ‘, hut by iess than a first-order dependence PTo 7h, experiments have not, as yet, bee„Pdescribed, rvenyon, Partridge, and Phillips,, , J. Ckem. Soc.,, , “ ««, , OH-; and that in the, , °f, , hVdrolys‘s varies with (OH-),, , 8' °f, 1937, 207., , PreSent ai»h°r>
Page 361 :
Alkyl-oxygen Cleavage in Neutral and. Basic Media, , 345, , between about 2 and 8. In strongly basic solution, this cyclic ester undergoes, saponification by the BAC2 mechanism; and, as we have seen (p. 327), hydrolysis, in strongly acid solutions proceeds by the less ordinary Aac 1, , mechanism., , Hydrolysis in H2018 at extreme pH values thus yields /3-hydroxybutyric acid, with essentially no O18 in the /3-OH group,29 but at intermediate pH values, this, hydroxyl group in the resulting acid is labeled. It might appear that here we, are simply observing another example of BAl\ hydrolysis, but for one addi¬, tional observation: hydrolysis at intermediate pH values results in inversion, of configuration about the /3-carbon \6S that is, the resulting hydroxy acid is the, enantiomorph of that resulting from hydrolysis in strong base or in strong acid., h3c, , \ /, H20* + Z)-, , H, , h, c—ch2, I, , L-, , o*'+, , I, , o—c=o, , Me, , H, , C—CHo—C—O- —, , I, , II, , H, O, L- HO *—CHMe—CH2—COOH, , XXVIII, It is likely that /3-propiolactone is also hydrolyzed by this mechanism at inter¬, mediate pH values, but the absence of the stereochemical criterion in this case, makes us less certain., It has been remarked that alkyl-oxygen cleavage in ester hydrolysis is, favored by electron-attracting groups or atoms in that section of the molecule, derived from the acid. Since such groups invariably increase the strength of, acids, a rather good rule of thumb (not without exception) is that esters of the, strongest acids are most likely to suffer alkyl-oxygen bond breakage in hydroly¬, sis. This is consistent with our knowledge that the alkyl esters of sulfonic acids,, in most of their ordinary displacement reactions, undergo cleavage of the alkyloxygen bond rather than the S-O bond. Similarly, it has been shown that the, basic hydrolyses of a number of alkyl nitrates,-55 sulfates,-" and chromates67, (t at is, esters of strong acids) take place, at least in part, with alkyl-oxygen, c eavage. On the other hand, esters of somewhat weaker acids—for example,, 2rr, Plates - and hypochlorites^-undergo basic hydrolysi, with alkyl-oxygen bonds intact., y, "Okon and Miller, J. Am. Chem. Soc., 60, 2687 (1938), , 73’ 4273 “«’>■ ^ another, methoxide to yield dimethyl ether see Bunnett rT bctWeen methy1 benzoate and sodium, (1950)., y Ctner> see Bunnett> Robinson, and Pennington, ibid., 72, 2328, ee Burwell FrTv\ and Shadan’ J- Am- Chem- Soc., 77, 2512 (1955), BurweH and Holmquist, ibid., 70, 878 (1948)- 74 1462 (19571, , •' A.tab0™ “'I7’ Samud a"d, „, , cn> J• Chem. Soc., 1954, 1869., , J• cL sJ.’ l954 ’, , 3603, , •30U3>, , Blumenthal and Herbert, Trms. Faraday Soc.. 41, 611 (1945)
Page 362 :
346, , -, , Reactions of Carboxylic Acids and Esters, , Decarboxylation, Almost any carboxylic acid, RCOOH, can be made to suffer decarboxylation, if it is treated severely enough (provided, of course, that it is not first destroyed, in some other way). However, the decarboxylation of an acid should occur, most readily if, within group R—, there is a strongly electron-attracting substituent, such as —NO2, —CCI3, —C=N, or —C=0. This is to be expected, for a, decarboxylation is ordinarily a heterolysis in which group R— departs with an, electron pair /R—jC—O- —» R:_ + O-C, , 0N, , ( io, One of the first decarboxylations to be studied kinetically was that of, acetoacetic acid.70, CH3—C—CH2—COOH, , CH3—C—CH3 +, , o, , co2, , o, , This reaction, as well as the closely related decomposition of a,a-dimethylacetoacetic acid into methyl isopropyl ketone and C02,7/ follows a rate law of, the following type:, rate = k (keto acid) + k! (keto-acid anion), , (5), , This suggests that there are two distinct modes of decarboxylation, neither of, which involves the enol form of the acid or its salt (since a,a-dimethylacetoacetic acid cannot exist in an ordinary enol form). It seems likely that the two, terms in the rate law correspond simply to unimolecular decompositions of the, keto acid (equation 7) and its anion (equation, , ), respectively. There is one, , 6, , minor difficulty: with both keto acids, k (for the keto acid itself) is much greater, than k' (for the anion), whereas one would expect the proton on the —COOH, group to inhibit, rather than to accelerate, decarboxylation. It is likely, there¬, fore, that there is a partial transfer of the carboxyl proton, through intra¬, molecular hydrogen bonding, XXIX, to the keto group, where it should aid, decarboxylation., , C^-C-CH-jC-^or ^ [cH-C-CHr-CH-^H, , o, , o, , o, , OXXX, , + H,0, *, fast, , (CH3)2C=0 +, , Widmark, Ada m,i. Scaad., 53, 393 (1920); Chm. Abstr.,1IS. 2763 O’21)„ Pederson, ./. Am. Om. Sac., 51, 2098 (1929); 58, 240 (1936)., , OH, , W
Page 364 :
348, , Reactions of Carboxylic Acids and Esters, , far out of the plane of the remaining three. (This is a case covered by Bredt's rule,, which stipulates that except for very large ring systems introduction of a double, bond at the bridgehead of bicyclic systems is prohibited.73) Since acid XXXII, cannot, except with rearrangement, yield the necessary enol or an enolate ion, intermediate, it resists decarboxylation., The rates of decarboxylation of a number of additional acids are propor¬, tional to the concentrations of the respective carboxylate ions, and, in the, absence of evidence to the contrary, we may assume that these anions undergo, heterolysis to a carbanion and C02 in the same manner as the anions of acetoacetic and a,a-dimethylacetoacetic acids. Among these acids are a-nitroacetic, and a-nitroisobutyric acids,74 dibromomalonic acid,73 phenylpropiolic acid, (Ph—C=C—COOH),73 the trihaloacetic acids,73-77 and 2,4,6-trinitrobenzoic, acid.73 In the first three of these cases, the carbanion intermediates have been, “trapped” with Br2 (as with the 0-keto acids). With none of these is the rate of, decarboxylation of the acid itself appreciable; but malonic acid, for which a, hydrogen-bonded structure analogous to XXIX may be drawn, decomposes, almost ten times as rapidly as its monovalent anion., Furthermore, a number of nitrogen-containing acids exist that undergo, 74 Bredt, , Ann., , 437, 1 (1924). This rule does not apply to systems such as XXXIV in, , which one of the bridges is merely a covalent bond. Neither does it apply to systems having, bridges with five or more atoms, for these are sufficiently flexible to conform to the geome ry, of the double bond without excessive strain in the ring. Thus., readily at 250° in quinoline (Prelog, et al, Helv. Chun. Acta 31, 92 (1948), 32, 1284, although the enol derived from it has a double bond at a bridgehead., , n Pederson, J. Phys. Chem., 38, 559 (19^)74 Muus, ibid., 39, 343 (1935); 40, 121 (1926)., 76 Fairclough, J. Chem. Soc., 1938,, ” Verhoek, et al, J. Am. Chem. S»c„ 56, 78 Verhoek, J. Am. Chem. Soc., 61, 186 (, )■, nitrobenzoic acid constitutes a furter in ica 10"_b, carboxylation is a unimolecular splitting o, order in carboxylate ion is consistent, , 79, , 299 (1950), decarboxylation of 2,4,6-trirate.determining step in the de-, , late ion A decarboxylation that is first, , mechanism but it does not demand such a, which the rate-determin-, , mechanism. Such a rate law *^?_C°Ton the ciSxvl group of the free acid; for, as, ing step in the attack of an O, 10, •, is directly proportional to the product, was pointed out (p. 338), the concentration o, mechanism for decarboxylation should be, (acid) X (OH-). However ^the -COOH group is, subject to steric hindrance, an, ’ ’, rWarhoxvlation slowly or not at all. Since this i, shielded from outside attack should un ergo, f, t,iis acid is untenable, and unless, , —- *“•also be Msum, to be unimolecular.
Page 365 :
Decarboxylation, , -, , 349, , first-order decarboxylation themselves but form anions that do not.79 Typical, of these are quinaldic acid (XXXVI) and thiazole-2-carboxylic acid (XXXVIII)., Since these acids are in equilibrium with their respective zwitterions (XXXVI', and XXXVIII', respectively), kinetics alone does not tell us which form of the, acid is undergoing decarboxylation. Since, however, the dipolar ion XXXIX, (which cannot tautomerize to a “non-zwitterionic” form analogous to XXXVI), is found to undergo decarboxylation readily, it seems very likely that for the, other acids also, it is the zwitterion that is being decarboxylated. This is as it, , XXXVIII, , XXXVIII', , XXXIX, , should be, for an added positive charge on the carboxylate group should hinder, decarboxylation, but an added positive charge on the remainder of the molecule, should aid decarboxylation. Here again, the anionlike intermediate XXXVII, may be trapped, this time by addition of an aldehyde or ketone.79** Decar-, , slow, , CO,, , r2c=o, , XXXVI, , boxylation should take place even more readily if the_C, beta to the — COOH grouD- for in tV, ,, h, C, , „, •, •, grouP 1S seated, , results directly in neutralization orchard deCarb°Vation of the zwitterion, , •, , m.257,
Page 366 :
350, , Reactions of Carboxylic Acids and Esters, N /, N- C, , v\, , -->, , +, , ^, , 1, , /, —N=C—CH, \, , —, , o, , i, , \/, fC, , II, o, II, , H, i, , i, /, —NH—C=C, , 0, , 1, , -G, , 2-Quinolylacetic acid, XL, and 2-thiazolylacetic acid, XLI (shown as the, zwitterions), are acids of this type and are somewhat more easily decarboxylated, than acids XXXVI and XXXVIII.79 Similarly, we see why primary (but not, , H, XLI, secondary or tertiary) amines may catalyze the decarboxylation of /?-keto, , /\, acids/0 The keto group is converted by the amine to an imine, , _, , N, ^—, , linkage, and the resulting /3-amino acid (for example, XLII) presumably, , O, Me-C—CMe,—COOH, , Ph—N==CMe, -COs, , PhNH5, , ^ ===, fast, , H, , (CMe,, , .CV, , /, , slow, , "O^C, , o, XLII, Me, H20, Ph—NH—C=CMe2, , fPhNH2, , fast > l Me—C—CMe2, O, , undergoes decarboxylation more rapidly than the original keto acid., , H, In the, , example shown, the decarboxylation of a,a-dimethylacetoacetic acid is oun, to be over ten times as fast in 0.05 molar aniline as in water A companion, between the intermediate for the catalyzed (XLII) and uncatalyzed (XXIX), reactions indicates why this should be so. Since the imino nitrogen ts far more, to Pederson, J. Am. Chem. Soc., 60, 595 (1938).
Page 368 :
352, late,, , -, , Reactions of Carboxylic Acids and Esters, , XLIV, can be demonstrated without a “trapping reagent,” since this, , enolate is unusually stable, and its formation and destruction can be followed, spectrophotometrically.sl(c), The decarboxylations of an additional class of acids, /?,7-unsaturated acids,, are formally quite similar to those of /3-keto acids and /3-imino acids. This can, best be seen by comparing XLV below to structures XLII and XXIX. (Note, the shift of the double bond, leading to rearrangement.) Such decarboxylations, , R—CH=CH, , R-CH—CH, , t>R., , H, , -c, II, o, , heat, , >200° >, , I, , ^, , H, , CR2, , +, , o=c=o, , XLV, , require rather severe conditions, for here the acid cannot exist in a zwitterionic, form (as can jS-imino acids), nor can there be appreciable proton transfer by, intramolecular hydrogen bonding at the, , 7-carbon., , It is also likely that the, , decarboxylations of a:,/3-unsaturated acids take the same path, , that is, that, , these acids first arrange to /3,y-unsaturated acids5 for it has been shown that a, number of a,/3-unsaturated acids are in mobile equilibrium with the correspond¬, ing /3,7-unsaturated acids at temperatures necessary for the decarboxylations,, provided that interconversion between the two requires transfer of only a, pro ton.82, 900°, r—CH2—CH=CR'—COOH, , R—CH=CH—CHR'—COOH —>, R—CH2—CH=CHR' + C02, , Thus what appears to be a decarboxylation without rearrangement is instead, a pair of rearrangements, the second, in a sense, nullifying the first. We should, then expect those a,0-unsaturated acids that cannot rearrange by a proton, shift to d,7-unsaturated acids to undergo decarboxylation only with great, difficulty. Indeed the acid (CH^C-CI^CH-COOH remains unchanged, after 2 hours’ heating at 300V5, Like nucleophilic substitutions,, , decarboxylations display a duality, , reaction mechanism. The large majority of decarboxylations are ummolecu ar, but a number of bimolecular decarboxylations are now known These are, effect, displacements of the carboxyl group (without its pair of bonding elec¬, trons) by a proton-that is, S„2 reactions-and almost always take placein, strongly acid solutions. Such reactions generally occur at unsaturate, .. Linstead,, al„ J. Cheat. So,, 1925, 616; .929, ™*-***™;, « Arnold, Elmer, and Dodson, J. Am. Chem. Soc., 72,, (VW).
Page 369 :
The Decarboxylation of Silver Salts with Halogens, , p, S*H+ +XC-COOH, , H-Cyc/, , A, , / \, , £, y \, , 353, , O, \, ^4h-G + C + H+‘, , V-H, , A °, , {Sv — solvent), XLVI, atoms; for in such cases, the new C—H bond may form without necessity for, simultaneous breakage of the old C—C bond. We should expect them to be, first order both in carboxylic acid and in H+, and they should be favored by, electron-donating substituents and aromatic rings bound to the /3-carbon, since, such groups should stabilize the intermediate carbonium ion, XLVI (and,, presumably, also the transition state leading to XLVI). The decarboxylations,, for example, of acids XLVII,*4(o) XLVIII,^(6) and XLIX,*^(c) in strongly acid, solutions, appear to be bimolecular., , COOH, , I, XLVII, , XLVI 11, , XLIX, , The Decarboxylation of Silver Salts with Halogens, Silver salts of carboxylic acids may be converted to alkyl or aryl halides by, treatment with elemental bromine or iodine in an inert solvent. In this, the, so-called Hunsdiecker reaction, C02 is evolved, but the reaction is related in only, a formal sense to the decarboxylation reactions that we have thus far consid¬, ered. There is strong evidence that the Hunsdiecker reaction usually (although, perhaps not always) proceeds by a mechanism involving free radicals.*5 The, initial step in the reaction is heterogeneous (for the silver salt is generally not, appreciably soluble in the solvent used) and we will not attempt to speculate, as to its intimate details. The net result of this step, however, is that only a por¬, tion of the halogen is precipitated as a silver halide. If care is taken to keep the, mixture cold so that further steps in the transformation do not occur, an equal, , rrr se: in;°iution may be sh°wn to, , °*^ng zt, , halogen; If an olefin be added tQ the ^, Rudin, Hdv. Chim. Ada,, ” For a, , brief survey of this evidence,, , j'j ^rv^' f' Schenkel and Schenkelse^Johnson and^Inghtunj GiirrLfievs.^56^ 250^(1956).
Page 370 :
354, , Reactions of Carboxylic Acids and Esters, , halogen atom becomes affixed to one double-bonded carbon, whereas the, carboxylate anion becomes affixed to the other.56 In short, the halogen in such, solutions has “positive” character—much like the halogens in hypohalites—, and such solutions are generally considered to contain acyl hypohalites of the, O, , I, , type R—C—O—Br and R—C—O—I, although no pure acyl hypobromites or, , O, hypoiodites have, as yet, been isolated., , O, — AgHal, , R—COO-Ag+ + Hal2, , R—C—O—Hal, , —>, , R—Hal + C02, , The decompositions of the “hypohalites” in solution to form C02 and alkyl, or aryl halides exhibit a number of features typical of free-radical reactions, (Chap. 16). They are promoted by radiation57 and exhibit induction periods.55, In addition, the “side products” arising from such reactions can be easily, explained by only assuming the intervention of free radicals. For example, when, silver benzoate is treated with Br2 in, , CCI4,, , the expected product, bromobenzene,, , is formed, but appreciable quantities of chlorobenzene and BrCCl3 may also, be isolated.59 These products point to the following reaction sequence:, Ph—COOAg + Br2 -> Ph—C—O—Br, , O, , cci, , PhCl + CC13- 5 BrCCli, , When the Hunsdiecker reaction is carried out on silver picolmate (L) m hot, nitrobenzene, 2,2'-dipyridyl is one of the products, strongly suggesting the, pyridyl radical as an intermediate.90, , Br,, COOAg, , “ See, for example Edward, and, ^“’JdHscher, Ann., 146, 4S, n Bockemuller and Hoffman, Ann., 519, 16b U J&)ie, (192*«Conley, J. Am. Chem.Soc 75, 1148 (1<^53)3, (1950), « Dauben and Tilles, J. Am. Chtm Soc^72,l\V> (1, ), »o Kuffner and Russo, Monatsh., 85, 1097 (1954).
Page 371 :
The Decarboxylation of Silver Salts with Halogens, , -, , 355, , Similarly, bibenzyl is obtained when silver trifluoroacetate is treated with bro¬, mine in toluene;5' the following reaction sequence is indicated:, Br?, , CF3COOAg -> CF3—C—O—Br ->, , O, , PhCH3, , dimerizes_, , _, , BrCF3-, , ^ HBr + PhCH2-> Ph—CH2, , co2, , Ph—CH2, , At present, then, the most satisfactory mechanism for the Hunsdiecker, reaction (barring the initial formation of the acyl hypohalite, about which we, know very little), is simply, , B 2 II Br-, , RCOOBr, , -> Br2 -)- C02, , etc., , R-->, , If this mechanism is correct, we should expect the silver salt of an optically, active acid—in which the a-carbon is asymmetric—to yield, upon treatment, R', with halogen, a racemic halide; for the radical R—C-, , should be symmetric, R", , (p. 149). Virtually complete racemization has indeed been observed when the, Hunsdiecker reaction is carried out on n-Bu—CHEt—COOAg,®*W PhCH2_, CHEt—COOAg,5*<6> and Et-CHMe—COOAg.«w, It is interesting that the silver salts LI and LII, having bridgehead carboxylate groups, readily undergo the Hunsdiecker reaction.55 If the proposed, ^, , Me, , Me, , » M Tldi"e and SharPe> J■ chem. Soc., 1952, 993, , onal case has been described by Arcus, , Camnbell, , . ,,, , lat aPPears to be an excep-, , Here ,he reaction of /f-Me-CHPh-cooT/w 'h T, e^°n> J' Chm' S°‘- 19«. 1510., *‘th « Percent inversion of configuraiblfoThe ‘TV” “ reported to yield a bromide, alternate mechanism for the Hunsdiecker, *■H extent that these results are correct an, recent att, to repeat ,h' ^b' “T"*' However, a more, » wna, , "on'of "he expected MeCHPh^, , 72. 5228‘(tST WmSt°n’ ^ ^ Chm' S«- 75’, , ', , ' °'s- Chem ’ «*. 1570, 0953); Cope and Synerholm M
Page 378 :
362, , Reactions of Carboxylic Acids and Esters, , The conversion of the adduct to the observed products, an amide and a carboxylate, ion, may then proceed by one of the following two paths:, o-, , I, , O, , *,, , ||, , Path 1: Ph—C—O—C—Ph ^ Ph—C—O" +, , I, , I, , O, , *», , C—Ph, , II, , NH.R, , >, , |, , O, , NH,R, , +, , +, PhCOO- + Ph—C—NHR, , I, O, , OI, , Ornh2, , ,, , I, , kh, , Path 2: Ph—C—O—C—Ph-> RNH+ + Ph—C—O—C—Ph ->, , I, o, , I, , II, , nh2r, +, , o, , I, nhr, , o, Ph—C—O- + Ph—C, II, O, , I, NHR, , (a) Benzoic anhydride having just one carbonyl group labeled with O18,, Ph—C—O—C—Ph, , I, , o, , II, , o*, , o*, is prepared from Ph—C—Cl and PhCOOAg. This singly labeled anhydride is, treated with NH3 at —33°., , One half, , of the labeled oxygen is found in the benz-, , amide formed. With which of the two mechanisms above is this consistent?, (b) When the anhydride, labeled as above, is treated with aniline in ether, only one, third of the labeled oxygen is found in the amide, the remaining two thirds being, in the benzoate. With which of the two mechanisms above is this consistent?, , (c), , Explain., When the experiment in (b) is repeated, using aqueous acetone as a solvent, about, 45 percent of the labeled oxygen is found in the benzanilide. Account for the, , (d), , variation in behavior as the solvent is changed., The results of an experiment using a deuterated cyclohexylamine, C6HU, , , in, , 2, , ether are essentially the same as those for aniline in ether, but if nonlabeled cyclo¬, hexylamine, C.HnNH,, is used, about 40 percent of the labeled oxygen ,s found, in the amide. Explain., , (e), , If labeled />-nitroberizoic benzoic anhydride, Q2N, , V, , \, , o*, , o, , ii, , ii, , C-O-Cr-Ph’, , is treated with aniline in ether, 53 percent of the labeled oxygen is found in the
Page 380 :
364, , -, , Reactions of Carboxylic Acids and Esters, , (h) The rates of hydrolysis of alkyl acetates in IN HC1 lie in the order:, /-Bu— > Me— > Et—, , > z-Pr—, , (i) The saponification of /?-NH2—CeH4—COOEt is accelerated by incorporating a, methyl group meta to the —COOEt group but retarded by a methyl group ortho to, the —COOEt group., (j) 2,4,6-Trimethoxybenzoic acid cannot be esterified in the same way as mesitoic, acid., (k) A mixture of Ph—COOEt, Me2CH—COOEt, and NaCPh3 in ether yields the, keto ester Ph—C—CMe2—COOEt after standing a short time, but yields the, , I, , O, keto ester Me2CH—C—CMe2COOEt after standing several days., O, (1) Acid LIV undergoes decarboxylation much more slowly upon heating in H2S04, than does acid LV., , CMe=C(COOH)2, , LV, , fVa, LVI, , LVI I, , (m) Silver salt LII (p. 355) yields a mixture of bromide LVI and chloride LV II when, treated with bromine in CC14, but yields only LVI when the reaction is earned out, fn) The bTsk hydrolysis of CF3CONHPh is not first order in OH", but tends toward, a limiting value at high hydroxide concentrations, the observed rate law being, L2(OH-) (amide), rate =, , 1 + /f(OH“)“
Page 381 :
CHAPTER, , lO, , Ccxrbanions and Enolization, , The C—H bond is relatively strong and its polarity is ordinarily very slight. The, removal of a proton from a —C—H linkage to form a carbanion, —C:~, should, , therefore be difficult; indeed, a hydrogen atom bound to a carbon may, in, most cases, be considered to have negligible acidity. This need not be so if one, or more strongly electron-attracting groups lie near the C—H bond under, consideration, or, more particularly, if removal of the hydrogen ion leaves a, carbanion in which the negative charge may be spread over a number of atoms, rather than being confined to a single carbon. Given one or both of these con¬, ditions, conversion to a carbanion may be significant, and, as we shall see, a, number of important reactions may proceed through carbanion intermediates., , Ionization of Carbon-Hydrogen Bonds and Prototropy, If dilute sodium hydroxide is added to a solution of a /3-diketone (such as, acetylacetone), to a /3-keto ester (such as acetoacetic ester), or to an aliphatic, mtro compound (such as nitromethane), an equimolar quantity of base is con¬, sumed, indicating a neutralization. Unlike the conventional neutralizations, owever, these reactions are not immeasurably fast, for each involves the breakage, a, H bond and requires appreciable activation energy. In each of these, cases, ionization of the C-H bond is favored both by the presence of adjacem, electron-attracting substituents (\=0, -COOEt, or -NO,) and by the, delocalization of negative charge in the resulting anion., 365
Page 382 :
366, , -, , Carbanions and Enolization, -HH, , ch3—c—ch—c—ch3 <->, , ch3—c—ch2—c—ch3->, II, , o, , Ao, , II, , o, , oi, , CH3—C=CH—C—CH3 ♦-> CH 3—C—CH=C—CH 3, , I, , II, , o_, , II, , o, , I, , o, , o_, , —H+, , CH3—C—CH2—C—OEt —4, , II, , o, , II, , CH 3—C—CH—C—OEt, , o, , O, , O, , CH3—C=CH—C—OEt, , o_, , o, , O, , O, , ->, , \, , o, , ✓, -:CH2—N, , _o, o-, , O, —H+, , CH3—N, , CH3—C—CH=C—OEt, , CH2=N, , /, , o J, , O, , The original diketone, keto ester, or nitro compound may be regenerated by, acidifying the solution of the respective anion; but in the third case, the con¬, version of anion to nitroalkane is slow and may be shown to proceed through, , _, , OH, , /, , an isomeric intermediate—a so-called an-nitroalkane, CH2—N, , —which, , \, O, , is very much more acidic and more highly colored than the original nitroalkane., The nitro-fla-nitro pair constitutes an example of tautomerism—that is,, the coexistence of two (or more) compounds that differ from each other only, in the position of one (or more) mobile atoms and in electron distribution., It is also an example of the more specific phenomenon of prototropy—that is,, tautomerism in which interconversion between forms may be achieved (at, least in thought) merely by the shift of a hydrogen ion and a redistribution of, electron density. As previously pointed out, both acetylacetone and acetoacetic, ester represented above as their keto forms, may also exist in tautomeric forms—, the ends, CH3-C=CH-C-CHJ and CH3-C=CH-COOEt.* (For both of, OH, , O, , OH, , i For detailed treatments of tautomerism see (a) Wheland, Advanced Organic Chemistry^, John Wiley and Sons, Inc., New York 1949 pp. 580-646; <*> Ingold■, £$££, in Organic Chemistry, Cornell University Press, Ithaca, N.Y., 1953, pp. 473 529, and (, ), son, Quart. Revs., X, 27 (1956)., _ _, • Additional tautomeric forms of these compounds, such as CH2 C, , °, CH, , OH, , °", , 2, , _„_CHiand, ^, , 0, , J, , c—CH=COEt, are conceivable, but there is no convincing evidence that they exist
Page 383 :
Ionization of Carbon-Hydrogen Bonds and Prototropy, , -, , 367, , these keto-enol pairs, interconversion between the forms is far more rapid than, between the forms of nitromethane under comparable conditions.) Further¬, more, a sample of acetylacetone, acetoacetic ester, or nitromethane will, unless, special precautions are taken, consist of an equilibrium mixture of the two, respective tautomeric forms., Since those compounds forming the most stable carbanions are, with a, few exceptions, those which exist in tautomeric forms, the question of carbanion, stability is tied in with the phenomenon of prototropy. The prototropic forms of, a compound (let us say a ketone, ke, and an enol, en) necessarily have a common, conjugate base. Any attempt to determine the acid strength of form ke, using, the conventional method of measuring the pW of a partially neutralized solu¬, tion, will give instead an apparent ionization constant, Kapp, which is related, to the ionization constant of the keto form, Kke, by the equation:, Kke = K,app, , (ke) + (en), , (1), , (ke), , Thus, Kapp will approach Kke when the keto form is present in much larger, concentration than the enol; but even if the mixture is 90 percent enol, the, apparent equilibrium constant will differ from the acidity constant of the ketone, only by a factor of 10., , Table 10-1. Apparent Acidity Constants for Some Acids in Which, the C—H Bond Undergoes Ionization, Acid, , pKa, , CH2(CN)2, , 11.2, , Acid, ch3—c—ch2—c—ch3, , O, CH 3—C—CH Me—C—CH 3, !l, ii, 3, , o, , 11.0, , CH3—c—CHoCOOEt, , II, , 10.7, , 9.0, , O, , CH3—C—CH2—C—CF3, -*, o, , 4.7, , o, CH2(N02)2, , 3.6, , I, , o, , pKa, , II, , o, , o, ch3—no2, , 10.2, , CH3—C—CH2—C—Ph, , A, o, «ve, , o11, , 9.4, , hc(no2)3, HC(S02CH3)3, , <1, <1, , apparent^ values (water, 25°) for some representa-, , ;-~that “■ actds m whieh a C-H bond slowly undergoes
Page 384 :
368, , Carbanions and Enolization, , ionization.5 All but the last of these may exist in tautomeric forms, and to, change the given values to actual pKa values, the concentrations of the indi¬, vidual tautomeric forms in solution must be known (p. 376). In each case, however, the predominating form of the acid in aqueous solution is that listed,, so the apparent pKa values are only slightly greater than the values corrected for, tautomerism. It is to be noted that the relative strengths of these acids are, governed in part by the same factors that determine the strengths of the more, ordinary acids in which the acidic hydrogen is derived from an O—H bond., For example, substitution of three electron-attracting fluorine atoms for terminal, hydrogens in acetylacetone boosts its acidity by a factor of 20,000. Conversely,, substitution of a methyl group on the center carbon of this diketone lowers its, acidity by a factor of about 100, an effect due partially to the electron-repelling, action of this group and partially to steric inhibition of resonance in the anion, (as described on p. 378). More subtly, acetoacetic ester, in which one of the, carbonyl groups is bound to an —OEt group, is only one fiftieth as strong an, acid as acetylacetone. Since this carbonyl group acquires electron density as a, result of conjugation within the ethoxy group (F), it can absorb a smaller meas¬, ure of negative charge from the negative carbon in the anion (I") and is thus, less effective in stabilizing the anion. As with carboxylic acids and phenols, the, nitro group is more strongly acid strengthening than are the acetyl and cyano, , o, o, o), o, Et—o-c-c—C-CH3 Et-o^-c—c—C-CH3, I, , (o _ o, Et-O-C-C-C-CHJ, , I, , I, , groups; indeed, the three nitro groups attached to a single carbon in trimtromethane make it a strong acid. The bicyclic diketone II is not appreciably, more acidic than ordinary monoketones (for which PKa values approach 19 or, 20), since the conjugation-stabilized anionic form of II would have a doub e, bond at a “bridgehead,” in violation of Bredt’s rule (p. 348)/, , -O', , » These apparent pKa values have been compiled by Pearson and Dillon, J. Am. Chem., Soc.y 75, 2439 (1953). See this paper for ^dlt^n^9r3C3fe(7"^', l Bartlett and Woods, J. Am. Chem. Soc., 62, 2933 (1J4U;.
Page 385 :
Ionization of Carbon-Hydrogen Bonds and Prototropy, , 1 ', , -, , 369, , /T\, , The sulfonyl group I —S— I is also strongly acid strengthening, a laige, , part of its electron-attracting action being due to its inductive effect. It is also, likely that a sulfur atom bound to a carbon from which a proton has been re¬, moved can absorb some of the negative charge into one of its outer d orbitals,, thus stabilizing the resulting anion. This type of delocalization is often repre¬, sented by structures such as III', in which the sulfur atom has “expanded its, , o, .., , O, , o, , I, , I, , -c-sI, , III, , I, , o, , —c=s-, , ., , 1, , o, , III', , -Hh, , o, , o, , V, , P, , a/l-7\/s:, Me, , O, , V, valence shell” since, by classical count, it now shares 10 electrons. However,, we should not expect the C=S “double bond” (which involves both d and, p orbitals) in III' to be a replica of an ordinary double bond between first-row, , elements (which involve only p orbitals). In particular, the requirement that, all atoms attached to a double-bonded pair must lie in or near a common plane, no longer holds. Thus, the trisulfone IV, unlike the diketone II, is strongly, acidic, presumably because anion V, despite the “double bond” at the bridge¬, head, is an acceptable structure.5, We must turn away from hydroxylic solvents if we wish to study the, acidities of such very weak acids as acetophenone, phenylacetylene, and triphenylmethane, for the anions derived from these acids are such strong bases, that they may not exist in appreciable concentrations in water or alcohol., (Such very weak acids may generally be converted to their conjugate bases by, treatment either with a very active metal or with the alkali metal derivative of, a still weaker acid-for example, butylsodium.) Absorption spectroscopy is a, ^ ‘r* °tr compar^nS acidities, for «he spectrum of a carbanion is differed, ‘ Pa"ent aC‘d (esPec,al‘y whl:n conjugation effects in the two species, are d.fferent), and the concentration of both the acid and its anion in a given, author, the evidence for’/c=S “double"'bond” ^ (1955)' In the °Pinion of the present, *rons 03 that which applied ,o sulfiteJ,, JV“ dfrived from -Ifones is no, as, argued, that the acidity of IV is due almost u n, ^xamP^e» P- 372). It may, in fact, be, sulfone groups adjacent to the bridgehead caTbon7akh^, inductive effect of three, an appreciably weaker acid than H—CfSO CH I ’l, oug 1 tllls does not explain why IV is, and merits further investigation., In any event, the question is a difficult one
Page 386 :
370, , -, , Carbanions and Enolization, , solution may generally be estimated. If one weak acid, HA, is dissolved in an, inert solvent containing the alkali metal derivative of a second weak acid,, say Na+.4'~, protons will slowly be transferred from HA to A'~, and eventually, equilibrium will prevail., HA + A'~, , A- + HA', , By measuring the concentrations of the four species involved in this equilibrium,, we can compare the acidities of HA and HA', just as we might compare the, acidities of two indicators (p. 103). Comparisons of this sort, which have been, carried out for solutions in ether and liquid ammonia, have only an approxi¬, mate significance, for the activity coefficients of the various species are unknown., We may not yet, in good conscience, put these acids on the same quantitative, scale as those studied in water or alcohol. However, experiments such as those, described enable us to list the following weak acids (and many others) in order, of decreasing acidity:^, , O, , II, , >, , EtOH >Ph —C —CH3 >Ph—C=CH >, VI, , VII, >Ph3CH>Ph2CH2, , A benzene ring adjacent to a C—H linkage facilitates the ionization of this, bond by its capacity to absorb negative charge in the resulting carbamon, (VIII <-> VIII'); however, it is not as effective in this respect as a carbonyl, group (Ex. 8a). A carbanion may acquire an extra measure of stability if its, parent acid contains a cyclopentadiene ring (for example, indene, VI, and, fluorene, VII). In such cases, the negative charge is distributed over each car-, , IX, , bon in the five-membered ring, as well as over the six-membered nng()., shown below in the canonical forms of anion IX. Although there are h«, benzene rings adjacent to the C-H linkages in trypttcme (X), this hydio, e Conant and Wheland, J. Am. Chem. Soc., 54, 1212 (1932).
Page 387 :
Ionization of Carbon-Hydrogen Bonds and Prototropy, , -, , 371, , is not appreciably more acidic than ordinary aliphatic hydrocarbons.7 Again,, , forms such as XI (analogous to VI11'), which might be expected to stabilize, the trypticide anion, are prohibited by Bredt’s rule., Although the alkali-metal derivatives of hydrocarbons are ionized in, ether, they are, as might be expected, largely nondissociated in that solvent;, that is, they exist mainly as ion pairs and other ionic aggregates.8 They appear, to be stiong electrolytes in liquid ammonia,9 with which, however, they react, slowly—for example,, Ph3C~ + NH3 -> Ph3CH + NH^Another indication that a compound forms carbanions, although perhaps, to a small extent, is the conversion of C—H bonds to C—D bonds when the, compound is dissolved in DjO or OD-labeled alcohol under basic conditions, (or, alternately, the conversion of C—D bonds to C—H bonds when the, deuterium-labeled compound is dissolved in ordinary water or alcohol). We, ave seen, for example, that hydrogen exchange at the a-carbon of esters in, the Clatsen condensation suggests the existence of a carbanion intermediate, “ this reaction (p. 335)., , Similarly,, , hydrogen exchange may be shown, , litHl7r "I, ,r “-Carb?n at°mS in mOSt ket0nes’ aldehydes, nitroalkanes,, well ; al, h, r U" er, conditions> and often in neutral solutions as, HCBr, , HPR r, , T, , " *hat 3 nUmber of the flhalomethanes (HCC1S, , show fog, UnderS° hydr°Sen, eXCha"*e, g thaftfe’, that these too °the-S), are in equilibrium, with carbanions., HCA\, , -h+, , ba- s°ltt«tons,/3®, , d2o _, , -» GXj-> DOT, + OD, , The trimethylsulfonium ion, (CH,)3S+ readily undergoes hydrogen exchange, , _ Kraus and Kahler, J. An,. Chen,. sL. si,'3I37 ('933), 827 0954)1, , ’, , 6°’ 193 <1938>-, , 73’ '376 (,951> « Hine, „, , at., «*„ 76.
Page 388 :
372, , -, , Carbanions and Enolization, , under similar conditions.*' Here, the intermediate is not actually a carbanion,, but rather a sort of zwitterion, XII, in which a partial neutralization of charge, is possible by electron drift from the carbon atom into the d orbitals of the, sulfur atom (XII')., , CH3—S—CH3—%, , ~:CH2—S—CH3, , I, , ch3, , CH2=S—CH3, , I, , ch3, XII, , I, , ch3, XII', , (Note that in form XII', the sulfur atom is associated with ten, rather than eight,, valence electrons.) Hydrogen exchange is similarly rapid for the cyclic sulfonium ion XIII, suggesting again that a carbon-sulfur bond having consider¬, able double-bond character may exist at the bridgehead of a bicyclic system., The tetramethylammonium ion, however, undergoes hydrogen exchange in, , basic solutions exceedingly slowly; the “zwitterion,” “:CH2—N(CH3)3, cannot, be stabilized by a form analogous to XII', since nitrogen cannot expand its, valence shell beyond eight electrons.", , Base- and Acid-catalyzed Halogenations of Ketones, The bromination or iodination of almost any enolizable ketone in the position, adjacent to the carbonyl group is accelerated by addition of bases. Such bases, are not catalysts in the strictest sense of the word, for they are stoichiometnca y, consumed; Nevertheless, these base-promoted halogenations are kinettcally, similar to reactions subject to general base catalysis, since if several bases ate, present, the rate law will contain several terms, one for each base (p. 112)., simplicity, let us first consider a system in which the action of only one base, B:, , is significant—for example, the bromination of acetone, promote, , OH- « The rate of this reaction is proportional to the concentrations of ketone, and base, but is independent of the concentration of bromine. This means that a 1 ou, bromine is consumed in the overall reaction, it becomes tnooloei after common, “ Mamalis and Rydon, J. Chem. Soc, 1955, 1049., i* Doering and Hoffmann, J. Am. Chem Soc., 77 521 (1955)., is Bell and Longuet-Higgins, J. Chem. Soc., 1946,
Page 389 :
Base- and Acid-catalyzed Halogenations of Ketones, of the rate-determining step., , -, , 373, , The rate laws for the chlorination75 and iodination^, , of acetone in the presence of hydroxide are the same as for its bromination;, moreover a given basic solution of acetone reacts at the same rate with chlorine, or iodine as with bromine. Clearly, the three halogenations have the same rate¬, determining step, the transition state of which contains acetone, an OH- ion,, and an indeterminate number of solvent molecules. A very simple mechanism, fulfilling these conditions comes to mind:, OH-, , :CH2—C—CH3 <-► CH2=C—CH3', , CH3—C—CH3 ^, , fast, , slow, , o, , o, , o_, ach2—c—ch3 + X-, , II, , o, Here X may be Cl, Br, or I, and the “indeterminate number of solvent mole¬, cules” is taken as zero. If the mechanism proposed for acetone also applies to, the bromination of the optically active ketone XIV, a given basic solution of, this ketone should undergo racemization in the absence of bromine at the same, rate that it would undergo bromination in the presence of bromine; for the, presumed rate-determining step in the bromination reaction (the conversion, of ketone to carbanion XV) destroys the asymmetry. Although the rates of, , O, Et, , CH, , G, , O, Ph, , [Et—C—G-Ph^-Et—G=C-Ph]—Et-CBr-C—Ph + Br~, , Me, , . I, , Me, , Me, , XIV, , Me, , O, , XV, , Et-CD-G-Ph, .Me, , have not been compared in water, they have, the base)116 And'that ^, aCetic acid <with ac«ate ™ added as, j, at ls not all> slnce the concentration of anion XV is small, h unTerwJ, r way. HCa‘i0n, It anion fv, XV ■is ^, srenerateH in °f f°™ati°"; *u, i, , z .-r4 -, , vH ^, Hsu, , Am', , Chem•, , and Wilson,, , Soc->, , 56> 967 (1934), , J. Chem. Soc.,, , 1936, 623.', , **, „ initial step, ^, , -.-Xit:; “
Page 390 :
374, , -, , Carbanions and Enolization, , pio\ ided reaction conditions (solvent, basicity, and temperature) are the same., It has, in fact, been found that the rates of racemization and deuteration of, ketone XIV are, within experimental error, equal in h^O-dioxane mixtures in, the presence of NaOD.76, The halogenations of enolizable ketones are also accelerated by acids, and, the characteristics of the acid-catalyzed reactions are somewhat the same as, those promoted by base. Such reactions have been found to be subject to, general acid catalysis;77(o) their rates are proportional to the concentrations of, , ketone and added acid, but independent of the concentration of halogen;17w the rate, of halogenation of a ketone wdth an asymmetric a-carbon atom is the same as, its rate of racemization in the absence of halogen ;77(c) and the rates of acidcatalyzed bromination and deuteration of a given ketone are the same in heavy, water.17(d) Again, we may conclude that bromination, iodination, racemization,, and deuteration proceed through a common intermediate, formed in a common, rate-determining step. Since we are now considering reactions in acid solutions,, we may be quite certain that this common intermediate is not the conjugate, base of the ketone; it is far more likely to be the enol form, XVI, of the substrate., I, , I, , H-C-C=0, , H.'l, , fast, , 1, , 1, , H-C-C = 0-H/J, |, , B’, , I, , I, , 1, , H-C1C=0-H-/-^fi:H+C=C-OH+^, , slow, , |, , I, , XVI, , I, , !, , Hal2orDzO, , then C = C—OH-;->■, , , ., , ., , ,, , ., , (2), , halogenation or deuteration, , lclSl, , XVI, , Note that the enolization is not simply a “proton jump” from the a-carbon to, the carbonyl oxygen atom (for this would allow no role for the acid catalyst)., Of the several items of evidence pointing to the rapidity of the initial step,, the most convincing is the very rapid oxygen exchange that ketones undergo, when dissolved in aqueous acidic solutions,75 an exchange that may be studied, using H2018. The oxygen exchange reaction of acetone, for example, is many, times as fast as its acid-catalyzed bromination under similar conditions. Now, this exchange, whatever its detailed mechanism may be, almost certainly pro¬, ceeds through the hydrogen-bonded complex (CH3)2C=0 • • HA, the same, type of intermediate as that proposed for acid-catalyzed enolization. Since the, it Hsu Ingold, and Wilson, J. Chem. Soc., 1938, 78. The deuteration of ketone XIV in, heavy water and its racemization in ordinary water would be expected ^an..different, procJed at slightly different specific rates; for the strength of a given base is shghtly different, in the two solvents, and their solvation charactenstics are not identical, , ^, , n (a) Dawson et at., J. Chem. Soc., 1926, 2282, 1928, 2844, 1, ,, ', Hammett, J. Am. Chem. Soc., 61, 2791 (1939). (c) Bartlett and Stauffer, ibid., 57, 2., , W R:s«.tfe(fmptHamm«VandM7u)ger, 7. An,., anH' "and, , Om. So,, 55, 4079, , 60, 679 (1938)., , (1935)., (, , (1933); «d *-*-
Page 391 :
375, , Base- and Acid-catalyzed Halogenations of Ketones, , overall exchange reaction cannot proceed any more rapidly than any individual, step we may conclude that the formation of the complex, if it is indeed a step, in the exchange reaction, is likewise many times as fast as the enolization., This ketone-acid association (which, if the acid is sufficiently, amounts to the formation of the conjugate acid of the ketone, H, , strong,, , I, , +, , \, , C, , C, , OHJ, , draws electron density away from the a-C—H bond toward the carbonyl group,, facilitating the removal of the a-hydrogen by a basic species. The role of the, base (designated noncommittally as B:) in sequence (2) may be played by a, molecule of solvent, or, if the solvent is hydroxylic, by a hydrogen-bonded, aggregate of solvent molecules. If other basic species are present in appreciable, quantity, these may also attack the hydrogen-bonded acid-ketone complex, formed in the initial step and the rate law for enolization will contain one or, more terms of the type £cat.(ketone) (HA)(£). In a solution such as an aqueous, acetate-acetic acid buffer, the rate law becomes even more complex, since there, are three acids (H30+, HOAc, and H?0) and three bases (OH-, OAc-, and, HoO) which may participate. The rate law for the iodination (hence the, enolization) of acetone, applicable over a large range of buffer compositions, is*0, rate = (acetone) [ki + £2(H30+) + *3(HOAc) + £4(OH-), + £5(0 Ac-) + *6(HOAc)(OAc-)], , (3), , (For a given solution, only two or three of these terms may be significant.), Each of the terms may be considered to represent the action of a different, acid, or base, or both; for example, the term containing k2 represents the action, of the acid H3O, , and the base HoO (which does not appear in the kinetic, , expression), whereas the term containing k6 represents the action of the acid, HOAc and the base OAc-. Terms containing (H30+)(0H-), (H30+)(0Ac-),, and (HOAc) (OH-) seem, at first glance, to be missing, but they are hidden in, the terms containing kh kz, and kb, respectively. Each of these three terms,, because of the mobile equilibria involved, represents the combined action of, two acid-base pairs." for example, the term containing k3 represents the action, of both the HOAc—H20 and the H30+—OAc- couples., At this point it may occur to the reader that the mechanism just considered, or acid-catalyzed enolization (which we may refer to as a concerted mechanism), qUeStT °f enolization Promoted by base; for sequence (2) obsh becomes a base-induced enolization if HA is the solvent and B: is its, , 21Tk, , jugate base. In contrast to the mechanism proposed on page 373 (which, , :T: 7r, , - r)a"son, Da i and Spivey, J. Chem. Soc.,In1930,, the Carbani0n, „, 2180, swain, J. Am. Chem. Soc., 72, 4578 (1950)., , meChanism’attach
Page 392 :
376, , -, , Carbanions and Enolization, , without help from an acidic species, whereas in the concerted mechanism, as its, name implies, simultaneous action by both acid and base are necessary (al¬, though the acid arrives at the reaction site first). Kinetics offers no choice if the, “acid,” which comprises the chief difference between the two possible transition, states, is one or more solvent molecules. The question is certainly not settled,, but as we shall presently see, the relations between structure and reactivity in, baseipromoted enolization are significantly different from those in acidcatalyzed enolization, suggesting that, at least in some cases, different mecha¬, nisms operate., , Keto-enol Equilibria, While considering the details of the enolization process under various conditions, we must not lose sight of one fact; the conversion of ketone to enol or to enolate, ion, whatever the mechanism may be, is much slower than the halogenation of, the enol. This means that if we quickly titrate an equilibrium mixture of ketone, and enol with Br2, we should observe an end point when all of the enol, but, practically none of the ketone, is brominated. (The bromine color at the end, , Table 10-2. Enol Content in Some Ketones, Diketones,, and Keto Esters", Compound, , ch3—c—ch3, , Compound, , 0.00025, , Ph—C—CH2—C—CH3, , I!, , O, , o, cyclohexanone, CH 3—C—CH 2—C—CH 3, O, O, CII3—C—CH —C—CH3, , I, , O, , Percent, , Percent, Enol, , I, , Me, , 0.02, 80, , Enol, 99, , O, , 1,3-cyclohexanedione, CH 3—C—C—CH 3, , i n, , 95, 0.0056, , o o, 33, , 1.2- cyclopentanedione, , 99, , 1.2- cyclohexanedione, , 40, , II, , O, , CH3—C—CH2—COOEt, , 7.5, , I, , o, CH3—C—CH(i-Pr)—COOEt, , CH2(COOEt)2, , 0.1, , II, O, .. For more extensive data on keto-enol equilibria, including .he «*■»*>«i of J®-, , Inc.. New York, 1956, pp. 446-453.
Page 393 :
Keto-enol Equilibria, , -, , 377, , point should, however, slowly fade as -the remaining ketone is converted into, enol and the latter brominated.) It is thus possible to estimate, by selective, bromination, the quantity of enol in a keto-enol mixture, although the deter¬, mination as described is rather unsatisfactory in practice unless it is modified, in one of several ways, which do not, however, alter its essential principle/5, The ketones, diketones, and keto esters listed in Table 10-2 exist as mixtures of, tautomers in which the enol content may be as low as a fraction of a percent, or, in other cases, may approach immeasurably close to 100 percent. The enol, contents given refer to equilibrium mixtures at 25° in the absence of solvent., The first four entries in this table emphasize the striking difference between, the extent of enolization in mono- and diketones. Incorporation of a second, carbonyl group, in the beta position to the first, increases the length of the conju¬, gated system in the enol form from three to five atoms,*4, , 1, 1, +, —C=C-OH ■<—> —C—C=OH, Li, 1, J, , O H-O, vs., , II, , 1, , “O H-O, -+, , 1, , II, , but this effect almost certainly accounts only for a portion of the observed dif¬, ference. Intramolecular hydrogen bonding (XVII) occurs in the enol forms of, O-H-O, , ,C, R, , C", , R, , R', XVII, most acyclic /S-diketones and fJ-kcto esters, but obviously cannot occur in the, keto forms. Since the energies of most O—H • • • O hydrogen bonds lie near, , ssrS-x, 11, , CBr-C-O + H+ + 21-, , 2, , ,:™, , II, , GH—C=0 + Br~ + I2, , ““ be'elegant procedure, which., llt^er>, , Helv., , Chim. Acta, 30, Tsy bbTfm?)!1 qUantitiCS °f Cn01’ SCe Schwarzenbach and', , results in an extension of th^con/ugMed/stem"^1 fh°UP '? acetylacetone or acetoacetic ester, o, an effect which should (and does) increase the degree oTe^oUzati^' 'h', , rinE
Page 394 :
378, , -, , Carbanions and Enolization, , 6 kcal per mole, we may expect hydrogen bonding, when present, to stabilize, the enol form with respect to the keto by about this figure. In the absence of, other effects, this would favor the enol in the equilibrium by about 4 powers of, 10. Such hydrogen bonding loosens the O—H bond and allows somewhat more, electron density to shift from the hydroxyl oxygen of the enol into the remainder, of the conjugated system. As a result, the polarity of the enol is lowered; and,, indeed, in practically all cases where a diketone or a keto ester has been sepa¬, rated into tautomeric forms, the enol has been found to be the more volatile,, hence, presumably, the less polar form (although hydroxyl compounds as a, class are commonly considered to be more polar than carbonyl compounds)., The atoms forming an effectively operating conjugated system should, as, we have repeatedly seen, lie in or near a common plane. Therefore any struc-, , I, , tural feature that hinders the coplanarity of the —C—C—C— group should, O, , OH, , decrease the stability of the enol form and lower the degree of enolization. A, scale model of XVII shows that group R and group R' will interfere with each, other when either is as large as or larger than a methyl group unless the, model is twisted so that the planarity of the conjugated system is destroyed., We thus see why incorporation of a methyl group in the 3-position ofacetylacetone (XVII, R' = CH3) lowers the degree of enolization, in this case by, over 50 percent. The effect appears to be less for acetoacetic ester, for in the, preferred conformation of the terminal, , COOEt substituent (p. 74), the, , O-H—-O, , O, , R, R', XVIII, , bulky ethyl group is far removed from substituent R' (XVIII). Nevertheless,, substituents R and R' may still interfere, and the isopropyl derivative of aceto¬, acetic ester is appreciably less enolized than acetoacetic ester itself., « In contrast to the effect of alkyl substitution on the methylene group of acetyjacetone,, substitutions on the metkyl groups of this diketone increase the c^ecofeaata.^ H mrnond., Ref 23 D 447) It has been suggested that the preferred conformation ot the keto, *leLPcsi x,X, in which .he cacbony, oxygen »conformation becomes much less favorable, however, when the R groups, R, , R, , O, , O, , b, , .
Page 395 :
Keto-enol Equilibria, , -, , 379, , On the other hand, when the ends of the enol form of a diketone are tied, back into a ring, the intramolecular rotations, , which, in an acyclic structure, , result in large departures from planarity—are prohibited. This being the case,, the enol becomes stabilized and the degree of cnolization increased. Thus, the, 1,3-cyclohexanedione derivative, dimedon (XXI), exists almost completely as, the enol form XXII, in which all carbon atoms in the ring except C5 are coplanar. (This predominance of enol is all the more striking since the geometry, of the ring system prohibits intramolecular hydrogen bonding.), , O, , O, , XXI, , XXII, , Cyclization has a most remarkable influence on the degree of enolization, of a-diketones. Diacetyl, CH3—C—C—CH3, exists almost exclusively as the, , O, , O, , keto form whereas 1,2-cyclopentanedione is almost 100 percent enolized. This, is almost certainly a conformational effect. In the most stable conformation, of biacetyl (XXIII), the C=0 dipoles are pointing in opposite directions with, the negative oxygen atoms as far from each other as possible. Because of the, rigidity of the cyclopentane ring, however, the two carbonyl dipoles in 1,2-, , XXIII, , XXIV, , cyclopentanedione lie at an angle of only about 65°,, the enol form (XXV), presumably in an attempt to, , XXV
Page 396 :
380, , Carbanions and Enolization, Since, as we have remarked, the keto form of a diketone or keto ester is, , almost invariably more polar than the enol form, we need not be surprised that, the enol: keto ratio for a given pair of tautomers at equilibrium in solution, depends markedly on the polarity of the solvent, and that this ratio tends to, be greatest in the least polar solvents. Acetoacetic ester is typical of many, tautomeric materials for which such a dependence has been observed; the, following figures represent the enol contents at equilibrium for dilute solutions, of this keto ester in various solvents (at 18°).*6, Solvent, Percent enol, , H20, 0.4, , HOAc, 5.7, , EtOH, 10.5, , Benzene, 16.2, , Hexane, 46.4, , Pure ester, 7.7, , Structure and Rate in Enolization, The rate of enolization of a ketone, or, more precisely, the rate of its conversion, to the enol-enolate system may be obtained by measuring its rate of halogenation, or rate of deuterium exchange, or if the carbon bearing the acid hydrogen is, asymmetric, its rate of racemization. The rates of numerous enolizations (and, related prototropic shifts such as nitro to aa-nitro conversions) have been, determined under a variety of conditions and in a number of solvents. Typical, values (for reactions in water at 25°) are recorded in Table 10-3.*7 These, reactions were carried out in the absence of added acid or base, being therefore, first order. It is evident that the stronger acids tend to enolize more rapidly than, the weaker. This is not surprising, for any structural feature that serves to, stabilize the enolate anion (extended conjugation, coplanarity, or the presence, of electron-attracting groups) would be expected to stabilize a transition state, in which the a-hydrogen is in the process of being removed. Yet there are, enough discrepancies to remind us that the correlation between the two types, of phenomenon is only a rough one. Acetoacetic ester enolizes about 2000 times, as rapidly as nitromethane although the compounds have nearly equal acidities;, an a-bromo group boosts the rates of enolization of acetylacetone and malonic ester but decreases that of benzoylacetone; benzoyltrifluoroacetone,, ph_q_ch2—C—CF3, enolizes only one half as rapidly as acetylacetone, , O, , O, , u Meyer Ann., 380, 212 (1911). Since the polarity of a solution of a lcetone or keto esttf, , sbsnd nitro., 67, 2027 (1945); and Reid and Calvin, ibid., 72, 2948 (19b ).
Page 397 :
381, , Structure and Rate in Enolization, , Table 10-3. Rate Data for Ionization of Acids Involving C—H, Bond Breakage, , 50, , ch2(no2)2, , ch3—c—ch2—c—ch3, , II, , o, , Acid, , ki, , CHs—C—CH—COOEt, , 4.5 X 10-4, , ki (min x), , Acid, , II, , 1, , O, , Et, , 1.0, , ch3no2, , 2.6 X 10~6, , 0.36, , ch3—c—ch2ci, , 3.3 X 10-6, , II, , o, , CH3—C—CHBr—COOEt, , II, o, , o, , CH3—C—CH2—COOEt, , 0.072, , I, o, , ch3—c—ch3, , 2.8 X 10“8, , o, , although its acidity constant is about 70 times as great. Moreover, in the nitroalkane series the weaker acids appear to enolize the more rapidly (although differences, within the series are small.*5 Apparently those substituents that show strong, electron-withdrawing power by induction are, in general, relatively more, effective in stabilizing the anion than the activated complex leading to it., Thus nitro compounds and trifluoromethyl ketones, almost without exception,, undergo tautomerization more slowly than nonhalogenated ketones, diketones,, or /3-keto esters of comparable acidity., The rates of conversion of haloforms to their carbanions constitute a puzHCX3 + B:^±B: H+ + CXj, zling series, for deuterium-exchange experiments*® indicate that the heavier, haloforms (which one would expect to be less acidic) are the more reactive., Ihe following relative rate constants (water, 0°) for base-catalyzed hydrogen, transfer illustrate the trend:, Haloform, k, ^DCCl,, , DCC12F, , DCC13, , DCBr2Cl, , DCBr3, , DCI3, , 0.019, , 1.0, , 31.0, , 121, , 130, , boxyTationrlf thinS -°, “ CqUaUy dis,urbinS ,rend in the rates of decar., of the anions of tnhaloacetic acids (CX3COO-) where, once again, «.,}“a3?74 095o(' Am■ ®"■ S0C•' 6°’ 2558 <1938>1 «. ™ 0943). Peai*o„ and Dillon!, , 7helan8^/^ ^^™®e"'',^*^**’1^^d*''V'1haloforni^fDC^C 'ami ‘4d6, 'n ’Ws CaSe’ ,h', The conclusions resulting from these experiment.:, , d ordinary water was studied., , ordinary, , since the hydrogen isotope effect, although sizable oroh hi° 3PP^ t0, haloforms,, throughout the series., 8, b ’ Probably remains very nearly constant
Page 398 :
382, , Carbanions and Enolization, , the derivatives of iodine react most rapidly, and those of fluorine least rapidly.50, The activated complexes in the two types of reaction have one obvious feature, in common: the CX3 group has assumed considerable carbanion character, (p. 346) in both. It thus appears that a CTj anion (and an activated complex, leading to it) is most stable if X is iodine and least stable if X is fluorine, but, explanations to account for this trend have not been convincing. One suspects, that because the halogens (except fluorine) may “expand their valence shells,”, the C—X bonds in each of these anions (except CFj) have some double-bond, , 1 +, character (XXVI «-> XXVI'), being in this respect similar to the, , :C—S—, , bond in the sulfonium derivative XII (p. 372). Nevertheless, why this double-, , X, X—C—X, , X, .., I, :X=C—X, • •, , XXVI', , XXVI, , bond character should be more important for the CI3 ion than for the CC13, ion (if indeed it is) is a baffling point., Returning now to more familiar ground, let us reconsider the basecatalyzed bromination of ketones. In the absence of, , complicating effects, , we should expect the rate of such brominations to fall when an electronrepelling alkyl group is substituted for an a-hydrogen. This has been found to, be the case, for example, in the acetate-catalyzed bromination of acetophenone;, here, the substitution of a single methyl for a hydrogen in the acetyl group of, the ketone results in a, , -fold decrease in specific rate (at 75°), whereas, , 6, , substitu¬, , tion of two such methyls results in a 33-fold decrease.5' Furthermore, in the, , halogenation of an unsymmetric ketone (for example, RCH,-C, , CHR,) the, , alpha carbon bearing the fewer alkyl groups (in this ease, the one on the left), should be preferentially attacked." Conversely, substitution of an electronattracting halogen atom for an a-hydrogen in a ketone should accelerate,, further halogenation. On this basis, we may understand why the major pro, from the base-promoted bromination of acetone (with^the, , ctone in, , excess) is the unsymmetrical tribromo compound, CH,-C-CBr3. The monobromo compound first formed is more, , readily, , brominated thar. acetone, , and the dibromo compound is still more read.ly brommated. Thus,, • See, for example. Brown, !\134 <>«'>•, 31 Evans and Gordon, J. Chem. Soc., 1938,, ** Cardwell, J. Chem. >Soc., 1951, 2442.
Page 399 :
Structure and Rate in Enolization, , 383, , tion of the tribromo compound takes place in a series of steps, the slowest of, which is the enolization of acetone itself., The picture changes markedly when we turn to acid-catalyzed halogena_reactions whose rates are determined by acid-catalyzed enolizations., , tjons, , Here, the substitution of alkyl groups for a-hydrogens accelerates halogenation,, whereas a-halo substituents retard it. As a consequence, we often find that an, unsymmetric ketone suffers halogenation (or any other enolization-controlled, reaction) mainly at one position in strong acid and at another position in base., In the following four ketones, for example, the letters a and b designate the, favored positions of attack under acidic and basic conditions:, , ch3-ch2— c-ch3, /1, , II, o, , ♦, , a, , Similarly, when acetone is brominated in the presence of strong acid rather, than in base, the monobromo derivative becomes isolable; for under these, conditions, it enolizes more slowly than acetone itself., The change in structure-reactivity relationships as an enolization reaction, is transferred from basic to acidic solutions suggests (although it does not de¬, mand) a change in mechanism. To account for the accelerating action of a-alkyl, groups on acid-catalyzed enolization, we may remind ourselves that an alkyl, group adjacent to a C=C double bond stabilizes the latter by about, , 2.5, , kcal per mole, an effect which is generally attributed to hyperconjugation,, f RCHo—C=C—, , H+CHR=C—C—\ Since a C=C double bond is being, , formed in an acid-catalyzed enolization, it is reasonable that the process should, be aided by a-alkyl groups. This explanation, however, obviously has a hollow, ring unless we can explain why the inductive action of a-alkyl groups which, presumably is important in base-induced enolization, suddenly becomes unthat, he breakage, bond>CatalySt', wh.ch requ.res, a goQdwe, port.on, ^ the, ZTh, mJhe Pof, 7the, T“a.c_H, °7n addk, In tWs reSard>, maV assume, p“ fner?Hy rry f°r a base-induced en°lization, is facilitated in the, P, of acid; for here the -C=0 group has been converted either to a, OH ‘“age or a -C=0 • • ■ Hd linkage, and electron density has, , breankaPgUeleof,rC-H0T3 ‘d ^ ^ Weakeni"g, In add> the"., of Charge around «h“ c ,, ", tha" the -organization, ge around the o-carbon atom in forming the double bondj and fac(ors
Page 400 :
384, , Carbanions and Enolization, , that control the ease of breakage of this bond likewise become less important.55, Although inductive effects overshadow hyperconjugation in base-induced, enolizations, it is still possible to show that the latter exist. In hyperconjugation,, as we ordinarily consider it, a double bond is stabilized by the “loosening” of, a G—H bond adjacent to it; and if such a C—H bond is replaced by a somewhat, “tighter” C—D bond, hyperconjugative effects diminish slightly (see, for ex¬, ample, p. 285). Thus, replacement of the four /3-hydrogen atoms in phenyl, cyclopentyl ketone, yielding the deuterated ketone, XXVII, lowers the rate for, acid-catalyzed enolization (HC1 in HOAc) by 20 percent, and since the same, , O, XXVII, isotope effect arises in the acetate-catalyzed enolization of this ketone, we may, conclude that hyperconjugation also operates here.'5'* It has been suggested5’*, that we have here evidence that a concerted, rather than a carbanion, mecha¬, nism (p. 374) for base-catalyzed enolization is operating. For it may be argued, that hyperconjugation should not significantly stabilize an anion such as, XXVIII or a transition state leading to it, since the high concentration of, negative charge on the carbonyl group in hyperconjugated structure XXVIII", would be expected to make this a very unstable structure. Although this argu¬, ment is interesting, it is indicative, rather than conclusive., , ?, -<—=—v, , C-Ph, , I, , o_, xxvnr, The Concerted and the Carbanion Mechanisms for Tautomerism, The mechanistic ambiguity which we have noted for base-promoted enoliza¬, tion does not apply to all base-accelerated tautomerizations. A rather clear, » The retarding action of c-halo substituents in acid-catalyzed cnohaanon. may be m, directly explained. Since such substituents are electron attracting, they reduce the bMicity, me ketone thus lowering the equilibrium concentration of its conjugate ac.d-.he, intermediate in the enolization., - fl956), n Emmons and Hawthorne, J. Am. Chem. Soc., 78, 5593 (1930;.
Page 401 :
The Concerted and the Carbanion Mechanisms for Tautomensm, , 385, , decision is possible, for example, in the conversion of the 0,7-unsaturated nitrile,, XXIX, to its a,(5-unsaturated isomer, XXX. If this reaction is initiated in, EtOD in the presence of OEt“, but the unconverted starting material is reiso¬, lated while the reaction is still in its early stages, considerable deuterium,, , XXXI, , XXIX, , may be found in nitrile XXIX; this indicates not only that the isomerization, proceeds through anion XXXI, but also that this anion, once formed, is more, likely to revert back to nitrile XXIX than proceed to nitrile XXX.55, An equally clear decision is possible (although the answer in this case is, different) for the tautomerization of such imines as XXXII to the isomeric, imines—for example, XXXIII. Here, if the reaction is carried out in EtOD,, Ph, , Ph, , \, , CH—N=C, , /, , /, , Ph, OEt", , -^, , \, , Me, , Ph, , \, , /, , C=N-CH, , /, , Ph, , \, , Me, , XXXII, , Ph, , XXXIII, , the rate of isomerization is found to be equal to the rate at which deuterium, enters the imine (at least in the early stages of the reaction), and if the reaction, #, , «|c, , is carried out on an optically active imine, Ph—CHMe—N=CPh2, the rate of, isomerization is also equal to the rate of racemization “ The observed corre¬, spondence in rates may be interpreted in one of two ways. The isomerization, may proceed through a carbanion intermediate, Ph—CMe—N=CPh2, which, is converted almost completely to imine XXXIII, in which case the isomeriza¬, tion would be very nearly irreversible; or alternately, this reaction may proceed, by a concerted attack of the acid EtOH (or EtOD) and the base OEr on the, imine (note that, according to this mechanism, each removal of a hydrogen, EtO', , H-OEt, , EtO-H, , sT, , Ph- C—N—C—Ph, Ph, , H, , OEt", , Ph-C=N— G—Ph, , I, , Ph, , Ph, , I, Ph, , „, , from the “left” side of the moiecule is accompanied by an addition of a hydrogen, , 319., , c->, , 4Z6., , (b), , ds|, , de Salas and Wilson,, , ibid.
Page 402 :
386, , -, , to the, , Carbanions and Enolization, light, , side). The first possibility, the formation of a carbanion and its, , irreversible destruction, is eliminated by our knowledge that the isomerization is, reversible (Keq 0.47 at 85°) ;S7 hence we may conclude that the concerted mecha¬, nism operates for this reaction. If the reaction were irreversible, neither alter¬, native could, on the basis of the evidence quoted, be excluded. Indeed the, base-catalyzed bromination of ketones is, as we have emphasized, mechanis¬, tically ambiguous, largely because of its approach to irreversibility., By using similar methods of investigation, it should be possible to obtain, information concerning the paths of a number of additional tautomerizations—, for example:, Ph., , Ph, , O, , Ph\_ .OH, , .nXH, O, Ph, , MeO, , EtCH2—CH=CH—COOEt, , MeO, , v BEt-CH=CH-CH2-COOEt, , Yet, although the literature of tautomerism is extensive, there remain many, tautomeric conversions for which we cannot say which, if either, of these, mechanisms applies. One rule of thumb is, however, of some use: For relatively, acidic substances in strongly ionizing solvents, tautomerization is likely to in¬, volve initial dissociation of a proton, whereas if the acidity of the substance,, or the ionizing power of the solvent, are lowered, the concerted mechanism, becomes more likely., , Carbanion Character in the Phenoxide and Pyrrolyl Anions, The conjugate base of a ketone, a keto ester, or a nitroparaffin is, as has been, emphasized, generally formed by ionization of a C—H bond. Since, howeser,, the negative charge is not confined to a carbon atom or group of carbon atoms,, but is instead spread over a conjugated system including an oxygen atom as, well, it may be argued that such anions are not “true carbanions'’ in the sam, sense as is, say, the triphenylmethide ion, PhiCt- The definition of a jue, carbanion,” is, like many definitions, somewhat arbitrary; but, regardle, definition; such conjugate bases obviously have much of the character of carb* Hsu, Ingold, and Wilson, J. Own. Soc., 1933, 1493; 1934, 93; 1935, 1774.
Page 403 :
Carbanion Character in the Phenoxide and Pyrrolyl Anions, , 387, , anions; that is, they are anions in which a significant concentration of negative, charge lies on one or more carbon atoms. On this basis, there are other anions, that, although not generally formed by ionization of a C, , H bond, also have, , unmistakable carbanion character. The most important of these are the con¬, jugate bases of phenol and pyrrole; ambident reagents (p. 296) in which the, carbon atoms, as well as the more electronegative oxygen or nitrogen atoms, are, nucleophilic. Indeed these anions are often represented as resonance hybrids, in which one or more of the individual structures have negatively charged, carbons., , The phenoxide ion displays carbanion character, for example, in its reactions, with bromine, with COo (the Kolbe reaction), and with chloroform, , (the, , Reimer-Tiemann reaction); for in each of these reactions, the attacking atom, is carbon rather than oxygen., , Similarly, in the carboxylation and ethylation of the pyrrolyl anion (XXXIV), the attacking atom in the ring is carbon, rather than nitrogen., , H, , - H+ shift, , fC-cr, 1, , ^COOH, , H, .Et, xEt, , XXXIV, , ■>T, , —H, , shlft -, , n, 'N', , I, H
Page 404 :
388, , -, , Carbanions and Enolization, , Geometry of Carbanions, We have noted on several occasions that the bonds about the tervalent carbon, ot a free carbonium ion, —C+, prefer to lie in a common plane and that, as a, consequence, the conversion of an asymmetric carbon atom to a positive carbon, during the progress of a reaction results in a racemic product (unless other, centers of asymmetry are present in the reactant). It then may be asked whether, the same situation applies to carbanion intermediates, but here a satisfactory, answer is more difficult to obtain. The suggested correspondence between basecatalyzed racemization of a ketone and carbanion formation (p. 373), which,, at first thought, would seem to imply that carbanions (like carbonium ions), may not be asymmetric, is not of help here. Aside from the ambiguity in mecha¬, nism (which might mean that free carbanions do not intervene at all here),, it must be remembered that the anion, if formed, would be in mobile equilib¬, rium with the enol, in which the asymmetry of the a-carbon is surely destroyed., , CH-C-, , *, , II, , O, , B:, slow, , —c-c•• II, o, , I, , BH+,fast, , — C = C—, , (symmetric), , .BH: .fast, , OH, , In all probability however, the carbanion, if formed, is symmetric, not neces¬, sarily because it is a carbanion, but rather because it is part of a conjugated, system that should be planar. The same is true for such carbanions as the, conjugate bases of triphenylmethane and indene; for in these also, the carbon, atom from which the hydrogen is removed becomes part of a conjugated system, (p. $70). If we then narrow our inquiry to the geometry of carbanions that are, , not stabilized by conjugation, we may be left with no problem at all, for it is, doubtful that such carbanions exist. We would be most likely to find them in the, alkyl derivatives of the very basic metals—for example, alkylsodium and alkylpotassium compounds; but such compounds, unless the alkide portion is sta¬, bilized by conjugation, appear to be insoluble in all solvents except those with, which they react.** Thus, if carbanions are present in such compounds, they are, incorporated into a crystal network, the detailed structure of which is unknown., .. Alkylsodium compounds dissolve in diethylzinc (and presumably in other alkyffinc, derivatives) to give conducting solutions; but here again, it is extremely likely that a chenuca, reaction occurs—for example,, C2H6Na + (C2H6)2Zn-> Na+ + (C2Hs)»Zn“, For a brief discussion of .his problem see Sidgvriek, CM.,, Oxford University Press, Oxford, 1950, pp. 277 279., , Bemtn,S, , W M, , <**-*<
Page 405 :
The Aldol Condensation and Related Reactions, , -, , 389, , Stereochemical studies with such organometallic derivatives will tell us little, about the geometry of the “free” carbanion, since there is no reason to suppose, that the carbanions present (if indeed they are present) in such networks have, the same configurations as the corresponding solvated carbanions., , At any, , rate, there is little point in speculating about the geometry of a carbanion that, is not stabilized by conjugation until such a carbanion can be demonstrated, to exist., , The Aldol Condensation and Related Reactions, The most important group of reactions proceeding through carbanion inter¬, mediates are base-catalyzed aldol condensations, including such variations as, the Knoevenagel and Perkin reactions. Our present picture of the aldol con¬, densation is rather similar to that of the Claisen condensation (p. 334); for in, both, the characteristic step involves the attack of a carbanion on a C=0, group., , OH, , II, , OH, , I, , OH, , I, , I, , II, , B:, , —C—C—H, , I, , " —C—C:, , O, , O, , H, , II, , II, , I, , —C—C:~ + C—R ?± —C—C, R, , (XXXV) (carbanion formation), , O-, , O, , I, C, , H, , OH, , £H +, , R, , R, , XXXV, O, , H, , — —C—C—C—R (addition and, proton transfer), R, XXXVI, , o, , OH, -h2o, , ^, , ^, , |, , R, , ^, , ^, , ^^^ (dehydration; does not always occur), , R, , XXXVI, As with the Claisen condensation, the overall reaction is reversible, but in, practice, it is often pushed to completion by destruction (dehydration) of the, condensation product, XXXVI. A number of bases (designated above as B:),, , 2-iodoo'^ne'^S?"72’4842 0 950)) that °Ptically active, responding carboxylic acid ar low, correspondmg lithium compound and thence to the corthe reactions of alkyllithium compoundseratjores, , partial retention of optical activity. If, , mediates, the conversion described would" MicatTaJT^^1^ !hrouSh,yrbani°n tnteractmty and is therefore nonplanar However alkvllhh, a Carban,on could retain optical, rewalent compounds, and there is no strong ’evidence, Ad?k“T '° ^ typkal, The conversion of optically active chlorides*^ nce that their reactions involve carbanions., to Grignard reagents, thencecarbon atoms, compiete racemiration (Goering and JcCarrot’,, , 80" £T<55?)^
Page 406 :
390, , -, , Carbanions and Enolization, , have been found useful, among them hydroxide and alkoxide ions, amines, and, carboxylate ions. Let us consider the reaction sequence as represented above, (aside from the final step, which is not actually part of the condensation itself)., There are six individual steps (three forward and three reverse), and since the, apparent rate law will depend upon the relative magnitudes of the various rate, constants, we might expect a number of rate laws for specific condensations., However, there are simplifying assumptions that can be made. The third step,, in which a proton is transferred to an oxygen atom from another oxygen atom,, may be assumed to be much faster than either of the steps preceding it. Further¬, more, the rate constants for the first two steps and the respective back reactions, must have values that prohibit the collection of significant quantities of anion, XXXV in solution; for the concentration of this carbanion is generally observed, to remain small. There are two kinetically simple, limiting cases. In the first,, carbanion formation is very slow, hence rate determining; in the second, car¬, banion formation is rapid but its reversal is also rapid. If the latter is so, we have, the familiar situation in which a slow step is preceded by a rapid equilibrium., The rate law becomes, , (_?_, , _), , ., , rate = &(carbanion)(R2C—O) = kKe<l-——*~^2., , ——, , (4, , (If B: is the conjugate base of the solvent, (RH+) in the denominator vanishes.), A typical condensation in this category is the hydroxide-catalyzed self-condensa¬, tion of acetaldehyde when the concentration of the aldehyde exceeds 1 molar., CHO H, OH”, slow, , OH OHO, , 1, , !, , CHO, h.O, , I, , CH3CHO 7-*, ~:CH2CHO —-—>, ch2—c—o—->, ch2—ch—oh, -fast, I, fast, °, , *, , ch3, , ch3, (aldol), , At high concentrations of acetaldehyde, the condensation is first order in alde¬, hyde, whereas if the initial carbanion formation were rapid and reversible, the, reaction should, as indicated in equation (4), be second order in aldehyde.^, The reaction is also, at least to a good approximation, first order in hydroxide.-*0(o) Furthermore, if the formation of the carbanion were significantly, reversible, the condensation, when carried out in D20, should yield an aldol, with C—D bonds, and, if the reaction were halted in the early stages, the un¬, reacted aldehyde should likewise have C—D bonds. This test has also been, <0 For kinetic studies of the aldol condensation of acetaldehyde see (a) Bell J Oum. Soc^, 1937, 1637; (b) Bell and Smith, ibid., 1958, 1691; (e) Broche and Gilbert, Bull. Soc. chi ., , France, 1955, 131.
Page 407 :
The Aldol Condensation and Related Reactions, , -, , 391, , applied and the reaction has been found to be accompanied by practically, no C—D bond formation, either in the product or in the unreacted aldehyde,, thus confirming the conclusions drawn from the observed rate law?*, On the other hand, the formation of the carbanion may be rapid and re¬, versible, but its subsequent attack on the carbonyl group relatively slow. This, is the case, for example, for the ethoxide-catalyzed condensation of benzaldehyde with acetophenone?* and for the hydroxide-catalyzed condensation of, H, OEt-, fast, , :CH2—C—Ph PhCH°> Ph—C—CH2—C—Ph ^, , CHs—C—Ph ^, , II, , EtOH, fast, , slow, , o_, , o, , o, , I, , ||, , fast, , o, H, , O, , Ph—C—CH2—C—Ph, OH, malonic ester with formaldehyde?* both of these reactions exhibiting rate laws, OH ~, , H2C=0 + CH,(COOEt)2-» H2C—CH(COOEt) 2, OH, corresponding to equation (4). Much the same picture is appropriate for the, base-catalyzed self-condensation of acetone, but here the evidence is of a less, direct sort (Ex. 3). It is also possible that the concentration of the carbanion, may be kept small, both by reversal of its formation and addition to the 0=0, group, with the two modes of destruction occurring at comparable rates. The, rate law then becomes somewhat more complex but may readily be derived, using the steady-state approximation (Ex. 4). This is the situation, for example,, with the aldol concentration of acetaldehyde at low aldehyde concentrations?0^’, Asthe concentration of aldehyde decreases, the attack by the carbanion on the, O group becomes less likely and its destruction by solvent assumes significance. At such concentrations,, , the reaction is accompanied by deuterium, , xcnange at the a-hydrogens if carried out in D2O?0(6>, In the fa™liar Perk‘" "““ion, a carbanion derived from an acid anhydride, a~, , °r a SUbStitUted, , Uehyde to yield a cinna^, , a salt of?, k, neCCSSary f°r rem°Val °f thC Pr°,on may, an a™ine, the lit of,?, ‘?°?niC, (PhosPhate. carbonate, etc), or, more usually, the salt of the carboxylic acid related to the anhydride., ‘;B„nh, , ffer, , d Wa|twSj z, , ?htm, , « w*h, ;nr*Evacns', ■, weich, J., Chem. Soc., J\, 1931, 653., , i9u’12M.
Page 409 :
The Aldol Condensation and Related Reactions, , -, , 393, , The product from the initial addition, the enolate ion, XXXVII, is rapidly, converted to the corresponding keto form, and the overall reaction is often, viewed merely as an addition of the ester to the C=C double bond. The con¬, versions of both malonic and cyanoacetic esters to their anions by base are known, to be rapid.3 It is very likely therefore that for Michael reactions involving these, esters, the rate-determining step is the second—the formation of the new C—C, bond (although this has not yet been demonstrated kinetically). The reaction is, reversible,45 and as was the case with the Claisen condensation (p. 334), the, maximum possible yield is governed by the basicity of the reaction mixture., In the Claisen condensation, it will be recalled, the product was a stronger acid, than the reactant and the reaction could therefore be pushed toward completion, by a large excess of base. With the Michael condensation, the product is a, weaker acid than the reactant (since a C—H bond has been converted to a, C—C bond); hence a large excess of base will favor the reactant. It follows then,, that although some base is needed to get the Michael reaction under way, a, high yield of product requires that the basicity of the reaction mixture be kept, relatively low., When both of the double-bonded carbons are bound to benzene rings (for, example, as in a-phenylcinnamic ester, XXXVIII), the Michael reaction fails,, presumably because the structure has been stabilized by incorporation of, the otherwise reactive double, , bond into an extended, , conjugated, , system, , (XXXVIII')., , A compound having two or more C=C double bonds in conjugation with, ar jony, , group (for example, the unsaturated ester, XXXIX) may in prin, , ■Pie, undergo the Michael reaction in several different ways. As shown, , the, , rbam n may attack at either the f> or a position; moreover, in the else of, bound «oWthP, are,P0SSible dePe"di"S uPon whether the proton becomes, bound to the a- or ,-carbon in the final product. As seen, ester XLI is the ory, iS Ingold and perrcn, J. Chem. Soc., 121, 1414 (1922).
Page 410 :
394, , -, , Carbanions and Enolization, CH(COOEt)2, , CH(COOEt)2, , BH+, , CH2=CHCHCH=COMe, , " CH2 = CHCHCH2COMe, , I, , o_, 5, , 7, , P, , a, , o, , o, , (|3 attack), , XL, , ||, , CH2=CHCH= CHCOMe, XXXIX, CH(COOEt)2, , I, , o, , II, , CH2CH2CH=CHCOMe, (XLI), , BHh, , and /or, , CH2CH=CHCH=COMe, , I, , CH(COOEt)2, , o_, , CH2CH=CHCH2COMe, , I, , II, , CH(COOEt)2, (5 attack), , O, (XLI I), , one of the three possible products that retains the —C=C—C=0 conjugated, system. Hence, it is not surprising that this ester is the predominant product.^*0, On the other hand, if a phenyl group is put into the 8 position (that is, if the, reaction is carried out using PhCH=CH—CH=CH—COOEt), the chief prod¬, uct is formed by/? attack; for only in this product would the, , \—CH=CH-, , conjugated system be retained.46(6) The y,8 double bond may also be stabilized, by hyperconjugation since reaction of MeCH=CMe—CH=CH—COOEt, with sodiomalonic ester proceeds by /?, rather than by 5 attack. In the preceding, discussion, we have assumed implicitly that of several possible products, the, most stable will predominate—that is, that the course of the reaction is thermo¬, dynamically controlled. This is by no means the case for all organic reactions; often, a given product will predominate because it is formed more rapidly than the, others—that is, the reaction may be kinetically controlled. However, in the case, of the Michael condensation, as it is usually carried out, thermodynamic con¬, trol of products is to be expected; for the steps are reversible and the reaction, mixture is generally allowed to come to equilibrium before being “worked up.”, , The Benzoin Condensation, The condensation of two benzaldehyde or substituted benzaldehyde molecules, has long been known to be catalyzed by cyanide ion, but not by any of the, more usual stronger bases., O, CN-, , II, , OH, I, , Ar—CHO + Ar'—CHO — Ar—C—CH—Ar, « (a) Farmer and Mehta, J. Ch,m. Soc., 1931, 1904. (*) Vorlander and Groebel, Am.,, 345, 206 (1906).
Page 411 :
The Benzoin Condensation, , 395, , (HCN, Hg(CN)2, and nitriles are ineffective as catalysts). Without knowing of, the specific cyanide catalysis, we would be likely to regard the benzoin condensa¬, tion as similar to the aldol condensation, with the anion derived from a benzaldehyde molecule (Ph—C=0) attacking the C=0 group of a second molecule., • •, , But we have no evidence that the hydrogen of a —CHO group is significantly, acidic; certainly it is not sufficiently acidic to be removed by such a weak base, as CN-. However, the cyanide ion can boost the acidity of this hydrogen, markedly by converting the aldehyde to the conjugate base of a cyanohydrin, (that is, an a-hydroxynitrile), the a-hydrogen then becoming similar in char¬, acter to the a-hydrogens of nitriles or esters. This conversion is known to occur, readily under the conditions necessary for the benzoin condensation. The rate, O(b), , Ph—C=0 + CN~ ^, , Ph-C-CN, , ^, , Ph-C-CN, , s, , @, , OH, , k, nonacidic, , acidic, XLIII, , XLIV, , law for the formation of benzoin itself (which is generally presumed to be typical, of reactions of this type) has been found to be.-*7, rate = £(PhCHO)2(CN-), telling us that the activated complex in the rate-determining step contains two, molecules of aldehyde and a cyanide ion. We then may conclude thatcarbanion, LI V, once formed, attacks the C=0 group of a second molecule of benzaldehyde, yielding anion XLV—a cyanohydrin which should be easily converted to, OH, , Ph, , OH, I, |, _ (c), Ph- -C:~ + C= o — Ph—c-, , I, , I, , CN, , H, , XLIV, , CN, , Ph, , -c—oH, , O- Ph, (d), , I, , I, , (e), , Ph—C—C—OH, , I, , I, , CN H, , XLV, , n, , O, , H, , II, , I, , Ph—C—C—OH + CNi, , ., , ., , Ph, fT tHe OVera,‘ C°ndenSa,i0n U —ib'^S each step must be
Page 412 :
396, , Carbanions and Enolization, , reversible. The rate law by itself does not tell us whether (c), (d), or (e) is the, rate-determining step, but we may assume that (d)—a transfer of a proton, between oxygen atoms—is not. Furthermore, (e) is simply the reverse step in a, ketone-cyanohydrin interconversion, a reaction in which, for simple cases,, equilibrium is known to be very rapidly established.49 Thus, if the mechanism, proposed above is correct, step (c) is almost certainly rate determining.50, The limitations of the benzoin condensation are of some interest.5* When, the strongly electron-donating —NMe2 group is substituted para to the —CHO, group in benzaldehyde, the reaction fails. Due to conjugation (XLVI <->, , XLVI'), the carbonyl group acquires additional electron density; that is, it, becomes less electrophilic. As a result, steps (a) and (c), both of which involve, addition of an anion to the carbonyl group, are hindered. Apparently, however,, the effect of the —NMe2 group on initial cyanohydrin formation is less im¬, portant than is its effect on the rate-determining condensation step (c); for if a, mixture of benzaldehyde and />-dimethylaminobenzaldehyde (XLVI) is treated, with NaCN, the “mixed benzoin,” XLVI I (but not the isomeric mixed ben¬, zoin), is formed. In the mechanism proposed, the carbonyl group on the final, product is that involved in the initial cyanohydrin-anion formation; we may, therefore conclude that the cyanohydrin anion formed from />-dimethylaminobenzaldehyde may add to benzaldehyde, but that the cyanohydrin anion, derived from benzaldehyde will not add to XLVI itself. The benzoin condensa¬, tion is also inhibited by strongly electron-attracting groups; />-bromobenzaldehyde, forms a benzoin slowly and incompletely, and ^-nitrobenzaldehyde does not, undergo the benzoin condensation. Both of these aldehydes readily form, cyanohydrins,49 but in the anions corresponding to XLIV, electron density, has been pulled away from the attacking carbon atom, making it a far less, effective attacking site. The/>-nitro group is particularly effective in this respect, for here the nitro group and the attacking carbon are in conjugation (XLV11I, , OH, , VQ-c7'An, XLVIII, , X, XLVIII, , * Sec, for example, Baker, et at., J Chern. Sac, :, , "'I, , 1942, 191; 19-, , bJc, in Or^R'^ns, Vo,. IV (edi.ed b,, , Adams), John Wiley and Sons, Inc., New York, 1948, p.
Page 413 :
Hydrolysis of Haloforms, , -, , 397, , XLVIII')- On the other hand, o-nitrobenzaldehyde undergoes the benzoin, condensation; presumably because the conjugation between the —NOz group, OH, group, which prohibits the condensation of the para isomer,, , and the —, CN, , is rendered ineffective in the anion of the ortho isomer (XLIX) because copla¬, narity between the two groups is no longer possible., It is now known that certain thiazolium and imidazolium salts (having, cations such as those shown below) also act as catalysts for the benzoin condensa¬, tion.^ Here, it appears that the condensation proceeds through an intermediate, +, , Et—N-, , Me-N-, , L, , Et-NR H, XC', , Me, , X., , I, OH, L, , adduct of the type L, in which the a-hydrogen, like that in a cyanohydrin, has, been rendered acidic by the strongly electron-withdrawing character of the, ring.5*(6), , Hydrolysis of Haloforms, The basic hydrolyses of chloroform, bromoform, and other trihalomethanes, are of considerable current interest because of the evidence that these reactions, proceed not only through carbanions but also through intermediates having, bwalwt carbon—that is, carbon atoms with only two covalent bonds. It should be, recalled that methylene chloride, CH2C12, is hydrolyzed by base much more, slowly than is methyl chloride, and we should therefore expect chloroform to be, hydrolyzed still more slowly. However, the hydrolysis of chloroform is a rela¬, tively fast reaction, leading us to suspect that the hydrolyses of the three halides, o not proceed by the same mechanism (although all three reactions are first, order in halide and first order in base).™® The abnormally high reactivity, of chloroform toward OH" does not extend to all other nucleophiles- it read, very slowly, for example, with the thiophenolate ion, PhS~, which is weakly, 3719"(m*)5*’ et **, , Pharm- SoC•, , 63> 269 0943); (A) Breslow, 7. Am. Chem. Soc., 80,, , c, ,, ^etrenk°-Kritchenko and Opotskv Ber 59 oi-ri MoOiO /t\ u., s°c-, 72,2438 (1950). (c) For studies of the hw u'V i 2 r O92^). (b) Hme, J. Am. Chem.
Page 414 :
398, , Carbanions and Enolization, , basic but generally considered strongly nucleophilic (p. 260). However, if, chloroform is treated with a solution containing both the OH~ and the PhS~ ions,, the reaction is rapid and the product is phenyl orthothioformate, HC(SPh)3.53w, The high reactivity of chloroform thus requires the presence of a strong base, and, in view of the base-catalyzed hydrogen exchange that is observed to occur, between haloforms and hydroxylic solvents (p. 381), it seems very likely that, the hydrolysis of chloroform proceeds through the C13C:~ carbanion, even, though only small quantities of this anion exist in solution at a given instant., Furthermore, the fact that the hydrogen exchange reaction is much faster than, the basic hydrolysis^ indicates that this carbanion is formed rapidly and re¬, versibly in the attack by OH- on the haloform.5^ Thus, most of the C13C:~, anions formed are reconverted to chloroform, but some react further, ultimately, yielding the observed hydrolysis products, carbon monoxide and formate., Although we know little about the final stages in the hydrolysis of the trichloromethide anion, it seems rather clear that this ion first loses Cl~ to form carbon, dichloride, CC12, an intermediate containing bivalent carbon., , -C^ cclj, , oh-, , hyd.oly.i-, co + HCOO_, , HCC13^=±C13C:, slow, , fast, , If the initial steps in the hydrolysis are those indicated, the reaction may be, said to proceed by an a-elimination mechanism—that is, a mechanism in w hich a, hydrogen atom and a chlorine atom are detached from the same carbon. In, the far more usual ^-elimination reactions, atoms are lost from adjacent carbons., A similar sequence may be written for the reactions of haloforms wuth alkoxides, in alcoholic media, for these alcoholyses are also unexpectedly fast.50 Since, species containing bivalent carbon are unusual (although not unknown), we, should be reluctant to suppose that carbon dihalides intervene in such reactions, if the evidence were not rather convincing. In the first place, it may be argue, that the hydrolysis of the C1,C;- ion is much more likely to occur by loss of a, Cl- ion than by direct displacement of Cl- by OH". With chloroform, such a, displacement is very slow, and it should be even slower for the conjugate base, of chloroform since there is electrostatic repulsion between the reacting ions., Secondly, the basic hydrolysis of chloroform in the presence of 'od.de yield, appreciable quantities of HCCW, a product that is not formed at a f'f"', rate by displacement of chloride by iodide (in neutral solution, , and, , is almost certainly not formed by such a displacement on CUC., ., , (Hi„:, verted to, , CF^, , ., , ., , ,, , ,, , ■, , <m, , r nrF Rr is however not accompanied by deuterium exchange, , by^on^erled process, , H, , ., , ., , <»*»• ><, , . Br)-, , CF), , “,s co", HOH + CF, + Bf, , analogous to the E2 process for (3 elimination (Chap. 12), “ Hughes and Preling, reported by Ingold, Ref. 1(b), p. W n., Chem. Soc., 80, 3002 (1958)., , her, , Tanabe. J. Am.
Page 415 :
Organometallic Compounds, , -, , 399, , solution). Again, it is far more likely that this product is formed by reaction of, iodide on neutral CChThe best evidence for carbon-dihalide intermediates arises, however, from, experiments in which such species are “trapped" before they can be further, hydrolyzed or solvolyzed; as may be supposed, such dihalides are extremely, reactive. Like the methylene “diradical,” CH2 (Chap. 16), they add to double, bonds, forming cyclopropane derivatives. Thus, when chloroform is treated, with base in the presence of cyclohexene, the bicyclic dichloride, LI, is formed.5e(o), Similarly, the action of £-BuO~ on bromoform in the presence of 2-butene yields, dibromide LII.56W>67, , HCCI3, , HCBr,, , H* >, , —hh, , - ^ >, , -Br', , LI, , CC12, , LII, , - CBr2, , It also seems likely that the condensation of chloroform with phenols in, basic media (the Reimer-Tiemann reaction, p. 387) proceeds through CC12,, since, once again, direct attack of phenoxide on the chloroform molecule with, displacement of chloride would be expected to be very slow., O", CHC12, HCCL-^U CCL, , PhO‘, , further ^, hydrolysis, , o, , A^cho, , Organometallic Compounds, We have already seen that solutions of the sodium and potassium derivatives of, such hydrocarbons as triphenylmethane, cyclopentadiene, and indene contain, carbanions, although in nonpolar solvents such carbanions exist almost exclustvely as ion pa.rs and higher aggregates. Carbanions probably exist also in, the solid state ,n the sodium and potassium derivatives of the lower hydrocar¬, bons. The reactions that these metallic derivatives undergo are similar in nature, «62 (,954). (*, Skell and Garner., the methyl groups in the product are trans where* e.re°specifi^ that is, if /r<zn.r-2-butene is used,, the product are aV., ’ WhereaS lf the, is u*ed, the methyl groups in, Paired electrons per molecule.'Fo^considerLi^n^f ^er carbon dihalides have zero or two unChem. Soc., 78, 5430 (1956). See also Chap. 13, qUCSt,°n’ see Ske11 and Garner, J. Am.
Page 416 :
400, , Carbanions and Enolization, , to those of the more usual organometallics, and it is perhaps not surprising, that the reactions of Grignard reagents and organolithium compounds are, often represented as involving anions also. Thus, the reactions of, say, methyllithium are sometimes depicted as reactions of the methide ion., , O', , /, -> h3c-c, , o, H3C:, , +, , CH,— CH., , HoC-CHo—CH,—O', , O, H3C:, , +, , Br-Hg-Br ->■ H3C—Hg—Br, , +, , Br', , It is possible that, in some cases, the reactions of organolithium compounds and, Grignard reagents do indeed proceed through preliminary ionization to a, carbanion (although this has not yet been demonstrated), but this is certainly, not true in all instances and is probably not true in most. Where a carbanion, R:~ is an intermediate, we should expect the reagents RNa, RLi, and RMgBr, to yield the same product (or the same mixture of products) on reaction with a, given electrophile. Of the many cases where such a correspondence is not ob¬, served, the following three are typical:, , (i-Pr)3C-OH, (*-Pr)2C=0, (t-Pr)2CHOH, , +, , CH3CH=CH2, , (CH2=CH-)2CH-COOH, , 68b, , CO., CH ,=CH- CH= CH- CH2-COOH, , Ph^C-NHPh, , 68c, , CHPh-NHPh, Ph2C=NPh, , It thus appears that the reactions of organolithium compounds and Grignard, « (a) Young and Roberts, J. Am. Chem. Soc^ 66, 1444\265 (1933)., Compt. rend., 224, 1118 (1947). (r) Gilman and K.rby, J. Am. Chem. Soc., 55,
Page 417 :
Organometallic Compounds, , -, , 401, , reagents involve the entire molecule (or more than one entire molecule) of the, organometallic., The majority of the reactions of organolithium compounds appear to be, too fast for convenient kinetic study, but the limited and not too satisfactory, data available indicate that the activated complexes in such reactions contain a, single molecule of RLi. The reactions of organolithium compounds with ketones, and with alkyl halides are first order in RLi and first order also in the reagent, undergoing attack. The rate law for the reaction with ketones55 suggests that, it proceeds through an addition compound such as LI 11, which decomposes to, the observed product, the alkoxide LIV; and it is possible that other “addition, R'-Li, , R', , r\KH, c=o, , R, ^)c=0 + R'—Li, , R, -*■, , R, , R, , R, , \b-0~Li, , LIV, , LIII, , reactions” of organolithium compounds proceed by analogous paths. The, manner in which organolithium derivatives attack alkyl halides (RLi + R'Z, + LiZ) is not clear, but scattered experiments indicate that during, such reactions the alkyl group from the halide assumes considerable carbonium-ion, character. Thus, the reaction of rc-butyllithium with optically active ^e-butyl, bromide yields an almost completely racemic octane.50, Et, *, I, tz-BuLi + Et—GH—Br -> n-Bu—CH + LiBr, , I, , I, , Me, , Me, (racemic), , Moreover, phenyllithium reacts with a-methylallyl chloride to give the same, mixture of allylic isomers (in very nearly the same proportions) as result from, the reaction of y-methylallyl chloride,« a situation bringing to mind the hy¬, drolyses of very active allylic halides (p. 286), where, it will be recalled, the, intercession of an allylic carbomum ion was indicated., MeCH=CH—CH2 Cl, , 0, 0, , 0, 0, , 0, 0, , PhCH2—CH=CH—CHS, , MeCH-CH—CH., MeCH—CH=CH2, Cl, , (90 percent), , Ph—CHMe—CH=CH2, , w ZooknanddG^Hnt’, 72’ 518, <1950)not and Goldey,J'tA7, J. Am. Chem. Soc75, 3975, M95Ti, Cristol, Overhults, and Meek J Am Ch, c, k, ’, j. Am. them. Soc., 73, 813 (1951), , (10 percent)
Page 418 :
402, , Carbanions and Enolization, , The following mechanism, in which one molecule of the lithium compound, acts as a Lewis acid, seems consistent with the facts:, / ,, slow, , R +, , R'—X + Li—R ->, , R—Li, , -> R—R + Li+, fast, , T:Li:RIn contrast, the reaction of an organosodium derivative with an optically, active alkyl chloride (having an asymmetric a-carbon) generally yields an, optically active hydrocarbon. Although we cannot be sure, it seems very likely, that inversion of configuration about the asymmetric carbon has occurred, during the attack by the carbanion.^ In a typical case,S;?(o), Ph, , *, , I, , Na+Ph2CH- + Z)-CH3CHPhCl -> L-Ph2CH—CH (probably) + NaCl, Me, Concerning the mechanisms of reactions of Grignard reagents, we can do little, more than speculate, for the Grignard reagent, which is often blandly desig¬, nated as “RMgA,” is now known to consist of a number of different species in, solution. In addition to RMgX, evidence has been obtained for the existence, of R2Mg, MgA2, MgZ+, R2MgA~, and polymeric species, in a solution that, would ordinarily be employed as that of a “single Grignard reagent.,,<?s(o) In, view of these complexities, the assumption that RMgAr is the only active species, (or even that it is the most active species) in a solution of a Grignard reagent, takes a great deal for granted.55(6) Moreover, any kinetic study of the reaction, of a Grignard reagent must be very carefully devised so that there is no am¬, biguity concerning the species whose disappearance is being followed. Thus,, aside from experimental difficulties, rate studies of Grignard reactions offer, vexing interpretive problems., One point, however, seems quite certain, in a number of cases the addition, of a Grignard reagent to a carbonyl group requires two molecules of magnesiumcontaining species. There are a number of instances known* where careful, admixture of “equivalent” quantities of ketone and Grignard reagent yields a, - (a) Bergmann, IIclv. Chim. Ac,a, 20, 590 (1937). (b) For further examples, see Letsinger,, TA, ni, Lr 70 400 fl048V 73 2373 (1951); and LeGofT, Ulrich, and lJennty,, elbh ’’go 622 (1958) (c) The stereochemistry of the reactions of alkyl sodium derivatives win, ,, , “At- s tz^t si*, , cm.
Page 419 :
Organometallic Compounds, , -, , 403, , precipitate, which, upon treatment with water, yields the ketone once more., Generally in such cases, the addition takes place in the usual manner if excess, Grignard reagent is employed. Thus it appears very likely that one molecule of a, magnesium-containing species (which we may designate for convenience as, RMgAr) acts as a Lewis acid. Here, the electrophilic magnesium atom is bound, to the oxygen of the carbonyl group, pulling electron density from the carbon, and making it more susceptible to attack by a second “molecule of RMgV”:, R, , \, , ;C, i = 0 + “RMgX’, , \, , -XG—CMMg'7 . RMgy\*_o;Mg/, , _, , - /, , h2o, , C-O-Mg, ——, /|, \, R, X, , /, , + MgX*, , N, , Mg-X, ^G-OH + RH, /|, R, , If this sequence is, in the main, correct, the addition reaction should proceed, with a single equivalent of RMgT, provided an additional Lewis acid is added, to perform the coordinating function in the initial step. It has indeed been, shown that a number of additions of Grignard reagents to C=0 groups are, facilitated by addition of anhydrous MgBr2, a typical Lewis acid.e5, The reader is doubtless aware that there are several ways in which the, reactions between ketones and Grignard reagents may “go astray.” With a, branched ketone or a branched Grignard reagent (or both), the ketone may, be reduced to the corresponding carbinol while the alkyl group of the Grignard, is converted to an olefin. Isotopic studies have shown that this reaction involves, , a transfer of a /5-hydrogen from Grignard to ketone.** It is probable that this, reduction proceeds through a cyclic activated complex, LV, similar in nature, to that thought to be involved in the Meerwein-Ponndorf (aluminum iso-, , V + H-c-0, I, MgBr, , C-H, , A, , ->, , i, o-A, , G—a, , /, O, , O, , /, , C—, , G, +, , o., , -Mg 1, , G, Mg, , Br, , Br, LV, , propoxide) reduction of ketones (Chap 13) Note, iispair °f euarms; tha*is’the, , ^, , a Mridfs:“"g
Page 420 :
404, , Carbanions and Enolization, , to this mechanism, a molecule of Grignard reagent must become coordinated, with the oxygen atom of the ketone in order to transfer its 0-hydrogen to the, carbonyl group. This being the case, the reduction reaction should be inhibited, by addition of a Lewis acid to the reaction mixture; for any ketone molecules, coordinating with the added Lewis acid would be unavailable for coordination, to RMgX But the addition reaction of Grignards is, as we have seen, facilitated, by such Lewis acids as MgBr2 and Mgl2. We should then expect that in cases, where both addition and reduction occur, the addition of anhydrous magnesium, halides will increase the extent of addition but decrease the extent of reduction,, and this is what is found. Thus, in the reaction of H-propylmagnesium bromide, with diisopropyl ketone, the ratio of addition to reduction is 0.48 in the absence, of added Mgl2, but rises to 2.5 when excess Mgl2 is added.55, , EXERCISES, , FOR, , CHAPTER, , 10, , 1. Predict which member in each of the following pairs is the stronger acid. Justify your, choice in each case., (a) CH3N02 or «-PrN02?, , (b) CH, | —S -Ph | or CH., , (c), , CHPh2 or Ph2CHCH2Ph, , (d) Methyl acetoacetate or acetoacetamide?, , (e) CH3COCH2COCH2F or CH3COCHFCOCH3?, , ?, , ?
Page 421 :
Exercises for Chapter 10, , 405, , 2. Predict which member in each of the following pairs is the more extensively enolized., Justify your choice in each case., (a) CH3COCH(f-Pr)COCH3 or f-PrCH2COCH2COCH3?, (b) 1,2-Cycloheptanedione or 1,3-cycloheptanedione?, (c) Cyclopentanone or diethyl ketone?, , o, , o, , (1) Benzoylacetone in toluene or in nitrobenzene?, , o, (h) PhCH2COCH3 or PhCHMeCOCH3?
Page 422 :
406, , Carbanions and Enolization, , 3. The position of equilibrium in the hydroxide-catalyzed aldol condensation of acetone, is such that the forward reaction (ketolization) is much less easily studied than the, reverse reaction (deketolization)., , oh-, , _, , ^, , Me2C—CH2—C—CH3 (ketol)-> 2Me,C—O, , I, , II, , OH, , O, , The rate law for deketolization is, — (/(ketol), , = £(ketol)(OH ), , dt, , (a) Show that this rate law indicates that the conversion of acetone to its conjugate, base is a rapid reversible reaction, but that the conjugate base condenses wtth a, second acetone molecule slowly. (Note that according to the principle of rntcrc,scopic reversibility, the steps in the deketolization are the reverse of the steps in, fb) Whi'chdo you think would proceed more rapidly (given the same concentration of, base): the ketolization of acetone or its base-catalyzed brom,nation?, 4, , Consider the mechanism for the base-catalyzed aldol condensation (p. 390),, H, B:, ki, , —C—CH2, , II, O, , v, , BH+, k-1, , RiC=0, ki, , -c-, , I, , -CH—, , —C^R2C=0, k-i, , o, , (a) "ra:—z, , O, , R, , -c—C—R s B:,k->, - ~=* aldol, o_
Page 426 :
410, , Carbanions and Enolization, , +, , CH3Li, , ->■, , 7. In the Mannich reaction, a compound having a labile C—H bond, an aldehyde, a, primary or secondary amine, and the ammonium ion derived from the amine react, to form a so-called Mannich base (LV)., H, NR2Hj+, —C—H + RCHO + HNR-> — C—C—NR2 + H20, , I, , I, , I, , R, LV, , Propose a mechanism for the Mannich reaction consistent with the following observa¬, tions:, (a) The reaction conditions used are often significantly milder than those needed to, carry out an aldol condensation between the first two reactants., (b) Generally, when the aldol condensation product from the first two reactants is, prepared independently and is treated with the amine, the Mannich base, if it, forms at all, forms more slowly than when the three components are mixed., (c) The reaction is third order overall, first order in each of the three components., (d) The reaction is inhibited in strongly basic and strongly acidic solutions., (e) Reactions requiring the collisions of three molecular species in solution within a, very short time interval are generally extremely slow., 8. Explain the following:, , O, (a) The C=0 group in — C—CH— is more effective in increasing the acidity of the, a-hydrogen than is the benzene ring in C6H5, , CH, , ., , (b) Tri(o-tolyl)methane is less acidic than tri(/?-tolyl)methane., (c) The hydrogen exchange between 2-picoline and water is far more rapid than that, between 3-picoline and water., (d) When ICH(COOEt)2 is treated with HI, elemental iodine is formed., (e) The a-methylation of ketones with methyl iodide is much more effectively cata¬, lyzed by dimethylamine than by trimethylamine., (f) The base-catalyzed bromination of ketone LVI is far slower than that of ketone, LVII., , Ph, COPh, NO 2, LVI, , from the basic hydrolysis of HCKr3;, ana oor, much less than that for HCBr3., , LVIII
Page 427 :
Exercises for Chapter 10, , 411, , (h) The preferred position of attack in the bromination of 3-methylcyclohexanone, is the same in acidic as in basic media. With isopropylcyclohexyl ketone, the pre¬, ferred positions of attack are different in the two media., , ^, , (i) The attack of Na+CH(COOEt)^ on CH3CH=CHCH=CHCH=CHCOOMe, occurs at carbon 3 or carbon 7 but not at carbon 5., (j) When a Grignard reagent can undergo both 1,2 addition and 1,4 addition to an, a,/3-unsaturated ketone, addition of MgBr2 to the reaction mixture generally, lowers the ratio of 1,4 to 1,2 addition., (k) The effect described in (j) is not observed in the reaction of PhMgBr with ketone, LVIII although both 1,2 and 1,4 addition occur., (l) There is a linear free-energy relationship (that is, a Br0nsted-catalysis-law rela¬, tionship) between the rates of amine-catalyzed bromination of nitroethane in, water and the basicities of the amines in chlorobenzene, but no such relationship, exists between bromination rates and basicities of the amines in water., (m) When PhCHMe—N=CPh2 and NaOEt are added to EtOH enriched with, EtOD, the rate of enrichment of the imine with deuterium is at first equal to its, net rate of isomerization to PhCMe=N—CHPh2, but as the reaction proceeds,, the rate of deuterium enrichment runs ahead of the net rate of isomerization., (n) Cation LIX, unlike cation LX, is ineffective as a catalyst in the benzoin con¬, densation.
Page 428 :
CHAPTER, , 11, , Electrophilic and Nucleophilic, Substitutions in Aromatic Systems, , Aromaticity, Benzene may be said, , to be the great grandfather of all aromatic compounds., , In the development of our present ideas concerning aromaticity, this hydro¬, carbon played a unique and immensely important role;1 for once the essential, nature of the benzene ring was understood, the behavior of other aromatic, compounds could be interpreted, for the most part, with little difficulty. To, many chemists, in fact, the term “aromatic compound” is synonymous with, “benzene derivative” (since even such fused-ring hydrocarbons as naphthalene, and anthracene are actually benzene derivatives). But a number of heterocyclic, systems have long been known which exhibit behavior sufficiently similar to, that of benzenoid compounds so that they too may be classified as “aromatic”;, the most familiar of these are thiophene, pyrrole, pyridine, and quinoline., There are, in addition, certain nonbenzenoid compounds of more recent vintage, that, at least in some respects, may be considered aromatic., Of the various criteria for aromaticity which have been suggested, two, seem the most useful. The first is resonance energy. This quantity, it will be recalled,, may generally be obtained from the heat of combustion or heat of hydrogena¬, tion, and tells us how much more stable a compound is than a second (h>po■ The early attempts Co propose a suitable single structureTor benzene., , Structure and Mechanism in Organic cnemisiry,, , 7, , v, , ., , 1ROq, , ChaD 3., , (c) Lachman, The Spirit of Organic Chemistry, Macmillan, New York, 1899, Chap., , 412
Page 429 :
Aromaticity, , -, , 413, , thetical) compound having the same set of bonds as the first but without the, cyclic conjugated system (p. 37). The resonance energies of pyridine, pyrrole,, and thiophene, for example, are each close to 25 kcal per mole—comparable, with the value for benzene (36 kcal per mole) and at the same time much greater, than the resonance energies of conjugated but noncyclic dienes (<5 kcal per, mole) which may be regarded as nonaromatic., A second feature pointing to aromaticity in a compound is its tendency to, undergo substitution reactions with a number of reagents that ordinarily simply, “add across the double bond” in olefins. The most familiar and important of, such reagents are electrophilic in character. Thus Cl2 or HOC1 will chlorinate, benzene derivatives with displacement of H+ but generally convert olefins to, dichloroalkanes or chlorohydrins; similarly, HBr adds to the C=C double, bond of a number of olefins, but with aromatic compounds only substitution, (hydrogen exchange), , occurs., , Concentrated sulfuric acid, which ordinarily, , converts aromatic compounds to sulfonic acids, often converts olefins to alkyl, hydrogen sulfates instead:2, RCH=CHR + H2S04 -» RCH2—CHR—0S03H, The reluctance of aromatic compounds to undergo addition reactions is, like, their resonance energies, a reflection of the extra stability associated with cyclic, conjugated systems; for in addition reactions these systems are destroyed whereas, in substitution reactions they are preserved., As with other effects related to conjugation, aromaticity requires that the, atoms comprising the conjugated system lie in or near a common plane. We, might expect the hydrocarbon cyclooctatetraene (I) to exhibit aromatic proper¬, ties since, at first glance, it appears to be a “benzenelike” ring with eight in¬, stead of six members. However, if the molecule were planar, with each of the, on, , angles in the ring the same, these angles would be 135°, that is, 15° greater, , *“ri*C C~C7C ^ond a"S‘es ^ “normal” conjugated systems. It is not, be nla, k, 'hat the cyclooctatetraene molecule has been found not to, be planar, but rather “tub shaped” (II) . Four of the bonds are 1 34 A, , II, , III, , IV, , * Nori-is and Joubcrt, J. Am. Chem. Soc., 49, 873 (1927)., Mark, Nature, 161, 165 (1^48))^^, asis of spectral studies (Lippincott et at, , diffractlonstudies (Bastiansen, P, , at.,, , Natur'l, , (Kaufman, Fankuchen and, J Am Ch * C „Stru^ur® has been proposed on the, 160 128^1947^, ’ 1^° (1951)’ and ^ctron-, , sent appears to favor the “tub” form. At any rate the, 7^, the bulk of evidence, At any rate, the molecule is assuredly not planar.
Page 430 :
414, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , (about the length of ordinary double bonds), whereas the remaining four are, 1.54 A, the length of ordinary single bonds. Moreover, with respect to its chem¬, istry, the compound has been found to be nonaromatic. Its resonance energy is, about 5 kcal per mole/ and it does not undergo the usual aromatic substitution, reactions (although its tendency to rearrange in the presence of electrophiles, prevents our classifying it as a “typical nonaromatic polyene”)., The observed lack of aromaticity is in accord with the nonplanar structure,, but there is some theoretical basis for the belief that even if the molecule were, planar, it would not have properties akin to those of benzene. A quantummechanical treatment of conjugated cyclic systems indicates that the unusual, stability which we term “aromaticity” should exist only for rings associated with, (4n + 2) 7r electrons, where n is an integer; that is, that rings having 6, 10, and, 14 7r electrons may be aromatic (provided they are planar) whereas those with, 4, 8, and 12 may not.5 Without the supporting calculations, which will not be, reproduced here, such a result is necessarily accepted on faith, even though it, looks a little like a rule of numerology. However, it is of less practical than, theoretical interest for, aside from azulene (III) and its derivatives, very nearly, all actual systems falling within its scope involve just six tt electrons or may be, broken down into six-electron units which “share edges.”5 Thus, excluding a, few very unusual compounds, the characteristic structural feature of aromatic, systems may be taken to be the 7r-electron sextet. With benzene and pyridine,, each atom in the ring supplies a single % electron to the sextet; whereas with, furan, thiophene, and pyrrole, the “hetero atom” supplies two tt electrons and, the four carbon atoms one ir electron each. The “7r-electron sextets” in pyridine, and pyrrole may be represented schematically as shown in Figure 11-1. Note, that pyridine has an extra pair of p electrons (in the shaded orbital) with which, it may coordinate with an . acid without disrupting the “aromatic sextet”;, whereas with pyrrole, in which both tt electrons of nitrogen have been incor¬, porated into the sextet, this is not possible. Thus, pyridine coordinates much, more readily with acids (that is, is a much stronger base) than pyrrole., Removal of a proton from cyclopentadiene yields the cyclopentadiemde, 4 Springall, White, and Cass, Trans. Faraday Soc., 50, 815 (1954)., 4 Htickel, Z. Electrochem., 43, 752, 857 (1937), Craig, J Chem. Sac., 1951, 3175 For bn f, summaries of this question in nonmathematical terms, see Longuet-Higgins, Proc. Chem., t957, 157; and Wheland Resonance in Organic Chemistry, John Wiley and Sons, Inc., Ne, ,955; Such aromatics as naphthalene and anthracene (having,, , respectively, , trons) may obviously be broken down into edge-shaung benzene rings., , 10, , and, y have, , cyclic polyenes as cyclodecapentaene C i »H > a^aimos^certaTHy not planar molecules. More-, , »^^r^,bldien^ JV;, e1ec.ronF:/r^,^wL^dCrkr0dUt. aOT., and Ind., 1956, 1306., , into 78, 653 (1,56,; and Petit,
Page 431 :
Aromaticity, , anion (V), whereas removal of a H:, , -, , 415, , ion from cycloheptatriene (in principle), , yields the cycloheptatrienylium cation (tropylium ion),r which is, however, in, , o o, V, , VI, , practice obtained in a different manner (see below). Since both of these ions, have six tt electrons in a cyclic system (5 + 1 in anion V, and 7 - 1 in cation, VI), we might expect to find aromaticity in both. But it is a little difficult to, know where to look; for the reactions which these ions undergo are very much, different from those of ordinary aromatic systems in which the rings bear no, net charge. Moreover, the resonance energies of the ions cannot be determined, in a straightforward manner. Nevertheless, it does appear that these ions are, significantly more stable than other carbanions and carbonium ions. The, tropyhum ion, which results when tropylium chloride or bromide is dissolved in, alcohol or water/ is probably the most stable carbonium ion yet preparedfor although the ion is eventually converted to ditropyl ether in water or ethyl, ropy et er in alcohol, solutions containing it may be studied without undue, diffi u‘ y, although at a not too leisurely pace. Infrared studies indicate that, JV / T“ m Th S0‘U,i0nS iS P'anar and that a11 of the carbon-carbon, , r“, , in whichVe, , ,S3me, , :::rzr’ ~, zzzssr, g‘h,> ‘hUS bdn? C°nSiStem With an “aroma'ic cation”, , ......., eagents tha, carry out subsutution reactions on benzene and thio• Fatd"veam!JrKn°X’ J' Am' Ch,m' Sx'> 76’ 3203 (1954), 7 and I-ippincott, J. Am. Chm. Soc., 77, 249 (1955).
Page 432 :
416, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , phene. As we shall presently demonstrate, the active intermediates in most of, such substitution reactions are cations (for example, Br+, NO^, and R—C=0),, and such cations would tend merely to form a new covalent bond with the, cyclopentadienide ions, yielding a neutral (but nonaromatic) product. But, there is a group of compounds, closely related to the cyclopentadienide ion,, that is unquestionably aromatic in character. These compounds are the ferrocenes, in which an iron atom is “sandwiched” between two cyclopentadiene, rings.5 The simplest member of the series, ferrocene itself (dicyclopentadienyliron, VII) may be prepared by the treatment of FeCl2 with cyclopentadienyl-, , VII, sodium; it may be considered to be a pair of parallel C5H" anions from which, enough, , r-electron density has been drained toward the central Fe2+ ion to, , 7, , allow formation of two very strong 7r bonds.i0 Ferrocene undergoes FnedelCrafts-type alkylations and acylations in much the same way as do benzene,, thiophene, and naphthalene, and also undergoes attack by the benzenediazonium ion in the same way as do the more reactive of the usual aromatics., Ferrocene has not yet been halogenated or nitrated directly, for its iron atom, is readily oxidized to the “+3 state,” yielding the so-called ferricinium ion,, (C5H6)2Fe+, Azulene (III) and its derivatives may be nitrated, halogenated, and made, to undergo Friedel-Crafts acylations and alkylations, but the reaction conditions, must be mild, for the azulene ring system is readily destroyed., , In add,non, , to taking azulene as an example of a “ten ^-electron aromatic system,, , it may, , be regarded as a combination of a tropylium cation and a cyclopentadien, anion sharing a “common edge.” Such a view is consistent with the rather, , ig, , • For summaries of ,he chemistry of ferrocene, Quarl. Revs., 9, 391 (1955); and Fischer,, , a, number of different, •‘sandwich compounds,” structurally sira'l“t°off'^th^ium compound (CsHd.Ru, these, am'far^'esTstabie^tha^fcrrocene^aml'its derivatives, and little is known concerning then, Chem»The nature of the hybridization of orbitals, open question. See Moflitt, J. Am. Chem. St*., 76, 3386 (W54, a, _, , ., , 75, , 171, 121 (1953)., , _, , For, , =u^:se?Gor^’<£-, , r., , o ., , 4980, , wZ, , (1950); 77, 6321 (195^50, ,39 (1953).
Page 433 :
Aromaticity, , -, , 417, , dipole moment of azulene (p. 64) and with its tendency to undergo electro¬, philic substitution on the five-membered (the more negative) ring, rather than, on the seven-membered (the more positive) ring.'*, An additional interesting cyclic system having some aromatic character is, found in the sydnones,ls a set of compounds so named because they were first, studied at the University of Sydney. The available electrons in the sydnone, CH2—COOH, , /, , -h2o, —-—>, , Ph-N^, , +/, , Ph—N, , N=0, VIII, , \, , CH=C-0“, , N-O, IX, , ring of N-phenylsydnone (IX), a compound prepared by the dehydration of, N-nitroso-N-phenylglycine (VIII), are represented schematically in Figure 11-2., Once again, six 7r electrons are associated with the heterocyclic ring. It is thus, not surprising that the sydnone ring undergoes halogenation;*(6) and nitration^(c), , Fig. 11-2. N-Phenylsydnone, , at the hydrogenated carbon. Additional electrophilic substitution reactions (for, example, acylation and sulfonation) on the sydnone system have not as yet been, reported, but only a few attempts seem to have been made., At this point it may be asked why electrophilic-substitution reactions, which are so rare in aliphatic systems, become much more commonplace in, aromatic chemistry. In the great majority of such reactions, the leaving group, G-RPK°T’mud thC reaCti°n then recluires the breakage of a rather strong, bond. The activation energy, which must be supplied largely by the, attacking electrophile, is necessarily high, and there is, moreover, no particular, an r T Y a" elCCtrophile should at'ack a relatively nonpolar C-H bond of, an aliphatic compound. With aromatic substrates, however, there is one itn, tant circumstance that favors electrophilic substitution. The asking elec, , J, , J- Chem. Soc., 1950, 1542., , ’’, , ’, , f "1 °llis-, , Qu°rt- X™-, , XI., , 80 0956), and (c) Baker, Ollis, and Poole,
Page 434 :
418, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , trophile may form a new bond before the old bond breaks, and the energy that becomes, available when the new bond forms may be used to help bring about the break¬, age of the old bond. We might then anticipate that the attack on an aromatic, system by an electrophile A+ will pass through an intermediate (or a transition, state) of the type X, in which both the attacking species and the leaving group, are bound to the substrate. An intermediate analogous to X in an aliphatic, , X, substitution would have five bonds to a single carbon atom and is therefore, prohibited., With nucleophilic substitutions, the tables are turned, for these are much, more common among aliphatic compounds. The bonds between carbon and the, more effective leaving groups (for example, C, , Br, C, , I, and C, , OTos) tend, , to be weak and are considerably more polar than C—H bonds. The polarities, of such bonds favor the initial nucleophilic attack on the carbon, whereas the, weakness of the bonds eases the departure of the leaving group. However, when, • •, , such a leaving group (which we may represent generally as —Y:) is attached to, a benzene ring, one of its p orbitals lies parallel to the tt orbitals of the ring. The, resulting conjugation allows electron density to drift from —T: into the ring, , (XI ^ XI'), lowering the polarity of the C-F bond and allowing it to assume, some double-bond character. Clearly then, this conjugation should decrease, ease of nucleophilic substitution at the carbon atom bound to group, , •., , for that carbon has become less positive and the C -1 bond has been stl“S, ened. Moreover, as has already been pointed out (p. 217), attach*e, , £, , second nucleophile to this carbon (as is necessary in the transition s, substitution) cuts down the .-electron system from seven atoms, , ^, , although nucleophilic substitutions on aromatic substrates have long, known, , they generally require much more severe conditions than, , hose, , aliphatic substrates. They occur most readily when the electron density m
Page 435 :
Nitration. The Nitronium Ion, , 419, , aromatic system is diminished by the presence of one or more strongly electronattracting groups (most often nitro groups)., In considering electrophilic aromatic substitution, we shall devote our, attention largely to nitration, halogenation, sulfonation, and the Friedel-Crafts, reaction. The mechanisms of other electrophilic substitutions, among them, hydrogen exchange,13w, ArH + D+, , ArD + H+, , diazo coupling,mb), ArH + Ar'N+ -> Ar—N=N—Ar' + H+, and bromodesulfonation,7S(c), ArS03H + Br2 + HoO -> ArBr + Br~ + H+ + H2S04, have also been investigated. However, despite some interesting differences in, detail, these reactions appear, at present, to follow the broad patterns set by the, more usual aromatic substitutions., , Nitration. The Nitronium Ion, Of the various types of electrophilic-substitution reactions, nitration currently, presents the clearest mechanistic picture. As has been hinted several times in, earlier chapters, the attacking electrophile in nitration is often (but, as we shall, see, not always) the nitronium ion, NCkf. Since this cation is not found in the, ordinary “store-room reagents,” we should first make sure that it exists before, apprehending it as the culprit in nitrations. The most convincing evidence arises, from the structure determination (by x-ray crystallography) of solid N02CKV'(a) and N206.^(« The first of these compounds was found to contain the, familiar CIO7 ion, and the second was found to contain the NOj ion- but in, addition, both were found to contain what appeared to be the same linear, inaiome species. From the stoichiometry of these compounds, this species must, , Nolrin, 2, , ,, , 4, , ^, thC C°mpounds may be represented, respectively, as, and NO+NO3 • Other nitronium salts, all necessarily derived from, , Na,ur, >62 M, , “ («) Cox, Jeffrey, and Truter, , Acta Cryst., 3, 290 (1950)., , fuming sulfuric acid., , ’, , P, , Y, , ’, , '■, 2927 <1957>2’ 258 ^1948)- W Gnson, Ericks, and de Vries,, , acids ^HS*o’, The ‘a"er ,wo, aCldS H2S2°7 and H^3O10. These acids are found in
Page 436 :
420, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , pounds have not been reported, spectral evidence indicates that each assuredly, contains the nitronium ion; for the Raman spectrum of each shows a line at, Raman frequency 1400 cm-1, a line that is also found in the Raman spectra of, NO2CIO4 and N205,1<J(o) and in the Raman spectra of mixtures of concentrated, nitric and sulfuric acids.I6(-b) Some ionic or molecular species must be common, to each of the solid compounds and must also be present in the mixture of acids., Since the anions are different in each case, we may conclude that the cation is, the same—that is, that these materials, like N02C104 and N205, contain the, NO* ion., Finally, cryoscopic measurements show that nitric acid in solutions of con¬, centrated sulfuric acid exhibits a v-factor (p. 98) of very nearly A,17 indicating, that sulfuric acid reacts with nitric acid in the same way as with triphenylcarbinol, breaking off a hydroxyl group as water, then converting it to H30+, HNOs + 2H2S04 -+ NOt + HsO+ +, , 2HSOj-, , (v = 4), , Having established the existence of the nitronium ion, we should next, clarify its role in substitution reactions. The nitrations, in concentrated sulfuric, acid, of such aromatic compounds as nitrobenzene, benzoic acid, and benzenesulfonic acid have long been known to exhibit second-order kinetics, , first order, , each in aromatic substrate and in “added nitric acid.’ 1S In 90 to 100 percent, sulfuric acid, virtually all of the nitric acid has been converted to NO*. There¬, fore, for these solutions, the reaction may also be said to be first order in, nitronium ion. For example, for nitration of nitrobenzene,, rate = yt(PhN02)(“HN03”) = *(PhN02)(NO*), , (1), , This rate law is obviously consistent with an activated complex containing the, substrate and one nitronium ion, but it does not demand it, for these kinetics, would also be observed if the attacking reagent were any species having a con¬, centration (large or small) that remains proportional to (NO*) during the course, » M Millen, J. Chem. See., 1950, 2600, 2606. (» Medard, Cempt rend->1W61M1»£, , Furthermore, Millen fo^XmanSe”^ rf'tfbraUon Thl selection rules for, 1400 cm- possesses no oher, , .f the responsible species is diatomic or .t, The nitronium ion is, as we have seen, an ,on, , °f th» GmespTii of, , Nate,,, 158, 480 (1946); 7. Om. Sec, 1950, 2473, 2493. The small ap¬, , parent deviation from the fourfold value of »*«l.n ‘^'“‘"“Svely nonpolar, « (a) Martinsen, Z. physik. Chem., 50, 385 1904),, 0U3 U ’>, in sulfuric, , aromatics such as benzene, toluene,, acid, and their rates of nitration are gen, Anisole and phenol are, , y, , bythe rates at which they dissolve,, raDidlv for convenient kinetic, results have since been confirmed byj, , rXTf^r^e foTetample; VVcstheimer and Kharasch, 7. dm. Oem. ^., 68,, (1946).
Page 437 :
Nitration. The Nitronium Ion, , -, , 421, , of the reaction (for example, HNO,, H2NOJ, N02HS04). Moreover, this rate, law tells us little about how the attacking species is formed., Thus it is clear that sulfuric acid, despite its value as a solvent in preparative, nitrations, complicates the mechanistic picture. More detailed information has, been obtained by the study of nitrations in the organic solvents, acetic acid and, nitromethane, in which such aromatics as benzene, toluene, and xylene are, easily soluble. Moreover, the nitrations of such compounds in these solvents, proceed at rates that may be conveniently measured. To minimize complica¬, tions due to the water formed in the reaction, a large excess of nitric acid is, often used, although this makes the determination of the kinetic role of HNO3, more difficult.19 Under these conditions, aromatic substrates (aside from amines, and phenols, which comprise a special problem) fall into three categories, based, upon three different modes of kinetic behavior., The rates of nitration of such aromatics as jfr-dichlorobenzene and ethyl, benzoate (that is, compounds which, in classical terms, undergo nitration “with, difficulty”) are, as may be expected, proportional to the concentration of, substrate. Since nitric acid is present in large excess, it does not appear in the, kinetic expression, and these nitrations are pseudo-first order., For aromatics that are “more active” than benzene (for example, toluene,, the xylenes, jfr-chloroanisole), nitration rates are independent of the concentra¬, tion of substrate. With nitric acid in large excess, these nitrations are pseudo¬, zero order. Their rates remain constant during the progress of a given nitration, , and suddenly drop to zero when all the substrate is consumed. Moreover, the, nitration rate for each of the compounds in this category is the same in a given, nitrating mixture., Thirdly, as we might anticipate, there are a number of aromatics of inter¬, mediate activity which may be considered “borderline cases.” Benzene and the, monohalobenzenes, for example, exhibit pseudo orders between 0 and 1 in, nitration, that is, their rates depend upon the concentration of the substrate, but are not directly proportional to it. In this intermediate region, certain, of the nitrations may be made to follow either pseudo-first- or pseudo-zero-order, rate laws (but generally not both) by suitable adjustment of the reagent, concentrations., One may now ask, “Do these kinetic differences reflect differences in, mechanism?” The answer is almost certainly “no”; for although'three cate¬, gories are listed above for descriptive convenience, we observe, in perspective, a, n inuous range of kinetic behavior. It is far more likely that a single mechaS* fOT thesuh;stra,es "S wftMn, , S anSe, , three “categories” and that the observed, , ere"CeS in thC rclativc rates °f the -dividual, , •iid rrrreac sequence-For nitrations ha™s, g, , , « <d; J. Chem. Soc., 1938, 929; 1950, 2400; ffatur,, 158, 448 (1946)., , mde.
Page 438 :
422, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , pendent of the substrate concentration, the substrate must not become involved, in the reaction sequence until after the slow step; the slow step must then be the, formation of an active intermediate (that is, the attacking electrophile) with, which the substrate then reacts very rapidly. This means that the attacking, species is neither nitric acid itself nor its conjugate acid H2NOJ (the formation, of which from HNO3 should be very rapid since it requires only a proton, transfer between oxygen atoms) but is almost certainly the NO£ ion, the forma¬, tion of which from HNO3 or H2NO^ requires the breakage of a N—O bond., It is possible, however, by using a less reactive aromatic, to slow down the, step in which electrophilic attack occurs so that this attack becomes rate deter¬, mining. The formation of NO^ would then become the more rapid step;, however, this ion may not accumulate in solution, for solutions of nitric acid in, organic solvents are known from other evidence to contain only small quantities, of NOt. The formation of NOt must then be reversible. The rate of nitration, in such cases is proportional to (NOt) (substrate); and (NOt) remains nearly, constant since it is in equilibrium with a constant concentration of HNO3., Pseudo-first-order kinetics are therefore observed. If we represent the sequence, of steps in nitration as follows:, formation (2 steps), , 2HN03, , reversal, , -, , NOt-» ArNOo + H+, No- at,ack, , h2o, , we may summarize the conclusions thus far as follows: if the “attack” step is, much faster than the “reversal” step, pseudo-zero-order kinetics are observed;, whereas if the opposite is true, pseudo-first-order kinetics are observed. Finally,, it may be shown, using the steady state treatment (Ex. 2c), that when the, “attack” and “reversal” steps have comparable rates, the nitration will exhibit, an intermediate nonintegral order that will depend upon the concentrations, of the reagents., As indicated in the sequence above, the formation of the nitromum ion, proceeds in two distinguishable steps rather than a single step. We may draw, this conclusion by noting the action of added nitrate on the pseudo-zero-or er, nitrations; for nitrate greatly retards these reactions but does not alter the,, kinetic character.'9 Suppose that the nitromum ion were formed from, in a single step., 2HNOs ^ NO+ + NOj + H20, If this were so, the action of nitrate in reversing the formation of NO+ would, Iw down the reaction (as is observed). However, NOy and the, be competing for'NOf, and the reaction rate should be increase, the concentration of substrate, since in doing sowemae,^», competitor for NO+. Furthermore, added water should ) c, , y ■, as
Page 439 :
423, , Nitration. The Nitronium Ion, , added nitrate in retarding the formation of NO£, hence in retarding the nitra¬, tion. But the rates of nitration of toluene, ethylbenzene, and the xylenes are, as, we have emphasized, independent of the concentration of substrate, whether or, not additional nitrate has been added; and the retarding action of water,, although observable, is far less than that of nitrate. We must conclude then that, the formation of the nitronium ion from nitric acid does not occur in a single, step., The following two-step process for the formation of nitronium ion is in, reasonable accord with the kinetic data:, , 2HN03^ {, , H2NO+ ^ NOt + HoO, NO? ', , Here again the nitrate ion retards the overall reaction by reversing the initial, step, but there is no longer any competition between nitrate and substrate, since, the former reacts with H2NO£ whereas the latter reacts with NO^. By increasing, the concentration of substrate we merely increase the quantity of material that, remains waiting to consume the small supply of NO£ as it is generated, and the, nitration rate is, as observed, unaffected.*0, The process of nitration requires not only the attachment of the nitro group, to the substrate, but also the breakage of a C—H bond and the departure of a, proton. Earlier in the chapter it was suggested that nitrations (as well as other, electrophilic substitutions) occur much more readily in aromatic than in ali¬, phatic systems because the new bond may form before the old bond breaks;, that is, the bond-making and bond-breaking steps are not contemporaneous., This has been demonstrated, not by classical kinetics, but by experiments, chosen to show the absence of hydrogen isotope effects in nitration.*7 In our, initial consideration of isotope effects (p. 194) it was noted that deuterated, mtrobenzene, CeDtN02, is nitrated at the same rate as ordinary nitrobenzene, Un H2SO4, , HNO3 mixtures),*"-) showing that replacement of a C—H bond, , wtth a C -D bond in nitrobenzene does not affect the activation energy for, attack by NOJ. Moreover, returning to nitrations in organic solvents, we find, at substitution of deuterium or tritium for “light” hydrogen in benzene, ouene, or naphthalene does not retard the nitrations of these compounds- this, observation is of little value, however, for here the formation of No, , ift, , *an us attack on the aromatic, is known ,o be rate determining. In these cases, hydrogen-,so,ope effects (or their absence) are best demonstrated by' compel, " may be asked why water is so^uchtL'efftctbe^han'n't'r' f°rmafion of the nitronium ion,, T^ueslion is perhaps bes, answered, , and Noland,, , I953>26051 Lauer, V, , )■, , (C), , Melander, Arkiv Kemi, 2, 213 (1950).
Page 440 :
424, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , don” experiments in which a C—H and a C—D (or G—T) bond in equivalent, positions are allowed to compete for NO£ as it is generated. For example, when, benzene that has been “labeled” at just one position with tritium (C6H6T) is, mononitrated, just one sixth of the tritium atoms are removed. Similarly, when, toluene labeled in an ortho position with tritium (XII) is nitrated, just half of the, Me, , *T, 29 percent, , 29 percent, Me, , -f traces of meta isomers, , ortho-hydrogens in the resulting o-nitrotoluene are labeled.*/(c) These experi¬, , ments, together with similar ones involving deuterium-substituted aromatics,, show that a C—T bond (or a C—D bond) is just as likely to be broken during, nitration as is a C—H bond in an equivalent position, telling us that, as with, nitrobenzene, substitution of heavy for light hydrogen at a given position does, not affect the activation energy for substitution by NO£ at that position. But we, have emphasized that the stretching of a C—T bond or a C—D bond is signifi¬, cantly more difficult than the stretching of an equivalent (but necessan y, “looser”) C—H bond. We may therefore infer that the activation process in the, attack of NO* at a C—H bond involves very little, if any, stretching of that, bond, and that the C-N02 bond is, as we suspected, very nearly formed before, the C-H bond begins to break. Nitration by NO* thus proceeds through a, 14-, , no5, cationic species,, , , and if we assume that the H+ ion cannot depart, , Ar, H, , J, , on its own but must be pulled off by some basie species in, admit that this cation exists in solution long enoug, , P, , ^, , attack of the base-that is, that it is an intermediate, rather than, a, , ^ ^, , to, , a o\, , complex. The following sequence then summarizes our presen p, usual mechanism for aromatic nitration*, , ^, , " Aromatic nitrations in ™ncer^^, about 90 percent H2SQ4 (see Refs. 18a, 18b). Ueiow, , °f
Page 441 :
Nitrous Acid and Its Derivatives in Nitration, , 425, , no2~i+, H+, , —H20, , HNOs ^ H2NO+, , /, , ArH, , NO^->, , B:, , ArN02 + BH+, , Ar, fast, , *\, , H, , J, , Nitrous Acid and Its Derivatives in Nitration, Aromatic nitrations are, as we have seen, generally retarded by addition of, nitrate. We should therefore expect that any substance which rapidly forms, nitrate when added to a nitrating mixture will likewise lower nitration rates., Nitrous acid, is, for example, known to be almost completely ionized in con¬, centrated HNO3, not as an acid, but rather as a base.*5, 0==N—OH + HNO3 ^ [0=N]+ + H20 + NOy, A similar equilibrium exists in solutions of nitric acid in organic solvents, but, the degree of ionization is slight (most of the nitrous acid having been con¬, verted either to N204 or N02, neither of which is of direct concern here)., At any rate, we need not be surprised that nitrations, both in pure nitric acid, and in organic solvents, are generally decelerated by addition of nitrous acid, (or nitrites), this action being similar to that which would result from the addi¬, tion of almost any base. The inhibition of nitration is particularly marked at, high concentrations of nitrous acid, for here a second pair of equilibria assumes, some importance., 2HN02, , N203 + HoO, , NO+ + NOy + H20, , Now the nitrite ion, like the nitrate ion, can reverse the formation of H2NO+,, a necessary intermediate in the formation of NOJ., H2NO+ + NOy — HN02 + HNQ3, , for removal of a proton from the, , N204 + H,0, , ' SulflJriC acid of (HS°4)> which is necessary, , hydrogen-isotope effect in nitration shows, 1S n° °nger tenable since the absence of a, activation process., WS that pr°t0n removal is, a significant part of the, ** Goulden and Millen, J. Chem. Soc., 1950, 2620.
Page 442 :
426, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , Moreover, since the nitrite ion is a stronger base than the nitrate ion, it is more, efficient in lowering the concentration of H2NO+ and is thus a more efficient, nitration inhibitor than is nitrate. Kinetic studies of nitration in the presence, of both large and small quantities of tripositive nitrogen have been carried out,, and the data are in reasonably good accord with the two modes of inhibition, described above*4 (Ex. 3)., Having considered the anticatalytic effect of nitrous acid on a number of, nitrations, the reader may be disturbed to learn (if he is not already aware), that nitrous acid actually accelerates the nitration of such very active aromatic, compounds as mesitylene, phenol, anisole, and aromatic amines. Such sub¬, strates are known to undergo nitrosation by nitrous acid, and the resulting nitroso, compounds are, in turn, known to be very rapidly oxidized by nitric acid to, the corresponding nitro compounds.*5, , ArH ^ Ar—N=0 ^ ArN02 + HN02, — H2O, , fast, , The kinetics of the nitrations of />-chloroanisole, />-nitrophenol, and mesitylene, in solutions containing both nitrous and nitric acids point clearly to this addi¬, tional mode of nitration—that is, nitration through prior nitrosation.*5 The, rate laws consist of two terms, the first corresponding to the “ordinary” zeroorder nitration by NOJ (which, as we have seen, is inhibited by HN02), the, second term being first order each in substrate and added HN02.*7 In concen¬, trated nitric acid, and, more particularly, when pains are taken to eliminate, nitrous acid and the oxides of nitrogen from the reaction mixture, nitration, through nitrosation assumes little or no importance. In more dilute nitric acid, (the nitration medium often used for very active compounds), in the presence, of significant quantities of nitrous acid, this mode of nitration can be made to, predominate, although nitration by the nitronium ion cannot be eliminated, entirely. Insofar as the “nitrosation-oxidation” sequence can be distinguished, kinetically from the nitration by NO+, the former appears to follow the rate, law (in acetic acid),, *4 Hughes, Ingold, and Reed, J. Chem. Soc., 1950, 2400., u See for example, Westheimer and Schramm, J. Am. Chem. Soc., 70, 1782 (1948), Viebel, Bn., 63, 1577 (1930)., ::ThrSeU:';ud“T.hfn;«?io^26ofphe„cl itself, the cr^ois, anda-Wfg, sents some difficulty. Independent of their, with the formation of nitrous acid. Since the, , niLion, the, ^, , reaction appears to be autocatalytia During a Rivcn, , „, , mncentrations of two of the, independent reactions (oxida-, , less ac,ive, aromatics such as the haloamsoles or the mtropheno s.
Page 443 :
Nitrous Acid and its Derivatives in Nitration, rate (of nitrosation), , -, , 427, , (2), , (ArH)(“HN02”), k + (NOj), , where (“HN02”) represents the total concentration of species that can be, titrated as tripositive nitrogen (present here very largely as N204). The form, of this expression indicates that there are two attacking species in the nitiosation, each having a concentration proportional to that of N204/5 It therefore, seems likely that these two attacking species are N204 itself (corresponding to, k) and the nitrosonium ion, NO+ (corresponding to k'). 4 he latter is generated, , from N204 in the following equilibrium:, N204, , NO+ + NOj, , and its concentration should therefore be inversely proportional to (NOj). The, nitrosation rate is apparently independent of the activity of water in solution,, indicating that attack of substrate either by HN02 or H2NO^ is not of impor¬, tance (Ex. 3). The oxide N0O3 is likewise excluded (in this case) as the nitrosating, agent, for its action would give rise to a term second order in (“HN02”)., It may be asked why the phenomenon of nitration through nitrosation, has been observed almost exclusively with the more reactive aromatic com¬, pounds and not with, say, nitrobenzene or the halobenzenes. To answer this,, we should remember: (a) that the NO+ ion is a much weaker acid and therefore, a much less powerful reagent in electrophilic substitution reactions than is the, NO;)- ion; and (b) that substantial quantities of NO+ may exist in solutions in, which all but tiny amounts of NO£ are destroyed (for example, nitrous acid, in moderately concentrated nitric acid). Now the nitrations of very reactive, aromatics are generally carried out in solutions in which (NO£) is very small, since otherwise polynitration will become troublesome; however, if nitrous acid, is present in such solutions, much may be converted to NO+ and N204, and, nitrosation assumes importance, not because these are efficient reagents but, because they are available. If a less reactive aromatic is treated with the same, nitration mixture, neither nitration nor nitrosation may occur at an appreciable, rate. But by raising the acidity of the mixture and decreasing its water content, it is possible to boost (NO+) greatly, while at the same time (NO+) and (NjO,), are increased only slightly (since tripositive nitrogen exists largely in these forms, already in less acidic mixtures). Thus, under conditions where attack by, ! °f a,re‘atlVely unreactive aromatic compound proceeds at a significant, ate attack by NO* may still be very slow, and attack by N,0< (which is even, , T', , Wi“ ^ S'OWer, Stm', Wh“, tack becomes the only important, path for, nitration., , k, , nitrol:, , In addition to the two types of mechanism that we have considered, may. jrOCeed in leSS USUaI Ways' Nations with nitrogen, i, , g es, and Ingold, J. Chem. Soc., 1952, 28
Page 444 :
428, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , pentoxide, acetyl nitrate (N02OAc), and benzoyl nitrate (N020C0Ph) in non¬, polar solvents have been described,30<o) and there is some evidence which indi¬, cates (although it does not prove) that the H2NO£ cation is a nitrating agent in, HNO3—HCIO4—H20 mixtures.ss(6), , Reactivity and Orientation in Electrophilic Substitution, During our discussion of nitration, we have described some aromatics as being, “reactive” and others as “less reactive,” referring simply to the rates at which, they are attacked by a given electrophilic species (in this case NO^ or NO+)., It might seem that we could arrange a group of aromatic compounds in the, order of their reactivities merely by measuring their nitration rates under, similar conditions, but this will not work for compounds undergoing pseudo¬, zero-order nitrations; for here the observed rate of nitration is the rate of forma¬, tion of the nitronium ion. For these, competition experiments (p. 172) prove, useful. Two (or more) substrates in a single solution compete for the NO£ ion, as it is formed, and from the ratio of the various nitration products, the relative, (but not the absolute) specific rates of attack by the nitronium ion may be, calculated.30 The results of a few of such comparisons are listed.3/(o), PhN02 PhCOOEt PhBr PhCl PhF PhCH2Cl PhH PhCH3 PhOH, , k, , <icr4, , 0.0037, , 0.03, , 0.03, , 0.15, , 0.30, , 1.0, , 25, , 1000, , 6 6, , ^c h, , These values need not be considered at length here, for, qualitatively, speaking, they reflect combinations of the inductive and resonance effects of, substituents as discussed in Chapter 7. The activation process involves the, attachment of the positive NOf ion to an “electron-rich” site in the ring. There¬, fore, nitration is facilitated by electron-releasing substituents such as alkyl, groups but retarded by electron-attracting substituents (for example, —NO2, and _COOEt). Similar statements may be made for other types of electro¬, philic-substitution reactions. When the / and R effects of a substituent are m, opposite directions (as in the case with the halogens and the -OH,, , NR* an, , „ (a) Gold, 't d„ J. Cbm. See., 1950,2452. (4) Halberstadt, Hughes, and Ingold, ibid., 1950,, , 244'» The applicability of ^"^d^tpec.t, competing substrates are attac, , apparent rate laws for the overall reactions be the, , :trHoweVet"raigMfoqrward interpretation of the results is possible only if there is a stngle, , 918; 1948, 575. No more than two significant, , gurcs a, , tions were not the same in all competition expenmen ^0, and empirical) of relative reactivities of polycychc in >, , treatment (both theoretical, l, , drocarbo„s in nitration, see, Y rough) rule of thumb is that
Page 445 :
Reactivity and Orientation in Electrophilic Substitution, , 429, , -SR groups) we cannot predict, in the absence of further information, whether, that substituent will activate or deactivate the ring, but Hammett’s a value for, the substituent will generally tell the tale. Thus the halogen atoms, which have, positive a values, are deactivating, whereas the, , OH and, , NR2 substituents,, , which have negative a values, are activating., An important and related question is that of orientation in aromatic substitu¬, tion, for an aromatic ring may generally be attacked in two or more different, and nonequivalent positions. As with reactivity, we shall base our discussion, of orientation on nitration reactions, bearing in mind that the arguments we, use may be applied equally well to other types of electrophilic substitutions., Primarily, we are interested in explaining why some substituents (among them, the —OH, —OR, and —NR2 groups, and the alkyl and halo substituents), will direct incoming electrophiles predominantly to the ortho and para positions,, whereas other substituents, , (among them the —N02, —CN, and —COOEt, , groups) direct such electrophiles to the meta position. This is a problem involving, relative rates, for, in actuality, a given electrophile will attack all available, positions of an aromatic ring, but reactions at the “favored” sites are more, rapid than those at the less active positions. Of the activated complexes resulting, from attack at each of the available positions, the complex having the lowest energy, is that associated with the “favored ” position for attack. As we have seen, the activated, , complex in aromatic nitration is intermediate in character between the sub¬, strate and the intermediate cation XIV; we may then represent this complex, as XIII. However, to avoid the inconvenience of working with structures having, , pamal bonds, we shall discuss the orientation problem in terms of the energie<, of the various possible cation intermediates. This procedure is not entirely free, from objection/-™ bu, the conclusions reached will not be significantly difren, from those that would be based upon consideration of the activated, complexes themselves., Our first example is anisole. Representative structures for each of the
Page 446 :
430, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , possible cation intermediates are shown as follows:, , PhOMe + NO2, , Two additional structures, of importance comparable to those shown, may be, drawn for each intermediate, but these do not affect the argument. Of the, structures above, XV (for ortho attack) and XVI (for para attack) are of greatest, interest in the present discussion. Try as we may, we cannot draw a structure, analogous to either of these two for meta attack, this difficulty being a conse¬, quence of the “alternating character” of the atoms comprising a conjugated, system. (As we have previously noted, this “alternating character, , is readily, , described, although not in any sense “explained,” using the language of reso¬, nance.) In the intermediate for meta attack, therefore, the positive charge carried, in by the NO+ ion must remain on the benzene ring until the proton departs,, completing the substitution reaction. On the other hand, with ortho, , 01, , para, , attack, Tr-electron density drifts from the electron-rich oxygen atom into the, ring, dispersing the positive charge, and lowering the energy both of the inter¬, mediate and the transition state leading to it. We may thus conclude t at in, the nitration of anisole, attack should take place more readily at the ortho and, para positions than at the meta position, although we cannot say, on the basis o, , this simplified picture, whether it is the ortho or the para position that is the mor, reactive. Analogous arguments may be used to, by halogen atoms and by the -OH, -OCR, and, , O, of these substituents, the atom, , NHCR groups. In, , 0, , ., , bound to the ring has at least one unshared and
Page 447 :
Reactivity and Orientation in Electrophilic Substitution, , -, , 431, , available pair of electrons, and in each case electron density from this pair drifts, , into the ring to help disperse the positive charge carried in by the electrophile, attacking either at the ortho or para position. Ortho-para direction by alkyl sub¬, stituents may be explained in a similar way. Here, it may be argued, the inter¬, mediate cation (and the activated complex leading to it) may be stabilized by, such hyperconjugated structures as XVII' and XVIII', which, again, are not, possible for intermediates in meta attack., H+, , H+, , There are a few, less usual, oxygenated substituents, among them the, , O, , I, }~° and, , ^, , OH groups, which constitute special cases. Although the, , O, atoms by which these groups are attached to the ring have unshared electron, pairs, they also bear positive formal charges and tend therefore to withdraw, e ectron density from, rather than supply it to, the benzene ring. In this respect,, , contribute to the cationic intermediate. However the ratio of orthn 4- IP rC°"JU5atlon) may, tack in the nitration of /-butyl benzene is about’ 10 (CnZ , ft ^ a“aCk t0 meta at’, whereas in the nitrations of ethylbenzene and isooroov ben ’, 1”, 169’ 291 (1952»>, greater (although the difficulty^ analyzing fnrP, ZCne’, corresPonding ratio is far, presence of lar|e quantities of the or/t and, °f the, isomers in the, ratios in these cases). I, appears then that'^, precise CTal“«i°» of the, although not negligible, is significantly less importan/than C In ° ° h''PcrconjuSation,, The question of hvperconiuoJinn, k, tha"C—H hyperconjugation., co-workers (J. Am. Chem. Soc., 79, 505 (1957))C ffiaTci1Hen1 ^, by Swain and, the same specific rate as “ordinary” toluene Tf w, ° tCD3 1S mtrated at almost exactly, (P. 285) tha, the importance of C-V“T ? °Ur °rigi”al ^omption, deuterium for hydrogen, we would have to conclude hTt hY ^ deCreased by substitution of, tor in stabilizing the transition state in the ortho nr /,, lyperconJUgation is a negligible fac-, , IfZ SV, , -er, tha, this assumption was too brold a„d th^, conjugation is more subtle than we at present suppose., , "U, , p0ssi“e, how., P'C substituti°n on hyper-
Page 448 :
432, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , they resemble the nitro group, and like the latter, they direct incoming elec¬, trophiles predominantly to the meta position., The —NH2 and —NR2 groups are special cases of a different sort. Although, they themselves are strongly ortho-para directing, they are converted, in the, acidic media necessary for nitrations, to the —NH£ and —NR2H+ groups,, which are just as strongly meta directing. We may account for the mi?ta-directing, character of these positively charged substituents (and of such substituents as, —SR£, —PR+, and —SeR£ as well) by considering, as we did for the nitration, of anisole, representative structures for the possible nitration intermediates,, which, for this group of compounds, must be doubly charged., , Two additional structures, comparable in importance to that shown, may be, drawn for the intermediate in meta attack, and one additional structure may, be drawn for the intermediates in both ortho and para attack but "one of these, affects the argument. Directing our attention to structures XX and » I, see that in both of these, the positive chargee are s,mated on adjacent atoms _Th, feature lowers the stabilities of these structures; for, electrostatically speaking,, a polyatomic dipositive ion is most stable when the positively charge, , atoms, , as far removed from each other as possible. In the language of resonance d*, contributions of structures XX and XXI to their, ", must be small; this is equivalent to saying that in ortho and/™.atack,, tive charge carried in by the NO+ group is not distr.but, ing but tends to be concentrated over a small area °PP“‘ e *, , osi., sub., , fH, , stituent. In contrast, however, none of the three structures contributing
Page 449 :
Reactivity and Orientation in Electrophilic Substitution, , 433, , intermediate for meta attack is analogous to XX or XXI; none has positive, charges located on adjacent atoms. In this intermediate, therefore, the incoming, positive charge may spread itself far more evenly about the ring than is possible, for ortho or para attack and, as a consequence, this intermediate (and the acti¬, vated complex leading to it) is the most stable of the three.33, By a similar argument, any substituent attached to the benzene ring by an, atom bearing a positive charge should likewise be meta directing. Even if the, positively charged atom is one position removed from the benzene ring, the, effect of the pole may be transmitted inductively to the junction between ring, and side chain, causing meta orientation. Thus, the —CH2NMe{ and the, —CHoSMe^ substituents are meta directing (although —CH2—CH2—NMc£, and —CH2—PMe| are not34). We may use very similar reasoning to account, for the 7tf<?ta-directing action of the nitro group, for the nitrogen atom in this, , o, group bears a positive formal charge, , I that is, —N, , /•, , and therefore, , 0_/, structures XXII and XXIII, for the intermediates in ortho and para nitration, of nitrobenzene, are open to the same objection as structures XX and XXL, This question may be viewed in a somewhat different light, this time considering, , XXII, , XXIII, , XXIV, , XXIV', , XXIV", , the electron distribution in the substrate itself. We have noted on a number of, occasions that the nitro group withdraws electrons from the benzene ring, and, ., , ..**An aher"atl^e explanation for the rorto-directing character of the —NH+ group and, , icast deactivated, position, , “~N-^^5 +, , deficiency due purely^othffndictive^ffect be^t T explanation requires that an electron, in much the same way as, ^ the rin^ to Ornate carbons, «« Baker and Muffin rn, ^hciencies due to resonance effects., 1280, 1930> 1722; Ingold, Shaw, and Wilson, ibid., 1928,
Page 450 :
434, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , as is shown by structures XXIV' and XXIV", this electron withdrawal is, most marked at the ortho and para positions. Since an electrophile tends to seek, out the site of highest electron density, it will tend to attack preferentially at, the meta position, where electron density has been depleted least, rather than, at the ortho or para positions. Completely analogous considerations apply to, the —COOH, —COOEt, —CHO, —C=0, and —C=N groups, all of which, R, are meta directing in electrophilic substitutions.55, We have chosen to discuss orientation in aromatic substitution in terms of, the manner in which substituents affect, either by induction or resonance, the, distribution of electron density in the substrate and in the activated complex., We have ignored kinetic-energy effects, implying either that they are negligi¬, ble or that they parallel the electronic (potential-energy) effects described., These are the same conditions that we proposed earlier in justifying the success, of the Hammett equation (p. 220), which stipulates that the comparative ability, of a given substituent to attract electrons from, or supply them to, a reaction, site on a side chain may be expressed by a single parameter, independent of, the nature of the reaction occurring at the side chain. It now may be asked, whether a similar statement holds true if the reaction occurs, not on the side, chain, but rather on the ring itself. If the answer is “yes,” and if, as we have, implied, kinetic-energy effects quantitatively parallel potential energy effects, (or are negligible in comparison to them), a Hammett-like equation should, apply to aromatic substitution reactions also. A recent compilation by Brown, and co-workers55 indicates that the Hammett treatment, with but slight modiu The predominant position of substitution in a number of aromatic molecules (both, benzenoid and nonbenzenoid) may be predicted by the following rule: Of the various inter¬, mediates resulting from attack at each of the possible sites, the intermediate corresponding to, the favored site of attack will be that having the largest number of acceptable canonical forms. Al¬, though this rule is based on oversimplified reasoning, and although the inclusion of the word, “acceptable” introduces a certain element of arbitrariness, the rule is surprisingly usefu^, We have for example, listed three contributing structures for the cationic intermediate in, meta nitration of anisole, but four structures each for the totermri, tion. Ignoring structures of higher energy, such as XXV, the lule, as stated, predicts, , XXV, para nitration. In the nitration of nitrobenzene, however the, , positions of nitration in naphthalene and thtophene (hx, » Brown and Okamoto, J. Am. Chem. Soc., 79,, , 1., , McGary Okamoto, and
Page 451 :
Reactivity and. Orientation in Electrophilic Substitution, , 435, , fication, may indeed be used to correlate the rates of electrophilic substitution, on the benzene ring. The Hammett-Brown equation,, , log (£) = pa+, , (3), , is very nearly the same as the Hammett equation itself, and the symbols have, similar significance. The k values, however, refer to the specific rate of substitu¬, tion at a single site; thus for use in this equation, the observed specific rate for a, reaction of benzene must be divided by 6 (since there are 6 equivalent positions, in benzene) and, for the same reason, the rate constant for meta attack on a, mono-substituted benzene derivative must be divided by 2. The notation a+, indicates that the attacking reagent is cationic. Equation (3) is applicable to, meta and para substitution, but not generally to ortho substitution, suggesting, that in ring substitution, as in side-chain reactions, ortho substituents often, introduce steric effects which are unrelated to their polar character. The <r+, values chosen by Brown for meta substituents are virtually the same as Hammett’s, a constants, but this is not true for para substituents. Below are compared the, <r+ and cr constants of some typical substituents (for the para position):, —OCH3, , —ch3, , -C(CH3)3, , —F, , —Cl, , —Br, , —I, , —NO, , <, , -0.76, , -0.31, , -0.25, , -0.07, , +0.11, , 0.15, , 0.13, , 0.78, , <jv, , —0.27, , -0.17, , -0.20, , +0.06, , 0.23, , 0.23, , 0.28, , 0.78, , Aside from the nitro group (the only “-R” substituent for which a reliable, value of a+ is at present available), all <r+ constants appear to be significantly, more negative than the respective ap constants. Since a highly negative value of, a is, by the Hammett convention, associated with effective donation of electron, density from the substituent, we may infer that, compared to meta substituents,, -\-R, , para substituents are more powerful electron donors during substitution, , on the benzene ring than during reactions on side chains. This is obviously, because a substituent is in direct conjugation with a ring carbon in the para position, (XXVI) but not, except under special circumstances (p. 225), with a side, chain attached to this carbon (XXVII):, , Y, XXVI, , XXVII, , the Ha'mmett-Brown^treatment^of5aromatiTsubsdtu5° ^ constants) support, original Hammett equation to side-chain reaction^ove^Oi^r^e anti equiUbrium^constants)^
Page 452 :
436, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , For aromatic substitutions to which the Hammett-Brown treatment applies, P, , values (which mfeasure the sensitivity of the various reaction series to changes, , in the polarity of the substituent) lie between — 4 and —10. These are much, greater than the p values for the vast majority of side-chain reactions, in which, one or more atoms may “shield” the reaction site from electrical disturbances in, the ring. Typically, the replacement of a —N02 group by a —OCH3 group, boosts the rate of ring bromination by a factor of about 1015, whereas a similar, replacement in /^-substituted ethyl benzoates lowers the rate of saponification by, a factor of only 100., There is no evident reason why this treatment should be confined in its, use to substitution reactions in which the departing group is H+. Indeed, the, rate constants for the acid cleavage of substituted phenyltrimethylsilanes, (equation 4)37 and those for the brominolysis of substituted benzeneboronic, acids (equation 5)ss may be correlated satisfactorily by the Hammett-Brown, equation., ArSiMe, + H+, , -"C'°~ „ . ArH + SiMe+[-^ (Me,Si)20], MeOH H-H2O, , p50o = -4.59, , ArB(OH)2 + Br+--> ArBr + B(OH)+[-^ B(OH),], H’°, , (4), , , PA, , (5), , to- = -4.30, , and it is likely that this correlation will be found to extend to additional reaction, series., , The ortho to para Ratio, Much of our discussion up until now suggests that the availability of electron, density is the same at the position para to a given substituent as at the positions, ortho to this substituent. If this were so, and if other effects were absent, we should, expect the ratio of ortho to para isomers resulting from all electrophilic aromatic, substitutions on mono-substituted benzenes to be 2.00 since there are two posn, tions ortho, but only one para, to any single substituent. In actuality this ra, is practically never observed, for there are good reasons why it should notFirst and most obvious, the para carbon is in a more favorable position for auac, than is the carbon ortho to the substituent, for the substituent acts as a h, which blocks a portion of the volume through which the attacking gro P, , «, , otherwise pass. This s.eric effect, which should lower the ortho to pareH, detectable for small substituents and becomes quite importan in the eve, substituent is bulky. The activated complex for the ortho nitration of e.hy, : KoMla,, , 5068 (1952); 77, 4838 (1955).
Page 453 :
The ortho to para Ratio, , -, , 437, , benzene is sketched in Figure 11-3 (the departing hydrogen is partially hidden, , Fig. 11-3. Steric Hindrance in ortho Nitration of Ethylbenzene, , by the attacking nitrogen atom). In the conformation shown, there is interference, between one of the hydrogens on the /9-carbon and the attacking NO^ ion. This, interference may be eliminated by rotation of the ethyl group about the bond, that links the a-carbon to the ring, but when we limit the number of possible, conformations in which the substrate may react, we lower the probability of, reaction. As expected, ortho substitution becomes still more difficult with further, branching at Ca. The following ratios then speak for themselves:55, , o:p ratio in nitration, , PhCH3, , PhCH2CH3, , PhCH(CH3)2, , PhC(CH3)3, , 1.57, , 0.93, , 0.48, , 0.22, , Similarly, we should expect the ortho to para ratio to be less if the attacking group, is bulky than if it is small. Typically, the chlorination of chlorobenzene (in which, the attacking reagent is probably Cl+) yields a mixture of dichlorobenzenes in, which the ortho to para ratio is about 0.7. Its nitration (in which the attacking, reagent is the more bulky NO^ ion) yields an ortho to para ratio of about 0.4., Finally, the ortho to para ratio resulting from the sulfonation of the same sub¬, strate (here the attacking reagent is probably the still more bulky S03 molecule), is less than 0.01. Similar trends have been observed in the substitution reactions, of such aromatics as toluene, bromobenzene, and phenol.40, Sometimes we must reckon with the inductive effect. It has been empha¬, sized that the intensity of this effect decreases sharply with distance from the, “primary pole” (in contrast to resonance effects, which are transmitted mainly, to alternate atoms in a conjugated system with no significant decrease in, intensity with distance). Thus, irrespective of resonance effects, the electron, density at the positions ortho to an electron-attracting group must be less than, hat, , to this group. The difference, which depends upon the magnitude of, inductive effect of the group, should be reflected in a preference by the, ‘Brown and Bonner, J. Am. Chem. Soc., 76, 605 (1954), , (1925), , 5 C°mpilal,on of data rela,inS <° <hb question, see Holleman, Chem. Revs., 1, 218
Page 454 :
438, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , attacking electrophile for the para position. Thus, the ortho to para ratio resulting, from the nitration of fluorobenzene is 0.14; that resulting from the nitration of, chloiobenzene (in which the directing group is less electronegative) is 0.43, For nitrations of bromo- and iodobenzene, the ratios are 0.61 and 0.70, respec¬, tively.^0-^ This trend is clearly in line with the inductive effects of the respective, halo substituents and would probably be even more pronounced if stcric effects, did not operate in the opposite direction. A similar trend appears in the series, below,'S9’^ in which the steric and inductive effects presumably reinforce each, other:, , o:p ratio in nitration, , PhCH3, 1.57, , PhCH2Cl, 0.75, , PhCHCb, 0.53, , PhCCl3, 0.23, , By an extension of this reasoning, we would expect the ortho to para ratio in, the nitration of such compounds as benzaldehyde, benzoyl chloride, and nitro¬, benzene to be very low; for, aside from their resonance effects, the —CHO,, —COC1, and —N02 substituents have strongly negative inductive effects, and, the steric requirements of each of these groups are by no means negligible. But, this is not the case at all, for the ortho to para ratios from these compounds are, found to be large—about 2 for the nitration of benzaldehyde, about 10 for ethyl, benzoate and benzonitrile, and about 20 for nitrobenzene and benzoic acid.-*0--*3, Quite obviously, one or more additional effects which we have hitherto over¬, looked are activating the ortho position or deactivating the para position (or, both). It has been suggested that such groups as —CHO, —COOH, and, —N02 facilitate nitration at the ortho position (in comparison to that at the, para position) by dipole-dipole interaction with the incoming nitronium ion, (for example, XXVIII and XXIX).^ However, this effect cannot be solely, N, III, , XXIX, responsible for the high ortho to para ratios in these nitrations, for a similarly, high ratio is obtained from the nitration of benzonitrile. Here because of the, 4i Sandin and Williams, J. Am. Chem. Soc.,, , 69, 2747 (1947)., , It is possible,, , by carrying out, , the nitration in concentrated HN03 at 0°, to decrease the ortho to.para ratio rating^om, nitration of iodobenzene to 0.52, a value less than that for the nitration of bromobenze, , f0.59), , under the same conditions. The reason for this reversal of order is not clear., , 4* Holleman, et at., Rec. trav. chim.., 33, 1, , (1914);, , Ingold, et at., J. Chem. Soc., 1931,195V,, , ^^ker, ,/ at., J. Chem. Soc., 1927, 836; 1928, 436; 1931, 314. The figures quoted are ap¬, proximate, for the o :p ratios in these cases are rather sensitive, 44 Hammond, Modic, and Hedges, J. Am. Chem. Soc., 75, 1388 (1953).
Page 455 :
The ortho to para ratio, , -, , 439, , inearity of the —C=N group, the positive end of the N02 dipole is far removed, rom the negative end of the C=N dipole, and dipole-dipole intei action, inalogous to the type indicated in XXVIII and XXIX would be negligible,, ndeed, the lack of independent evidence that such interaction exists, even in, :favorable cases” suggests that the high o:p ratios in the nitrations which we are, :onsidering are best explained in a different manner.'*5, Let us compare, for example, the canonical forms of nitrobenzene, XXIV', tnd XXIV", which depict, in the language of resonance, the depletion of elecron density at the ortho and para positions. The question as to which of these, , orms is the more important is a quantum-mechanical one and cannot be, tnswered reliably by inspection. It appears, however, that XXIV" (the so¬, iled “/wa-quinoid” structure) has a significantly lower energy than XXIV', the so-called “orMo-quinoid” structure), although this difference seems never, o have been explained satisfactorily in nonmathematical language.-*5 On this, )asis, electron density is more effectively depleted at the para position than at, he ortho, and ortho substitution should predominate over para (with meta subtitution predominating over both ortho and para). Similar reasoning may be, ipplied to the nitrations of benzoic acid, acetophenone, and benzonitrile., Extending the argument to such activated aromatics as anisole, we may, econsider structures XV and XVI. Both depict the stabilization of the cationic, MeO*, , lUr'NO*, , +OMe, , X, yj, , H, XV', , r, , -, , NO, , XVI, , isr, , resuit, A, , ^, , °r, , —, , ron structure of the quinone itself and the «mil? ? a number of factors other than the elecortuitous., 4, nC ltSClt’, the similarity between the two effects may be merely
Page 456 :
440, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , intermediate in nitration by the drift of negative charge from the methoxide, oxygen to the ring. Here, the intermediate for ortho nitration (XV) is of the, 01, , Mo-quinoid type, whereas that for para nitration (XVI) is of the jfrara-quinoid, , type. If the />ara-quinoid structure is again considered the more stable, we may, infer that just as “ — R” groups deactivate the para position more than the ortho,, groups activate the para position more than the ortho. But the direction, , of this effect is the same as that of the steric effect, and often the same as that of, the inductive effect, the three effects are difficult to separate. In a typically, perplexing case, the ortho to para ratio resulting from the nitration of phenol is, , close to %.40 The difference between this value and the “statistical value” of, 2.0 almost certainly cannot be attributed solely to the steric effect, for the steric, requirements of the —OH group are very modest. However, we cannot say at, present just how much the ortho to para ratio is lowered by the inductive effect, of the —OH group, and how much it is lowered by the tendency of the activated, complex to approximate a />ar<2-quinoid (rather than an or/Ao-quinoid) structure., , Halogenations, In analogy with the picture of aromatic nitration which has been presented, we, would expect the most effective species in aromatic halogenations to be the, electron-deficient halogenonium ions, :C1:+, :Br:+, and :I:+. However, unlike, the nitronium ion, these ions have not been characterized in salts, nor have, they been detected in solution by physical methods.*7 Nevertheless, there is, strong kinetic evidence that such cations participate in many (but certainly not, all) aromatic halogenations., Consider, for example, the chlorination of aromatic compounds in acidic, solutions of HOC1 (with Ag+ added to remove chloride, which, when present,, introduces mechanistic complications). As with nitrations in nitromethane, the, rate law depends upon the reactivity of the substrate.** With relatively un-, , 47, , The appreciable conductivity of liquid iodine (Rabinovitsch, Z. pkysik. Chem., 119, 82, , (1926)), , is sometimes taken as evidence for its ionization into 1+ and I, , xh, but this conductiv¬, , ity appears to be metallic, rather than ionic, in character since it decreases with increasi g, P, , Salts of substituted halogenonium ions such as XXXI, XXXII, and XXXIII are known, , electron deficient in the same sense as are the simple: na.cjj«.*, •, 4« (a) de la Mare, Hughes, and Vernon, Research, 3,192, 242, jwj. (oj o, J. Aw. Chem. Soc., 77, 3410 (1955)., , and Ket,cy>
Page 457 :
Halogenations, , -, , 441, , reactive (but water soluble) aromatics such as nitrophenols and anisolesulfonic, acids, chlorinations are third order,, rate - A(ArH)(HOCl)(H+), , (6), , but with more reactive substrates (for example, phenol and anisole) the reactions, become very nearly zero order in substrate—that is, second order overall., rate = Jfc(HOCl)(H+), , (7), , The observed duality in kinetics suggests that the steps in chlorination are, analogous to those comprising the usual nitration mechanism. For aromatics, which are chlorinated in accordance with rate law (6), the activated complex, in the rate-determining step in which the ring is attacked consists of a molecule, H, + /, of substrate and either a Cl—O, , ion or a Cl+ ion. Since the interconversion, , \, H, between the two cations is probably far more rapid than the attack of either, ion on the aromatic ring, kinetics alone cannot direct a choice. Suppose, how¬, ever, that a much more active substrate is used. The specific rate at which the, ring is attacked becomes much greater, and the attacking species may be con¬, sumed very nearly as rapidly as it is generated. The rate of the overall reaction, then becomes essentially equal to the rate at which the attacking reagent,, C10H+ or C1+, is formed from HOC1 and H+, the chlorination proceeds ac¬, cording to rate law (7), and we may expect to find a number of active aromatic, substrates which are chlorinated at the same rate in a given solution., The fact that a number of aromatic chlorinations are observed to follow, rate law (7) strongly suggests that the attacking reagent in such cases is C1+, rather than C10H+. For it is very unlikely that the formation of C10H+ from, “d H30+ (a proton transfer between two oxygen atoms) could be slow, enough to be the rate-determining step in a chlorination reaction. Furthermore, , InDo'tha, , f°Il0W ratC'laW (7) pr°Ceed more rapidfy, , Proton tranf ^, m, ^, evidence aSainst a rate-determining, sfer, for D transfers are known almost invariably to be slower than, , z;;• *—»■ ■"», * *, bond L aO^ tI, , S:, , **P is the breakage of the Cl-O, , summarizedmecha, T T, mtrati°n “ VirtUal'Y ComPletei, we may, the mechanism for chlorination in acidified HOC1 as follows:, , H, T_ _, , +, , H+, fast, , HOC1 ^==1 C10H+, (1>, , C1+ ArH>, (2), , (3), , — H+, fast, , Ar, CL, XXXIV
Page 458 :
442, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , As with nitration, the apparent rate law will depend upon how step (3) and the, reversal of step (2) compare in rate.49 The intervention of the cationic inter¬, mediate XXXIV could be demonstrated by showing the absence of a kinetic, hydrogen isotope effect in step (3)—the attack on the ring. This has not as yet, been done for aromatic chlorination, but experiments similar to those described, for nitration (p. 423) have shown that the hydrogen-isotope effect in aromatic, bromination is negligible., Kinetic studies of aromatic bromination in acidified solutions of HOBr50, indicate that here also the attacking reagent is a halogenonium ion (in this case, :Br:+). However, we should remind ourselves that the halogenation conditions, which we have been discussing are not the usual ones used in preparations, for, aromatic brominations and chlorinations are ordinarily carried out using ele¬, mental bromine or chlorine, often in the presence of a Lewis acid such as zinc, chloride, aluminum chloride, or ferric chloride. It has been suggested by some, authors51 that the function of such an acid is the heterolysis of the halogen, molecule (for example, Br2 + ZnCL ^ Br+ + ZnCLBr-), prior to the attack, on the substrate by the halogenonium ion. However, there seems little, if any,, reliable evidence that such a heterolysis takes place at ordinary temperatures,, and it is far more likely that the attacking species is the diatomic halogen molecule, itself. If this is the case, the Lewis acid must await the formation of a complex, between substrate and halogen molecule before assisting in the breakage of the, bond between the halogen atoms. At present, then, the most likely mechanism, for such halogenations appears to be, H‘ +, — H+, -^2, , ArZ, , Ar, , ZnCh, AlBr3, etc., , fast, , ArH ^ ArH-A^-*, fast, , slow, , ZJ, ZnCl2Z- AlBr3Z- etc., , in which the rate-determining step is a simultaneous nucleophilic “push” by, the substrate and an electrophilic “pull” by the catalyst on opposite ends of the, halogen molecule with heterolytic breakage of the X—X bond. Recent kinetic, studies5* of such halogenations are consistent with this mechanism, for there, The higher chlorination rate in D20 may be explained in terms of the ^hi^ eqmhbn O, a weaker base than H20 (Wynne-Jones, Trans. Faraday Soc., 32, 13)1, (19?6)f a greater concentration of the conjugate acid of hypochlorous add mayrna, given acidified solution of HOC1 in heavy water than in the correspondmg solutton m ordtnary, Wate<® See, for example, (a) Wilson and Soper J, , 1949, M7*i (»> Derbyshire and, Wiley and Sons, Inc, New, , YOrk<'. wVh^been found, for example, , 4549, , (1956)), that brominations of a number of aromat, , y
Page 459 :
Halogenations, , -, , 443, , appear to be no cases where a halogenation of this sort is zero order in substrate, (which would be a symptom of the preliminary conversion of X2 to a more, reactive form)., It is not necessary that the role of electrophilic catalyst in halogenations, be assumed by a metallic halide. Interhalogens, such as ICl5S(o) and IBr5*<6>,, and iodine itself5S(6) may act in a similar capacity, and the same is true for HC1,, carboxylic acids, and even water.55(o) However, the pulling action of water is, feeble and becomes effective only for the more active substrates. The participa¬, tion by water molecules in brominations carried out in bromine water cannot, be demonstrated kinetically, but it has been shown that brominations in glacial, acetic acid may be strikingly accelerated by small amounts of added water.54, It should be clear that a halogenonium ion, X+, when it can be generated,, is a more powerful halogenating agent than the corresponding halogen mole¬, cule, X2, since halogenation with the latter requires an extra measure of activa¬, tion energy to break the X—X bond. Since —OH is known to be a far poorer, “leaving group” than the less basic halide ions (p. 261), we should expect the, hypohalous acids, X—OH, to be even less powerful halogenating agents than, the respective halogen molecules, X—X. Indeed, it has been found that the, hypohalous acids, unless converted to their own conjugate acids, X—OHt,, will halogenate, at appreciable rates, only the most electron-rich aromatics., The phenoxide ion, for example, is chlorinated at a readily measurable rate, with HOG1, the reaction exhibiting simple second-order kinetics.55, C6H50, , -f- HOC1 —> C1C6H40 (3 isomers) T HaO, rate = £2(PhO-)(HOCl), , (8), , This reaction is accelerated strikingly by addition of HC1, for this acid reacts, with HOC1 to form Cl2, which, we have remarked, is a more effective chlorinat¬, ing agent., HOC1 + HC1 ;=± CI2 + H20, Elemental iodine is a less powerful halogenating agent than are chlorine, by Znd2, are first order each in substrate, in bromine, and in ZnCl2. Sec also (b) Yeddanaoalli, “S°C•’1956’ 49M’ and « TsU™ta - <, , cLL Soc, 74, 5P996, , " (a) Keefer and Andrews, J. Am. Chem. Soc., 78, 5623 (1956) • 79 5169 M9571, , •>rtecSiS; S, , (k\ p •, , T Hhile one <OT more), , since this interhalogen is appreciably* dissociated intoTa^T Wi'.h ‘Br ar', above; this compound acts primarily as a bromimf, 2, d Br2 3t r°°m temPerature and, , iodinating agent., , P"™nly as a brommatmg agent, not, as one might suspect, an, , ** *5eefer and Andrews, J. Am. Chem. Soc., 78, 3637 (’19561, Soper and Smith, J. Chem. Soc., 1926, 1582.
Page 460 :
444, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , and bromine, for the metal halides usually used as halogenation catalysts co¬, ordinate only weakly with iodine, and their “pulling power” is of little avail., Aromatic iodinations in organic solvents are often carried out instead with, iodine, , monochloride, I Cl,™-53 although iodinations with iodine itself are, , possible in the presence of AgC10457(o) since univalent silver is unusually elec¬, trophilic toward iodide.™ Elemental iodine has also been used for iodinations., of phenols and aromatic amines in aqueous media, but for these reactions there, is strong evidence that the attacking reagent is either 1+ or IOH£, rather than, 12 or Ij. The iodination of aniline in the presence of base and excess iodide, was discussed in an earlier chapter (p. 176), where it was shown that a rate law, , rate _ MglXPhNH^), , (9), , rate = A'(PhNH2) (B:)(!+), , (10), , could be interpreted as, , (For each additional base, B:, present in solution, an additional term may be, added.) Rate law (10) immediately suggests an activated complex consisting, of a molecule of base, a molecule of aniline, and an 1+ (or IOH£) ion, but there, are two additional interpretations possible, based upon alternate forms of the, ✓, , rate law, as follows:, , and, , ', , rate = *"(PhNH+)(£:)(HOI), , (11), , rate = *'"(PhNH2) (B: H+) (HOI), , (12), , The transition state suggested by expression (11), which implies attack upon, the anilinium ion, may be excluded on chemical grounds; for the — NHJ, group is strongly meta directing, whereas iodination of aniline yields mainly, the />-iodo compound. Expression (12) is not so easily disposed of, for it implies, a mechanism similar to that which we have proposed for halogenations in, organic solvents:, fast, , PhNH2 + HOI, +HOH +B:, , while expression (10) suggests the following mechanism:, , n:, H.N H0 + V^, , H,N, , eq, , slow, , '*NH0-, , XXXV, ..Lambourne and Robertson. J. Chm. So,, 1947, 1167; Andrews and Keefer, J., , Am., , Chem. Soc., 79, 1412 (1957)., f'1933) • 67B, 917 (1934). (t) Craig, et al.,, 57 (a) Birkenbach and Goubeau, Ber., 66B, 128U (iJjo), 0/, v, J. Chem. Soc., 1954, 332.
Page 461 :
Sulfonation, , 445, , To distinguish between these two mechanisms, we may note that the rate¬, determining step in sequence (14) involves breakage of a C—H bond in the, ring, whereas that in (13) does not. Thus, if (13) is correct, iodination of ringdeuterated anilines should proceed at very nearly thfe same rate as that of, ordinary aniline, whereas if (14) is correct there should probably be a kinetic, hydrogen-isotope effect. Although this test has apparently not yet been applied, to aniline, it has been applied to the iodination of phenol in buffered aqueous, media, for which the kinetic picture is very similar to that for aniline.55 It has, thus been found55 that 2,4,6-trideuterophenol is iodinated at only one fourth, the rate of unlabeled phenol, strongly suggesting a mechanism analogous to, sequence (14) for the iodination of phenol, and, by implication, mechanism, (14) itself for the iodination of aniline. It would be interesting to see whether, the hydrogen-isotope effect in iodinations persists in nonpolar solvents where the, existence of intermediate XXXV is more precarious., Virtually nothing of what has been said regarding chlorination, bromination, or iodination may also be applied to fluorination; for it is certain now, that the reactions of elemental fluorine with aromatic compounds, except, perhaps at very low temperatures, proceed through free fluorine atoms, formed in, a preliminary homolytic cleavage of F2 molecules.50 These reactions are addi¬, tions, rather than substitutions, and lead to fluorinated cyclohexane derivatives., The very low F—F bond energy (about 37 kcal per mole) favors conversion of, F2 into atoms; whereas the very high ionization potential of the fluorine atom, (about 100 kcal per mole greater than for chlorine) renders the formation of, the :F:+ ion in chemical systems impossible. Furthermore, fluorinations with, hypofluorous acid, HOF, have not been carried out, since neither this acid, nor its salts have been prepared., , Sulfonation, The intimate details of aromatic sulfonation reactions are, at present, known, with much less certainty than those for nitration, chlorination, and bromination, Ihere is little doubt, however, that sulfonation, like halogenation and nitration, may proceed by more than one mechanism, depending chiefly on the medium, t';7'mOSt hk,dy attackin§ reagent for sulfonations carried out in concenmolerT UnC ^ ('C°ntaining sma11 amounts of water) appears to be the S03, e or its solvate,” H2S207 (that is, S03-H2S04). This guess is based, argely on studies of the effect of added water on aromatic sulfonations, which, **®erhner’ J■ Am■ c/le™- Soc.t 73, 4307 (1951), Grovemtei” and Kilby, J. Am. Chem. Sac., 79,, , 2792 (1957)., , Chem. Revs., 40, 51 (1947yCaCU°nS of dcmental fluorine with organic compounds, see Bigelow,
Page 462 :
446, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , indicate that sulfonation rates are inversely proportional to the square of the, concentrations of added water.61 Suppose that the added water reacts very nearly, completely with the sulfuric acid that is present in excess., H20 + H2S04 -> HsO+ + HSO7, Neglecting autoprotolysis of sulfuric acid, the concentrations of both the H30+, and HSOj ions should be equal to the concentration of added water. Further¬, more both of these ions are produced when SO3 is formed from H2S04; hence,, we may write the equilibrium constant for the reaction as follows:, , 2H2SO, ^ SO, + H,0+ + HSOt; K = —(15), , (ri2oc>4), , telling us that if the concentration of unionized H2S04 can be kept very nearly, constant and if activity coefficients may be neglected, then the concentration, of SO 3 should be inversely proportional to the product (H30+)(HS07) or to, (H20)^dded. If a single molecule of S03 is involved in the rate-determining step,, the sulfonation rate should likewise be, as is observed, inversely proportional, to (H20)*dded. On the other hand, it may be shown (Ex. 5) that if the attacking, species were H3SO£, HSO|, S2Oe, or H2S3Oi0 (all of which may be present in, concentrated sulfuric acid),6* the rate of sulfonation would show a different, type of dependence on (H20)added. The answer appears unequivocal until, we remind ourselves that in the concentration range studied (92 to 99 percent, H2S04) the concentration of unionized sulfuric acid is by no means constant65, and that the activity coefficients of one or more of the species in equation (15), may change significantly in going from 92 to 100 percent H2S04.6* Thus,, there exists, , the disturbing possibility that the rate data, when corrected, , for both of these trends, will approximate a rate law pointing to attack by a, different species. Analysis of rate data for sulfonations in fuming sulfuric acid65, is similarly difficult, due largely to the plurality of readily interconvertible, species present in such solutions. However if SOs is the attacking reagent for, sulfonations in 92 to 100 percent sulfuric acid, it is also very likely to be at least, one of the attacking reagents for sulfonations in fuming sulfuric acid., There are two additional features of the sulfonation reaction that must be, accommodated by any proposed mechanism. First, aromatic sulfonation, as has, « Cowdrey and Davies, J. Chem. Soc., 1949, 1471. Similar (as yet unpublished) studies by, Davenport and Hughes have been reported by Ingold, Ref. 1(b), p. 2))., et Gillespie, J. Chem. Soc., 1950, 2516., « Deno and Taft, J. Am. Chem. Soc., 76, 244 (1954)., ?, « These objections are essentially those expressed by Gold andffatcheU J. Ch, 1956 1635 However, on the basis of a more complex treatment, these authors a so sug|0, (ahhough with reserve) that the most likely attacking species in such sulfonaUons in the SOa, molecule., gs Brand, J. Chem. Soc., 1950, 1004.
Page 463 :
Friedel-Crafts Reactions, , 447, , long been known, is reversible in polar solvents. Secondly, aromatic sulfonation, exhibits a pronounced kinetic hydrogen-isotope effect,*'<«> suggesting that the, , C-H bond in the ring undergoes breakage in the rate-determining step. We, may then represent the mechanism for sulfonation as, H, slow, , fast, , ArH + S03, , I3St, , ^ ArSOi- + H+ ^ ArSOaH, , +Ar, , slow, , fast, , SOr_, XXXVI, In accordance with the observed isotope effect, the rate-determining step in the, forward reaction is the breakage of the Ar—H bond. If we assume the first step to, be fast, its reversal must also be fast since intermediate XXXVI does not accu¬, mulate. Since one of the steps in the reverse reaction (desulfonation) must be, slow and since the ionization of the sulfonic acid is very fast, the slow step here, must also be the second. Note also that the evidence so far available does not, tell us whether adduct XXXVI is an intermediate or merely an activated com¬, plex in a one-step reversible substitution.66, , Friedel-Crafts Reactions 67, The Friedel-Crafts reaction, the alkylation or acylation of an aromatic ring, in the presence of such catalysts as A1C13, BF3, SnCl4, or I2, is not an easy, reaction to study quantitatively. Such condensations are usually sensitive to, traces of moisture and are often accompanied by isomerization of the product, or by polymerization. Moreover, the “catalytic halide” may form addition, compounds with the aromatic substrate, with the alkylating or acylating reagent,, with the products, or with combinations of these, and some of these adducts take, part in the Friedel-Crafts reaction, whereas others do not. Finally, many, Friedel-Crafts reactions are heterogeneous, and these do not lend themselves, to straightforward kinetic treatment. Although olefins, alcohols, and (in special, cases) esters and ethers may serve as alkylating agents, and although acylations, may be carried out with esters, anhydrides, and carboxylic acids, we shall con¬, sider mainly the more usual Friedel-Crafts reagents—, alkyl and acyl halides., hese are generally used in conjunction with the halides of aluminum, tin, or zinc; but for quantitative studies, use of the anhydrous halides of gallium, , SOs (HiX^ood^1^; y. cLlS2:.;ni91t 1372C 194h4S4^6 4^ Us n" tT", , SEE, , ,he reac,i°" °f;, , tTso1; r,, , ", , r„s e, , ^.!’vFlTl,a355Ce("t954n)d CritiCa‘ diSCUSsion of the Friedel-Crafts reaction, see Baddeley, Quart.
Page 464 :
448, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems;, , is sometimes preferable, for with these, undesirable side reactions are often, minimized., The function of the metal halide is to loosen or break the carbon-halogen, bond in the alkyl or acyl halide, making the latter reagent more electrophilic, and therefore a more effective attacking reagent for the aromatic compound., , 5+, , 5-, , R—X + AlBr3 -> R—AT—AlBr,, The chlorides and bromides of, , or, , aluminum6S(°), , R+A’AlBrj, , and gallium6S(6) are known to, , form polar addition compounds with many (although not all) alkyl and acyl, halides. Typically, the addition of anhydrous aluminum bromide to a solution, of ethyl bromide in cyclohexane causes a striking increase in the apparent dipole, moment of the solute,65 although aluminum bromide itself has no dipole, moment. Such solutions appear to conduct the electric current, , 6S(°) with, , feebly,, , aluminum migrating to the anode. Furthermore, if aluminum bromide having, radioactive bromide is dissolved, along with unlabeled ethyl bromide, in an, inert solvent, the radioactive bromide is found to enter the ethyl bromide.es(c), These observations are consistent with the formation of a polar complex,, Et—Br: AlBr3; this is ionized to an unknown degree into the ion pair Et+AlBr4,, , which is, in turn, slightly dissociated into Et"*" 4* AlBr4. Halide exchange is, more rapid for secondary and tertiary halides but slower for methyl halides,, these differences being in line with what we observed in Chapter 8, , that a, , carbonium ion becomes more stable as branching at the a-carbon increases., , In Friedel-Crafts reaction mixtures where the concentration of carbonium, , ions (in ion pairs) is significant, it is likely that this ion is an attacking species, , and possible that it is the only important attacking species. Kinetics does not, , tell us whether the alkylating species is the complex RT:ALY3 or the ion pair, , R+ALYi; but it does assure us that if the alkylating species is the ion pair, then, , the formation of this ion pair cannot be rate determining. For if the ionization, , were the rate-determining step, the alkylation rates of a series of aromatics with, , a given alkylating mixture would be independent of the eoncentrat.on and, identity of the aromatic; whereas alkylations in the presence of GaCb,, , *, , A1C1,™<» and AIBr,™1' have been found to be first order in the aromatic, anc, “ W S«> <°r, , 69 Fairbrother, Trans. Faraday Soc, , 7*5*6275 (h?53)**78,6247*(195<S)^Wongand^Brown, , 37, 763 (, , 70 (a) Ulich and Heyne, Z. Elektrochem., 41, Chem. Soc, 75, 6285 (1953). (c) Jungk, Smoot, Brown, rbid., 78, 6245 (1956). In the, kept in large excess, so the kinetic order w‘th rtspec, , )., , ^ Grayson> j Am, >■, , ^ ^, , (1956) (d) Smoot anc, wa, , are much slovver thar, , However, the fact that alkylations o xnztnc, excess) tells us that the two alkyla, the corresponding alkylations of toluene (with toluene in excess;, tions do not have a common rate-determining step.
Page 465 :
Friedel-Crafts Reactions, this is probably also true for alkylations catalyzed by GaBr3., , 449, , ( > Alkylations, , catalyzed by A1C13 and AlBr3 in the slightly polar solvents nitrobenzene and, 1,2,4-trichlorobenzene, , are first order also in catalyst and in alkylating agent,, rate = A:(ArH)(ALY3)(R.<Y), , (16), , This rate law is consistent with attack on ArH either by RA. A1A 3 or by, R+AIA7, since the concentrations of both of these are proportional to (RA) (A1A 3)., However, it excludes attack by free R+ ions, since the concentration of R+ is, proportional to V(RA')(ALY3). Moreover, if the attacking reagent were the, free carbonium ion, the ratio of products (for example, the ortho to paia iatio, from alkylation of a monosubstituted benzene) should be the same whether, RC1, RBr, or RI is the alkylating agent; this, likewise, is not what is found.7*, The mechanism of Friedel-Crafts alkylations in nonpolar solvents is, at, present, obscure. It has been found that alkylations in benzene catalyzed by, GaBr3 are second order in catalyst,70(d) but there is no guarantee that this is true for, other solvents or for other catalysts.7* It would not be surprising however if, alkylations, like some aromatic halogenations in nonpolar solvents, require, “pulling action” by more than one molecule of Lewis acid., The familiar skeletal rearrangements that occur during Friedel-Crafts, alkylations may be assumed to arise from alkyl shifts or hydride shifts in the, carbonium ion while the latter exists as a portion of an ion pair. (Carbonium-ion, rearrangements are known to occur under a variety of conditions and will be, discussed at some length in Chap. 14.) Since we know that, in general, secondary, and tertiary carbonium ions are more stable than primary, we need not be, surprised that alkylation of benzene with neopentyl chloride yields /-amylbenzene (XXXVII). Similarly, alkylation with primary chloride XXXVIII, yields mainly the rearranged hydrocarbon XXXIX.73 It is, however, somewhat, Me, , Ph, , Me, , I, Me-C-CH2C1 —318> Me-C-CH^:, , I, , Me, , Me, , +, I, Me—C—CHS, I, , PhH, , Me, , A1C17, , I, , Me, , I, , -> Me-C-CH, I, Me, , A1CI7, , XXXVII, (H\, Me3C-CH2-CH2Cl, , Me3C-C-CH+ ^ Me3C-C-CH8-^^Me8C-CHPh-CH3, H, , H, A1C14, , A1C17, 4, , XXXVIII, , XXXIX, , n ?Tr,°Wn and JunSk> J■ Am■ c/tem. Soc., 77, 5586 (1955)., enceofeamum^^T’, , Ga,a, , A, , ,, , stlg, , 70(a?’ rep°rt f°r examPle> that alkylations in CS2 in the pres-, , f, , the work is rhi^ p-^VeM Chie“y “, , hmerling and West, J. Am. Chem. Soc., 76, 1917 (1954).
Page 466 :
450, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , disturbing to find examples where alkylation with a tertiary halide yields a, pioduct derived from a secondary carbonium ion—for example, alkylation, with chloride XL to give product XLI75. To account for the rearrangement in, , MeCH,—CMe,Cl, , A1C!;, , (SX, , Ph, , H, , + _, >Me-C-CMe 2 'V, , +, , I, , Me—CH— CMe 5, , Me—CH—CMe2H, , (mainly), , H, , Aicir, , Aicir, , XL, , XLI, , this direction, we should recall that a secondary carbonium ion, besides being, less stable than a tertiary, is also more reactive. Therefore, if we assume that the, two carbonium ions are allowed to reach equilibrium, we may represent the, interconversions involved in the “skeletal rearrangement” as follows:, secondary, halide, , secondary carbonium ion, (small equil cone), , tertiary carbonium ion, (larger equil cone), , PhH, , PhH, , fast, , slower, , ‘secondary” product, , tertiary, halide, , ‘tertiary” product, , As may be seen, the ratio of “secondary” product to “tertiary” product depends, not only upon the relative stabilities of the carbonium ions, but also upon the, rates with which these ions react (as ion pairs) with the aromatic substrate., On the other hand, there seem to be a number of substrates that are so reactive, they consume the carbonium ion originally formed, before the latter has a, chance to undergo significant isomerization. Thus, alkylation of mesitylene with, rc-PrBr yields 91, , percent of the n-propyl product under conditions where, , alkylation of benzene yields only, , 21, , percent of the /z-propyl product.70(d), , Having implied that isomerization requires ionization of the alkylating, agent whereas alkylation itself does not, we may expect the milder catalysts, that stretch the R—X bond without completely breaking it to bring about, alkylation without isomerization. It appears that ferric chloride is a catalyst, of this type, although it is often difficult to determine whether a bond has been, broken or merely loosened in solution. At any rate, Friedel-Crafts alkylations, of benzene with chlorides XXXVIII and XL in the presence of FeCh yield, unrearranged products/5 whereas rearrangement predominates in the presence, of AICI3., .. ., Friedel-Crafts acylations74 are similar in a number, of respects to al '>’ alions. Here, the attacking electrophile is an acylmm ion, R, , 0=0, a type of, , carbonium ion which we have already encountered in our discussion of esten ca¬, tion (p. 325). Besides being formed from the action of aluminum halides ana, n For recent quantitative studies of Friedel-Crafts acylations, , see: Brown^, , J. Am. Chem. Soc., 80, 2291, 2296 (1958); Denney and Klcmchuk, ,M„ 80, 3285 (195 )
Page 467 :
Friedel-Crafts Reactions, , 451, , other Lewis acids on acyl halides, such cations exist in high concentrations in, solutions of acid anhydrides in concentrated sulfuric acid, 75(o) or may be formed, by the action of a soluble silver salt (for example, AgC104) on an acyl halide in a, polar but nonhydroxylic solvent such as nitromethane.75( }, , O, +, , (RC0)20 +, RCOC1, , 2H2SO4, , II, , 4-, , —> R—C=0 + R—C—OH2 + 2HS04, , + Ag+, , -> R—C=0 + AgCl, , Solutions of either type are effective acylating agents., Acylations, unlike alkylations, are not accompanied by rearrangement of, the attacking electrophile, but are subject to a different complication. In some, cases, especially when “R” of the R—C=0 ion is highly branched, decarbonylation (loss of carbon monoxide) occurs.76 Trimethylacetyl chloride, for, example, behaves in this way:, , O, Me3C-C-Cl:AlCI3-> [MesC-^=0]+, , -v MesC+ +, , + AIC17, , C^O —rH> ArCMe^, , +A1C17, , This is a “side reaction,” and, as is the case with rearrangements during, alkylation, it may be minimized by using a very reactive substrate. For the more, reactive the substrate, the more likely it is to be attacked by the acylium ion, before the latter undergoes decarbonylation. Given similar conditions, the, reactions of trimethylacetyl chloride and aluminum chloride with the three, aromatics, benzene, toluene, and anisole result in yields of CO of about 90, 50,, and 10 percent, respectively.76, It is of interest that Friedel-Crafts acylations require somewhat more than, stoichiometric amounts of electrophilic catalysts, whereas alkylations proceed, nicely in the presence of much smaller amounts. The product of an acylation is, a ketone which is itself basic and which may inhibit the reaction in one of two, ways. It may “tie up” the catalyst, thus preventing the ionization of the acylat¬, ing agent, or, alternately, it may “tie up” the acylium ion, preventing its at¬, tacking the ring. More recently, halide exchange experiments have shown77, hat ketones when added to Friedel-Crafts reaction mixtures, do not prevent, mzation of the acyl halide. Therefore we may conclude that, in the absence, ror2_ag, , b“", , rSh/T CXa?!e’ Rothstein and Saville, J. Chem. Soc. 1949 1946, Baddeley and Voss, J. Chem. Soc.y 1954, 418., ’
Page 468 :
452, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , of substantial amounts of catalyst, the action of the acylium ion, rather than, that of the catalyst, is blocked. The competition between the ketone and the, substrate for the acylium ion may thus be represented as follows:, R, , \, yC—O: A1C13 + HC1 (desired products), , O, , O, •fl, , R—C—Cl: A1C13 ^, , R-C + A1C17, , O, , Ar, .., /, R-C:Q=C, , II, , \, R, , Aicir, , o, , (not possible when Ar—C—R is, coordinated with A1C13), , It is probable that if the substrate is sufficiently reactive, acylation may, O, occur through attack by R—C—Cl:A (where A is a Lewis acid) without prior, ionization of the latter. Thus, the acylation of anisole in the presence of iodine75, O, almost certainly proceeds through the complex R—C—CLL rather than the, ion pair R—C=0 Clly; for although there is evidence that the Clly anion, might exist,75 it is, at best, extremely unstable under ordinary conditions., , Nucleophilic Aromatic Substitution. The Bimolecular Mechanism'0, The duality of mechanism which we have considered with respect to nucleo¬, philic aliphatic substitution (Chap. 8) also extends to aromatic systems. The, two most familiar types of aromatic substrates that undergo such substitutions, under nondrastic conditions are nitrohalobenzenes and diazonium salts, typical, reactions of which are shown:, , + cr, , + r2nh, , h2o, M Kaye, Klein and Burlant, J. Am. Chem. Soc., 75, 745 (1953)., , 79, , von Kiss and Urmanczy, Z. anorg. Chem., , »(„), , 202, 18, , (, , , Zahler,, , %%Tow^CkJ,,,y, McGraw-Hill Book, ^ Hb), pp., , 12, 1 (1958)., , )., , -8,5; Bunnott,
Page 469 :
Nucleophilic Aromatic Substitution, , -, , 453, , Studies of reactions of the first type leave little doubt that they are bimolecular, in character, whereas certain reactions of the second type are just as unmistak¬, ably unimolecular., Second-order kinetics have been observed for many substitution reactions, of ^-nitrochlorobenzene,, , 2,4-dinitrochlorobenzene,, , and, , the corresponding, , bromo compounds,'81 and the reactions of a given substrate are fastest with the, most basic nucleophiles.* While these observations strongly indicate bimolecularity, they do not demand a mechanism analogous to the direct displacement, mechanism in aliphatic systems in which bond making and bond breaking are, synchronous: For, if the reaction center is an aromatic, rather than an aliphatic,, carbon, an attacking nucleophile, like an attacking electrophile, may form a, new bond before the old bond breaks. Indeed, as we have already suggested, (p. 237), a nitro group ortho or para to the position under attack facilitates the, reaction by dispersing the negative charge brought in by the attacking group, before departure of the leaving group (XLII, , XLII')., , For evidence in support of this picture, consider the specific rates at which, 2,4-dinitro compounds of type XLIII react with piperidine:**, , XLIII, , ?, , C-*=-Cl,—Br, — S-PhS—Ph), , O, , O, , the breakage of the C—A’ bonds were important in the activation process for, t ese substitutions, the chloro compound should react much more slowly than, t e, , romo compound (since a C—Cl bond is more difficult to break than a, , XAm.CtaL^263 n95”l, active bromide (ie Roux j 1, , “dr °f^’^‘dinitrochlorobenzene (Reinheimer, „ „l.,, , ” Cta tC794T^ 4hdm,tr°br0m0b'nZen' with radi°-, , s sar***a ssssstas&vsss,, “Bunnett, Garbisch, and Pruitt, J. Am. Ch,m. Soc., 79, 387 (1957).
Page 470 :
454, ^, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, bond), and the, , SOPh and —S02Ph derivatives should react at very, , different rates. It is found, however that the chloro, bromo, and —SOPh, derivatives react at almost exactly the same rate, and the —S02Ph compound reacts, about two thirds as fast as these. Moreover, the activation energies for the four, substitutions are very nearly equal, telling us that the activation process in the four, cases is substantially the same and not affected by the nature of substituent A'., , It is likely, therefore, that the activation energy required in these substitutions, is utilized in desolvating the piperidine molecule, in redistributing electric, charge in the benzene ring and about the attacking nitrogen atom, and in, pushing the C—A bond out of the plane of the ring, but not in breaking the, C—X bond. It might also be noted that the corresponding substitution reaction, on 2,4-dinitrofluorobenzene is about 750 times as fast as those on the chloro, and bromo compounds (whereas if breaking of the C—F bond in the activation, process were important, the fluoro compound should react more slowly). In, line with the picture we have presented, it may be inferred that the strong, electron-attracting power of the fluorine atom lowers the electron density at, the center of reaction so that attack by the piperidine molecule becomes some¬, what more favorable energetically., We may next ask whether a species such as XLII is an intermediate or, merely an activated complex. The isolation of salts having anions that are very, probably analogous to XLII indicates that at least in some cases, and possibly, in all, XLII is an intermediate. The most familiar of such salts is XLV,, formed by the action of KOEt on 2,4,6-trinitroanisole (XLIV). As is indicated,, OMe, , 02N^k^N02, , no2, XLIV, , this salt appears identical to the salt formed under similar conditions from, KOMe and 2,4,6-trinitrophenetole (XLVI); on acidification, both products, yield the same mixture of trinitroethersSimilar salts have not yet been pre¬, pared from the reactions of nitrohalobenzenes, and it may be that, because o, the greater effectiveness of the halide ions as leaving groups, such salts are too, unstable to be isolated., There are, as may be expected, groups other than the —NOo, facilitate nucleophilic substitution in the ortho and para positions. The —, 83 Meisenheimer, Ann., 323, 205 (1902).
Page 471 :
455, , Nucleophilic Aromatic Substitution, O, , —CN, and —S—CH3 groups are in this category, but each is a far less potent, O, activator than the nitro group. The relative rate constants for the reactions of, the following substituted bromobenzenes with piperidine (in benzene at 99°), are typically illustrative of the greater effectiveness of the nitro group.54, , o, I, i, *BrC6H4N02, , o, 0.053, , In addition, the diazonium group, —N=N+, is an effective activator for nucleo¬, philic substitution, although this group is itself so easily displaced that quantita¬, tive studies of substitutions activated by it are difficult. There are, however,, many instances where a nitro or halo substituent, ortho or para to the diazonium, group in a diazo salt is displaced during a reaction that is intended to alter only, the diazonium group.55, As is seen from structure XLII', the action of the nitro group in facilitating, nucleophilic substitution requires that the bond linking this group to the ring, assume some double-bond character and therefore that the nitro group lie in, or near the plane of the benzene ring. The nitro group becomes a much more, eeble activator when it is forced out of the plane of the ring. We have already, noted (p. 238) that 2-nitro-5-bromo-tn-xylene (LI) undergoe! substitution reac¬, tions much more slowly than /.-nitrobromobenzene simply because the methyl, groups stencally prohibit the approach to planarity that is necessary for effective, IJUnnett an.d Levitt> J- Am. Chem. Soc., 70, 2778 (1948), , (a reaeUoV^To^, , IndNPo“-n,'y Pr°dUCG XUX a"d 4, , and XLVIII with ethanol
Page 472 :
456, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , activation by the nitro group. A similar effect boosts the activation energy for, the reaction of l-nitro-2-bromonaphthalene (LII) with piperidine; AHt for, this reaction has been found to be 12.3 kcal as compared to a AH* of 10.4 kcal, for the same reaction of the l-bromo-2-nitro isomer (LI II).56 Here it is the, so-called peri hydrogen (shown in the figure) that interferes with coplanarity, of the nitro group in LII. On the same basis we may understand why the nitro, group in the 1 position of compounds LIV and LV is more easily displaced, than that in the 4 position, even though the former, classically speaking, is, “hindered.” In both of these compounds, the 4-nitro substituent may facilitate, , Me, 87a, , o2n-/~Vno2 + nh3, , +, , NOf, , Me7, LIV, , -b Nor, , 87b, , displacement of the 1-nitro substituent, but the latter, because it is kept out of, the plane of the ring, may not facilitate displacement of the 4-mtro substituent., , " Berliner, Quinn and Edgertor^/O-j, *7 (a) Ibbotsen and Kenner, J. Chem. i>oc., 1ZJ,, , 39’ nKspUcement of the 1-ni.ro substituent in, steric hindrance than might be supposed; for m the, the, , incoming and leaving groups are well out, , • »£““ Eman, Rcc. ,ra, chi,n.,, , \ i, , and, tKtn^ d, , P££ it, , pi, , f, , ^ ,w0, , may be shown that unless the, , substituents “sandwiched” between them. U, g, exccssively bulky, there is little steric, incoming group or the ortho substituents (or both) are excessively, y,, interference in the transition state.
Page 473 :
The Reactions of Diazonium Salts, , 457, , The Reactions of Diazonium Salts. The Aromatic Sn 1 Reaction, The formation of diazonium salts was discussed briefly in an earlier chapter, (p. 175), where it was shown that the kinetic picture suggested the following, sequence of steps:, 4fast, , 2HN02 ^ N203, eq, , fast, , ArNH2—N=0 -> -> ArN+ + H20, , ArNH2, -», slow, , NOF, , with the detailed nature of the steps following the slow step still in doubt. Once, formed, aromatic diazonium salts react with a host of different nucleophiles, under a wide variety of conditions, the nucleophile attaching itself to the ring,, and the diazonium group departing as N2. Many of these reactions, particularly, those carried out in basic solutions or in nonpolar media, proceed by free-radical, mechanisms; certain of these will be treated in Chapter 16. On the other hand,, the reactions of diazonium salts in acidic polar media are often heterolytic in, character, and a number of these provide us with what is, at present, the only, important group of examples of the 6V1 reaction on the aromatic ring., The evidence for unimolecularity is much the same as that discussed for, aliphatic systems. The decompositions of substituted benzenediazonium chlo¬, rides in water are first order55 and, aside from solvent and salt effects, their rates, are not affected by added alcohol or added chloride. Nevertheless, these reagents, partially divert the product from a phenol to an aromatic ether or substituted, chlorobenzene. We thus may infer that the rate-determining step in these con¬, versions is the bond breaking in the diazonium ion, followed by a rapid com¬, bination with the attacking nucleophile which determines the identity of the, product., 7, , ArN, , -N,, slow, , > ArH, , ArOMe, , +, , H, , rate determining, ArCl, product determining, , ^Zrlw^gsXlhfemfin diP|a‘iC ^ reaCti°nS’ WC Sh°U'd eXpeCt el«tron., , ad,cal pa,h;, , also via a fre':
Page 474 :
458, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , the specific rates of decomposition of the m-Cl and p-N02 substituted benzenediazonium salts are, respectively,, , 3^4, , and, , ^40, , of that for the parent salt (in, , water at 29°), whereas the m-CH3 and m-OH groups boost the specific rate of, decomposition (by factors of 4.5 and 12, respectively).89{b) It is, however, some¬, what disturbing to find that the /)-OH and />-CH30 groups greatly reduce the, rates of decomposition of aromatic diazonium salts. Here, it appears that, because of direct conjugation between the —group and the —OH or —OMe, group (LVI <-> LVI'), the bond linking the N£ group to the ring acquires, , some double-bond character, and therefore becomes more difficult to break., Likewise, />-toluenediazonium chloride decomposes more slowly than the benzenediazonium salt, presumably because hyperconjugation (LVII) strengthens, the C—N bond in the tolyl compound., It is emphasized that the kinetic work on the reactions of diazonium salts, in polar media has been concerned in the main with attack by weakly nucle¬, ophilic species—that is, H20, Cl", and alcohols. It is quite probable that if, studies were extended to reactions with more effective nucleophiles (for example,, I-, CN~, PhS-, and SCN-), bimolecularity would be observed. In fact, it has, recently’been found that the addition of the moderately strong nucleophile,, Br“, to a solution of />-nitrobenzenediazonium ion not only diverts some of the, product to /,-nitrobromobenzene, but also accelerates the heterolysis of the, diazonium salt. In this case, the rate of decomposition is equal to the sum o, two terms,, (17), , = *i(ArN£) + *2(ArNJ)(Br-), dt, , the second term corresponding to direct displacement of N, by the brom, ion.» The more usual way of replacing an aromattc diazonium group with, bromide (or with chloride or cyanide) involves reaction w.th the cupro, complex CuBr? (or with CuClj or Cu(CN),), the, This reaction, which is thought at present to procee, , ro, , mechanism, is discussed in Chapter 16., , The von Richter Rearrangement, The reaction of meta- and pam-nitrohalobenzenes with oqoeous alcoholic KCN, above 1 50° results predominantly in loss of the mtro group as NO, frathe, 90, , Lewis and Hinds, J. Am. Chem. Soc., 74, 304 (1952).
Page 475 :
The von Richter Rearrangement, , -, , 459, , displacement of the halogen atom) and the acquisition of a carboxyl group., This reaction is often called the von Richter rearrangement, for it was this worker, who found that the position taken by the incoming carboxyl group was not, the same as that vacated by the nitro group.w(o) More specifically, the reaction, of a />-nitrohalobenzene yields a meta compound, whereas the reaction of a, m-nitrohalobenzene yields a mixture of ortho and para compounds., , COOH, Br—\, , N02 + CN~ _ 150'200O._>, H2Q-EtOH, , Br, , Br, , / \, , /—COOH +, , (, , +, , no:, , y—COOH + no;, , Br, The migration of a cyano group is not involved here, for meta- and para-halobenzonitriles do not rearrange. What might appear to be the most reasonable, mechanism is, , LVIII, ^^/COOH, , X, , X, , ^r“e^>Caacnd0t be COTPl:tCly COrreCt’ f°r hal°benZonitri,es related, employed. Moreover, , the'T'd, , UndCr the reaction conditions, , independently likewise resist^h H "i ^, , '° the, acid’ when Prepared, tty, likewise resists hydrolysis under, vonfinal, Richter, conditions, , «, Chm-'5’481 «** M, , more reactant studies of ,his reaction see., and McKay, T. ote.
Page 476 :
-, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , A more acceptable (but still tentative) mechanism is as follows:"<•*>, , Once again it is proposed that the initial intermediate is adduct LVIII, formed, by attack of CN~ on the electron-deficient ring carbon ortho to the nitro group., It is further suggested that the nitro group assists the hydrolysis of the cyano, group, possibly by formation of the cyclic anion, LIX. As shown, the resulting, species LX is tautomeric with the substituted benzoyl nitrite, LXI, which is, in, turn, readily hydrolyzed to the resulting carboxylic acid. This scheme is an, interesting one, but further investigation is obviously desirable., , Animations of “Nonactivated” Aryl Halides. Benzynelike, Intermediates, Aryl halides having no activating group such as —NO, or —CN may be con¬, verted to arylamines using metallic amides (for example, NaNH, or LiNEt.)., However, it is now clear that the mechanism by which these conversions gener¬, ally proceed is very different from the substitution mechanisms that have thus, far been considered. Indeed, these reactions are not simple substitutions at a, but are instead combinations of elimination (of the hydrogen halide) and additionAs indicated, the intermediate in such a sequence is a substituted benzyne LXH,, a species of unusual interest since, despite many attempts, no compound having, a triple bond in a six-membered ring (or, for that matter, in a three- fou^, or five-membered ring) has been isolated."' For this reason, we would be, " Ideally, in an acetylenic derivative, the two, , ^arrangement,, , We triple bond is, , or even a, , incorporated into, , a
Page 477 :
Aminations of "Nonactivated" Aryl Halides, , -, , 461, , skeptical of the existence of an intermediate such as LXII if the evidence in its, favor were not quite compelling., , R', , R', , A, , ^, , -H*, (elimination), , rj, , + R2NH, , (18), , J, , -■■—^, s, (addition), , a benzyne, , LXII, , V., , The sequence above suggests that the aminations of aryl halides should be, accompanied by rearrangement, but that these rearrangements need not be, complete (as in the von Richter reaction). Such rearrangements have long been, known to occur.55 Furthermore, the entering nitrogen atom should not be found, further than one carbon atom away from the position of the departing halogen, atom; that is, orMo-substituted halobenzenes should not give rise to parasubstituted arylamines. This also is observed. Typically, treatment of o-chlorotoluene with KNH2 in liquid ammonia yields a mixture of o- and m-toluidines, (but no/>-isomer); similar treatment of />-chlorotoluene yields a mixture of mand />-toluidines (but no o-isomer); and treatment of w-chlorotoluene yields a, mixture of all three toluidines.Wa), For reactions in which only one benzyne intermediate is possible, sequence, (18) demands that the ratio of isomeric amines in the mixture of products be, independent of the nature of the halogen atom that departs in the elimination, step. This has been shown to be the case for the mixtures of amines resulting, rom the action of NH3 and KNH2 on the o-halotoluenes^(o> and also for the, mixture of a- and /3-naphthylpiperidines resulting from the action of piperidine, , LXIII, A — Cl, Br, I, , <», , *- *, , <**>•
Page 478 :
462, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , and sodamide on a-chloro-, a-bromo-, and a-iodonaphthalenes.95 Moreover, if, the benzyne intermediate is symmetric* the incoming nitrogen atom should be, just as likely to attack at the position from which the halogen atom departed, as at the adjacent position. In this regard we have already pointed out (p. 144), that when chlorobenzene having the 1-carbon labeled (LXIV) is treated with, KNH2 in liquid ammonia, almost exactly 50 percent rearrangement (actually, , NH,, , +, 48 percent, , LXIV, , 'NH2, 52 percent, , 52 percent rearrangement) occurs.96 Finally, if sequence (18) represents the, correct path for amination, halobenzenes having no ortho hydrogens should not, undergo amination unless reaction conditions are made sufficiently severe so, that direct displacement of the halogen may occur. It has, in fact, been found, that bromodurene (LXV), bromomesitylene (LXVI), and 2-bromo-3-methylanisole (LXVII) are inert toward KNH2 in liquid ammonia,^®-97 although, steric hindrance to attack on the ring should not be prohibitive.66, , A, , Br, , Br, , Br, Mcx, , ,Me, , Me^/L/Me, , Me/, , "Me, , V, , MeO^J\, , o, , .Me, , Me, , LXV, , LXVII, , LXVI, , None of the observations concerning the animations of halobenzenes will,, when taken alone, prove the intervention of the benzyne intermediate, but,, when taken together, they comprise a rather convincing case. Furthermore, we, now have evidence that benzyne intercedes as an intermediate in other reactions,, for this species apparently is a powerful “dienophile” in the Diels-Alder reaction., Thus, if o-bromofluorobenzene is treated with lithium or magnesium in the, presence of furan96(o) or anthracene,96(6) products are isolated that may, , e, , rationalized most easily by assuming a Diels-Alder condensation with benzyne,, which, in turn, results from dehalogenation of the starting material., « Bunnett and Brotherton, 7. Am. Chem. Soc, 78, 6265 (1956). Here, the intermediate is, presumably naphthalyne, LXIII., " The slight departure from 50 percent rearrangem, unexpected, for reactions a, C--labeled posit,o„s are, , . /, (, , kinetic isotope effect) is not, P slightly slower, y, to, sl.ght y, , than those at unlabeled positions. See Ropp NucUoma.10, ^ O95 >■, ” Benkeser and Buting, J. Am. Chem. Soc., 7 ,, /1955) (*) Wittig and Ludwig, ibid*, 9* (a) Wittig and Pohmer, Angew. Chem., 67, 348 (19W). W, 68, 40 (1956).
Page 479 : Aminations of "Nonactivated" Aryl Halides, , 463, , The elimination-addition (benzyne) mechanism is an important one for, the substitution reaction of “nonactivated” aryl halides, but there is evidence, that substitution by direct displacement (the .5V2 reaction) may occur as well., For example, although a-chloro-, a-bromo-, and a-iodonaphthalene yield the, same mixture of naphthylpiperidines on treatment with piperidine and sodamide,, the a-fluoro compound yields much more of the unrearranged product than do the, other three halides.53 This suggests that a-fluoronaphthalene is reacting by both, the benzyne and Sn2 mechanisms, whereas the heavier halides react almost, exclusively by elimination-addition. This is not surprising, for we have seen that, aryl fluorides undergo SN2 reactions much more easily than aryl chlorides,, bromides, and iodides (p. 454). On the other hand, fluorides are known to be, converted to benzynes only with difficulty™) (presumably because such a con¬, version requires the breaking of a C-F bond with considerable double-bond, character)., The basic hydrolyses of halobenzenes and halotoluenes are negligibly slow, at room temperature, but occur readily at 250 to 350°. It has recently been, -T /o656 17dr°lySeS may’ dePending upon reaction conditions, proceed, fclntr by,the “benZyne mechanism.” or by a combination of, hv ,1, k hUS’ /’-bromoto,uene 15 hydrolyzed by 1 N NaOH at 340° largely, mechl -enZyne 7eCha"1Sm (5° perccnt rearrangement), and at 250° by both, nydrolyzed at 250, , (22, whereas /Modotoluene, almost completely by direct displacement (less than, , has been, Chl°robenzene, with the chlorinated carbon labeled, has been shown to be hydrolyzed by both mechanisms a, 340°. f, ,, resulting phenol molecule,, „, uecnamsms at 340 ; for among the, , /ft, , with on* 42 perc”m atekd at rm ", amination of chlorob nzene, ^ ** a Slight, , k, , ' “, , k P°Sm°nS ^ ‘°, , carbon, , * <As, , “ ::;7ly by, , Botum and Roberts, J. Am. Chem. Soc., 79, 1458 (1957), , h-e -en, the, benzyne mecha-
Page 480 :
464, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , It should be emphasized that the benzyne mechanism may operate only in, the presence of strong base, since only strong bases can remove protons from the, benzene ring. Thus, /?-bromotoluene, when treated with aqueous sodium, acetate at 340° yields/?-cresol but none of the ortho or meta isomer,95 indicating, that the destruction of the halide has, in this solution, occurred only by the, Sn2 mechanism., , EXERCISES, , FOR, , CHAPTER, , 1 1, , 1. Predict the position most likely to be taken by the incoming nitro group in the mono¬, nitration of each of the following. Justify your choice in each case., (a) C6H5CF3, (b) Thiophene, (c) Naphthalene, (d) C6H5SCN, (e) C6H5Se02H, , Me, , (m), 'NHAc, (n), O", , (o), , XT, , 2. Consider the presently accepted mechanism for aromatic nitration:, , 2HN03, , ^ H2NOt + NO^
Page 481 :
Exercises for Chapter 11, , 465, , A NOf + H >0, , H,NO+, , (fast), , > ArNO, + H+, , NOJ + ArH, , ~TT, H, , J, , (a) Assuming that, very shortly after the beginning of the reaction, the concentration, of NOj reaches a steady state value, show that the reaction rate is given by the, expression:, , k, , k2 (HNOs)2, k.-i, (NOF), , ■*_2(H2o), _*3(ArH), , (b) Assume that k\ and £_i are large (that is, that the first step is readily reversible), and that the concentrations of nitric acid, water, and nitrate ion remain nearly, constant during the course of the reaction. Show that the rate law in (a) predicts, a nitration rate very nearly independent of (ArH) when the attack by NO^ on, ArH is much faster than its return to H2NO^. Show also that if the attack by NO^, on ArH is much slower than its return to H2NO£, the nitration rate becomes, proportional to (ArH)., (c) Show that if the rate of attack by N02 is comparable to its rate of reconversion to, H2N03, the nitration rate will depend on (ArH) but will not be directly pro¬, portional to it., (d) Show that if both (NCFf) and (H20) are small, doubling (NO^) will be far more, effective in retarding nitration than will doubling (H20)., (e) For the nitration of a very reactive aromatic, show that a large increase in (H20), will result in a transition from pseudo-zero- to pseudo-first-order kinetics., 3. Consider the inhibition of nitration by HN02. Assume that this inhibition occurs by, the action of N03 generated in the reaction:, HNOz + HN03, , NO+ + HoO + NOJ, , (a) Assume further that the concentration of nitrate in such a solution may be esti¬, mated by adding (NOa) present in the solution, before HNO, is added to the, concentration of nitrate generated by addition of HNOz. Show that the’rate of, nitration in the presence of HN02 is, (rate)o, , rate =, , 1 +, , a, , VC'HNOo”), , ™r^1r;;trati0n ra'e in ,he absence of HNO=> « '» a constant, and, IZo,, concentration of tripositive nitrogen present very largely, (b) According to (a), a plot of (rateo/rate vs. a/(“HNO ”1 ,1,,, ij •, l.ne of slope a. It is found, however that a, Xn, ’ } Sh°U'd g,Ve a straiSht, acid, this plot is “concave downward”—that, of added nitrous, closer to unity than would be predicted hv th, ’ h, ^, (jate^/rate is, predicted by the expression in (a). Explain.
Page 482 :
466, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , (c) Consider the additional mode of inhibition of nitration by HN02 in the presence, of moderate concentrations of water. This presumably occurs by the action of, small amounts of NOT with H2NO^:, NO7 + H2NOl, , hno2 + hno3, , Assume that the NOT arises from N203> which is, in turn, generated from N204:, 2N204 + H20^ N2Os + 2HN03, N203, , NO+ + NOT, , and assume also that most NO+ present in solution is generated from N204, N204, , NO+ + NOT, , Show that (NOT) is very nearly proportional to (“HNOT’) ", and that if this, mode of inhibition is considered to augment additively the inhibition by NOT, discussed in (a), then the rate of nitration may be approximated by the expression:, (rate)p__, ratC =, , 1, , + a (“HN02”)^ + b (“HNO2”)*, , (d) Show that the rate law for nitrosation in nitric acid (equation 2) is consistent with, attack by NO+ or N204, but inconsistent with attack by HN02, H2NOt, or, n2o3., 4, , The rates of aromatic sulfonations is concentrated sulfuric acid containing small, amounts of added water are inversely proportional to the square of the concentiation, of added water. Show that this is consistent with attack by S03 or H2S207, but not, with attack by H3S04”, HSO3T or S2C>6., , 5. The para to meta ratio resulting from the nitration of toluene with acetyl nitrate in, acetic acid is 8.5., (a) Estimate p for aromatic nitration in this medium., (b) Estimate the para to meta ratio resulting from the nitration of amsole under, similar conditions. (Ignore nitration via nitrosation.), , 6., , Propose mechanisms for each of the following conversions:, (a) HNOs + H2018 -> H30+ + [02N0'8], (b) N205 + C6H6, , PhN02 + HN03, , (c) C6H6 + DC1, * C6H5D + HC1, (HC1 and A1C1S do not react in the absence of other substances. ), , OH, hno2, , a-naphthol ^, , (d) PhNH2, N=N—Ph
Page 486 :
470, , -, , Electrophilic and Nucleophilic Substitutions in Aromatic Systems, , . Consider the sequence of steps in the animation of halobenzenes by the “benzyne'’, mechanism:, , 8, , (a) What evidence is there that the rate-determining step does not occur after the, formation of benzyne (that is, that the formation of benzyne from the halobenzenes is neither fast and irreversible nor fast and reversible)?, (b) What evidence is there that the first step is not fast and irreversible with the second, step slow?, (c) From (a) and (b), it follows that the initial step must be either fast and reversible, (with the second step slow) or slow (with the second step fast). Outline an experi¬, ment with an o-deuterohalobenzene to decide which possibility is correct., (d) When o-deuterofluorobenzene is treated with KNH2 in liquid ammonia it is con¬, verted to unlabeled fluorobenzene, but no appreciable amination occurs. Explain., (e) When o-deuterochlorobenzene is treated with KNH2 in liquid ammonia and the, unreacted halide is reisolated before amination is complete, the deuterium content, of the aryl halide is found to decrease. If the corresponding experiment is carried, out with o-deuterobromobenzene, the deuterium content in the unreacted halide, is found to increase. Explain., (f) The rates of amination of aryl halides with KNH2 in liquid ammonia lie in the, order: PhBr > Phi > PhCl » PhF. Explain., (g) The aminations of both o-chloroanisole and m-chloroanisole yield only m-anisidine. Explain., 9. Explain each of the following:, (a) The rates of nitration of nitrobenzene in H2S04—HNO3 mixtures parallel the, degree to which indicator LXVIII is converted to its acidic form but do not, parallel the conversion of LXIX (anthraquinone) to its acidic form., , O, C-OH, , LXVIII, , LXIX, , (b) Iodination of mesitylene in CC14 with IC1 is third order with respect to 101, but, the same reaction in liquid CF3COOH is first order with respect to ICl., (c) Aromatic nitrations with N205 in CCL may be accelerated by addition of pure, , HNO3.
Page 488 :
CHAPTER, , 12, , Beta-elimination Reactions, , The introduction of unsaturation, , into a molecule generally involves the loss, , of two substituents from a pair of adjacent atoms in a chain or ring. Commonly,, a proton is lost from one of two bound carbon atoms and a nucleophile, X:,, is lost from the other., “I, , 10, , -HA', , V, , 0/, , /, , \’, , c=c, , X—C—C—H->, , (-Z, , —halogen, —NR^", —OHJ, —SR^, —O—C—R, etc.), , O, This is a beta elimination, the carbon atom bearing substituent —X being desig¬, nated the a-carbon, and the carbon atom from which the proton is removed, being the 0-carbon. A number of additional types of beta elimination have, been studied in some detail, among them, — HCl, , ArCH=N—Cl, RoCH—O—N02, , HN°4 R2C=0'<», OH-, , CH3CCH2CMe2—OH, , and, , » ArC=N/(o), , > CH3CCH3 + Me2C=0/(c), O, , O, , However, in this chapter, we shall be concerned only with those beta elimina¬, tions in which olefins and acetylenes are formed.*, . («) Hauser, LeMaistre, and Rainsford, 7., Easty,, , J.Chm.Soc.,, , An, Cten Soc, , 57,1056 (M35) (4) Bjk«, , 1952,1193, 1208 (r)LaMer and ^E|imi(natic«., , p‘"Mtnd ^71 x aZmL,W, 472
Page 489 :
473, , Unimolecular Eliminations, , Unimolecular Eliminations. The E1 Mechanism, In discussing substitution reactions, we noted that 5V1 reactions are often, accompanied by olefin formation, which does not, however, alter their kinetic, character (p. 263). In such cases, the carbonium ion formed in the initial slow, step may be attacked in more than one way by nucleophilic species present in, solution. Consider, for example, the ethanolysis of /-butyl bromide, which,, according to our present picture, proceeds through the /-butyl cation (I):, CH,, CH3—C—CH3, , (81 percent), , O, /+\, H, , Et, , (CH,),C-Br, CH,, , Et, \+, O-H, , +, , CH2=C-CH<, , H7, (19 percent), , As indicated, a molecule of ethanol may attach itself to the positively charged, carbon, completing the substitution process, or it may extract a /0-hydrogen, from the carbonium ion, forming isobutene. With /-butyl bromide, in which all, 0-hydrogens are equivalent, only one olefin is formed, but with more compli¬, cated substrates such as /-amyl bromide, EtCM^Br, elimination may take, place in two (or even three) directions, forming two (or three) olefins., The slow step in the elimination reaction is the ionization of the substrate., Since this step is unimolecular, this elimination and others like it are often, designated unimolecular elimination (El) reactions, in obvious analogy to, unimolecular nucleophilic-substitution (SN1) reactions. The El mechanism is, an important one for olefin-forming eliminations; it has been established for a, number of alkyl halides and sulfonium salts,3 and it is almost certainly the path, by which acid-catalyzed “dehydrations” of many alcohols proceed. We would, expect to encounter this mechanism in cases where the substrate may yield a, relatively stable carbonium ion—that is, when the n-carbon is secondary or terrvnd; m°re PartiClUlarly- When an “-Phe"y! or a-vinyl group is present,, eedless to say, the substrate must also have one or more /3-hydrogen atoms., the comD«VS h‘, ‘"““l S'°W heteroIysis <rate «™tant *„„) is irreversible,, leads m*, Umm0‘eCular substitution and elimination reactions, leads to the rate expression for elimination:, (rate)elim = kh.t(KX), , ^_, , ks + kE, • Hughes, Ingold, „ a,., J. Chem. Soc, 1937, 1271-1285; 1940, 2038-2093., , (1)
Page 490 :
474, , Beta-elimination Reactions, , where ks and kE are the rate constants for the substitution and elimination reac¬, tions which consume the carbonium ion. This expression emphasizes the close, relationship between the rates of elimination and heterolysis and tells us that, any environmental factor that affects the rate of heterolysis must also affect, the rate of elimination to about the same extent. Thus, the E\ reactions of, alkyl halides will be faster in polar than in nonpolar solvents and should be, accelerated by addition of noncommon-ion salts., The proposed mechanism for unimolecular elimination requires that the, elimination to substitution ratio associated with a given alkyl group in a sub¬, strate be independent of the leaving group; although RBr, RI, and RSMeJ, react at different rates in a given medium, the ratio of substitution product, to olefin should be the same for each. As we have already seen (p. 263), this, is true only to a first approximation, for changing the leaving group sometimes, introduces minor but unmistakable variations in the .SW-El ratio/ It therefore, appears that in some cases the leaving group remains close enough to the, carbonium-ion intermediate to influence slightly (in a manner not yet clear), the mode of subsequent reaction., Broadly speaking, unimolecular elimination, even more than unimolecular, substitution, is favored by branching at the /3-carbon atoms. Table 12-1 lists the, percent olefin produced in the unimolecular solvolysis of a number of alkyl, chlorides (“80 percent” alcohol at 25°).5 Note that at one end of the scale,, , Table 12-1. Percent Olefin Formed in the Solvolyses of Some, Percent, Olefin, , Alkyl Chlorides, , Alkyl Chlorides, , Percent, Olefin, , H, , CH3CH2C(CH3)2C1, , 16, 34, , (CH3)2C—C(CH3)2C1, (CH3)3C—C(CH3)2C1, , (CH3CH2)2C(CH3)C1, , 41, , 0-Pr)2C(CH3)Cl, , (CH3CH2)3CC1, CH3CH2CH2C(CH3)oC1, , 40, 33, , (CH3)3C—C(C2H6)2C1, , (CH3)3CC1, , 62, 61, 78, 90, —, , /-butyl chloride (in which none of the /3-carbons bear aikyi suosiw, solvolyzed with only 16 percent elimination, whereas t-BuC(Et)2Cl (in wh.ch, there are five methyl groups bound to U-carbons) is solvolyzed wtth 90 percent, elimination. The yields of olefins from the remaining chlorides (which have on, t For example, in 80 percent EtOH, the 5„1/£1 ratio for '-BuC1 a‘, ‘sob'dX'by, that for ,-BuSMef but the ratios for the corresponding t-anayl ,denvat.ves (at 0 », bJ, about 30 percent. Likewise, the *1 /E\ ratios for t-AmBr and t-Aml (at 25 ) are the same,, are about 35 percent greater than the ratio for !■_AmU, *, , (a), , Brown and Fletcher,, , 75, 10 (1953)., , J. Am. Chem. Soc.,, , 72,, , Berneis, W,, W
Page 491 :
Unimolecular Eliminations, , 475, , to four alkyl groups bound to /3-carbons) lie between these extremes. Without, looking too closely at the intermediate values, we might suspect that this trend, in olefin yield is steric in origin. We have already seen that Syl reactions of, crowded substrates are subject to steric assistance (p. 279) since strain is released, when the leaving group departs in the rate-determining step, allowing the bond, angles about the a-carbon to increase from about 109° (in the substrate) to, 120° (in the carbonium-ion intermediate). Extending this reasoning, it might, be supposed that the fate of the carbonium ion, once formed, is also subject, to steric influences; for if the carbonium ion loses a /3-hydrogen (elimination),, the bond angles about the /3-carbon also increase, and the molecule becomes, even less crowded, whereas if the carbonium ion becomes bound to another, nucleophile (substitution), crowding is again increased. It has thus been argued, that steric crowding should not only favor 6^1 reactions over Sn2, but should, also favor E\ reactions over Sn\.6, Yet, this cannot be the entire story, for there are a number of cases where, an increase in crowding results in a slight decrease in the E\ /Sn\ ratio. We, see, for example, that Et3CCl gives slightly less olefin than Et2CMeCl, that, w-PrCMeoCl gives less olefin than EtCMe2Cl, and that /-BuCMe2Cl gives a, little less than t-PrCMe2Cl; whereas considering only preferential steric assist¬, ance to elimination, we would predict the opposite in all three cases. Quite, obviously, the elimination to substitution ratio associated with a carbonium, ion is being influenced by another factor, which may best be understood by, briefly digressing to the related problem of orientation in E\ reactions., Although only a limited number of cases have been studied, it now seems, clear that when two different olefins may result from an E\ reaction, the olefin, bearing the larger number of alkyl substituents will predominate. (To this statement, must be added the familiar and conveniently vague reservation, “In the absence, of complicating effects.”), , Typically, when /-amyl bromide,, , EtCMe2Br,, , is, , solvolyzed in 80 percent ethanol, roughly four fifths of the resulting olefin is, compound II (with three alkyl groups) and the remaining fifth is compound, ill (with but two alkyl groups).6, CH3, CHa—CHo-C—Br -° perccnt ElOH ., CH3, CHa, , /, , f-AmOH, , CH3—CH=C, , + ch3—ch2—c=ch2 +, CH,, , ,Dh, , H, , fc, , CH3, , 11, , III, , (32perCent), , (8 percent), , Uhar, Hughes, and Ingold, J. Chem. Soc., 1948, 2065., , , /-AmOEt, , (60 percent)
Page 492 :
476, , Beta-elimination Reactions, , This preference for the “more substituted” olefin brings to mind the evidence, that from heats of combustion and hydrogenation the stability of a double bond, may be significantly increased by alkyl substitution on the double-bonded, atoms (p. 49)—a stabilization which we attributed to hyperconjugation. More¬, over, since C, C, , H hyperconjugation (IV) is presumably more important than, , C hyperconjugation (V), a methyl group, with three hydrogen atoms avail¬, , able for hyperconjugation, is a more effective “stabilizer” than other alkyl, , \, , \, , H—C, , H+, , / \, , /, , C=C, , /, , C, , / \, -/, c—c, / "\, , <—>, , \, IV, , \\, , /, , /, , \, , o, o, , /, , I, , \, /, —c—c, , /, \, —C+ c, /, l\, -/, c—c, / ”\, , V, groups with fewer available hydrogens., On this basis, we can understand why Et2CMeCl, which yields principally, olefin VI, having two methyl groups in hyperconjugation with the double bond,, exhibits a somewhat higher E\/Sn^ ratio than Et3CCl, which yields olefin VII, with only a single methyl group in hyperconjugation. Similarly EtCMe2Cl,, CH2CH3, , ch2ch3, , I, , /, , ch3ch2—c—ch3 -» ch3ch=c, Cl, , ch3, VI, , ch2ch3, CH3CH2—C—CH2—CH3 -> CH3CH=C(C2H6)2, Cl, VII, , Which forms a triple-methylated olefin, exhibits a slightly larger E\/SN\ ratio, than n-PrCMe2Cl, which forms a double-methylated olefin. It is thus reason¬, ably clear that the degree to which a carbonium ion may undergo elimination, rather than substitution depends both on steric and hypereonjugative effects,, although there is some disagreement concerning the relative importance of the, two types of effects.7, , .
Page 493 :
477, , Unimolecular Eliminations, , Although orientation in elimination by the E\ mechanism is generally, determined by hyperconjugation, here again there are cases where steric effects, are evident. For example, in the mixture of olefins resulting from the solvolysis, of brosylate VIII in acetic acid, olefin IX (an “expected” product) predomi¬, nates, olefin X is a minor product, and olefin XI (the cis isomer of IX) is present, in only trace amountsA(o) The very low yield of olefin XI is almost certainly due, HOAc, , t-Bu—CH2—CHMe, OBs, VIII, /-Bu, , H, C=C, , /-Bu—CH,, , /-Bu, , +, , Me, , c=c, , c=ch2 +, , H, Me, IX, 75 percent, , H, X, 24 percent, , H, H, XI, 1 percent, , to the operation of a steric effect, for a scale model of this isomer shows that there, is significant steric interference between the /-butyl and methyl groups situated, , cis to each other, and it is very likely that the transition state leading to olefin, XI is, despite hyperconjugation, somewhat less stable than that leading to olefin, X.s Moreover, with the very crowded chloride XII, loss of HC1 in the direction, dictated by maximum hyperconjugation must yield olefin XIII, in which one, of the methyl groups lies cis to the /-butyl group. It is therefore not surprising, that this elimination leads predominantly to the 1-olefin, XIV, yielding only, minor amounts of the more crowded olefin, XIIIA<«-s<6>, /-Bu, /-Bu, , Me, , c=c, , CH2—C(CH3)2—Cl, , /, XII, , \, , H, Me, XIII (minor), , /-BuCHo, , H, , \, , +, , ,, , C=C, , /, , Me, H, XIV (predominant), , ,n support °f thc ,at", substitution reactions of any of the chlorides in Tahl I? i f, Slg"lh,cant stenc assistance in tb, the present author that ifSteric^crowdintt no7?, the IaSt two' lt is the opinion o, I'kely to be important in the carbonium fon d° T ant,in a Partmular alkyl halide, it is no, generally less crowded than the halide from which itT derived^6, 77., , 3607, , 0 9T5).and °kam0t°- ^ ^ ^, , 36,9 O933)- W Brown and Moritani,, , a, , strate through, transition states leading to these olefins Later, state in the El reaction^ rather ^^0 that, 1011, , ^ Carb°nium ion 1, , i *n, of th, , 1S’m the larSe majority of cases, justified (p., , 16, , UU., , “‘f1"5 d'riV'd fr°m a Stagle sub, ‘ ? the Same °rder as the respectiv<, that the geometrV of, transitior, , )^ ^ °lefln ^ that SUch an assumP, , 494
Page 494 :
478, , Beta elimination Reactions, , Bimolecular Elimination. The E2 Mechanism70, An important diagnostic feature of eliminations by the E\ mechanism is that,, aside from salt effects, their rates are not affected by the addition of bases. On, the other hand, as the reader is probably aware, there are a large number of, olefin-forming eliminations that are accelerated by base, many of these being, exceedingly slow in the absence of base. Most eliminations of this type which, have been investigated kinetically exhibit second-order rate laws—first order, each in substrate and in base, B, rate = /^(substrate) (£:), , (2), , Two simple mechanisms are consistent with this rate law. In the first, a, one-step or “concerted” mechanism, the base removes the /3-hydrogen while, the leaving group departs from the a-carbon., , -X, , ->-, , BH+, , +, , C=C, , /, , +, , \, , X, , (3), , In the second, the base removes the /3-hydrogen to form a carbanion which,, after a significant length of time, loses the leaving group to form the olefin., , B: + H—C—C—X, , II, , -;q—q—x, , *“, , II., , ^C=C^ + :X, , /, , (4), , \, , carbanion, There is no evidence that measurable amounts of carbanion form during the usual, elimination reactions. It may therefore be presumed that if the second (“twostage”) mechanism operates, the carbanion must be either rapidly reconverted, to the substrate (making the initial step reversible) or else rapidly converted to, the olefin., One of the substrates most likely to undergo elimination by preliminary, reversible carbanion formation is /3-phenylethyl bromide, PhCH2CH2Br; for, as we have already emphasized (p. 370), the ionization of a C—H bond to, form a carbanion is very much facilitated by a benzene ring bound to the carbon, atom. If the first step in the base-catalyzed conversion of this halide to styrene, were the reversible formation of the anion ~:CHPh, , CH2Br, this conversion,, , SSSSSSSA-
Page 495 :
Bimolecular Elimination, , 479, , when carried out in D20 or in a deuterium-labeled alcohol, should be accom¬, panied by the incorporation of deuterium into the unconverted alkyl halide,, and thence into the styrene formed., PhCH2CH2Br, , :CHPhCH2Br ^ PhCHDCHoBr, :CDPhCH2Br —PhCD=CH2, , However, it has been found that when this conversion is brought about by, the action of NaOEt in EtOD, both the unreacted bromide and the styrene, formed are free of deuterium.“(o) We may then rule out elimination via, reversible carbanion formation for this alkyl halide; and similar studies with, Et2CD—CH2Brn(6) and with the tetra-alkylammonium iodide XVn(c) allow, , cht-ch2-n(ch3);, , I~, , XV, the same conclusion to be drawn for these substrates also. On the other hand, it, is quite likely that the base-catalyzed elimination reactions of trialkylsulfonium, salts or of substrates having —CN, —N02, or —C— groups bound to the, , O, /3-carbon atom do proceed through reversible carbanion formation, for such, substances are known to be very readily converted to carbanions in basic, media (p. 361).1S, The possibility of an initial slow formation of a carbanion, followed by a, rapid conversion of the carbanion to an olefin, is not easily dismissed. Such a, mechanism would be very similar in character to the one-step concerted, elimination process (2); a reliable distinction between the two mechanisms may, be drawn only by making available to the carbanion intermediate (if such, exists) some fate other than immediate conversion to the olefin—that is, by, “diverting” the carbanion. The fact that such a “diversion” has not yet been, accomplished leads us to believe that in most cases a carbanion as such is not, ormed; and, as we shall presently see, the stereochemistry of base-promoted, , (iw). w!'hoITu, , 2SZ&Z&, ™ wS'imT,1'% 1T5)'.<J!Hi"* *'■*.*“•76>5129, , promoted elimination is accomZ H, and Fix, ibid., 75, 2647 (1953).P", , ( n, )- 00 °" 3 Smgle instance in which baseY * Sma amount of hydrogen-exchange, see Cristol, , (iMV))1: tt,, PhCH2CH2SMe+Br- results in a neelirihl, K, process in the rate-determinine sten invol, , #—, , zraga\j-j, M-cw, base-catalyZed elimination, , tc~.tS0°Pe, , ^1612, , reaction of, indicating that the activation, , substrate. This result suggests although itdoeT^t ? n° StJe;ching of the C~S bond in the, carbanion intermediate that is rapidly converted toolefim, , P‘ ^ ^ intervention of a
Page 496 :
480, , Beta-elimination Reactions, , elimination reactions likewise suggests that, except for a few special substrates, a, concerted, rather than a stepwise, process operates., Whether or not a carbanion intermediate intercedes, eliminations that, require the participation of an ion (or molecule) of-base are bimolecular and their, mechanism is designated E2. This path is the most usual one for olefin-forming, eliminations, embracing not only many of the dehydrohalogenations of primary,, secondary, and tertiary alkyl halides,i0(o) but also the important “Hofmann, elimination” of tetra-alkylammonium hydroxides,/0(6), , HO<:rTpUH^^C!^NMe3 ->, and in addition,, , HOH, , + \;=(/ +, , NMc,, , the base-promoted elimination reactions of less familiar, , substrates such as tetra-alkylphosphonium salts, R4P+T_,/s(o) and alkyl sulfones,, , O, RCH-.—CHR'—S—R.;s(6) The E2 mechanism is favored by strongly electron-, , I, , O, attracting groups in the substrate (for example, —NR£ and —S02R), for these, ease the removal of the /3-hydrogen. It is also favored by attachment of a, benzene ring to the /3-carbon, for this not only increases the acidity of the, /3-hydrogen, but also stabilizes the resulting olefin.^ However, the manner in, which alkyl substitution and branching in the substrate affect bimolecular, elimination is not completely straightforward, for the direction of such effects, depends upon the nature of the leaving group and, to some extent, on the, identity of the attacking base. This question is best considered along with the, orientation problem in such reactions., , Orientation in £2 Reactions. Hofmann- and Saytzeff-like Eliminations, It has been recognized for many years that two distinct empirical rules are, needed to summarize orientation in base-promoted eliminations. When more, than one olefin may form from a tetra-alkylammonium or Irialkylsulfomum sail, the, olefin bearing the smaller number of alkyl groups will predominate, whereas, in situations where a pair of olefins may result from the dehydrohalogenauon, of an alkyl halide, the olefin bearing the larger number of alkyl groups will often, (but not always) predominate. Bimolecular eliminations of ’on,urn salts ar, said to proceed according to the Hofmann rule,«<•> whereas those el.mtnat.o, m (a) Fenton and Ingold, J. Chm. See., 1929, 2343; Hey and Ingold, ibid , 1933. 531-, , (6) Fenton and Ingold, iiirf., 1928, 3127; 1929,, ,, « Ingold, el al„ J. Chrn. See., 1927, 997; 1948, 2072, , ,, , « (a) Hofmann, Ann., 79, 11 (1851). <») Saytzeff, .bid., 179, 296 (1875).
Page 497 :
Orientation in, , E2,, , Reactions, , 481, , that lead preferentially to the “more substituted” olefin are said to be governed, by the Saytzeff rule.15™ Typical Hofmann- and SaytzefT-type eliminations are, illustrated below, together with an “atypical” dehydrohalogenation in which, the “less-substituted” olefin predominates over the “more-substituted” olefin, in the product:, Et—CH—SMe2, , °E-—> EtCH=CH2 + MeCH=CHMe (Hofmann)16W, (74 percent), , (26 percent), , Me, -> EtCH=CH2 + MeCH=CHMe (Saytzeff), , Et—CH—Br, , I, , (29 percent), , (71 percent), , Me, * EtCH=CH2 -f- MeCH=CHMe (Hofmann)-16™, , Et—CH—Br, , I, , (73 percent), , (27 percent), , Me, We have already encountered Saytzeff-type elimination (although we did, not refer to it as such) in considering unimolecular eliminations; for it was, emphasized that E\ reactions, regardless of the nature of the substrate, will, yield mainly the olefin having the larger number of alkyl groups. It seems likely, that the reasoning used at that point to account for preferential formation of, the more highly substituted olefin also applies here. Once more it might be, argued that since the olefin having the larger number of alkyl groups (attached, to the double-bonded carbons) is—because of hyperconjugation—the more, stable, the activated complex leading to that olefin should also be the more, stable. We are here assuming, in effect, that all factors aside from the stability, of the double bond being formed are of secondary importance., Such an assumption is obviously not valid where Hofmann-type elimination, is observed, for here the less stable olefin predominates. There is some disagree¬, ment as to the reason for this. Until recently, it was generally felt that an, elimination reaction would proceed according to the Hofmann rule if its direction, , were determined by the acidity of the proton being removedThus, it was argued of the, five /3-hydrogens in a substrate such as XVI, the two on the methylene group are, H, , H, , R—c—CIIX—G—H, *|, , «, , H, t, , ., , less acidic, , (1956)., , 10, H, , +, (XVI: R = alkyl; — X = —NMe3, *—SMe2), , T ., more acidic, , f”/, W Dhar' HuSh«, Am. Chem. Soc., 78, 2193, , waErow„d’ Montam>, Moriuni ^d, Ok'’ ‘948’, and Okamoto, J., , V ;, , 2M3;'See, for example, I„gold, Ref. 2(b), PP. 429-434, and Dhar, e< W,, , Ota,, a,, 194.,
Page 498 :
482, , Beta-elimination Reactions, , significantly less acidic, than the three on the methyl group, due to the acid¬, weakening inductive effect of the alkyl group, R— (p. 205). This being the case,, the removal of a methyl hydrogen by the attacking base is more likely than the, removal of a methylene hydrogen, and Hofmann-type orientation results., Extending this reasoning, we may make the more general statement: Alkyl, substitution at the /3-carbon of an ’onium salt reduces the acidity of the remain¬, ing /3-hydrogens and retards bimolecular elimination at this site. This statement, should apply not only to different sites in the same substrate (in which case it, becomes a principle governing orientation), but also to different substrates, being treated under equivalent conditions.18,19, The suggestion that Hofmann-type elimination stems, at least in part, from, inductive effects is not unreasonable, for we have independent evidence that the, rates of E2 reactions are subject to inductive influences. It has been found, for, example, that the specific rates of the E2 reactions of ring-substituted 2-phenylethyl bromides (ArCH2CH2Br) may be correlated with Hammett’s a values for, the ring substituents, which, as we have seen, are measures of the electronattracting or electron-donating power of these substituents.20 The specific rates, of the E2 reactions of the corresponding dimethylsulfonium bromides (ArCH2CH2SMe^Br-) may be similarly correlated, and as may be expected, the, reactions of the sulfonium salts (for which p is found to be 2.7) are more sus¬, ceptible to polar influences than are those of the alkyl bromides (p = 2.1)., However, the difference is a small one. From it we would certainly not predict, that when steric and hyperconjugative effects are brought back into the picture, by alkyl substitution at the 0-carbon, inductive effects continue to dominate, for the sulfonium salts but not for the alkyl bromides. Yet, this, in effect, is the, view that we must defend when we assume the Hofmann rule to arise prin¬, cipally from inductive effects, for Hofmann-like eliminations are observed for, sulfonium salts but not (except under special circumstances) for alkyl halides., Reexamining the question, it is to be noted that the leaving groups from, , substrates, , that, , undergo, , Hofmann-type, , elimination
Page 499 :
483, , Orientation in E2 Reactions, , _NMe+) are decidedly bulkier than halide ions, which depart during the course, of Saytzeff-like eliminations. It has, in fact, been found that the fraction of, “Hofmann product” in the mixtures of olefins resulting from the E2 reactions, of 2-pentyl derivatives progressively increases as the bulk of the leaving group, increases.^, , X, , Et, , I, , OE,-, , H, \, , Et, , /, , EtCH2—CH—CH3-> PrCH=CHa +, , \, , C=C, , /, , +, , /, , C=C, , \, , ', , H, , Me, , /, Me, , \, , H, , H, , 1-pentene, , —X, 1- pentene, 2- pentenes, , —Br, , —SMc+, , 0.45, , 6.7, , 2-pentenes, -SOoMe, —NMe+, 7.7, , ca. 50, , Furthermore, we now know that the E2 reactions of alkyl halides can be made, to exhibit Hofmann-like, rather than Saytzeff-like, orientation by sufficiently, increasing either the degree of branching in the substrate or the steric require¬, ments of the attacking base./e(c), , Thus, when the tertiary bromides listed, , below are dehydrohalogenated under conditions that insure bimolecular elimi¬, nation (concentrated KOEt in dry alcohol at 70°), the ratio of 1-olefin (“Hof¬, mann product”) to 2-olefins (“Saytzeff product”) increases with branching, as shown below:*2, , EtCMe.Br, , (n-Pr)CMe2Br, , (f-Bu)CMe2Br, , (n*o-C6Hu)CMe2Br, , °'43, , 1 00, , 1-17, , 6.1, , 1-olefin, ^olefins, , The effect of increasing the bulk of the attacking base may be even more, striking. In the mixture of olefins resulting from the action of alkoxide bases on, the bromide Me2CHCMe2Br, the ratio of 1-olefin to 2-olefins has been found to, range from 0.25 (when OEt, , is used) to 11.4 (when EtsCO“ is used)."<“>, , f •' 78’,21" <1956)- In‘h‘= *b“"“ of quantitative, bulkiness of these groups increases as branching is incr^ldTtlhe'aSm thatTTth 7", , very smaU snd^Uie^arTder Waal^^radius^^he^methT^611^, of an oxygen atom. The —0S02C6H CH ft " 'l, gr°Up, , °f f°rmally P^ive^ulfurts, considerably greater than that, , branched,” and the ratio of 1-pentene^to^Z-nentenlg^6 T^’ “ thc present ^nse, “single, tosylate with ethoxide is 0.97. While it is true that the^t ^ 10m the ^ reactlon of 2-pentyl, most of its bulk is in the benzene ring and the *-CH 1, Y uhaS considerable bulk,, far removed from the remainder of the substrate, 3 S °P W 1Ch molecular models show to be, " Brown. Morilani, and Nakagawa,, , J. Am. Ch,m. Soc., 78, 2190, 2197, 2203 (1956).
Page 500 :
484, , Beta-elimination Reactions, EtO~, , /-BuO-, , EtCMeoO-, , Et3CO“, , 0.25, , 2.7, , 4.3, , 11.4, , 1- olefin, 2- olefins, , Similarly, the ratio of 1-olefin to 2-olefins in the mixture obtained from the E2, reaction of /-amyl bromide is 0.43 when the attacking base is OEt~ but 7.7, when the attacking base is Et3CO~.i6(c), To account for these trends, each of which indicates a shift from Saytzefflike to Hofmann-like elimination as steric crowding is increased, let us compare, the transition states for the two modes of reaction. According to our present, picture of the activated complex in the usual E2 reaction, the five most import¬, ant atoms involved in the elimination lie in or near a common plane. These, are the “attacking atom” of the base, the /3-hydrogen, the a- and /3-carbons,, and the departing halogen (or oxygen, sulfur, or nitrogen) atom. Moreover,, the attacking base and the leaving group are situated trans (or “anti”) to each, other. (We shall shortly review the evidence for this picture.) Such an activated, complex may be represented as shown in Figure 12-1, or, more conveniently., , Fig. 12-1. Activated Complex in an E2 Reaction, , using the Newman convention (p. 73), as XVII. In the latter, it will be re¬, called, the bond between the a- and /3-carbon atoms is placed perpendicular, to the plane of the page.*3 Suppose now that the substrate is RCH2—C(CH3)2., , « The view of the activated complex is an E2 reaction shown in Figure 12-1 is commonly, abbreviated in two ways:, , In the present text, the first of these, rather than the second, will be used, since it indicates the
Page 501 :
Competition between Elimination and Substitution, , -, , 485, , The activated complex for Saytzeff-type elimination (in which a methylene, hydrogen is removed) is represented as XVIII, whereas that for Hofmann-type, elimination (in which a methyl hydrogen is removed) is XIX. From the, “Newman pictures” of these (XVIIV and XIXr, respectively), it is evident, that steric interference might arise between groups R, , and X, , if either or, , B:, , I, , H, , B:, , B:, , \, , \, , H, , Me, , XVII, B:, , B:, , i, H, , i, H, , both become bulky. Under these conditions, transition state XIX (in which, R— is free to move out of the way of X—) will be favored, and Hofmann-type, elimination will be observed. Similarly, if the attacking base, B:, is excessively, bulky, it will interfere more with group R— in XVIII than in XIX, and again, Hofmann-type elimination will be favored., In summary then, there can be scarcely any question that steric effects, markedly influence orientation in E2 reactions. It is possible that inductive, effects also play a part; but except for a few special cases (see, for example,, Ex. 6d), this role seems not to have been unequivocally demonstrated., , Competition between Elimination and Substitution, Lhly“ °f °'efi" “ a. Siven PreParati°n will depend largely upon how effeceeneLl, , ellm‘nation reaction competes with the SK reaction(s) that, , ehrninaf "“TT-™S qUeSti°", considered when both, on and subst.tut.on are unimolecular; for under these conditions the, - 'Y, , 15 Very nearly mdependent of the identity of the nucleophilic leaving, , ,£££* ofncMfusingThTdthedltaer!e, dotted lines (reprae„,i„g bonds, , *°, , ^ and since, °f«"*«>
Page 502 :
486, , Beta-elimination Reactions, , group (p. 474), and, in practice, very nearly independent of the concentration, and identity of bases present in solution.** The variations in the £1 /SN1 ratio, with structural changes in the substrate may, as we have seen, be easily inter¬, preted, largely in terms of hyperconjugative effects. The problem becomes, more complex when unimolecularity is no longer guaranteed (that is, in the, presence of moderate or high concentrations of base); for now E/SN ratio may, depend upon the identity of the attacking base, the alkyl group (R—) of the, substrate (RA), and the leaving group (—X). Moreover the possibility exists, that a change in substrate or a variation in reaction conditions may bring about, a change in the predominating mechanism for elimination, for substitution,, or for both., Let us begin with the alkyl group, R—. It has been emphasized (p. 275), that, due to the operation of steric and ponderal effects, branching at either, the a- or the /3-carbon generally retards Sw2 reactions. At the same time, how¬, ever, branching at the a-carbon accelerates E2 reactions; for an alkyl group on, an a-carbon is in a position to stabilize the double bond of the resulting olefin, (and presumably the double bond being formed in the activated complex), through hyperconjugation.*5, , r, , —CH, , I, , I, , 0, , «|, , —CH—C—X, , -HX, ->, , i, , / i, , C—H, , \, , L/, , c=c, , /I, , C, , ^, , V, , /, , c—c, , H+, , / \, , j, , It follows then that the E2/SN2 ratio must increase with branching at the a-carbon atom—that is, this ratio tends to be greatest for tertiary halides and ’onium, salts and least for primary halides and ’onium salts. It is less safe to generalize, concerning the effect of branching at the /3-carbon, for although hyperconjuga¬, tive effects arising from such branching tend to accelerate elimination, steric, (and perhaps inductive) effects tend to retard it. Data are not plentiful, but it, appears that in those cases where /3 branching retards E2 reactions, it retards, SN2 reactions even more effectively. Thus, ft substitution, like a substitution, gen¬, erally favors bimolecular elimination at the expense of substitution., When a substrate is undergoing bimolecular substitution and bimolecular, elimination at the same time, the E2/SN2 ratio should be independent of the, « It is, of course, conceivable that the addition of a weakly basic but strongly, species (such as, , Nf), , would lower the, , E/SN, , nucleophilic, , ratio by diverting some of the carbonium-, , intermediate (in this case to an aliphatic azide). However, such species, except, , y acci e, , ,, , not generally present when elimination reactions are carried out, already, - The effect of « substitution on the rates of £2 reactions of sulfomum sal s h«, £, been, , noted, , (Ref., , 20)., , Similarly,, , the specific rates of the El reactions of, , (CH3)2CHBr, and (CH3)3CBr (with NaOEt at 25 ) he in the ratio, Refs. 6 and 16b)., , ., , ., , (, , »
Page 503 :
Competition between Elimination and Substitution, , 487, , concentration of the added base, B:, for the rates of both reactions are propor¬, tional to the product (RX)(B:). When both the elimination and the substitu¬, tion reactions are unimolecular, the E\/Sn\ ratio is also independent of (B:),, indicating that the carbonium ion intermediate suffers attack by solvent rathei, than by B:. However the E2/SN2 ratio is not (except by extraordinary coin¬, cidence) equal to the E\/SN\ ratio for the same substrate, and it may then be, asked whether the transition from uni- to bimolecularity as the concentration, of base is increased favors elimination or substitution., Again, considerably more data would be desirable, but it appears at present, that the E/Sn ratio may be either raised or lowered, depending upon the, nature of the base involved. If the base is weak but is strongly nucleophilic, toward carbon (for example, PhS-), it will be very effective in bringing about, bimolecular substitution (where attack on carbon is required), , but far less, , effective in bringing about bimolecular elimination (where extraction of a, proton is required). Hence, when such a base is gradually added to a solution, of the substrate, it will soon become involved in bimolecular substitution,, lowering the, , E/Sn, , ratio. At much larger concentrations of base, bimolecular, , elimination may occur at an appreciable rate, but the substitution will benefit, more from the change in molecularity than will the elimination. Quite obviously,, it would be foolish to use such bases as PhS-, N^, and CN~ as “catalysts” in, the preparation of olefins., With very strong bases, the reverse is true. As hydroxide or alkoxide is, added to a solution of a secondary or tertiary halide undergoing unimolecular, elimination and substitution, it is the elimination reaction that first changes its, molecularity. Substitution becomes predominantly bimolecular only at higher, concentrations of base and benefits less from the change in mechanism than does, elimination. Typically, for i-PrBr at 55°, the elimination to substitution ratio has, been found to be 0.03 in dry alcohol but 3.8 in 2.0 N NaOEt; the corresponding, ratios for *-BuBr are 0.37 in alcohol and 13 in 2.0 N NaOEt.** It is thus apparent, why synthetically useful conversions of alkyl halides to olefins in solution gen¬, erally employ strong bases at high concentrations., For each-secondary or tertiary halide, there should be a range of base concentratton within which elimination with strong base is bimolecular but sub¬, stitution is unimolecular, and within which the E/SN ratio will depend on the, concentration of base. This range of concentration depends upon the base, mp oyed, the polarity of the medium, and the nature of the substrate In par, , ughes, I^gold, Masterman, and MacNulty,, , J. Chem. Soc., 1940,, , 899.
Page 504 :
488, , Beta-elimination Reactions, , are treated with dilute (about 0.05 N) NaOEt, although the tertiary halide, gives more olefin both in the absence of base and in concentrated NaOEt.*®, In 0.05 N NaOEt, f-PrBr has apparently adopted the E2 mechanism (resulting, in a higher olefin yield) while £-BuBr continues to undergo elimination by the, , E\ mechanism., The elimination to substitution ratio associated with a given substrate, may also be influenced by the polarity of the solvent. There are two distinct, ways in which this might come about. Let us compare, for example, the transi¬, tion states for the Sn2 and E2 reactions of an alkyl halide, brought about by a, negatively charged base (such as hydroxide or alkoxide):, 6-, , |, , 5-, , RO-G-Hal, ', , H-C-C-Hal, i, , +, , (5^2), \, , RCT, , i, , \, , \, , For both reactions, a negative charge originally concentrated on one atom be¬, comes dispersed over several atoms in the transition state. Both reactions should, therefore be favored by a decrease in solvent polarity; but, since the charge in, the E2 reaction is dispersed over the greater area, low solvent polarity will favor, elimination even more than substitution. Suppose further that the reaction is, carried out in or near the region of mechanistic change. A decrease in solvent, polarity will tend to favor the SN2 and E2 reactions in relation to the SN1 and, , E\ reactions (which require separation of unlike charges in the transition state)., This factor will also tend to increase the E/SN ratio, for we have seen that when, the base is strong, the shift to bimolecularity favors elimination over substitu¬, tion. Regardless of which of these two effects predominates, we would expect, the E/Sn ratio resulting from the reaction of an alkyl halide with hydroxide or alkoxide, , to increase as the solvent is made less polar. Typically, the reaction of f-PrBr with, NaOH (at 55°) results in an E/Sn ratio of 1.17 when carried out in 60 percent, ethanol, 1.44 in 80 percent ethanol, and 2.45 in dry ethanol.*7, It is to be noted that the activation process for an E2 reaction requires, , Et3S+ (at 100°) in 60 to 80 perce, difference may be due in part to, in aqueous alcohol the more str<, ion.
Page 505 :
Stereochemistry of Eliminations in Noncyclic Systems, , -, , 489, , the stretching of a rather strong C—H bond, whereas the activation process foi, an Sn2 reaction does not. Although this extra expenditure of energy for elimina¬, tion is eventually refunded, at least in part, by the conversion of a C—C single, bond to a double bond, we should nevertheless expect the energy of activation, for elimination to be somewhat greater than that for substitution. This means, that both elimination and substitution arc accelerated by an increase in tem¬, perature, but elimination is accelerated more; that is, the E2/Sn2 ratio associated, with a given substrate (regardless of type) should increase with an increase in, temperature. This is what is observed/5 Typically, the E2/SN2 ratio for f-PrBr in, 60 percent ethanol is 1.14 at 45° but 2.42 at 100°. A similar differential tempera¬, ture effect should be observed for unimolecular reactions, for the El/S^l ratio, depends upon an analogous competition. Here the carbonium ion may be con¬, verted either to an olefin (requiring the stretching and breaking of a C—H, bond) or to a substitution product (requiring no such stretching or breaking)., Typically, the E1/Sn1 ratio for /-BuCl in 80 percent ethanol rises from 0.20, to 0.57 when the reaction temperature is raised from 25° to 65°., In summary then, bimolecular elimination is facilitated (at the expense of, substitution) by: (a) branching at either the a- or /5-carbon atom, (b) the action, of strong bases at high concentrations, (c) nonpolar solvents, and (d) high, temperatures., , Stereochemistry of Eliminations in Noncyclic Systems. “Eclipsinq”, Effects, We have already noted (without, however, citing supporting evidence) that, , E2 reactions are, except in special cases, trans eliminations; that, is the activated, complexes in these reactions adopt conformations in which the departing nucleo¬, , phile is as far removed as possible fromf the proton being extracted by the attacking base,, (Fig. 12-1)/® Many examples of trans eliminations, some leading to olefins,’, others leading to acetylenes, have been reported;5® of these, we need consider, only a few. The dehydrobrominations of the two diastereomeric stilbene diroimdes XX and XXII give two different a-bromostilbenes, XXI and XXIII, ibromide XX, the meso form, gives a cis olefin (XXI), whereas the d l-di-, , lead to confusion only in the discussions, 7, ^ ThiS dual USaSe is ^ely to, derivatives) in which trans 1,2 substituents need not'C°mP°Unds (Particularly eydohexane, elimination., e in the positions necessary for facile, 1912, 65" 3 Sl,mmary of t,am eliminations in the early literature, see Frankland, J. Chem. Soc.,
Page 506 :
490, , -, , Beta-elimination Reactions, , bromide (XXII) gives the tram olefin (XXIII)." Both conversions, shown below, using two conventions, are stereospecific:, HO“, , Her, , I, , \, H, , Ph, , H, Ph, TT, , H —>, , Ph, \, , /, , G=C, , b/, , Xh, , XXI, ds, , HO-, , \, H, , H, , H, , Ph, , PI, , /, •, , C=C, , /, , \, , Br, , Ph, , XXIII, trans, Similar stereospecificity has been observed in the base-promoted elimination, reactions of compounds XXIV, XXV, and XXVI. Each of these has two, PhCHMe—CHPh—NMe+I-, , PhCHMe—CHMe—OTos, , XXIV, , XXV, Ph—C—CHBr—CPIPhBr, , II, , O, XXVI, adjacent asymmetric carbons and therefore exists in threo and erythro modifica¬, tions (which are closely analogous to d,l and meso forms, respectively, except, that the two asymmetric carbons have two, rather than three, substituents in, common). In each case, the erythro (“mwo-like”) form has been found to give, the cis olefin, whereas the threo (“</,/-like”) form gives the trans olefin. The reac¬, tion of one member of each of these pairs is shown as follows\se-ss, 31 Pfeiffer, Z.physik. Chem., 48, 40 (1904)., ka 74, * (a) Cram, Greene, and DePuy, J. Am. Chem. Soc., 78, 790 (1956). (b) Cram, ibid., 74,, 2149 (1952). (r) Lutz, Hinkley, and Jordan, ibid., 73, 4647 (1951)., ,, 33 It may be asked how we may be certain that the configurations of the reactants an, products in these conversions are as represented. With the olefimc products there is g, little difficulty, for enough cis-trans pairs of olefins have been prepared and charactemed, that we may say with confidence that of-such pairs, the trans compound almost always has, higher melting point, the lower boiling point, the lower solubility in a given solvent, and, lower dipole moment. In addition, spectral evidence is occasionally of use here, For the reactants, the problem becomes more complex. Sometimes a, , form may, , V
Page 507 :
Stereochemistry of Eliminations in Noncyclic Systems, H, , {traits) 32(,a), , II, , (XXIV', threo), , 491, , Ph, , V, , OEt~, , -, , AMe, , Ph, , +NMe3, , H, , (XXV', erythro), , OEA, ->■, , Me, , V, I, AMe, Ph, , {cis), , 32(i), , o, II, , O, H, Et,N, , (XXVI', threo), , G—Ph, , V, , {trails), , 32(c), , /c\, Ph, , Br, , Moreover, it has long been recognized that haloalkenes and related com¬, pounds in which a hydrogen and halogen are trans to each other (for example,, Me, , Br, , c=c, H, , HOOC, , and, Me, , _, , H, C=C, , C1, , are much more rapidly conCOOH>, , verted by base to acetylene derivatives than are the corresponding cis compounds., The overwhelming predominance of trans elimination in most E2 reactions, has not yet been explained in a fully satisfactory manner. It cannot be ascribed, to electrostatic repulsion between the attacking base (which is often negatively, charged) and the departing nucleophile (which frequently bears a partial, negative charge), for onium salts such as XXIV also undergo trans elimination, although the leaving group in the activated complex bears a partial positive, charge. Some workers prefer to regard the change at the a-carbon during the, elimination process as sort of a Walden inversion in which the attacking species, , instances, thecomplete structure ArimnalLn' t ', , into enantioraorphs; and, in isolated, , of diastereomers „!ay allow, T ” b°,h, asymmetric carbons. However since 1950, n of rclatlve configurations about the, erythro pairs have been based upon their methods of a11 J;0nfiSuratl0nal assignments for threoon unsaturated compounds of known configuration SurhS1S’ .genci ally via addihon reactions, the mechanisms, or at least the stereochemical rules ’ f, ^gnments require that we know, treated in the following chapter), ", ^ SU"h addition rcactions (a topic to be
Page 508 :
492, , Beta-elimination Reactions, , is the pair of electrons derived from the /3-carbon atom when a /3-hydrogen is, extracted. On this basis, it might be argued that, as in the SN2 reaction, the, attacking species must approach the reaction center (that is, the a-carbon), from one side while the leaving group departs from the opposite side. However,, the geometry of the activated complex in an SN2 reaction is considerably dif¬, ferent from that in an E2 reaction; and, according to our present picture, the, electrons from the /3-carbon do not “attack” the a-carbon in the same way as, does an external nucleophile but may be said, more accurately, to “drift into, place., , It therefore seems that the analogy between the stereochemistries of, , the E2 and Sn2 reactions is, at best, a very broad one and is in no sense an, “explanation.”, ^Quite naturally, when the members of a erythro-threo pair (or a meso-d.,1 pair), of diastereomers undergo elimination reactions to give different products, their, reaction rates will be different. The threo form of methiodide XXIV, for exam¬, ple, reacts with NaOEt over 50 times as rapidly as the erythro form5^(o), corre¬, sponding to a difference in AF* of 2.3 kcal per mole. Since it is extremely, improbable that the difference can be due to a large difference in stabilities of, the reactants,54 it may be inferred that the activated complex leading to the, , trans olefin is considerably more stable than that leading to the cis olefin. Com¬, paring the two activated complexes (XXVII and XXVIII), we see that the, ~OEt, , OEt, , 1, , H, , H, , Ph', , Q, , Ph, , 1, , H, , Ph, , V, , P\AH ., , Me, , A, , piryfSvie, , +, NMe+, 3, , Ph, , Me, , (NMeJ, , V, , 1, , A, , Ph, , Me, , cis, , trans, , XXVII, , XXVIII, , (less crowded), , (more crowded), , XXIX, , large phenyl groups interfere with each other in XXVIII but not in XX\ II., Thus, the elimination proceeding through transition state XXVII takes place, more readily than that proceeding through XXVIII. The relatively slow reac¬, tion leading to the cis olefin is sometimes said to be due to the “cis effect” or an, , “eclipsing effect” (for as the reaction proceeds from the transition state to the, olefinic product, one phenyl group appears to “pass in front of” the other)., Exactly the same argument may be used to explain why dibromide XX, , ,, , the threo form of benzalacetophenone dibromide (XXVI), undergoes dehy ro« Abd Elhafez and Cram, J. Am. Chem. So,, 75, 339 (1953) have studied1 the, between the threo and erythro forms of the formates related to XXIV (in which the, N, , » g, , ^"“abS S the, , has been replaced by an -OCHO group), and have found AFfor the, be only -0 14 kcal per mole. It seems very unlikely that the difference in sta, corresponding trimethylammonium iodides is very much greater than this.
Page 509 :
Stereochemistry of Eliminations in Noncyclic Systems, , -, , 493, , bromination with Et3N about 100 times as rapidly as does the corresponding, , erythro form;5*(c) the reaction of the latter must proceed through activated, complex XXIX, in which the large phenyl and benzoyl groups are becoming, “eclipsed.”, Eclipsing effects may also determine the predominant product resulting, from elimination from a single substrate when more than one /3-hydrogen is, present. Thus, when 2-bromopentane is treated with KOEt, almost three times, as much trans-2-pentene is formed as is m-2-pentene,*(6) for in the transition state,, XXX, leading to the cis olefin, the methyl and ethyl groups have begun to, “eclipse each other.”, , The same effect is even more marked when the leaving group is very bulky., Ester XXXI, for example, undergoes an E2 reaction (with breakage of the, alkyl-oxygen bond) when treated with t-BuOK. The olefinic product is almost, exclusively fram-stilbene “ The predominance of trans olefins in the products, Et, f-BuO~, ->■, , PhCH2CHPh—O —C, , II, O, , coo~, , Et, XXXI, , resulting from £1 reactions (p. 477) may be explained on much the same basis, Here however, we must consider the activated complexes lying be,weenie, mum ion intermediate and the various olefinic products., '* Curtin and Kellom, J. Am. Chem. Soc., 75, 6011 (1953).
Page 510 :
494, , -, , Beta-elimination Reactions, Ac, , H, , CMC,, , XoTh-M, , ►, , /, , OBs, , H, , Me, , -OBs, , /-BuCH-CHCH, 3, , jC=C, , /, , H, , I, , /CMe8, , \, , Me, , H, , > /-BuCH2CHCH,, HOAc, , H, , (main product), Me, , CMe,, , Me, , )C=C\, , H, , CMes, , H, , H, , H, (trace), , As shown, the methyl and /-butyl groups are becoming eclipsed in the transition, state leading to the cis olefin but not in that leading to the trans olefin, Aside from eclipsing effects of this type, we would perhaps expect unimolecular eliminations to be nonstereospecific, for there is a time lag between, the departure of the leaving group and the extraction of the 0-hydrogen. How¬, ever, there is evidence that some E\ reactions are complicated by neighboring, , group effects that lead not only to stereospecificity but also to the possibility of, rearrangement (see Chap. 14)., , Trans elimination has been observed for other reactions which do not, how¬, ever, involve loss of a proton. Vicinal (a, 0) dibromides are known to react, with iodide ion, yielding 12, Br-, and the corresponding olefin. The reaction is, second order,56, \, , —C—C— + 21-, , I, , /, , I, , /, C=C, , + 12 + 2Br~;, , rate = £2(I-) (dibromide), , (5), , \, , Br Br, and appears to be stereospecific,56(o)-57 for ^w-2,3-dibromobutane and mesostilbene dibromide yield trans olefins, whereas the corresponding d,/-dibromides, yield cis olefins. The stereochemistry, at first glance, seems opposite to that ob¬, served for dehydrohalogenation reactions (where meso dihalides yield cis olefins),, but the apparent contrast is due to the nature of the reactions rather than to, differing stereochemistries; for a dehydrohalogenation involves removal of one,, rather than two, halogen atoms., , Br, , H, , Me, , V, 11, /Cx, Me, , Br, , MCAH, , V, , -A, , ||, , A,, H, , Me, , Br, , H, , cis, , d,l, , « (a) Winstein, Pressman, and Young, J. Am. Chem. toe., , 54., , 61, 164^J^, , 952 (1932); (c) Weinstock, Lewis, and Bordwell, ibid., 78, 6U4Z U M)., , ’37 Young, et at., J. Am. Chem. Soc.,, , 61,, , H, , Me, , trans, , ibid, , Me, , 1640 (1939);, , 65,, , 2099 (19, , )., , (
Page 511 :
Stereochemistry of Eliminations in Noncyclic Systems, , -, , 495, , (As with dehydrohalogenations, the reactions leading to the as olefins are con¬, siderably slower than those leading to trans olefins., The following mechanism is consistent with both the rate law and the, stereochemistry:, , R, H, , sl°™ >, E2, , IBr, , +, , H, , XsG=C/, , /, , H, , I", , +, , I^Br, , I-I, , +, , \, , Br', , R, , +, , Br’, , If this mechanism is correct, an iodide ion extracts Br+ in the rate-determining, step in much the same way as a base extracts a proton in an E2 reaction of the, more usual type. This sequence, however, does not explain satisfactorily why, debrominations leading to terminal olefins (RCH—CH2), , are considerably, , faster than those leading to olefins of the type RCH=CHR,33(6) and it has been, suggested that dibromides such as MeCHBr—CH2Br are first converted to, bromoiodides, which are then attacked by iodide in the manner shown for, dibromides in sequence (6).55 Dibromide XXXII would be more readily conRCHBr—CHoBr —> RCHBr—CH2I -U Br~ + RCH=CH2 + I2, XXXII, , (6), , XXXIII, , verted to a bromoiodide than would, say, 2,3-dibromobutane, since one of the, bromine atoms in XXXII is bound to a primary carbon at which Sn2 reactions, readily occur., , Furthermore,, , the iodine atom, , in bromoiodide XXXIII is, , more susceptible to nucleophilic attack than is a bromine atom, for iodine, readily “expands its valence shell” (activated complex XXXIV), allowing, formation of a new I—I bond before the I—G bond breaks. The initial step in, , :I:, , + :I:C—C—Br, , : I—I—C—C—Br, XXXTV, , .. .., , \, , : I—I: +, , *• ••, , X, C=C, , X, , + Br~, , \, , sequence (6) would result in an inversion of configuration about the terminal, carbon, , but this could not be detected unless the reaction were carried out, , using the deuterium-labeled dibromide, RCHBr—CHDBr (Ex. 7), Still another example of a reaction involving tram elimination is the “decarboxylative debromination” of j8-halo acids.39, », , and,Brader> Am- J- Chem. Soc., 77, 361 (1955)., , ibid., 75,2?39a953), , exception,, , WCurPnlnd, , h, , l*’ ^ 2645 (1953)' W Grovenstein and Lee,
Page 512 :
496, , Beta-elimination Reactions, RCHBr-CHR-COOH, , RCH=CHR + Br~ + C02, , This elimination, if carried out in ethanol or in a-less polar solvent, is stereo¬, specific; anions XXXV and XXXVI yield trans olefins, whereas the diastereomeric forms, XXXVII and XXXVIII, respectively, react more slowly to give, , cis olefins.40, COO', , COO-, , COO‘, , ,4c:, , Y"., , XXXVII, , XXXVIII, , H, , Ph, , "w, , 1, , H^Y^Br, , H Y ^Ph, , Br, , Br, , XXXVI, , Ph, , *, , H, , V, , COO-, , Ph *, , /C\, H, Ph, , H, , /G\, H, , Br, , Y, , Br, , Ph, , *, , H, , Xc/, , Xc/, , II, , II, , /Gx, H, Ph, , /G\, H, Ph, , We may then represent this reaction, which is first order in carboxylate anion,59(o), simply as, , 1 Step-->■, , 0=C=0, , + XC=C/-f, , /, , \, , Br~, , When carried out in water, this reaction tends to lose its stereospecificity. In, aqueous solution, anion XXXVI continues to yield a trans olefin, but anion, XXXVIII yields a mixture of cis and trans olefins with the trans isomer consti¬, tuting about three fourths of the product. It seems likely that in the more polar, medium the reaction proceeds, at least in part, by preliminary ionization of, the C—Br bond to give a “zwitterion” (XXXIX), which then undergoes, Br, PhCHBr —CHBr —COO, , -^*Ph, , ch-^htc, , O, • PhCH= CHBr 4- C02, , XXXIX, <« The fact that anions XXXV and XXXVI both yield l,am olefins even though XXXV, is an erytkro form while XXXVI is a «W form has no, , t^fn““aXed in the, , £ “Te “o4e4hri4”TXh^s^l Won appl.es ,o substrates XXIV, XXV, and XXVI, for with each of these, a thrto compound yields a Irani o e
Page 513 :
Cyclic Systems, , -, , 497, , decarboxylation. In this sequence, the asymmetry about the 0-carbon is lost in, the initial step, and the intermediate is converted principally to the trans olefin,, very probably because the formation of the cis olefin must pass through a transi¬, tion state in which the bulky bromo and phenyl groups are being eclipsed, (p. 494)., , Cyclic Systems. Possibility of Elimination by a Carbanion Mechanism, Thus far, we have considered only noncyclic systems, in which rotation about, the C«—Qj single bond is permitted (although the molecule may prefer some, conformations to others). When this bond is incorporated into a ring (having, fewer than about nine members), rotation becomes restricted. In a cyclohexane, ring, for example, two equatorial substituents on adjacent carbons (which lie, trans, but in an unfavorable position for bimolecular elimination) may be moved, into the axial positions (trans, but in a favorable position for elimination) simply, by “turning the ring inside out”—a process which is permitted, except in special, cases (p. 241). However, cis substituents on adjacent atoms (one substituent, axial, the other equatorial) may not be moved into positions favorable for an, E2 reaction without introducing considerable strain in the ring, and such sub¬, stituents do not participate in the usual type of bimolecular elimination reac¬, tions. We may thus appreciate why menthyl chloride, XL (which has only one, 0-hydrogen trans to the chlorine) yields only one olefin upon treatment with, NaOEt. In contrast, neomenthyl chloride, XLI (with two 0-hydrogen atoms, trans to the chlorine) yields a mixture of two olefins with olefin XLI 11 pre¬, dominating (as would be predicted by considering hyperconjugative effects).^, , XLI, , XLI I (23 percent), , XLI 11 (77 percent), , Furthermore the formation of 2-menthene (XLII) from chloride XLI is over, mes as fast as its formation from chloride XL. Both reactions must pass, isopTotl'grorin XU ", "M H „, o, Lesutke,, , Am.?J«. m (wS?, , '"V110™6 * alda1’ but the ™,hyl and, equatorial, whereas in XL they are axial. Since, ^, , ^ 19M>, , 3839- (»> Huckel, Tappe, and
Page 514 :
498, , -, , Beta-elimination Reactions, , bulky groups prefer to occupy equatorial positions (p. 241), chloride XLI, yields a transition state of lower energy than does chloride XL and therefore, reacts more rapidly., The flexibility of the cyclohexane ring may be very much diminished by, incorporating it into a bicyclic system such as that appearing in compound, XLIY or that in XLV. Both of these chlorides undergo bimolecular eliminations, , at very low rates,^ even though both have hydrogen and chlorine atoms in, trans positions on adjacent carbons; for the rings in both compounds are rigid, and cannot be twisted so that the four atoms involved in the elimination (two, carbons, an a-chlorine, and a /3-hydrogen) lie near a common plane. Moreover,, the trans isomers of XLI V and XLV, in which the hydrogen and chlorine atoms, lie cis to each other, react somewhat more rapidly than XLIV and XLV them¬, selves, although the differences are. not sufficiently large to be attributed with, confidence to any single factor. Since the usual structural requirements for an, E2 reaction cannot be met in these cases, it may be asked how the slow, but, observable, elimination reactions of these compounds proceed. A number of, workers feel that the alternative mechanism for bimolecular elimination, the, “carbanion mechanism” (sequence 4, p. 478), comes into play here. If this is so,, the rate-determining step is the conversion of the substrate to its carbanion,, followed by the rapid loss of Cl- from the carbanion,, H, , LI, , P, , i, c, V, , H, a, , Q., , slow, , Cl, , Cl, , ., , —, , 1, C', , (', , [, , Cl Cl, carbanion, , \, , C=C, , /, \, , H, , and we would therefore anticipate any structural modification that boosts the, acidity of the /3-hydrogen (without introducing other complicating effects) to, accelerate the elimination. Thus, it is not surprising that substitution of the, 4f, , (a) Cristol and Hause, J. Am. Chem. Sac., 74, 2193 (1952). (h) Cristol and Hoegger, ibid.,, , 79, 3438 (1957).
Page 516 :
500, , -, , Beta-elimination Reactions, , Although the action of the sulfone group in accelerating eliminations is, consistent with a two-step carbanion mechanism, it does not demand it; for an, elimination reaction proceeding by the usual one-step mechanism should also be, facilitated by groups that greatly increase the acidity of the /3-hydrogen(s')., Indeed, the two types of mechanism merge into each other as the lifetime of the, carbanion intermediate becomes progressively shorter. For the usual E2 reac¬, tions, that lifetime is experimentally indistinguishable from zero; and even when, eliminations are facilitated by a sulfone group, there is strong evidence that a, carbanion intermediate, if it forms at all, is extremely short lived. It is now, known that the elimination reactions of substrates such as XLVII and L are, catalyzed by amines (as they should be) but not retarded by the conjugate, acids of these amines.46 However, if the initial formation of the carbanion were, reversible, then the concentration of carbanion LI (hence the rate of the overall, , o, j, , o1, , 1, , —S—C H—C—X + R3N:, I, |, , o, , 1, ••, —s—c—c, 1, 1, , o, LI, , reaction) should be inversely proportional to (R3NH+). Since this inverse, dependence is not observed, we may conclude that if carbanion LI is formed, its, conversion to an olefin is much faster than its return to the substrate, a reaction, which itself would be expected to be very rapid. In the opinion of the present, author, the question of a carbanion intermediate in E2 reactions (except, perhaps in eliminations facilitated by nitro, cyano, or keto groups) remains, open.47, , Intramolecular (cis) Eliminations, Of the many compounds that undergo pyrolytic eliminations, at least three, types almost certainly react unimolecularly through cyclic transition states., These are carboxylic esters:, , —C—C—O—C—R, H, , C=C, , O, , + HO—C—R, , O, , *e Weinstock, Pearson, and Bordwell, J. Am. Chem. Soc., 78, 3473 (1956); Pearson and, Vogelsong, ibid., 80, 1048 (1958)., ,■ . ;n rn, * For further arguments in favor of the intercession of a carbanion intermediate in M, reactions promoted by a sulfone group, see Goering, Relyea, and Howe, J. Am Chan. 5 ^, , 79, 2502 (1957). For arguments to the contrary, see Bordwell and Landis, ibid., 7 ,, (19*57); and Shell and McNamara, ibid., 79, 85 (1957).
Page 517 :
501, , Intramolecular (cis) Eliminations, xanthates, (the “Chugaev reaction”):, A, , \, , —C—C—O—C—SMe, , c=c, , /, , I, , /, , + O, , C=S + HSMe, , \, , S, , H, , and trialkylamine oxides, (the “Cope reaction”):, R, , I, I, , I, I, , /, l\, , —C—C—N, H, , \, , /, , c=c, , /, , + R2N—OH (an N,N-dialkylhydroxylamine), , \, , OR, , The decompositions of carboxylic esters and alkyl methyl xanthates at high, temperatures in the absence of solvent obey first-order rate laws and exhibit, negative entropies of activation.45 Thus, motion is more restricted in the transi¬, tion state than in the reactants, and we may suspect that a ring forms during the, activation process (p. 182). However, the strongest evidence in support of a, cyclic transition state is stereochemical in character; for unlike the eliminations, taking place in polar solvents, the elimination reactions of esters and xanthates, are cis eliminations. Typically, the predominant product obtained from xanthate, LI I is olefin LIII, showing that the hydrogen lost in the reaction lies cis to the, xanthate group; whereas pyrolysis of xanthate LV yields chiefly the isomeric, olefin, LIV.49 As shown, there are two /3-hydrogens cis to the xanthate group, , LIV 7 percent, , 4~, , LIV, , 86 percent, , ■n LV, but the predominant olefin is LIV, in which the double bond is con¬, jugated with the benzene ring. Moreover, as elimination may also be observed, (a) O’Connor and Nace, J. Am. Chem. Soc., 74, 5454 (’1952') • 7 4 911ft, mD, Head, and Williams, J. Chem. Soc., 1953, 1715., K, }> 75,2118 (1953>- (*) Barton,, Alexander and Mudrak, J. Am. Chem. Soc., 72, 1810 3194 t195fn, ments in the menthyl and neomenthyl series see Ref 41 (M A, 5, , described by, , Bordwell and Landis, J. Am. Chem., , •, F, , -i, ar CXpen', , Soa. 80, *50 (1958)!“* 'XCepti0nal Case is
Page 518 :
502, , Beta-elimination Reactions, , in noncyclic systems by comparing the olefins formed from diastereomcric:, xanthates.50, Ph, , H, , \ /, C, , II, G, , / \, Me, , Ph, , trans, , (Xan = —O-C-SMe), , II, , S, , Similar experiments point to cis elimination in the pyrolysis of carboxylic, esters. In the pyrolyses of the diastereomeric forms of the labeled acetate, PhCDH—CHPh—OAc,55 for example, the erythro-acetate, LVI, retains almost, all of its deuterium, whereas the threo-acetate (not shown) loses most of its, deuterium. Although both forms may conceivably lose either the /3-hydrogen, or the /3-deuterium, one reaction path involves eclipsing of the two phenyl, groups and is therefore less favored.61, Ph., , ,H, (phenyl groups eclipsed, in the transition state, —less likely), , -DOAc ^, , H, , — HOAc ;, , Vph, , (more likely), , c, p/, , H, , It seems likely that the mechanism for the pyrolytic conversion of esters, to olefins is similar to that for the Chugaev reaction. It is often assumed that, the breakage of the Ca—O bond is synchronous with that of the Q?—H bond;, that is, the transition states for these two reactions may be represented respec, lively as LVI I and LVIII. While these activated complexes account for the, so Cram, J. Am. Chem. Soc., 71, 3883 (1949); Cram and Abd Elhafez, ibid., 74, 5828, , 095 "A, , more detailed analysis of this situation must also consider the fact that a C-D bond, , is more difficult to break than a C-H bond in the corresponding location. This comphcat, does not, however, alter the qualitative conclusion derived from these experime, elimination is very much more likely than trans elimination. For further evidence t, conversion of esters to olefln, involves rir elimination (based lately on reactrom mtb, and triterpenoid series) see Barton, et al., J. Chem. Soc., 1949, 2174, 2459, 1952, 455., Marvel and Williams, J. Am. Chem. Soc., 70, 3842 (1948)., , hc
Page 519 :
Intramolecular (cis) Eliminations, , -, , 503, , H, , +, O, , R, , LVII, , H, , S, , II, olefin + C + MeSH, C-S-Me, , C-S-Me, , O, , O, , LVII I, predominance of cis elimination, they do not explain the small, but detectable,, degree of trans elimination. Assuming that the latter does not arise from isomer¬, ization of the substrate before the elimination or isomerization of the olefin, after the elimination, we must consider the possibility that, in some cases at, least, the breakage of the C—O bond occurs before that of the C—H bond. If, this is true, the reacting system exists, for a significant time interval, as an, intimate ion pair or (less likely)5* as a pair of radicals. During this interval,, molecular rotation may allow a /3-hydrogen to approach the xanthate or, carboxylate ion (or radical) which originally lay trans to it, and to undergo, extraction. This reaction path would almost certainly require a higher energy, of activation than an elimination passing through a cyclic intermediate such, as LVII or LVIII, and cis elimination would therefore predominate, although, trans elimination would become increasingly important at higher temperatures., Conversely, trans elimination would be expected virtually to disappear at, sufficiently low temperatures. For example, at about 120° (a temperature at, which most ester and xanthate pyrolyses are far too slow to follow conveniently),, a number of amine oxides (R3N—O) undergo almost exclusive cis elimina¬, tion,"3^) very probably through a transition state such as LX. Thus, the threo
Page 520 :
504, , Beta-elimination Reactions, , amine oxide, LIX, decomposes to yield a mixture of olefins, 93 percent of which, is the cis olefin, LXI, and 7 percent of which is the “terminal” olefin, LXII., , >, , II, , + Me2N—OH, , Me, LX, , Ph, , LXI (93 percent), , Ph, , I, , Me-CH, , \, , H, , /, , \, , ■NMe,, O, , H, H, , H, , MeCHPh, , H, , "c'', >, , ||, CH2, , 4- Me2N—OH, , LXII (7 percent), , Only traces of the trans isomer of LXI could be detected. Thus, stereospecificity, is more extreme here than in the pyrolysis of the corresponding Mra>-xanthate,50, for the latter yields a mixture of olefins in which the cis to trans ratio is only 3.3., Similarly, the trans to cis ratio resulting from the erythro amine oxide (diastereomeric to LIX) is 30, whereas the same ratio from the corresponding xanthate, is only 9.6SW, , The Possibility of Ionic Reactions in the Vapor Phase, In proposing a mechanism for the trans elimination that almost invariably ac¬, companies the more favored cis elimination in the pyrolyses of carboxylic esters, and xanthates, it was suggested that the initial bond breaking in the substrate, was heterolytic rather than homolytic; that is, the reaction proceeded through an, ion pair rather than a radical pair. It should be noted, however, that before, the 1950’s it was generally felt that ionic organic reactions could not proceed in, the vapor phase. Consider, for example, isopropyl bromide, a substrate that, undergoes heterolysis rather easily in polar solvents. It has been esumated54(o), that dissociation of this compound into Me2CH+ and Br“ in the vapor state, requires an energy input of about 150 kcal per mole, whereas dissociadon into, Me2CH- and Br- requires an input of only 60 kcal per mole.^6) Comparing these, values, it might be argued that in this case, heterolysis may not compete effec¬, tively with homolysis in the vapor phase, except perhaps at exceedingly, , tg, , .< (a) Stevenson, in Streitwie^V «£*£ “£"*^<£^^£5, 614. (b) Farmer and Lossing, Can. J. Chem., 33,, (., ), • i <• tu isopropyl, these two values represents the difference between the ton,sat,on potent,al of the tsopropy, group and the electron affinity of the bromine atom.
Page 521 :
The Possibility of Ionic Reactions in the Vapor Phase, , 505, , temperatures. The situation should be even more extreme for substrates that, undergo heterolysis less readily. However, an argument of this sort overlooks, the fact that much of the energy input presumably necessary for heterolysis is, used in separating the ions after their formation in a medium having a dielectric, constant of very nearly unity (for example, increasing the separation between, o, , o, , two univalent ions of opposite charge from 4 A to 100 A requires an energy, input of about 80 kcal per mole). It may thus be seen that gas-phase reactions, proceeding through ion pairs are not energetically prohibited., , Indeed, there is evidence which suggests (although it does not prove) that, the dehydrohalogenations of alkyl chlorides and bromides in the vapor state, may proceed through a rate-determining conversion to an ion pair—that is,, that they are, in effect, E\ reactions in the vapor state., , 4-., , -c—c, H, , Cl, , C=C, , \, , H, , fast, , /, , + HC1, , \, , ci-, , The study of such reactions is often complicated by free-radical side reactions, that may be accelerated not only by elemental oxygen or halogen, but also by, contact with glass surface. However, when precautions are taken to minimize, these complications, it is found that these dehydrohalogenations are frequently, first-order reactions, many of them giving good yields of olefins.55 The kinetics, are obviously consistent with a one-step mechanism proceeding through a cyclic, transition state, LXIII (analogous to that proposed for the pyrolyses of car-, , c—H, I, , —C—H, , C—Cl, , —C—Cl, , LXIII, boxylfc esters, xan,hates, and amine oxides); but the relationships between, tructure and react.v.ty in these dehydrohalogenations are strikingly similar to, those observed for conventional S„1 and El reactions *» The follol, that, , a, ,J .“w"£££££, , C ton, whereas branching at the 0-carbon is very much less effective•*>«?£.(,949); 46-114
Page 522 :
506, , Beta-elimination Reactions, EtBr, , w-PrBr, , Me2CHCH2Br, , f-PrBr, , ^-BuBr, , EtCMe2Bn, , 1, , 3-5, , 6.3, , 170, , 32,000, , 46,000, , These figures indicate that the activation process in the rate-determining step, , involves partial bond breaking at the a-carbon, but probably not at the /3-car¬, , bon, and that the bond breaking is greatly facilitated by hyperconjugative, , effects arising from a branching. Furthermore, it has been found that bromides, , LXIV and LXV undergo dehydrobromination at almost the same rate, even, , though the reaction of the latter involves the removal of an allylic hydrogen tc, form a conjugated diene., Me—CH—CH—Me, , I, , MeCH=CHMe;, , I, , H, Br, LXIV, CH2=CH—CH—CH—Me, , I, , H, LXV, , CH2=CH—CH=CHMe, , I, , Br, , This again suggests that the /3-hydrogen has not been appreciably loosened, nor, , has the new double bond begun to form, during the activation process of the, , rate-determining step, for both hydrogen loosening and double-bond formation, should be greatly facilitated by the y,8 double bond in LXV., , Finally, let us compare (Table 12-2) the energies of activation, EA, for, a number of these gas-phase eliminations to two sets of bond-dissociation ener¬, gies: first, to the energies needed to split the various substrates into free radi¬, , cals;57 and, second, to the energies needed to break them into ions.54(o) We see that, , Table 12-2. Comparison of Activation Energies for Vapor-phase, Eliminations with Homolytic and Heterolytic, Bond-dissociation Energies, EtBr, , /-PrBr, , /-BuBr, , EtCl, , f-PrCl, , /-BuCl, , 60.2, 80.9, 192, , 41.4, 78.3, , 183, , 63.8, 132, , 50.5, 82.2, 168, , 158, , AE (for R.T —* R+ + X~), , 47.8, 67.6, 156, , 42.2, , AE (for RT —> R' + X'), , 53.9, 67.2, , £U(elim)(kcal), , 2), , but refer^o the^breakag”of indivHual^onds^n specified compounds^The^eterminahon^, , these values, which is generally moredifhcultthanthe ^^nimation o, discussed by Szwarc, Chem. Revs., 47, 75 (175UL ana oy warn g,, Wiley and Sons Inc., New York, i957, pp. 40-53., , ,Tohn
Page 523 :
Exercises for Chapter 12, , 507, , the trend in activation energies for elimination is similar to that for heterolytic, bond-dissociation energies; that is, both sets of energies decrease as branching at, the a-carbon is increased. On the other hand, the energies of activation bear no, obvious relationship to the homolytic dissociation energies that are greater for the, isopropyl halides than for the ethyl and ^-butyl halides. From these figures, we, would strongly suspect that the activation process in these gas-phase dehydrohalogenations is heterolytic (ionic) rather than homolytic (free radical) in character., , EXERCISES, , FOR, , CHAPTER, , 12, , 1. (a) It has been noted that the replacement of Me— by Et— as group R— in the, R', 1, , tei tiary halide R, , C, , Cl raises the rate at which the double bond enters group, , R", R— in an E\ reaction but lowers the rate at which the double bond enters groups, R^ and R— if the latter two are not altered. Explain., (b) Assume, more specifically, that the replacement of Me— by Et— as group R—, raises the rate at which the double bond enters R— by a constant factor, a; lowers, the rate at which it enters R— and R— by a constant factor, b; but does not, affect the rate at which the carbonium-ion intermediate, RR'R"C+, reacts with, the solvent to give a substitution product. From the following yields of olefin, (80 percent EtOH, 25°, Ref. 5a), calculate the factors a and b:, Me3CCl, , 15 percent, , EtCMe2Cl, , 34 percent, , Et2MeCl, , 40 percent, , (c) Calculate the percent yield of olefin from Et3CCl. (Observed 39 percent), ( ) Calculate the ratio of olefins obtained from EtCMe2Cl and Et2CMeCl., 2. Predict which reactions in each of the following pairs results in the higher ratio of, elimination to substitution:, 8, ratio ot, (a) t-PrCMe.Br or a-PrCMe*; heat in aqueous alcohol, Et, (b) EtoCMeCl or EtCH2, , C, , Br; heat in aqueous dioxane., , Me, (d) PhrHMrM oL'"f,riS+I~; trcat with alcoholic NaOEt., (d) PhCHsCMesSMe^- or PhCH^CMe*; heat in aqueous alcohol., , (e) Et-S-Ph or Me.CH, , S, , Me; treat with alcoholic KOH., , (g) lPBuCHCCMrBr''PrBr °r'3t W"h a'C0h0lic NaOE<(') ‘-PrsCClEt; heat in t-BuOH or in EtOH, , aqUe°US alC°ho1'
Page 526 :
510, , -, , Beta-elimination Reactions, , OH, NaNH2, , (s) Ph2C=CHBr-» PhC=CPh, CH2Br, (t) Me2C, , /, \, , + OH- -» CH2=CMe2 + HCHO, ch2oh, , 4. (a) Excluding optical isomerism, there are eight possible benzene hexachlorides, (1,2,3,4,5,6-hexachlorocyclohexanes). Sketch these., (b) Of these, five have been prepared, with the following structures assigned:, , Isomer, Structure, , (cis, , Cl's), , oc, , (3, , y, , 8, , e, , 1:2:4, , 1:3:5, , 1:4, , 1:3, , 1:2:3, , Each of these, on treatment with base, loses 3 moles of HC1 to give one or more, trichlorobenzenes. However one of the five isomers above reacts about 1000 times, as slowly as the remaining four. Which one is this? Explain., (c) Two of the remaining four prepared isomers lose their second molecule of HC1, much more rapidly than they lose the first. Which are these?, (d) Of the remaining two isomers, one isomer loses the second molecule of HC1 about, one half as rapidly as the first HC1, whereas the other isomer loses the second HC1, about one twelfth as rapidly as the first HC1. For which of these two isomers is the, ratio ki/k\ the smaller? Explain., 5. Predict the predominant products in each of the following cases, specifying, where pos¬, sible, whether an olefin formed is the cis or trans isomer:, Me, (heat, HOAc), , (a) n-Pr—C—Br ->, Et
Page 528 :
512, , Beta-elimination Reactions, , 6. Explain each of the following observations:, (a) The, , ratio, , of m-2-olefin, , to, , trans-2-o\tfin, , obtained, , from, , the, , unimolecular, , elimination reaction of EtCHMe—OBs is about twice that obtained from, z-BuCHMe—OBs under comparable conditions., (b) Although the —SRt and the —S02—R groups are easily removed in elimination, reactions, the —SR group is not., (c) The, , ratio of 2-olefin, , to, , 1-olefin obtained from the action of NaOEt on, , «-PrCHBrMe is less than that obtained from EtCHBrMe., (d) In the bimolecular dehydrochlorination of CHC1zCH2C1, the chlorine lost is the, “single” one (that is, the product is CC12=CH2); but in the bimolecular dehydrobromination of MeCBr2CHBrMe, the bromine lost is one of a pair (that, is, the product is MeCBr=CBrMe)., (e) The mass law effect (p. 256) is much more commonly observed in S.vl reactions, than in E\ reactions., (f) There are two very different concentrations (c and c') of NaOEt (in dry ethanol, at a given temperature) at which the elimination to substitution ratio for z'-PrBr, is the same as that for /-BuBr. Between c and c', the E/Sn ratio is greater for f-PrBr,, but at concentrations less than c or greater than c', the E/S.v ratio is greater for, <-BuBr., (g) When the threo form of ester LXVI is treated with base, practically no deuterium, is lost. When the erythro form of this ester is so treated, almost all of the deuterium is, lost., ., (h) When the ammonium ion LXVII is treated with base, the major products are an, olefin and Me,N; but when ion LXVIII is so treated, the major products are, amine LXIX and MeOH., i) When the three isomer of acid LXX is treated with pyridine, ,t undergoes, , deca-, , boxylative denomination,” but when the erythro form of LXX rs treated tn the, same way, it undergoes simple dehydrobromination.
Page 530 :
CHARTER, , 13, , Addition Reactions, , Among the hundred-odd, , chemical elements, only carbon, nitrogen, oxygen,, , and sulfur form double bonds in the usual sense;1 and of these only carbon and, nitrogen form true triple bonds. Although nine different types of multiple, bond may be formed from the various combinations of these elements, our, knowledge of addition reactions is based largely on studies of the transforma¬, tions of C=C and C=0 bonds., The reader is reminded that addition reactions of the C=0 group are, often initiated by nucleophilic attack on the carbon atom, although in strongly, acid solutions the oxygen atom may be protonated first. On the other hand,, the addition reactions of olefins are generally initiated by electrophilic attack,, although initial nucleophilic attack is possible if strongly electron-attracting, groups lie near the C=C bond. Moreover, additions to the C=C bond may, also be initiated by free radicals, but such reactions are not to be considered, until Chapter 16., , The Hydration of Olefins, The dehydration of alcohols to olefins, generally carried out in solutions of, strong acids, is reversible; treatment of the olefin with dilute acid may regen¬, erate an alcohol., / In addition, a few compounds with B=N bonds are known, and there is some (incon¬, clusive) evidence that a number of bonds in inorganic compounds (for example, the P—O, bonds in phosphates, and the metal-carbon bond in some cyano complexes) have ‘double¬, bond character.” Although such bonds resemble the more usual double bonds in certain, respects (being especially strong and unusually short), they do not undergo addition reaction, in the same banner as do C=C, C=0, and C-C bonds, and are thus not ‘unsaturated, from the organic chemist’s point of view., , 514
Page 531 :
The Hydration of Olefins, , h3o+ +, , \, , c=c, , /, If, , we, , /, , \, , II, , -H+, , I, , +H, , 515, , ^ —c—C— ^==f —c—c—, , I, , h2o+ h, , I, , I, , oh, , h, , assume that the dehydration is a simple unimolecular elimination reac¬, , tion, the principle of microscopic reversibility (p. 319) requires us to assume, that the hydration proceeds by a similarly simple two-step sequence. The con-, , \, /, II, II, c=c + h3o+ —c—c— + h2o -> —C—C—, /, \, 1 I +, 2 I I, H, , H, , (1), , OH2, +, , I, , centration of carbonium ion I must remain low in an aqueous medium, so we, may infer either that this ion is consumed (in step 2) essentially as rapidly as, it is formed (in step 1), or that its formation is reversible (in which case, step 2, would be rate determining). The second of these possibilities is excluded by a, number of observations, of which we need consider only two. It has been found, that il the hydration of an olefin is initiated in acidified D20 but stopped in its, early stages, the unreacted olefin has acquired no deuterium and has undergone, no isomerization,~ whereas, as shown below with methylenecyclobutane, both, deuteration and isomerization should be expected if the reversible formation, of a carbonium ion were involved., , (not observed), , (not observed), e are then left with the alternate possibility, the slow formation of carbonium, ■on I but this appears to be excluded by the observed rate law. For it has been, melt’s A T thC ra'! °f hydration is Proportional, not to (H30+), but to Hammett s h0 iunction,5, ‘u, , 79, 3724 (195*7). lnd.hlt’perime»^'”', , ), , 5807 <1956>’ » ^esz, Taft, and Boyd, ibid, , a"°w «Um between the olefin, w Agreement’* not eompTtethat, otervldrate |W, transfer from H.O+ to the o.ehn; see, for, , mUCh '°° ^, , ^, , 73’ 7392 <1951>^
Page 532 :
516, , Addition Reactions, , indicating that the rate-determining step involves a molecule of olefin and a, proton but no other species. It has thus been proposed S(o) that sequence (1) must, be expanded by addition of another step in which the solvent does not take, part. Now, it is probable that when the proton first attaches itself to the olefin,, it is attracted to the x electrons of the double bond rather than specifically to, one of the two unsaturated carbons—that is, that it first forms a x complex, (p. 119). However, the added proton is obviously associated with just one of the, carbon atoms in the resulting carbinol and in the carbonium ion, I, leading to, it (that is, the carbonium ion is a “<r complex”). The observed rate law then, suggests the following mechanism for the hydration of olefins, with the rate¬, determining step being the conversion of the x complex (II) to the carbonium, ion, I:, H, \, , /, , C=C, /, , ,, , +, , fast, , \, , fast, , /CTC\^f, , H, , \, , j, , /, , slow, , +H, , -c-cI, , I, , I, , + H20, fast, ^—H20, slow, , H, , >°x, , H, , II, x complex, , carbonium, ion, , I, , C-C-, , H, (3), , If this mechanism is correct, we must infer that the x complex II is also an, intermediate in the reverse reaction, the acid-catalyzed dehydration of the, carbinol. However, this complex would not be detected in kinetic studies of the, latter, for, as is indicated above, it is formed and destroyed after the rate-deter¬, mining step in the dehydration., In contrast to the hydration of simple olefins, the rates of acid-catalyzed, hydrations of the double bonds in n,£f-unsaturated aldehydes and ketones have, been found to be proportional to (H30+), rather than to h0A This, along wit, other evidence, suggests that these additions proceed by a mechanism other, than that represented in sequence (3) (see Ex. 1)., , Additions of Hydrogen Halides. Markownikoff’s Rule, The addition of hydrogen halides to double bonds is not conveniently studied in, , hydroxylic solvent; for hydrogen halides are largely dissociate, , in such med,, , V , ,L catalvzed hydration (or “alcoholation”) of the olefin will compel, and acid-catalyzed hyarat, t, addition. When such reactions are, , St'- — ,, , r, , ^, , rh.m, , W, , 4 Lucas, et at., J• Am. Chem. ooc.,, , 59, , >, , 1461 (1937); 66, 1818 (1944).
Page 533 :
517, , Additions of Hydrogen Halides. Markownikoff's Rule, , tion. Specifically, the addition of HC1 to Me2C=CH2 and that of HBr to, CH3CH=CH2 (both in pentane and both at 0°) have been observed to be, first order in olefin and of an indefinite order (near three, but not equal to it), in hydrogen halide.5 Such a rate law is similar in nature to that for the alco¬, holysis of acyl chlorides in CC14 (p. 334) and that for the iodination of mesitylene, with I Cl (p. 443). It indicates that besides the molecule of HX actually in¬, volved in the addition, the activated complex for the slow step is associated, with about two “extra” molecules of hydrogen halide. It is probable that the, function of these two (or more) “extra” molecules is to increase the polarity of, the medium in the immediate vicinity of the reaction site, thus facilitating the, ionic cleavage of the attacking molecule. They may even serve to “pull” the, halide ion away from the hydrogen-halide molecule, as indicated in the mecha¬, nism proposed below:, , \ /, C, II, , \ /, , fast, , + 3HX, , HA H-X(, , eq, , A, , H-JT, , C, , HX, , V/, , /', , HA H+ A -X</, , c, , C, , / \, , HX, , / \, , H-X, , III, , \ /, C, , HA H, , (slow?), , --A, , C, , I, -C +, , |, , H-C—, , A \, , X■, , fast, , I, , —c-x, , ->, , I, , H-C—, , Of the stepwise nature of these additions there can be no doubt, for rearrange¬, ments may accompany the reaction, even though neither the adduct itself nor, the olefin itself rearranges under the reaction conditions employed. Thus, the, addition of HC1 to z-PrCH=CH2 yields some rearranged product, IV,, whereas addition of HI to /-BuCH=CH2 yields a large amount of iodide V,, likewise a rearranged product.™ These rearrangements, as shown, almost, certainly proceed through carbonium-ion intermediates., , Me2CH—CH=CH, -^X, , ?, , Me2C-CH-CH3-j-Me2C-CH2-CH3^>Me2C-Et (IV), , Me, Me3C-CH=CH, —, , I, —> Me2G—CH—CH3—>Me2G —CH—CH3-H-Me2C—CH—Me, , i, , (V), , < ni'1 aL’ L Am' Chm- S,IC-' 69> 1339. 1348 0947), , Whit il, ,w”72, ImAsO)"' y', , ^^', , S°‘•’ 55> 5020 (1933)' « Eck'. C°ok, and
Page 534 :
518, , Addition Reactions, , From the evidence available at present, we cannot tell whether the formation, or the destruction of the tt complex (III) is rate determining., When the original olefin is unsymmetric, two (unrearranged) addition, products are possible, for the hydrogen ion may attach itself to either the a- or, the /3-carbon. In such instances, the predominant direction of addition will, depend upon which of the two possible carbonium-ion intermediates is the more, stable. Specifically, since tertiary carbonium ions are more stable than second¬, ary, which are, in turn, more stable than primary, an addition reaction pro¬, ceeding through a secondary or tertiary carbonium ion should be favored over, one proceeding through a primary carbonium ion. This is the basis of the, familiar Markownikojf rule, which stipulates that in the addition of HX to an, olefin, group —X becomes bound to the more highly substituted of the unsaturated, carbon atoms. Typically, the addition of HI to l-butene7(a) and the heterolytic, addition of HBr to 2-methylpropene7(6) yield secondary and tertiary halides VI, and VIII, respectively, rather than primary halides VII and IX. Similarly, the, , EtCH—CH3 (more stable) -U EtCHICH3 (observed), , H, , I, , I, , VI, , EtC^CH2, H, , EtCH2—CH2 (less stable) -U EtCH2CH2I, , (not observed), , VII, Me?C—CH3 (more stable)-Me2CBrCH3 (observed), , X, , "'2, VIII, Me2C-|-CH2, Tutl2\, H, X Me2CH—CH2 (less stable) -^Me2CHCH2Br (not observed), IX, , addition of HC1 to CH2=CHBr proceeds through the carbonium ion, CH3CHBr+ (with three /3-hydrogens), to give CH3CHBrCl, rather than through, +CH2CH2Br (with only two /3-hydrogens), which would give ClCH2CH2Br.s, Apparent exceptions to Markownikoff’s rule may be encountered where factors, other than hyperconjugation influence the relative stabilities of the two car¬, bonium ions derived from a single olefin. Thus, both neurine (X)«» and tnfluoromethylethylene (XIII)S<6) undergo “anti-Markownikoff addition”; for in, both cases, the carbonium ion intermediates for Markownikoff addition (XI, and XIV) are energetically less favorable than the respective alternative, carbonium ions (XII and XV)., r (,) Kharasch, Hinckley, ibid.,, , 56,, , and, , Hannum, J. Am. Chem. Sac,, , 56,, , 1782 (1934). (A) Kharasch and, , 1212 (1934)., , 1 (a)^Schmidt, Ann.’, (1950)., , Carbonium ion XI would be expected to, , 267,1300, , (1891). (b) Henne and Kaye, J. Am. Chem. Soc.,, , 72,, , 3369
Page 535 :
Additions of Hydrogen Halides. Markownikoff s Rule, , ', Me3N—CH =^= CH2, H, +, , X, , 519, , -> Me„N—CHI—CH., , MesN—GH—CH3-, , Me,N—CH=CH2, , -, , (Mark’fF product,, not observed), , XI, , N, , + V, , Me3N—CH2— CH2, , Me3N—GH2-CH2I, , XII, , (anti-Mark’ff product,, observed), , f3c-ch-ch3-^-* f3c-chci-ch., , F3C-CH =ch2, , H", , H, +, , XIII, , (Mark’fF product,, not observed), , XIV, , ■>F3c-CH=j=cH2, , f3c-ch2-ch+, , cr, , >F3C-CH2— CH2C1, , XV, , (anti-Mark’fF product,, observed), , rather unstable since there are positive charges on adjacent atoms. Similarly,, despite hyperconjugation, ion XV appears to be somewhat more stable than, ion XIV; for in the latter, the positively charged carbon is adjacent to the, carbon atom of the F3C— group which, because of the strong electron-attracting, action of three fluorine atoms, also bears a partial positive charge., Anti-Markownikoff addition of HBr to olefins is often observed when no, effort is taken to exclude peroxides or elemental oxygen from the reaction, mixture. This, the familiar “peroxide effect,” is due to a fundamental change in, the nature of the addition that, under such conditions, proceeds by a freeradical mechanism, with Br«, rather than H+, attacking the double bond., , \, , C=c, , /, , Br-, , HBr, , -> — C—C—-> —C—C-b Br., I, , •, , /, , \, , » etc. (see Chap. 16), , ||, , Br, , Br H, , There are indications that occasionally the addition of a hydrogen halide, may bypass the carbonium ion; that is, the halide ion may attack the tt com¬, plex directly. For example, the additions of HI to the cis-trans pair of acids—, tighc acid and angelic acid—in chloroform have been found to be stereospecific;, tiglic acid gives an erythro 0-iodo acid, whereas angelic acid gives a threo /3-iodo, , Me, , COOH, , XT, Me, , xH, , Me, HI, , ., , M'YDfCOOH, , H, H, , COOH, , ^, H/GxMe, , tiglic acid, , cis, , to, , erythro, , angelic acid, trans, , ^oung, Dillon, and Lucas, ./. Am. Chem. Soc., 51, 2528 (1929)., , M'3vrCOOH, H, , y, H, , threo, , 'Me
Page 536 :
520, , Addition Reactions, , Such stereospecificity would not be expected if the addition proceeded through a, carbonium ion analogous to I; for free rotation about the Ca—Cp single bond, in such an ion is presumably permitted (XVI, , XVII), and, contrary to what, , Me-^—^CXDOH, , Me"XlX h, H, XVI, , XVII, , is observed, a mixture of erythro and threo /3-iodo acids would result from both, unsaturated acids. Since the additions are trans, they cannot be one-step, processes in which the hydrogen attacks one unsaturated carbon while, at the, same time, the iodide attacks the other. Instead, it is likely that the halide ion, attacks the tt complex (XVIII) on the side opposite to that where the proton lies., I, Me, , Me, , 1", , N., , I, , Me, , x, , I, , Me, , /, , +, , H, , 'CTC', H+, , COOH, , H, , ^, , COOH, , XVIII, , Hydrogen halides may add to triple bonds in much the same way, for with, these reactions also (in aqueous media) trans addition is observed. It has long, been known, for example, that both acetylenedicarboxylic acid (XIX) and, propiolic acid (XX) undergo trans addition with hydrogen halides., Br, , HOOC, \, HOOC—C=C—COOH + HBr, , c==C, , 11(a), , \, , /, , XIX, , /, , COOH, , H, , H, , h3c, \, CHs—C=C—COOH, , + HC1, , XX, , c==C, , /, Cl, , /, , \, , 11(b), , COOH, , Additions of Halogens'2 and Hypohalous Acids, A worker wishing to study the addition of halogens to unsaturated linkages, if, he chooses to carry out his experiments in nonpolar solvents, will soon become, « (a) Michael and Brown, Bn., 20, 550 (1887) (4) Friedrich Am 219, 520 <1«>it For reviews of halogen additions, see (a) Williams, T,m,. Faraday Soc., 37,749(11, and (b) de la Mare, Quart. Revs., 3, 126 (1949).
Page 537 :
Additions of Halogens and Hypohalous Acids, , 521, , convinced that he is dealing with a very complex subject. If he purifies his, solvents carefully and works in the dark, the addition reactions will be very, slow; however, they are accelerated by light and are strikingly catalyzed by, small quantities of such polar substances as hydrogen halides and water,, although generally there is no simple algebraic relationship between the con¬, centrations of such catalysts and the rates of catalyzed addition. Moreover the, addition rate may easily be increased by adding glass beads to the reaction, mixture, showing that the reaction is sensitive to glass surfaces and suggesting, the possibility that additions in the dark in the absence of polar catalysts take, place largely (or perhaps exclusively) on the glass walls of the container., On the other hand, these complications virtually disappear when such, additions are transferred to polar solvents, and most of what is known at present, concerning the mechanisms of halogen-addition reactions in solution is based, upon studies of reactions in hydroxylic solvents, particularly water and the, alcohols.13, The additions of halogens, like those of hydrogen halides, are stepwise, reactions; that is, they proceed through intermediates in which only one of the two, halogen atoms has become attached to the olefin. The intervention of such an inter¬, , mediate in the bromination of ethylene has been clearly demonstrated by, carrying out the addition in the presence of NaCl or NaN03, for under these, conditions a bromochloride (XXII) or a bromonitrate (XXIII) is formed, along with the expected dibromoethaned4 We would then expect the “halfbrominated” intermediate to be carbonium ion XXI which, when formed,, may react either with one of the anions present in solution or with the solvent., BrCH2CH2Br, BrCH2CH2Cl (XXII), Br, + CH2=CH2, , Br~ + [BrCH,CH, XXI, BrCH»CH20N02 (XXIII), BrCH2CH2OH2, , In interpreting these results as diversions of a common cationic intermediate,, we are assuming that the monobromo compounds XXII and XXIII are not, ormed in substitution reactions on ethylene dibromide; this assumption is, certainly justified, for it may be shown that although BrCH2CHoBr does react, both with Cl- and NOj, these reactions are far slower than the formation of the, With, , « almost invariably homely,ic in, , since F, reacts readily with the available 1,1,1'!!°' iP ''7' 7 mfa nonP°*ar balogenated solvent,, 40, 51 (1947)., y, ailable polar solvents- s". f” example, Bigelow, Chem. Revs’, i Francis, J. Am. Chem. Soc., 47, 2340 (1925).
Page 538 :
522, , Addition Reactions, , bromochloride or bromonitrate by the action of bromine on ethylene. We are, further assuming that the monobromo compounds are not formed by the direct, addition of BrCl and Br0N02 to the C=C double bond; this assumption is, likewise justified since BrCl is too strong an oxidizing agent to be formed under, these conditions and the species Br0N02 is, to date, unknown., Moreover, addition of bromine to stilbene (PhCH=CHPh) in methanol, yields, along with stilbene dibromide, a /3-bromo ether, XXV./5(a) It can, Br2 + PhCH=CHPh -», v'^PhCHBr—CHBrPh, Br~ + PhCH—CHPh, , I, , Br, , PhCHBr—CHPh—OMe + H+, XXV, , XXIV, , be shown that ether XXV does not result from the (very slow) reaction of, methanol with stilbene dibromide, nor does it arise from the addition of methyl, hypobromite, CH3—O—Br, to the double bond.15 Similarly, when the sodium, salt of dimethylmaleic acid is treated with aqueous bromine or aqueous chlorine,, a /3-lactone (XXVII) is formed.17 In this case, the addition reaction is com¬, pleted when the single-halogenated intermediate (which we may represent for, the time being as XXVI) reacts “with its own tail.” The salt of dimethylfumaric, , Me, V, , Me, , coo-, , COCT, v., , -> Me, , Me, , X, , COO, , O, , o, , XXVII, , XXVI, , .. (a) Bartlett and Tarbell, J. Am. Chm. S,c., 58, 466 (1936). <») For additional expen-, , SC S, , fn^eof and Se^r and’, 519,165 (1935, (additions m-he acidb, « The unimportance of the addition of methyl hypobromite, MeOBr, to the doubleIxma, may be established by the following argument: The formation of ^ ^romtK from, :, anol and bromine cannot be rate determining for, , the, , s“”e large, , first order in stilbene; that is, stilbene is isw, addition, it must be, quantities of hypobromite are not detected during the course of the addition,, formed reversibly in the following reaction., Br2 + MeOH, , MeOBr + H+ + Br“, , This being the case, the^oncentrationji^ MeOBr^ hence^the^rate^ofTeaction^should^be^iii, , of acidity^we may^ionclude th^'^riulthy^ hypobromite, if it is formed a, all, is no, an important, 59, 407 (1937).
Page 539 :
523, , Additions of Halogens and Hypohalous Acids, , acid reacts in the same way, and for both of these reactions it may be shown, that the lactone does not form from the anion of the a,fl-chhalo acid., In representing the intermediate in halogen additions as a halogenated, carbonium ion, we have ignored one important feature—these reactions are,, in a number of cases, stereo specific. For example, it has already been noted that, the addition of bromine to maleic acid yields </,/-dibromosuccmic acid, whereas, the addition to fumaric acid yields m^o-dibromosuccinic acid (p. 150). This, stereochemical situation is inconsistent with intermediate carbonium, , ion, , XXVIII (if free rotation about the C—C bond in this carbonium ion is as¬, sumed), but consistent with the cyclic bromonium ion XXIX/18 Using NewmanCOOH, , HOOC, , H, , HOOC, free, , H, , rotation, Br, , H, , H, , HOOC, , H—TC, Br, , COOH, , H, , XXVIII, , XXVIII, , Br, +• •, XXIX, , COOH, , type projections, we may depict the addition of bromine to fumaric acid as, follows:, HOOC, , +, , H, , Br, , Br, , \ /, C, , Br2 +, , ||, , /G\, , H, , COOH, , Note that the second step, the attack on the bromonium ion, is in essence a, Walden inversion. The net result, like the reactions of hydrogen halides with, unsaturated acids (p. 519), is trans addition. Assuming that other halogen addi¬, tions follow a similar path, we may provisionally rewrite proposed intermediates, XXI, XXIV, and XXVI as cyclic bromonium ions XXX, XXXI, and, XXXII, respectively., , co7, , -o2c, H2C-, , -CHS, Br, +, , XXX, tu, for ., , PhCH-, , -CHPh, , C-(, , \ /, Br, +, , XXXI, , Me, , / \ /, Br, +, , Me, , XXXII, , Jh<; ^sumPtlon that intramolecular rotation in a carbonium ion is much more rapid, ,, an externcal nucleophile on such an ion is, at present, open to serious doubt (see, , I, , he cTP heSchaefrcr’„and Colhns> J• Am- Chem. Soc., 79, 6160 (1957)). This being, acceptance mnsfhf, "T, ^ hal°gCn additi°n’ deSPite itS wid«Pread, tions Ire dkeTiH c°nsiderfed tentative. Although in the present chapter a number of addir-action mrH, d in terms of thls mechanism, the reader should bear in mind that few organic, mechanisms have been accepted so widely while supported with such limited datf.
Page 540 :
524, , Addition Reactions, The question as to whether the formation of the initial cation (from the, , olefin and Br2) or its destruction (by attack of Br~) is rate determining is not, an easy one to answer. In either case, the transition state would be composed of a, molecule of substrate and (disregarding electron distribution) two bromine, atoms. Hence, kinetic studies which indicate that, in polar solvents, the additions, are first order each in olefin and halogen'5^9 are of little direct help here. A, tentative answer has been obtained for the addition of bromine to stilbene in, methanoF5(o) which, we have noted, gives a mixture of stilbene dibromide and, /?-bromo ether XXV. From a careful examination of the relationship between, the ratio of products and the concentration of added bromide ion (see Ex. 2 for, details), it is concluded that, for this reaction at least, the formation of the initial, cation is rate determining and its destruction is rapid., , For additions of bromine, iodine, and the interhalogens, BrCl, IBr, and, I Cl, in poorly ionizing solvents (such as acetic acid and nitrobenzene) another, term must be included in the rate law. This term is second order in halogen and, becomes predominant at moderate halogen concentrations:50, , The second term suggests a picture similar to that proposed for the additions of, hydrogen halide in nonpolar solvents (p. 517). In poorly ionizing media, the, formation of the presumed halogenonium-ion intermediate (a reaction requiring, charge separation) is difficult, but is aided by a second molecule of halogen, which helps disperse the negative charge by formation of the large trihahde, ion, Xj., \ /, , C, , II, , G, , / \, , + X2, , fast, ieq, , I, , V, , -Ik X—X, , slow, , x+, , fast, + Xs, , —G, , x-cI, , +, , X*, , -G-X, , (5), , /C\, , In keeping with this suggestion, no third-order term has been found for the, additions of chlorine, which forms the C\j ion only with extreme difficulty., The additions of bromine to allyl halides, vinyl halides, and vinylacetonitrile in glacial acetic acid have been found to be catalyzed by added bromide, and more effectively, by added chloride.55 This is not merely a salt effect, tor, the additions of acetates, nitrates, or bisulfates to the reaction mixtures cause, ,, t nu, ioq7, 1950 1624. (b) Berthoud and Mosset,, a (a) Robertson, et at., J. Chem. Soc., 1937, 555, isou,, w, , J. chim. phys., 33, 272 (1936)., , 179.1945, , 129- 1947, 628; 1950, 812, 2191., , « ChatTaway and Hoiyle,! Chem. Soc\ 123, 654 (1923); Sherrill and Izard, J. Am. Chem., , Soc., 53, 1667 (1931)., M9421, « Nozaki and Ogg, J. Am. Chem. Soc., 64, 697-716 (1942).
Page 541 :
Additions of Halogens and Hypohalous Acids, , -, , 525, , only small increases in rate. Assuming that, in the absence of added halide, the, addition would proceed according to sequence (5), we must suppose that the, halide ion intervenes before the rate-determining step. Probably it competes, with a halogen molecule in destroying the intermediate 7r complex, XXXIII., , V, I, , Xc/, , fast, , + Br2, , Br-Br, eq, , X, , -, , x-c+, , slow, , (6), , I, , /\, , /G\, , Br', , — C—Br, , XXXIII, , We then may ask why Br-, which is almost always more nucleophilic toward, carbon than is Cl-, is less potent in catalyzing the addition. The answer is, probably that although Br- is more efficient than Cl- in attacking the inter¬, mediate complex XXXIII, the addition of Br- converts a significant fraction, of Br2 in solution to Brj\ This lowers the equilibrium concentration of this, intermediate, thus tending to nullify the catalytic action., Although it is quite possible that the addition reactions of chlorine and, HOC1 pass through cyclic chloronium ions, , analogous to, , bromonium ions XXX, XXXI, and XXXII, evidence is accumulating which, indicates that “classical” chlorinated carbonium ions (—C—C—cA are also, , V + 1, , /, , involved. Consider, for example, addition of HOC1 to isobutene, Me2C=CH2,, which is known to yield as the predominant product the a-chlorohydrin, Me2CCH2Cl.*s If the product were formed by attack of water on the chloronium, OH, ion XXXIV, we would expect the alternate chlorohydrin XXXV to predomi¬, nate, since “SN2” attack at the primary carbon would be favored (as is the case, with substituted ethylene oxides, p. 292). We may, however, account for the, predominance of chlorohydrin XXXVII by assuming that the chloronium, Cl, , I, CH2OH, , (minor product), , XXXV, , Me2C=CH2 -C°ffiuMe2C-CH,, , \ /, +C1, XXXIV, , OH, h2o, , Me2C, , CH2C1 —^—> Me2C —CHoCl, , XXXVI, , XXXVII, (main product), , ** Michael and Leighton, Ber., 39, 2157 (1906).
Page 542 :
526, , Addition Reactions, , ring opens to give a tertiary carbonium ion XXXVI before H20 or OH~ attacks., If carbonium ion XXXVI is indeed an intermediate, we must consider the, possibility that, instead of reacting with a molecule of water to form chlorohydrin XXXVII, it may lose a proton, forming an olefin. Moreover, since this, carbonium ion is the same as that which would result from ionization of iso¬, butene dichloride (Me2CCl—CH2C1) in water, the ratio of olefin to chlorohydrin, resulting from the treatment of isobutene with HOC1 should be very nearly the, same as that resulting from the hydrolysis of the dichloride under the same con¬, ditions. The two reactions have, in fact, been compared (at 45°), and in both, cases the mixture of products contains about 85 percent chlorohydrin XXXVII, and 10 percent of the olefin CH2=CMe—CH2C1.*4, Although the cyclic chloronium ion (for example, XXXIV) is assumed to, intervene largely on the basis of analogy with stereospecific bromine additions,, there is independent evidence of its existence (the necessary reasoning is a little, involved, and the reader may have to go over the following section more than, once before he feels completely at ease with the subject). The addition of, HOC1 to CH2=CMeCH2Cl gives principally chlorohydrin XXXVIII, with, minor amounts (about 6 percent) of the isomeric chlorohydrin XXXIX,, which, for our present purpose, is the more interesting product. If the addition, is carried out on a substrate “labeled” with Cl36, it may be shown that about, one third of the chlorine in the 2 position in chlorohydrin XXXIX was originally, at the end position in the substrate.*5, Me, cioh2+, , CH2=C—CH2C1*->, OH, C1CH>—CMe—CHoCl* (XXXVIII) major product, h0CH2—CMeCl—CH2C1* (XXXIX) minor product, HOCH2—CMeCl *—CH2C1 (XXXIX') minor product, Now, if the observed reaction passes through a “classical” carbonium ion, that, ion, , is, , almost, , certainly, , the, , symmetrical, , species,, , ClCH2-CMe-CH2Cl, , (XLI); for there seems no way in which the rearranged chlorohy, could form from the alternate carbonium ion, +CH2—CMeCl—C, cept perhaps by preliminary rearrangement of the latter to ranged chlorohydrin could easily form from, labeled, , chlorine, , through, , XLI, , the cyclic structure,, , X, , ct a, , u De la Mare and Salama, J. CW S*., 1956, 3337 The, , be identical, for both isobutene and its dichloride may react to a small extent throu0, P, , « Ballinger and de la Mare, J. Chem. Soc., 1957, 1481., , _, , ‘, , ^, , )•, , simply by, , ., , *
Page 543 :
Nucleophilic Additions to, Me, , I, , ch2=cch2ci, , C—C, , 527, , Bonds, , * CIOHt, unrearranged chlorohydrins, , Me, , (XXXVIII and XXXIX), , ■G-CH2Cr, , h2c, , Me, , \/, , .Cl, , CH,, , x ClCH2CCH2a*—>C1CH2C, XL, Me, XLI, , HoO, , -> <, , V, , XXXVIII, ana, XXXIX', , XL', , XXXIX' cannot be formed through carbonium ion XLI, since a symmetrical, carbonium ion should yield (except for a negligible isotope effect) equal, quantities of XXXIX and XXXIX', whereas the unrearranged chlorohydrin, XXXIX is found to predominate. Hence a fraction of chlorohydrin XXXIX, must arise in a different way—very probably, as indicated above, by direct, attack by water on chloronium ion XL. The fact that ion XL (which, aside, from labeling, is identical to XL') may undergo two different reactions indi¬, cates that XL (hence XL') is an intermediate rather than a mere transition state,, since we would not expect to find two different reactions proceeding through, the same activated complex., , Nucleophilic Additions to C=C Bonds, From the mechanism proposed to explain the catalysis of halogen addition, by halide ions in acetic acid (sequence 6), it would be supposed that HBr or, HC1 would be no more effective catalysts than the corresponding lithium, halides. This is the case when the substrate contains no oxygen (for example,, in additions to vinyl and allyl halides); but for additions to a,/3-unsaturated, ketones, aldehydes, acids, and esters, the hydrogen halides are found to be, much more potent catalytically than are the alkali halides.** Since the sub¬, strates that undergo additions preferentially catalyzed by hydrogen halides are, compounds that may be converted (at least in part) to their conjugate acids, under the conditions employed, it appears likely that these conjugate acids, have become involved in the catalyzed additions. Moreover, such additions,, unlike “ordinary” additions, are retarded, rather than accelerated, by alkyl, substitution on the olefinic carbons,*6 suggesting that the attack on the C=C, bond is nucleophilic' rather than electrophilic. The following mechanism then, seems consistent with what is known about additions of this type:, t6 Morton and Robertson, J. Chem. Soc., 1945, 129.
Page 544 :
528, , Addition Reactions, Br, , H+, —C-=C-C- +=*, , +, , —c—c=c-, , Br~, —>, , I, , —C—C=C—, , I, , HN, OH, , V, , I, , Br, , OH, , Br,, , L -H +, , Br, , I, , I, , G—G—G— ^=± -C-C-C-, , I, , OH, , Br, , Br, , These additions are catalyzed, although much less effectively, by strong acids in, the apparent absence of added halide ion (for example, by HCIO4, H2S04, and, HNO3).*7 Here a similar mechanism may be written, except that the conjugate, acid is attacked by a molecule of solvent (or by HSO7 or NO7) rather than by, a halide ion during the initial stages of the reaction, thus diverting a portion of, the substrate to a single-halogenated product. As the reaction proceeds and, more halide ion is generated, more and more dihalide is formed., In these reactions, initial attack on the olefinic carbon is nucleophilic, rather than (as is more usually the case) electrophilic, even though the attacking, species present in solution are not strongly nucleophilic. This is because the, normally strong electron-withdrawing action of the C=0 group has been, augmented by the acquisition of a proton. If more powerful nucleophiles are, available in solution, nucleophilic attack at a C=C (or a C=C) bond in, conjugation with a C=0 or C=N linkage may occur without prior protona¬, tion. Of the many and varied “1,4 additions” that take place in basic media,, most conform to this description. In addition to the Michael condensation (p., 392) and the 1,4 additions of Grignard reagents to a,/3-unsaturated ketones, in, which the attack is by a strongly nucleophilic carbanion, the following may be, considered typical:, C=N, , C=N, PhCH=C, , + CN~ -> PhCH—C, , \, , COO-, , I, , CN, , \, , COOCN, S8, , PhCH—CH, , \, , CH2=CH—CN + PhNH—C=NPh, S_, , COO", CN, h2o, CH 2—CH 2—CN*9, S—C=NPh, I, NHPh, , " Robertson, et al., J. Chem. Soc., 1945, 888; 1948, 980., ss For kinetic studies of this addition, see Jones, J. Chem. Soc., 105, 15, t, >■, ” The 1,4 addition reactions of acrylonitrile (cyanoethylations), avc, N, by Bruson in Organic Reaction, Vol. V (edited by Adams), John Wtley and Sons, Inc., Ne, York, 1949, p. 79.
Page 546 :
530, , Addition Reactions, , Additions to Conjugated Polyolefins, In the absence of extensive information concerning the kinetics and stereo¬, chemistry of additions to polyolefinic systems, it is convenient (and it may even, be correct) to assume that such additions proceed by the same types of mecha¬, nisms as do additions to mono-olefins under corresponding conditions. Re¬, cently, it has been demonstrated that the 1,4 addition of chlorine to butadiene, is a trans addition, to each other., , that is, that the chloromethyl groups in the product lie trans, , C1CH2, CH2=CH—CH=CH2 + Cl2 ->, , ^CH^CH, , \, , CH2C1, , This shows the addition to be a stepwise process (that is, it excludes a transi¬, tion state such as XLV) and indicates also that the addition does not proceed, through chloronium ion XLVI (in which the chlorine is incorporated into a, five-membered ring), for both such routes would lead to a cis dichloride. In¬, stead, as with mono-olefins, it appears that addition occurs through the chloro¬, nium ion XLVII, the chlorinated carbonium ion XLVIII, or through both., , HC^CH, //J, , H,C, , ACHj, , (, , HC=CH, , /, , \, , h2c, , HoC, , XH,, , Cl — -Cl, , +C1, , XLV, , XLVI, , 0, 0, , CH—CH=CH,, Cl, XLVII, , 0, 0, , 0, 0, , CICH5—CH—CH—CH2, , XLVIII, The 1,2 addition of chlorine to butadiene (giving C1CH2CHC1CH=CH2), accompanies the 1,4 addition considered in the preceding paragraph and, at, room temperature predominates.55^ As the reaction temperature is raised,, 1,4 addition becomes increasingly important, since the activation energy for, 1*4 addition is greater than that for 1,2 addition. Nevertheless, the 1,4 dichlonde, about the extent that cyclopropane resembles ethylene, more data concerning cyclopropenc, and its derivatives must be obtained before we can be at all confident that such an analogy, well taken., ., ., 3* Mislow and Heilman, J. Am. Chem. Soc., 73, 244 U'^l)., .,, ,, « (e) Muskat and Northrup, J. Am. Chem. Soc., 52 4042 (1930). (*) Pudovik, J. Cen, Chem. (U.S.S.R.), 19, 1179 (1949); Chem. Abstr., 44, 1005 (1950).
Page 547 :
Additions to Conjugated Polyolefins, , -, , 531, , u actually the more stable isomer; for if the mixture of isomers obtained at room, temperature is heated to 200° or merely treated with ZnCl2 to bring about, equilibration by loss and return of Cl', the resulting mixture contains about, 70, , percent of the 1,4-dichloride .«<» Thus, we have here a situation in which, , the less stable product is formed more rapidly than the more stable—that is,, where the addition at moderate temperatures is kinelically, rather than thermo¬, dynamically, controlled. .This is also the case for the additions of bromine to, cyclopentadiene,5e(a> l,3-cyclohexadiene,M<“> and butadiene (in nonpolar sol¬, vents),“<*> the addition of HC1 to isoprcne,’"'* and the addition of HC1 and, HBr to butadiene.s<?(d) In each of these instances the mixture of products resulting, from addition at room temperature or below contains significantly more 1,2, adduct and significantly less 1,4 adduct than corresponds to a mixture of the, two at equilibrium. We may now ask: (a) why the 1,4 adducts in these reactions, are more stable than the 1,2 adducts; and (b) why the 1,2 adducts tend to form, more readily than the 1,4 adducts. Neither of these questions is easy to answer,, and any attempted explanation based on structural considerations must be, regarded with skepticism since we are probably dealing with energy differences, of well less than 1 kcal per mole. In the noncyclic series, it may be that the in¬, creased possibilities for hyperconjugation stabilize 1,4 adducts such as XLIX, with respect to 1,2 adducts such as L,, Xi, , CH2=C—CH=CH5, , (or HA"), , ♦ CH2—C=CH—ch2z + ch2—cx— ch=ch2, H(X), , H(X), XLIX, , whereas in the addition of halogens to cyclic dienes, it may be that the 1,4, adducts (for example, LI) are slightly more stable than the 1,2 adducts (for, example, LII), because in the former, dipole-dipole repulsion resulting from, the two C—X linkages is somewhat less than in the latter. As for the kinetic, , advantage in 1,2 addition, let us suppose that the course of the addition of a, halogen X2 to a diene is determined by the position at which X~ attacks the, halogenonium-ion intermediate LIII, whereas the course of a hydrogen-halide, addition is determined by the position of attack by X~ on carbonium ion LIV, 36, , J. Chem, J. Or% • Chem., 2, 489 (1938)., , ’, , ’, , irT*tand Thorpc-, , ( ) Kharasch> Kntchevsky, and Mayo,
Page 548 :
532, , Addition Reactions, , It is reasonable that attack at the 2 position of LIII (leading to 1,2, I, , \, , |2, , addition), , 4, , -C—c—CH=CH2 ^ —G—C=GH—CH,, , /2, , I, , X, , I, , H, ^AAi, , H, LIV, , LIV/, , will be slightly favored over attack at the 4 position (leading to 1,4 addition)for the former involves only the breakage of the three-membered ring, whereas, the latter requires, in addition, a rather extensive redistribution of x-electron, density. Turning now to the attack on the hybrid ion LIV, , LIV', we may, , assume that nucleophilic attack will occur preferentially at the carbon having, the highest density of positive charge. Recalling that a secondary or tertiary, carbon will, in general, tolerate positive charge more easily than will a primary, carbon, it seems probable that the positive charge in LIV, although spread, over three carbons, will be most concentrated at the 2 carbon and that, once, again, 1,2 addition will be kinetically favored over 1,4 addition., If we consider the additions that we have just described as typical, we are, apt to suppose, as have a number of workers,37 that additions to dienoid sys¬, tems at moderate temperatures result in a predominance of a less stable adduct., If we adopt this view, then we must infer that those additions that are observed, to result predominantly or exclusively in the more stable adduct (for example,, reactions 7 through 10) first yield the less stable adduct, which then isomerizes, before the product can be identified. The alternate (unobserved) adducts are, shown in square brackets in these sequences; each has at least one very labile allylic, bromine and would be expected to undergo allylic-rearrangement or internalreturn reactions (or both) (p. 286) with ease. Note that the addition of HG1 to, less stable adduct, (not observed), r Br(H), Br2, , CH2=CMcCH=CH2, , CHo—CMe—CH =CH2, , (or HBr), , Br, more stable adduct, (observed), Br(H), CH2—CMe=CH—CH,Br»«(“«, , (7), , r Br(H), Br2, , CH2=CMeCMe=CH2, , CH2—CMe—CMe=CH2, , (or HBr), , Br, Br(H), CH2—CMe=CMe—CH2Br, 37, , c), , (°>, , See, for example, Hughes, Ingold, dc la Mare, and Catchpole, J. Chm. Soc., 1948, 8,17.
Page 549 :
Cis Additions to the C=C Bonds. The Diels-Alder Reaction, Bn, , PhCH=CHCH=CH2, , -*, , LV, , (or HBr), , r—, , PhCH—CH=CH—CHo, Br(H) j, PhCH=CH—CHBr—CH2 •»«■*>, , Br, , PhCH=CHCH=CHPh, , 533, , (9), , [PhCHBr—CH=CH—CHBrPh] -», , LVI, , PhCHBr—CHBr—CH=CHPh«(a), , (10), , isoprene results largely in the kinetically favored 1,2 adduct, whereas the addi¬, tion of HBr results in the thermodynamically favored 1,4 adduct, presumably, because Br- is much more easily detached and more easily refastened to the, carbon chain than is Cl-., In the author’s opinion, however, there is considerable doubt that the less, stable adduct invariably forms first, then isomerizes (either slowly or rapidly), to the more stable adduct. Especially in cases where the 1,2 adduct is the more, stable, it seems quite possible that the thermodynamically favored product is, also favored kinetically. It would be interesting to carry out the addition of, bromine to dienes LV and LVI in methanol and examine the methoxy bromides, formed during the early stages of the reaction. If 1,2 addition predominates, here also, it could be assumed that the 1,4 adduct had not intervened, for a, rapid allylic rearrangement involving a methoxide group is extremely unlikely., One further point: when addition to an extended conjugated system is thermo¬, dynamically controlled (and this is generally the case), orientation will be such, as to preserve conjugation to the greatest extent possible. Thus, addition to, 1,4-diphenylbutadiene (LVI) yields a 1,2-dibromide rather than a 1,4-dibro¬, mide, since in the 1,2 adduct (but not in the 1,4 adduct) the remaining C==C, bond lies in conjugation with one of the benzene rings. Similarly, the addition, of 1 mole of bromine to hexatriene LVII yields a mixture of 1,2 and 1,6 adducts, (both of which have two double bonds in conjugation), but no 1,4 adduct:39, BrCH2CHBrCH=CHCH=CH2 +, BrCH2CH=CHCH=CHCH2Br, (no BrCH2CH=CHCHBrCH=CH2), , CH2=CHCH=CHCH=CH2, LVII, , Cis Additions to the C=C Bond. The Diels-Alder Reaction, We have referred to addition of bromine via the bromonium-ion mechanism, as a tram addition (p. 523), but this description is appropriate only because the, second step, in which the bromonium-ion ring is destroyed, involves an inversion, mi,SSdiy 'Alt’ Swrent"' ^h V* 756 '’922)' (i), (1909). (,) Mbe(r;, 9 Farmer, et at., J. Chem. Soc.,, , "» Sc°«., , 1927, 2937., , and Marshall, J. Chm., «30, 510. M Strauss, B„., 42, 2866
Page 550 :
534, , -, , Addition Reactions, , ot configuration. The initial step—the formation of the ring—is itself a as;, , addition, for substituents lying as to each other in the parent olefin also lie;, as to each other in the cyclic intermediate. Of the other types of as addition, , known, most involve cyclization; two, in particular, result in three-membered:, , rings which, unlike that in the bromonium ion, can be readily isolated. Olefins', may be converted to substituted ethylene oxides (“epoxidation”) with peroxy, acids., R, , R, , \, , /, , O, , R, , II, , C=C, , \, , + R"C—O—OH, , R', , R', as, , /, C-C, , R', peroxy acid, , r, , O, as, , + R"COOH, R', , This addition is facilitated by electron-releasing alkyl groups on the olefin, sug¬, gesting that the attack on the double bond is electrophilic. In polar media,, particularly in solutions of formic or acetic acid, the attacking reagent may be, the hydroxonium ion, OH+, formed in the equilibrium., R—C—OOH + AcOH, , OH+ + R—C—OH + OAc~, , I, , I, , o, , O, , However, epoxidation occurs readily in nonpolar solvents also. Here, the most, satisfactory mechanism (which may apply as well to polar solvents) is'*0, , O, H, , O, , The addition of carbon dibromide, CBr2, to olefins also appears to be, stereospecific. From m-2-butene, a derivative of m-dimethylcyclopropane,, LIX, is obtained, whereas trans-2-butene yields a /ranj-dimethyl derivative,, LVIII.^ At present we do not know whether CBr2 has zero or two unpaired elec¬, trons—that is, whether it is a “diradical” or merely an electron-deficient frag¬, ment. The principle that electrons associated with a given energy level remain, unpaired if possible, which applies straightforwardly to free atoms (p. 9),, does not help us here; for there is undoubtedly some drift of electron density, from the filled p orbitals of the bromine atoms to the electron-deficient carbon,, and it is difficult to predict how this will affect the energies of the two possible, forms. It has been found, however, that the relationships between the structures, i° Bartlett, Rec. Chem. Prog., 11, 47 (1950)., V Skell and Garner, J. Am. Chem. Soc., 78, 3409, 5430 (1956).
Page 552 :
536, , Addition Reactions, , with a G—C bond)—resulting in the formation of a six-membered ring:**, , V, , C, , G., , c, , G, , 4-, , -c, , /c\, , \, diene, , \, dienophile, , adduct, , The Diels-Alder reaction is, in essence, a reorganization of electron density,, , during which two single bonds are created, two double bonds are converted tc, , single bonds, and one single bond is converted to a double bond. At presem, , we cannot say whether the electrons shift one by one or in pairs—that is, whethei, , the reaction is homolytic (involving a diradical such as LX) or heterolytic.^, , v, , \ ^, G, , \ /, XC, , +, /C^G, , /C\, , /, , >, , LXI, , 4J The role of the diene in the Diels-Alder reaction may be assumed not only by the, usual acyclic and cyclic dienes, but also by such compounds as anthracene (which is attacked a, the center ring) and furan. Typical dienophiles are maleic anhydride, acrolein, acetylene, arb^Hc 3. and p-benzoquTnone. For review, of the ^Aider reacbon ,ee *£•£*, articles* in Organic Reactions (edited by Adams), John Wiley and Som, Inc., New, Kloetzel, 4, 1 (1948); Holmes, 4, 160 (1948); Butz and Rytina, 5, 136 (19 )., ,, « Because Diels-Alder reactions exhibit characteristics in common with both free rad, and ionic reactions, it is sometimes said that such reactions are partially homolytic, alternatively, “partially heterolytic” (see for example, Ingold, Structure and Mechanism in rg, , Chemistry, Cornell University Press, Ithaca, 1953, p. 718). The significance, are, is not, not clear,, clear, tor, for quantum, £t some, is, quantum mechanics, mecnames f suresus thatbound, -, dectro^K, T •, c, _cdKle* that, some, either paired or unpaired; there ean be no middle ground. It u, of comae,, , Diels-Alder reactions are homolytic and other,, duct may besomed by simultaneous operation of homolytic and heterolytic additions., , ad.
Page 553 :
Cis Additions to the C=C Bonds. The Diels-Alder Reaction, , 537, , The rate law, which stipulates that the additions are first order each in diene, and dienophile,^(a) is noncommittal on this question, and the highly negative, entropies of activation which are observed (from -20 to -25 cal per deg mole), could conceivably be due either to loss in freedom of motion when noncyclic, reactants form a cyclic transition state (p. 182) or to the necessity for an electron¬, unpairing process of low probability. The fact that the Diels-Alder reaction is, not accelerated by such radical-producing agents as peroxides does not exclude, homolytic addition; for these so-called “initiators” are effective only in the, formation of monoradicals that almost certainly would become involved in, side reactions before being converted to diradicals. Moreover, the rates of, Diels-Alder reactions appear to be more sensitive to the polarity of the solvent, used than are those of many typical free-radical reactions, but less sensitive than, are the rates of the large majority of heterolytic reactions.45(6), The Diels-Alder reaction is a cis addition, both with respect to the diene, (otherwise a stable six-membered ring could not form) and with respect to the, , dienophile, , that is, groups lying cis to each other in the dienophile also lie, , cis to each other in the adduct. This suggests that if the reaction passes through, LX, this diradical must undergo ring closure before rotation about the C—C, bond derived from the dienophile can occur. If the reaction is heterolytic, the, interval between the formation of the first and second of the new G—C bonds, must be similarly short; indeed, a number of workers feel that these bonds, form simultaneously w°)—that is, that “zwitterion” LXI does not intervene at, all. However, the observed relationships between structure and reactivity sug¬, gest that, in some cases at least, there is an interval early in the progress of, the reaction, during which considerable electron density has been transferred, from diene to dienophile but little, if any, is transferred from dienophile to, diene. Electron-donating groups (alkyl and alkoxyl) on the diene generally, facilitate the addition, and electron-attracting groups on the dienophile almost, invariably have the same effect. This is consistent with the formation of inter¬, mediate LXI, but it may merely reflect the preliminary formation of a complex, between the two reactants*™ before the series of bond breakings commences., In either case, the general acid catalysis of the Diels-Alder reaction which is, sometimes (but not always) observed may be attributed to coordination of the, acid with the dienophile, drawing electron density away from the reaction site, and thus making the dienophile more electrophilic.^6), , 510 (1937)., , ’, , ’, , • (c) Alder and Stein> An&ew- Chem., 47,837 (1934); 50,, , -»?• ^, , i’, , complex formation between d^ene and Hien^ha ’■, r, -?■ ,, (i) For evidcnce of preliminary, Keefer, J. Am. Cta, Sod,, , see Andrews and, (1956)., ’, ”, ’ Z84 (1955); and Berso", Reynolds, and Jones, ibid., 78, 6049
Page 554 :
538, , Addition Reactions, , Hydroxylation of Olefins with Permanganate, The oxidation of olefins with permanganate at high pH values may be easily, controlled so that the predominant product is a 1,2 diol. This is another cisaddition reaction, for cyclic olefins are invariably oxidized to cis glycols.^, Moreover, it has been shown, using MnOj labeled with O18, that both oxygen, atoms in the glycol are derived from the permanganate, rather than from the, solvent.47(6), , *0, , \, C=C, , *, HO, , O *"»-, , + 2, , /, , h2o, , Mn, /, , OH, , \, , L*0, , *, OH, , C, , *, , C-b 2MnOj, , (cis addition), , Q*J, , The occurrence of cis addition suggests the intervention of the cyclic complex, LXII in a mechanism such as:, , 0, , 0, i_, , \ X, G, , \, , V, , /, , Mn, , II, C, , o, \, , \, , 5, , /, , -c—o, , /, , **\, , i, 2H20, , Mn, , —>, 0, , 0, _l, , /, , r—c—o, i, , —, , o, , LXII, O', , 'HO, -C—OH, I, -C—OH, , s, , +, , S, , /, , Mn, .HO, , oJ, , then,, f 3St, , H2MnO^ + UnOj + 20H~ -> 2MnOi~ + 2H20, As indicated, the formation of complex LXII involves the transfer of a pair of, electrons from the olefin to the manganese atom through an oxygen. The +5, oxidation state of manganese (such as exists in H2MnOr) is relatively unfamiliar,, but it has been shown to be stable in cold 8 to 10 N alkali.^'(f) At lower basicities,, it would be expected to react with MrxOj to form the (green) MnOf” ion. The, latter oxidizes olefins to glycols also/7(c) but much more slowly than does, MnO^; this is likewise a cis addition, and for it a sequence very similar to that, above may be written., No salt containing an anion such as LXII has as yet been isolated, nor as, <» („) Boeseken, Ra. Irav. Mm.; 40, 553 (1921); 47, 683 (1928). (4) Wibcrg and Saegebarth, J. Am. Chem. Soc., 79, 2822 (1957). (c) Pode and Waters, J. Chem. Soc., 1956,, /•
Page 555 :
Additions to the C=0 Double Bond, substantial, , 539, , spectral evidence for the intervention of this intermediate been ob¬, , tained. However, osmium-containing esters having the ring structure, , I, , —c—o, , o, , —c—o, , o, , I, have been isolated from reaction mixtures in which olefins are undergoing, hydroxylation with osmium tetroxide, 0s04.4S This hydroxylation, like that, with permanganate, is a cis addition, and is often assumed to proceed by an, analogous path.49, , Additions to the C=0 Double Bond, Unlike the usual C=C bonds, the C=0 bonds are strongly polarized and, may be attacked either at the electron-rich oxygen atom by electrophiles or at, the electron-deficient carbon atom by nucleophiles. Of the familiar carbonyladdition reactions, most involve weak acids in which the hydrogen is bound, to carbon, nitrogen, or oxygen., , \, , C=0 + HB ^ —C—OH, , /, , B, When the acidic hydrogen is bound to a saturated carbon atom, the attacking, nucleophile is necessarily the conjugate base, B:~, and the addition is subject to, base catalysis (either general or specific); we have seen this to be the case, for, example, in the aldol condensation (p. 389) and the benzoin condensation, (p. 394). When the acidic hydrogen is bound to oxygen or nitrogen (that is’, w en the addendum is water, an alcohol, an amine, or a substituted hydrazine), initial nucleophilic attack by the addendum itself is also possible, for such com¬, pounds are Lewis bases. Addition of a stronger acid, HA, to the reaction mixture, wi, , partially convert the carbonyl compound to its conjugate acid, , C—OH, , /, «OvSe’ Mar^hand> and Wannowius, Ann., 522, 75 (1936); 550 99 (19381, , there is evidence for ^roccurrence'oTa numbe™"^diff^ C°nStitute.a comPlex topic, for, operating in a particular case may depend not onlv, A, mechamsms- rhe mechanism, also on the acidity of the solution and its temperature FoTabrVf ', ?ubstrate but, Ladbury and Cullis, Chem. Revs., 58 403 ('1958') Fnr, ,, review ol this topic, see, Waters, J. Chem. Sac., 1953, 435, 440, 2836, 3119; 195472456^1955,797’ ** DrUmm°nd and
Page 556 :
540, , Addition Reactions, , or perhaps to a hydrogen bonded complex,, , \, , C=0 • • • H—A. In either, , case, electron density is drawn away from the carbon atom, and attack by a, basic oxygen or nitrogen atom is facilitated. Thus, carbonyl additions in which, the addendum itself, rather than its conjugate base, attacks the carbon atom, should be catalyzed by acid, provided, of course, the attacking nucleophile is, not completely converted to its own conjugate acid., The hydration of acetaldehyde and its reversal, OH, ch3cho + h2o ^ CH3CH, OH, have been studied in some detail.50 Hydration is extensive (about 60 percent, at, , 20°)5*, , in distilled water, but much less complete in aqueous acetone, and, , both the forward and backward reactions are subject to general acid and gen¬, eral base catalysis. We should have little difficulty in suggesting a mechanism, for, , the, , acid-catalyzed reaction, for here, the hydrogen-bonded complex, , CH3CH=0 • • • HA is evidently involved. The formation and dissociation, H, , H, MeCHO l, Me(t.=0, + HA I JZt, LXIII, , + HjO, slow, , fast, , I, , H.4 ,, , HA, , MeC—O, , fast, , — H2O, slow, , o, '+N, , H, H H, MeC—O, , I, , H, , HA, fast, , f MeCH(OH)2, + HA, , O, H, LXIV, of the hydrogen-bonded complexes LXIII and LXIV and the transfer of a, proton from one oxygen atom to another are assumed to be rapid, whereas t e, dc, 1Q7A 141 Cl949V Trans. Faraday Soc., 46, 34 (1950). The re10 Bell, el al., Proc. Roy. Soc., 197A,, v., ; ’, ,, ., ■, measurine the changes in, , rZSFa,2; L, 43 >, hydration when dissolved in water but the iso a ion o, However, stable and ^^dily isolab e ydrates^re, electron-attracting groups lying near tne, , 0=CH—CH(OH)i)., , difficT, y |dehydes having strongly, CllCH(OH), and
Page 557 :
Additions to the, , C=0, , Double Bond, , -, , 541, , formation of a new C—O bond (in the hydration) and its breakage (in the, dehydration) are assumed to be rate determining., For hydration in the presence of OH-, the slow step is almost certainly the, attack by the OH- ion at the C=0 group of the aldehyde. However, when a, weak base (which we may designate B:) is also present, a second mode of attack, may occur, for this base also accelerates the hydration, even though (OH-) is, kept constant by simultaneous addition of conjugate acid i?:H+. Here it is, likely that the attacking reagent is a water molecule that has been made more, nucleophilic by coordination with the base, forming the hydrogen-bonded, complex, B‘. • • H, , O, , H,, , ^, , C=0, , slow, , B: • • H—O, , +, , I, , fast, , BH+ + HO—C—O" ^, slow, , |, , fast, , H, , I, I, , B: + HO—C—OH, H, As might be expected, complex B: • • H—O, , /, , is a much less powerful, , nucleophile than is OH-, and weak bases are considerably less effective catalysts, than is hydroxide ion. Indeed, for the hydration of acetone, complexes such as, , B: • • HOH are ineffective as catalysts, and this hydration appears to be sub¬, ject to specific hydroxide ion catalysis.5*(o) However, acid-catalyzed hydration, of this ketone appears to follow the same course as the hydration of acetaldehyde,, for like the latter reaction, it is subject to general acid catalysis. Furthermore,, a very similar mechanism may be written for the reaction of methanol with, acetaldehyde to form hemiacetal, LXV, for general acid catalysis operates, here also.5*(o*, OH, , CH3CHO + MeOH ™ CH3CH^, , \, OMe, LXV, The reverse reaction, , the splitting of a hemiacetal into its parent alcohol, , and aldehyde—is conveniently studied when the hemiacetal is optically active, ue (partially or solely) to an asymmetric a-carbon atom, for the incorporation, hydrate it^mn.aK, y’ J- im- Chem■ SoC•’ 60’679 <1938)- Since the concentration of acetone, tion^rate lsqH t "a"" .mUCh t0° Sma11 t0 be measured *Y conventional methods, the hydra0‘*, , replaced', , Trans. Faraday Sac., 48, , T7b)°M, , ^eton, , 1015 (1952), , rate at which, J. Chem., 30, 501 (1952);, , £ H2°'8 ^ measuring the, , ^ 3, , Darwcnb
Page 558 :
542, , Addition Reactions, , of that carbon atom into a C—O group destroys its asymmetry. The sugars ini, , their cyclic forms are hemiacetals of this sort, as they may be considered derived!, from a, , CHO group at one end of the molecule and an —OH group near the, , other end. The scission of the hemiacetal linkage in sugar solutions results im, , the familiar phenomenon of mutarotation—that is, a drift in the optical rotation, , of the solution to a constant value (complete loss of optical activity does not:, occur since a number of additional asymmetric carbon atoms in the sugar, , molecule are not affected). The mutarotation of glucose (LXVI) in aqueous;, solutions55(a) and that of 2,3,4,6-tetramethyl glucose (LXVII) in organic sol-, , OH, LXVIII, vents5S(6) have been studied carefully. In an earlier chapter (p. 139) it was seen, that mutarotation of the tetramethyl compound required catalysis by both, an acid and a base, and that 2-hydroxypyridine (LXVIII), in which the acidic, and basic centers are held rigidly in positions favorable for attack, is a particu¬, larly effective catalyst. The mutarotation of glucose itself in aqueous solution, is subject to general acid and general base catalysis. Assuming that a similar, mechanism operates here, we may assign to the solvent the role of a base if the, , reaction is carried out in the presence of added acid (reaction 11), and the role of, an acid if the reaction is carried out in the presence of added base (reaction 12)., , (ID, , LXVI, , +, , OH', , +, , BH +, , -> LXIX, , (12), , J., , - (.) Basted and Guggenheim, 7. Am. Cm., 49. 2554, 1925, 1385, 2883; 1927, 2554; Swam and Brown, J. Am. Chm. Soc.., , 2538 (1952)., , ,
Page 559 :
Additions to the C=C Double Bond, , -, , 543, , As yet there is no evidence for catalysis of glucose mutarotation by an acid-base, pair that does not include at least one molecule of water., Next let us consider the conversions of carbonyl compounds to semicarbazones (LXX)oximes (LXXI),«<6) and phenylhydrazones (LXXII).^(c), , —C.(OH)—NHNHCONH2, , —C=N—NHCONHa, (LXX), -h2o, , -C(OH)—NHOH, , -C=N—OH, (LXXI), , I, , -h2o, , ; —C=N—NHPh, (LXXII), , -C(OH)—NHNHPh, , These reactions probably proceed in much the same manner as does the forma¬, tion of hemiacetals. Again, general acid catalysis suggests that the nucleophile, (which we may designate RNH2) may attack either the conjugate acid of the, carbonyl compound or a hydrogen-bonded complex, LXXIII (analogous to, LXII)., \, , slow, , C=0 • • • HA + NH2R —> —C—O • •, , /, , LXXIII, , HA, , fast, , I, , NHoR, H, , I, , —C—O •, , fast, , • • HA ->, , I, NHR, , -h2o, , \, , C=NR + HA, , /, , (RNH2 = NH2NHCONH2, NH2OH, or NH2NH2), The relationships between reaction rate and acidity are complicated here, by an additional factor: addition of appreciable quantities of acid protonates, the basic nitrogen atom of the addendum, converting it to an inactive species., Suppose, for example, that there is just one catalytically active weak acid, HA,, present in solution. Then the rate of the reaction is, rate = (RNH2) (, , 0=0 ) [*H+(H+) + k*A(HA)], , (13), , where kH+ and kHA are the catalytic constants for the hydronium ion and weak, acid, HA, respectively. However, the concentration of nucleophile RNH2 is not, the concentration added to the solution (which may be designated (RNH») „ .), or this base has been partially converted to RNH+. It may be shown (Ex.'4), (1934)! Crossed Fuat'S’ ^22^(1949^(1) B, , (1908). (,) Stempel afd, , Westheimer’ ibid> 56> 1962, ^ ^^ ^^
Page 560 :
544, , -, , Addition Reactions, , that the rate expression may be rewritten, , r\c=o i, L/ J, , (RNH4, nn a, \ J*vx ^ A A l) added, , *h+(H+), “h (H+)/-^rnhj+, , ^ha(H^), , 1, , 1 T" (H+)//Trnh,+, , where /lrnh,+ is the acidity constant of the conjugate acid RNH|. As in equation, (13), the first term in parenthesis refers to catalysis by H30+ and the second term, to catalysis by HA. If we attempt to accelerate the reaction by gradually adding, a strong acid, the first of these terms will increase rapidly at first; but as (H+), approaches, then exceeds, /CRNH,+ (which is 4 X 10-5 for NH2NHCONH2 and, 1.5 X 10-6 for NH2OH), the first term will tend to approach the constant, value, £H+/CRNHa+- At the same time, the second term will be decreasing. On the, other hand, if we gradually add acid HA, the first term will be only slightly, affected, but the second term will increase steadily, since (H/l) in its numerator, will be increasing more rapidly than (H+) in its denominator. Moreover, the, second term may be made to increase still more markedly by adding, along, with acid HA, its salt, Na+T-, for now (£L4) will increase while (H+) may be, kept constant. It should thus be clear why the addition reactions under con¬, sideration proceed most rapidly in the presence of a high concentration of, slightly ionized acid, together with a correspondingly high concentration of, its salt. It may further be shown (Ex. 4) that the most effective acidic catalyst, for such reactions is one having an acidity constant equal to /CRNHa+The formation of oximes is catalyzed not only by acids but also by strong, base,54(6) the base-catalyzed reaction almost certainly involving the attack of, the conjugate base NH2-0~ (or its tautomer HO—NH~) on the carbonyl, group. The additions of semicarbazide and phenylhydrazine should, in prin¬, ciple, also be subject to base catalysis, but such catalysis seems not to have, been reported. It may be noted, however, that semicarbazide and phenyl¬, hydrazine, having no O—H bonds, are much weaker acids than hydroxylamine. At concentrations of base necessary to convert appreciable quantities, of these addenda to their conjugate bases, side reactions such as the aldol con¬, densation, the Cannizzaro reaction, and destruction of the addendum itself, may compete significantly with the desired addition., , Hydride-transfer Reactions, In a number of reductions of aldehydes and ketones to carbinols, the key step, is the transfer of a hydrogen, together with its pah of electrons, from the reducing, agent to the carbon atom of the 0=0 group, thereby converting the carbony, compound to an alkoxide. We have already seen that the reduct,on of ketones, with the ^-hydrogen of a Grignard reagent or a dialkylmagnestum, , compound
Page 561 :
-, , Hydride-transfer Reactions, , 545, , is a hydride-transfer reaction, for which we proposed the cyclic transition state,, LXXIV (p. 403). A very similar transition state, LXXVI, may be written for, the familiar reduction of carbonyl compounds with aluminum isopropoxide, (the so-called Meerwein-Pondorf reduction), but in this case, an a-hydrogen, , is, , transferred. Thus, when the reduction is carried out with deuterium-labeled, isopropoxide, LXXV, the deuterium becomes attached to the a position of the, resulting carbinol.55 The same reaction may be used to oxidize secondary, , \ / DCMe,, , 91 + 91, o, , LXXIV, , I, , —>, , I, Al(«-PrO),, , LXXV, , D\, , CMe,, , l/D, C, , + Me2C=0, , Cxo, , cx, , o, , X, , .XL, , LXXVI, , alcohols to ketones. When the alcohol is added to (XBuO)3A1, it forms an alu¬, minum salt, liberating XBuOH, a weaker acid than the secondary alcohol., When the aluminum salt reacts with acetone, it transfers an a-hydrogen (as, hydride) to the latter, converting it to a salt of isopropyl alcohol, and is itself, oxidized to a carbonyl compound (the Oppenauer oxidation)., — 1-BuOH, , R2CHOH + (/-BuO)3A1, , Me2C=0, ->, excess, , * R2C—o, H, , A1(/-BuO)2, H, , Al(f-BuO)2, , I, , I, , R2C=0 + Me2C—O, Reductions with optically active Grignard reagents (in which the /3-carbon, is asymmetric) are stereospecific.56 For example, when the Grignard reagent, derived from the optically active chloride, ( + )-EtCHMeCH2Cl reduces the, O, ketone MeCCMe3, the resulting mixture of D- and L-MeCH(OH)CMe3 con¬, tains an excess (of about 14 percent) of the dextrorotatory carbinol. This stereo¬, specificity is consistent with the cyclic transition state LXXIV; for it may be, argued that conformation LXXVII, in which the ethyl and f-butyl groups lie, , Irons to each other, is somewhat favored over the alternate activated complex,, LXXVIII, in which these groups lie as. Similar stereospecificity in reductions, “ Williams, Krieger, and Day, J. Am. Chem. Soc., 75, 2404 (1953). For further evident, at this reaction involves a hydride transfer, see Doering and Aschner ibid, 0^°" ^ LXXVI, , 75, , 393 (1953), , Woodward, W^ndter, and*7, 1425, , " Mosher> et a/-, J- Ar"- Chem. Soc., 72, 3994 (1950); 78, 4374, 4959 (1956).
Page 562 :
546, , Addition Reactions, Me, , I ^°—-M$, Et, I, , G, , t-Bu, , SH-, , -Cl, , 'CEL, , -C, , I, Me, LXXVII, , LXXVIII, , with optically active aluminum alkoxides (in this case, with the a-carbon asym¬, metric)57 is likewise consistent with the cyclic transition state, LXXVI. In, either reaction, however, stereospecificity does not, in itself, demand a cyclic, activated complex; for it is quite possible that one conformation of a noncyclic, transition state be energetically favored over others.58 In the opinion of the, present author, the strongest evidence for the cyclic transition state in Grignard, reductions is the inhibition of reduction by added anhydrous magnesium, halides (p. 404). If the attacking hydride were to come from a Grignard mole¬, cule other than that already coordinated to the 0=0 group, magnesium, halides (being more effective Lewis acids than Grignard reagents) should, facilitate the reduction, as they do the Grignard addition reaction which com¬, petes with the reduction. Similarly, it would be of some interest to determine, the effect of added anhydrous aluminum halides on the rate of reductions by, aluminum isopropoxide (in, let us say, toluene or xylene)., The Cannizzaro reaction—the reaction of two aldehyde molecules in strongly, basic solutions to give a carboxylate anion and a molecule of carbinol, , is, , doubtless a hydride-transfer reaction.58 The hydrogen that converts one of the, aldehyde molecules to an alcohol must arise directly from the second aldehyde, molecule, rather than from the solvent, for if the reaction is carried out in, 17 Doering and Young, J. Am. Chem. Soc., 72, 631 (1950)., , « Stereospecificity has been observed—for example, in the conversion of carbonyl com¬, pounds to cyanohydrins—when the reaction is catalyzed by the conjugate acids of optically, active amines (Prelog and Wilhelm, Helv. Chim. Acta, 37, 1634 (1954)). This conversion very, probably proceeds through the noncyclic transition state LXXIX., , \, +/ \, “N=C—^G=0—H-N^, , I*, -> n=C-C-OH, , / \, +, , LXXIX, ..There is evidence that, in some instances .he Cannizzaro, lyrically (see,Tor example, Kharaseh and Foy T., 0.^5^, 5, may be catalyzed, in a manner which is not, oxides or metallic, , ciried oi., , Jnder, , radical initiators., , ,, , V, , ■, , ^, 1100 (1946v j Org., , he^ only with Cannizzaro reactions, .he mo” lal iondirions-tha. is, homogeneously, and m .he absence ol
Page 563 :
Hydride-transfer Reactions, 0,0, , 547, , the alcohol formed has no C-D bonds.- The reaction in some cases, , exhibits third-order kinetics (rate law 15) ,"<•> in some cases fourth-order kinetics, (rate law 16),",w and in still other instances kinetics of mixed order., 2RCHO + OH- -> RC—O- + RCH.OH, , rate = (RCHO)!(OH-), , II, , °r, , o, , rate = (RCHO)2(OH-)2, , (15), (16), , In the third-order reaction (which, the kinetics suggests, passes through an, activated complex consisting of two aldehyde molecules and a hydroxide ion), the rate-determining step is very probably a hydride transfer from the aldehvdeOH, hydroxide adduct, R—C—H, formed in an initial rapidly established equilib-, , o_, rium, to a second molecule of aldehyde., OH, 1, R—C—H, 1^, :0*, , R, , R, , OH, , |, , slow „, ', fast, ^c=o -> R-C + H-C-CT -> RCOO" + RCH2OH, 1, II, (17), H, O, H, , Here, the hydride is, in a sense, “pushed off” by the negative charge on the, carbonyl-hydroxide adduct. The fourth-order rate law (reaction 16), which, applies to the Cannizzaro reaction of furfural and formaldehyde at high con¬, centrations of base, is consistent with two mechanisms. In the first of these, (reaction 18), a hydride is transferred from one carbonyl-hydroxide anion to, another; in the second, the hydride is transferred from the doubly charged, anion LXXX to an aldehyde molecule (reaction 19)., , -V, , OH, R, I, R-C + H-C-0~+ OH-, , o, , or, , O', I, R-C-H, , K, :0:}, • •, , A, , R, + "C^-, , I, H, , ?‘, R—C +, , R, I, H-C-O', , II, , I, , O, , H, , LXXX, 60 Fredenhagen and Bonhoeffer, Z. physik. Chem.,, , ^ Emiy;, , (18), , 181 A, 379 (1938), , (19)
Page 564 :
548, , Addition Reactions, , Path (18) seems extremely unlikely, for when an aldehyde molecule accepts a, hydroxide ion, it loses almost all of its electrophilic character and becomes very, much less subject to further attack by hydride. Moreover, no instances are known, , where a hydride attacks a saturated carbon atom, displacing OH-" On the:, other hand, path (19) is quite reasonable: for anion LXXX, although present, onh in low concentration, would be expected to be an extremely potent hy¬, dride donor. Where mixed third- and fourth-order kinetics are observed, the, reaction presumably proceeds by simultaneous operadon of sequences (17), and (19)., A reaction similar in nature to the Cannizzaro reaction, but occurring in, acid solution, is the “disproportionation" of the diaryl carbinol LXXXII into, a ketone and the parent diarylmethane:54, , 2(MeO)2CHOH C'iCL^°~'> (MeO^-)aC = O + (MeO2CH,, LXXXII, , (not deuterated), , This reaction almost certainly proceeds through hydride transfer to the carbonium ion LXXXII I, derived from the diaryl carbinol,, , )2c ■+*, OH, , OH, LXXXIII, , In contrast to the Cannizzaro reaction, in which the hydride is “pushed off’, by a high concentration of negative charge on the hydride source, the hydride, in the reaction above is “pulled off’ by the positive carbon of the carbonium, ion. In the Meerwein-PondorfF reduction, the hydride involved in the transfer, is subject both to push and pull; for, as shown in structure LXXXIV, the carbon, atom to which the hydride becomes attached bears a partial positive charge,, « \ path sometimes proposed for the (third-order) Cannizzaro reaction (sec, for example,, Geissman in Organic Reactions, VoL 2, John Wiley and Sons, Inc., New \ork, 1944, p. 94),, , C^OH, R, , °, , R, , -> o=C—O —CH2R, , RCOO-+RCH.OH, , R, , LXXXI, involving transition state LXXXI, is open to similar objection. In this case also,, , Cornell University Press, Ithaca, 1953, p. 08, t> Bartlett and McCullom, J. Am. Chem. Soc., / 8, 466 (19d )., , °nJ|
Page 565 :
Stereochemistry of C—O Additions. Cram s Rule, , -, , 549, , whereas the hydride departs from a region bearing a partial negative charge., , LXXXIV, Among the additional hydride-transfer reactions which have been studied, recently are the following, D, , r==, , \, , N—CH2Ph, , Ph2C=S +, , > Ph2CDSI, , _, , /, , A-//, , A+, , NT—(, N—CH2Ph^(a), , (20), , J, , H, , CONH,, , CONH,, 64(b), , (21), , (22), Reaction (20), which requires neither strong acid nor strong base, is thought to, be similar to certain enzyme-catalyzed reductions in which the reducing agent, (the so-called, , co-enzyme”) has a reduced pyridine ring system. Reaction (21), , is a fnm-annular (across-the-ring) hydride shift, somewhat analogous to the, pinacol rearrangement (Chap. 14). Finally, reaction (22) is the first step in the, reduction of ketones by the very useful hydride-transfer reagent, sodium borohydnde; this reaction has been found to be first order in each reactant,', but further information is needed before a reliable guess as to its mechanism, may be made., , Stereochemistry of C=0 Additions. Cram’s Rule, At present we cannot meaningfully speak of tram or as addition to a 0=0, , TV, , pTit, The’ - T;, , '° ^ r0ta“0n 3bOUt the C-° Si"8‘e b°"d - ‘he, , after, conformation of this product persists for only an instant, r its formation. There is, however, a different type of stereochemical probKung,afeO“w^'1 79,712 (1957>- (>), 2H (1957)., , ’, , ’, , (1956)* (c) Brown> Wheeler, and Ichikawa, Tetrahedron, 1,
Page 566 :
550, , Addition Reactions, , lem associated with C=0 additions. Unless the substrate is formaldehyde or a, symmetric ketone (such as acetone or benzophenone), such additions may con¬, vert the carbon of the carbonyl group to an asymmetric center. If the reaction, occurs under “symmetric conditions,” equal numbers of D and L molecules, should form; that is, the product will be racemic. On the other hand, if the, substrate is optically active, particularly if the a-carbon is asymmetric, the, addition reaction may be said to take place in an asymmetric environment., Since there will be two asymmetric carbons in the addition product, two diastereomeric forms (threo and erythro) will be possible and one of these will, generally predominate. In short, the presence of one asymmetric center in the, substrate influences the ratio of isomers formed when a second asymmetric, center is created. Generally, the predominant product is that resulting from an, activated complex in which steric interference is minimal. Consider, for exam¬, ple, the conversion of the ketone PhCHMe—C—Me to a carbinol, using a, , O, Grignard reagent or a hydride donor (such as LiAlH4, NaBH4, or Al(z-PrO)3)., Regardless of the detailed mechanism of the reaction, we may suppose that the, oxygen of the carbonyl group becomes involved in complex formation (with a, magnesium, aluminum, or boron atom) before the carbon of the carbonyl, group is attacked. We would, on this basis, expect the carbonyl group, with, whatever Lewis acid is complexed with it, to take a position as far as possible, from the bulky phenyl group (LXXXV); and we would further expect the, incoming nucleophile (which we may designate as A;) to attack preferentially, on the side near the a-hydrogen, leading to diastereomer LXXXY I, rather, than on the side adjacent to the more bulky a-methyl group, which would, result in the formation of diastereomer LXXXVII. This type of argument may, , (LXXXVI, major product), , (LXXXVII, minor product), LXXXV, A, , / \
Page 567 :
Stereochemistry of C=0 Additions. Cram's Rule, , -, , 551, , be applied in many similar situations and has been summarized by Cram’s, rule of “steric control of asymmetric induction,” which may be stated as o lows- When an addition reaction generates an asymmetric center adjacent to, one already in existence, the double bond in the preferred transition state will, be “flanked” by the two least bulky alpha substituents, and the attacking, nucleophile will approach from the least hindered side of the double bond.'", Schematically, if we designate the substituents on the a-carbon as L (large), M, (medium sized), and S (small), we may represent an addition proceeding according to Cram’s rule as follows:, , This rule would not be expected to apply to heterogeneously catalyzed addi¬, tions, (for example, to catalytic hydrogenations), for the stereochemistry of, these reactions is often determined by the orientation of the substrate as it lies, on the surface of the catalyst. Moreover, the rule may be applied to the addition, reactions of substituted cyclic ketones only with reserve; for here the relative, stabilities of the possible transition states will be determined not only by the, interactions between the a substituents and the addendum, but also by the, conformations (axial or equatorial) of the various ring substituents.*6, An asymmetric center may influence the stereochemistry of an addition, reaction, even if it is several atoms removed from the reaction site. If, for exam¬, ple, the phenylglyoxylic ester of an optically active alcohol (LXXXVIII) is, treated with a Grignard reagent, the a-hydroxy ester formed (LXXXIX) has, a new asymmetric center and yields, upon hydrolysis, an optically active, a-hydroxy acid.67 The situation may be analyzed in much the same way as, , Ph, , O, , O, , II, , II, , C, , C, , R, , */, O—C-f- RMgBr, , h2o, , O, , * I, , II, , OH, LXXXVIII, , ,, , ibid., 74,, , *./, , -> Ph—C—C—O—C—, , 5835 (1952), and Curtin, et at.,, , LXXXIX, , ibid., 74,, , 67 Prelog, et at., Helv. Chim. Acta, 36, 308, 320, 325 (1953)., , 2901 (1952)
Page 568 :
552, , Addition Reactions, , may those additions falling within the scope of Cram’s rule, again designating, the three substituents on the asymmetric carbon in ester LXXXVIII as S,, , M, and L. Because of repulsion between the C=0 dipoles, the most stable::, conformation of the phenylglyoxylic ester may be assumed to be LXXXVIII',, in which the C, , O groups lie at an angle of 180°. Again, we may suppose that:, , the alkyl group of the Grignard reagent will attack the keto group preferentially, from the least hindered side, that the predominant form of the a-hydroxy ester, formed will be LXXXIX, and that this ester will give, upon hydrolysis, acid, , XC. This conversion is of interest, for it allows us to relate the configuration, about the asymmetric carbon in acid XC with that about the asymmetric car¬, bon in carbinol XCI, and further allows us to compare the configurations of, two optically active alcohols having asymmetric a-carbon atoms. Let us sup¬, pose that dextrorotatory ^c-butyl alcohol (the configuration of which is thought, to be XCII)/* and an alcohol of an unknown configuration are converted to, their respective phenylglyoxylic esters. Suppose further that both esters are, treated with CH3MgBr, that the a-hydroxy esters which result are then hydro¬, lyzed, and that the directions of optical rotation of the acids formed in the, hydrolyses are compared. If rotation of the two hydroxy acids is of the same sign,, showing that the direction of stereospecificity is the same in both cases, we may, assume that the unknown alcohol, like (+)-**-butyl alcohol, has a configuration, which may be represented as XCIII. Conversely, if the direction of stereo¬, specificity is different for the two series of conversions, the configuration of the, unknown alcohol would, in all probability, be XCIV., , XCIV, XCII, ti Kirkwood, J. Chem. Phys., 5, 479, , (1937); Kuhn, Z.physik. Chem., B31, 23 (1936).
Page 569 :
Exercises for Chapter 13, , EXERCISES, , FOR, , CHAPTER, , -, , 553, , 13, , 1. Propose a mechanism for the acid-catalyzed hydration of acrolein,, CH2=CH—CHO + H20 ^, , HO—CH2—CHoCHO, , consistent with the following observations:, (a) The rate is much greater than the rates for hydration of 2-butene or crotyl alcohol, under similar conditions., (b) The rate is proportional to (H30+), rather than to ho., (c) The entropy of activation is much more negative than that for the acid-catalyze, hydration of 2-butene or crotyl alcohol., , 2., , The rate of addition of Br2 to stilbene in methanol is decreased by adding Br~ to the, reaction mixture, due to the formation of the less electrophilic Br3 anion. The appar¬, ent specific rate for this addition may be defined, — d(S)/dt, SPP “, , (S)(Br2)added, , where (S) is the concentration of stilbene and, , (Br2)ftdded, , is the stoichiometric concentra¬, , tion of added bromine (present both as Br2 and Brj). It is found'7 that £app varies with, (Br~) as follows:, fl(Br-) + bK, RPP, , (23), , K + (Br~), , where a and b are constants and K is the dissociation constant of Br3 into Br2 and Br ., (a) Show that this relationship is consistent with the following mechanism:, , OMe, and that k\ = a and k2 = b., (b) Show that the fraction of product diverted to the a-bromoether is, , 1 + Y (Br~), , (c) When (Br~) = 0.1, the fraction of a-bromoether is 0.90. When (Br~) is raised to, 0.2, this fraction falls to 0.82. Calculate two values for the ratio k»/kt., (d) The value of K under the reaction conditions employed is 0.0024. Values of k&i>p in
Page 571 :
Exercises for Chapter 13, , (r) Br2 to PhCH=CH—CH=CHPh or to PhCH2—CH, , CH, , 555, , CH-.Ph in nitro, , benzene?, , in dilute HO Ac ?, (t) Hydroxylamine to cyclohexanone in the presence of an equimolar mixture of, HOAc and OAc- or an equimolar mixture of Cl2CHCOOH and Cl2CHCOO~, in water?, (u) Maleic anhydride to furan or to benzofuran in chlorobenzene?, 4. (a) Consider the general-acid-catalyzed addition of nucleophile RNH2 to an aide-, , /, , hyde or ketone ^designated, , \C—O, _ .J. Show that under conditions where the, , added nucleophile exists both as RNH2 and RNH^, the rate of addition is given, by equation (14), page 544., (b) Assume that in this case, the first term in parenthesis in equation (14) (represent¬, ing catalysis by H30+) may be neglected in comparison with the second. Suppose, now that this reaction is carried out in the presence of a fixed quantity of weak, acid, HA, with enough base added to convert a portion of this acid to A~. Show, that under these conditions, the rate will be a maximum when enough base has, been added so that, , (H+), , —, , RNHjj^, , (Hint: remember that (H<4) + {A ) remains constant, and differentiate.), (c) Assuming that a buffer system may work effectively when (H.4) lies near (A~),, show that the most effective catalytic acid for this reaction is one for which Kha, lies very near A'hnh^., 5. Suggest mechanisms for the following conversions:, , +, , HBr, , PhNO,
Page 575 :
Exercises for Chapter 13, H, , 559, , Me, +, , EtMgBr, , O, Me, , OA1(s-BuO)2, , (q), , +, , H, , >, , Ph—C—CH3, , II, , o, , Et, , 7. Outline a series of experiments to show that the addition reaction, R' O, , ,/, , I II, , ./, , R—G—C—O—C-h R'MgBr -* R—C—C—O—C—, , II II, , O, , O, , *\, , I, , OMgBr, , *\, , proceeds stereospecifically in the direction dictated by minimum steric interaction, between R' and the groups on C« while the keto ester lies in conformation LXXXVIII', (p. 552). Assume that the absolute configurations of the reactants and products are, unknown., 8. Explain each of the following observations:, (a) The rate of formation of ethylene bromide from the reaction of bromine with, ethylene in water is increased by the addition of NaBr, but the rate of disappear¬, ance of the olefin is decreased., (b) The rate of addition of hydrogen halides to olefins in nitrobenzene is increased by, the addition of SnCl4., (c) The rate of addition of Br2 to cinnamic acid in water is decreased by the addition, of small quantities of dilute HN03., (d) IC1 adds more rapidly to stilbene than does Br2 under comparable conditions., (e) Cyclopentadiene undergoes the Diels-Alder reaction more readily than does, 1,3-butadiene., (f) The low temperature addition of HC1 to isoprene gives CH3CMeClCH=CH2, when equimolal quantities of reactants are used. However, with a slight excess of, HC1, the chief product is CH3CMe=CHCH2Cl., (g) The dimerization of cyclopentadiene is catalyzed by trichloroacetic acid more, effectively in CC14 than in dioxane., (h) The conversion of an aldehyde to a hemiacetal in water exhibits general acid, catalysis, but the further conversion to an acetal is catalyzed only by H30+, (i) When benzaldehyde undergoes the Cannizzaro reaction, one of the products, is benzyl benzoate, although benzyl alcohol and benzoic acid do not react to, °rm this ester under the reaction conditions employed., (j) The rate of hydration of methylenecyclobutane (in aqueous HNOs) is nearly the, same as that for methylenecyclopentane, although the equilibrium constant for the, former hydration is about 10,000 times that for the latter.
Page 577 :
CHAPTER, , 14, , Participation of Neighboring Groups, in Nucleophilic Substitution, Reactions and in Rearrangements', , In Chapter 7 we considered the ways in which the reactions of an organic, molecule could be accelerated or retarded by groups that lie near the reaction, site but, nevertheless, do not participate directly in the reaction. We attributed, the action of such groups to inductive, conjugative, and steric effects, or to com¬, binations of these. In the present chapter, however, we shall be concerned with, reactions where neighboring groups become more deeply involved—where, in, fact, such groups become bonded (fully or partially) to the reaction center for an, interval of time during the reaction’s progress., There are three principal types of evidence that point to neighboring group, participation. First, if such participation occurs during the rate-determining, step, the reaction is almost certain to be significantly more rapid than other, reactions that are similar but do not involve such participation. Typically,, the 0-chloro sulfide, ClCH2CH2SEt, is hydrolyzed over 10,000 times as rapidly, as is the corresponding ether, ClCH2CH2OEt (in aqueous dioxane).* This rate, difference is far too great to be attributed to differences in inductive, conjugative,, or steric effects, but suggests rather that the hydrolysis of the sulfide (but not, *e ether) proceeds through a cyclic ’onium ion (in this case, sulfonium ion I)., Ihe intermediate, because of the strained three-membered ring, is, hydrolyzed to the observed products., , », , ,, , $5“’, , BU“■, , readily, , 18> 055 <>«»• and, , ' Bohme and Sell, Ber., 81 123 (1948)., , S61
Page 578 :
562, , -, , Participation of Neighboring Groups, , •S*‘", , ~^CH2-^C1, , -CH2, ,, /, ch2, , /\CH2/, Et, , H2o, very fast, , EtSCH2CH2OH ■+■ H+, , I, Secondly, the stereochemistry of a reaction might suggest that neighboring, groups become involved. In an earlier chapter, we saw that the hydrolysis of the, a-bromopropionate ion in water or dilute base yields lactate with retention of, configuration about the a-carbon (p. 270). Since nucleophilic substitutions at, secondary carbon atoms almost invariably result in partial or complete inversion, of configuration, it was assumed that two displacements were actually involved:, , O, , I, , the first a displacement of bromide by the neighboring —C—O- to form the, nonisolable a-lactone (II), and the second a very rapid cleavage of the lactone, by water. Inversion presumably occurred in both displacements, the second, inversion “nullifying” the first., Me, , Me, , Finally, neighboring group participation may lead to molecular rearrange¬, ment when the neighboring group remains bonded to the reaction center but, breaks away from the atom to which it was originally attached in the substrate., Thus, the chlorinated amine III yields, upon basic hydrolysis, the rearranged, aminohydrin, V, presumably because the intermediate imonium ion, IV, is, attacked preferentially at the primary a-carbon atom rather than at the second¬, ary /3-carbon atom.5, , CHEt-^-Cl, , •Et2N, , OH", CHEt ->- Et,N-, , CH,, , \u/(i, CH*, , III, , IV, , \«, , /», , C1-> E.t2N, , -CHEt, Nch2oh, , a, , A reaction that is accelerated by neighboring group participation is some¬, times said to be anchimerically assisted,' for such a reaction proceeds throng, what is, in effect, an ‘•internally attached form” of the substrate (for example,, structures I and IV)., * Ross, J. Am. Chem. Soc., 69, 2982 (1947)., I Winstein, Lindegren, Marshall, and Ingraham, J. Am. Chem. Soc., 75,, , (1953).
Page 579 :
Intramolecular Displacement by Oxygen, , Intramolecular Displacement by Oxygen, The nature of neighboring-group reactions was first elucidated through experi¬, ments involving participation of the carboxylate group /—C—O \, particular y, , O, through studies of the hydrolysis and alcoholysis of the anions derived from, a-halocarboxylic acids. In discussing a typical reaction of this sort, the hydrolysis, of a-bromopropionate (p. 270), it was suggested that the carboxylate group does, not displace the a-bromine directly, but rather that the formation of the a-lactone (II) follows the ionization of the C—Br bond. This conclusion arises from, the observed catalysis of the reaction by Ag+,5(o) its acceleration when the degree, of branching at the a-carbon is increased,5(6) and the observed positive kinetic, salt effect5(c) (each of these features being characteristic of ^Arl-like, rather than, iSi\r2-like solvolyses). However, although direct displacement may be difficult,, due to the strain involved in forming a three-membered ring, the departure of, Br- is doubtless facilitated by a combination of the inductive and field effects of, the —COO- group. As soon as the bromide ion has receded a bit from the, a-carbon, the energy needed to form the three-membered lactone ring becomes, available from the neutralization of charge upon formation of the Ca—O bond.*5, As the distance between the halogen atom and the carboxylate group is, increased, the cyclic structure involved in direct intramolecular displacement, becomes much less strained. Reaction by displacement becomes favored, and, solvolysis is no longer facilitated by branching at the halogen-bearing carbon or, by the addition of Ag+. Thus the conversion of, , 7-bromovalerate, , in water to the, , five-membered ring lactone, VII, proceeds by direct displacement,7(o) and the, same is probably true for the formation of /3-butyrolactone (VI).7(6) But, as is the, case with other types of cyclization, formation of rings of seven or more members, entails some difficulty, due to the low probability for collision between the, opposite ends of a long chainlike molecule. When e-bromocaproic acid (VIII), is treated with Ag20 in water, ordinary solvolysis to give the. €-hydroxy acid, Ch‘ W, “nd InSold" J- Ch"»• «*•. W38,1243. (») Lane and Heine, J. Am., Chem. Soc., 73, 1348 (1951). (c) Grunwald and Winstein, ibid., 70, 841 (1948)., a fa t Jt, T*? I°U ud ^Ref' 5? that S°lvolysis of the «-bromopropionate ion is accelerated by, factoSr„?I °n y I i, "t ", transferred from methanol to water, in contrast to rate increases by, fared fmm m«i 1,, ^ Solv°lySeS of "dinary secondary alkyl bromides are transerred from methanol to water; that is, the presence of the —COO' group lowers the sensitivitv, of the solvolysis to the dielectric constant of the medium. This inSes tha" Jhe degree of, arge separation m the transition state is much less for solvolysis of the bromopropionlte ion, whTeHf0:,S0lV°lyS1S °f’uSay’ 1SOpr°Pyl chl°ride and confirms the picture of a transitfon state in, * ‘h‘, , 55 MHimw.nS Mori, J Biol. Ch,m, 4364 (19(16)7 ' ( >, , <• partially neufraliaed by interaction, , 78, 1 (1928). (b) Johansson and Hagman Be,, , arUd’ “ < J- Am- Ctum■, , «. 2«8 (1924). W von Braun,
Page 580 :
564, , -, , Participation of Neighboring Groups, , competes with neighboring-group participation (which gives the seven-membered ring lactone, IX)/<‘> Anion X yields only a hydroxy acid (no eightmembered ring lactone).7^), , o, H2o, , CH3CHBrCH2COO~, , -o, Me, VI, , CH3CHBr(CH2)2COO, , VII, , O, Br(CH2)5 COOH, , Ag2Q, , +, , h2o, , HO(GH2)5COO‘, , VIII, IX, h2o, , *-, , Br(CH2)6COO", , H0(GH2)6C00~, , (no lactone), , X, When the carboxylate group is converted by protonation to the carboxyl, group, —COOH, it becomes very much less nucleophilic and loses a great deal, of its effectiveness as a participant. The substitution of a carboxyl group for an, a-methyl in (CH3)2CHBr, , lowers, , the rate of solvolysis in water about a hundred¬, , fold,5^ whereas a similar substitution in PhCHMeBr, , retards, , solvolysis by a factor, , of about 105.5(6) If any anchimeric assistance effects are present, they are com8 Although neighboring-group participation in the solvolysis of salts of simple 5-halocarboxylic acids has not yet been demonstrated, it almost certainly occurs in the hydrolysis of, anion XI (which is analogous to the salt of a 5-halocarboxylic acid in that the chlorine atom is, separated from the carboxylate group by a chain of four atoms). It is found that anion XI may, be easily hydrolyzed in boiling water although its para isomer, XII, is inert under the same, conditions (Bordwell and Cooper, J. Am. Chem. Soc., 79, 916 (1957)., , c (a) Winstcin, Grunwald, and Jones, J. Am, J. Chem. Soc.,, , 1937,, , 343., , . Chem. Soc., 73, 2700 (1951). (b) Taylor,
Page 581 :
Intramolecular Displacement by Oxygen, , 565, , pletely overshadowed by the strongly negative inductive effect of the —COO, group. It has been found that the “deaminations” of a-amino acids proceed with, retention of configuration/0 but in this case it is not clear whether participation, D-RCH-COOH + N2, I, OH, , D-RCHC1—COOH + N2, by the —COOH or the —COO“ group is involved, since the intermediate,, Z)-RCH—COOH, is a strong acid., I, +N=N, Whether or not an ester linkage functions as a neighboring group depends, largely upon how it is situated with respect to the reaction center in the substrate., The solvolyses of CH3CHBrCOOEt6(o) and the deamination of CH3CH(NH2)COOEt10 result in inversion of configuration about the a-carbon, suggesting, that there is negligible participation by the —C—OEt group when the reaction, , O, center lies alpha to the C—O bond of the ester. This is what we might expect, for, such participation would require the formation of a strained three-membered ring, -C—OR\without any compensating charge neutralization. On the, , —C-, , O, other hand, ester participation has been observed in a number of substitutions in, which the reaction center is located in the alkyl, rather than the acyl, section of, the ester molecule; in each of these, the anchimeric intermediate contains a, five-membered ring. We have already considered the stereochemical evidence, that the conversion of /nm-2-acetoxycyclohexyl brosylate, , (XIII), , to trans-, , diacetoxycyclohexane proceeds through the symmetric intermediate acetoxonium, i°n’ XV (P- 271)> and similar evidence points to the same intermediate in the, reaction of toj-2-acetoxycyclohexyl bromide (XIV) with silver acetate.11 As, indicated, addition of small quantities of water to either reaction mixture diverts, some of intermediate XV to the monoacetate of m-l,2-cyclohexanediol (XVI), presumably through intermediate XVI., The formation of XV from brosylate XIII or bromide XIV almost certainly, , ;r, , \rct dispiacemem, , ^ by ** acet<*y iyi„g trms ,0, , neighboring group being in a position favorable for attack. In contrast, the, n wCWStCr’ Hir°n’ Huehes> Ing°H, and Rao, Nature 166 178 HOtm, m„ 80,, , Am- CW**•. ««. 2780, 2787 (1942). See also Roberts,„ a,..
Page 582 : Participation of Neighboring Groups, Me, , XIV, , XVI, , XVII, , solvolyses of the cis isomers of XIII and XIV are governed by the rate at which, these substrates lose brosylate or bromide ion, such ionizations being, in these, cases, anchimerically unassisted. Hence, it should not surprise us that XIII and, XIV are solvolyzed several hundred times more rapidly than their respective, cis isomers under the same conditions., There is evidence that the acetoxonium ion XV is also an intermediate in, the conversion of the aV-monoacetate XVII to ^rani'-2-acetoxycyclohexyl chloride, (XVIII) with concentrated HC1, a reaction for which the following path has, been proposed:1S, HO, , ch3, 1, G, , \, , IX, , ✓ xo, 1, , yf, , H+v, , °\ /X,, , \, XVII, , Me, 1, , Me, , Me, , \ /, , K, , 1, y/, , /, , <A>, T, , 1, , /, , «, , Cl, , XVI, , XV, , XVIII, , The difficulty in carrying out a similar conversion with the Irons isomer of XVII, (which cannot form intermediates XVI and XV) and the difficulty in converting, cyclohexanediols to chlorohydrins in the absence of acetic acid tend to support, this mechanism., ., The negative oxygen in alkoxides is one of the most familiar neighboring, groups, its participation leading to alkene oxides, many of which are readiy, isolable. The basic hydrolysis of ethylene chlorohydrin, HOCH,-CH,C,, yields ethylene oxide, and the hydrolyses of more complex, 1,2-chlorohydr.n, » Boschan and Winstein, J. Am. Chem. Soc., 78, 4921 (1956).
Page 583 :
567, , Intramolecular Displacement by Oxygen, , often yield substituted ethylene oxides. These ring closures exhibit second-order, rate laws:13, , -c- +, , -C, , I, , OH-, , I, , OH, , -C-h Cl, , —C-, , + H2O, , O, , Cl, , (OH-), , (1), , which point to the following mechanism:, , slow, ■->, , OH~n, , 4- cr, , fast, eq, , Intramolecular displacements forming epoxides should occur more readily than, those forming a-lactones, for alkoxide oxygens are far more basic than carboxylate, oxygens, and the epoxide ring is somewhat less strained.14 Moreover, it has been, found that intramolecular displacements leading to epoxides are, as a group,, thousands of times faster (under comparable conditions) than their intermolecular counterparts, the attack of alkoxides on alkyl chlorides.ls(6,c) Although the, formation of an epoxide is (because of ring strain) somewhat less favored ener¬, getically than the formation of a noncyclic ether, far less restriction of motion is, necessary in closing a small ring than in bringing together two free molecules, into a single activated complex. The ease of formation of epoxides is then an, entropy effect, and the same effect facilitates the formation of three-membered, rings containing nitrogen or sulfur (for example, cations I and IV)., Somewhat surprisingly, epoxide formation is markedly facilitated by alkyl, substitution, as shown by the following comparison of rate constants (water, , 18°) :1S(c), HOCH2CH2Cl, , HOCHMeCH 2C1, , HOCMe2CH2Cl, , HOCMe2CMc2Cl, , 250, , 1370, , k, *HOCHjCHjCI, , *‘0, , This trend is contrary to what would be expected from consideration of intermolecular substitution, and there is some question as to how it may best be, 2449'7o1o^mt81nAnNIrUCaS,, MTu -948 ‘ ( ) Nlkson and Smith,, bonH, , 1, , 1, , S°,C•’ 61» 1576 (1939>- <*> Warner>, 166A, 136 (1933)., , 70,, , Z.physik. Chem., , mcourporatiion °f a carbony1 carbon into an a-lactone ring requires squeezing of the, , fr°m 120°‘° ?OU,,60°- ^ncorP°rat'°n of ag„ a.qkox;deqca'ta„, poxiae ring requires a decrease in bond angle from 108° to about 60°., , fmo, , an
Page 584 :
568, , -, , Participation of Neighboring Groups, , accounted for. At present, the most satisfactory explanation is that steric crowd-., , ing in the parent chlorohydrin is somewhat relieved as the epoxide is formed, for ■, reduction of one of the bond angles about a carbon to 60° allows the other bond, angles about that carbon to “open up" well beyond their “normal value” of, 108 . There is thus less interference between groups bound to the same carbon, in the epoxide than in the chlorohydrin, and the greatest relief of strain will, presumably result from ring closure of the most crowded chlorohydrin.75, In intramolecular displacements, as in their intermolecular analogs, the, entering nucleophile must approach the reaction center from one side while the, leaving group departs from the other, resulting in an inversion of configuration., This means that at the instant before the closure of chloroalkoxide ion XIX, the, oxygen atom must lie trans to the chlorine. Thus, fra/zj-2-chlorocyclohexanol can, be converted by base to cyclohexene oxide, but its cis isomer cannot.16 Due, to freedom of rotation about C—C single bonds in noncyclic molecules, both the, erythro and threo diastereomers of an aliphatic 1,2-halohydrin may be converted, to epoxides; but since both forms must, during reaction, adopt conformations in, which the oxygen and halogen lie trans to each other, the conversions are stereo¬, specific. As shown below for the 3-bromo-2-butanols,73(o) the erythro-halohydrin, yields a trans-oxide whereas a //zrai-halohydrin yields a m-oxide:, \, Me, , HO, , H^S)—(§^, , H, , Br, B, , Me, , erythro, , H, ->, , q, , C-C, , /, , \, , Me, , H, , Me, , \ / \ /, H, , \, , /, , -”1, , >, , C, /, Me, , trans, , H, , C>, , Me, CIS, , If the basic hydrolyses of the above bromohydrins are allowed to go to comple¬, tion—that is, if the epoxides are not isolated—opening of the epoxide rings will, occur with inversion of configuration. The result of the overall conversion will be, the transformation of the ^ry/Aro-bromohydrin to a meso-glycol, and that of the, Mn?o-bromohydrin to a d,/-glycol—in short, net retentions of relative configura¬, tions.77 This stereochemical outcome would have led us to suspect intervention of, the epoxide, even if the latter were not isolable. As with the carboxylate group,, protonation of the —0~ group to —OH greatly lowers its tendency to partici16 In support of this explanation, apparently first suggested by Beesley, Ingold, and Thorpe, (J. Chem. Sac.,, , 107,, , 1080 (1915)), it has been found that the H—G—H angles in, , cyclopropane, , are 118°, rather than the “usual” 109° (see Donohue, Humphrey, and Schomaker, J. Am., , Chem. Soc., 67, 332 (1945))., 16 Bartlett, J. Am. Chem. Soc., 57, 224 (1935)., ti n, n The term “relative” is used here to distinguish the stereochemical result from retenno, of absolute configuration such as occurs, for example, in the hydrolysis of the salts o a- a, carboxylic acids by dilute base. Although we may start with an optically active ^r«>-bro, , ^, , hydrin, the resulting glycol must be racemic, for the intermediate m-epoxi e, P, ,, symmetry. However, the configurations about the two asymmetric carbons ***JJP^, will be the same (both D- or both L-), in contrast to the erythro-bromohydnn and t, glycol derived from it, in which the two asymmetric carbon atoms in a molecule, configurations., , PP
Page 586 :
570, , -, , Participation of Neighboring Groups, , nucleophilic substitution results in retention of relative configuration about the, , reaction center.^ The anchimeric intermediate, as shown below for the:, , acetolysis of ^nj-2-methoxycyclohexyl bromide (XXII) is a bridged methoxonium ion (XXIII)., , Neighboring Nitrogen, Sulfur, and Halogen, , Without considering the possibility of neighboring-group interaction, we might!, expect such halogenated amines as Br(CH2)4NH2 and Cl(CH2)2NMe2 to behave, in substitution reactions in much the same way as do ordinary alkyl halides., Specifically, we might predict that these amines be hydrolyzed in aqueous base, at about the same rates as are simple primary halides (for the inductive effects of, the —NH2 and —NMe2 groups are slight), and that these hydrolyses be, first order in hydroxide. It is found, however, that treatment of these amines, with aqueous base releases halide ion thousands of times as rapidly as the hy¬, drolyses of nonaminated alkyl chlorides, and that the formation of halide is, zero order in (OH").40'*1 Yet, in spite of their kinetic character, these reactions, are not ordinary Sn^ solvolyses, for they are not accelerated by the addition of, Ag+; moreover, there is no obvious reason why the presence of a nitrogen atom, in the substrate should greatly facilitate the usual type of solvolytic ionization., Instead, the rapid release of halide ion in these cases is due to internal attack by, the nitrogen atom at the halogenated carbon, forming a cyclic ammonium ion, (such as XXIV or XXV). As indicated, the five-membered ring ammonium, , h2c, , H, , H, , H, , /N:>chA, , \, , h2c, , *, , /, -ch, , +, , H, , H,C, \, , CH2, j, , (stable), , +, , Br, , h2c—ch2, , 2, , XXIV, , -cr, CH„— Cl, , Me,N:, ch;, , ^, , =>, , +, , OH" or, , CH., , Me2N;, , H,0, , ch2, XXV, Me,N—CH2—CH2OH, , ru, , to For studies of the reactions of bromoalkylamines, see: (a) Freundlich el'al., Z-P^', 7K, a01IV 79 681 0912) • 87, 69 (1914); 101, 177 (1922 ; 122, 39 (1926); 166,161, (1933); 17, 851 (m4>;19,, , -Jo, , 11 The substitution reactions of 2-dialkylaminoalkyl chlon “(eso, Cfow»., mustards”) have been investigated by a number of workers See. W, A 948) - 74, 1875,, ‘, ca 0071 0077 (\ 947V 71 1415 (1949); (b) Cohen, et al., ibid., 70, 2ol, /*», 69,2971, 2977, /, 550 0946);, 1787 (1952); (c) Golumbic, et al., J. Org. Chtm., 11, 518, 500, oou, Can. J. Research, 26B, 181, 193 (1948)., , (d), K ), , Thompson, et al.,
Page 587 :
Neighboring Nitrogen, Sulfur, and Halogen, ion (the pyrrolidinium ion, XXIV) is stable. Ethyleneimonium ions (such as XXV), are, as a class, somewhat more reactive than ethylene oxides, although salts containing such ions can, with care, be isolated., As might be expected, the rate at which the various cyclic ammonium ions, are formed depends markedly on the number of atoms in the ring. Rings of five, and six members are readily formed, not only because they are stable, but also, because collision of a nitrogen on one end of a molecule with a halogenated, carbon atom four or five atoms removed is a fairly frequent occurrence (thus the, formations of such rings have high entropies of activation). On the other hand,, since the bond angles in a four-membered ring lie near 90°, the formation of this, ring will involve considerable strain (that is, a high energy of activation) and, will therefore be slow. Formation of rings having seven or more members will, likewise be slow, not because they are unstable, but rather because collisions, between two groups separated by six or more atoms in a chain are rare. Finally,, despite the instability of three-membered rings, ethyleneimonium ions, like ethyl¬, ene oxides, will form with ease since the neighboring group engaged in the, attack lies so conveniently close to the reaction center. Like ethylene oxides,, ethyleneimonium ions are of particular interest, for they are formed readily and, react readily, thus furnishing easy paths for reactions that would otherwise, proceed much more slowly. Intervention of ethyleneimonium ion intermediates, accounts not only for the great ease with which 2-haloalkylamines are hy¬, drolyzed, but also for the readiness with which they react with a host of nucleo¬, philes, among them amines, carboxylate ions, and such inorganic anions as, S202- and SO|~.w These imonium ions undoubtedly also intercede in rearrange¬, ments, not only of such aliphatic haloalkylamines as III (p. 562), but also of, such heterocyclic amines as N-ethyl-2-chloromethylpyrrolidine (XXVI) :gs, , I, , — presence of base, are zero order in, ** Ross> J- Chem. Soc.,, , 1949,, , 2589., , 1 uson and Zirkle, J. Am. Chem. Soc.,, , 70,, , 2760 (1948).
Page 588 :
572, , -, , Participation of Neighboring Groups, , base.*-*4 These reactions doubtless proceed through cyclic sulfonium ions of the:, type XXVII and XXVIII, analogous to the cyclic ammonium ions XXIV and, XXV:, R, , R, , \XS^/A, , /, , > R-S, , /G, , nTA, , C —Cl, , xc +, , \ / \ /, -> G, G, , ' \, / \, -—G—C-—, , x \, / \, —-C—C—, \ /, , \ /, XXVII, , XXVIII, , The most familiar (and the most extensively studied) chloroalkyl sulfide is, “mustard gas” (j8,/3'-dichlorodiethyl sulfide), the hydrolysis of which in/olves, the formation of the ethylene sulfonium ion, XXIX, followed by the much faster, destruction of this ion by water. As befits a reaction sequence of this type, the, first-order rate constant for the hydrolysis will “drift downward” as the reaction, proceeds, due to the reversal of the cyclization when intermediate XXIX is, attacked by Cl- accumulating in solution (the mass law effect, see p. 256). More¬, over, if the reaction is carried out in the presence of an additional nucleophile,, even one as weakly nucleophilic as formate or chloroacetate, much of inter¬, mediate XXIX may be converted to a substitution product (for example,, formate XXXI) rather than to the hydrolysis product, XXX:, starting material, , CH, 2-ch-ci, , _cl-, , \, , CH2— ch2—Cl, , CH., +/ "\CH, , * s-, , CH2— ch2-oh, > S, , l, CH0- CH, , \, , ch2—ch2-ci, XXX, , o, , mustard gas, XXIX, , ch2-ch2-o-c-h, /, , rate determining, , \, , product, determining, , CH2—CH2—Cl, XXXI, , The fact that such relatively weak nucleophiles as chloride and formate may,, , even at low concentration, compete with the solvent water for the cation*, intermediate is further evidence that this intermediate is indeed a sul, « For mechanistic studies of the reactions of chlorosulfidra see^ W Bartlett Md Swam,, , n,, r., 7| 140'S n 949V (b) Oeston, et at., Trans. Faraday Z>oc., 44,, t, ., ttCffZJ cd!a£! 1927, l«76i 1929, 25671 .938, 813; M Goering and Howe, X, _, , ., , Chem. Soc., 79, 6542 (1957).
Page 589 :
Neighboring Nitrogen, Sulfur, and Halogen, , -, , 573, , ion, rather than the very much more reactive primary carbomum ion, C1CH2CH2SCH2CHt. While the latter is not excluded by the kinetics, its lifetime, would be extremely short, and all but a very small fraction of such ions would be, converted to the hydrolysis product, XXX (by action of a solvent molecule fiom, their solvent “cages”) before colliding with a nucleophilic species present in, small concentration in solution., As we have noted, anchimeric assistance by negatively charged oxygen or, by an ester linkage requires that the assisting group and the leaving group lie, tons to each other. The fact that the trans forms of 2-chlorocyclohexyl phenyl, sulfide (XXXII) and the corresponding cyclopentyl derivative are hydrolyzed, 105 to 106 times as rapidly as their respective cis isomers indicates that trans, orientation is required for /3-sulfur participation a\so.S4(d) But trans orientation,, SAr, , although necessary, is apparently not sufficient; for it has been found that the, trawj-chlorosulfide XXXIII is solvolyzed (in 85 percent aqueous ethanol) at, very nearly the same (very low) rate as is its cis isomer, and that the transchlorosulfide XXXIV is solvolyzed similarly slowly.*5 The reader will recognize, the ring systems in compounds XXXIII and XXXIV as those present in certain, dichlorides that were considered in Chap 12. It will be recalled that the rigidity, of the ring systems in these compounds prohibited the formation of the transition, states necessary for trans elimination, thus leading to very low rates of bimolecular dehydrochlorination, even when hydrogen and chlorine atoms on adjacent, carbons lay trans to each other. Since the same rigidity of structure in compounds, XXIII and XXXIV will prevent a coplanar arrangement of the four atoms, involved m formation of the sulfonium bridge (the sulfur atom, the chlorine, atom, and the a- and 3-carbons), it is reasonable to infer that the stereochemical, , ~",S, <3'SUlfUr, (and Probab,y, by 3 n troge f°rr“, and d-halogen aSSiS‘anCe, also) are by, similar, in naturc, (Q those for, for assistance, ^ b;_, ten,a 7, ‘"T"' AUhouSh more data are clearly needed, it may be, ten,attvely assumed that effective anchimeric assistance requires a transition, Bristol and Arganbright, J. Am. Chem. Soc., 79, 3441 (1957).
Page 590 :
574, , -, , Participation of Neighboring Groups, , state in which the a- and 0-carbons, the leaving group, and the “assisting atom” ', lie in or near a common plane., Neighboring-group participation by a 0-bromo or 0-iodo group results in, , the formation of a cyclic (three-membered ring) bromonium or iodonium ion of the :, type discussed in connection with the addition of halogens to olefins (p. 523)., Although such halogenonium ions are too reactive to be isolated, the stereo¬, chemistry of conversions of the diastereomeric 3-bromo-2-butanols to dibromobutanes,*6 and that of the acetolyses of these dibromides in the presence of, Ag+,n point to the intervention of bromonium ion XXXV in the erythro-meso, series, and to the intervention of the cis isomer of XXXV in the threo-d,l series., , XXXVI, erythro, As shown, the conversion of the bromohydrin to the dibromide and the conver¬, sion of the latter to the acetoxy bromide (XXXVI) involve retention of relative, configuration. Retention of configuration is also observed for the corresponding, reactions in the threo-d,l series., Since chloride is considerably less nucleophilic than bromide, the (3-c, group would be expected to be a less effective “participator” than the, , o, , 0-bromo, , group. Although some reactions appear to proceed through a chloromum 1, bridge (see, for example, p. 525), there is no evidence that a chloro group rende, appreciable anchimeric assistance. On the other hand, part.ctpat.on by the very, « Winstein and Lucas, J. Am. Chem. Soc.,, , 61,, , 2845 (1939).
Page 591 :
Aryl Participation. The Phenonium Ion, , -, , 575, , nucleophilic iodine atom may greatly accelerate a reaction. It has been found, for example, that the acetolysis offrani-2-iodocyclohexyl brosylate (XXXV11), is almost 3 million times as rapid as that of its cis isomer, quite obviously because, the reaction of the trans (but not that of the cis) compound may proceed through, iodonium ion XXXVIII/7 With the corresponding bromobrosylates the trans, , XXXVII, , XXXVIII, , isomer is likewise solvolyzed more rapidly than the cis, but here the effect, {ktranjkcu, , = 800) is much less striking., , Aryl Participation. The Phenonium Ion, The neighboring groups considered thus far are nucleophilic in the classical, sense, for each has unpaired electrons available for coordination or displace¬, ment. As we have emphasized on several occasions, the aromatic ring also has, nucleophilic character, and we might therefore expect that, under favorable, conditions, neighboring-group participation by aryl groups might be observed., Strong evidence for phenyl participation arises, for example, from the stereo¬, chemistry of the acetolyses of the stereoisomers of 3-phenyl-2-butyl tosylate,, PhCHMe—CHMe—OTs/8 Acetolysis of the threo tosylate (XXXIX) yields a, threo acetate which is, however, racemic whether or not the original tosylate, , was optically active. On the other hand, acetolysis of the erythro tosylate (XL), proceeds with retention of configuration (both relative and absolute), the result¬, ing erythro acetate having very nearly the same optical purity as the erythro, tosylate. The stereochemical facts are consistent with the intervention of phenonium-ion intermediates XLI and XLII which may suffer attack either at, , C* or Cp. Note that phenonium ion XLI has a plane of symmetry (if the plane, of the six-membered ring is assumed to be perpendicular to that of the threemembered ring) and therefore may yield only racemic products, whereas ion, XLII is asymmetric. Similar evidence points to phenonium-ion intermediates, “, , acetolysis of the tosylates and brosylates of PhCHMeCHEtOH and, , hCHEtCHMeOH/8 Here, however, the substituents at Ca in the intermediate, wald"£.^mt5«8a"’ an<1 Ingraham’ J' Am■ ^' S0C821 (,948>” Cram» J- Am. Chem. Soc., 71, 3863 (1949)., Cram, J. Am. Chem. Soc., 71, 3875, 3883 (1949); 74, 2159 (1952).
Page 592 :
576, , -, , Participation of Neighboring Groups, , Me, , OTos, , L threo, XXXIX, , Ph, , Me, , Me, , OTos, , L erythro, XL, , are not the same as those at Cp, and attack at Qs results in the formation of a, rearranged acetate (for example, XLIV)., Ph, , H, , XLIII, , threo, tosylate, , AcO, , Et, XLIV, , rearrangement, product, , It may then be asked whether the third criterion for neighboring-group, participation, the enhancement of reaction rate, is also applicable to cases 1 voicing benzonium-ion intervention. It is difficult to answer this for stmple, monophenylated substrates such as XXXIX, XL, and XLIII, since anchnnenc, assistance by a single phenyl is slight and is opposed by a rate-retarding mductive, effect associated with the same phenyl.® However, it has been found that, >» Using a modification of the Taft treatment (p. 227), Strei.wieser estimates (Ref- tWj, p. 718)tTal introduction of a 3-pheny. group in 2JZ'„7 stce i o^en tosla.es, acetolysis by a factor of about 8 (in the absence of, Jc, assume a rate, XXXIX and XL are solvolyzed about half as rapidly as 2-butyl tosylate, we may, enhancement of a factor of 4 due to participation.
Page 595 :
Aryl Participation. The Phenonium Ion, corresponding, , 579, , ratio from the erythro tosylate indicates predominant retention., , This suggests that some of the acts of substitution in these acetolyses procee, through a classical carbonium ion, whereas others proceed only through a, phenonium-ion intermediate., In special cases, participation by aryl groups lying three atoms removed, from the reaction center (5-aryl participation) may occur. It has recently been, found that substitution of methoxy groups into 4-phenyl-l-butyl brosylate,, Ph(CH2)40Bs, increases its ease of solvolysis. Incorporation of methoxy groups, at the 2 and 4 positions of the benzene ring boosts the rate of formolysis by a fac¬, tor of 10, whereas incorporation of methoxy groups at the 3 and 5 positions, accelerates formolysis about sixfold. Somewhat smaller increases are observed for, acetolysis rates.56(o) Since these methoxy groups are too far removed from the, reaction center to exert appreciable inductive or steric effects, it may be inferred, that they assist the solvolyses by facilitating aryl participation. The solvolysis of, the 2,4-dimethoxy compound very probably proceeds through the tetramethylenephenonium ion, LVI, analogous to the ethylenephenonium ions that inter¬, vene in the course of (3-aryl participation. It is considerably less likely, however,, that the solvolysis of the 3,5-dimethoxy compound involves the corresponding, tetramethylenephenonium ion, LVII; for in this ion the methoxy groups lie, , MeO, , f I « v, MeO, , Vx, , CH2-CH, OBs, , MeO, , LVI 11, , EZSS-Z, , phel(”LCr, , ibid., 77, 92, 99 (1955)., , Ph,CHCHPhOA(‘) ^'1“' *, aceioiysis ot Ph2OHCHPhOAc, see Collins and Bonner,, , (a) Heck and Winstein, J. Am. Chem. Soc., 79 3105 3114, /n d • j, j up, stein, ibid., 79, 756, 4238 (1957)., *, ’ 5114 (1957)- W Baird a"d Win-
Page 596 :
580, , -, , Participation of Neighboring Groups, , meta to the reaction center, and due to the “alternate” character of atoms com¬, , prising a conjugated system (p. 214), the methoxy groups in LVII cannot aid in, the dispersal of positive charge as can the methoxy groups in LVI. It is instead, more probable that the intermediate in the solvolysis of the 3,5-dimethoxy, compound is cation LVIII, in which the methoxy groups lie ortho and para to the, reaction center. Indeed, the major product from the latter solvolysis is 5,7-dimethoxytetralin (LIX), which may be formed merely by the removal of a, proton from cation LVIII. As indicated, tetralin LIX is also obtained from the, solvolysis of the 2,4-dimethoxybrosylate,56(6) very probably via rearrangement of, intermediate LVI (this rearrangement being similar in nature to the dienonephenol rearrangement, p. 639). Apparently, aryl participation is unusually, sensitive to the size of the ring formed, for no significant increase in solvolysis, rate results when methoxide groups are incorporated into the benzene rings, of Ph(CH2)3OBs or Ph(CH2)6OBs.57, , Intimate and Solvent-separated Ion Pairs, Let us return momentarily to the reaction we called upon to introduce the, phenonium-ion intermediate (p. 575), the acetolysis of the diastereomers of, PhCHMeCHMeOTos. In particular, we may reconsider the acetolysis of an, optically active threo form of this tosylate to the racemic threo acetate LXII, bear¬, ing in mind that in the poorly ionizing solvent, acetic acid, the proposed phenonPh, , Ph, , I, , /Hrai-MeCH— CHMe, I, OTos, (active), , ->, , Ph, , 'v, , MeCH-'CHMe, , ->, , I, , threo-MeCH—CHMe (racemic), , OTos", , I A, OAc, , LX I, , LXII, , 37 Since five-membered rings are, in many cases, formed more readily than six-membered, rings of the same type, it may be asked why aryl participation leading to an intermediate, analogous to LVIII does not assist the solvolysis of 3,5-(MeO)2C6H3(CH2)30Bs. The answer, is probably that the transition state, LX, leading to such an intermediate is rather strained., .Assuming that the five-membered ring in LVII is nearly planar, that the bond angles at U, MeO, , -OBs", , H, , H, LX, , membered ring in the transition state leading to, certainly, less strained., , cation LVIII would be puckered and, almost
Page 597 :
581, , Intimate and Solvent-separated Ion Pairs, , iiim-ion intermediate should exist mainly as an ion pair (designated LXI). If, the first step in this sequence (ionization to a symmeti ic phenomum ion, were slow and irreversible, then the loss of optical activity should parallel the, destruction of the tosylate. In actuality, however, if the reaction is interrupted, when about half the tosylate is solvolyzed, the remaining tosylate is found to be, almost completely racemic, suggesting that the initial ionization is reversible. Similar, , experiments with the corresponding threo brosylate point to the same conelusion.58 Moreover, the reversal of the ionization is mainly an “internal return, (p. 288), rather than an “external return”; that is, a phenonium ion is much, more likely to recombine with its original anionic partner than with a different, anion. This may be demonstrated by carrying out the acetolysis of the alkyl, brosylate in the presence of a large excess of OTos- ion, under which conditions, only a small fraction of the substrate is converted to the (more stable) alkyl, tosylate., , D,L-ROBs, slow, , £>,L-ROAc, , L-ROBs -> R+OBs-, , D,L-ROTos (small amount), Since internal return is a phenomenon connected with ion pairs (or higher ionic, aggregates), we would expect, and we find, that internal return becomes signifi¬, cantly less important in formolysis,59 for the ratio of 6olvated ions to ion pairs is, much greater in formic acid (dielectric constant 38) than in acetic acid (dielec¬, tric constant 6)., Some information concerning the structure of ion pairs such as LXI may be, obtained by studying the behavior of sulfonate esters in which the alkyl-bound, oxygen has been labeled. Such experiments have been carried out using labeled, 2-phenyl-l-propyl brosylate (LXI 11), which undergoes ionization and internal, return in acetic acid to yield brosylate LXI V.40 The three structures that come, most readily to mind for the ion-pair intermediate in this reaction are LXV,, LXVI, and LXVII. If the ion-pair intermediate is of type LXV, all of the, labeled oxygen in LXI 11 will become alkyl-bound oxygen in the rearrangement, product, LXIV; if the intermediate is of type LXVI, none of the labeled oxygen, , Th, PhCHMeCH 2— 0*-S, , I, , MeCH-, , O, , •CH;, , MeCH—CH2Ph, , OBs", OBs, , LXIII, LXIV, 58 Cram, J. Am. Chem. Soc., 74, 2129 (1952)., ;:^inste- and Schreiber, J. Am. Chem. Soc., 74, 2165 (1952)., Denney and Goldstein, J. Am. Chem. Soc., 79, 4948 (1957)., , (0*=018)
Page 598 :
582, , -, , Participation of Neighboring Groups, +, , MeCH--CH2, MeCH-^CH2, \, , ^, , z, , SO", , "O'*, , i, , I, , Ar, , SOz Ar, LXV, , LXVI, , LXVII, , will become alkyl bound. Finally, it is possible that due to rotation within the ion, , pair, the three oxygens become chemically equivalent before the completion ol, , the rearrangement (LXVII), in which case one third of the labeled oxygen will, , become alkyl bound. Since it is found that in acetic acid nearly two thirds of the, labeled oxygen atoms become bound to the alkyl group in the rearranged, brosylate, LXIV, we may infer that rearrangement in this solvent proceeds in, , large part through ion pair LXV, but partially also through ion pairs LXVI or, LXVII (or through both). On the other hand, the results of a similar experi¬, ment with the anisyl derivative, jfr-MeOC6H4CHMe—CH2—0*Tos, indicate, that here the rearrangement passes through an intermediate in which all of the, sulfur-bound oxygen atoms have become equivalent. Thus, between the instant, of ionization and the time of return, the anisonium-bridged ion may become, more nearly independent of its original anionic partner than may the less stable, phenonium-bridged ion., The same conclusion emerges from a different type of study—that of salt, effects on acetolysis. Since the activation process for these solvolyses requires, separation of unlike charges, the reactions should be accelerated by increasing, the ionic strength of the medium. Typically, when an optically active threo form, of PhCHMeCHMeOTos is dissolved in acetic acid, both its rate of ionization, (as measured by loss of optical activity) and its rate of destruction (as measured, by the appearance of TosOH) are increased by addition of LiClO-j. If the salt, concentration is kept below 0.1 molal, both rate constants are, to a good ap¬, proximation, linear functions of the concentration of added salt.^(o) Suppose,, , however, that the active threo form of the anisyl compound, />-MeOC6H4CHMeCHMeOBs, is treated in the same way. Again, the rate of ionization (loss, of optical activity) is essentially a linear function of the concentration of added, LiCICb, but the acetolysis exhibits what is said to be a “special salt effect., Specifically, very small concentrations of added salt bring about striking in41 (a) Fainberg and Winstein, J. Am. Chem. Soc., 78, 2780 (1956). It has already been, noted (p. 185) that for reactions in which the activation process involves separation of unlike, , chareesP the observed rate constant is an exponential, rather than a linear function o, strength. However, if only a narrow range of salt concentrations is being considered, the e- xp, nential curve relating the two functions does not deviate greatly from a straight line. ( ), stein and Robinson, J. Am. Chem. Soc.,, , 80,, , 169 (195 ).
Page 599 :
Intimate and Solvent-separated Ion Pairs, , 583, , creases in rate, but as more and more salt is added, further increases in rate are, merely those typical of “normal” salt effects. The following rate constants illus¬, trate the contrast :41(6), 106£, sec 1 (25°), , Added LiC104, , Ionization, , Solvolysis, , none, 0.03 M, , 81, 111, , 20, 84, , 0.10 M, , 200, , 160, , It is apparent that the rate of solvolysis of the anisyl compound (formation of, HOBs) is increased more than fourfold when the concentration of added, LiC104 is increased from 0 to 0.03 M under which circumstances the rate of, ionization increases by less than 40 percent. On the other hand, the specific, rates for both ionization and solvolysis are increased somewhat less than twofold, when the concentration of added salt is boosted from 0.03 to 0.10 M. Moreover,, “special” salt effects are not observed when the added salt is LiOTos or LiOBs., Since the marked acceleration of the acetolysis of the brosylate by small, quantities of LiC104 is apparently not due to a great increase in the rate of, ionization of the substrate, the salt must exert its effects mainly after the ionization, has occurred. At first glance, it might be supposed that the perchlorate reacts, with the ion pair, R+OBs~, converting it to R+C10^, which must undergo rapid, solvolysis, since it cannot “internally return” to an unstable alkyl perchlorate,, , ROCIO3. (This would account for the fact that neither LiOTos nor LiOBs, exerts a “special” salt effect; anion interchange between R+OBs~ and either of, these salts would give an ion pair which should readily undergo internal return.), If this explanation were correct, the addition of large quantities of LiC104 should, suppress internal return, causing the rate of acetolysis to approach the rate of, ionization of the substrate. However, there is no indication that these rates ap¬, proach each other, and it therefore seems probable that an additional step must, be included in the solvolysis scheme to account for the “special” salt effect., Winstein feels that two distinct types of ion pairs intervene in acetolyses that, involve anisonium-bridged ions. Initial ionization of the substrate yields an, “mtmate" ion pair, in which the cation and anion lie in a common solvation shellinternal return is simply the collapse of this pair, with or without rearrangement’, As the solvolysis proceeds, the “intimate” ion pair is converted to a solventseparated ion pair, , that is, to an ion-pair in which the cation and anion are in, , ifferent so vation shells. Such an ion pair may revert back to an intimate ion pair, or, alternatively, it may extract a solvent molecule from the solvation shell of
Page 601 :
Alkyl and Cycloalkyl Participation, , -, , 585, , These rearrangements (and many similar conversions) proceed most readily, under conditions that, for an ordinary alkyl derivative, would favor substitution, via a carbonium-ion intermediate. The reactions of neopentyl halides, for exam¬, ple, are catalyzed by such electrophilic species as Ag+ and HgCl2 and take place, at rates similar to those at which the corresponding ethyl halides undergo SN1, reactions under ionizing conditions.^ It is very likely then that these rearrange¬, ments proceed through the neopentyl ion, Me3CCH+; but since nonrearranged, products are apparently not formed, it is equally likely that this carbonium ion, rearranges to the /-amyl ion, Me2CCH2Me, very soon after the initial heterolysis., — y-, , feist, , t-, , Y•, , MeoC—CH2Ar —* Me2C—CH+ —> Me2C—CH2Me —> Me2C—CH2Me, slow, , Me, , |, , *ast, , Me, , Y, , For an instant (or perhaps for an appreciable period of time) during the reaction,, the migrating group lies equidistant from the /3-carbon (the migration source), and the a-carbon (the migration terminus). At this point, the system has as¬, sumed the configuration LXIX which brings to mind the phenonium-ion interMe, , ✓, , ✓, , s, , N, , Me 2 C ——-—-vCH2, LXIX, mediate through which many aryl shifts proceed. Yet there is an important, difference; since the carbon of the migrating methyl group is bound to three, hydrogen atoms, it cannot be fully bound (in the usual sense of the word) to, both the a- and /3-carbons simultaneously. In this respect, a bridged carbonium, ion such as LXIX is even less “classical” than a phenonium-bridged ion. This, does not mean that alkyl-bridged carbonium ions are automatically excluded, as intermediates in alkyl shifts; indeed, stable compounds having alkyl bridges, of this sort are known “ We would expect, however, that much less anchimeric, ** Perhaps the most familiar compound having alkyl bridges is the dimer of trimethylaluminum, LXX (Snow and Rundle, Acta CrysL, 4, 348 (1951)). For a brief discussion of the, , IT, , H, , I, , H,C., , -CEL, Al, , H,C, , 'Ah, , CH„, , :c:, , H, , |, H, , H, , LXX, 0nC,UdinS th°Se With M-gen bridges), see
Page 602 :
586, , -, , Participation of Neighboring Groups, , assistance be rendered by neighboring alkyl groups (which are essentially nonnucleophilic) than by neighboring aryl groups under similar circumstances., Typically, neopentyl chloride is hydrolyzed (with rearrangement) somewhat, more slowly than ethyl chloride (without rearrangement), whereas Ph3CCH2Gl, is hydrolyzed (with rearrangement) over 10,000 times as rapidly.45 We may as¬, sume the hydrolysis of the neopentyl halide to be virtually unassisted anchimerically, but hydrolysis of the triphenylethyl halide is quite obviously assisted.45, In very few cases is there substantial reason for supposing that an alkyl shift, in a noncyclic system proceeds through an alkyl-bridged intermediate (as dis¬, tinguished from an alkyl-bridged activated complex). The strongest evidence to, date for a methyl-bridged intermediate arises from the deamination (with, HN02 in acetic acid) of /Ara>-3-phenyl-2-butylamine, PhCHMe—CHMeNH2., , Me, -Nj, , 44% PhCHMeCHMeOAc (57% threo, 49% erythro), (from simple substitution and phenyl shift), 24% PhCMeCH2Me (from hydride shift), , PhCHMe—CH—N? ——>, , 1, , HOAc, , threo, , OAc, 32% PhCHCHMe2, about, optically active, (from methyl shift), OAc, LXXIII, , Here we are concerned with the methyl shift that, as indicated, competes favor¬, ably with the phenyl shift, and which is stereospecific in the direction of inversion, of configuration about Q3. If the methyl shift were to proceed wholly through the, “open” carbonium ion LXXVI, the resulting acetate LXXIII should be race¬, mic, whereas if none of the acts of methyl rearrangement were to proceed, through ion LXXVI, complete inversion of configuration at Q> would be ex¬, pected. The observed partial inversion of configuration then suggests the inter¬, vention of the methyl-bridged carbonium ion LXXV, which may be attacked at, Ca to give unrearranged product, or attacked at, , to give rearranged product, , V Charlton, Dostrovsky, and Hughes, Nature, 167, 986 (1951)., *e The solvolyses, with rearrangement, of such very crowded substrates as: LXX, , an, LXXII are also unusually rapid (Bartlett, el al., J. Am. Chem. Soc., 77, 2801, 2804, 2806 (, ))•, The high rates in such cases are frequently attributed to steric acceleration (release of steric stra, , O, , CHMe,, , I, MesC —C—Cl, , (LXXI), , (MegC—)sc|, , O, , C, , $, , \, , N02, , (LXXII), , CHMe., when «hc reactant is converted to the transition state) rather than .<-nchim'rit: t^tanctn, There is, at present, no way of distinguishing between these two effects for react,on, which alkyl groups migrate.
Page 604 :
588, , -, , Participation of Neighboring Groups, , result, respectively, in phenyl, hydrogen, or methyl migration. It should also be, , LXXVIII, , LXXIX, , borne in mind that the bridged ions resulting from acetolyses of tosylates and, brosylates are present in ion pairs, whereas the carbonium ions formed in, deamination reactions are more nearly free, and therefore considerably more, reactive., When an ethyl or n-propyl group is substituted for one of the (9-hydrogens, of an alkyl halide or alkyl arenesulfonate, solvolysis rates suffer only minor, changes, for anchimeric assistance by /3-alkyl groups is negligible. If, however,, the ethyl or propyl group is “tied back” onto the /3-carbon, forming a cyclopropyl, or cyclobutyl ring (which does not include the a-carbon), the specific rates for, solvolysis (under ionizing conditions) rise markedly and may exceed those for, noncyclic substrates by factors of several hundred or more.49 Here it is probable, that solvolysis is being facilitated by a combination of field and anchimeric, effects. It is thought (although agreement is not complete on this point) that the, field effect associated with the cyclopropyl group arises because the regions of, maximum electron density lie well outside the ring instead of on the straight lines, connecting the carbon nuclei. For maximum bonding interaction (that is,, maximum overlap), the bonding orbitals of each carbon atom should be directed, toward the attached atoms, necessitating that one of the angles between orbitals, in cyclopropane (for each carbon), , be 60°. However, quantum mechanics, , stipulates that the orbitals about any first-row element may not form an angle, of less than 90°, and that, in particular, when a carbon atom forms four bonds,, angles approaching 109° are favored. Now, substantial overlap may still occur, when the angle between orbitals is 109°, but it has been calculated that the best, compromise (energetically speaking) involves an interorbital angle of 104 *, w (a) Typically, benzenesulfonate LXXXI suffers ethanolysis over 500 times a, rapWly, , as does ethyl benzenesulfonate (Bergstrom and Siegel,, , J. Am. Chem. Soc., 74,, , CHMeOBs, , CH20S02Ph, LXXXI, , LXXXII, , (b) Brosylate LXXXII undergoes acetolysis over 500 times as rap, , brosylate (Winstein and Marshall, ibid., 74, H20 (1952))', « Coulsen and Moffitt, Phil. Mag., 40, 1 (1949)., , y
Page 605 :
Alkyl and Cycloalkyl Participation, , -, , 589, , (although there is, at present, no direct way to verify this value experimentally)., This, in effect, means that the bonds between carbons in cyclopropane are, “bent,” somewhat like bananas, as shown in Figure 14-1-, and that three regions of, high electron density lie outside the triangle of carbon nuclei. (The bonds be¬, tween carbons in cyclobutane are thought to be bent also, but to a lesser degree.), When a cyclopropylcarbinyl derivative, in which —X is the leaving group,, adopts the conformation shown, one of the electron-rich regions outside the ring, approaches Ca from the side opposite —X, facilitating the departure of the latter., , cyclobutyl, derivative, , allylcarbinyl, derivative, , Fig. 14-1. Formation and Rearrangement of a Cyclopropylcarbinyl Cation, , As evidence for neighboring-group participation by the cyclopropyl ring, we, may note the extensive rearrangement that occurs when cyclopropylcarbinyl, derivatives are solvolyzed in weakly nucleophilic solvents. A cyclopropyl¬, carbinyl cation in the conformation shown may rearrange to a cyclobutyl or, allylcarbinyl cation merely by redistribution of the electrons comprising the, “banana bond” nearest the positively charged carbon. Thus, it has been found, that hydrolysis of cyclopropylcarbinyl benzenesulfonate (LXXXI) or of the, corresponding chloride, or deamination of aminomethylcyclopropane LXXXIII, (with HNO,) yields mixtures of cyclopropylcarbinol, cyclobutanol, and allylcarbinol, (CH2= CHCH.CH .OH i.5' Moreover, if the hydrolysis of the benzene¬, sulfonate or chloride is halted in its early stages, rearranged benzenesulfonate, or chloride is found in the mixture, suggesting ionization, rearrangement, and, internal return. Rearrangement during solvolysis of the chloride is much less, extensive in the more poorly ionizing solvent, ethanol/9, hv ,hIf CO"VerSion °f a cyclopropylcarbinyl to a cyclobutyl derivative proceeds, bLrinDh ;UggeS,ed ^ FigUre ^ thC Carb0" at°m ^S-ted C.) originally, , pou d, •, , l 7 gr°UP r-Vsh°Uld be“me a, - ‘be cyclobutyl con,has, however, been shown that if the aminomethylcyclopropane is, , " R°bertS and Mazur> J■ A™- Chem. Soc., 73, 2509, 3542 (1952).
Page 606 :
590, , -, , Participation of Neighboring Groups, , labeled at Ca with C14 (LXXXIII), very nearly one third of the labeled carbon, in the resulting cyclobutanol lies in the 7 position, virtually none in the a posi¬, tion, and two thirds in the 0 position.*' This points to the intervention of an, intermediate in which the two carbons which ultimately become 0-carbons, and, the single carbon which ultimately becomes a 7-carbon, in cyclobutanol are, geometrically equivalent. No classical structure fulfilling this requirement comes, readily to mind, but Roberts has proposed carbonium ion LXXXIV, which is, very “nonclassical” indeed. Such an intermediate could conceivably be formed, if the —CH^ group of the cyclopropylcarbinyl cation moved back of the cyclo¬, propane ring and became attracted to the electron-rich portion of the “banana, bond” directly opposite it. Under such an electrostatic stress, the two electrons, comprising this relatively weak bond could be redistributed in the regions be¬, tween the three methylene groups, resulting in partial bonding (as represented, in LXXXIV by the broken lines).52 Conversion of this very strained inter¬, mediate to the cyclobutyl cation requires breakage of one of the “full” bonds, (a, b, or c) and one of the “one-third” bonds (a\ bf, or c'). Disregarding isotope, effects, the three modes of conversion are equally probable. In examining the, , HNO,, CH2NH2, , CH,N^, , LXXXIII, , LXXXIV, , second major product from this deamination, cyclopropylcarbinol, it is found, that much of the labeled carbon has entered the cyclopropane ring. If all of this, product had been formed through intermediate LXXXIV, the ratio of labeled, carbon in the side chain to that in the ring should be 34 since there are two, methylene groups in the ring. Since this ratio is found instead to be 0.83, we may, assume that although some of the cyclopropylcarbinol forms through LXXXI ,, some also forms from the classical cyclopropylcarbinyl cation without interven¬, tion of LXXXIV., » Proposed intermediate LXXXIV is somewhat analogous to alk^"^., ions that have been, , dkyl-bridged cation., , two electrons, , form two partial bonds, whereas In LXXXIVtwo,elc, the original “banana bond”) form three partial bonds. The bonds represented, lines in LXXXIV thus have a bond order of one third., , y
Page 607 :
Neighboring Hydrogen, , 591, , Anchimeric assistance to ionization by saturated rings of five or more, members has not been shown to be significant for simple systems. However, after, ionization has occurred, a ring carbon may migrate from a p to an « position, resulting in an expansion of the ring (if the positive charge lies outside the ring), or a contraction (if the positive charge lies on the ring). Two typical 1,2 shifts, involving monocyclic systems are shown:, , However, the most interesting examples of 1,2 rearrangements involving ring, carbons occur in the chemistry of polycyclic systems. A number of these are, discussed in a subsequent section., , Neighboring Hydrogen, Although hydride shifts in reactions involving carbonium-ion intermediates are, well known, we should expect a /3-hydrogen, like a /5-alkyl group, to be quite, ineffective in rendering anchimeric assistance toward ionization at the a posi¬, tion. As with neighboring-group participation by other species, hydride migra¬, tion generally becomes more important as branching at the /8-carbon is increased., Thus, reactions of isobutyl (Me2CHCH2—) compounds under ionizing condi¬, tions often yield predominantly products derived from the f-butyl cation,, Me2CCH3, but practically no hydrogen migration occurs in reactions proceeding, through the ethyl cation, CH3CH^ (as may be demonstrated by examining the, reactions of/3-deuteroethyl compounds5^). Nevertheless, the reactions of isobutyl, compounds are not significantly faster than those of ethyl compounds, showing, that in the isobutyl-to-/-butyl transformation the migrating hydrogen does not, exert appreciable pushing action on the leaving group. This is probably because, c*a’^6).'' 405’ 129 (19,4); 4t7’ 255 (19,8)- (i) Ru2icka a"d B^r, He,,, , fTdf?: ™mpl?' <hat the decomP°si'ion of the deuterated diazonium ion., , CH.DCH N^r„en, That w„,dH h.ryef, 5943^9“^, P, , a" contaminS'«<>>*" 2 per cent CH.CHDOH, the produc, “ deu,enum skift (R<*em a"d Yancey, J. Am. Chem. L. 74,
Page 608 :
592, , -, , Participation of Neighboring Groups, , the bond to the leaving group is completely or nearly broken beforj the hydrogen, migration gets under way. The hydrogen-bridged carbonium ion LXX\\, H, , I, , MetC-CHtNj, , H, , H, , I, , MeaC —C.H j, , Me,C, , 'CH*, , Me.C-CH,, , LXXX\', t-BuOH, , (3):, , (which may be either an activated complex or an unstable intermediate) i:, , structurally similar to a tt complex formed by protonation of an olefin. Such si, complex was proposed as an intermediate in the acid-catalyzed addition of water, to olefins to account for the observed dependence of rate on /j0 (p. 516). Yet ii, may be shown if olefin-hydration proceeds through a w complex, tire hydrogen:, shift (sequence 3) does not. It will be recalled that the equilibrium between an, olefin and its protonated ?r complex is rapidly established, whereas the conver-sion of the complex to a classical carbonium ion was presumed to be slow. There--, , fore, if LXXXY were a rr complex, it should exchange protons rapidly with the!, solvent water. Now, when the decomposition of the diazonium ion is run in D20,., , the resulting alcohol is essentially free of carbon-bound deuterium, proving that,, , the hvdrogen that migrates does not become equilibrated with protons of the', , solvent.55 We may thus infer that the “hydrogen bridge” and the “protonated:, , x complex" are different species although the positions of the atoms are veryi, , nearlv the same in both. It seems likely that the proton in a tt complex lies near,, the “edges" of the 7r-electron clouds and distorts them only slightly (LXXXVI),, whereas the proton in a hydrogen bridge is deeply imbedded in a single cloud of, electric charge that occupies space above both carbon atoms (LXXXVII)., , LXXXVI, 7r complex, , LXXXVI I, hydrogen bridge, , To date the clearest evidence that participation by /3-hydrogen may ap¬, preciably (although not spectacularly) affect reaction rates has been obtaine^, from studies of cyclohexyl derivatives. Compare, for example, the as and trans, isomers of 2-methylcyclohexyl tosylate. For both, the bulky methyl group preier., Si, , Cannell and Taft, J. Am. Chem. Soc., 78, 5912 (1956).
Page 609 :
Neighboring Hydrogen, , -, , 593, , to lie in an equatorial position, rather than a more crowded axial position (p., 241). As a result, the tosylate group assumes an axial position in the avore, conformation of the cis isomer (LXXXVIII) and an equatorial position in the, favored conformation of the trans isomer (LXXXIX). We should then expect the, cis tosylate to undergo unimolecular solvolyses about two or three times as, , OTs, , OTs, , Me, , rapidly as the trans tosylate (Chap. 7, Ref. 54), since more crowding is relieved, when an axial —OTs group departs. However, the acetolysis of the cis tosylate, has been found to be over 70 times as fast as that of the trans tosylate.5'7 It is, likely then that the ionization of tosylate LXXXVIII is anchimerically assisted, by the hydrogen at the 2 position, for this lies trans to the —OTs group and both, substituents occupy axial positions in the favored conformation of the tosylate., In tosylate LXXXIX, the methyl lies trans to the —OTs group but does not, significantly assist its departure; in the favored conformation of this tosylate, the, bonds to the —Me and —OTs substituents are not coplanar with the bond link¬, ing Ci and C2, and the system may not, without considerable strain, assume the, transition state necessary for anchimeric assistance (p. 573). Assistance by hydro¬, gens at the 6 position may be ignored in both cases since these are bound to a, secondary, rather than to a tertiary, carbon atom. In a similar way we may, understand why neomenthyl tosylate (XC) is acetolyzed almost 80 times as, rapidly as is menthyl tosylate (XCI),Si and why the ratio of solvolysis rates for, the corresponding chlorides in 80 percent ethanol is 41.57, The stereochemistry of the formolysis of neomenthyl tosylate, XC, points, clearly to hydrogen participation. If only the “classical” carbonium ion, XCV,, were involved, there should be formed, in addition to neomenthyl formate, two, olefins, 2- and 3-menthene, and both should be optically active since the asym¬, metric center at Ci would not be affected. It is found, however, that the only, olefin formed is 3-menthene (XCIV), and that this olefin is racemic,5' although, optically active 3-menthene is not itself racemized under the solvolysis condi¬, tions. This suggests the intervention of the hydrogen-bridged carbonium ion,, XCII, which is converted to the symmetric carbonium ion XCIII., fortlTattLXXXVnr1^; 1Z'9° ('955)' Presumably. <h= favored conformation, HI18 dlctated by the stenc requirements of the methyl group rather than, mm, Tr, * Br°UP' A"h°Ugh ,he lat,er is the 'arger, most of its bulk is hdd far a„av, from the cyclohexane ring and does not interfere with th!? hydrogen atoms bound to the £7, Hughes, Ingold, and Rose, J. Chem. Soc., 1953, 3839., g’
Page 610 :
594, , -, , Participation of Neighboring Groups, , OTs, , > should give 2- and 3- menthenes, (both active), XCV, , Bicyclic Systems, In the reactions of bicyclic systems, 1,2 shifts are frequently encountered. Al¬, though these are similar in principle to the alkyl shifts that occur in noncyclic, compounds, they often appear more complex; for migration of a ring carbon,, together with three of the bonds attached to it, may result in the formation of a, new ring system. The following are typical examples:, , A number of rearrangements in the bicyclo(2.2.1)heptyl series (parent, hydrocarbon XCVI) have been studied in some detail. The conversion, for, , bicyclo (2.2.1) heptane, (norbornane), XCVI, , camphene, hydrochloride, XCVII, , « (fl) Doering and Farber, J. Am. Chem. Soc.,, Emster, Ber., 53, 1815 (1920); 55, 2500 (1922)., , isobornyl, chloride, XCVIII, , 71, 1514 (1949). (b) Meerwein and van
Page 611 :
Bicyclic Systems, example,, , -, , 595, , of camphene hydrochloride (XCVII) to isobornyl chloride (XCVIII), , is a classic example of the Wagner-Meerwein rearrangement. The bicyclo(2.2.1)heptyl system appears both in the reactant and the product but, as is, shown below by numbering the carbon atoms, there is a “reshuffling” of four of, the seven carbons in the rings (those designated as 1, 2, 3, and 7). This trans¬, formation proceeds most readily in polar solvents such as sulfur dioxide and, nitromethane, which promote the ionization of tertiary halides without destroy¬, ing them. Moreover, it is catalyzed by such Lewis acids as HgCl2 and SnCL, and, also by HC1, but not by ionic chlorides.55(6)-50 Since the function of these, catalysts is evidently to remove chloride from the substrate rather than to supply, chloride, we may suppose that the rearrangement involves ionization of XCVII., However, the ionization cannot be rate determining, for it has been found that, the rate of exchange between HC136 and camphene hydrochloride is much, greater than that of the rearrangement.60 Instead, it appears that the ionization, of XCVII is rapid and reversible but the subsequent “collapse” of the ion, , pair (XCIX) to give isobornyl chloride, XCVIII, is slow. (Similar paths may be, written for this rearrangement, using HgCl2, SnCl4, or FeCl3 as the electrophile, that facilitates removal of the chloride.) As indicated, the isomerization is, reversible, the reverse reaction passing through the same series of transition states, and intermediates as does the forward reaction (as is required by the principle of, microscopic reversibility). If, as we have suggested, the conversion of ion pair, XCIX to camphene hydrochloride is rapid, then the slow step in the reverse, reaction is the ionization of isobornyl chloride., According to the sequence above, camphene hydrochloride is converted to, the bridged carbonium ion (in ion pair XCIX) without intervention of a classical, carbonium ion. Since the solvolyses of this chloride (under ionizing conditions), proceed from 300 to 8000 times as rapidly as those of ordinary tertiary chlorides,151, it seems very likely that the carbon atom at the 6 position is rendering anchimeric, assistance to ionization at the 2 position-that is, that the bond from C6 to C2 is, orming at the same time the C2-C1 bond is breaking. Likewise, no classical, carbonium ion is inserted between isobornyl chloride and the bridged ion pair, 820 (<;^)Mecrwein» ^ ^3, .16 (1927). (A) Bartlett and Pockel, J. Am. Chem. Soc, 59,, « S'S4, , Sa[as> and Wilson, J. Chem. Soc., 1939, 1188., , 168, 65 (1951). ^, , *, , em' S°C’’ 1935, 255 5 (6) Brown> HuShes> Ingold, and Smith, Nature,
Page 612 :
596, , -, , Participation of Neighboring Groups, , XCIX in the reverse reaction, since solvolyses of this chloride are thousands of3, times as fast as those of ordinary secondary chlorides,strongly suggesting, concerted ionization process here also. Furthermore, if carbonium ion C were to., iniei\ ene in this reaction, we would expect the formation of appreciable quanti-, , +, , C, ties of bornyl chloride (Cl), for at room temperature this halide is very nearly, as stable as the isobornyl isomer.5S(6), As we might predict, isobornyl chloride (XCVIII) is solvolyzed far more, rapidly than bornyl chloride (Cl), a 36,000-fold difference in rate having been, noted in 80 percent ethanol.62 In the isobornyl halide, the assisting ethylene bridge, (—C5—C6—) is held trans to the departing chlorine—a position favorable for, intramolecular attack—whereas in the bornyl halide, this bridge lies cis to the, departing chlorine and ionization must proceed without assistance. Similar, differences in reactivity have been observed among norbornyl compounds:, £*o-norbornyl chloride (ClI), in which the assisting ethylene bridge lies trans, to the departing chloride, is solvolyzed 70 times as rapidly in 80 percent ethanol, as is m/o-norbornyl chloride (CIII).^ The ratio of acetolysis rates for the cor7, , X, , exo, , endo, , CII, , cm, , CIV, , responding isomeric brosylates (at 25°) is 350.The differences in the nor¬, bornyl series are considerably less striking than those in the bornyl-isobornyl, series, suggesting that anchimeric assistance is more powerful in the latter., Comparing the bridged carbonium-ion intermediates for the two series (XCIX, and CIV), we see that one of the carbon atoms sharing the positive charge in the, bornyl cation (carbon 2) bears a methyl group. This, by reason of its hyperconjugative effect, would be expected to make the bridged bornyl cation (and, the activated complex leading to it) more stable than the bridged norbornyl, cation, CIV (and the activated complex leading to it)., When ^-norbornyl brosylate (CII, V = OBs) undergoes solvolysis via the, " Roberts, Urbanek, and Armstrong, J. Am Chem Soc., 71, 3049 (1949)«» Roberts, Bennett, and Armstrong, J. Am. Chem. Soc., 72, 33-9 (195 ),, ,, , .
Page 613 :
Bicyclic Systems, , -, , 597, , bridged cation CIV, the “rearrangement” product is also an ^o-norbornyl deriv¬, ative in which, however, the carbon atoms have become “shuffled.” Now cation, CIV has a plane of symmetry, which, as may be more clearly seen from CIV',, passes through C4, C6, and C6, midway between Cx and C2, and midway be¬, tween C3 and C7. Therefore, the product of a reaction which passes through CIV, should be racemic, whether or not the starting material is optically active. The, rate of racemization may be taken as the rate of ionization of the brosy late, and, this has been found to be considerably greater than the rate of solvolysis (produc¬, tion of H+OBs-).*4 Thus, the ionization of the brosylate, while necessary for, solvolysis, is not a sufficient condition for it, and we may once more attribute, the lag between ionization and solvolysis to internal return (p. 581)—that is, to, collapse of the ion pair R+OBs~ to form ROBs once again., ^n^ROBs (racemic), ionization, , ., , ROBs-> R+ (racemic), OBs(R+ = ion CIV), , ', , ,, ^ROS (racemic) + H+OBs~, , As in other such cases, internal return is less important for solvolyses in such, nucleophilic solvents as ethanol and aqueous acetone than in acetic acid; for the, more nucleophilic the solvent, the more likely it is to attack the ion pair before, the latter has a chance to collapse to starting material.65, An additional complication associated with the acetolysis of brosylate CII, (and very probably with the solvolyses of other bicyclic substrates as well) is the, occurrence of an intramolecular 1,3-hydride shift (from C6 to C2), which may be, detected by acetolyzing a sample of this brosylate which has been labeled with, C14 in the 2 and 3 positions (CV).66 If the sole ion-pair intermediate in this, , CVII, , ej winStem 3nd Trifan> J- Am■ Chem. Soc., 74, 1147, 1154 (19521, M R 1"stem and Schreiber, J. Am. Chem. Soc., 74, 2165, 2171 (19‘, Roberts, Lee, and Saunders, J. Am. Chem. Soc., 76, 4501 (1954
Page 614 :
598, , -, , Participation of Neighboring Groups, , solvolysis were CVI, we would expect half of the labeled carbon originally ati, C2 and C3 in brosylate CV to turn up at Cx and C7 in the resulting acetate, fon, acetic acid would be just as likely to attack CVI at Ci as at C2. Since, however,, substantial amounts of labeled carbon are found also at C6 and C6 (CIX), [u, , appears that a portion of the solvolysis proceeds through a hydrogen-bridged ion:, pair such as CVIII., , The 1,4-endoxocyclohexyl system, present in halides CIX and CXI 11, is-, , structurally similar to the norbornyl system. Again it is found that the exo--, , halides (CIX) are very much more reactive than the m/o-halides (CXIII), the-, , hydrolysis rates differing by factors of from 200 to 300.67 Rearrangement,,, , analogous to those occurring in the norbornyl and isobornyl series, is very nearly, complete, but the product isolated is 3-formylcyclopentanol, rather than hemiacetal CXII, which is unstable under hydrolysis conditions. The fact that non-, , rearranged products are not formed suggests that the bridged carbonium ion,, CX is converted irreversibly and almost immediately after it is formed, to the, rearranged oxonium ion, CXI. The hydrolyses of the mfo-halides (CXIII), also yield 3-formylcyclopentanol, and it may well be that these too proceed, through the bridged ion CX. Since, however, the ethylene bridge in CXIII lies, ds to the halogen, it cannot assist the breakage of the carbon-halogen bond, and, , formation of the bridged ion may not occur until after the ionization. Here, then,, the classical carbonium ion, CXIV, is probably a precursor of the bridged ion., Very spectacular anchimeric assistance by the tt electrons associated with, a G=c double bond has been noted in the reactions of <™h-7-norborneny, derivatives, , (CXV), , « The anff-7-tosylate (CXV,, , X, , = OTs) is solvolyzed over, , a, , billion times faster than its saturated analog (CXVII, X = OTs), and the ratio, of rates for ethanolyses of the corresponding chlorides is about a million,, shown in CXVIII, the 7r-electron lobes in the tf/ih-7-norbornenyl derivatives i, near, , C7, , on the side opposite the leaving group. Not only may these e ectron, , 97 Martin and Bartlett, J. Am. Chem. Soc,.79, 2533, « (a) Winstein and Shatavsky, J. Am. Chem. Soc., 78,, Roberts, ibid., 78, 5653 (1956)., , Woods, Ca.boni, and, u w). \ )
Page 615 :
Transannular Rearrangements, , cxv, , X, , X, , X, , CXV1, , CXVII, , 599, , cxviii, , cxix, , help “push off” the leaving group, but after its departure, they may help accom¬, modate the positive charge by drifting toward C7 and allowing this charge to, spread over three atoms (CXIX) rather than being confined to just one. Assist¬, ance by the tt electrons requires that these lie opposite the leaving group; thus,, the syn forms of 7-norbornenyl halides and sulfonates (CXVI) are solvolyzed, more slowly than are the corresponding anti forms by factors of more than 105., Finally, it should be noted that the 7r-electron lobes associated with the, double bond between C2 and C3 in the norbornenyl system are “tipped” so that, they lie much nearer to C7 than to C6 or C6. We should therefore expect (and we, find) that such a double bond renders much less assistance to ionization of, substituents at C6 or C6 than to ionization at C7. The <?xo-chloride CXX under¬, goes solvolysis in 80 percent ethanol about 150 times as rapidly as the endochloride, CXXI,65 a difference in reactivity not much greater than that observed, for the corresponding saturated chlorides., , Cl, , CXX, CXXI, , Transannular Rearrangements, Group shifts in which the migration origin and the migration terminus are, separated by one, two, or more atoms are rare, except in certain cyclic systems, where the puckering of the ring allows the close approach of positions that are, not bonded to each other.69 The norbornyl system, in which a 1,3-hydride shift, has been established, is one of these, and the cyclooctyl system is another. It has, been found, for example, that formolysis of ^rans-cyclooctene oxide (CXXII), gives predominantly the transA,4-monoformate (CXXIV),™<“> suggesting that
Page 618 :
602, , -, , Participation of Neighboring Groups, , the 1,2 shifts that we have considered.7^ The conversion is generally carried, , oi, , in the presence of strong acid and may be assumed to proceed through at lea's, one carbonium-ion intermediate. Whether this is a “nonclassical” (bridged) or, classical, , (open) carbonium ion, and whether there are two, or perhaps three, , such intermediates (that is, CXXXVI, CXXXVII, and CXXXVIII, 0, any two of these) interceding between reactant and product are matters tha, should be investigated for various individual cases. Along with the pinacol re, R, , R, , I, , R, , I, , —C—C—, OH2, +, , HO, , R, , '+ ', , —C—C— (??) -► —c, • I, HO, , +, , (??), , I, HO, , L cxxxvi, , -H!, , —C—C— (??), HO, , CXXXVII, , CXXXVIII, , J, R, , I, —C—c-, , I, , I, , o, , arrangement, we may consider the reactions of /3-halo-hydrins with Ag+ o, Hg2+:, RI, —C, OH, , R, -C —, , Ag+, , > —c, , C— + H+ + Ag-T (or HgA+), , (or Hg2+), , X), X)f, , O, , and the deaminations of /3-aminohydrins:, R, I, , cI, OH, , R, I, , HNO,, , -> —C, II, , C—, I, nh2, , c—, , + H+ + No, , o, , The products resulting from these reactions are equivalent to those resultins, O, , OH, , V Such conversions as that of PhCH—CPhoOH to PhCCPh2H may proceed by a pat, analogous to that of the pinacol rearrangement, but with hydride, rather than an a, , y °, , aryl group, shifting. However, an alternate path, indicated below, apparently has not ye, been ruled out for such reactions:, OH, H+, , PhCHOH—CPh2OH-> PhCHOH—CPh^, , —H+, , I, , * PhC—CPh2, , w, PhC—CPb2H, , In view cfthis ambiguity, only those pinacol rearrangements, to be considered. For kinetic studiesof the pinacol, j Am. Chem. Soc., 76, 3925 (1954); Duncan and Lynn, J. Chem. Soc., 1956, 351^, 33, , 3674, .
Page 619 :
The Pinacol and Related Rearrangements, , -, , bOd, , rom .he pinacol rearrangement, and it is very likely that the mechanisms are, CryThe stereochemical requirements for these rearrangements are most readily, ipparent from the study of substrates in which the C„—C„ bond is incorporated, n,0 a ring, thus restricting freedom of rotation. From such systems we find that, t is desirable, if not necessary, that the migrating group and the leaving group, ie trans to each other during the early stages of bond breaking. Thus, the as, orm of 1,2-dimethylcyclohexane-l ,2-diol (CXXXIX) (in which the methyls lie, rans to the hydroxyl groups) undergoes the pinacol rearrangement easily in, , iilute H2SO4 with migration of a methyl group; whereas the trans glycol, CXL, in which the methyls lie cis to the hydroxyl groups) undergoes rearrangement, jnder similar conditions with contraction of the ring.'"(a) Similarly, only the as, , CXXXIX, , CXL, , form of the isomeric l,2-dimethylcyclopentane-l,2-diols undergoes rearrange¬, ment with methyl migration. Both the cis and trans forms (CXLI and CXLII) of, 7,8-diphenylacenaphthene-7,8-diol undergo the pinacol rearrangement with, phenyl migration; however, the trans isomer is converted to ketone CXLI 11 only, one sixth as rapidly as is the cis. Moreover, since the trans diol is readily con¬, verted to the cis under the reaction conditions used, it is likely that the latter, is an intermediate in the transformation of the trans diol to CXLIII.75(6), OH, , OH, , Ph, , O, , CXLI, , The evidence that the migrating group approaches the migration terminus, (C“) from the Slde opposite the leaving group strongly suggests that an inversion, of configuration occurs at this carbon. While this apparently has not yet been, demonstrated directly for the pinacol rearrangement as such, it has been shown, or the deamination of the optically active aminohydrin CXLIV, which yields, predominantly ketone CXLV:™, , Z, , (l\, H)^artlf“’ Pocke1’ and BavleV, J■ Am• Chem. Soc., 59, 820 (1937)- 60 2416 (19381, W Bartlett and Brown, ibid., 62, 2927 (1940)., V, ’ ou’, U758)., n Bernstein and Whitmore, J. Am. Chem. Soc., 61, 1324 (1939).
Page 620 :
604, , -, , Participation of Neighboring Groups, , CXLIV, , CXLV, , (For proof that this conversion proceeds as shown, see Ex. 3.), , However, although inversion predominates, about 25 percent of the resulting, ketone is racemic, showing that the rearrangement proceeds, at least in part, , through the open carbonium ion, MeCH—CPh2OH, which has no center oi, , asymmetry. Recently, this reaction has been reexamined, using aminohydrir, CXLVI stereospecifically labeled in only one of the benzene rings with C14,7, , and it has been found that virtually all migration of Ph* takes place with inversiot, of configuration about Ca, whereas very nearly all migration of unlabeled pheny, , occurs with retention of configuration about Ca. This rather unexpected resul, , indicates not only that the aminohydrin greatly prefers conformation CXLV], , (in which the hydrogen bound to Ca lies between the two bulky phenyl groups or, , C/3), but also that the carbonium ion reacts preferentially in conformatior, , CXLVIII, leading to the observed inversion at C« and migration of Ph*. A, , small fraction (about 12 percent) of the carbonium-ion intermediate react!, , in conformation CXLIX (leading to retention at Ca and migration of unlabelec, , Ph), but practically none of the intermediate reacts in conformations CL 01, , CLI. It is reasonable that the carbonium-ion intermediate, when initially, , formed, should adopt conformation CXLVIII (which is similar to the preferrec, , conformation of the substrate). Some conversion to CXLIX (requiring a 60', , rotation around Ca—Qj) occurs, but conversion to CL (requiring a 120° rota¬, , tion) or to CLI (requiring a 180° rotation) is negligible. If this interpretation i:, , correct, the time interval between formation of the carbonium ion and migration of thi, , n Benjamin, Schaeffer, and Collins, J. Am. Chem. Soc., 79, 6160 (1957). Stercospedfo, labeling was accomplished by treating the SnCh adduct of benzene-labeled 2-aminopro, , piophenone (CXLVII) with unlabeled PhMgBr. As stipulated by Cram’s rule (p. 549), th(, , incoming (unlabeled) phenyl group should attack preferentially on the less crowded side (thai, , is on the left side in the drawing shown). Here the degree of stereospecificity is much grea e., than that which would be observed in the corresponding attack on a nonchelated substra, e, , for chelation assures us that virtually all molecules of the aminoketone assume conformatior, CXLVII., SnCl8, / \, O, NH, PhMgBr, , CXLVII, , H2Q, , ?, , CXLVI
Page 622 :
606, , -, , Participation of Neighboring Groups, , conformations for the erythro and threo forms of CLII are CLIII and CLV, respectively, for in both, the smallest substituent on Ca (—H) lies between the, two large aryl substituents on Cp. This is in contrast to the alternate conforma¬, tions CLIV and CLVI, in which the more bulky methyl group lies between the, two aryl substituents. Suppose, as seems to be the case with aminohydrin CXLVI,, that the carbonium-ion intermediate which is first formed resembles the more, favored conformation of the substrate, and suppose also that the aryl migration, is more rapid than rotation about Ca—Cp; phenyl migration should then pre¬, dominate in deamination of the erythro aminohydrin (as is observed), whereas, anisyl (An-) migration will predominate in deamination of the threo aminohydrin., The migrational preferences in these deaminations may also be accounted, for by assuming the intervention of aryl-bridged carbonium-ion intermediates., It may be argued that the favored bridged intermediates from CLIII and CLV, are CLVII and CLVIII, respectively; for in the alternate structures, CLIX and, CLX, there is aryl-methyl eclipsing. Opening the bridge at Cp in CLVII and, , 11, , (j^j), CLVII, , CLIX, , OMe, +, , CLX, , +, , CLVIII, , CLVIII leads the observed ketones. Since, however, deamination of aminohy¬, drin CXLVI has been shown to proceed partly, if not wholly, through an open, carbonium ion, it seems likely that the deaminations of CLIII and CLV likewise, proceed, at least in part, through open carbonium ions. There is, at present, a, similar ambiguity in interpretation of the rearrangements (in the presence o, each of the two carbons as L (largest),, , S, , (smallest), and M (medium-sized), then apply the, , threo and erythro prefixes in the usual way (p. 490)., , threo, , ery'hro
Page 623 :
Migratory Aptitude, , fn-/, Ag+) of the diastereomers of, \, , -, , 607, , V- r-Pti—nHPhRr 75 As with thedeamina/ y, '-•rirnjir, OH, , tions just considered, the identity of the migrating aryl group is different for the, threo- and <?ry/Aro-bromohydrins. Once again, we cannot say whether the observed, stereospecificity is due to conformational preferences in the reactant or to such, preferences in a bridged carbonium-ion intermediate (or perhaps to a combina¬, tion of both)., , Migratory Aptitude, During the 1930’s, a number of workers felt that a “migratory aptitude'’ could, be assigned to each of the various alkyl and aryl groups, representing the ability, of that group to shift between adjacent bound carbon atoms in molecular re¬, arrangements. It was hoped that, by applying these values to substrates having, two or more groups which might migrate, the ratio of possible products could be, estimated; implicit in this hope was the assumption that the ratio of migratory, aptitudes for two different groups was independent of the substrate at hand., From the previous sections it should be clear that the migratory aptitude of a, substituent is linked to its effectiveness as a neighboring group and this is, in, turn, governed largely by its nucleophilicity. Typically, the extent of migration, of substituent R, , in the deamination of RCH2—C*H:2NH2 decreases progres¬, , sively in the following series :80, r-ch-c*h2-nh2 —V HO-CH2-C*H2-R, , R-, , MeO-, , o, o2n-, , H3G, , o, , ; 33 percent rearrangement, , -; 24 percent rearrangement, , - ; 5 percent rearrangement, , ; 1 percent rearrangement, , Z pU[tin aad Meislich, J. Am. Chem. Soc., 74, 5905 (1952)., Roberts, et al., J. Am. Chem. Soc., 74, 5943 (1952); 75, 2069, 5759 (1953).
Page 624 :
608, , -, , Participation of Neighboring Groups, , rear¬, For exam¬, , Now, the difficulty in evaluating migratory aptitudes by examining the, rangement of unsymmetrically substituted pinacols should be evident., , ple, a methyl rather than a phenyl group migrates in the rearrangement of, glycol CLXI, not because methyl is a more effective migrator, but instead be¬, cause the phenylated carbon can support a positive charge in the carbonium-ion, intermediate, CLXII, more readily than can the methylated carbon. If thefor-, , Ph2C-CMe2 ^ Ph2C—CMe2, OH, , OH, , CLXI, , Ph2C-CMe -> Ph2C-CMe, , OH, , Me, , OH, , Me, , O, , CLXII, , mation of the initial carbonium ion is essentially irreversible, the group migratHO, , OH, , I, , I, , ing in the rearrangement of a pinacol R2C—CR'2 will be determined merely, by the relative stabilities of the two possible carbonium ions that may be ob¬, tained by removal of an -—OH group.81, To avoid this difficulty, we may consider systems having only one possible, migration terminus—that is, a-halohydrins (in their reaction with Ag+) or, a-aminohydrins (in their reaction with HN02). Here again, however, the pic¬, ture may not be straightforward, for, as we have seen, the group migrating in, the reaction of an erythro substrate may not be the same as that migrating in the, reaction of its threo isomer. In such cases, the identity of the migrating group, may depend on the conformation assumed by the reactant or the transition state, (or both) in order to minimize crowding., The most successful evaluation of migratory aptitudes has been accom¬, plished with “symmetric” tetraaryl pinacols of the type, Ar—C(OH)—C(OH)—Ar,8*, , I, Ar', , I, Ar', , in which the two possible reaction sites are chemically equivalent. The ratio of, the two ketones formed by migration of groups Ar— and Ar'— may be taken as, the ratio of the migratory aptitudes of Ar and Ar', and by studying the rear¬, rangements of a series of such “symmetric” pinacols, a number of migratory, aptitudes may be evaluated. The following are typical values (MA for Ph, , ,, , is, , «' Bachmann and Steinberger, J. Am. Chem Soc.,56,’1™ 0,934)., thc giyCols, « Bachmann and Ferguson, J. Am. Chem. Soc., 56 208 (, )•, ., However,, were not separated into the m«» and d,l forms before bang, thesame,, The steric requirements of the various aryl groups would be expected to be very nearly, so conformational effects should be of minor importance.
Page 625 :
Migratory Aptitude, , -, , 609, , fixed at 1.0):, OMe, , These values are not only reasonable, they are, to a remarkable degree, inter¬, nally consistent. For example, the migratory aptitudes of the />-tolyl group and, the jfr-biphenylyl group have been found (from examination of the rearrangements, of pinacols CLXIII and CLXIV to lie in the ratio 16/12 or 4/3). We should then, , ft, , V-/ V-CPh-CPh, s=j\=/, I, I, , OH, , OH, , OH, , OH, , CLXIV, , CLXIII, , expect rearrangement of pinacol CLXV to result in a mixture of ketones CLXVI, and CLXVII, in which the former (formed by migration of the jfr-tolyl group), predominates by a factor of 4/3. This is in excellent agreement with the observed, ratio of ketones:, , Me, OH OH, , 57 percent, , The very low migratory aptitude of the c-anisyl group may be attributed to steric, interference between the o-MeO group and the nonparticipating aryl groups in
Page 626 :
610, , -, , Participation of Neighboring Groups, , the transition state (CLXVIII):, , It is emphasized that migratory aptitudes derived from studies of one type, of rearrangement cannot be applied quantitatively to other rearrangements., More specifically, it appears that as the importance of “nucleophilic push” by, the migrating group increases, the migration becomes electronically more selec¬, tive. We have already noted that the identity of the migrating group in the, rearrangements of aminohydrin CLII and bromohydrin CLXIX is determined, largely by conformational effects, that is, the group migrating in rearrangement, of the threo form is in both cases different from that migrating in rearrangement, of the erythro form. In contrast, both the threo and erythro forms of brosylate, , MeO—y—CPh—CHMeNlI2, , Cl—f, , 7— CPh—CHPhBr, OH, , OH, CLXIX, , CLII, , CHPh—CHMeOBs, , MeO, , CLXX, , CLXX undergo acetolysis with the shift of the more effective migrator, the, /j-anisyl group, despite conformational effects.55 The latter reaction is of the, type for which evidence pointing to an aryl-bridged carbonium-ion intermediate, is more convincing.54, , Neighboring-group Participation in Elimination and Addition, Although almost all of our discussion thus far has been concerned with the role of, neighboring groups in substitution reactions (with or without rearrangement),, it is reasonable to expect neighboring-group participation in other types ol, reactions that proceed through carbonium-ion or carboniumlike intermediates,, in particular in E\ -type elimination reactions and in electrophilic additions., M Curtin and Bradley, Rec. Chem. Prog., 15, 121 (1954)., « For further work demonstrating the variations in relative ™gra ory, , ., pt, , nature of the reaction, see Raaen and Collins, J. Am. Chem. Soc., 80, 1409 (1958)., , ., , w:t'R the
Page 634 :
CHAPTER, , 15, , Further Molecular Rearrangements, , The pinacol rearrangement,, , the semipinacolic deamination, and the Wagner-, , Meerwein rearrangement were considered in the preceding chapter. Each of, these is, in essence, a 1,2 shift of a group (alkyl, aryl, or hydrogen), together, with its pair of bonding electrons. In each, both the migration origin and the, migration terminus are carbons. We now turn to a group of analogous 1,2 shifts, in which, however, the migration terminus is a nitrogen atom, the more impor¬, tant of these being the familiar Beckmann, Hofmann, and Curtius rearrange¬, ments. Known, but less usual, are 1,2 shifts in which the migration terminus is, an oxygen atom; these most often involve peroxy derivatives., , The Beckmann Rearrangement, The transformation of ketoximes to amides /R—C—N, \, , R', , OH —> R, , C, , NHR, , O, , /, , is catalyzed by such acidic reagents as H2S04, P2O5, SO3, SOCl2, PC15, and, sulfonyl chlorides. In moderately concentrated sulfuric acid, rates of rearrange¬, ment are proportional to Hammett’s h0 function/ suggesting (p. 190) that the, rate-determining step of the rearrangement involves the conjugate acid of the, oxime, R2C=N—OH+. With other catalysts, the rearrangement very probably, proceeds through an ester of the oxime (for example, R2C=N—0S02Ph); for, when such esters are prepared and purified, they are found to undergo rearrange¬, ment with ease in neutral solvents and in the absence of added catalysts. •, ' Hammett and Deyrup,, ' (a), , cZm., , 618, , 73%9*c551>:, , J. Am. Chem. Soc., ^’2721(1932), , 448., , fcu-, , Lampert and Bordwell,, , J- Am., , W Fo! evidence’that the Beckmann rearrangement may P—
Page 635 :
The Beckmann Rearrangement, , rearrangement, sidered typical, , 619, , of the benzenesulfonate of benzophenone oxime«‘> may be conThe rate-determining step in such sequences is doubtless t e, CHaCN, , Ph2C=N—0S02Ph-> Ph, , „ —N_ph, , c, , 0S02Ph, , -H^> Ph—C—NHPh + PhSOsH, , I, O, , heterolysis of the N—O bond, for the reactions have been found to be first order, in substrate/(o>fc) and to proceed most readily in polar solvents. Rearrangement, is facilitated by electron-attracting groups in the esterifying acid (which make, the anion from this acid a more effective leaving group), but inhibited by the, presence of such groups in the parent oxime (which would tend to strengthen the, N—O bond)., At present we cannot say whether or not the breakage of the N, , O bond, , in the Beckmann rearrangement and the shift of the alkyl (or aryl) group ate, concerted, but we may be certain that if these two acts are separated by an, interval of time, this interval is very short. For it is now firmly established that, the migrating group in this rearrangement approaches the nitrogen atom from the side, opposite to the departing oxygen atom. This has been demonstrated in several cases/, of which we need consider only the one indicated below:* * 5, , III, As shown, ketoxime I may be converted to benzisoxazole III with base, thus, through the oxime anhydride, R >C= -N—O—N—CRo, see Stephen and Staskun, J. Chem. Soc.,, , 1956, 980., , 5 (a) Catalysis by HC1 is complicated by the apparent necessity for small quantities of, amide; it has been suggested (Ref. 2a) that in this case the rearrangement proceeds through, oxime-amide adduct, R2C=N-0-CR=NR. (b) Beckmann rearrangements in liquid SO,, in the presence of Br2 or I2 have been described (Tokura, Asami, and Tada, J. Am. Chem. Soc.,, 31ri-1957].;A .h°Ugh mechamsms have been proposed, further quantitative investiga¬, tion of this modification is highly desirable., t■, , ‘, , ^ hor summaries of this question, see: Blatt, Chem. Revs., 12, 20 (1933) and Wheland, vanced Organic Chemistry, John Wiley and Sons, Inc., New York, 1949 p. 338., eisenheimer, Zimmcrmann, and von Kummer, Ann., 446, 205 (1926)
Page 636 :
620, , -, , Further Molecular Rearrangements, , indicating that the nitrated benzene ring lies on the same side of the C=N link¬, age as the —OH group. However, when ketoxime I undergoes the Beckmann, rearrangement, the phenyl group, rather than the nitrated benzene ring, shifts., The necessity for backside attack by the migrating group in the Beckmann re¬, arrangement suggests that migration occurs while the nitrogen atom is still, shielded by (and perhaps partially bonded to) the leaving group., A second stereochemical feature of this rearrangement deserves comment;, when the migrating carbon atom is asymmetric, the configuration at that carbon, is retained. Typically, ketoxime IV undergoes the Beckmann rearrangement (with, H2S04 in ether) with about 99 percent retention of configuration at the asym¬, metric carbon:5, Et, , O, , I*, , W-Bu —C!, , Me, , Me, , Et, , I, w-Bu— Ci, , H, N, , H, , OH, , (retention), -N, , H, , IV, The stereochemical evidence thus suggests structure VII as an activated complex, (or perhaps a high-energy intermediate) for the Beckmann rearrangement in, poorly ionizing solvents. As the solvent is made more polar, the complete de¬, parture of the OR— group (and its equilibration with other anions in solution), during the course of the reaction becomes more likely. Note that the migrating, group does not become completely detached from the remainder of the molecule, and that the breakage of the C—C bond and the formation of the new C—N bond take, place on the same side of the asymmetric carbon. This accounts for the observed reten¬, tion of configuration about the migrating carbon. Similar retention of configura¬, tion is very probably a feature of all 1,2-alkyl shifts in which the migrating, « Kenyon and Young, J. Chem. Soc., 1941, 263. To show that configuration was maintained, in this rearrangement, acid V (R— = n-BuEtCH—) was converted to amide VI by two, routes:, Hofmann, H2O, , (-J-)R—CONHo->, , ( + )R—COOH, , AcsO, , ,, , (+)R—NH2->, , (-)R—NHAc,, VI, , V, Me2Cd, , (+)R—COOH, , and, , ., , ,, , Beckmann, , (+) R—COBr-» (+)R—C—Me -» (+)R—C—Me->, , V, , O, , N—OH, (-)R—NHAc, VT, , These sequences prove that if the Hofmann rearrangement proceeds with retention of con¬, figuration, then the Beckmann rearrangement does also. Retention of configuration in, Hofmann rearrangement has been independently demonstrated (p. 622).
Page 637 :
The Hofmann Rearrangement, , Bu, , Bu, , Et, , /, , /, , “, , \, , +, , \, , Me, , Me, , Et, , V, , RO, , VIII, , Et, , Bu, , O, , ROH, , I, , R, , H, , MeG-NCH, II, \, , -> ^C=N, , C=N, , 621, , Bu, , H, , 6, , ", , -, , J, , VII, , (R=H0S022, , ClCH2C-} etc.), 2II, O, , group retains its pair of electrons, although such retention has not yet been, demonstrated in the pinacol and Wagner-Meerwein rearrangements., , The Hofmann Rearrangement iV', The rearrangement of N-chloro or N-bromoamides to isocyanates in basic, solutions, R—C—NH.Y + OH- -► R—N=C=0 + H20 + X~, , {X = Cl, Br), , O, is most often used to convert amides to amines (having one less carbon); for, N-haloamides are easily prepared from amides themselves, and amines are, readily formed when isocyanates are treated with water. This rearrangement, almost certainly proceeds through the conjugate base of the N-haloamide,, R, , C, , NA-; for with care, salts containing anions of this type may be prepared, , O, and are found to rearrange rapidly to isocyanates.7 Moreover, halogenated, N-alkylamides ^R, , C—NR'A^, which cannot form such an anion, do not un¬, , dergo the rearrangement., The rearrangement of N-bromobenzamide, PhC—NHBr, is accelerated bv, , I, , Y, , O, incorporation of electron-donating groups in the benzene ring; for such groups, not only ease the departure of the Br~ ion, but they facilitate the shift of the, benzene ring (p. 607).* Thus, we may represent the Hofmann rearrangement as, a simple 1,2 shift. Unlike the 1,2 shifts previously discussed, the migration termi7 Mauguin, Ann. Chim., [8], 22, 301 (1911)., * Hauser and Renfrow, J. Am. Chem. Soc., 59, 121 (1937).
Page 638 :
622, , Further Molecular Rearrangements, , R, NHX, \ /, , oh 7, , N— X, , R, , \ /", , slow, , C, , ?, , ^, , O, , R, /\, / \ x—►, CtN-X, 11 vO, , r, ^ \, p XT T? H20 j, > O—O—N—R -W, , rnh2, , lco2, , IX, , nus does not acquire a positive charge during the course of the rearrangement., However, the loss of a halide ion from anion IX would leave behind the species, R—C—N:, in which the nitrogen atom has only a sextet of valence electrons and is, , O, thus electron deficient in much the same sense as is the boron atom in B(CH3)3., There is some doubt that this electron-deficient nitrogen species has an inde¬, pendent existence, for it may be argued that such a molecule would react rapidly, with water to give a hydroxamic acid, R—C—NHOH, whereas no hydroxamic, O, acids or products closely related to them have been found in the reaction mix¬, tures.5 For this reason, it is suggested that the departure of the halide ion and, the shift of the migrating group in the Hofmann reaction are simultaneous or very, nearly so., As in the Beckmann rearrangement, we would expect retention of con¬, figuration at the migrating atom. This prediction has been confirmed most, directly with amide X, derived from camphoric acid. When this amide, in, which the —COOH and —CONH2 groups lie cis to each other, is subject to, the Hofmann rearrangement, amino acid XI results. In the latter, the, , COOH, , and —NH2 groups also lie cis to each other, for this amino acid readily forms, lactam XII. (The trans amino acid (XIII), which would result from inversion of, configuration at the migrating carbon, should not form such a lactam.^) More-, , Me, , Me, , Me, , yL, , NH2, , VV/^COOH, , COOH, , /, , h^CONH, , Me, , Me Me, , X, , over, the bicyclic amide XIV readily undergoes the Hofmann rearrangement. In, this case, the rigidity of the ring system prohibits inversion of configuration, at Ca.”, 0 Hauser and Kantor, J. Am. Chem. Soc.,72, 4284 (1950)., 10 Noves et at J. Am. Chem. Soc., 34, 1067 (1912); 36, 118 (1, ibii, , H915). Archer,, ),, », (, ‘ r r the, 62, [972 (1940). Amide X is known to be derived from a m-dicarboxylic aci ,, , parent acid forms a cyclic anhydride on heating., /10-*cn, u Bartlett and Knox, J. Am. Chem. Soc., 61, 3184 (1939)..
Page 639 :
623, , Reactions of Azides, , OBr", , h9nc, , o, XIV, The Lessen rearrangement, which involves N-acyl derivatives of hydroxamic, acids (XV), is obviously analogous to the Hofmann rearrangement, but here the, leaving group is a carboxylate ion, rather than a halide. As expected, the reaction, R—N=C=0, B:, , R, , NH-O-C-R', , XO, , \/, C, , R, N-O-C-R', \ /II, , c, o, , o, , ’ ^ ^, G-1-NL0-C-R', II *0*, II, , o, , +, O—C—R', , o, , II, O, , XV, is facilitated by electron-supplying substituents in R— but retarded by such, substituents in R'—.ls Again, virtually complete retention of configuration in, R— is observed in its migration.*3, , Reactions of Azides. The Curtius and Schmidt Rearrangements14, Inorganic and organic azides, aside from a few ionic azides such as NaN3, are, relatively unstable substances. Often, in their decompositions, two of the three, nitrogen atoms depart as N2, leaving the third nitrogen atom and the remain¬, der of the molecule to stabilize themselves as best they can. Acyl azides, , (, , R—C—N—N=N\, which may be prepared either by the action of NaN3 on, O, , /, , acyl halides or by the action of nitrous acid on acid hydrazides (RCONHNHo),, decompose when, , heated,, , yielding, , nitrogen, , and, , isocyanates. The Curtius, , rearrangement may then be represented:, R. ^N-^NsN, , X X', , c, , —eat > 0 = C=N-R, , +, , N,, , 2, , II, o, '* Renfrow and Hauser, J. Am. Chem. Soc., 59, 2308 (1937)., /’ W*llis’ Nagel, and Dripps, J. Am. Chem. Soc., 53, 2787 (1931) • 55 1701 (19331 Th^
Page 640 :
624, , Further Molecular Rearrangements, , As shown, the mechanism is very similar to those for the Hofmann and, , Lossen, , rearrangements. Once again, retention of configuration is observed in the, migrating group15 and we have represented the departure of the leaving group, and the shift of the migrating group as being synchronous., The Curtius reaction is subject to acid catalysis.This, at first glance, may, seem surprising; for it might be argued that protonation of either the acyl group, or the nondeparting nitrogen should pull electron density from the middle nitro¬, gen in XVI toward the nitrogen on the left, thus making the heterolysis of this, bond more difficult. However, from the 7r-electron diagram, XVII, we see that, the N—N bond broken in the reaction has double-bond character. Protonation, , XVII, of the acyl group draws 7r-electron density from the nitrogen at the left even, more effectively than it draws <r-electron density (for 7r electrons are the more, polarizable). This reduces the double-bond character of the N—N bond at the, left and facilitates its breakage., The Schmidt rearrangement is actually a group of reactions, each employ¬, ing as a reagent hydrazoic acid (HN3) in sulfuric acid. We shall be concerned, here only with the action of this reagent on carboxylic acids and carbonyl com¬, pounds. Carboxylic acids are converted to amines (having one less carbon atom),, doubtless through an isocyanate, which is, however, generally not isolable from, the reaction mixture. This reaction is thus closely related to the Curtius rearrange¬, ment and may be assigned the following mechanism:, OH, , HO, \, , /+, , slow, , C=N—N=N, , XVIII, , XIX, , XX, , XXI
Page 641 :
Reactions of Azides, , 625, , Cation XVIII is the conjugate acid of the carboxylic acid that is taken as the, substrate, whereas cation XXI is the conjugate acid of the isocyanate from which, the resulting amine is presumably derived. The third step in the sequence is, assumed to be rate determining, both by analogy, and in accord with the obser¬, vation that electron-donating groups in R— increase the rate of rearrangement., Cation XX, rather than XIX, has been designated as the active intermediate, in the rearrangement; for if XIX could undergo rearrangement directly, there, seems no reason why N-alkylated cations of the same type (for example, XXIV), could not also undergo rearrangement, and if this were the case, alkyl azides, (for example, CH3N3) should enter into the Schmidt reaction, converting, carboxylic acids to secondary amines. Such a modification of the Schmidt reac¬, tion is generally unsuccessful. As with the other rearrangements considered thus, , /, , -Ns, , HoO, , -> 0=C-N, , COz + R'NHR, , \, R', (not observed), , far in the present chapter, the migrating group in the Schmidt reaction shifts, with retention of configuration.18, n Briggs and Lyttleton, J. Chem. Soc., 1943, 421. The ease with which mesitoic acid under¬, goes the Schmidt rearrangement has aroused some interest. Since mesitoic acid is known from, cryoscopic measurements (p. 99) to be very nearly completely converted to the mesitylium, ion XXII, in concentrated sulfuric acid, it is often suggested that the reaction proceeds, through this cation:, , H, RCOOH, , H2S04, , I, , HN, , ♦, , +, , > R—G=0-R— C—N—N=N, , II, , /, , O, XXII, , -Na, , C=N—N=N, , HO, , XXIII, , XX, , Me, , R—N=C-OH, , (R = Me—^—, Me, However this view is inadmissible if it is assumed that the rate of rearrangement is determined, by the rate of decomposition of cation XX; for if XX were in equilibrium with both XXII and, XXHI any stenc and electronic factors that would favor the conversion of mesitoic acid to the, , Z, , otheOiand, .T “ “J'***” 'h', °f'“«*•» XXIII to the mesiTyltu™ ion o“ £, deV^rm d’ h, obser1ved reactivity of mesitoic acid is consistent with a path in which the rateeterminmg step is the attack by the mesitylium ion XXII on HM HicrJa •, vr t \, , Campbell and Kenyon, J. Chem. Soc.,, , 1946,, , 26.
Page 642 :
626, , -, , Further Molecular Rearrangements, , When a symmetric ketone (R2C=0) is treated with hydrazoic acid in, sulfuric acid, an amide results. With an unsymmetric ketone (RR'C=0), a mix¬, ture of two amides is obtained, whereas an aldehyde generally yields a mixture, of an N-alkylated formamide and a nitrile. Although little quantitative study, has been devoted to this phase of the Schmidt reaction, the following mechanism,, based largely on analogy, may be considered a likely one:, R H, R\+, hn3, ,11, +, C-OH=?=* R-C-N—NsN, , HoO, , I, , R/, , o, , OH, XXV, , XXVI, , XXVIII, , Here, cation XXV is the conjugate acid of the aldehyde or ketone used. Once, again we have suggested that N2 is lost from cation XXVII rather than its, hydrated form, XXVI, because of the presumed reluctance of alkyl and aryl, azides to undergo an analogous reaction.19 When an aldehyde is used, the inter¬, mediate, , corresponding, , to, , XXVIII,, , formed, , from, , hydride, , migration,, , is, , R—C=N—H. This cation may be assumed to form the observed nitrile, directly (by loss of a proton), rather than being first converted to the amide, for, under the conditions of the Schmidt reaction, amides cannot generally be, dehydrated to nitriles., With unsymmetric ketones the question arises as to which of the two possible, amides will predominate. The intermediate from which nitrogen is presumably, lost (cation XXVII) has a structure very much like that of an oxime. In analogy, with the Beckmann rearrangement, we should then expect the migrating group, to approach the migration terminus (the left-hand nitrogen in XXVII) from one, side while N2 leaves from the opposite side. When group R— lies trans (anti), to the azido group, then R— will migrate; when group R' lies trans to the azido, group, then R'— will migrate. Cation XXVII will probably exist in both possi¬, ble forms, but that form will predominate in which the larger of the two groups, lies trans to the azido group (for in this form, steric interference will be less)., We would thus expect the bulkier of the two groups in the parent ketone to, « However, it has been found (Boyer and Hamer, J. Am. Chem. S°e., 77 951■, , that, , a azido alcohols (XXIX) react with aromatic aldehydes in concentrated sulfu i, , 2-aryloxazoHnes (XXX). I. is possible ,ha, this reaction proceeds by a path analogous to that, proposed for the Schmidt reaction., , ArCH + HOCH2CHR-N3-, , I, , r., , (SA, •ArC-N-N=N, I, , \, , HO, , OH, , XXIX, , CHR, CH, , I, OH, , OH, -N, 1, _—A ArC—NH, +, \, , |, , — Hh, + ArC=N, -HoO, , ^CHR, , ho-ch2, , XCHR, , /, , o-ch2, , XXX
Page 643 :
627, , Reactions of Azides, , migrate preferentially in the Schmidt reaction, and this, generally speaking is, what is observed." The ethyl group migrates preferent.ally in methyl ethyl, ketone, the phenyl group in acetophenone, but very nearly, , equal, , quantities ol, , the two possible amides are obtained from isobutyrophenone. Moreover, unsymmetrically /.^-substituted, , benzophenones generally give, , nearly equal, , quantities of both amides, for para substituents are held well out of the way of the, departing nitrogen molecule and their steric effects in this reaction are negligible., When an excess of HN3 is used in the Schmidt reaction, a substituted tetrazole (XXXI) may be formed.*' Since amides themselves do not react with HN3, to form tetrazoles, we may assume that this product is formed by the action of a, second molecule of hydrazoic acid on intermediate XXVIII., , fc-N, , R—N=C—R' + H-N-N=N, , —H+, , N=N-N-H, +, , XXVIII, , C-R', , R—N=C-R'-* R—N, , N, , N, V, , N, , XXXI, (a tetrazole), , Both the Gurtius and Schmidt rearrangements have analogs in which the, leaving group is a N2 molecule and the migration terminus is an electrondeficient but uncharged carbon atom. The Wolff rearrangement of diazoketones**, is formally similar to the Curtius rearrangement, and may be assigned the, mechanism:, , “, , R, , +, , -No, , CH-N=N —, , V, ii, o, , a diazomethyl, ketone, , R^^CH:-^ RCH=C=0, G, , II, , RCH2COOH, , a ketene, , O, XXXII, , "Sanford, Blair, Arroya, and Sherk, J. Am. Chem. Soc., 67, 1942 (1945). See also Alex¬, ander, Principles of Ionic Organic Reactions, John Wiley and Sons, Inc., New York, 1950 p 71, F°r recem evidence that electronic effects play little part in determining migratory aptitudes, in the Schmidt reaction, see Saunders and Ware, J. Am. Chem. Soc., 80, 3328 (1958), " Spielman and Austin, J. Am. Chem. Soc., 59, 2658 (1937)., , methaSiaGHeNneSTh^ w~GJ*—N2, ^re generally prepared from acyl halides and diazosynthesis by which a carboxylic, , R-cIltoOH
Page 644 :
628, , Further Molecular Rearrangements, , The following evidence supports this mechanism: (a) the carbonyl carbon in the, diazomethyl ketone becomes the carboxyl carbon in the resulting acid (as has, been shown in studies with C13);*3^ (b) group R— migrates with retention of, configuration ;~3(6) and (c) in favorable cases and in the absence of water or, alcohols, the intermediate ketene may be isolated.*3^ We have suggested here, that the loss of nitrogen occurs before, rather than concurrently with the migra¬, tion of R, , , for one of the products sometimes resulting from the decomposition, , of the diazo ketone in water is the hydroxy ketone, R—C—CH2OH. While it is, , O, possible that this could result directly from the diazo ketone itself, it seems more, likely that it is formed by hydrolysis of the electron-deficient species XXXII dur¬, ing the short time interval between the departure of N2 and the rearrangement, to the ketene., O, , OHO, , I, , R—C—CH: + H20, , I, , .., , +/, , ”, , \, , I, , R—C—CH—O, , XXXII, , -> R—C—CH2OH, H, , There is, however, one important feature of the Wolff rearrangement which, the proposed heterolytic mechanism does not explain; it is often effectively, catalyzed by silver compounds (for example, Ag20, Ag(NMe3)^OBz ) under, basic conditions. Since univalent silver may act as a single-electron acceptor, we, cannot dismiss the possibility that this rearrangement, at least in some cases, may, proceed by a free-radical mechanism such as the following :eJ>, , O, R—C—CHN2, , O, , o, B:, , S', , R-C-C-N =N^> R'V^G^N=N, , 0=C=G-R, , H, H,0, , O, then R—G—C—O + R, , C, , R—c=c=o, CHNr, , RCH2COOH, -N,, , R-C-C—N=N, , etc., , II, O, Such a sequence, in the absence of further investigation, must be regarded as, tentative., ., ..., The reaction of ketones with diazomethane is similar to their reaction with, , Warn0^1 )TvViI:ry, Hirzel, Be,., 49, 2523 (1916); Schroeter, Be,., 42, 2346 ( 909) 49,, u Newman and Beal, J. Am. Chem. Soc., 72, 2438 (, )•
Page 645 :
629, , Rearrangements of Peroxy Derivatives, , hvdrazoic acid (CH2N2 and HNj have like structures and have the same numbe., of electrons). With diazomethane the chief product is generally a new ketone, having an extra -CH2— linkage adjacent to the carbonyl group, with a substi¬, tuted ethylene oxide formed in minor amounts.*5 In the absence of quantitative, data concerning this reaction, we may provisionally assign to it the following, mechanism (in analogy with the Schmidt reaction):, , r-G-R' + :CH2—N=N, , II, , O, , Here, it is suggested that loss of nitrogen from intermediate XXXIII may be, accompanied either by shift of R (or R'), yielding a ketone, or by attack by the, negatively charged oxygen atom, forming the epoxide ring., , Rearrangements of Peroxy Derivatives, Depending upon conditions, an O—O bond may break homolytically (—O—O—, ••, , ••, , —► 2-0—), or heterolytically (—O—O-> ~:0-1- +0—), and in some cases, ••, , ••, , both modes of breakage may operate together. This duality complicates the, study of the reactions of peroxides and peroxy derivatives, although there are z., number of diagnostic tests which, when taken together, will tell with some relia¬, bility whether homolytic or heterolytic cleavage is occurring., , ( J|, Peroxy acids \R, , C, , ), O, , O, , HJ attack ketones in a manner which is, , formally analogous to attack by diazomethane or by hydrazoic acid. Diazo¬, methane, as we have seen, interposes a —CH2— group between the —C— group, , II, O, and an a-carbon, whereas hydrazoic acid interposes an —NH— group in this, position. Peroxy acids insert an -O- linkage, thus converting the ketone, , o, ^, , 6*, , °, , R to the ester R—C—OR/, c1 od or to, . a mixture, ., ^, wa , to, io the, me ester, ester R', K —G—UR,, of., , O, , 52, 3456 (1930).and Kenner’ J' Chem* Soc-> 1939> 1815 Mosettig and Burger, J. Am. Chem. Soc.,
Page 646 :
630, , Further Molecular Rearrangements, , the two. It has been shown, using labeled oxygen, that the carbonyl oxygen im, , the ketone becomes the acyl oxygen in the resulting ester,2e(o) and, in furthen, analogy to the Schmidt reaction, the migrating alkyl group has been found to, shift with retention of configuration.2^ We thus may infer that, as with HN3, the, peroxy acid first adds to the C=0 linkage, and that the adduct (XXXIV) then, suffers heterolytic cleavage at the 0—0 bond while (or just before) a 1,2 shift, H, , xo, I, Ac-OOH + R2G+, peroxyacetic, acid, , HO, , R, +, , —H+, , =±R2C-0-0Ac^=^, , C-O^OAc — Ac >, , i, , H, , Rqh, XXXIV, , O, , OH, , I, , -h+, , II, , R-C-OR -RCOR, +, , occurs from carbon to oxygen. (As yet, we cannot say whether the formation of, the adduct or its rearrangement is rate determining.), We would anticipate that when an unsymmetric ketone is subjected to this, reaction, the identity of the predominant migration group will depend on the, “migratory aptitudes” of the two substituents, and, when these substituents differ, significantly in bulk, on conformational effects as well. Migratory aptitudes alone, should govern the direction of rearrangement of unsymmetrically /><2rtf-substituted benzophenones; for, as in the Schmidt reaction, the steric effects of para, substituents should be negligible. Accordingly, it is not surprising that when, this reaction is carried out with peroxyacetic acid in glacial acetic acid, the, jf?-anisyl group is found to be a “better” migrator than the phenyl group, which, is, in turn, a better migrator than the /?-nitrophenyl group.v' As in the pinacol, rearrangement, the more nucleophilic substituents appear to shift more easily,, probably because of the ease with which they form “bridged” intermediates or, transition states. Under the same conditions, benzoylcyclohexane gives very, nearly equal (but small) yields of the two possible esters. Here, it may be sup¬, posed that although the benzene ring has a much greater “intrinsic ’ migratory, aptitude than the cyclohexyl group, the peroxy intermediate prefers conforma¬, tion XXXV, in which the bulky cyclohexyl group lies trans to the acetate, and, therefore in a position favoring cyclohexyl migrationOn the other hand, when, , uj;, XXXV can affect the ratio of products only if rotation abou
Page 647 :
Rearrangements of Peroxy Derivatives, , 631, , the same reaction is carried out in chloroform, cyclohexyl migration greatly, predominates over phenyl migration/3 although neither migratory aptitudes nor, conformational effects should vary with the solvent used. It is probable then, that a free-radical mechanism has become predominant in the less polar solvent,, and it may well be that a significant, though smaller, fraction of the reaction, proceeds by a radical path in the more polar solvents also. We may assume that, adduct XXXIV' is common to both mechanisms, a possible path for its homolytic decomposition being as follows:, AcO* +, R, , \, C—O—OAc, R', , R, , /I, , OH, , ^C—O- (R> Shlfts> R'-C-OR, , XXXIV', R', , /I, , I, , OH, , OH, , then., , O, , OAc, , C—OR —^ R'_c—OR, , R'—, R, , \, R', , C, , OR +, , OH, C—O—OAc, R, , OH, , R', , OH, etc., , c—o-/I, , XXXIV, R', , OH, , Fiiess and Farnham, J. Am. Chem. Soc., 72, 5518 (1950)
Page 648 :
632, , -, , Further Molecular Rearrangements, , (Chain-reaction mechanisms, of which this is one, are to be considered in the, following chapter.), A similar duality in mechanism is associated with the decomposition of, /?-methoxy-//-nitrobenzoylperoxide, , (XXXVI).50, , In, , nitrobenzene, , and, , in, , thionyl chloride, this unsymmetric diacyl peroxide rearranges predominately to, carbonate XXXVIII. The preferential mode of heterolysis of the O—O bond is, , XXXVIII, in the direction shown (XXXVII), rather than in the opposite direction (which, would result in formation of/>-nitrophenol and />-methoxybenzoic acid), doubt¬, less because the nitro group is a much stronger electron attractor than the, methoxy group. Conversely, the/>-anisyl group is far more effective in stabilizing, the positively charged fragment. However, when the decomposition of peroxide, XXXVI is carried out in benzene, very nearly equal quantities of/>-nitrobenzoic, and /?-methoxybenzoic acids are formed, along with little, if any, phenolic, O, , - w u-v-^ «-•, , a similar study of the decomposition ol 4-metnoxy j ,, and Petropoulos, ibid., 79, 3068 (1957).
Page 649 :
Rearrangements of Peroxy Derivatives, , -, , 633, , products.S0(o) Under these poorly ionizing conditions, radical decomposition has, evidently taken over., The benzoate and substituted benzoates (XXXIX) of trans-9-decalyl, , v-, , dropcroxide constitute an interesting series. These rearrange with migration of a, bridgehead carbon, yielding derivatives of 1,6-epoxycyclodecane, , (XL)., , This is obviously a heten .lytic reaction, for the rates of rearrangement may be, greatly increased by transferring the reaction from nonpolar to polar solvents., Electron-attracting substituents on, , “departing, , effectiveness”, , the, , l, , benzene ring, , ), , of Ar—C—O /, , accelerate, , (which improve the, , the, , rearrangement,, , whereas electron-donating groups retard it.s;(6) At present it appears that the, movement of the benzoate ion (or substituted benzoate ion) is concurrent with, the formation of the new C—O bond in the ring—that is, that the transition, state may be represented as XLI. For if the rearrangement of the peroxybenzoate, , XL1II, , is carried out in the presence of excess /,-nitrobenzoate ion, no /i-nitrobenzoate is, incorporated into the resulting ester, showing that the benzoate ion stays asso¬, ciated w ith the decalyl system during the entire course of the rearrangement.s*(c), oreover, if the carbonyl oxygen of the peroxy ester is labeled with O18, virmaUy, g oup,, , an, , of the, , 0>», , in the resulting epoxy ester,, , XL,, , is found in the carbonyl, , whereas if the anion had strayed far enough from the decalyl system, , o al ow the two benzoate oxygens to become equivalent, only half of the 0“, should remain m the carbonyl group. In retrospect, it is not surprising that this, , j., , Z. SLfefe ntiYlTa, , A""" 5?d27’ 135 (1948>- (*> Bartlett and Kice,, and Denney, ibid., 77, 1706 (1955), 79, 4806^57). ^, ?5’ 5853 (1953)' ^ Denney
Page 650 :
634, , -, , Further Molecular Rearrangements, , rearrangement does not pass through “classical” (nonbridged) cations. An, oxonium, , ion such as XLII, having a positive charge solely on an oxygen atom,, , would be expected to be quite unstable, and a carbonium ion such as XLIII, would have a positively charged carbon at the bridgehead of a bicyclic system—, a very unfavorable location, since coplanarity of the bonds to that carbon would, involve considerable strain (p. 280)., The geometric relationship between the positive and negative ions in the, rearrangement of the unsymmetric diacyl peroxide XXXVI is more complex., When the carbonyl oxygen of the />-nitrobenzoyl group in the peroxide is, , labeled with O18, only about 66 percent of the labeled oxygen is found in the, , jfr-nitrobenzoyl group, the other 34 percent being found at the position between, , the two carbonyl groups in the product. Thus, the rearrangement cannot pro¬, , ceed solely through ion pair XLIV nor wholly through XLV; for if the reaction, , chose the first of these paths, none of the labeled oxygen would move to the posi¬, , tion between the carbonyl groups, whereas with the second path, all of the, , labeled oxygen would so move. Moreover, the reaction cannot proceed solely, , through an ion pair in which the carbonyl oxygens in the nitrobenzoate ion have, , +, , OMe, , OMe, , +, , OMe, , (XLVI), , no2, , no2, , no2, , become equivalent, for in this case, just one half of the labeled oxygen would be, , found at the position between the two carbonyl groups. The observed degree of, , O18 rearrangement is, however, consistent with a combination of two or all three, , of these paths, or alternately, with an intermediate such as XLVI, in which, the carboxyl oxygens of the /,-nitrobenzoate group are not equidistant from t, , carbonyl carbon of the ^-anisoyl group. In the latter case two modes of forma¬, tion of the new C-O bond would be possible, but attachment by the unlabeled, oxygen would predominate.
Page 651 :
The Benzilic Acid and Related Rearrangements, , -, , The Benzilic Acid and Related Rearrangements, The benzilic-acid rearrangement—the conversion in strong base of aromatic, 1,2-diketones to salts of a-hydroxycarboxylic acids—is a 1,2 shift in which the, migration terminus is a slightly electron-deficient carbon atom of a C, , O group., , Ar, , ai, , 1, o, 1, , II, , OH-, , ©%, , 1, , slow, , fast, , —, , -Ar-vHO-G-C-Ar -►HO —c- C-Ar -> O, , II, , o o, , fast, eq, , a benzil, , |k, , |, , o_ oj, , 1, , o_, , O, , _, , „., , C, II, O, , CAr2, 1, OH, , anion of, benzilic acid, , XL VII, , In our present picture for the mechanism of this rearrangement, the migration, may be said to be due more to a “push” from the electron-rich migration origin, than to a “pull” by the migration terminus. The second-order rate law for the, reaction (first order each in diketone and hydroxide)33(a) does not tell us whether, the formation of adduct XLVII is fast or is rate determining (or, indeed,, whether this anion is actually an intermediate or merely an activated complex)., However, the observation that the diketone undergoes O18 exchange in oxygenlabeled water in the presence of base, at a rate much greater than that of the, rearrangement,'S3(6) indicates that XLVII is an intermediate and that its forma¬, tion is rapid and reversible.33 (The reasoning here is the same as that used to, OH, demonstrate the existence of the intermediate R'—C—O- in ester saponificaOR, , tion, p. 317.), 880 a93)8)WeStheimer’ J' ^‘, , S0C'' 58’ 2209 (1936)- W Roberts and Urey> ibid., 60,, , 77 3280^9^, ^ Hendley> and Neville> J&*.,, sim.of ° 1955 I, hC rate-determining step of the benzilic acid rearrangement involves, simultaneous aryl m.gration and proton transfer. However, the rearrangement of benzil has, Ar, -Ar, , o=c, I, _o, , I, , c-, , -Ar, , I, H—O, , been found (Hine and Haworth, J. Am. Chem Soc 80 2274 nossn *, , ,, " rapidly in D20-dioxane fin which On’, ’ Th1 958)) to Proceed almost twice, HjO-dioxanc (in »h“h OHbecomes^LhediThi^H, '° 'he ™*ratio" °r«in) “ *, is not involved in the rate-determining, r, ’ h indlc?tes strongly that proton transfer, to be faster than deuteron transfers (see forT, The observed increase, , t.^.?sfers are known almost invariablv, ^g’ CW, 55> 713 0955))., , tainly due to the fact that OD~ is a considerablvstron''^^ ^ *I*° ~ S* CaSC iS alm°St cer*, Nelson and Butler, J. Chem. Soc., 1938, 957)., §Cr 53Se than, ’ (see> for example,
Page 652 :
636, , -, , Further Molecular Rearrangements, , The benzilic ester rearrangement is a closely allied reaction, but here the initial, attack is by alkoxide (for example, MeO~, *-BuC>-) rather than by, , hydroxide, , As may be expected, the product is an ester, Ar2G(OH)—COOR, rather than, the salt of a benzilic acid.54 This variant does not proceed well with EtO~ or, z-PrO-,55 for these anions readily reduce benzil to benzoin. The reaction also, fails with phenoxides and substituted phenoxides,53(o) presumably because these, anions are too weakly basic to carry out the initial attack., In a similar manner, an aryl group is “pushed” from one carbon to another, in the reaction of o-tolylmagnesium bromide with benzil. In this case, the prod¬, uct is the a-hydroxy-ketone XLIX, formed by phenyl migration :36, , /^/MgBr, , +, , \^Me, , PhC—CPh, , II, , ->, , II, , o o, , Me, , Me, , ,G—CPh 2, A, o o, , I, MgBr, , It is likely that this rearrangement is facilitated by chelation with the magnesium, atom (XLVIII), which increases the electrophilic character of the carbonyl, group in the initial adduct. Note that the phenyl group migrates in, , preference, , to, , the somewhat more nucleophilic o-tolyl group, for, as is often the case, migration, of an ^-substituted aryl group involves an unduly crowded transition state., Nucleophilic “push” and electrophilic “pull” probably operate together, also in the rearrangement of the bromomagnesium derivative of the cts-c, , oro-, , hydrin, L.57 As shown in structure LI, the (partial) positive charge on, magnesium “pulls” at the chlorine while the (partial) negative charge on, oxygen facilitates the shift of the methyl group. In accord with t is pic u, H Doering and Urban, J. Am. Chem. Soc., 78, 5938 (1956)., as T arhman, , - Roge^nd, , J Am. Chem. Soc., 45, 1509 (1923)., . uas been, McGregor, 7. Chm. Soc, 1934, 442. An analogous rearrangement, , observed with mesitylmagnesium bromide (Ref. 34), 37 Geissman and Akawie, J. Am. Chem. Soc, 73, 199, , (, , )■
Page 653 : Rearrangements of Aldehydes and Ketones in Acids, , -, , 637, , trans isomer of L (chlorohydrin LI 11) gives only traces of ketone LI I when, treated with MeMgBr., , Rearrangements of Aldehydes and Ketones in Acid, Protonation of a carbonyl group greatly increases its electrophilic character and, may initiate rearrangement when nearby alkyl or aryl groups are in a position, favorable for migration (and when a shift is energetically advantageous). Little, quantitative study has been devoted to such rearrangements, but they are be¬, lieved to proceed in much the same way as the pinacol rearrangement. The, simplest occur with a-hydroxy-aldehydes and a-hydroxy-ketones. In the first of, Ph, , Ph, , Ph2C—CH % Ph—C—C—H -> Ph—C—C—H —^ Ph—C—CHPhs*<a>, HO, , O, , HO OH, LIV, Me, , HO OH, LV, , O, , OH, , Me, , Me2C—C—CHMe2 ^ Me—C—C—CHMe2 -> Me-C-C-CHMe, — ■>, HO, , O, , HO, , OH, LVI, , HO, , OH, , LVI I, Me, Me—C—C—CHMe2SS(6), O, , OH, , these, the system gains stability as a result of the conversion of LIV to LV, since, in the latter the positive charge lies adjacent to a benzene ring and may distribute, itself over the ring (p. 253). In the second rearrangement, the system gains, hyperconjugative stability when LVI is transformed to LVII, for in the LVII, 58 (a) Danilov and Venus-Danilova, Ber., 59, 377, 1032 (1926)- 60 2390 i'1927'i m, Oumnov, Bull. Soc. chim. France, 43, 568 (1928)., ', ’, (1927). (6)
Page 654 :
638, , -, , Further Molecular Rearrangements, , there are three hydrogens in the /3 positions, whereas in LVI there is only a single, hydrogen. (A bridged carbonium-ion intermediate may intervene in the first, rearrangement, but probably not in the second.), The rearrangements of nonhydroxylated aldehydes and ketones in strong, acid are more complex, for each of these reactions requires at least two shifts; for, example,, Ph, , Ph, , H, , Ph2CHCHO ^ PhCH—CH -» PhCH—CH -> PhCH—CPh, OH, , OH, , OH, PhCH2—C—Ph5s(o), , II, , O, Me, , Me3C, , I, , H+, Me3C-C-CMe3 —*■ Me2G— C-CMes, O, , OH, , Me2C—C^CMe, , +, , -H+, , Me2C —C—Me, , I, , OH, , OH, , LIX, , LVIII, , Me3G, , 39, , I, Me2C—C—Me, , O, LX, , These isomerizations are, in essence, exchanges of substituents between the, carbonyl carbon and the carbon alpha to it. In the first of the rearrangements, above, that of diphenylacetaldehyde, a phenyl group moves to the carbonyl, carbon, after which a hydrogen shifts back to the a-carbon. In the second re¬, arrangement, that of hexamethylacetone (LVIII), a shift of a methyl is followed, by the “back shift” of a *-butyl group. A second path for conversions of the type, LVIII —> LX might be considered;^0 intermediate LIX could conceivably, “close” to a protonated epoxide ring, LXI, which might then undergo ring, opening with a methyl shift (rearrangements of this type have been observed, with substituted ethylene oxides***’)- However, this alternate path is ruled out, , Me, Me3G-G-CMe8, , Hh, , +, , I, , -Me,C—C-CMe3, , I, O, , OH, , /Me, MeC-C-CMe,, , ?♦, H, , LXVIII, , LIX, , LXI, , -Me —C—CMe2-CMe3, HO, , —H+, , ■> LX, , , r> *, r ri.a™ Cnr 1456 2483• (b) Zook and Paviak, J. Am. Chem., (a) Barton and Porter, J. Chem. i>oc., 14ao,, t >, Soc., 77, 2501 (1955)., 4436 (1957). (b) See, for example,, 4o (a) Zook, Smith, and Greene, J. Am. Chem. i>oc., 74,, U, \, Lagrave, Ann. chim., 8, 363 (1927).
Page 655 :
639, , Rearrangements of Aldehydes and Ketones in Acid, , (at least in the example shown) by carrying out the rearrangement on a sample, of ketone LVIII that has been labeled at the carbonyl carbon with C14, for it, has thus been demonstrated that the carbonyl carbon in ketone LVIII remains, a carbonyl carbon in ketone LX.55(o) If the rearrangement proceeded through, a protonated ethylene oxide as shown, the carbonyl carbon in LVIII would, become an a-carbon in LX., In the dieneone-phenol rearrangement, the migration terminus is not the carbon, of a protonated carbonyl group, but rather a carbon in conjugation with it., As the name implies, this reaction results in the transformation of a quinoid, structure to a benzenoid ring. Typically, treatment of ketone LXII with sulfuric, acid and acetic anhydride yields acetate LXIIL4/(o) In a like manner, ketone, , LXII I, , LXIV, when treated with acid, yields naphthol LXV;4'<6> as we might have, predicted, the phenyl, rather than the methyl group, has shifted. Similarly,, dienone LXVI yields 6-hydroxytetralin (LXVII).*®, , O, , OH, , LXV, *• (a) Arnold, Buckley, and Richi, Buckley, ibid., 71, 1781 (1949)., *• Winstein, Heck, and Baird,, , J., , O, , OH, , LXVI, , J. Am. Chem. Soc., 69, 2322 (1947)., Chem. Soc., 79, 3109 (1957)., , (b), , Arnold and
Page 656 :
640, , -, , Further Molecular Rearrangements, , Rearrangements Proceeding through Carbanions or Related Species, One of the simplest rearrangements involving a carbanionlike intermediate is the, Stevens rearrangement,45 an early example of which is shown below:, OH', , slow, , Ph-C-CH,, , Ph —C—CH—CH2Ph, , II, , \+/, , O, , O, , /\, Me, , I, , NMe2, , Me, , (Note that the proposed intermediate, LXVIII, is not actually a carbanion, but, , instead a type of “zwitterion” in which the negative pole has carbanion char¬, , acter.) The evidence pointing to the mechanism above is relatively straight¬, forward. First, the rearrangement does not proceed readily when the benzoyl, , group is replaced by a phenyl or alkyl group, these being relatively ineffective in, , “labilizing” the hydrogens on an adjacent methylene group.44 Secondly, al¬, , though the reaction is accelerated by base, a limit is reached when slightly more, , than one equivalent is added, at which point virtually all of the substrate has, , been converted to its conjugate base, LXVIII. Thirdly, when the rearrangement, is carried out on the optically active ammonium ion, LXIX, the a-phenylethyl, group migrates with retention of configuration,45 indicating that the new C—C, , bond is formed and the old C—N bond breaks on the same side of Ca. Finally,, Me, , Me, CH—Ph, , Ph—C- -ch2, , Ph—C—CH-, , s+/, , NMe2, , O, , N, , O, , ^I, -CH—Ph (retention), , Me, , Me, LXIX, , the rearrangement is retarded by incorporation of electron-attracting substituent!, (, , cl, —NOz, etc.) into the benzoyl group,>’ for these lower the electron density, , at the negatively charged attacking carbon., 7, T ru, CVw 1Q-2R 3193'1930, 2107, 2119; 1932, 55, 1926; 1934, 279., ::, , or, , PhLi, , Ph—CH, , T •, , Me-> Ph—CHLi, , v, , Me, , ><, , Me, , PhCH—Me + Li+, , I, , NMe2, , Me, Me, Me, ■ rr, t ru,™ Snr 1947 93; Brewster and Kline, J- Am, V Campbell, Houston, and Kenyon, J. Chem. Soc., 194/,, Chem. Soc., 74, 5179 (1952).
Page 657 :
Rearrangements through Carbanions or Related Species, , -, , 641, , Similar rearrangements have been observed with sulfonium salts, and, under, sufficiently drastic conditions, with benzyl or allyl ethers., OH, CH,Ph, , Ph-C-CH, , ^Ph-C-CH:, , Y, , 11, , \e/J, S +/, , O, , knh2, CH3, , o11, , », , K+, , PhCH:, , \ /, , CH,, , \ /), , V, , O, , -> PhCH — CH <, , 46, , Y,, , s, , I, Me, , Me, , Me, , PhCH2, , Ph-C-CH-CH2Ph, , CH2Ph, , 47, , _o, K+, , (the Wittig rearrangement), , The Stevens and the Wittig rearrangements, which are thought to involve, migration of a group without its pair of bonding electrons to an electron-rich, carbon atom, are said by some workers to be electrophilic rearrangements, in contrast, to the far more usual nucleophilic rearrangements, in which a group migrates with, its pair of bonding electrons to an electron-deficient terminus. We shall, however,, find it more convenient to regard carbanion rearrangements as internal nucleo¬, philic substitutions (Sai reactions), with the negatively charged carbon atom the, attacking nucleophile., The Sommelet-Hauser rearrangement48 is a SNi'-type analog (internal nucleo¬, philic substitution with allylic rearrangement, p. 296) of the Stevens rearrange¬, ment. When benzyltrimethylammonium salts are treated with NaNH2 in liquid, ammonia, removal of the benzyl hydrogen, if it occurs at all, is reversible;, whereas removal of a methyl hydrogen results in an attack on the benzene ring., The initial rearrangement product of ammonium ion LXX is the exomethylene, derivative LXXI, which is readily isolable/s<c> However, when the benzene, , ivie, , LXX, , ring is not methylated, the corresponding rearrangement product, LXXII, is, * Thompson and Stevens, J. Chem. Soc., 1932, 69., Happe, Ann., 550, 260 (1942) f557^205 ^1947)3’, , (1951)‘ See also Wittig> Lohman, and, , 73- 4172 C»1)i 76, 1264, 79, 5512, 6274, 6277, 6280 (1957), , ('ComPt■ Tend-> 205, 56 (1937)), but, co-workers., , (d) This reY, , hi, , ' W HaUSCr and Van Eenam- ^id.., been most^ntf™"?, dlScovered bY Sommelet, ost Pensively investigated by Hauser and
Page 658 :
642, , •, , Further Molecular Rearrangements, , not isolated; for it may be readily converted, merely by the loss and recovery of a, proton, to an aromatic structure,, , .49, , LXXIII, , An interesting variation of, , H2C, , —H +, , ■HH, , Me,N—CH 2, , Me, NCI L, Me2NCH2, , H, LXXII, , LXXII I, , this rearrangement has been achieved, using as a substrate the l,l-dimethyl-2phenylpiperidinium ion, LXXV.4S(6) The resulting product has a nine-membered ring containing one nitrogen atom (LXXVI)., , LXXVI, , In the Favorskii rearrangement, an a-halo ketone is transformed to an ester,, using alkoxide, or to a carboxylate salt, using hydroxide. The mechanism which, R, , I, „, , +, , EtO-C-C —, , cr, , o, R-C-C-^, , R, , "o-c-c—, o, , +, , cr, , i, , 4f So pronounced is the tendency toward aromatization in, , this type of, , rearrangement that, , even amine LXXI forms an aromatic system (LXXIV) on being heated to, CH2CH2NMe2, , CH:, CH2NMe2, Me, , .Me, , heat, "Me, , Me, LXXI, The path by which this isomerization proceeds is, at present, unknown.
Page 659 :
Rearrangements through Carbanions or Related Species, , 643, , first comes to mind for such a rearrangement is very similar to that proposed, for the benzilic acid and benzilic ester rearrangements; the basic anion attacks, the carbonyl group, “pushing” substituent R— to the adjacent carbon where, it, in turn, displaces halide. Yet this mechanism cannot be correct for the, , R, , I, , EtO', , ,, , 1/, , + c-c ^, II, , I, , O, , Cl, , ©v, EtO-C-C—, , R, I, -> EtO-C-C -, , Co: ibV, , + cr, , o, , rearrangement of 2-chlorocyclohexanone, as has been demonstrated by carrying, out the reaction on a sample of this ketone labeled with C14 at the chlorinated, carbon, LXXVII.50 As shown, if the reaction proceeds by a path analogous to, the benzilic-acid rearrangement, all of the labeled carbon should be present at, the a position of the resulting ester, LXXVIII:, , LXXVII, , LXXVIII, , Instead, just half of the labeled carbon is actually found at Ca in the ester, with, the other half at C0. This suggests the intervention of an intermediate in which, Ca and C^ are equivalent, specifically the substituted cyclopropanone, LXXX,, formed by internal substitution within carbanion LXXIX. The “cyclopro-, , panone mechanism” for the Favorskii rearrangement is obviously excluded for, tones such as 1-benzoylcyclohexyl chloride, in which the nonhalogenated, -carbon bears no hydrogen atom. Nevertheless, this ketone is found to re¬, arrange with ease to ester LXXXI.« It thus appears that there are a. least two, 60, 61, , Loftfield, J. Am. Chem. Soc., 72, 632 (1950), Stevens and Farkas, J. Am. Chem. Soc., 74, 5352 (1952).
Page 660 :
644, , Further Molecular Rearrangements, , distinct paths by which the Favorskii rearrangement may proceed, and that a, mechanism analogous to the benzilic-acid rearrangement (sequence 1), may, operate when the “cyclopropanone” mechanism may not. We would also expect, some ketones to rearrange by a combination of the two mechanisms, but this, has not yet been confirmed.BS, , The Claisen Rearrangement and Related Reactions, In the Claisen rearrangement, the allyl or substituted-allyl group of an aryl allyl, ether migrates to the ortho position of the benzene ring; if there are substituents, at both ortho positions, the migration is to the para position. In an earlier chapter, (p. 143) we noted that experiments with phenyl-allyl ether labeled at C7 (LXXV), showed that the allyl group becomes attached to the ortho position in the ring at, C7 rather than Ca (that is, that the ends of the allyl system become interchanged, during the migration) .55(°) I n a similar way, it has been shown, despite some early, confusion on the question,55(6) that such an interchange does not occur in migra¬, tion to the para position.5S(c), 11 Another carbanion rearrangement thought to proceed through a three-membered ring, intermediate is the basic rearrangement of the tosylates of certain ketoximes to a-amino ketones, (Neber et at., Ann., 492, 281 (1932); 515, 283 (1935)). The proposed intermediate is an azinne, (a three-membered cyclic imine, LXXXIII). In at least one case—the rearrangement of, oxime LXXXII—this intermediate has been isolated and its structure verified by reduction to, the substituted ethyleneimine, LXXXIV (Cram and Hatch, J. Am. Chem. Soc., 75, 33, 38, (1953))., ArCH-CHMe, , I, , V, no2, , ArCH: ^, , Me, , I, °2N, , GH2, , C=N, , b:, , \ X, C=N, , OTs, , LXXXII, , N, an azirine, , Me, , { OTs, , H, , ArCH—CMe, , LXXXIII, , LXXXIV, , <«o, , 'ArCH-C-CHs, , I, , II, , nh2 o, , . Acta, 35, 1879 (1952); 36, 489 (1953). (b) Mumm,, (a) Schmid and Schmid, Helv. Chim,, (1939). (c) Rhoades, Raulins, and Reynolds, J. Am., Hornhardt, and Diedrichsen, Ber., 72, 100, , Chem. Soc., 75, 2531 (1953).
Page 661 :
The Claisen Rearrangement and Related Reactions, , 645, , (reversal of the allyl group), , (no allyl reversal), , All attempts to detect fragments (ions or radicals) as intermediates in the, Claisen rearrangement have failed. When the rearrangement is carried out in a, nucleophilic solvent (for example, phenol or dimethylaniline), no product re¬, sulting from attack by the allyl ion on the solvent is formed. Nor is the reaction, generally retarded by addition of free-radical reaction inhibitors. When two, ethers having different aryl groups and differently substituted allyl groups are, allowed to rearrange in the same solution, no “cross products” are obtained.54, All this, together with the observed first-order rate law for the rearrangements, and the negative entropies of activation (ranging from —2 to —16 kcaj per, degree, depending upon substituents),55 suggests a cyclic mechanism in which, H Typically, rearrangement of a mixture of ethers LXXXVI and LXXXVII gives a, mixture of phenols LXXXVIII and LXXXIX, but no detectable quantities of phenols XC, or XCI (Hurd and Schmerling, J. Am. Chem. Soc., 59, 107 (1937))., , C, , /0-CH2CH=CHrh, J, , LXXXVII, , CHPhCH=CH2, CH2CH=CH2, , OH, no, , no, XCI, , in* IdlaSot, , ^ (TO9); 62> 1728 0»«). (4) Goer.
Page 662 :
646, , -, , Further Molecular Rearrangements, , the new C—C bond forms while the old C—O bond breaks; thus, for ortho, migration,, , Here we have represented the electrons as moving in a “counterclockwise”, direction (as they would move if the reaction were the attack of the allyl cation, on the ortho position of the phenoxide anion). Yet, in a concerted cyclic mecha¬, nism of this sort it is often difficult to determine whether the direction of electron, flow is that indicated, or the opposite, or, for that matter, whether the electrons, move in pairs or become unpaired during the progress of the reaction.56, It is interesting that the Claisen rearrangement of optically active ether, XCIII (in which C« is asymmetric) leads to an optically active phenol, XCIV,, in which a new carbon has become asymmetric.57 This is a type of asymmetric, , OH, , — CH—CH=CHMe, , 'CHMe—CH=CHMe, *, , XCIV, , XCIII, , induction; it probably results because the transition state XCV, in which the, methyl groups are trans, is sterically favored over transition state XCVI, in, which they are cis., , XCV, .. For recen, evidence .ha, suggest, ,ha, the direcdon of electron, rearrangement is oppositeto that shown in> st™c u1, >, the’ claisen rearrangement,, and Fife, J. Am. Chem. Soc., 80, 3271 (1958). Reactions siicn as, as, nQ decisive, the Diels-Alder reaction, and the Chugaev reaction, experimental evidence for ions or radica s as m erme i, , classified by some workers as, ^ und, oi, COvalency change at, , dme!ysee,tr /(ample, Hine, Physical 0,'an* Ch'.is,ry, McGraw-, , Hi“ ^a?„dern/„dN^u,Yb0err"71d“’., , 73, 4305 (1951).
Page 663 :
The Claisen Rearrangement and Related Reactions, , 647, , It is now clear that the “para Claisen rearrangement,” which occurs when, both ortho positions are substituted, takes place in two stages—the allyl group, being “reversed” in the first stage, then “unreversed” in the second. The initial, stage is closely similar to that in the “ortho Claisen” rearrangement, but the, intermediate dienone, XCVII, is considerably more stable, since it cannot un¬, dergo aromatization merely by tautomerization. This intermediate may be, , R, , XCVII, , trapped as a Diels-Alder adduct by carrying out the reaction in the presence of, maleic anhydride.55 Adduct C, when heated, gives the product (Cl) resulting, from “para rearrangement.” Moreover, dienone XCIX has been prepared, OCH2CH=CH2, , Me^, , ^'Me, , Me, ch2ch=chs, , XCVIII, , o, o, , XCIX, , o, , Me, , :o, CH2CH=CHj, , to, , Cl, ndependently and has be'n found, on heating, to isomerize to a mixture of, Phenol Cl and allyl ether XCVIII. Thus, the initial stage in the para Claisen, rearrangement (XCVIII _ XCIX) is, as indicated above, reversible », , ." SrrddcF:rrd'; *, , point, see Curtin and Johnson,, , ibid., 76,, , t- %, , 2276’(1954), , (, , 2™r6)-, , ^ 1 F°r further evidence on this
Page 664 :
648, , Further Molecular Rearrangements, The rearrangement of ethers such as CI1 and Cl 11 is, at first glance,, , formally analogous to the Claisen rearrangement. Mechanistically, however,, , there is little similarity, for there is strong evidence that the rearrangement of, such nonallylic ethers involves the breakoff and migration of an alkyl cation., , Such rearrangements, although they sometimes occur merely on heating, are, often catalyzed by protonic or Lewis acids, and may yield significant amounts, of “fragments” derived from the alkyl group (thus the rearrangements of ethers, Cl I and Cl 11 yield styrene as a by-product). Moreover, when the rearrange¬, ment of the ether derived from one phenol is carried out in the presence of a, second phenol, alkylation of the second phenol (formation of a, , cross-product ), , may be observed., We would accordingly be quite justified in assuming the simple mechanism, given above for such rearrangements were it not for one difficulty; the rearrangement of an optically active ether (with an asymmetric a-carbon atom) often, yields an isomeric phenol in which the configuration of the migrating carbon is, partially retained,60 although cross-products are racemic. It thus may be concluded, that these rearrangements proceed, at least in part, through “intimate” ton, pairs (p. 583), in which the environment of the carbonium ion is not symmetric., These ion pairs may “collapse,” yielding an optically active product, or they, may dissociate, yielding a symmetric carbonium ion in a symmetric environ¬, ment which, in any subsequent reaction, forms an optically inactive produc ., Typically, the rearrangements of ethers CII and CIII are about 20 percent, stereospecific.e0(6), On the other hand, a number of reactions are known, , ,, which, , appear to, , more authentic analogs of the Claisen rearrangement. Three of these are lisle ., » („) Sprung and Wallis, J. Am. Chen,. Soc, 56, 1715 (1934). <» Har, and Eleu.eno,, ibid., 76, 519 (1954).
Page 665 :
649, , The Claisen Rearrangement and Related Reactions, CH2, , V, , OJ, , CH2, O, , CH-Me, , CH, , 61(a), , Ph-Cx^, , AN—Ph, , Ph-C, , CHMe, , CIV, , 'N', , /, , Ph, , CH2, , NC., , A*, , C, , 61(b), , CH, , (A S* = -12 cal//®), Me—C^, , ^CHMe, CH2, , 61(c), , (AS* = -8 cal//®), , CVI, Note that the migrating allylic groups in CIV and CV are “reversed” in the, rearrangements, this presumably also being the case with CVI (although we, cannot be sure). The rearrangements of CV and CVI display, as they should,, first-order rate laws and sizably negative entropies of activation consistent with, the cyclic mechanisms indicated. A similar cyclic rearrangement has been pro¬, posed as a step in the oxidation of phenols with benzoyl peroxide.6* This reac¬, tion, which generally yields o-benzoyloxy phenols, displays none of the char¬, acteristics of a free-radical reaction. The following path is suggested:, , OH, , o, It, , —, , -PhCOO", , -C-Ph, II, , o, , CVII, , (Presumed intermediates CVII and CVIII are enclosed in parentheses since, there is as yet no direct evidence for them.), (a) Mumm and Moller, , Ber, , 70, , 79Id fiooT\, , (1940); 63, 1843, 1852 (1941); 69 1893 (19471 (, Schuler and Murphy, ibid., 72, 3155 (1950)., ‘, , /l\, , u, 90p*’ tf, J‘ Am’ Chem' Soc-> 62, 441, Hurd ^ Pollack> ibid-> 60, 1905 (1938);, , Walling and Hodgdon, J. Am. Chem. Soc., 80, 228 (1958).
Page 666 :
650, , •, , Further Molecular Rearrangements, , The Rearrangements of N-Haloanilides and Related, Aromatic “Rearrangements”, When N-chloroacetanilide or a ring-substituted derivative of this chloroamide is, treated with HC1 in a hydroxylic solvent, the chlorine appears to “migrate”, from the nitrogen atom to the ortho or the para positions (or both) in the ring., , This reaction is sometimes called the Orton re arrangement.6 3 The specific action, of HC1 in this reaction immediately suggests the possible intervention of chlo¬, rine; and indeed, if the reaction is carried out while air is bubbled rapidly, through the solution, elemental chlorine is found to be carried away, leaving, acetanilide behind. However, if no effort is made to remove chlorine from the, solution, it chlorinates the benzene ring of acetanilide, a reaction that may inde¬, pendently be shown to be much faster than the rearrangement of the N-chloroanilide.^ Moreover, the ortho to para ratio in the mixture of ring-chlorinated, anilides resulting from direct chlorination is the same as that in the mixture, obtained in the rearrangement (provided that the solvent is the same).63 The, Orton rearrangement may therefore be assumed to proceed by the following, path:, Cl, , Cl, , I, , I, Ph-N-Ac, , +, , Ph-NH-Ac, , H+, , Ph, , cr, , I, +, , Cl—NH -Ac, , slow, , fast, , >, , Cl-Cl, , +, , PhNHAc, Cl, , Cl, , $, , \—NHAc, , +, , —NHAc, , This rearrangement, then, unlike the rearrangements we have thus far con¬, sidered (except perhaps the rearrangement of alkyl aryl ethers), is mtermolecular. The “migrating group,” the chlorine, becomes completely detached from, the remainder of the molecule for a significant interval of time during, , P, , %, , ress of the reaction, and there is no guarantee that the ch orme returmng to a, , „, , given molecule is that chlorine which has departed from it. In a sense, t e, .. Orton,, , d„ J. Chem. So,.., , 95,, , 1456 (1909);, , Soper, J. Phys. Chem., 31, 1192 (1927)., , 99,1185, , (1911);, , 1927,986; 1928,, , 782, 99 ,
Page 667 :
The Rearrangements of N-Haloanilides, , *, , 651, , rearrangement is not a rearrangement at all, but simply a heterolysis of the, N_Cl bond, followed by a (completely independent) electrophilic attack on the, benzene ring., #, Other observations are consistent with the mechanism shown. It an JN-chioroanilide with a “deactivated” ring, such as N-chloro-2,4-dichloroacetanilide,, (CIX), is treated with HC1 in the presence of acetanilide, the “labile” chlorine, does not migrate to the ring of its parent molecule, but instead “crosses over”, into the acetanilide ring, emphasizing the intermolecular character of the reac¬, , tion. Furthermore, if the rearrangement is carried out with chlorine-labeled, N-chloroacetanilide, only a very small fraction of the radioactive chlorine is, found bound to the benzene ring in the product.65 Finally, the observed rate, law also fits into the picture, for the reaction has been found to be of the third, order.66, rate = £3(ArNAcCl)(HCl)2, , (2), , Assuming complete ionization of HC1, rate law (2) may be rewritten as follows:, H, rate = *3(ArNAcCl)(H+)(Cl-) = k', , ArNAcCl, , (Cl"), , (3), , This is obviously consistent with the proposed mechanism if the attack by, chloride on the conjugate acid of the substrate is the rate-determining step., Various modifications of the Orton rearrangement are known. The N_Cl, bond in N-chloroanilides may be broken also by HBr and HI, but the resulting, products are bromo and iodo compounds." The intermediates in such cases are, 76 n938),0w/n, Soc- 5S- 2212467 <1936); 59, 1613 (1937); J. Org.Chm, /0 u 738). Since the two chlorine atoms n the intermediate (the M, ' u, 1 \, &, , 3, , we might expect just Act, of the labeled chlorine to bT'ot, H^eve ’ exca^c’bcTwXrf "d', Cl, tn solution is known to be extremely rapid, due to the rapid formation, , *C1—ci + cr ^ [*ci—ci—ci](a) Harned and Seitz, J. Am. Chem. Soc, , *ci~ + ci—ci, , 44 147S <'10771, , J927> 2761. fr) Dawson and Millet, ibid1932 1920, 61 Richardson and Soper, J. Chem. Soc., 1929, 1873., , <k\ c, *, , , _, ^, , yde’ J' CW
Page 668 :
652, , -, , Further Molecular Rearrangements, , Br, +, (<* I'), , Cl-NH-Ar, , -> BrCl +, (or IG1), , NHAr ->, , [c, , <f~\-NHAc, Br/W, , +, , HQ, , (i), presumably BrCl and IC1 (rather than Cl3). As with direct halogenation (p., 443), the more electropositive halogen of the interhalogen intermediate attacks, the benzene ring. The rearrangement may also occur, although at a much di¬, minished rate, in aqueous solutions of strong “oxygen acids” (HNO3, H2S04,, and HC104).55(6)*es Here, it seems likely that the, , N—Cl bond is broken in, , attack by a water molecule, forming the H2OCl+ ion as an intermediate., H, , Ac, , Ac, ., , 4", , O—Cl + NH—Ph, , H20 + Cl—NH—Ph, +, H, , j- and /?-AcNH—C6H4—Cl + H2Q, It has even been found that the rearrangement of N-chloroanilides is promoted, by light,55 suggesting that the Orton rearrangement, like the Cannizzaro reac¬, tion, the Wolff rearrangement, and the decompositions of diacyl peroxides,, may proceed by more than one class of mechanism., There are a number of additional instances known in which a group at¬, tached to the nitrogen atom of an N-substituted aniline appears to migrate to the, ortho or para position of the aromatic ring. Three of the more familiar of such, reactions are shown:, Me, , H, , 0=N|-N—Ph + H30+—^0=N-0X, , ,—., + MeNHPh-> MeNH—^, , 70(a), , Vn=0, , H, , o, II, , Ph_N=N>—NHPh + RCOOH—> Ph—N=N+, , O, , G, , R + NH2Ph, , H2N, , n-Buj— NH2— Ph, , Cl, , n-BuCl T NH2Ph, , I, , ** Similarly, , N-haloacetanilides have been found to undergo rearrangement in ehloro-, , benzene in the presence of strong carboxylic, , 2676;:, , 55°- 192
Page 669 :
The Rearrangements of N-Haloanilides, , -, , 653, , As indicated, each of these reactions is, like the Orton rearrangement, intermolecular. In each case, the migrating group may be intercepted by carrying out, the reaction in the presence of an additional aromatic amine with a more, “active” ring than that in substrate. In the third of these reactions (the so-called, Hofmann-Martius rearrangement) the alkyl-halide intermediate is probably con¬, verted to a carbonium ion before it attacks the ring, since rearrangements in the, migrating alkyl group are often observed., The mechanisms here suggested for the rearrangement of N-substituted, aniline derivatives are at variance with the views held by a number of workers71, (principally in the late 1940’s), who felt that some or all of these rearrangements, proceeded through a, , r complex in which the migrating cation could move, , 7, , rather freely over the 7r-electron lobes associated with the aromatic ring before, becoming bound to the ring at the ortho or para position. While it is possible that, the system passes through such a complex before the departure of the migrating, fragment or before the formation of the new bond (or before both), the un¬, mistakably intermolecular character of these rearrangements would seem to, rule out the possibility that such a complex is the sole intermediate between these, steps. Moreover, now that the extreme lability of the t complex is recognized, (p. 119), it seems very likely that any such complex, if formed, would be in, mobile equilibrium with the fragments here proposed as intermediates., When a phenyl ester is treated with anhydrous aluminum chloride, the acyl, group migrates to the benzene ring, principally to the ortho position (when this, position is available). This reaction, the useful Fries rearrangement, is likewise, intermolecular, at least in part. If, for example, the rearrangement is carried out, on a mixture of esters CX and CXI, four o-hydroxy ketones may be easily, isolated from the reaction mixture.7* The observed transfer of acyl groups from, , OBz, , OH, , Me, , Me, , OH, , OH, , CXI, , ring to another suggests the following mechanism:, , ^n7'103 (,92o>;, Press) Ojrford^, , i*., ^, , Rosenmund and Schnurr, Ann., 460, 96 (1927)., , wSn&i, , CWr** Oxford Univerei.y
Page 670 :
654, , Further Molecular Rearrangements, , OA1C13, , +, , R—C=0, , *-'C—R, CXI 11, , II, O, , /Aid,, , O : AlCIs, , CXII, , o' \, , o, , ( J, , c=o, , C-R, , —HCl, ->, , 1, R, CXIV, , As indicated, the function of the aluminum chloride is the withdrawal of electron, density from the acyl-oxygen bond by coordination (CXII), which facilitates, the initial heterolysis. The acyl cation thus formed may attack the benzene ring, of anion CXIII. We have assumed that, in the absence of water, the orthomigration product is a chelated complex such as CXIV, from which the hydroxy, ketone itself may be formed upon hydrolysis. Indeed, it is probable that chela¬, tion with tripositive aluminum significantly stabilizes the orMo-migration prod¬, uct; for the para-migration product (which obviously cannot form such a, chelate) is known to be formed reversibly, whereas ortho migration is essentially, irreversible. This effect accounts for the predominance of ortho migration often, (but not always) observed in the Fries rearrangement., But there is evidence that the mechanism above is somewhat oversimplified., Suppose o-cresyl acetate is rearranged in the presence of 2-hydroxybiphenyl. The, predominant, , product, , is, , 3-methyl-4-hydroxyacetophenone, , (CXV),, , accom¬, , panied, as we might expect, by significant quantities of the acetylated biphenyl,, CXVI. What is surprising, however, is that if the quantities of each of the, , reagents are kept constant but the volume of solvent used is allowed to increase,, the ratio of ketone CXVI to ketone CXV steadily decreases.” This means that, the reaction in which the “cross-product” (CXVI) is formed has a higher aver¬, age kinetic order than that in which the “main product” (CXV) is formed, an, » Baltzley and Phillips, J. Am. Chem. So,, 70, 1491 (1948)., reagent are used, but the volume of solution ra.sed from 500 to 2000 cc, th,, CXVI:CXVI drops from 0.62 to 0.28.
Page 671 :
The Rearrangements of N-Haloanilides, , -, , 655, , suggests that the acyl ion is present in part as ion pairs, RC=0 ArOAlCl^. Such, ion pairs may “collapse” with rearrangement to give CXV or its ortho isomer, (first-order reactions), may react directly with hydroxybiphenyl to give CXV1, (a second-order reaction), or may dissociate to a free acyl cation that may at¬, tack either of the two available ring systems., A final rearrangement which, in a formal sense, is similar to those we have, been considering, is that of N-nitroanilines (and related N-nitro amines) to, o-nitroanilines in strong acid. On the basis of analogy to the rearrangements of, , N-chloroanilides (p. 650) and N-nitroso-N-alkylanilines (p. 652), we might, suppose that this reaction is intermolecular also—that is, that the —NO2 group, is removed from the substrate (for example, as H2ONO^ in water) and that,, independently, nitration of the aromatic ring follows. However, a closer look, at this reaction makes it apparent that such an analogy is poorly taken. Suppose, that the nitrating species were to become completely independent of the re¬, mainder of the molecule (in this case an aromatic amine) before attacking the, ring. We should then expect the same mixture of products from the rearrange¬, ment as is obtained when the amine is treated with an outside source of the, nitrating agent under rearrangement conditions. Specifically, the same mixture, of 0-, m-, and /?-nitroanilines should result from the rearrangement of N-nitroaniline (in, say, sulfuric acid) as is obtained from the nitration of aniline itself, with HNO3 (again in sulfuric acid). This comparison has been carried out in, 85 percent H2S04 at IO0,74 under which conditions the rearrangement gives, almost exclusively o-nitroaniline, but direct nitration gives a mixture of isomeric, nitroanilines in which the o:m:p ratio is 1 :6:10. Furthermore, it has been found, that when the nitroamine CXVII undergoes rearrangement in the presence of, dirnethylamline, the ring of the latter compound (which is much more suscepti¬, ble to attack than the ring of /;-nitroaniline) is not appreciably nitrated.™, , na, , °2, , _/-nh~no2, , +, , CXVII, , (no Q2N-^~~^-NMc2), 7< (a) Hughes and Jones, J. Chem. Soc 19«>0 9A7S u, assumed to be the NO? ion (p 419) (b) For fr\a', Hartogs, and van der Linden, , 44, 704 (,9U), ^, , he mtratin& aSent may be, prccis' ”udy- ** Holleman,
Page 672 :
656, , Further Molecular Rearrangements, , Finally, it has been found that if the rearrangement of N-nitroaniline is carried, out in the presence of N-labeled nitrate, none of the labeled nitrogen is in¬, corporated into the product.75 We may thus rule out an intermolecular mecha¬, nism as playing a significant role in the nitroamine rearrangement., It is not easy to visualize a simple mechanism in which the nitro group, is transferred in a single step from the amino nitrogen to the ortho carbon, but the, following tentative stepwise mechanism is not unreasonable:, , CXVIII, The rearrangement of the proposed “nitrito” intermediate, CXVIII, is a little, like the Claisen rearrangement and, as with the latter, the indicated “direction, of electron flow” is arbitrary. The presumed initial isomerization to CXVIII, brings to mind the Stevens rearrangement (p. 640)., , H, , I, , --N,-, , Ph-C\, , Ph-Nt, , H, , /, Me, , ., , / :CH—Bz, , \, Me, , Stevens, rearrangement, , nitroamine, rearrangement, , The Benzidine Rearrangement, The most familiar of aromatic rearrangements, the benzidine rearrangement, is,, in some respects, the most baffling. The most usual course of the reaction, the, transformation of hydrazobenzenes to derivatives of/>,//-diaminobiphenyl, ten s, to predominate with ortho- and mefa-substituted hydrazobenzenes and may also, occur with certain /mm-substituted hydrazobenzenes if the para substituent is, readily subject to electrophilic displacement. More generally, however, para, substituents divert the rearrangement into one or more alternate courses., Hydrazobenzenes having para substituents on both rings are common y, u Brownstein, Bunion, and Hughes,, , para Claisen rearrangement., , O-. *, , M, , 19*b ^"gTha."^
Page 674 :
658, , -, , Further Molecular Rearrangements, , the benzidine rearrangements must account for the formation of each of these, products., There is no doubt that the benzidine rearrangement is intramolecular. We, have seen how this has been demonstrated by the nonformation of cross-over, products when a mixture of 2,2'-dimethoxy- and 2,2'-diethoxy-hydrazobenzene, is treated with acid (p. 141). It may just as convincingly be shown by the rear¬, rangement of a mixture of methyl-labeled 2-methylhydrazobenzene (CXXI), and unlabeled 2,2'-dimethylhydrazobenzene (CXXII). In this case, the o-toli-, , f, , GXXII, , _yCHs, , CXXIII, , dine (CXXIII) formed is unlabeled.76 We here assume also that the formations, of semidines and diphenylenes are likewise intramolecular, although this ap¬, parently has not yet been rigorously demonstrated., The rearrangements of hydrazobenzenes to benzidines,77(a) to diphenyl¬, enes,77(6) and to o-semidines77(c) are third-order reactions, first order in sub¬, strate and second order in hydrogen ion. The rearrangements are subject to, specific hydronium ion, rather than general acid catalysis and proceed more, rapidly in D20 than in H20.77(c) It therefore appears likely that the rate-deter¬, , mining step involves only the “double conjugate acid, , of the substrate, , that is,, , the cation Ar—NH2—NH2—Ar.7S It has also been found that />anz-deuterated, hydrazobenzene (CXXIV) undergoes the benzidine rearrangement at very, “nearly the same rate as does nondeuterated hydrazobenzene,7J suggesting that, , D-^^— NH-NH—D, CXXIV, 16 Smith Schwartz, and Wheland, J. Am. Chem. Soc., 74, 2282 (1952)., ...., " W Ham-nd a^d Shine, 7., Chem. Soc 72 220 (1950); Croce and Gettler, iW., 75 874 (1953); Blackadder and Hinshelwood, J. Chem. Soc., \957, -%, • (, *, ’, Odioso, J. Am. Chem. Soc., 73, 1002 (1951). If) Bunton, Ingold, and Mhala, J. Chem. See., 1957,, ,906n Agreement is not unanimous on this point. For a differing view, see Cohen and Hammond, J. Am. Chem. Soc., 75, 880 (1953)., 2444 (1955), 7« Hammond and Grundemeier, J. Am. Chem. Soc., 77,, t, 1
Page 675 :
The Benzidine Rearrangement, , -, , 659, , the /.era-hydrogens become detached from the ring carbons after, rather than, during, the rate-determining step (p. 193)., The available data on the benzidine rearrangement then point to the fol¬, lowing cyclic mechanism:, , cxxv, , CXXVI, , Here it may be assumed that the easy breakage of the N—N bond in CXXV is, due largely to electrostatic repulsion between the two positively charged nitrogen, atoms. When there are substituents at the para positions of the substrate, inter¬, mediate CXXVII (corresponding to CXXVI in the “ordinary” benzidine re¬, arrangement) cannot aromatize simply by loss of two protons. We would then, expect CXXVII to survive long enough to undergo a second rearrangement,, yielding the customarily observed orf/zo-semidine., , The formations of /wa-semidines and diphenylenes probably also occur in, X°YS“m -the inhial Stage Presumab'y bei"S the rearrangement to cation, LXXVIII, in which the rings are joined ortho to the amino groups., There are two important questions in connection with the benzidine rear¬, rangement to which satisfactory answers have not yet been obtained. First it, may be asked what holds the benzene rings so close to each other in the tran’sinon State of the slow step. We have already noted that, due to the ir-electron, obes associated W‘th the benzene ring, the normal “half thickness” of the ring is, sho H ,f t (P r )' ^ tHe abSenCC °f SPCdal ^ts, considerable difficulty, should therefore be encountered in moving two parallel (or nearly parallel!
Page 676 :
660, , -, , Further Molecular Rearrangements, , benzene rings to within less than about 3.7 A of each other. Indeed, the benzene, rings in di-p-xylylene (CXXIX) have been found to be “buckled,” due presum¬, ably to interference between the 7r-electron clouds of the two rings.50 It has been, , CXXIX, suggested that the transition state in the benzidine rearrangement (and in, "related arrangements) is subject to exceptional stabilization by resonance.51 but, the nature of this “extra resonance” is not clear. The transition states in these, rearrangements are not even approximately planar, hence do not fulfill what, is ordinarily a primary requirement for extensive delocalization of 7r-electron, density over a conjugated system. The fact that a rather large number of hyperconjugated structures can be drawn for each of the various transition states is,, in the opinion of the present author, not relevant., The second question concerns the manner in which the substituent m, singly ^-substituted hydrazobenzenes directs the predominant course of the, reaction. Why, for example, does a p-C\ substituent favor rearrangement to a, 10, tt, , Brown, J. Chem. Soc., 1953, 3265, 3278., Hammick and Mason,, Hughes and Ingold, J. Chem. Soc., 1941, 608;, , ibid,, , 1946, 638.
Page 677 :
661, , Exercises for Chapter 15, diphenylene, a, , p-Me, , group favor rearrangement to an or^o-semidine, and a, , ^-NH2 group favor formation of a /wra-semidine?, Until more light is shed on these two (possibly related) questions, we cannot,, in good conscience, claim that we understand the benzidine rearrangement., , EXERCISES FOR CHAPTER, , 15, , 1. (a) Outline an experiment, starting with available materials, to show that in the, Wolff rearrangement (p. 627) the carbonyl carbon of the diazomethyl ketone, becomes the carboxyl carbon of the resulting acid., (b) Oudine a series of transformations to demonstrate that the migrating group in the, Wolff rearrangement shifts with retention of configuration. (Assume that the, relative configurations of reactants and products cannot be determined merely, by comparing their directions of optical rotation.), 2. Which member in each of the following pairs will undergo the indicated rearrange¬, ment more readily? Justify your guess in each case:, (a) Ph2C =N—OAc or Ph2C=N—OTs (Beckmann)?, , (Beckmann) ?, (c) Ph2C N OH, 2M H2S04; in EtOH or in H20 (Beckmann)?, (d) AcNH2 or AcNHMe (Hofmann)?, (e) PhCONH2 or PhCH2CONH2 (Hofmann)?, , (f), , or, , (Orton), , Me-(U\-Me, , AcN—Cl, , ?, , AcN-Cl, , (g) />-Nitrobenzamide or />-chlorobenzamide (Hofmann)?, , (h) MeO-^y-CO-NH-O-C-^^-QMe, , or, , O, MeO, , \, , ^, , CO-NH-O-C-Q (Lossen) ?, , °, (i) Cyclobutanone or cyclopentanone (Schmidt)?, , OMe
Page 688 :
CHAPTER, , 16, , Free-radical Reactions', , For our present purposes, we may define free radicals simply as species hav¬, ing one or more unpaired electrons.2 Familiar examples are the methyl radical, CH3*,, , the triphenylmethyl radical, Ph3C*, the chlorine atom, Ch, and the nitric oxide, molecule, :N.\O:, each of which has a single unpaired electron. Homolytic, ••, , ••, , reactions—that is, reactions passing through free radical intermediates—have, , been mentioned at various points in the preceding chapters; but because such, reactions are fundamentally different from the large body of heterolytic reactions,, a detailed consideration of them has been postponed to this point., , Part | — LONG-LIVED AND SHORT-LIVED FREE, RADICALS. FORMATION AND DETECTION, OF FREE RADICALS, , Triarylmethyl Radicals, The chemistry of organic free radicals may be said to date back to 1900, for it, was in this year that Gomberg reported the synthesis and identification of the, / For more detailed treatments of this topic, see: (a) Steacie, Atomic and Free Radical Reac¬, tions, , (2d Ed), , Reinhold Publishing Corp., New York,, , 1954;, , (b) Walling, Free Radicals in, , ^Solution, JohnWiley and Sons, Inc., New York, 1957; (0 LefHer,, Organic Chemistry, Interscience Publishers, Inc., New Y°rk 19f56’^T74’ 234^51 ’ W W3t, The Chemistry of Free Radicals, Oxford University Press, Oxford 1946, 7 We do not generally refer to ions of the transition metals or the rare earth metds for, pvamnk, vk?So, , Cr+3, , chemically, , 672, , Yb+3) or to complexes derived from these ions (for example, Cr(N *)• », ’free radicals although such species may have one or more unpaired dor/dec-, , from the organic free radicals Co be cons.dered ,n .ha chapter.
Page 689 :
Triarylmethyl Radicals, , 673, , -, , firs, organic free radical, triphenylmethyl.’ When triphenylmethyl chloride,, PhjCCl in benzene was treated with finely divided silver metal, there was ob¬, tained a yellow solution, which upon careful evaporat.on in the absence of air, gave the expected product, hexaphenylethane, a white solid. To what then could, the yellow color be attributed? This color disappeared slowly when the solution, was exposed to air, but in this case, evaporation yielded triphenylmethyl, peroxide, Ph3C—O—O—CPh3. Moreover, the color could be discharged by, treatment with iodine or with nitric oxide, resulting in the formation, respec¬, tively,, , of triphenylmethyl, , iodide or, , triphenylmethylnitrosomethane, , (Ph3-, , q_]\j=0). In short, the reactions of the yellow solution were those to be, , expected of the triphenylmethyl radical, although the molecular weight of the, white solid indicated that it was largely, if not wholly, hexaphenylethane. The, conclusion (much less obvious in 1900 than today) was that an equilibrium be¬, tween hexaphenylethane and the triphenylmethyl radical existed in solution,, and that this equilibrium was displaced when iodine, oxygen, or nitric acid, destroyed the radical., , Ph3C—O—O—CPh3, Ag, , Ph3C—Cl —► Ph3C—CPh3 ^ 2Ph3C-, , ii, , -Ph3C—I, , -Ao, , Th3C—N=0, Since 1900 a large number of hexaarylethanes have been prepared, and, most have been found to be more dissociated (under similar conditions) than, hexaphenylethane itself. The accurate determination of the various dissociation, constants is, however, somewhat of a problem. In principle, the degree of dis¬, sociation of such hydrocarbons could be evaluated from cryoscopic measure¬, ments of their apparent molecular weights, but any worker who has carried out, measurements of this type is well aware of the very approximate nature of the, results. Somewhat more success is possible using colorimetric methods^ (for the, radicals are colored whereas the hexaarylethanes are generally colorless), but, care must be taken to prevent interference by colored decomposition products.4(6), A third method, which is of considerable fundamental importance, makes, use of the magnetic properties of free radicals. Ordinary organic compounds are, diamagnetic; when they are placed in a magnetic field, tiny currents are set up in, , the filled orbitals, and these currents are associated with induced magnetic, fields opposed in direction to the applied field. As a result, the sample of the, * W Gomberg, , yimf, , Ber., 33, 3150 (1900); J. Am. Chem. Sac., 22, 757 (1900). (b) For reviews, , rC!fed, , John wn|P, y ind, radicals> see Wheland, Advanced Organic Chemistry (2d Ed ), John Wdey and Sons Inc, New York, 1949, p. 680; and Bachmann in Gilman’s Organic, Chemistry (2d Ed.), Vol. 1, John Wiley and Sons, Inc., New York, 1943, p 593, , cJ, , SMM?-“> s“>for “->*■*»•
Page 690 :
674, , Free-radical Reactions, , compound tends to be pushed out of the magnetic field. Ordinarily, the dia¬, magnetism of a molecule may be estimated from its structure simply by adding, together terms for each of the atoms, then making appropriate corrections for, conjugation, in much the same way as we estimate molar refractivities.5 How¬, ever, if the compound has one or more unpaired electrons, these act as small, electromagnets and tend to pull the sample into the applied magnetic field. In this, case, the material is said to be paramagnetic: the paired electrons continue to exert, their diamagnetic push, but for any individual radical this is overshadowed by, the paramagnetic pull. Quite obviously, then, paramagnetism in an organic, material is an excellent indication of the presence of relatively stable free, radicals, and by measuring the paramagnetism of solutions of hexaarylethanes,, it should be possible to estimate their degree of dissociation into triarylmethyl, radicals. We need not be concerned here with the types of apparatus used to, measure the susceptibility of materials to applied magnetic fields, nor with, the details of magnetochemical calculations.6 One difficulty should, however, be, apparent. The usual methods measure a force that results from the combination, of diamagnetic and paramagnetic effects; to obtain paramagnetic susceptibility,, corrections must be made for diamagnetic susceptibilities—not only of the nondissociated ethanes, but of the radicals themselves. When the degree of dissocia¬, tion is small, the paramagnetic term is small, and any error made in estimating, the diamagnetic contribution of the hexaarylethane will seriously affect the, estimated paramagnetic term and, hence, the estimated radical concentration., When the degree of dissociation is large, it is the diamagnetic contribution of the, radical that causes trouble, for there is good reason to suppose that the custom¬, ary method of estimating diamagnetic susceptibilities cannot be applied to, radicals of this type because of interaction (of as yet unknown magnitude) be¬, tween the benzene rings.7 Because of the uncertainties in interpreting magnetic, measurements, the dissociation constants of substituted hexaphenylethanes,, determined magnetically, are given in Table 16-1 only to one significant figure,, and even this figure should be regarded as tentative., In attempting to relate these dissociation constants to the structures of, the various hydrocarbons and to the radicals derived from them, we must first, decide why hexaarylethanes are dissociated at all at ordinary temperatures, whereas ethane itself and hexaatty/ethanes are not. Recalling the unusual stabil¬, , ity of triaryl carbanions, we may be quite certain that one important factor, . See, for -amp, Bh=ar a„d MatHur, S {Edited by Weissbergerh'vol. I, Indolence Publish^, Inc, New York, .*»., Cha^Fo>; detailed discussions of these points, see^Selwood, science Publishers Inc., New York, 1956, pp. 1 35, 135, 7 Selwood and Dobres, J. Am. Chem. Soc., 72, 380U, p. 695., , Magne,Ms,ry (2d Ed.), Inter■, , Wheland (Ref. 3b),
Page 691 :
Triarylmethyl Radicals, , 675, , Table 16-1. Dissociation Constants of Some Substituted, Hexaphenylethanes (benzene, 25°)s, , stabilizing triarylmethyl radicals is the capacity of the three benzene rings to, delocalize the unpaired electron. This is a conjugation or resonance effect which,, for the triphenylmethyl radical, may be represented pictorially by drawing, contributing structures I', I", and many like them. Substituents at the ortho, , etc., , and para positions may further delocalize the unpaired electron by electron, donation (II <-> II'), by electron withdrawal (III <-> III'), or by hyperconjugadon (IV <-> IV ). The radical derived from hexa-/>-biphenylylethane (radical V), is particularly stable, since the unpaired electron is “distributed over” nineteen, positions (three positions on each of the six rings, in addition to the a-carbon)., , 61,, , 77 (1939);, , 63,, , 1892* 0941^64* 1824 0942)- wf «Trt944? ^W°rkerS’, , ., , values, see Bowden, J Chem. Soc., 1957, 4235., , ’, , ’ 415 ^1944)- For, , J‘ Am' Chem' Soc->, , recent cryoscopic
Page 692 :
676, , Free-radical Reactions, , The fact that ortho substituents are generally more effective than para substit¬, uents in promoting dissociation indicates that steric assistance is also coming, into play. Using scale models, it may be shown that hexaphenylethane is a very, “crowded” molecule, due largely to interference between the ortho hydrogens on, the various benzene rings. When the molecule dissociates, the bond angles about, the aliphatic carbon atoms increase from 108° to about 120°, allowing consider¬, ably greater freedom of motion.5 When there are ortho substituents on the rings,, the hexaarylethane becomes even more crowded. Hence, the release of steric, strain accompanying dissociation into radicals becomes even more welcome, (although it is probable that even in the triphenylmethyl radical, steric interac¬, tion among the or^/zo-hydrogen atoms is still sufficient to keep the phenyl groups, from lying in the common plane/0 as would be desirable for maximum resonance, stabilization). It is interesting that hydrocarbon VI, in which two of the phenyl, groups on each nonaromatic carbon are “tied back ’ in a fluorene ring, is not, noticeably dissociated into free radicals except at high temperatures." Since the, , (at room temperature), , • Further evidence that the breakage of the C-C bond in hexaphenylethane » «eric^, assisted comes from the comparison of the following heats of hydrogenolysts (Bent and Culbe, son, J. Am. Chem. Soc., 58, 170 (1936))., , Ph£/chI, , + H, , V, , 2CH? + 13 kca^1, , (calculaMdltom heats of combustion), , Much Ire hea, is released in, , nt“ri?,:h.'h' 'Sar^ayn0be concluded therefore that the C-C bond in the, hexaphenyl compound is weakened by excessive crowdmg, « See, however, Karagoums, Helv. Chim. Acta 34, n Bent and Cline, J. Am. Chem. Soc., 58, 1624 (1936).
Page 693 :
Further Types of Stable Free Radicals, , -, , 677, , radical derived from this hydrocarbon should be no less “resonance stabilized, than a triphenylmethyl radical, we may conclude that dissociation does not, occur in this case because of inadequate steric assistance., , Further Types of Stable Free Radicals, We have seen that the two features favoring the dissociation of a hydrocarbon, into radicals are: (1) excessive crowding in the hydrocarbon itself, and (2) delocalization of the unpaired electron over a large area in the radical. The first of, these features is mainly responsible for the observed dissociation of hydrocarbons, VII;*(a) and VIII,whereas the second factor accounts for the stabilities of, radicals IX,^(c) X/*(d) and XI.**(e) With the help of pencil and paper (or per¬, , haps merely by inspection), the reader may see that the unpaired electron may, distribute itself over fourteen carbons in radicals IX and X, over no less than, twenty carbons in radical XI, but over only seven carbons in the radicals formed, by dissociation of VII and VIII., Not all atoms comprising the conjugated system within a stable free radical, “f., , “rbom' Careful oxidations of triphenylamine, tetraphenylhydrazine, , vv,/(», u Pe"'aPheny‘Pyrrole (XIII) yield aminium ion radicals XIV »», XV,«<» and XVI "<•> respectively:, , J. Am. Chem. Soc., 79, 4439 (1957), „, , (e) Muller -inH M u, , " («) Weitz and SchwechtenBer, , ., , ’ r>, , «-•, , »., , 434’ 34 (1923)- 00 Koelsch,, f?" 69®> 665 (1936)‘, , Biegleisen, J Am. Chem. Soc., 64, 2801, 2808 (1942) (c)Kx h, W LewiS’ Lipkin’ and, , ^ouo (U4Z). (c) Kuhn and Kamer, Ber., 85, 498 (1952).
Page 694 :
678, , Free-radical Reactions, , Ph3N:, , -e, , Ph, , Ph2N-NPh2, , PiV, , XII, , XIII, , —e~, , Y, , Ph, N, , ^Ph, , i>h, , V, , phN_/Ph, , +, , jrjf, , [Ph2N-NPh2]+, , Pir^N^Ph, , XV, , Ph, XVI, , Radical ion XIV is obviously analogous to the triphenylmethyl radical in the, carbon series, radical ion XV is not unlike the pentaphenylethyl radical, (Ph3C—CPh2-), and radical ion XVI is closely allied to the pentaphenylcyclopentadienyl radical (XI). Radicals XIV, XV, and XVI, like the corresponding, carbon radicals, are deeply colored and are paramagnetic. Each, however, bears, a positive charge because of the “extra” proton in the nitrogen nucleus. The, positive charges tend to prevent association of the radicals to their “dimeric, forms, for such dimers would be dipositive cations in which a normally weak, N—N bond is further weakened by positive charges on adjacent nitrogen atoms, (for example, Ph3N—NPh3)., Oxidation of/>-phenylenediamine or an alkylated derivative of this amine in, slightly acid solutions yields the so-called Wurster salts, in which the cations are, radical ions of the type XVII.4* Such cations, to which no simple carbon analogs, appear to be known, are presumably stabilized by distribution of the unpaired, electron over all six carbons in the ring as well as over both nitrogens. However,, , XVII (Wurster salt cation), , the fully methylated diamine, XVIII, does not form such a radical,«<« almost, .. (a) Rumpf and Tron.be, J. Mm. pHy35, , 110 (1938);, , (») Michael*, Schubert, and Granick, J. Am. Chm. &c.,, , 61,, , 1981 (1939)., , 206, 671 (1938).
Page 695 :
Further Types of Stable Free Radicals, Me, , -, , 679, , Me, , XVIII, certainly because the large departures from planarity resulting from interaction, between the N-methyl and C-methyl groups prohibit the necessary resonance, stabilization., Closely related to the Wurster salt cations are the semiquinones,16 the, simplest of which is XIX. These, however, are anions, and unlike the Wurstei, salts, are most stable in basic solutions. Semiquinones represent the oxidation, , state lying between hydroquinones and quinones, and, as indicated above, may, be prepared by careful oxidation of the former (in basic media) or careful, reduction of the latter (again in basic media). In neutral or acid solutions, semi¬, quinones tend to disproportionate to mixtures of quinones and hydroquinones or, to the familiar quinone-hydroquinone adducts (quinhydrones). In contrast to the, semiquinones, themselves, to which the unpaired electron may be distributed, over six ring carbons and two oxygens, resonance in the conjugate acids of semi¬, quinones is somewhat restricted. The unpaired electron may again distribute, itself over the nonprotonated oxygen and the positions ortho and para to it; but, structures such as XX, in which the unpaired electron is “placed” on the protonated oxygen (or on positions ortho and para to it) should not be significant, contributors, since these involve considerable separation of unlike charge. This, (19351, , °r the,ChcTltry °f the semicluin°nes, see Michaelis, Chem. Revs 16 243, and Michaelis and Schubert, ibid., 22, 437 (1938)., ’
Page 696 :
680, , Free-radical Reactions, , XX, , accounts, at least in part, for the marked difference between the stabilities of, the anionic and neutral forms of the semiquinones. On the other hand, radicals, XXP6(a) and XXII,;6(6) which are analogous to the neutral forms of semi¬, , quinones, appear to be relatively long lived. Neither of these may undergo dis¬, proportionation to a quinone and hydroquinone without breakage of a C—-O, bond; moreover, both have more extended conjugated systems than are present, in the simpler semiquinones., Tetraphenylhydrazine, PI12N—NPI12, and, more particularly, the tetra-/?methoxy and tetra-jfr-dimethylamino derivatives thereof, have been found to dis¬, sociate partially and reversibly into diarylamino radicals (A^N*) in nonpolar, solvents17, , /"AA N(r_h(iv)2 N_NKiy~R)2, , 2(r, , _/, , /, , These homolyses bring to mind the dissociations of hexaarylethanes into triarylmethyl radicals, resonance stabilization of the radical being once again more, important than that of the “dimer.” With four, rather than six, aryl groups per, molecule, steric assistance to dissociation is much less important than with the, hexaarylethanes, but dissociation is favored by the low N—N bond energy, (P ^ime additional free radicals of long life are listed in Table 16-2, together, with representative references. It is scarcely necessary to point out that in eac, of these radicals, the unpaired electron is incorporated into a conjugated system., . .., T ., rum Cnr 54 4095 (1942). (b) Goldschmidt, Vogt, and, u (a) Fieser and Young, J. Am. Chem. boc., 5 ,, {, 70 278 (1948)., , in color and much ,e» stable than, diarylamino radicals, for it decomposes above -160 .
Page 697 :
Further Types of Stable Free Radicals, , -, , 681, , Table 16-2. Some Types of Long-lived Free Radicals, \, , References, , Ziegler and Ochs, Ber., 55, 2257 (1922)., , (XXIV), , (Ph2C —O:)- , (a toy, XXV), , Schwarzenbach and Michaelis, J. Am. Chem., Soc., 60, 1667 (1938)., Bachmann, J. Am. Chem. Soc., 55, 1179, (1933); Wooster, ibid., 56, 2436 (1934); 59,, 377 (1937)., , Cook, et al., J. Am. Chem. Soc., 75, 6242, (1953); 78, 2002, 3797 (1956)., , NO,, , Ph,N—N—C-Ar, , *, , Ph2N—6:, , (diphenylpicrylhydrazyl, XXVII), , (XXVIII), , II, , (XXIX), , (XXXI), , I, , Goldschmidt, et al., Ann., 477, 194 (1924);, 473, 137 (1929)., , O, , (XXX), , S', , Turkevich and Selwood, J. Am. Chem. Soc.,, 63, 1077 (1941)., , Wieland and, (1914)., , Offenbacher, Ber., 47, 2111, , Cutforth and Selwood,, 70, 278 (1948)., , J. Am. Chem. Soc.,, , Fava, Sogo and Calvin, J. Am. Chem. Soc.,, 79, 1078 (1957).
Page 698 :
682, , Free-radical Reactions, , This becomes obvious even for XXVIII and XXXI when these are represented, by one of their alternate structures. Resonance stabilization is least in the tri-t-, , XXXI, , XXXI', , butylphenoxyl radical (XXVI), for which, however, dimerization is strongly, sterically hindered., , Detection of Short-lived Free Radicals, We have emphasized the difficulties in estimating accurately the concentration of, long-lived free radicals using cryoscopic, spectral, or conventional magnetic, measurements; and it should be clear that such methods are likewise unsuited for, the detection of very small quantities of short-lived free radicals that generally, intervene as intermediates in the usual homolytic reactions.'8 For this purpose,, i* A related technique for detecting and measuring the concentrations of free radicals, para-ortho hydrogen conversion—is interesting, but not significantly more sensitive than the, usua magnetic measurements. In this case, the terms ortho and para have nothing o do with a, benzene ring, but refer to the directions of nuclear spin in the H2 molecule Nuclei with odd, numbers of protons or neutrons have, as do electrons, quantized spin. In orMo-hydrogen the, , solution (see Wigner, Z. physik. Chem.,, , B23, , 38 (1933))., , by the slow conversion of the para to the or o, radicals, and the technique is not suitable for meas, , 40, 129 (1940).), , ^, , J, ^, , -n thc absence of free, , f racbcai concentrations of much, Ortho-hydrogen, Para-hydrogen,
Page 699 :
Detection of Short-lived Free Radicals, , 683, , however, other means of investigation are available. The first compelling, evidence for short-lived free radical intermediates in organic reactions arose from, experiments in the vapor phase. During the period 1929-1935, Paneth and his, coworkers75 studied the thermal decompositions of certain volatile organometallics (for example, Me4Pb and (PhCH2)4Sn) carried by an inert gas passing through, a glass tube, a small portion of which had previously been coated with a “mir¬, ror” of zinc, antimony, or lead. No appreciable reaction occurred in the absence, of heat, but if a small area of the tube were vigorously heated so that the, molecules of substrate passed through the hot zone before reaching the mirror,, a new metallic mirror (Pb or Sn) was deposited in the inside of the tube at the, point of heating. More significantly, however, the gaseous decomposition prod¬, ucts were found to remove the original mirror as they passed over it, although neither, the carrier gas nor the stable final decomposition products could do this. More¬, over, this mirror disappeared most rapidly when it was situated very close to the, point of heating. It was thus inferred that the decomposition of the organometallic compound yielded the parent metal and free alkyl radicals (for, example, CH3* or PhCH2-), and that these free radicals reacted with the original, metal mirror, converting this mirror to a volatile organometallic compound, (which, indeed, could generally be isolated by condensing the emerging gases)., However, if the mirror to be removed was too far from the site of heating, most, of the radicals produced in the decomposition dimerized before reaching the, mirror, and removal was slow., The Paneth mirr*r-removal technique has been used to detect alkyl radicals,, not only from the pyr#lyses of •rganometallic compounds, but also from the, decompositions of aliphatic azo compounds,5°(°> from the pyrolyses of paraffin, hydrocarbons,*0(6) from the photolyses of aldehydes and ketones,*°<c> and from, the action of sodium vapor on alkyl and aryl halides.*0w It is important that, only a few of the simpler alkyl and aryl groups can exist as free radicals in the, vapor state. (Although methyl, ethyl, phenyl, and benzyl radicals have been, detected by Paneth’s method, most of the higher straight-chained alkyl radicals, have not. If formed at all, they rapidly undergo breakage into smaller fragments.), The Paneth technique, as such, is not applicable to reactions in the liquid, phase but there are many substances that are known to react rapidly with short¬, lived free radicals in solution. Radical-consuming reagents whose disappearance, may be measured quantitatively are particularly useful. Among these are, diphenylpicrylhydrazyl (XXVII), whose consumption may be followed by the, , 2702, , ^1935, 380. ^, , °929)' W Paneth and Putsch, Ber., 64B,
Page 700 :
684, , Free-radical Reactions, , fading of its characteristic violet color,w(o) and FeCl3, which often transfers, chlorine atoms to active free radicals yielding the readily titratable FeCl2.**(6), Occasionally the presence of radicals in solution is inferred from the ability of, the solution to initiate polymerization of such unsaturated compounds as, acrylonitrile and methyl methacrylate (even though species other than radicals, are known to be capable of initiating such polymerizations). In any case, chemi¬, cal detection of free radicals is possible only if the “test reaction” proceeds at a, rate comparable to, or faster than, the reaction which the radicals would other¬, wise undergo (just as the Paneth technique is successful only when a large num¬, ber of alkyl radicals react with the metal mirror rather than undergoing, dimerization)., At present, the most promising method for studying small quantities of free, radicals is electron-spin resonance spectroscopy.M As was pointed out in Chapter 1,, the spin of an unpaired electron in the presence of an external magnetic field may, become oriented in two ways, either “against” or “with” the field. These two, orientations have slightly different energies and if radiation is supplied, a transi¬, tion from one orientation to the other may occur. The energy difference between, the two possible spin orientations is proportional to the strength of the applied, magnetic field, but even in a very strong field, this difference is very small (for, example, about 0.0026 kcal per “mole” of electrons at 10,000 gauss). To bring, about this type of transition but with no high-energy processes, radiation in the, microwave region (about 1010 cycles per second) must be used. Without at¬, tempting to describe the operational details/5 we may note that in practice,, the wavelength of the incident radiation is kept constant while the intensity of, the magnetic field that surrounds the sample is continuously varied in an at¬, tempt to detect sudden small changes in the adsorption of microwave energy by, the sample. No corrections for diamagnetism need be applied, and using this, technique, radical concentrations as low as 10~7 molar have been detected., Moreover, by suitable refinements in the method, the mean lifetimes of the radicals, may be estimated., c, Furthermore the observed absorption bands are often found to have fine, structure; that is, they are composed of two or more thin bands lying very close, to each other. This splitting results from interactions between the sfm of the un¬, paired electron and the spins of one or more nuclei. By analysis of the fine structure it, i 0f,en possible to learn whether the unpaired electron is ocal.zed on a sing, atom or “spread over” a number of atoms. Thus, it as leen oun, >, accordance with the “resonance hybrid” representation of tr.phenylmeth) ,, ..., HanlondHSef, , , rri., iqq (l955) For a criticism of this procedure, sec, J Am- Chem. Soe.. 77, 3244 (1955). (« Bamford, Jrntins, and, , Johnson, Nature, 177, 992 (1956)., , r.sF:,, , r,, , R, , 55 g29 (1955)., , * 1—. *. ^
Page 701 :
The Configurations of Free Radicals, , -, , 685, , the unpaired electron in this radical is spread over the benzene rings instead of, being localized at the nonaromatic carbon atom.**, , The Configurations of Free Radicals, Although steric interference among the ortho-hydrogen atoms in the triphenylmethyl radical probably prohibits planarity of the entire radical, it is extremely, likely that the three bonds to the methyl carbon in this and in related radicals lie, in a common plane. Movement of one of these bonds out of the plane of the, other two would lower resonance stabilization without relieving steric strain., For free radicals not stabilized by conjugation, we cannot at present say whether, the configuration about the trivalent carbon is planar (XXXII) or pyramidal, (XXXIII), or, indeed, whether all such radicals have similar configurations., , XXXII, , XXXIII, , It appears that the trivalent carbon in a carbon radical will generally not, support asymmetry if the radical is truly “free.” For example, the free-radical, chlorination of (+)-l-chloro-2-methylbutane yields the racemic dichloro com¬, pound, XXXIV.*5(o) Similarly, the free-radical decarbonylation of optically, , *, Z)-Et—CHMe—CH2C1, , Cl, ., I, Et—CMe—CH2C1 S Et—C—CH2C1, , (racemic), , I, Me, XXXIV, active aldehyde XXXV in the presence of f-butyl peroxide yields racemic, Me, , Me, <-BuO-, , 1, , •, , i-Bu—C *—CHO -> i-Bu—C *—C=0 —>, I, , I, , Et, , Et, XXXV, . c, , i-Bu, , r, , Me, , Me, , 1, rcho, ., C-- i-Bu, , CH, , Et, , Et, , (racemic), , + RCO, , (1958)e These authors point out, how-, , thirds of this electron” is concentrated a^t that carbon^ ^ methyl carbon, , that “about two
Page 702 :
686, , Free-radical Reactions, , 2,4-dimethylhexane.*5(6) It is of interest, however, that a number of optically, active diacyl peroxides of the type XXXVI, having asymmetric a-carbon atoms,, , Me, O, I, I!, R—CH-C-O, *, , Me, , O, , I, , II, , Me, , I, , R—CH-C-O •, , O, , II, , -> R—CH-G, *, v, O, , +, , -> <, , -CO,, , I, , II, , Me, , ^ R-CH- I, Me, , R-CH-G-O, , R-CH-C-O, O, , I, , II, , Me, , O, , XXXVIII, , XXXVII, , XXXVI, , -> R—CH, I, Me, , decompose to esters in which the a-carbon atoms, both in the alkyl and acyl, sections, retain their configurations.*6 Retention of configuration in the acyl, group is to be expected, for in this group the asymmetric carbon is not affected., Retention of configuration in the alkyl group is more surprising. If the mecha¬, nism indicated above is correct, we may suppose that in many instances alkyl, radical XXXVIII, formed in the decarboxylation of acyloxy radical XXXVII,, reacts very rapidly, while still asymmetric, with a second acyloxy radical., (Since partial racemization of the alkyl group is always observed, a short time lag, between decarboxylation and coupling may be inferred.) It seems probable that, the alkyl and acyl groups of the resulting ester molecule stem from the same, molecule of peroxide (although this has not yet been proven), and that the, alkyl and acyloxy radicals recombine before they break out of the “solvent cage, that surrounded the initial peroxide molecule. However, we cannot neglect the, possibility that a portion of the ester is formed in a cyclization without interven¬, tion of free radicals., Me, I, R-C-H, , Me, R-C-H, , O, , +, R, , C., , \ / Vi, , O, II, G, II, O, , C, o, / \, H, Me, , In any event, the stereochemical evidence now available does not, , ^, , guish between the planar and pyramidal configurations or ree ra ica, ', , . (a) Brown, Kharasch, and Chao, 7., , spr'»hs, , O-., , j. Am. elm. Soc, 77, 4809 (1955). « DeTar and Wexs,, , 62, 3435 (1940). (*) Doering, Farher,, , Mr’-(i) 0reenc', , 79, 3045 (195 ,
Page 703 :
687, , Formation of Free Radicals. Initiators, , assumed that a pyramidal free radical can “turn itself inside out” (XXXIX, XL) many times per second, as molecules of ammonia and amines are known to, do. Neither a planar radical nor a pyramidal radical that can undergo rapid, , XXXIX, , XL, , inversion should support asymmetry when free, but either type could if suitably, shielded., Note, however, that a pyramidal configuration is virtually mandatory for, free radicals in which the trivalent carbon is situated at the bridgehead of a rigid, tricyclic ring system. A species of this sort is the apocamphyl radical, XLII, an, intermediate in the decomposition of apocamphyl peroxide (XLI).27, , Me, , Me, , Me, , Me, , Formation of Free Radicals. Initiators, Most of the free-radical reactions with which we shall be concerned are chain, reactions; they are initiated by small quantities of reactive free radicals that may, produce relatively large amounts of final product. An initiator is, as its name im¬, plies, a substance that, under reaction conditions, furnishes a sufficient number, of radicals to get the reaction “under way » Organic free radicals are most often, produced in solution by heating (or in some cases, merely dissolving) compounds, having weak covalent bonds. Organic peroxides, which generally suffer homoly¬, sis at the weak 0-0 bond (ROOR -> 2 RO-), are obviously initiators. These, vary greatly in their stability; a-phenylpropionyl peroxide (XLIII), for exam¬, ple, is so unstable that it cannot be isolated under ordinary circumstances *•»>, whereas /-butyl hydroperoxide (XLIV) is stable for weeks in the dark at room, (’954>;, and, the “brommative decarboxylation” of the silver salt of Tn * Y theKSamc r^dlcal through which, (p. 355)., y, OI mC SllVCr Salt of aP°camphenecarboxylic acid proceeds
Page 704 :
688, , -, , Free-radical Reactions, , Ph—CHMe—C—O—O—C—CHMe—Ph, O, , Me3C—O—O—H, , O, XLIII, , XLIV, , temperature and decomposes only slowly even when heated to 100V* Azo com¬, pounds may likewise act as radical sources, for when these are heated, or merely, allowed to stand, radicals break off from both ends of the molecule, leaving, behind elemental nitrogen., 50°, , Ph3C—N=N—Ph -> Ph3C- + N2 + Ph-*s(o>, Ph—N=N—OH, , S Ph- + N2 (+ -OH)*9(6), , Substituted azoacetonitriles (for example, XLV) are particularly useful radical, initiators; here, fragmentation is favored not only by the stability of the N2 mole¬, cule, but also by delocalization of the unpaired electron over the —C—C=N, conjugated system., N=C-CMe2-N=N-CMe2-C=N, XLV, N=C-CMe2. + N2 + •CMe2—C=N30, Among the other types of compounds which yield organic free radicals when, heated are alkyl nitrates,5/(o) the carboxylic acid salts of Pb(IV),s/(6) and, organocobalt compounds.Sl(c), 200c, , RO—N02, 150c, , Pb/O—C—R^, , ■» RO* -p N02, * Pb/O—C—R\, , I, , o11, , o, , + 2-0—C—R, , ), /., , A, o, , 35°, , RMgV + CoCl2 -* R2C0 ^ Co + 2 RPhotochemical dissociation is an extremely important method for the produc¬, tion of radicals, for many molecules are split by the action of visible or ultraviolet, light, generally of a wavelength corresponding to their respective values of XmaIMany of these are the same substances that undergo homolysis on heating, but, some are not. Chief among the latter group are aliphatic carbonyl compounds,, ts Stannett and Mesrobian, J. Am. Chem. Soc., 72, 4125 (1950)., ., ..., » W Cohen and Wang, 7. Am. C!*m. So,., 75, 5504 (1953). (») This is a seep m .he famthar, Gomberg-Bachmann reaction (see p. 729)., so T fwis and Matheson, J. Am. Chem. Soc., 71, /4/, M Lew J Am. Chem. So,.. 76, 3254, 3790 (1954). <» Hey,, 7. 0,1, detail, , __c, , Shrlmg,, , and, , W.lhams, , S„7i954, 27471 1955, 3969. M Reactions, TJnmsf) Uif h^er, not'd Ji,. present whether the, , organArobaltinter mediate is RsCo, RCoCl, or both; for a discussion of this point, see W.lds and, McCormack, J. Org. Chem., 14, 45 (1949).
Page 705 :
Formation of Free Radicals. Initiators, , -, , 689, , which have been studied most extensively in the vapor state.55 Four typical, homolyses are shown below:, O, , O, , Me2CH—C—CHMe2, , Me2CH—C- + •CHMe255(o), , ‘ O, , O, , Me—C—CH2C1, , ^ Me—C—CH2. + Cl-55®, , Me2CH—CHO, O O, , ^ Me2CH* + -CHO5^, , CH3—c—c—CHs, , 2CH3C=055(d), , Organic compounds may also be converted to free radicals by high-energy, radiation, for when an a, (3, or, , 7, , particle, or an x-ray collides with a molecule,, , ionization (electron removal) may occur. The resulting energy-rich ions may, then break up, generally with homolysis of one or more bonds. Such effects,, which account in great part for the action of radiation on biological material,, are obviously of great importance. However, this topic is not directly related to, the present discussion and will not be pursued further here.5-*, In some instances, free radicals may be produced by electrolysis. Anodic, oxidations of carboxylic acid salts in aqueous solution, for example, yield carboxylate radicals, RCOO*. These undergo decarboxylation to alkyl radicals, (R*), which, in turn, rapidly dimerize. This set of reactions constitutes the, familiar Kolbe electrolysis.36, , RCOO- ^ RCOO. ^ R. ±merize> R-R, The carboxylate radicals may be diverted by addition of anisole to the reaction, mixture, whereupon an o-acyloxyanisole is formed.55^) Similarly, alkyl or aryl, radicals may be produced by electrolyses of Grignard reagents in ether., RMgBr —R. + MgBr+ —> R—R»«, , anH T, |i°tolys« of ^iph*ti<; c^onyl compounds has been reviewed by Noyes, and JoUey (CW to., 56, 49 (1956)), and, more briefly, by Pitts (7., 33 (a) Whiteway and Masson, J. Am. Chem Soc, ibid., 77, 5252 0955). (t) Blare, and, Faraday Soc., 50, 1067 (1954)., ’, , 77, , ’, , 1, , e., ’, , CoMZnanrsleadllow0™»;;™r S"®, “ W Weedon,, , ,, , (d) Nicholson, Trans., , by hiSh-ene^y radiation, tee, , 380^ (1952) 'S«, 471 (1956)., , Wdson and Lippincott, J. Am. Chem. Soc., 78, , «, nard reagents (p. 402)., , (1951)., , Porter, , 4290 0 956), , Y g, , ^ !‘Udy °f 'h'S reaction> see, , Chemt Soci6?-2574 (i94i>;, , the complex nature of solutions of Grig-
Page 706 :
690, , Free-radical Reactions, , Likewise, radicals are produced at the cathode in the electrolyses of ketones in, aqueous acid., HO, , R^,, , ^, , 2C=0->, , „, , *, , _, , OH, , dimerizes _, , R2C—O--> R»C—CR2s7, +2H+, , ", , J, , The same type of dimeric product is formed when benzophenone is electrolyzed, in pyridine to which sodium iodide has been added, using magnesium electrodes., Here, however, the pinacol is formed at the anode, the pole at which electrolytic, oxidation occurs. The reducing agent is almost certainly unipositive magnesium,, formed by oxidation of the magnesium anode, for the ketone is not reduced at the, electrode in the absence of the electric current.ss, _, , PtuC—O, , —e~, , Mg -, , _, , ,, , *, , dimerizes, , Ph2C—o-, , I, , > Mgl;r-> Mgl2 -f- Ph2C—O--■>, , Ph2C—oThe reduction of ketones to pinacols with a combination of magnesium and, magnesium iodide very probably proceeds in the same way., Another important method for the production of free radicals is the action, of inorganic “one-electron” oxidizing or reducing agents.50 The most familiar, reagent of this type is the Fe2+-H202 couple (“Fenton’s reagent”), from which, free hydroxyl radicals may be obtained., Fe2+ + HO:OH -> FeOH2+ + -OH, (Note that in this reaction, iron has been converted from the dipositive to the, tripositive state.) The free hydroxyl radical is very reactive and ordinarily would, react with another Fe2+ ion. However, its existence in solution may be inferred, from the ability of such solutions to initiate vinyl polymerization.^0 Similarly,, dipositive iron reacts both with the peroxydisulfate ion (S2Of-)^(0) and with, organic hydroperoxides (for example, cumene hydroperoxide)^(6) to give free, radicals, Fe2+ + o3S—0:0—SO2- -> FeSO+ + -O—SO^, pe2+, , PhCMe2—0:0H, , —> FeOH2+ + *OCMe2Ph, , 37 Spe, , for example, Haggerty, Trans. Am. Electrochem. Soc., 56, 421 (1929)., - Rausch McEw,;, Jf KlSnberg, J. A.n. Ctm. Soc., 76, 3622 (1954). For a rev.ew of, reductions by unipositive magnesium, together with a summary of the evidence, , hat th.s, , valence state exists, see Rausch, McEwen, and Kle.nberg, Chcm. R,v< 57, 417 (1957)_, « For more detailed treatments of this topic see Bacon, Quart. Revs., IX, 287 (1955). bee, also Walling, Ref. 1(b), pp. 564-579., *0 See, for example, Baxendale, Evans, and Park, , —, „, ,, r., Trans. EaradaySoc.,, , 49, , 15c, , 42,, , 155 (, , h, , 4/ (a) Kolthoff Medalia, and Raaen, J. Am. Chem. Soc., 73, 1733 0, )• ( ), ’, t, r i~\, i£ 1, For work on the reaction between, , Arimoto, and Nudenburg, J. Org. Chem., 16, 1556 (, - )., 61952) • 18, 322, Co2+ and cumene hydroperoxide, see Kharasch, et at., J. Org. Chem., 17, 207 (1952), 1#,, , (1953).
Page 707 :
The Types of Free-radical Reactions, , 691, , The reaction of tripositive titanium with hydroxylamine in acid solution appears, to generate amide radicals, NH2-, for such a solution likewise acts as a polymeri¬, zation initiator.4*, Ti+3 + HO—NH2 -> TiOH+3 + -NH2, Finally, there may be occasions when two or more molecules react in the, absence of initiators, resulting in the formation of new bonds and in the “unpair¬, ing” of electron spins. The proposed “homolytic” mechanism for the DielsAlder reaction (p. 536) is of this type, and analogous processes have been pro¬, posed to account for the “uncatalyzed” polymerizations of styrene and methyl, methacrylate.45, , The Types of Free-radical Reactions and, Some Common Characteristics, The two types of radical reaction in which we shall be most interested are radical, displacements and radical additions. Radical displacements seldom take place on, carbon atoms.44 Most often they occur on hydrogen or halogen atoms, for, example,, Cl- + H—CMe3, CH3- + Br—CC13, , -» Cl—H + -CMe3, CH3—Br + -CC13, , We may, of course, refer to radical substitutions on carbon, but such substitutions, are generally the result of two (or more) independent reactions, neither of which,, under ordinary circumstances, is a radical displacement on carbon—for exam¬, ple, the chlorination of isobutane., Cl- + H—CMe3 Me3C- + Cl—Cl, , Cl, , H -f- -CMe3, , ’ Me3C—Cl + Cl-, , (displacement on hydrogen by C1-), (displacement on chlorine by Me3C-), , On the other hand, in most radical additions, it is a carbon atom (or, rather a, "■-electron system associated with two or more carbon atoms) that suffers attack, Dy the radical., free, ,P™SOf !mPOrtanCe ^ ™y result from the collision of, e radicals Rad,cl coupUng is simply the combination of two radicals, with the, pairing of electrons, to form a new covalent bond,, , W nS’ EV«nS’ 3nd HiSSinson> d. Chem. Soc., 1951 2563, this type of radical formation, ^Waning @5 HbSfpp 'm~m °94<5)', <h>58);, , Mjfnl 5764, , (ml).', , 3 discussion of, , ^ 79> 6370 ««7); 80, 66
Page 708 :
692, , Free-radical Reactions, , whereas radical disproportionation is the transfer of an atom from one radical to, another, forming a saturated and an unsaturated molecule., 2CH3CH2- -> CH2=CH2 + CH3CH3, , (The latter type of reaction may occur with alkyl, but not aryl, radicals.) In, virtually all reactions that proceed through active free-radical intermediates, the, concentration of radicals at ordinary temperatures is much less than the concen¬, trations of other reactants (except perhaps during the final stages of reaction). It, might therefore be expected that the yields of products resulting from radicalradical reactions would be negligible in comparison to the yields of products, resulting from the action of radicals on nonradicals. Although this is sometimes, true, there are many examples of reactions in which the “coupling product”, forms a substantial fraction of the isolated material, probably because collisions, between organic radicals, although relatively rare, may be extraordinarily, efficient chemically. Since no bonds are broken during a coupling reaction,, little or no activation energy should be necessary unless steric hindrance is, significant. It has been calculated, for example, that over half of the collisions, between methyl radicals in the gas phase (at 165° and moderate pressures), result in coupling.45 The formation of disproportionation products, in greater, than trace amounts, is less frequently observed, although many cases have, been recorded in which disproportionation competes favorably with radical, coupling.45, Finally, we may note that in favorable instances, free radicals may undergo, eliminations, decarboxylations, or rearrangements., Br2CH—CHBr, CH3COO, , -> Br+ BrCH=CHBr47(o), -> CH3* + C0247(6), , Me2CPh—CH2. -> Me2C—CH2Ph47(c), By now it should be obvious to the reader that the free-radical reactions, observed in the laboratory are generally not single reactions but are rather com¬, posed of a number of steps. The “chainlike” character of most of such “com¬, posite” reactions arises from a simple mathematical principle—the sum of an, even plus an odd number is always an odd number. If a radical (having an odd, number of electrons) attacks a nonradical (having an even number of electrons),, one of the resulting species must have an odd number of electrons and must itself, be a radical, possibly different from the initial radical, but likewise capable of, ** Kistiakowsky, Gomer, and Roberts, J. Chem. Phys., 19, 85 (1951); 21, 1637 (1*53)., ** See, for example: Ivin and Steacie, Proc. Roy. Soc.,,208A, 25 (19, , ),, , ic e an, .,, , Rec. trav. chim., 69, 312 (1950); Overberger and Lombardino, J. Am. C e, , (1958)., , •, , rum Snr 74 4141 Cl952). (b) Kharasch, Rowe, and, V (a) Steinmetz and Noyes, J. Am. Chem., 74, 4141 t, ) \ ), Urry, J. Org. Chem., 16, 905 (1951). (c) Seubold, J. Am. Chem. Soc., 75, 2532 (195.).
Page 709 :
The Types of Free-radical Reactions, , -, , 693, , attacking a nonradical. Attack by the second radical should produce a third, radical, which may or may not be the same as the first, but which, in any case,, may participate in a third attack, producing a fourth radical, and so on. Such a, sequence (a chain reaction) may continue until the radicals are destroyed or the, reactants are fully consumed. Thus, a single radical may bring about the suc¬, cessive formation and destruction of thousands of radicals, and hence bring, about changes in thousands of molecules., On the other hand, radical chain reactions may be strikingly retarded, or, even halted, by reagents which react with active radicals, forming less active, radicals or nonradicals. Among such “chain breakers” (generally called re¬, , tarders or inhibitors) are elemental iodine and aromatic disulfides., active radical, R., , +, , I—I-> RI, , +, , R*, , + ArS—-SAr-> RSAr -f-, , less active radical, IArS*, , Moreover, such stable free radicals as nitric oxide and diphenylpicrylhydrazyl, (XXVII), which are presumed to couple with active radicals, are likewise re¬, tarders. Elemental oxygen constitutes a special case; as we shall presently, emphasize, it is a type of biradical (that is, it has two unpaired electrons per, molecule) and may act as an initiator when no better initiator is available., However, 02 generally reacts with carbon radicals to yield peroxy radicals;, , —c* + o2, , —c—o—o*, , and radicals of the latter type tend to be relatively unreactive. Hence, in the, presence of more effective initiators, oxygen may act as a retarder. As a result,, we find that a number of free-radical reactions exhibit an induction period if oxygen, is not rigorously excluded. More specifically, such reactions begin very sluggishly, but their rates suddenly increase when all of the oxygen has been consumed., The induction period and inhibition by retarders are two features that, distinguish homolytic from heterolytic reactions. There are others: heterolytic, reactions are not accelerated by light, whereas, as we have seen, homolytic reac¬, tions frequently are. On the other hand, free-radical reactions are much less, frequently subject to acid or base catalysis than are heterolytic reactions. Moreio, , he rates of free'radlcal reactions (except, perhaps, those involving radical, ns) tend to be much less sensitive to changes in solvent polarity and in ionic, ength than are the rates of most heterolytic reactions. Finally^ the relation, , clear cu, , forheTT"' T* 7“****, , ^, , at prCSent) s°mewhat more, , Jss?;reactions-A heteroiytic, reactants wil, generally (although not
Page 710 :
694, , Free-radical Reactions, , donating group similarly placed; however, it is not uncommon for a homolytic, reaction to be accelerated by substituents of both types., The kinetic treatment of free-radical reactions tends to be more complex, than that of most heterolytic reactions; for in addition to the initiation step(s),, there may be several propagation and several termination steps. Consider, for exam¬, ple, the addition of HBr to propylene, catalyzed by benzoyl peroxide (Bz202)., Although this is not a particularly complicated reaction (compared to many, other free-radical reactions), we may write two initiation steps, two propagation, steps, and three termination steps., , (1) Bz202 —2 BzO(2) BzO- + HBr —> BzOH, , (initiation), + Br-, , (3), , Br- + MeCH=CH2 X MeCH—CH2Br, , (4), , MeCH—CH2Br + HBr, , (5), (6), , 2Br-, , (propagation), , MeCH2—CH2Br + Br-, , Br2, , 2MeCH—CH2Br, , MeCH—CH2Br, (termination), MeCH—CH2Br, , (7), , MeCH—CH2Br + Br* X MeCHBr—CH2Br, , (Even this sequence is a simplification, for we have ignored reversibility of, individual steps and have assumed that all benzoate radicals react with HBr.), Now the net rate of reaction may be taken as the rate of formation of n-propyl, bromide, reaction (4), that is, *4(MeCH—CH2Br)(HBr). The concentration of, HBr is measurable; but the concentration of the MeCH—CH2Br radical is not,, and therefore must be expressed in terms of measurable concentrations and the, various rate constants. This is a difficult task since this radical is being created, (in reaction 3) and destroyed (in reaction 7) by action of the Br- radical, the, concentration of which is likewise immeasurable. Here then, we have a problem, that can best be handled by the steady-state approximation (p. 170); that is,, it may be assumed that very soon after the reaction gets underway, the concen¬, trations of each of the three free-radical intermediates (BzO-, Br-, and MeCH—, CH2Br) remain constant. If only the steps in the above sequence are of impor¬, tance, the three “steady-state equations” become, , ^(BzQj = 0 = 2yt1(Bz202) - A,(HBr) (BzO.), , (8), , dt, - o = £2(HBr) (BzO-) -, , /:3(Br-)(MeCH=CH2), , + *4(MeCH-CH2Br)(HBr) - i.(Br-)« - i,(Br-)(MeCH-CH.Br), , (9)
Page 711 :
Free-radical Halogenations, , tf(MeCH, , -, , 695, , - yt7(Br.)(MeCH—CH2Br), , (10), , CH2Br) = Q _ ^3(Br.) (MeCH=CH2), , dt, - £4(HBr) (MeCH—CH2Br) - £6(MeCH, , CH2Br)2, , If the concentrations of the radicals are considered “unknowns” but the rate, constants and the concentrations of added reagents “knowns,” the radical con¬, centrations may be obtained by solving three simultaneous equations (two of, them quadratic) in three unknowns. Thus, a rate expression may be derived., However, unless the problem is further simplified by assuming that certain of, these steps are of negligible importance compared to others (and this is fre¬, quently possible), the algebra is not simple. Moreover, the rate constants for the, separate steps become “lumped together” and cannot, generally speaking, be, individually evaluated. For photochemically induced reactions, however, special, methods are available with which the separate rate constants may sometimes be, determined.45, , Part II —HOMOLYSES AND FREE-RADICAL, DISPLACEMENTS, Free-radical Halogenations, The chlorination of aliphatic hydrocarbons is generally one of the first reactions, with which the student of elementary organic chemistry becomes familiar. It is a, typical free-radical reaction, for it may be initiated photochemically or by such, initiators as benzoyl peroxide, tetramethyllead, and azomethane,49(o) and may be, inhibited by traces of elemental oxygen.49^ The propagation sequence is, doubtless as follows:, is, , In the most important method for determining individual rate constants (the so rilled, , 2f TrT” mah°d>■ ,he miXtUre iS SUb-i“"d, -cccssiv. Short£=rio*of, and darkness by rotating an opaque disc, from which a sector of predetermined size hTbf™, cut, between source of light and the mixture The “licrht”, j ■, , ,, as been, the concentration of radicals does not IttTin the s efdv-su vlT th't, ‘ ,7^ S,° that, uninterrupted illumination, and the “dark” period is made shorentha* WOul.d pi?Vai1 with, concentration does not drop to a negligible value In favnnKl, enouSh so that the radical, evaluate individual rate constants from the m -,n ’ . •, , ,b eL CaSCS’ U becomes possible to, with the known duration of the “lieht” and “d L.^ m .w*jlcb tbe overall reaction rate varies, matical treatment of such data s'? Flo™ / A, X” " F<>r the (rather “"^1 ™the« < MJ., 71, 497 0 949) For descriotTonf of a' C^m\So‘- ». 241 (>537); and Ma.heson,, , 53’ 3728 (1931)- W Ritc’h^e 2d fe4^cl2iI0^)lM“3^WalZ* J• Am• Chem•
Page 712 :
696, , Free-radical Reactions, , RH + Cl-, , R- ^ RC1 +, HC1, , and, as we have seen (p. 685), if the carbon atom at the substitution site is, asymmetric, complete racemization occurs. At least two modes of chain termina¬, » RC1; and R- + R* —> R—R) are possible, even if the system, , tion (R- + Cl-, , is rigorously protected. In the presence of trace impurities, to which these reac¬, tions are extremely sensitive, other termination mechanisms may become impor¬, tant. This may affect the apparent rate law,40(c> but not the yields of the possible, chlorination products, for these should be determined by the rates at which the, various (nonequivalent) hydrogen atoms of a hydrocarbon molecule are re¬, moved in the initiation step., It has long been recognized50 that tertiary hydrogens are removed more, readily in free-radical chlorinations than are secondary, and that these are, in, turn, removed more readily than primary. Typically, chlorination of isopentane, (in the vapor phase at 300°) gives the following mixture of monochlorides:50, C1CH2, , Cl, , H,C, , S2, , 3, , 4, , ^CH—Et (34%) + Me2C—Et (22%) +, , cu, , CH—CH2—CH3 ->, H C, , /, Me2CH—CHC1—Me (28%) +, , H3C, , Me2CH—CH2—CH2—Cl (16%), Noting that there are six equivalent hydrogens at the 1 positions (primary), a, single hydrogen at the 2 position (tertiary), two at the 3 position (secondary),, and three hydrogens at the 4 position (primary), we may estimate that reactivi¬, ties of primary, secondary, and tertiary carbons lie in the ratio of about 1:3:4, (Ex. 2). Using these (or similar) values, it is possible to predict rather satisfac¬, torily the ratios of products formed in the monochlorination of other aliphatic, hydrocarbons under the same conditions.50 Similar treatments, with somewhat, different reactivity ratios, should apply to different temperatures., The observed sequence in reactivities (tertiary > secondary > primary), brings to mind a similar sequence governing the reactivities of aliphatic halides, in SN1 reactions (Chap. 8) and suggests that the formation of a carbon radical,, like that of a carbonium ion, may be facilitated by hyperconjugation involving, a-alkyl groups., H, , +, , H+, , H, , R-fc-«R-n. for example, Hass, McBee, and Weber,, See, (1936)., 50, , Ind. Eng. Chem.,, , 27, 1190 (1935), 28, 333
Page 713 :
Free-radical Halogenations, , -, , 697, , On this basis, we would likewise expect a chloro or cyano substituent to facilitate, ladical attack at the a-carbon atom, for “resonance-stabilized” radicals should, be obtained by removal of an a-hydrogen in either case, , -C—C=N: <->— C=C=N:, , -C—Cl: <->—C—Cl:, , We might therefore predict that alkyl chlorides and nitriles would undergo freeradical chlorination preferentially at the a positions. This is precisely what is, , not observed; it is now known that homolytic chlorinations of alkyl halides5*(a), and aliphatic nitriles5*(fc) generally yield only minor amounts of the products, resulting from a chlorination but much larger amounts of the products resulting, from /3, and where possible,, , 7, , chlorination. The same is true for chlorinations of, , carboxylic acids, esters, and acyl chlorides.5*(6)'5* Some factor (or factors) other, than the stability of the radical intermediate must then account for activation by, an a-methyl group but deactivation by an a-chloro, and a-cyano, or an ct-carboxyl substituent. It seems very likely that a series of polar effects are being ob¬, served—that is, that the attacking chlorine atom, besides being a radical, is an, electrophilic species that seeks out an electron-rich site in the substrate for pref¬, erential, , attack., , Thus,, , “+/ groups”, , whereas “ — I, , facilitate chlorination,, , groups,” as observed, retard it.55 Similarly, the side-chain chlorination of toluene, has been found to be retarded by electron-attracting substituents in the benzene, ring.5-*, As with many other reactions yielding a mixture of isomeric products, freeradical chlorination generally becomes less selective as the reaction temperature, is increased; that is, at a high temperature a chlorine radical becomes more likely, to extract the first hydrogen atom with which it comes into contact. Accordingly,, it is somewhat surprising to find that in the chlorination of alkyl chlorides, the, yield ^of 1,2-dichloro compound falls very nearly to zero at temperatures above, 37 50.55 This is almost certainly due to the tendency of radical XLVI (the preK.h<ira.sch and Brown. S. Am. Chem Soc, , 77■ 4919 <1955>; (b)]3™ylanB- ", , fit, , 9149, , n, , Bdg.: 58, 210 (1949)^2", , Kharasch and Brown, J. Am. Chem. Soc., 62, 925 (1940), , ^ 1Z ^XlnsinTcE, , *, , ’, , °" ™, , SeXS, has blTunZ™, various ring substituents., f Vaughan and, , Rust,, , 7?! 269 <»*>• A satisfactory correlation, 1C ° 1 orinatlon rates and the <r values for the, J. Org. Chem.,, , 6, 479 (1941).
Page 714 :
698, , Free-radical Reactions, , sumed intermediate in the formation of the 1,2-dichloro compound) to decom¬, pose to an olefin, which, under the reaction conditions does not add chlorine., Cl-, , R—CH—CH2C1, , •, , ^75°, , R—C—CH2C1-* R—C=CH2 + Cl., XLVI, , Elemental fluorine reacts violently with most organic compounds under ordi¬, nary circumstances, but fluorinations may be studied in the gas phase by diluting, the reactants with nitrogen, or in the liquid phase by dilution with such inert, solvents as, , CCI3CF3.56 Even, , under very mild conditions, however, C—C bonds,, , as well as C—H bonds, are often attacked. The ease with which fluorinations, occur in the gas phase and the frequent isolation of coupling products (as well as, fragmentation products) leave little doubt that these reactions are homolytic, and that they involve fluorine radicals. Since, however fluorinations often occur, in the dark, at low temperatures, and in the absence of initiators, we may ask, how these radicals arise. It has been suggested57 that fluorine radicals are formed,, together with alkyl radicals, when an energetic fluorine molecule collides with a, hydrogen atom of a hydrocarbon chain., , —C:H + F:F, , -C- + H:F + F, , Such a reaction is favored thermodynamically by the low F, , F bond energy, , (37 kcal) and the very high H—F bond energy (135 kcal) and, although it is, quite unlike any other initiation process presently known to occur at low tem¬, peratures, no better possibilities come readily to mind. In any event, elemental, fluorine, in the presence of compounds having C—H bonds, appears to be an, excellent radical source, for traces of fluorine have been found to catalyze the, homolytic chlorination and oxidation of such compounds very strikingly., Let us turn briefly to aliphatic bromination; although this is similar, in a number of ways to chlorination, there is one fundamental difference., Since H—C bonds are generally stronger than the H—Br bond but weaker, than the H—Cl bond, the hydrogen-removal step in bromination,, , —C—H + Br- —> —C- + HBr, , is often endothermic whereas that in chlorination,, « W Bigelow, Ckm. to., 40 , 51 (1947). (4) The fluorinations of, carbons (fo/example, CC1.CHC1,) are much less vigorous gX***?*, room temperature; see, for example, Miller, . at., em., ’ 49g2 (1956)., „ Miller, Koch, and McLafferty, J. Am. Chm. See., 78, 4174 (IM).
Page 715 :
Free-radical Halogenations, , C—H + Cl-, , -, , 699, , —C- + HC1, , is exothermic. This difference is reflected in the respective activation energies;, that for attack by bromine is about 14 kcal per mole more than that for attack, by chlorine.55 Consequently, a sizable fraction of the Cl- radicals present in, a hydrocarbon-chlorine mixture may attack the hydrocarbon successfully at a, temperature at which only the most energetic Br- radicals in a hydrocarbonbromine mixture may attack. A chain reaction initiated by a given bromine, atom proceeds relatively slowly and is likely to terminate when only a few, molecules have become brominated (it is sometimes said that kinetic chains, in bromination are “short,” whereas those in chlorination are much “longer”)., Many more acts of initiation are therefore needed to produce a given yield, of bromination product than to produce the same yield of chlorination product, under similar conditions.59, We have noted at several points during the earlier chapters that as a reagent, becomes less reactive, it generally becomes more selective. We should therefore, expect the secondary and tertiary halogenation products to predominate over, primary halogenation products to an even greater degree in bromination than in, chlorination, and this is what is found. Bromination of isobutane and cumene50(6), results in almost exclusive attack at the tertiary carbons, whereas the bromina¬, tion of H-pentane yields a mixture of 2- and 3-bromopentane (both secondary, bromides), but little, if any, 1-bromopropane.50^, Me, , Me, Br2, , I, , t>, , Me—C—Me -» Me—C—Me;, I, , H, , ‘, , PhCHMe2 -4* PhCBrMe2, , I, , Br>, , Br, , Br2, , ^-CgHi2 —>, Me—CH—Pr, -f- Et- -CH, Br., i, I, Br, , -Et, , (no Me(CH2)4Br), , I, Br, , Similarly, attack by Br- is, to a marked degree, isotopically more selective than, awjfZfSW Ar,Sda,Cn' and *“““"•**, , CW PHy,, ,0, 305 (1942); II, 6, , " Moreover, hydrogen abstraction by Br- is reversible; that is, the reaction, R* + HBr —> RH -f- Brstance which fulheHo^r ^ mes^f" homoli^brominat**, , irrevcrsible, a circum-, , chlorination. The attack of HBr by alkyl radical is an, at'°n t0 th°Se °f h°molytic, tion of HBr to olefins (p. 735), and^ in the homolytic addifor the difficulty in bringing about homolytic additionsVAci * *, , ^ aCCOunts in Part, , sell and Brown,^ Am^Che^Soc^^ Chm' ^ 20> 1659 (1952)- (*) Rus-
Page 716 :
700, , Free-radical Reactions, , attack by Cl*. Typically, the removal of either hydrogen atom from the side chain, of PhCH2D using Cl* is about twice as fast as removal of the deuterium atom, (that is, kB/kD = 2.0).61M When Br* is the attacking radical, the ratio kH:kD rises, to 4.6.61(f>), Free-radical halogenations may be carried out using reagents other than the, molecular halogens. The most familiar of these “halogen carriers” is N-bromosuccinimide, XLVII, the N—Br bond of which is easily broken.®* Although this, , H,C^ \, N—Br, , H2C^ /, , o, XLVII, is often considered a brominating agent specifically for allylic carbons (that is,, for carbons alpha to C=C double bonds), it may, in the presence of efficient, initiators, brominate saturated compounds as well.65 In keeping with their, homolytic character, brominations with N-bromosuccinimide (NBS) in nonhydroxylic solvents, may be accelerated photochemically or by addition of, peroxides or azonitriles; moreover, they are subject to inhibition by hydroquinone, elemental iodine, or large quantities of oxygen., Very probably, the initial attack on the substrate (RH) in these brominations is by the succinimido radical (XLVIII), after which the resulting carbon, radical attacks a second molecule of NBS. If the carbon radical R* is allylic, it, may react with the second molecule of NBS at one of two sites, leading to a mix¬, ture of bromides., , N-Br, s, , O, , o, , o, , O, , //, , hv or, , X, , initiator, , /, , N*, , R!I->, , o, , /, \, N-H, _/, , +, , \>, , XLVIII, , (,) Brown and Russell, J. Am. Chm. Soc., 74, 3995 (1952). (3) Wiberg, Chem. Revs., 55,, , 731, , « Brominations with N-bromosuccinimide have been reviewed I by' Djerassi, £££*■£, , 271 (1948) Heterolytic brominations with this reagent are discu, ., ., have, 43,-71 (1, )., y, 4327 0958) Other N-haloimidcs and N-haloamides hav <, and Peterson, J. Am. Chem. Soc., 80, 4, (, , Wnhl Rer 52 51 (1919), and Hebbelynck, been used as halogcnating agents; see .for^example Wohl,, 52,, 1, d'i, , and Martin, Bull soc. chm. Beiges, 59, 1 >3 (1» )., « Ford and Waters, J. Chem. Soc., 1952,, nuu, 18, 649 (1953)., , Dcmcrseman, J. 0,g. Chem.,
Page 717 :
Free-radical Halogenations, , +, , 701, , O, , O, R-, , -, , <N-Br, , etc., , 7«-, , RBr, , ->, , +, , O, _H-, , *, , *, , _, , NBS, , AinCH2CH=CH2-> [AmCHCH=CH2 <-> AmCH=CH—CH2]->, (Am = /z-C6Hi2), , !, , Br, AmCH—CH=CH2, AmCH=CH—CH2Br, cis and trans, , (17 %)6*, (83%), , Moreover, if there are two nonequivalent allylic hydrogens in the substrate, as, many as four different allylic bromides could conceivably be found in the result¬, ing mixture (disregarding cis-trans isomerism). As yet, however, too few cases of, this sort have been studied to allow us to say whether all possible bromides will be, formed in significant amounts.65, Both the attack of the succinimido radical (XLVIII) on the substrate RH, and the subsequent attack of the resulting carbon radical (R*) on a molecule of, NBS may be stereospecific. When the erythro form (XLIX) of the deuterated, bromide, PhCHD, , CHBrPh, is treated with NBS, deuterium atoms are found, , to be removed from the substrate nearly twice as rapidly as are hydrogen atoms.66, , Ph, , (predominant products), , LI, , his is a little surprising, since, under ordinary circumstances, C—H bonds are, ttacked about 2.5 times as often as are O n K, a, •, i, aS are C~D bonds m brominations using, „!ate™n and Gunneen, J. Chem. Soc1950, 941, , "c
Page 718 :
702, , Free-radical Reactions, , NBS.W(6) No doubt, this reversal of the “usual” isotope effect is partially a steric, phenomenon; the preferred conformation of the substrate is presumably LI, in, which the bulky phenyl groups are as far removed from each other as possible., We see that removal of a hydrogen atom from the carbon atom in back requires, that the attacking radical approach on the more crowded side of LI (near the, phenyl and bromo substituents), whereas to remove a deuterium atom, the, radical approaches on the less crowded side (near the phenyl and hydrogen)., Polar effects likewise favor removal of the deuterium atom, for it would be, expected that the electronegative nitrogen atom in the attacking succinimido, radical would keep as far as possible from the electronegative bromine atom in, the substrate., Nor is that all. As indicated above, the resulting dibromide, L, is largely, (about 90 percent) in the meso form, indicating that the formation of the new, C—Br is also stereospecific. This could be taken to mean that, in most instances,, the trivalent carbon retains the same asymmetric configuration during the time, interval between abstraction of the hydrogen (or deuterium) atom and formation, of the new C—Br bond; however, we have already considered the evidence that, such carbon atoms do not support asymmetry (p. 685). Instead, we may invoke, reasoning similar to that used in Cram’s rule concerning steric control of asym¬, metric induction (p. 549). The favored conformation of the radical intermediate, is probably LI I, in which, again, the phenyl groups lie as far apart as is possible,, , and it may be supposed that the MBS molecule, from which radical LII extracts, a bromine atom, approaches over the less crowded side (the left side in the drawing) of this radical., The mode of action of a second familiar “halogen earner,” sulfuryl Monde,, SO2CI2, is not nearly so clear. A number of its peroxide-catalyzed reactions, (such as the chlorinations of cyclohexane and the simpler carboxylic acids, undoubtedly involve the Cl- radical as an intermediate, but because of the d.ssociation equilibria,, SO2CI2 ^ Cl- + SO2CI. ^ 2C1- + S02, any mixture containing sulfuryl chloride at moderate, contain the SO2O radical and molecular S02. Either, or io, - Kharasch, , and Brown, 7. dm., , «., , £, ,, , (1939); 62, 925 (1940).
Page 719 :
Iodine Exchange Reactions, , 703, , may enter the reaction chain, and it is not surprising that chlorinations with, S02C12 are frequently accompanied by sulfochlorinations (that is, by conversions, of the type RH -> RS02C1).M Moreover, since hydrogen atoms may be re¬, moved from the substrate by S02C1* radicals, as well as by Cl* radicals (and since, the two types need not display the same selectivity), the mixtures of alkyl, chlorides obtained in chlorinations with SO2CI2 will not generally have the same, compositions as the mixtures obtained in chlorinations with elemental chlorine, alone., , Iodine Exchange Reactions, We should not expect homolytic iodinations of hydrocarbons with I2 to occur., , / l, , l, , \, , The hydrogen-removal step ^—C—H + I* —> —C* + Hly should be strongly, endothermic, for the H—I bond is a weak one. Organic iodine compounds are, how¬, ever, often attacked by elementary iodine (generally homolytically), with the, main result being the substitution of iodine atoms from the elementary iodine for, iodine atoms in the substrate—that is, iodine exchange., R—I + I*—I -> R—I* + I—I, As indicated above, such a reaction is generally followed by using “tagged”, iodine., An important case of this sort is the reaction of optically active jw-butyl, iodide with radioactive iodine.*3 If every act of iodine exchange involved a secbutyl radical, EtCHMe, racemization should occur at the same rate as exchange., However, it is found that racemization proceeds 1.5 times as rapidly as exchange., This calls to mind the reaction between optically active iodides and iodide ion, in, which, as we emphasized (p. 266), every act of substitution is a Walden inver¬, sion, resulting in a rate of racemization twice that of iodide exchange. With ele¬, mental iodine, however, it appears that exchange takes place by a combination, of at least two paths, the first a stepwise substitution., R—I + *!• —> R. -f I—-I, , R—I* + I., , This, if alone, should result in equal rates of exchange and racemization; and the, second a direct displacement by iodine atoms,, *1* + R—I —► *1—R -f-, , 1., , which, if alone, should result in a 2:1 ratio between the rates of racemization, , and, , ^am■ **, 66 Herrmann and Noyes, J. Am. Chem. Soc., 78, 5764 (1956)., , «•««
Page 720 :
704, , Free-radical Reactions, , and exchange. If this interpretation is correct, this reaction would constitute one, of the few examples of a radical displacement on aliphatic carbon.70, Two mechanisms appear to operate also in the exchange of iodine between, 12 and iodobenzene.71 The observed rate law is, rate = *x(PhI)(I2)0-* + *2(PhI)2, , (11), , The first of these terms is not significantly affected by a change in solvent (sug¬, gesting that it is associated with a homolytic process). The second term, which, does not involve (I2), is very sensitive to solvent and may be made to disappear, by transferring the reaction to nitrobenzene. Free phenyl radicals are not, involved in either process, for the exchange is not retarded by oxygen. The first, term is consistent with the following sequence:, , fast, eq, K, , *, I*, , Phi, slow, , fast, ->■, , ->, , +, , !•, , the rate of the slow step being, rate = A(PhI)(I-) = */C(PhI)(I2)0-5, , (12), , Here, it is assumed that, as in electrophilic aromatic substitution, the new bond, to the benzene ring forms before the old bond breaks., The second term in equation (11) has been ascribed to a heterolytic process, —the reaction of two iodobenzene molecules to form diphenyliodonium iodide, (LIII). Presumably, this rapidly exchanges its ionic iodine with I2, then decom¬, poses, as diaryliodonium iodides are known to do,'~ to two molecules of iodo¬, benzene. Although this sequence fits the observed kinetics, more convincing, , •T, I—Ph, LIII, 70 In the exchange of iodine with w-butyl iodide, it is likely that the only important in¬, itiation step, except at high temperatures, is the homolysis of I2. However, as the carbonradical intermediate becomes more stable, a second mode of initiation may'become, ., \ kinetic analysis (which is too complex to be considered here) of the exchange of iodine wit, benzyl iodWe (Gazi,h and Noyes, J. Am. Chm. Soc. 77, 6071 (.955)) ind.ca.es .hat .he mmati0n StCP’, , PhCHoI - PhCH2- + I-, , has assumed importance in this re^°n7/ I evine and Noyes, J. Am. them. Z>oc.,, , HU,, , z^ui, , m Beringer, et al., J. Phys. Chem., 60, 141 (1956)., , (1958); these authors also discuss the, ., „, fll_tupr
Page 721 :
Autoxidations, , 705, , evidence is needed to prove that the diaryliodonium iodide is formed under the, conditions used. If indeed it is formed, we should be able to observe exchange of, labeled iodine between one aryl iodide and another when the two are heated, together in the absence of elemental iodine., Finally, we may note that the exchange of iodine between I2 and acyl, iodides, R—C—I, takes place wholly heterolytically.73 Such exchanges are not, , I, , O, accelerated by illumination, but are catalyzed by traces of moisture and are very, sensitive to the dielectric constant of the solvent. The observed rate law,, rate = A2(RCOI)(I2) + *3(RCOI)(I2)2, , (13), , suggests two heterolytic processes. The first term corresponds to a rate-determin ing and reversible formation of an acyl tri-iodide (LIV),, , O, , O, , slow, R—C—I + I—I*-> [R—C=0]+[I—I—I*]- -> R—C—I* + I—I, LIV, and the second term may well correspond to the similar formation of an acyl, *4-, , pentaiodide, R—C=0 Ig\, , Autoxidations, The reactions of organic compounds with elemental oxygen under mild condi¬, tions are generally referred to as autoxidations, for such oxidations often take place, by themselves” when the (slightly impure) substrate is exposed to the atmos¬, phere. Compounds of many types, including alcohols,74(o) phenols,7^(6) enols 74M, ethers,™ amines,™ ketols,™ and Grignard, , reagents,™, , may undergo, , autoxidation, but we shall confine our attention in this section to the autoxidations of hydrocarbons and aldehydes., ■ I U"d" ‘he fdeSt COnditions in which a reaction can occur (temperatures, below 100 tn the presence of a free-radical initiator), oxygen often attacks, hydrocarbons to form alkyl hydroperoxides (compounds containing the -OOH, group). Some typ.cal hydroperoxides formed in the autoxidation of hydrocarbons are shown as follows:, y °, ” ?°^man and Noyes, J. Am. Chem. Soc., 79, 5370 (1957), (a) Brown, et at., J. Am. Chem. Soc 77 1756 (\occ\, i
Page 722 : 706, , Free-radical Reactions, , OOH, LVIII, , LIX, , Of these, LV is obviously derived from decalin,75<a) LVI from cumene,75® LVII, from tetralin,75(c) LVIII from cyclohexene,75(d) and LIX from the dimer of, cyclopentadiene.75(e) Attack on such hydrocarbons may be initiated by a radical, derived from an outside source (generally a peroxide), or, much less effectively,, by molecular oxygen,75 which is itself a “diradical.”, , R- H-C—H -» R—H d-C*, , I, , I, , or, , •6—6- d-C—H -* -6—OrH d-€•, , The propagation sequence is simple., _<LH, , O,, , —c- d- O2 —> —C—O—O-> —C—OOH d-c* -> etc., , One of the termination reactions is undoubtedly the coupling of two alkyl, radicals to form a alkane with twice the number of carbons, but the fate of the, _C_OO- radicals, in the absence of added inhibitors, is a question that must be, investigated for each individual instance.77-75, As is the case with hydrogen abstraction by other radicals, attack of a C, , H, , « (a) Criegee, Ber., 77B, (1944). (b) Armstrong, Hall, and Quinn, J Chem. Soc., 1950,, , 666. (c) Farmer and Sundralingham, ibid., 1942, 121. (d) Criegee, Pilz and Flygare, Ber., B72,, 1799 (1939). (e) Hock and Depke, Ber., 84, 356 (1951)., n, 76Kahn J. Chem. Phys., 22, 2090 (1954). Agreement is, however, not complete that Us, may react directly with C-H bonds to initiate radical chains at moderate tei"Per^tU^S 53°, an Loosing view, see Bateman, Hughes, and Morris, Discussions Faraday Soc.,U, 190, 953h, ”PPeroxy radical with a-hydrogens may, for example, disproportionate in the followi g, manner:, , I, , I, , I, , 2—CH—OO- —*• —C=0 d-CHOH, , +, , 02, , STMw* Mayo Mine, a’nd RusseU,, Chem. Soc., 80, 2500 (1958); Bateman, Quart. Reus., 8, 147 (19541., , d,„.
Page 723 :
707, , Autoxidations, , bond with ROO radicals proceeds more readily if the carbon is tertiary or, secondary than if it is primary, and removal of a hydrogen from an allylic or, benzylic carbon is still easier. Indeed, it has been estimated75 that incorporation, of an additional alkyl group at the reaction site (converting a primary carbon to, a secondary, or a secondary to a tertiary) facilitates the removal of a hydrogen, from that site by a factor of about 3; whereas replacement of an a-hydrogen, with a phenyl group or a vinyl group (converting the a-carbon to a benzylic or, an allylic carbon) raises the reactivity of the C—H bonds 23-fold and 100-fold,, respectively. Thus, it will be noted that in formation of hydroperoxides LV-LIX, by autoxidation, attack has occurred at alpha to a double bond, alpha to a ben¬, zene ring, or at a tertiary carbon., In the absence of added initiators, the autoxidations of many hydrocarbons, are autocatalytic. Hydroperoxides, as might be expected, may act as free-radical, initiators (although they are much less effective than dialkyl or diacyl peroxides)., During the early stages of such reactions, the concentration of hydroperoxide is, steadily increasing, and more and more kinetic chains are being initiated by the, hydroperoxide formed. In fact, early in the reaction, the oxidation rate may be, very nearly proportional to the quantity of oxygen that has already been ab¬, sorbed.50 Later, however, the reaction rate diminishes (as in an ordinary reac¬, tion) as the supply of hydrocarbon becomes depleted., The preparation of hydroperoxides in high yield from hydrocarbons offers, serious difficulties. If the temperature is too low, the chain reaction becomes, inconveniently slow, but at higher temperatures homolytic decomposition of the, hydroperoxide may greatly reduce the yield. Moreover, with unsaturated, hydrocarbons, the hydroperoxides, when formed, can attack the double bonds, of the remaining hydrocarbon molecules or those of olefinic hydroperoxides., Thus, although over 50 hydroperoxides have, to date, been prepared by autoxi¬, dation, only a few (including hydroperoxides LVI, LVII, and LIX) have been, obtained in yields of over 30 percent., The autoxidations of hexaarylethanes occur readily in the absence of out¬, side initiators. These hydrocarbons, which, as we have seen, dissociate into, triarylmethyl radicals, yield dialkyl peroxides,5' almost certainly by the follow¬, ing path:, Ar3G—CAr3, , Ar3C- S Ar3C—02. —C~CAr3>, LV, Ar3C, , 02, , CAr3 -f- Ar3C*, , etc., , " F°lla.nd’ TranS- Faraday Soc-> 46, 358 (1950)., Proc. Roy. Soc. London,, , Am*218, , 309 (1946); W Ba,^, Ga-d J,, Ziegler, Ann., 551, 127 (1942)., , See’, , f°r, , examPle:, , (fl), , Bolland,, , A1M’ ^, ’’ iya3» 3029-
Page 724 :
708, , Free-radical Reactions, , In the presence of inhibitors such as catechol, radical LV is diverted, and the, hydroperoxide, rather than the peroxide, is formed., , Ar8C—O—O*, , +, , LV, , The autoxidation of aldehydes proceeds with ease, even at room tempera¬, ture. This is likewise a chain reaction with steps analogous to those in the autox¬, idation of hydrocarbons., R—C* S R—C—O—O*, , RC-HO-> R—C—OOH + R—C* °’, etc., , In-, , R—C—H, , II, , O, , O, , o, , o, , o, , InH, , Here, the product, a beroxy acid /R—C—OOH\, is generally not isolated as, , II, , O, such, but may be converted, by adding acetic anhydride, to a diacyl peroxide, (LVI)., R—C—OOH + Ac20 -* R—C—O—O—Ac + AcOH, , O, , o, LVI, , Here, once again, the chain reaction may be initiated by a radical derived, from an outside source (designated In*), or, less effectively, by 02 itself;*5 and,, once again, removal of a hydrogen atom from the substrate is facilitated by the, presence of electron-donating groups.*4,55, At present, the substances that are known to be most effective in inhibiting, autoxidations are phenols and aromatic amines. In spite of the great importance, of these antioxidants, their mode of action is not yet completely clear. There is, little doubt, however, that they react with thcperoxy radicals in the kinetic chain,, rather than with the alkyl or acyl radicals; for such inhibitors differ considerably, « For a criticism of this interpretation, see Boozer, Hammond, Hamilton, and Sen, J. Am., , C|^53).^Th^e authors, t»owcver^pr«sent, ihe Ve'^Ton rnula«" d by an added radical source. This very in.eresiing, , question merits further investigation., , of the termination steps is not yet clear., , 9027 (1951)
Page 725 :
709, , Autoxidations, , from the species that are known to inhibit reactions involving only alkyl radicals, (for example, polynitrobenzenes, disulfides, and iodine).56 The simplest mecha-, , R-02- + HO, , OH, , ROOH + ‘O, , OH, , nism for inhibition is simply a hydrogen-atom transfer and this probably operates, in some cases.56(o) On the other hand, it has been found that the deuterated, amines PhNDMe and Ph2ND inhibit the autoxidation of cumene just as effec¬, tively as do their respective nondeuterated analogs,52 whereas if the inhibition, process were to begin with a reaction involving the breakage of a N—H (or, N—D) bond, the deuterated compounds should be less effective. Moreover,, autoxidation is similarly inhibited by tetramethyl-/>-phenylenediamine, , (p-, , Me2N—C6H4—NMe2), in which there are no N—H bonds. It therefore has, been suggested52 that the antioxidant action of aromatic amines (and possibly, that of some phenols as well) is due not to hydrogen-atom transfer but rather to, the formation of complexes with the peroxy radical; for example,, , ROO +, , “OOR, , LVII', Note that the cation in the ionic structure of the complex is the cation present in, Wurster’s salts (p. 678). Indeed, the characteristic blue color of this cation is, observed when solutions of cumene and the parent diamine are treated with, oxygen., Autoxidations are catalyzed by traces of metal salts. To be effective, the, salt should be derived from a metal having at least two readily accessible oxida¬, tion states differing by one unit (for example, Fe, Co, Cu, V, Mn). Peroxides react, with both the upper and lower oxidation states of these metals, for example,, Fe2+, , 4- ROOH -> Fe(OH)2+ + OR-, , Fe(OH)2+ + ROOH -> Fe2+, , + HOH + ROO-, , Hence, if a salt having the metal in the lower valence state is added to a solution, contammg peroxide, the metal ion will be oxidized, then reduced, then oxidized, radicals, has been demonstrated"by kirLaic'nwt'hods"s,'han with alkVl or acV*, Faraday Soc., 43, 201 (1947). (b) Fo/vvhat appears to he *** B° land a.nd ten Have> Trans., Waters, J. Chem. Soc., 1952, 2432., PP, b an excePtlonal case, see Moore and
Page 726 :
710, , Free-radical Reactions, , again, etc. Each of these changes will generate a radical capable of removing a, hydrogen atom from a molecule of hydrocarbon (or aldehyde). As a result, these, heavy-metal salts may greatly enhance the effectiveness of peroxy compounds as, initiators.57 On the other hand, since such salts, in effect, catalyze the decomposi¬, tions of peroxides, the yields of peroxides obtained from such accelerated autoxidations are generally reduced., , Thermal Decompositions of Hydroperoxides and Dialkyl Peroxides, As we have noted, the autoxidations of hydrocarbons may become quite complex, at temperatures above 100°, for the decompositions of the hydroperoxides, (which are generally the initial products) become increasingly competitive as, the temperature is raised. The nature of these complications may be more, clearly understood by examining the decompositions of hydroperoxides, , in, , the absence of their parent hydrocarbons. Decomposition studies of a number of, hydroperoxides, both alone and in solvent, have been carried out, with the, greatest attention to date directed toward the two tertiary hydroperoxides, tbutyl hydroperoxide (/-Bu02H) and cumene hydroperoxide (PhCMe202H)., /-Butyl hydroperoxide, either in the pure state at 100°s5(o) or in an inert, solvent at below 140°*S'S5(6) decomposes cleanly to /-butyl alcohol and elemental, oxygen. The decomposition is evidently a chain process, for which the following, mechanism5S(6) seems likely:, Me3C—OOH -» Me3C—O* + -OH, , (14), , Me3C—O* + HOO—CMe3 -> Me3C—OH + -OO—CMe3, , (15), , HOO-CMei, , 2MeaC—OO- -> 02 + 2MejCO-> etc., The first two of these steps, the homolysis of the O, , (16), , O bond and the transfer of a, , hydrogen atom, are of familiar types. The third is more interesting, for here,, two radicals collide, yielding two new radicals and a diradical (the 02 molecule)., In the decomposition of cumene hydroperoxide under similar conditions*5,ss two additional reactions of the alkoxyl radical (PhCMe2, , O) may, , assume importance. Coupling of these radicals gives appreciable yields of, peroxide LVIII, whereas at somewhat higher temperatures, a radical elimina¬, tion reaction occurs, yielding acetophenone. Broadly speaking, the dimerization, ar For studies of autoxidations catalyzed by heavy-metal salts see for ^mple: Robertson, « */., TV,. Faraday Soc, 42, 201, 217 (1^); Bawn, £—", Mesrobian, et al., J. Am. Chem. Soc., 72,, (, >, ’, (1946). (b) Bell, et al.. Discussions, ss (a) Milas and Surgenor, J. Am. Chem. Soc., 68, 2Uo, w, Faraday Soc., 10, 242 (1951)., n, ch, 16 113 (1951); Fordham and WilKharasch, Fono, and Nudenburg, J. Org. Chem., 10, iid k, liams, Can. J. Research, B27, 943 (1949).
Page 727 :
Thermal Decompositions of Hydroperoxides and Dialkyl Peroxides, 2PhCMe20- —>, , PhCMe2—O—O—CMe2Ph, , 711, , (in HOAc at 100°), , LVIII, , O, PhCMe20-, , —> Ph—C—Me + Me-, , (in decane at 120°), , of alkoxy radicals to dialkyl peroxides at temperatures above 100° is rather rare,, but the breakdown of alkoxy radicals to ketones (or aldehydes) and alkyl, radicals is often observed. An analogous homolysis is of importance, for example,, in the decomposition of hydroperoxide (LIX).50, , LIX, , Although the decompositions of a number of secondary hydroperoxides, have been investigated,57 these reactions do not, as yet, form a coherent picture., The decompositions of primary hydroperoxides are of considerable interest; for, in addition to the “expected” homolytic decomposition, a second type of, decomposition, probably proceeding by a cyclic mechanism, has been found to, occur, and this yields elemental hydrogen as a major product.5* The decomposition, of w-butyl hydroperoxide at 85°, for example, has been studied with some, care.5~(i) The products with which we shall here be concerned are hydrogen,, butyraldehyde, butyric acid, propane, n-butanol, and rc-butyl formate., /H,, /C3H8, \C3H7CHO, + \ n-BuOH, n-Bu-OOH -> < C3H7COOH, l H—C—OBu, , ', , (, , O, , yields increased by irradiation, When the decomposition is carried out with the aid of ultraviolet light at 25°, (conditions known to favor homolytic cleavage of the O—O bond), the yields of, the first three products drop sharply, whereas, as indicated, the yields of propane, "'butyl alcohol> and "‘butyl formate rise. This suggests that the latter three, 2 ^chmidt and Fisher, J. Am. Chem. Soc., 76, 5426 (1954)., , and Waters!”., ” (a) Reiche and S fit, , Ts^f^afem’, , ?2, 333 °950); Rob«tson, , ^2458, , Durham, J. Am. Ch,m. Soc., 77, 5451 (1955);, , 4392', , 80,’ 327, 332, , (1958).', , M, , Wurster' and
Page 728 :
712, , Free-radical Reactions, , products arise from a homolytic sequence, the first step of which is the breakage, of the O—O bond in the hydroperoxide; but the details of this sequence have, not yet been worked out.53 The reaction in which hydrogen is formed exhibits, an induction period, which may, however be eliminated by adding traces of, aldehydes to the reaction mixture. The hydrogen formed is derived from C—H, bonds rather than from the O—H bond; for if the reaction is carried out with, n-Bu—OOD, H2 rather than HD or D2 is evolved. Moreover, since the reaction, mixture exhibits none of the tests characteristic of atomic hydrogen, it may be, assumed that the hydrogen molecules are formed as such, rather than by com¬, binations of free atoms. The bulk of evidence points to the intervention of com¬, pound LX, a sort of “hemiacetal” derived from a mole each of aldehyde and, hydroperoxide, which may decompose via the cyclic transition state LXI to, hydrogen, butyraldehyde, and butyric acid. Indeed, a-hydroxyperoxides similar, , °T°., ^, , 0-0, /, \, PrCH2OOH + PrCHO, , PrCH,, , CH-Pr, , /C, Pr-CH ), , \, , HO/, , PrCHO, C-Pr, , ry, , \, , H 4 H, , OH, , PrCOOH, , H,, , LXI, , LX, , to LX have been prepared and have been found to decompose readily, evolving, hydrogen and leaving behind a mixture of aldehyde and carboxylic acid. In the, absence of added aldehyde, it is likely that n-butyl hydroperoxide is converted, slowly to butyraldehyde, thence to intermediate LX, by the following sequence:, OH, , PrCH2OOH -> PrCH—OOH, , PrCHO PrCH2°^ PrCH2—OO—CHPr, LX, , The slow build-up of the concentrations of butyraldehyde and intermediate LX, to their respective steady-state values accounts for the observed induction period., The decompositions of dialkyl peroxides appear to be considerably less complex than those of hydroperoxides. One of the most stable of these is d.-i-butyl, peroxide (Me3C—OO—CMe3), the breakdown of which has been studied by a, ts The following sequence of steps accounts for the observed products, , BuO-, , Pr- + HCHO —» PrH + H—C—O, -OH-, , BuOOH--* BuO--, , BuOH ■+• BuOO*, , As indicated, the fate of the BuOO- radicals £“^n, to form 2BuO* + 02 (a reaction analogous to reaction, of O2 have been isolated from the reaction mixture., , ????, , PP, , H—C—OBu
Page 729 :
713, , Thermal Decompositions of Hydroperoxides and Dialkyl Peroxides, , number of workers.** This decomposition is not a chain process, either in the, vapor state or in hydrocarbon solvents, for it displays clean first-order kinetics, and is not inhibited by addition of 02 or NO. The decomposition of the vapor, gives mainly acetone and ethane, and there is little doubt that the reaction, sequence occurring here is, 12Me2C—O, , slow, , Me 3C—O—O—CMe 3, , >, , 2Me3C—O, , 12Me* —> C2H6, , Minor amounts of methyl ethyl ketone and methane may be attributed to attack, of methyl radicals on the acetone formed., CH3. + H3C—C—CH3-+ CH4 + -H2C—C—CH3-^, , I, , II, , o, , o, h3c— ch2—c—ch3, , II, , o, In a hydrocarbon solvent, the Me3CO* radical may, instead of decomposing, to acetone and Me-, simply extract a hydrogen atom, forming /-butyl alcohol., Me3CO* + HR -> Me3C—OH + R., Generally, the ratio of acetone to /-butyl alcohol obtained from the reaction in, such a solvent increases with temperature, indicating that the activation energy, for the breakdown of the Me3CO radical is greater than that for hydrogen, extraction. As may be expected, the acetone to /-butyl alcohol ratio depends also, on the nature of the solvent, in particular on whether the hydrogen atoms avail¬, able are bound to primary, secondary, or tertiary carbons. Thus, when the, decomposition of di-/-butyl peroxide is carried out in cumene (PhCHMe2), in, which tertiary C—H linkages are available, the ratio of/-butyl alcohol to acetone, in the product is about 4; whereas if the decomposition is carried out under, similar conditions in /-butylbenzene (which has only primary and aromatic, C, , H linkages), this ratio drops to 0.6., It appears that the decomposition of di-t-butyl peroxide as a pure liquid, , proceeds, at least in part, by a chain mechanism,since the specific rates for, decomposition of the pure peroxide are several times those for the peroxide in, hydrocarbon solvents. Moreover, in addition to /-butyl alcohol, acetone, , and, , methane, there may be isolated from the reaction mixture large quantities of, and, , Ota,, , ., , 74, 6005 (1952); Williams, , Oberrifrht, , ’ n, , T, , been reviewed by Frost and Pearson Kinetics and M, York, 1953, p. 310-317., ’, ", , ’, , and SzWarc> J' Am• Chem. Soc,, a*, , 8} V9<? (1956)- This reaction has, John Wiley and Sons, Inc., New, , md Mecha™™,, , 95, , 70, 88 0,48); LoSsing, , Bell, Rust, and Vaughan, J. Am. Chem. Soc., 72, 337 (1950).
Page 730 :
714, , Free-radical Reactions, , isobutylene oxide, LXX, suggesting the sequence, Me3C—OO—CMe3 -f Me3CO, etc., , •CH2CMe2—OO—CMe3 -> CH2-CMe2 + .OCMe3 —>, , \ /, O, , LXII, /-BuOH, in addition to the “usual” decomposition yielding acetone and the methyl radical, (which may also remove hydrogen from the peroxide). The intramolecular, radical displacement on oxygen leading to epoxide LXII is unusual; this step, may be regarded somewhat skeptically until further examples of such displace¬, ments are found., The decompositions of a number of additional dialkyl peroxides have been, studied," but these appear to be more complex than that of di-/-butyl peroxide, and need not be discussed here., , Decompositions of Diacyl Peroxides, The decompositions of diacyl peroxides are of considerable importance, for such, peroxides are extensively used as chain initiators. The most familiar of these is, benzoyl peroxide /Ph—C—OO—C—Ph\, the decomposition of which in most, , V, , O, , o, , /, , organic solvents proceeds by at least two distinct paths.97 The first of these begins, with a “spontaneous” unimolecular breakage of the O, PhC—O—O—CPh, , II, , 2Ph—C—O-, , o, , o, , O, , O bond,, , whereas the second path (induced decomposition) presumably involves attack, on the O—O bond by a radical present in solution., R. + PhC—O—O—CPh —► R—O—CPh + -O—C—Ph, , A, , o, , O, , O, , ..See, for example: M Takezaki anc^^■•(^hyl(pcJJc);, S^boTdlnd Vaughan"./^., , ..For, lett, J. Am. Chem. Soc., 68, Io»o {, , ),, , cL. See., 72, 338 (1950), », , (on /-butyl alkyj peroxides);, , S ct,W, 68, ,976 (.946); 69, 0950); (d) Hammond and
Page 731 :
Decompositions of Diacyl Peroxides, , -, , 715, , The fate of the benzoyloxy radicals depends largely upon the solvent used. If, secondary or tertiary hydrogen atoms are available, they may be “extracted”, from the solvent molecules by PhCOO* radicals, forming benzoic acid and, radicals derived from the solvent (which may, in turn, induce the homolysis of, additional peroxide molecules). With less active solvents, decarboxylation of the, benzoyloxy radicals is likely to occur,, O, II, Ph—C—O* -» Ph- + CO2, and the resulting phenyl radicals may dimerize or may attack the solvent or per¬, oxide. If the solvent is aromatic, free-radical arylation by the phenyl radical may, occur. If the solvent is an olefin, polymerization may be observed. When the, decomposition of benzoyl peroxide is carried out in damp, , CCI4 in, , the presence of, , elemental iodine, the iodine reacts with the benzoyloxy radicals very soon after, they are formed, suppressing the induced decomposition. The resulting product,, benzoyl hypoiodite, PhC—OI, may be hydrolyzed to benzoic acid, which may, O, then be recovered in almost quantitative yield,S7(rf) thus indicating that virtually, all acts of “spontaneous” decomposition proceed through PhCOO* radicals., , / o, , \, , \PhCO, , /2, , o, , o, , 2Ph—C—O* X 2Ph—C—OI XX 2PhCOOH + 2HOI, , In relatively inactive solvents, such as aromatic hydrocarbons and CC14, the, induced decomposition very probably proceeds largely by the following path:, , /—\, , f—\, , — CO2, , B22O2, , PhCOO-> Ph-> PhOBz + BzO*, , -CO2, , -» etc., , which can be shown (Ex. 3d) to lead to a rate law for the overall decomposition:, -^(Bzo02), —2- = *(Bz202) +, , k\BZ2O2)3'*, , The rate constant for the “induced contribution,”, rn, CCh),, , (IV), , is nearly the same in a num-, , rKrS.(fcr example’ benzene’ ni‘rob“2^’ cyclohexane, and, > a further mdicat.on that radicals derived from the solvent are not, , .nvolved m these cases. However, when the decomposition reaction is trans¬, ferred from benzene to acetic acid, the rate of induced decomposition increases, about twelvefold, suggesting that radicals from the solvent have begun to enter, he picture. In aliphatic ethers and in alcohols having a-hydro^en atoms decom, shT" d To faStCr S,m> indiCating tha‘ ‘he indu“d decomposition has ovet, shadowed the “spontaneous.” Here, the chain reaction is probably propagated
Page 732 :
716, , Free-radical Reactions, , by such solvent-derived radicals as R—CHOH and RCH—OR'. Once again,, the rate-law is determined by the nature of the termination step(s); decomposi¬, tions of benzoyl peroxide in alcohols or ethers follow, to a good approximation,, first-order kinetics, which may be shown (Ex. 3c) to be consistent with termina¬, tion by the following reaction:, , R—CH—OR + BzO- -> RCH(OBz)—OR, Moreover, such benzoyloxy ethers may be isolated from the decompositions of, benzoyl peroxide in aliphatic ethers.97(6)-9S In cyclohexane also, induced com¬, position of benzoyl peroxide is caused by radicals derived from the solvent.97(l,), The effect of high pressures on the decomposition of benzoyl peroxide (in, acetophenone) has recently been studied.97(e) Just as activation energies, AHt,, may be obtained from the observed variation of reaction rates with temperature,, so may “activation volumes,” AV* (volume changes in going from reactants to, transition state), be obtained from the observed variation of reaction rates with, applied pressure." Unimolecular homolyses are generally retarded by applica¬, tion of high pressures, whereas bimolecular processes (for example, radical addi¬, tions and displacements) in which the transition state is more compact than the, reactants, are generally accelerated.700 The decomposition of benzoyl peroxide in, acetophenone is (as would be expected) gradually retarded as applied pressure, is increased, but at very high pressures decomposition becomes sharply acceler¬, ated. The latter effect is almost certainly due to the induced decomposition that, presumably requires a number of bimolecular processes in each kinetic chain., Bearing in mind that the breaking point in the benzoyl peroxide molecule is, the bond between two very electronegative oxygen atoms, we might expect the, decomposition of the peroxide to be accelerated by substituents that supply, further electron density to this reaction site and to be retarded by substituents, which withdraw electron density. This effect is indeed observed;97-707 moreover, the rates of “spontaneous” decomposition of meta- and /xzra-substituted benzoyl, peroxides have been correlated satisfactorily using the Hammett equation (p., *» The analogous product resulting from the decompositions of Bz202 in a primary alcohol, would be RCH(OH)OBz, which decomposes spontaneously to a carbonyl compound (., this case an aldehyde) and benzoic acid (see, for example, Gehssen and Hermans, Ber., 5i1, 765 (1925)), , ’, , J., , For the reaction of diacyl peroxides with phenols, which is almost certai, , ,, , y, , vie thp “fathering together” of solvent molecules about the separating cha g, *LS?Z2,X Tran, Faraway, 5,, ,497 (,955). and Brower,, , Am. Chem. Soc., 80, 2105 (1958))., r,, «, ‘oo See, for example, Walling and Pellon, J. Am Cfum Soc, , 7q, , a-jjc., , 79, 4776,, , tot Blomquist and Buselli, J. Am. Chem. Soc., 73, 3883 (1951)., , 4782 (1957)., (
Page 733 :
Decompositions of Diacyl Peroxides, , 717, , 220), suggesting that polarity is the dominant factor leading to rate difference, in this series. On the other hand, it appears that electron-withdrawing sub¬, stituents accelerate the induced decomposition of these peroxides in dioxane,0'(c), probably because such substituents increase the electrophilic character of the, derived benzoyloxy radicals, hence the ease with which they withdraw hydrogen, atoms from the solvent.*02, The decomposition of acetyl peroxide,, , CH3—C—O—O—C—CH3,, , II, , o, , at, , I!, , o, , moderate temperatures is similar in a number of respects to that of benzoyl, peroxide; again, both a “spontaneous” and an induced decomposition occur.*05, The chief difference between decompositions of the two peroxides is the very, short lifetime of the acetoxy radical, which, unlike the benzoyloxy radical, is not, stabilized by conjugation between the carboxy group and the benzene ring., Even in very reactive solvents, acetyl peroxide decomposes to give virtually, quantitative yields of C02, and acetoxy radicals cannot, at ordinary tempera¬, tures, be trapped by \2.10Jt Acetyl peroxide may thus be considered an excellent, source of free methyl radicals in solution., Methyl radicals attack most organic solvents, readily extracting hydrogen, atoms bound to aliphatic, allylic, or benzylic carbon atoms; if no such hydrogens, are available, attack on carbon-halogen bonds may occur.*05, CH3- + PhCH3, , -> CH4, , CH3- + CH3COOMe -> CH4, CH3- + CC14, , + PhCH2*, + • CH2COOMe, , -> CH3C1 + CC13*, , Such reactions are rapid, but their relative rates may be compared by allowing, acetyl peroxide to decompose in a known mixture of CC14 and a second solvent, that may furnish hydrogen atoms, then comparing the quantities of methane, and methyl chloride which are formed., , Revs. VII, 198 (1953)., , see Trotman-Dickenson, Quart.
Page 734 :
718, , Free-radical Reactions, , In this way, it may be shown, for example, that methyl acetate reacts with, methyl radicals over 20 times as rapidly as does CCI4, which, in turn, reacts 16, times as rapidly as does methyl benzoate.705, Moreover, it is found that chloroform, acetaldehyde, and methyl acetate are, attacked by CH3* radicals much more readily than are the ordinary aliphatic, hydrocarbons. More generally, when the attacking radical is CH3*, extraction of, a hydrogen bound to an aliphatic carbon is facilitated both by an a-halogen, atom and an a-carbonyl group, although, as we have seen, both such substituents, tend to retard hydrogen extraction by a chlorine atom. The essential difference, between the two types of attack is that the chlorine atom, besides being a free, radical, has electrophilic character, whereas the methyl radical has not. The ease, with which a substrate is attacked by a methyl radical is probably closely related, to the stability of the resulting radical intermediate, and, as has been shown, (p. 697) both an a-carbonyl and an a-halogen substituent tend to stabilize such a, radical., Methyl radicals also attack aromatic rings, but the nature of the reaction, is not yet clear; it is generally assumed to be some sort of an addition reaction, that does not yield methane. Thus, if acetyl peroxide is allowed to decompose in a, mixture of an aliphatic hydrocarbon (which yields methane) and an aromatic, substance (which does not), the yield of methane should drop as the reactivity of, the aromatic increases, for more methyl radicals undergo addition rather than, displacement. In this way it may be shown, for example, that anthracene is much, more readily attacked by methyl radicals than is phenanthrene, which is, in, turn, more readily attacked than biphenyl.707, The decompositions of a number of additional diacyl peroxides have been, studied,705 including peroxides XXXVI and XLI, which we have considered in, relation to the stereochemistry of free radicals (p., , ). Certain of these studies, , 686, , emphasize that just as a distinction between ionization and dissociation must, sometimes be drawn in polar reactions, so also must a distinction between, homolysis and dissociation be made in free-radical reactions. More specifically,, when a molecule within a “solvent cage” undergoes homolysis, the resulting, radicals remain within this cage for a significant time interval, during which they, Edwards and Mayo, J. Am. Chem. Soc., 72, 1265 (1950). Reactivity ratios obtained in, this way are in only fair agreement with corresponding ratiosSee, example, Trotman-Dickenson and Steacie, J. Chem. Phys., 18, 1097 (1950), 1, , ,, , (, , iUSO ^Levy and'szwarc, J. Am. Om. Soc, 77, 1949 (1955). These -ta -ej*-*, , nf, , the, , tvDe described to, , obtain, , quantities that they, , aromatics^However^ such parameters can have little, , and Szwarc, J. Am. Chem. Soc., 78, 3322 (1956)., , RemS, , term, , “methyl, , affinities, , for the, , vanou, , -£, , -Smid,, , <”57>'
Page 735 :
Decompositions of Diacyl Peroxides, , 719, , O, , II, (Et-CHMe—C—0)2, XXXVI, , XLI, may be quite likely to recombine.*05 The dissociation-recombination sequence, may sometimes be detected if one or both of the radicals are altered in some, manner during this time interval. With acyloxy radicals, this “alteration” may, be a decarboxylation that, in some cases, takes place very soon after the initial, homolysis. When, for example, propionyl peroxide (XLI 11) is allowed to, decompose in hydrocarbon solvents, the propionoxy radicals first formed rapidly, decarboxylate to ethyl radicals, and one of the products is, as expected, rc-butane,, formed by combination of ethyl radicals.108(f>) Now, the fraction of ethyl radicals, , o, I, , o, I, , —> | 2, , C-O-O-C-Et -, , O, I, , ', , Et-C-O-1, , -2CO,, 2 Et-, , -> CLH 10, , XLIII, , that form butane has been found to be very nearly independent of the initial peroxide, concentration, whereas if the ethyl radicals were to become independent of each, , other before combining, the extent of recombination should rise with peroxide, concentration. We may then conclude that the two ethyl radicals that combine, to form a given butane molecule are, in almost all cases, derived from the same, peroxide molecule; that is, the entire sequence shown above occurs in the same, solvent cage (represented by the dotted rectangle). Moreover, the addition, of quinone, an excellent “scavenger” for free alkyl radicals, does not affect the, yield of butane from this decomposition, again indicating that the ethyl radicals, giving butane do not become free of their solvent cage., In the decompositions of a number of diacyl peroxides (but not that of, benzoyl peroxide) the decarboxylation follows the initial homolysis so closely that, It may be asked whether these two steps are not concerted. While this question, cannot be answered definitively for diacyl peroxides, evidence for a “concerted”, mechanism has been obtained for the decompositions of the /-butyl esters of a, mber of peroxy ac,ds."<> The relative rate constants for a few of such decom-
Page 736 :
720, , Free-radical Reactions, , positions (in chlorobenzene at 60°) are given below:, f-BuO—O—C—R —► f-BuO + C02 + R-, , O, R—, , —ch3 —CH2Ph, , k/k BuOjAc, , 1, , 300, , —CMe2Ph, , —CPh2Me, , 40,000, , 80,000, , Quite obviously, the rate of decomposition rises sharply as the stability of radical, R* increases. It is difficult to see why this should be so if the rate-determining, step were merely the breakage of the O—O bond in the peroxy ester. It is, how¬, ever, quite consistent with a mechanism in which the breakage of the O—O, bond and the C—R bond in the ester are simultaneous; for if this were the case,, factors which stabilize radical R* should also stabilize the transition state leading, to, , it.111, , Arylation of Aromatic Rings^2, The decomposition of benzoyl peroxide (and substituted benzoyl peroxides) in, aromatic solvents results mainly in arylation of the solvent—that is, the formation, of substituted biphenyls. In analogy with the free-radical substitution reactions, thus far considered, we might anticipate that arylations proceed by a sequence of, steps such as, -CO2, , ( ?, , ), , \PhC—o—/, , ArH, , * Ar—Ph, , Bz202 —> BzO-> Ph-> Ar-, , or Ph-, , 111 It is probable that the decompositions of peroxy acids proceed by a different path., Rate data are available at present for the decomposition only of peroxylauric acid, C11H23COiH (Parker Witnauer and Swern, J. Am. Chem. Soc., 80, 323 (1958)). The decomposition o, this acid in beAzene, mainly to elemental oxygen and lauric acid exhibits first-order kinetics;, moreover, such solutions are very poor free radical initiators, , It is doubtful then that this, , decomposition is homolytic. A cyclic mechanism, proceeding through transition state LXIV,, is likely and is consistent with the greatly negative entropy of activation for this reaction., O-H, , /, R-C, , R-C, , 0-^0:, , \, , +, , 6:, , O, , LXIV, The five-membered ring (chelated) monomer b known to be ^Predominant formofPe™V, acids in nonhydroxylic solvents (Rittenhouse, Lobunez, Swern, and Mil, , ,, , •, , 8°’ 4m For Preview of homolytic aromatic arylation reactions, see Augood and Williams,, Chem. Revs., 57, 123 (1957).
Page 737 :
Arylation of Aromatic Rings, , -, , 721, , There are two lines of evidence against a mechanism of this sort. First, an aryla¬, tion proceeding through an Ar- radical derived from the solvent ArH should, yield, as one of the by-products, compounds of the type Ar2, which in fact have, not as yet been isolated. (More specifically, arylations carried out in chloro¬, benzene give no appreciable amounts of dichlorobiphenyls, and arylations in, nitrobenzene appear to give no dinitrobiphenyl.), Secondly, it has been found that such arylations are not subject to a hydro¬, gen-isotope effect. Thus, when 2,4-dinitrotritiobenzene (LXV) is arylated with, benzoyl peroxide, equal quantities of the labeled and unlabeled 2,4-dinitrobiphenyl are formed, whereas if attack had begun with the breakage of a C—H, (or C—T) bond, removal of a hydrogen atom should be favoredd/5 The absence, , 02N\n^%n/'N02, , 02N^^N02, BZ202, , +, , + GO,, , T^^^ph, LXV, , ~"V", , equal amounts, of a hydrogen-isotope effect also excludes a direct displacement mechanism for, arylation;, Ph- + Ar—H -> Ph—Ar + Hfor with this mechanism also, attack on the ring involves breakage of a C—H, bond. Moreover, since such a displacement sacrifices a C—H bond to form a, (somewhat weaker) G, , C bond, it is energetically unfavorable., , The most likely mechanism for arylation then appears to involve prelimi¬, nary addition of a phenyl radical to a solvent molecule, after which a hydrogen, atom is removed from the resulting radical LXVI by another radical or by a, second molecule of peroxide. If this scheme is, in the main, correct, we should, , Ph* + C6Hr AT, , X, , n, ‘, -“““cnzauon ot radical LXVI amone the, products. While dtmers of this sort have not ye, been isolated in large quantity, (alJu'ht, subject to hydrogen-isotope effects; however for an evamnT'V, ytlC -^tion, see Denney and KtanchST**, 80, 32^9 (1958), , 41°! ('958>- " * Probable, lft,ons are likewise not, “ 1SOt°Pe *** * h°mo'
Page 738 :
722, , Free-radical Reactions, , from such reactions, they may well comprise a portion of the “high molecular, weight, tarry material, , which is always formed in considerable amount, along, , with the substituted biphenyls, in such decompositions."4, The reader will recognize that although the incoming phenyl group is, shown entering the position para to substituent —X in the sequence above, ortho, and meta arylation is also possible. Indeed, it is now recognized that the arylation, of almost every monosubstituted benzene gives sizable amounts of all three, isomeric products. Typical isomer distributions resulting from a number of, phenylations with benzoyl peroxide are listed in Table 16-3. Quite obviously,, Table 16-3. Isomer Distributions Resulting from Phenylations, with Benzoyl Peroxide'*5, Substrate, , Percent ortho, , Percent meta, , Percent para, , 17, , Anisole, , 71, 24, 67, , 50, 18, , 12, 27, 15, , Chlorobenzene, , 62, , 24, , 14, , Biphenyl, Nitrobenzene, , 49, , 23, , 60, , 9, , 29, 32, , Toluene, /-Butylbenzene, , the effects governing orientation in these arylations are very different from those, governing orientation in aromatic nitration, halogenation, and other electro¬, philic substitutions. All substituents, regardless of their electronic character,, appear to direct the incoming phenyl group preferably to the ortho and para, positions. (In many cases, the yield of meta compound is somewhat greater than, that of para compound, but the meta to para ratio never is as great as 2, the value, that would result from random substitution.) Moreover, it may be shown by, competition experiments that nearly all ring substituents, irrespective of their, nature, facilitate phenylation."5-"5 Such rate increases are small, however, the, largest being the fourfold acceleration produced by the nitro group. This is in, marked contrast to electrophilic nitration and halogenation, where incorpora¬, tion of a single substituent may raise or lower the rate of attack on the ring by, several powers of ten (p. 428). In the same way, the orienting effects of substit¬, uents in phenylations are seen to be much less pronounced than in electrophilic, "4 See, however, DeTar and Long, J. Am. Chem. Soc., 80, 2742 (1958)$, us These values were determined mainly by Hey and co-workers JChem Soc ^1953,, 3412-1954, 3352; Disc. Faraday Soc., 14, 216 (1953); J. Chem. Phys.,, , “nLne is much S' readily, , luaSTd, , ,, , by .he tripheny.me.hyl radical than is benaene itself (see
Page 739 :
Arylation of Aromatic Rings, , -, , 723, , substitutions where, in many cases, only a few percent of the “less favored”, product is formed., From the small increases in phenylation rates brought about by ring substit¬, uents, and from the relatively slight orienting effects of such substituents, we, may infer that differences in activation energy due to the incorporation of substit¬, uents into the substrate are small, considerably less than 1 kcal per mole. It, would therefore seem that any proposed correlation between structure and re¬, activity in this series should be taken with reserve. Nevertheless, it is often sug¬, gested that substituents lying ortho or para to the incoming phenyl group in the, radical intermediate stabilize this intermediate (and the activated complex lead¬, ing to it), whereas no such stabilization is possible for meta substitution.-^7 (The, , MeO, , Ph-, , MeO, , +, , reader will note the similarity between structures LXVI I', LXVIII', and LXIX', Benkeser 3,nd Schrocdcr, J. Am. Chcrn, , Sor, , fin, , electron withdrawal frora the -.rival«by, , 4 /ioqqw t* •, the, , rin^pToM, , rich oxygen atotns in’,he nitro group rfiSCf, , in^he, , ‘T f, , *' '"Ctr0"-, , o:, PhN02 -(- *CPh3 ;=± Ph—N, , CPh,, , O:, , j", , for, A, nber °f ^ttemptS to treat this problem theoretically have been carried out. See,, for example, Brown, Quart. Revs., 6, 63 (1952)., y have
Page 740 :
724, , Free-radical Reactions, , on one hand, and the structures II', III', and IV' on page 675 on the other. The, latter three were employed in discussing the influence of substituents on the dis¬, sociation constants of hexaarylethanes, a situation in which the effect of the, various substituents likewise did not depend directly upon their electron-at¬, tracting or electron-donating properties.) The relatively high yield in many cases, of m^a-substituted products seems a little puzzling. However, it should be, remembered that the observed isomer distributions among the substituted, biphenyls need not reflect accurately the relative rates of attack at the various, positions, for the intermediate radical, formed by addition of Ph* to the substrate,, need not be converted to a biphenyl. It may dimerize, may disproportionate, or, otherwise be led astray; and the radical intermediate in ortho or para substitution, may (for a reason which is not yet clear) be somewhat more likely to become, involved in side reactions than the intermediate in meta substitution. Further, investigation of this problem is obviously desirable., Naphthalene is much more readily attacked by radicals than is benzene, and, substituted naphthalenes are more active still.“s(o) In fact, when benzoyl per¬, oxide is decomposed in the presence of naphthalene or substituted naphthalenes,, the benzoyloxy radicals attack before significant decarboxylation can occur, and, benzoyloxynaphthalenes, rather than phenylnaphthalenes, are the major prod¬, ucts. Benzoxylation, rather than phenylation, occurs also with a number of, polynuclear aromatics (for example, anthracene and 1: 2-benzanthracene)., , ), , Decomposition of Azo and Diazo Compounds, Aliphatic azo compounds, R—N=N—R', may be prepared by oxidation of, dialkylated hydrazines. The simplest azo compound, azomethane, decomposes, unimolecularly in the vapor phase at temperatures above 300°, yielding methyl, radicals and molecular nitrogen//s(o), H3C:N=N:CH3 -» H3C- + N=N + -CH3;, , rate = *1(MeN=NMe), , In the absence of other substances, the methyl radicals dimerize, yielding, ethane. As expected, the substitution of alkyl groups for methyl hydrogens in, azomethane stabilizes the resulting alkyl radicals, hence facilitates the decom¬, position. The same effect is even more pronounced when ary groups aie so, substituted. Typically, then, the energy of activation for the decomposition o, azomethane is found to be 50 kcal per mole,"»<*> that for the decomposition of, Me.CH—N=N—CHMej is 41 kcal per mole,"»»> whereas that for the decom.., , x, , (a), , Dannley and, , Gippin, J. Am. Om., , x a-. **., , Soc,, , 74, 332 (1952). <» Roil,, , and Waters,, , * <*> Ramspcrg'r'X, , Soc > 50,714 (1928). (c) Cohen and Wang, ibid., 77, 2475 (1955)., , Cto"
Page 741 :
Decomposition of Azo and Diazo Compounds, , -, , 725, , position of Ph2CH—N=N—CHPh2 (which decomposes readily in inert solvents, at 65°) is only 27 kcal per molediff(c) 1X0, In the same way, a-cyano groups may act as radical stabilizers, thus greatly, easing the decomposition of aliphatic azo compounds., , Indeed, substituted, , azoacetonitriles (for example, LXX), which, in many cases, may be readily pre¬, pared from ketones, HCN, and hydrazine, have become very important as, ! radical sources., RoC—N=N—CRo -» 2[R2C—C=N: <-> R2C=C=N.] + N2, I, N=C, , I, C=N, LXX, , The decompositions in solution of over thirty of such azonitriles have been, studied kinetically;757 almost all exhibit clean first-order kinetics with no signifi¬, cant induced decomposition. As may be expected, the specific rate for the decom¬, position of a given azonitrile is quite insensitive to solvent. A number of decom¬, position rates (in toluene at 80°) are compared below:, Me2C—N=N—CMe2, N=C, kX 104, sec —i, , I, , C=N, , I, , I, , N=C, , 1.53, , I, , C=N, , |, , N=C, , 0.84, , C=N, 1.15, , CN, , PhCMe —N=N— CMePh, I, |, N=C, G=N, k x IQ4, , (n-Pr) 2C—N=N—C(n-Pr) ■>, , Et2C—N=N—CEt2, , NG, , CN, , NG, , 'N=N', , “very fast5’, , 0.73, Note that the nature of the alkyl groups bound to the a-carbon has little effect on, the rate of decomposition, but that incorporation of a-phenyl groups, which, , “° The observation that the diisopropyl compound, Me2CH—N=N_CHM>, poses much more rapidly than the monoisopropyl compound' McCH-N-N, perger, J. Am. CW Soc. 51, 2.34 (.929),,, _, , R~N=N~-R, , LdJ, , U, , t, , to ru.?ouu^'ep" e ££££(RamS', , slow, , f, , t, , + R- + •N=N—R-> 2R- -f N2, , for the decompositions of aliphatic azo compounds, , Fnr, , m, , ., , rate-determining step involved breakage of only one G-N bond, , th^, , n* if the initial, , determined chiefly by the identity of the more stable of, alkyl radical lost first). Since, on the contrary the, f H, alkyl radicals, it may be supposed that therate, , ’ m ?V, rate Sh°uld be, ° alky radlcals (that is, the, influenced by both, , C-N bonds, resulting in thfconcerted formation of two 2, wise mechanism cannot be ruled out for the H, .^ky, , breaka*c of, radlcals- However, the step-, , PhcOTN=N-CPh, (ace Davi« He^Lwd^TlT” f 7^ “ “founds L, See, for example, (a) Overberger ,lil, ,, T '™,; S"c" 1956> '•397)., (W51);75,2078 (1953); 76 2722, 6186 (1954)’, W Lewis, LewUaLM'',b’’, 2661, (1949>, 1 79., 4880, t, , (Oj, and Matheson,, ibid.,, 71, 747, (1949).
Page 743 :
Decomposition of Azo and Diazo Compounds, , LXXVI, , ^ GHPh—N=N—CHPh^, (CH2)8, , CHPh— CHPh\, (CH2) &, , (CH2)s, , CHPh—CHPh ^, , CH Ph—N=N—CH Ph, LXXVII, The, , decomposition, , of, , unsymmctric, , phenylazo, , compounds, , such, , as, , Ph—N=N—CPh3 yields phenyl radicals,'*5 and if such decompositions are, carried out in aromatic solvents, the phenyl radicals may attack the solvent in, the same manner as do phenyl radicals derived from benzoyl peroxide (p. 720)., Aryldiazonium hydroxides, Ar—N=N—OH, decompose into aryl radicals, and, probably to OH* radicals as well., Ar—N=N—OH -> Ar* + N2 + OH*, Since these diazonium hydroxides are prepared with great ease merely by treat¬, ing solutions of aryldiazonium salts with base, they are extremely useful sources, of aryl radicals and are often used to convert benzene derivatives to substituted, biphenyls (the Gomberg-Bachmann reaction).1*6 Quantitative study of this, reaction is difficult, for it is generally carried out in a two-phase system. More¬, over, it is not a clean reaction; yields of the substituted biphenyls often drop, below 30 percent and large amounts of tars are produced. The fate of the OH*, radical (if indeed it is formed) has not yet been determined., Arylations may be carried out much more cleanly using substituted N-nitrosoacetanihdes /Ar, , \, , N, , Ac\, which generally yield nitrogen quantitatively when, , N=0 /, , allowed to decompose in aromatic solvents. Phenylations of toluene, nitro¬, benzene, cumene, pyridine, and 1-butylbenzene using this reaction give, to, within experimental error, the same ratios of isomeric biphenyls as are obtained, in the corresponding phenylations using benzoyl peroxide.'” Here again then, aryl radicals are intervening. The rates of decomposition of substituted N-nitrosoacetamhdes have been found to be very nearly independent of the solvent under¬, going arylation, indicating that the latter does no, participate directly in the, , (1955).SeC’ f°r CXampIe’ Cohen and Wa"B-, , J-, , Am-, , «"»■ ■Sec-. 75, 5504 (1953); 77, 3628, , John Wiley and Sons, Ine^New"?^' W4 ™ 224"d H°ffmann in 0rS«”“ Reactions, Vol. 2,, , (,957>', , a summary or isomer ratios
Page 744 :
728, , Free-radical Reactions, , rate-determining step.'** Moreover if /3-naphthol is added to a solution in which, the N-nitrosoanilide is decomposing, a portion of the latter reagent is converted, to dye LXXIX (almost surely a heterolytic reaction), but the rate of disappear¬, ance of the nitroso compound is virtually unchanged. It is thus extremely likely, that the nitroso compound is, in the rate-determining step, being converted to a, diazo acetate, LXXVIII, which may react rapidly, either homolytically or, (under appropriate circumstances) heterolytically. The exact manner in which, , Ar, \, , N, / X, , «N2 + AcOH + Ar—Ar', Ar—N=N—O, , Vj, °^, c', , xc=0, , ✓ \, , O, , Me, , Me, + AcOH, , LXXVIII, , (rate determining), , LXXIX, (product determining), , homolytic arylations occur in such cases is not yet clear. The first path that comes, to mind involves the decomposition of diazo acetate LXXVIII to molecular, nitrogen and to acetoxy and aryl radicals,, Ar—N=N—OAc -> N2 + OAc* + ArLXXVIII, followed by the attack by Ar* on the substrate. However, very little C02 is, evolved in such arylations, whereas, as we have pointed out (p. 717), acetoxy, radicals are thought to undergo decarboxylation readily. To account for this, apparent inconsistency, it may be supposed that for a significant time interval, after the above homolysis has occurred, the resulting radicals occupy a single, “cage” of solvent molecules. One of these molecules is presumably attacked by, the aryl radical, whereupon the acetoxy radical, before breaking out of its, “solvent cage,” attacks the composite radical, forming acetic acid (which is found, to be a major product)., , + N2, ns Huisgen, Horeld, and Nakaten, Ann, , ., 562,, , 137 (1949);, , 573,, , 163 (1951).
Page 745 :
The Sandmeyer Reaction, , 729, , The Sandmeyer Reaction, The conversion of aromatic diazo compounds to aryl halides by use of cuprous, halides (the Sandmeyer reaction) also appears to proceed through aryl radicals., Such radicals may be diverted from solutions in which this reaction is occurring, by use of nitrobenzenei25(o) or iodine.'*a(6) Moreover, with care, such solutions, may be used to initiate the polymerization of acrylonitrile./gg(c) Thexeaction is, first order both in diazonium ion and in “cuprous chloride” (actually CuCly)/5, indicating a rate-determining reduction of the aryldiazonium ion with uni¬, positive copper., ArNt + CuCkr -> Ar- + N2 + CuCl2, However, the reaction is retarded by a large excess of added HC1, suggesting, that the “higher” complex, CuClp3 is less effective (or ineffective) in converting, the diazonium ion to an aryl radical. Essentially no univalent copper is con¬, sumed in the overall reaction ;;S0(6) thus, the copper oxidized in the proposed, rate-determining step must be reduced in a subsequent step. The following, sequence then is consistent with what is known about the Sandmeyer reaction, (involving, , chloride):, slow, , Ar—N=N+ + Cl—Cu—Cl-, , fast, , > Ar- + Cl—Cu—Cl -> Ar—Cl + CuCl, -+*n2, +n2, , The transfer of an electron from the CuCl^ ion to the ArN^ ion in the initial, step probably occurs through a “chloride bridge” in an activated complex of the, type Ar, , N=N • • Cl—Cu—Cl. Electron transfers of this sort are well recog¬, , nized in inorganic chemistry.131 The second step in the indicated sequence is, merely a radical displacement by Ar- on the chlorine., If an attempt is made to carry out the Sandmeyer reaction in a solution, containing styrene, acrylonitrile, or a similar unsaturated compound, attack by, the aryl-radical intermediate on the double bond may occur, with the more, stable of the two possible radical adducts (for example, radical LXXX) being, formed. This aralkyl radical then reacts with cupric chloride (or a chloro com¬, plex thereof), resulting in the formation of a “chloroarylation” product (for, example, LXXXI)., , CH2, , ch, , cn -> ArGH2 CH, LXXX, , CN, , ArCH2—CHC1—CN + CuCl.w*, LXXXI, , (*) Kochi,^^r.^, , 3801 °956); 22’1070 <1957)', , J.^.C»L.Si,.,7»,aiM?a956f' w DictCrma9n WS48: Q7i', :, , *2, , Taubc!f, , a4n48, , XK.s . often caUed .he Meecwein ceac.ion (See.Vor, , 558 (195^' <»> Kochi., , "ftsT"'^ 0 9M>Mecewein, Buchnec, and
Page 746 :
730, , Free-radical Reactions, , PART III — Additions and Rearrangements, of Free Radicals, , Homolytic Additions, Energetic Requirements, The type of free-radical addition reaction that has been most intensively studied, is addition polymerization—that is, the build-up of long molecular chains from small, unsaturated molecules, often with the aid of a radical initiator. In addition to, the tremendous industrial importance of such polymerizations, much of our, present knowledge concerning the nature of chain reactions in the liquid state, is derived from investigation of such processesdss(a> Related to such reactions are, the additions of elementary oxygen to olefins, forming polymeric peroxides.t33(-b), , R2C=CH2 + 02 -> [—CH2—CR2—O—O—CH2—CR2—O—O—]x, We shall make no attempt to consider either of these important topics, even, briefly, but shall instead confine our attention here to addition reactions that, yield relatively small molecules., Suppose we are dealing with a chain process, initiated by species In-, in, which groups X and Y (derived from molecule X—Y) are adding to a C=C, double bond., , X—Y, , Y■, , /C, , \ Y—C—C- " Y-—C—C—X + Y- —-^, , Y—G—C- —► etc., , In order that a significant kinetic chain length be maintained, it is necessary that, both the attack by Y- on the double bond (the addition step) and the attack by, , “, , A number of works devoted nearly exclusively to consideration of, , and polymers are ^, Universny Press,, , Publishers Inc, , Itliaca, , 1953, , ^ Pnnciples, , ^'frey, , New York 1952 U, , ^, , few, , PolymeAation, , ^, pUynmiza*m, John Wiley, ^ Jo(m wiley and s, , Wpotym^ric peroxides are discussed in some detail by Mayo,, Miller, and Russell, J. Am. Chm. Soc., 80, 2465-2507 (1958).
Page 747 :
Homolytic Additions, Energetic Requirements, , 731, , radical Y— C—€• on molecule XY (the displacement step) be rapid in compari¬, son to all chain-termination steps. Indeed, it may be assumed that if either of, these two propagation steps is significantly endothermic, addition via a chain, mechanism is difficult or impossible, no matter how effective an initiator is used., If both propagation steps are exothermic, addition via a chain process is possi¬, ble; but it is not guaranteed, for even strongly exothermic reactions may have, high energies of activation or highly negative entropies of activation. With this, in mind, let us compare the energies of the propagation steps for some conceiva¬, ble homolytic additions. In Table 16-4, for which propylene is chosen as a, common substrate, the usual thermodynamic convention (negative AH values, for exothermic reactions) is observed. These values tend to rule out long-chain, , Table 16-4. Approximate Enthalpies of Propagation Steps, in Radical Additions to Propylene'^4, Y- + MeCH=CH2, , MeCH—CH2T+ XY->, , MeCH—CH2T, X—Y, , MeCHZ—CH2T + X-, , AH, , AH, , H—OH, , — 36 kcal, , H—Cl, , -30, , 9, , H—Br, , -9, , -7, , 3, , -23, , Cl—Cl, , -30, , -15, , Br—Br, , -9, , -13, , 3, , -10, , -20, , —7, , -18, , -4, , H—I, , I—I, H—CMe, , 26 kcal, , II, O, Cl—CC13, , processes in which water or hydrogen chloride adds homolytically to propylene, for these are reactions in which the displacement step is endothermic, due largely, to the difficulty in breaking the H-OH bond or the H-Cl bond. Similarly, excluded are additions of iodine or hydrogen iodide to propylene, for here the, addition step is endothermic, due largely to the weakness of the C-I bond., (Ref. i£hppC ^, , 'plated by Walling, , preciably affected by a /3 substituent Verv, , rhr, , f", are, largely fortuitous., , nen., , l, , ti, , ?y, , ^ the lsoProPyl radical is not ap-, , *, , fr om1P a u H ng s^bond e nergi^1^ p 37) U S j °n s, drawn if theL, ® es 'P", although this appears to be
Page 748 :
732, , Free-radical Reactions, , However, the additions, via chain mechanisms, of hydrogen bromide, chlorine,, bromine, acetaldehyde / as H-, , CMe, , V, , O, , /, , and carbon tetrachloride, are seen to be, , energetically feasible, since in these five cases, both the addition and the dis¬, placement step are exothermic. Each of these five additions has, in fact, been, observed., For the homolytic additions of these same reagents to styrene, PhCH=CH2,, the AH values for the various addition steps are lower (that is, the additions are, more exothermic) than those for propylene, probably by about 20 kcal each., This difference reflects the added stability of radicals of the type Ph—CH—, CH2Y, due to conjugation with the benzene ring. On the other hand, the AH, values for the various displacement steps are undoubtedly higher, probably by about, 20 kcal, than those for the propylene series, since the resonance-stabilized, Ph—CH—CH2Y radical necessarily attacks with less vigor than a radical of the, type MeCH—CH2Y. With styrene, each of the addition steps, including those, involving I-, becomes highly exothermic, but the displacement steps (except, those involving Cl2 and HI) have become endothermic. We thus see why a number of, homolytic additions that occur readily by chain processes with ethylene and, propylene must proceed by nonchain processes (if they proceed at all) with, styrene. A radical such as ArCH—CH2Y, although it may be formed with great, ease, is a relatively ineffective displacing agent., , Addition of Hydrogen Halides, In the previous discussion of the addition of hydrogen halides to olefins (p. 519),, the reader was reminded that the structure of the product resulting from the ad¬, dition of HBr often depended upon the reaction conditions employe ., , ore, , specifically, in polar solvents or in the presence of such inhibitors “'h.opheno, and hydroquinone, HBr adds to olefins in the same way as do HC1 and HI,, obviously by a heterolytic process. In nonpolar solvents in t e Prcs|n, oxygen or peroxides, or with the aid of illumination, the direction of addition is, frequently reversed, and the addition becomes just as obviously homolytic., Br, Br‘, , ,CH3—CHR, , -> CH3—CH—R, , (iheterolytic addition), , CH2=CHR, , etc., , pit i>r_CHR___* BrCH2, CH2t>r, , Cvnix, , (homolytu addltlon), , CH2R + Br
Page 749 :
Addition of Hydrogen Halides, , -, , 733, , The homolytic addition is the basis of the familiar peroxide effect, which, at a, relatively early date, strikingly pointed out the necessity for distinguishing ionic, from free-radical mechanisms.'55 The attacking bromine atom is probably gen¬, erated from peroxides present by a sequence such as, 2HBr, , RO—OR -> 2RO-, , » 2ROH + 2Br*, , To account for orientation observed in the radical additions of HBr, we, may invoke reasoning similar to that used in explaining orientation in heterolytic, additions. Just as the secondary carbonium ion,, , CH3—-CHR, is more stable than, , the primary carbonium ion, CH2—CH2R, so also is the, , secondary radical,, , CH2Br—CHR, assumed to be more stable than the “primary radical,” CH2—, CHR—Br; for, as we have seen, hyperconjugation appears to stabilize radicals, in much the same way as it stabilizes carbonium ions (p. 696). Thus, the differ¬, ence in orientations is not due basically to the fact that one process is heterolytic, and the other homolytic, but rather to the circumstance that in one case addition, begins with the formation of a new C—H bond, whereas in the other, it is the, C—Br bond that is formed first. Markownikoff-type addition would presumably, be observed for free-radical additions also if the reaction conditions could some¬, how be altered so that the initial attack on the double bond were carried out by a, H« radical., A glance at Table 16-4 indicates why the peroxide effect is observed for, the addition of HBr but not, generally speaking, for additions of HC1 or HI. By, the argument already presented, the addition of HBr (in which both propagation, steps are exothermic) may proceed readily by a chain mechanism, whereas the, additions of HC1 or HI (in which only one propagation step is exothermic) may, not. At ordinary temperatures, even in the presence of radical initiators, addi¬, tions of HG1 and HI, if occurring at all, will tend to proceed preferentially by a, heterolytic path.'se«*> In rare instances, radical additions of HC1 to olefins have, been observed, but, except at high temperatures, kinetic chains are short.186w, It is interesting that the homolytic additions of HBr to substituted cyclo¬, hexenes are stereospecific trans additions. Addition to 1-methylcyclohexene gives, m-l-bromo-2-methylcyclohexane (LXXXII, R=Me)«™ in which the incom¬, ing hydrogen and bromine atoms lie trans to each other, and similar trans addi¬, tions have been observed with 1-bromo- and l-chlorocyclohexene.'57W Stereochemically these additions seem similar to the heterolytic additions of halogens, Wain4! C^Urr27°,f r5°l ale9a40)ng * ^ elUddati°n °f ^, , ^ see, , «*>, , 56, ,782 (1954)., , (1950); Mayo, ibid., 76, 5392 (1954)., , «■, , ”-La] Gm°^unSy, , Abell> and Aycock,, , Sims, ibid., 77, 3465 (1955)., , tcke, Cook, and Whitmore,, J. Am. Chem Soc, °’, , 74, , /4’, , 3588 i'Ick?',, , /no, , 72, 1511, , ., , (1952). (b) Goering and
Page 750 :
734, , Free-radical Reactions, H, , LXXXIII, to double bonds (Chap. 13), and one is perhaps tempted to suppose that a, bridged intermediate (in this case radical LXXXIII) intervenes here also. How¬, ever, the bromine atom in this radical has nine valence electrons, a feature which, makes this intermediate unlikely.75* An alternative explanation of the stereospecificity757(6),,5a(o) is that the Br* radical attacks the double bond via the least, hindered route. From LXXXIV, we see that this approach, somewhat para¬, doxically, corresponds quite closely to an axial direction in the cyclohexyl radical, that results759(6) and we may suppose that the incoming bromine initially occupies, an axial position in this radical. If it is assumed that the bonds about the trivalent, , LXXXV, , LXXXVI, , trans adduct, , (preferred), carbon in the cyclohexyl radical are pyramidal rather than planar (as may often, be the case in strained systems), then substituent R may occupy either an, equatorial position (LXXXV) or an axial position (LXXXVI). Sterically,, conformation LXXXV is preferred, leaving an axial position open to attack by, HBr. Thus, the trans adduct should be formed preferentially. One proviso must, be made however; the stereochemical argument implies that the attack by, HBr on the cyclohexyl radical intermediate must occur before the ring has a, chance to “turn itself inside out” (p. 241), for such a ring inversion would throw, the bromine into the (thermodynamically preferred) equatorial position., The homolytic addition of HBr to noncyclic olefins at room temperature, and above is nonstereospecific, for attack by Be converts a C-C double bon, , to, , a single bond, about which rotation may occur. Moreover, add.uon of Be, ... Species in which bromine has an “expanded valence shell” dor example, BrF. and Br, 1, are known, but in these, the bromine ,s bound onlyto hahJgen >, m (a) Brand and Stevens, J. Chm. So,., i9St,, 1, S, , thc attacking, tend is not unani, , Leavitt, and Sawarc, ,. da„ O-, , 5621, , (1957).
Page 751 :
Addition of Hydrogen Halides, , 735, , -, , simple olefins is often reversible and has been found, in a number of cases, to, promote olefin isomerization.1', , H, , R, , V, ii, , /Cx, R', , H, , trans, At very low temperatures, however, rotation about the C—C bond in the radical, intermediate becomes much less free, and if a large excess of HBr is taken, stereo¬, specificity may be restored. Thus, the homolytic addition of HBr to the stereoisomeric 2-bromo-2-butenes at —80°, using liquid HBr as a solvent, has been, found to be very nearly completely stereospecific.141 As indicated, the cis olefin, , H, Br •, —80, , G, / \, Br, Me, , Me, , > Br, , H, , Me, , (92%), , Br, LXXXVII, , cis, , meso, LXXXVII I, , gives the mwo-dibromide, LXXXVII I, in over 90 percent yield, whereas the, , trans olefin (not shown) gives the corresponding d,/-dibromide. We suspect then, that the attacking HBr molecule approaches the intermediate radical, LXXXVII, (or its counterpart in the trans-d,l series), keeping as far removed as is possible, from the newly acquired bromine atom, this path presumably being the least, hindered. This is, however, not the only possible explanation; it may be that the, bromine radical, instead of attacking an olefin molecule, attacks a tt complex, , H, , Me, , Br, , Me, , V, I, A, , “fr0m, , Noyes, , °'efin and HBr' 3 Path, , -suiting in stereospecifi, , F°,aday Sx" 45' 749 <1949)- (*) Steinmetz an, Goering and Larsen, J. Am. Chem. Soc., 79, 2653 (1957).
Page 752 :
736, , Free-radical Reactions, , Homolytic Halogen Additions, There are several difficulties associated with the study of homolytic halogen addi¬, tions. We have already noted (p. 521) that such additions in polar solvents tend, to be overshadowed by heterolytic additions, and this may be the case also in non¬, polar solvents if pains are not taken to exclude traces of water or hydrogen halides., Even if the reactants are mixed in the vapor phase, the addition may still proceed, heterolytically on the glass walls of the container unless they are coated with a, nonpolar and (presumably) inert material such as paraffin. Some judgment, must be exercised in the choice of a reaction temperature, for at high tempera¬, tures homolytic substitution may compete effectively with addition. A final com¬, plication is the possibility that the addition of a halogen atom to the double, bond may be significantly reversible., The energy values in Table 16-4 suggest that the additions of chlorine and, bromine, via a radical chain mechanism, to ordinary olefins should proceed with, ease, whereas additions of iodine should not. Successful kinetic studies of addi¬, tions of fluorine are exceedingly difficult, for such additions are generally violent,, often resulting in the formation of a number of products.55, The usual chain mechanism for halogen addition, X, , )c=c(, —X, , hV, ->, , X-, , C-, , -c, I, , X,, + z=±, , X, etc., , 18, , -G-C- + AT-, , I, , may lead to one of several rate laws, depending upon whether the addition or the, displacement step (or both) are reversible, and depending also upon the nature, of the termination step. For example, it may be shown (Ex. 6a) that a termina¬, tion step involving the reaction between two — C—CX— radicals results m a rate, proportional to (X*)* (whether or not the addition step is reversible); the rate, laws for the chlorination of CH2=CHC1, CHC1=CHC1, CHC1=CCU, and, CClj=CCl2 in the vapor phase have been found to be of this type.'<!, , n t e, , other hand, a termination step involving the reaction between two unlike rad.cals leads to a rate law of the type, rate = kl* I, , C=C, , 1, , (X,), , <19>, , where / is the intensity of the radiation bringing about the ^, rate law for the chlorination of ethylene itself is of t is type., in Schumacher, e< */.,, (194m Schmitz, Schumacher, and Jager,, , B35' 285’ 455 ^, , Z. Phpik. Chm.,, , B51281 (1942)., , ^, , ”
Page 753 :
Homolytic Halogen Additions, , -, , 737, , actions is inhibited by oxygen, as are the halogenations of aliphatic hydrocarbons, (p. 695), and for the same reason, that is, the diversion of the chloroalkyl radical, to the very much less reactive peroxy radical, XC,, , —C—C—Cl, , —C—C-h O2, , I, , I, , I, Cl, , 0—0, , I, , •, , XC, Addition of Cl- to ordinary C=C double bonds appears substantially irre¬, versible at temperatures below 200°, but reversibility becomes increasingly, important as the temperature is raised. The rate of hydrogen abstraction likewise, increases, and at temperatures above about 450° we may expect substitution at, the allylic position to overshadow halogen addition. Typically, isobutylene is, converted to methallyl chloride, CH2=CHMe—CH2CI, in good yield by treat¬, ment with chlorine at 600°.*4#, In contrast to addition of Cl*, addition of Br- is often reversible at room, temperature. This reversibility is most strikingly demonstrated by the ability of, sources of bromine radicals to bring about cis-trans isomerization of olefins, a, transformation that we have discussed in reference to homolytic additions of HBr., This process is of considerable interest since it is related to the rotation of a, group about a single bond within the bromoalkyl radical. By using radioactive, bromine it is possible to compare the rate of such a rotation to the rates of gain, and loss of Br- by an olefin. 1Jf0(-b) Such comparisons have been carried out, using, as a substrate m-l,2-dibromoethylene, and it has been found that the prover¬, bially, , free, , rotation in radical LXXXIX is not extremely fast, but actually, , about half as rapid as the loss of Br- from such a radical., , Elemental iodine, when heated or irradiated in solution, likewise catalyzes, the as-trans isomerization of olefins,« indicating that, as expected, addition, of iodine atoms to ordinary C=C double bonds is also reversible. As with broand Hearne’ Ind■ Ens■ Chem., 31, 1530 (1939)., (f) N°yes> Dickinson, and Schomaker, J. Am. Chem Soc 67 mo fioam tu\ tv i ■, son, rf al., tbid., 61, 3259 (1939); 65, 1427 (1943); 71, 1238 (1949) ’, (, 5)* W DlCkm', m
Page 754 :
738, , Free-radical Reactions, , mine, the rate of iodine-catalyzed isomerization of m-diiodoethylene has been, compared to the rate of iodine exchange, using labeled iodine.^5(o) Due proba¬, bly to steric interaction between the large iodine atoms, rotation about the C—G, bond in the CHI2, , CHI radical is quite slow; for the isomerization is found to, , proceed only about one hundreth as rapidly as iodine exchange. With iodine, addition, the displacement step in sequence (18) is also reversible, making the over¬, all addition reaction reversible. In fact, although additions of I2 to olefins may be, carried out at low temperatures,it is the reverse reaction which is more, readily studied. 1^6(~b), , RCH—CHR i RCH=CHR + I2, , I, , I, , I, , I, , This homolytic elimination may be initiated by attack of I* on the bound iodine, atom,, RCHI—CHIR + I. ^ RCHI—CHR + I2, or simply by photochemical breakage of the C—I bond by ultraviolet irradia¬, tion. I*6(-c), With allylic halides, the reversibility of halogen atom addition may lead to, allylic rearrangement. Thus the conversion of 3-bromo-l-butene (XCI) to, 1 -bromo-2-butene (XCII) with HBr at -12° is catalyzed by peroxides, almost, certainly through the intervention of Br* radicals. ^7, CH2=CH—CHBr—CH3 % [BrCH2—CH—CHBr, , CH3], , XCI, BrCH2—CH=CH—CH3, XCII, No discussion of homolytic halogen additions should omit mention of the, addition of chlorine to benzene, for this is one of the few cases in which a nonactivated benzene ring undergoes addition rather than substitution under rela¬, tively mild conditions.^* The resulting mixture of 1,2,3,4,5,6-hexachlorocyclohexanes contains five of the eight possible stereoisomers of that compound; the, identification of these isomers has already been considered (Chap. 12, Ex., , )., , There can be no doubt that the initial attack is homolytic, for it is initiated by, f« (a) Forbes and Nelson, J. Am. Chem. Soc., 59, 693 (1937). (b) Schumacher, et at, physik. Chem.,, , Bit,, , 45 (1931);, , B12,, , 349 (1951). (c) Dc Right and Wng, J. Am. Chem. Soc.,, , W Young and Nuzak, J. Am. Chem. Soc., 62, 311 (1940)., us (a) For quantitative studies of this reaction, see: Noyes, r a .,, , Chem., B19, 190 (1932)., , ., ., , phem Soc, m., , Z., ,, , 54 161, ’ ^955).
Page 755 :
Additions to Dinitrogen Tetroxide, , 739, , -, , light and by the addition of peroxides. The sequence represented below is purely, schematic, for we know little about the intervening steps. Dichloride XCIV,, , mixture of, benzene, hexachlorides, XCIV, , (mixture), , J/, , <,o, , /, , Cl, , <, , XCVI, , which should be extremely reactive, has not yet been isolated from the reaction, mixture, but a mixture of four stereoisomers of tetrachloride XCV may, with, care, be obtained.740 These are readily chlorinated photochemically to mixtures, of benzene hexachlorides, but react very slowly with chlorine in the dark., As indicated, radical intermediate XCIII may be “trapped” if the chlorination, is carried out in the presence of maleic anhydride.750 The “trapped” radical is, isolated ultimately as phenylchlorosuccinic anhydride, XCVI., , Additions of Dinitrogen Tetroxide, The additions of N2O4 to olefins in nonpolar solvents are similar in a number of, respects to homolytic halogen additions and may be considered briefly at this, point. The N—N bond in N204 is extremely weak (13 kcal), and the tetroxide, is measurably dissociated into N02 radicals, even at 0° C. The initial attack on, the double bond is by N02, but it is not known whether the resulting nitroalkyl, radical then reacts with N02, with N204, or with both.757, , /, , c, , \, no2, , 1, 1, -no2, -C-->, or N2O4, , 1, 1, C, , no2, c, 1, , no2, , r, 1, 1' or —C-C—, ', V, no2, , 0, , C=C, , 1, , £, 1, 0, , •no2, , J, , As shown, the second nitro group may become bonded to the carbon by if, nitrogen atom, yielding a dinitro compound, or, alternatively, by an oxygen, 4243 (ImI)"63”'’ Gn*ng’ Kerr’ Kolka’ and Orloff, J. Am. Chem. Soc., 73, 5224 (1951); 75,, , 2, , TukC’ Buzbee> and Kolka, J. Am. Chem. Soc., 78, 79 (1956), , This question is discussed by Brand and Stevens, Ref. 139(a).
Page 756 :
740, , Free-radical Reactions, , atom, yielding the nitrite ester of an a-nitro alcohol.'5* The initial attack is, represented as being reversible, for it has been found that N02, like iodine and, bromine, catalyzes the cis-trans isomerization of olefins.'53 Until recently, a, number of workers felt that this reaction was ionic—that is, that the NO^ and, NO^- ions were involved. However, the observed “diversion” of the nitroalkyl, radical with iodine'5* and with BrCCl3'3i?(a) leaves little doubt that, in nonhydroxvlic solvents at least, the reaction is homolytic., , CH»=CH—Et —°4+I> 02N—CH2—CHI—Et, , N2Q4, in BrCCl3, , It is of interest that the addition of N0O4 to 1-methylcyclohexene, like the, homolytic addition of HBr (p. 734), gives a nearly pure trans adduct, whereas the, addition to cyclohexene itself yields a mixture of cis and trans adducts with the, , trans predominating slightly.'39(o) If the additions are presumed to proceed by a, path analogous to that for additions of HBr, the nitro groups in the nitrocyclohexyl radical intermediates should, at first, occupy axial positions. Assuming, once again that the configuration of bonds about the trivalent carbon is pyram¬, idal, it may be supposed that both the methyl group in XCVII and the hydrogen, 0-N=0, , shown in XCVII I prefer the less crowded equatorial position, but that this, preference is much stronger for the bulkier methyl group. A significant fraction, ., , r nu, , 1093 1096 1100:1948, 52; 1949, 2627. Somewhat puztwch ^Togl’ps a, attached through osygen a,on,, have, , not bein isolated in significant yields from such react,on matures., "* Khan, J. Chem. Phys., 23, 2447 (1955)., in Stevens and Emmons, J. Am. Chem. Soc., 80, 338 (1958).
Page 757 :
Additions of Thiols, , 741, , of the radical intermediate from cyclohexene may well assume conformation, XCIX, resulting ultimately in cis, as well as trans, addition.755, , Additions of Thiols, Although heterolytic additions of mercaptans and thiophenols are known156, homolytic additions are considerably more important. Like the corresponding, additions of HBr, these are initiated by peroxides, may be photochemically, induced, and may be retarded by such inhibitors as hydroquinone. Moreover,, the direction of addition is generally opposed to that predicted by Markownikoff’s rule; for example,, •, , D CU, , RS- + CH2=CHR' ^ RS—CH2—CHR'-» RS—CH2—CH2R' + RS-'57, For ordinary olefins, both propagation steps in the sequence above are exo¬, thermic. The addition step is about 10 kcal more exothermic than the corre¬, sponding addition of Br-, whereas the energy for the displacement step is very, nearly that for attack of a Br—CH2—CHR radical on HBr. Once again then,, reaction via a chain mechanism is possible and, indeed, is observed./57 The addi¬, tion step is, at least in some cases, reversible, for cis-trans isomerization of olefins, sometimes occurs during thiol addition reactions.158, Thiyl radicals (RS*), should, like chlorine and bromine radicals (p. 697),, be electrophilic, and accordingly, we may expect electron-donating groups in the, olefin to facilitate homolytic thiol additions. It has indeed been found that, incorporation of a />-MeO group into a-methylstyrene increases its reactivity, toward HOOC, , CH2, , S- radicals by a factor of about 100, whereas incorpora¬, , tion of a p-F group decreases reactivity by a factor of two.75S(o) By an extension of, the same argument, electron-attracting groups in the thiol should increase the, reactivity of the derived thiyl radical, and this too has been observed.154^, , ^, , ° ,Ac|Idlt‘onis of N2°4 to alkynes appear to be nonstereospecific. See Campbell, Shavel, , Chem' S°C:’ 75, 2400 (-1953); and Freeman and Emmons, ibid., 79, 1712, L r-,\, ,^0m0cytlc rpec^amsms have been suggested also for the addition of N203 and, N02C1 to olefins. See Schechter, et al., J. Am. Chem. Soc., 74, 3052 (1952); Chem. and Ind., 1955,, , «• Base-catalyzed additions of thiols to C=C or C^C bonds lying near strongly electronattracting groups are undoubtedly initiated by attack of RS~ on the double or tripVbond (see, for example, p. 529) On the other hand, acid-catalyzed additions of thiols to olefins (see for, example, Ipatieff, Pines, and Friedman, J. Am. Chem Soc 60 2731 no3«n, i, nary protonation of the double bond., ’, ’, (1938))’ mv°lve Prehmi-, , J., , Cm' J' Ctm- 32’ 1078 (l954>- 33’ 1034 <'955);, 118 Helmreich and Walling, Ref. 1 (b), , J. dL(a“4I',, , p 323, , J' A7 k*T-, , 7°’ 2559 <1948>- W Cunn«n,, , Alderman, and Mayo, J. Am. Chem. Soc 70°3740 (mwT™'T* i-” f°r.examPle> Gregg>, direction., ’, ’ /4U t1948)) do not indicate clear trends in this
Page 758 :
742, , Free-radical Reactions, The stereochemistry of thiol additions, like those of a number of homolytic, , additions, depends upon the substrate and, to a lesser extent, upon reaction con¬, ditions. Like the additions of N2O4 to alkynes, the addition of thio-^-cresol to, phenylacetylene is nonstereospecific.160 On the other hand, as we might predict,, addition of thiols to, , -chlorocyclohexene gives predominantly trans addition;, , 1, , some cis addition is observed, but it becomes less important as the ratio of thiol to, olefin is increased.*0* Here, it is likely that conformational effects analogous to, those proposed for the homolytic additions of HBr (p. 734) come into play, with,, however, the operation of one additional factor. Because thiols are considerably, less reactive than HBr as hydrogen-transfer reagents, the radical intermediate G, may exist for a significant time interval before attack by a second thiol molecule;, indeed, it appears that during this interval, a fraction of these radicals may, assume the alternate conformation, Cl. If it is assumed that the chloro group in, both C and Cl prefers an equatorial position, conformation C leads to trans, addition, whereas Cl leads to cis addition. Moreover, as the concentration of, , thiol is increased, radical C will be more and more likely to suffer attack by thiol, before isomerization to Cl can take place., Still another stereochemical situation is encountered in the addition of, thiols to the rigid bicyclic norbornene system (CII).*6~ By the use of scale models,, it may be shown that attack by a large radical on the “underside'’ of CII (endo, attack) is considerably more hindered than attack on CII on the, , flank, , (exo, , attack). As a result, attack on CII by the ^-thiocresoxy radical gives almost, exclusively the exo radical CIII, which, in turn, leads almost exclusively to the, , exo adduct CIV. But this transformation does not tell us whether the overall, , M Kohler and Potter, J. Am. Chem. Soc., 57, 1316 0935)i« Goering, Relyea, and Larsen, J. Am. Chem. Soc., / 8, 348, ibid., 79, 3493 (1957)., c, Sti99 11954)., tet Cristol and Brindell, J. Am. Chem. Soc., 76,, (, , }. Bordweu and Hewett,, no
Page 761 :
Additions of Polyhalomethanes and Aldehydes, , -, , 745, , over only two chlorines). Such an inference is in accord with known bond-dis¬, sociation energies., A complicating factor which, in principle, is associated with all homolytic, additions, may become particularly pronounced in additions of halomethanes., The radical intermediate —C—C—CA3 may, instead of attacking a second, , I, , I, , molecule of halomethane, attack the double bond in a second molecule of olefin, and the resulting radical may then attack a third molecule of olefin, and so on., , .1, , \, , II, , /, , —c—c—cz3 +, , c=c, , /, , -III, , -> —c—c—c—c—cx3, , \, , I, , I, , I, , I, , Such polymerization cannot, in the presence of excess halomethane, go far, for, soon the growing carbon chain must attack a halomethane molecule. Thus, there, , —C—C—C—C—CX3 + CA’4 -> X—C—C—C—C—CX3 + -CAa, , I, , I, , I, , I, , I, , I, , I, , I, (“chain transfer”), , may be formed a series of molecules in which a X— and a X3C— group are, separated by two, four, six, or more carbon atoms. Such “short-chain polymers”, are often called telomers. Telomerization, as may be expected, becomes important, when the halomethane employed is relatively unreactive (or when its concentra¬, tion is low). Thus, considerable quantities of telomeric materials are formed in, the reaction of CC14, HCCI3, or CF2Br2 with ethylene, but practically no telomer, results from the addition of the more reactive halides, CCl3Br and CBr4, to the, same olefin.^ Similarly, allyl chloride gives much telomer in its reaction with, CC14, but very little in its reaction with CCl3Br under similar conditions. Telomerization is likewise favored by increasing the stability, hence the selectivity, of, •, , I, , the — C—C—CX3 radical intermediate; a more stable radical is more likely to, survive in solution long enough to add to another olefin molecule, whereas a very, reactive radical is more likely to react with the excess halomethane. It is found, or example, that the photochemically induced addition of CF3I to acrylonitrile, (which proceeds through the resonance-stabilized F3C—C—C—CN radical), , CH=CHFtUre,H0f Fl0meriC Pr°dUCtS’ WhereaS addi“0n, the'same halide ,o, 2, >1C S the exPected monomeric addition product.'*'™ Similarly, CcJAso kS, Sehon, J. Chem. Phys., 19, 656 (1951 )).e G, , *«■ th. C-Br bond in, ^ b°nd m CHCl2Br 1S 54 kcal (see Szwarc and
Page 762 :
746, , Free-radical Reactions, , addition of bromoform to styrene (but not to 1-octene) results largely in telomer, formation. 765(a), The reactivity increase within the sequence, Cl—CC13 < Br—CC13 <, I—CCI3, is undoubtedly related to the corresponding decrease in activation, energies for displacements of the following type:, , \, , \, , /, , /, , —c- + x—ccu —► —C—X + -CCls, This, in turn, is related to the energies of the bonds being broken (p. 37). The, reactivity increase within the sequence,, , X—CAr3,, , X— CH3, , <, , X—CH2Ar, , <, , X—CHAT2, , <, , is almost certainly related to the corresponding increase in resonance, , energies associated with the radicals being produced. Indeed, halomethanes, having less than three halogen atoms undergo addition only with difficulty;, however, additions of such compounds as BrCH2CN and Cl2CHCOOMe (which, yield radical intermediates significantly stabilized by conjugation) have been, reported.767, A further complication may enter the picture if the olefin has one or more, allylic hydrogens, since the radicals produced in the initiation step may remove, an allylic hydrogen rather than adding to the double bond. The allylic radi¬, cal thus formed may dimerize or may react with another alkyl polyhalide, molecule.766 Finally, if the addition is carried out on an allylic halide, a halogen, , H, I, , I, , I, , c=c—c-, , -H*, , —c=c—c—c—c=c—, allylic, termination, , —C==C—C—A -f- *CA 3, , atom may be lost from the radical intermediate, resulting in an “allylic, rearrangement.”, , ch2=ch-ch,y ^4 *3C-ch2-ch-ch2a- ^c_CHj_CH=CH!,tt, - see, for example, Kharasch, Skell, and Fisher,, ... See, for example, Kooyman, Due. FaeaiaySoe1C>,163 (Uhlh, , of polyhalide, , Si maX“rabfe di^hyXan al.empt in this direction, see Melville, Robb,, , ^ ^Kharasch and SagelT^. CW, .4, 573 (1949).
Page 763 :
Additions of Polyhalomethanes and Aldehydes, , -, , 747, , Another recently developed addition reaction is that of aldehydes to olefins., This process, which is generally initiated by acetyl or benzoyl pci oxide, proceeds,, in all likelihood, via the following path:, In-, , •, , CH2=CHR' _, , _.TT, , • TTn /, , RCHO, , RCHO -> RC=0-» R—C—CH2—CHR ->, O, , •, , _ CH2=CHR', , R—C—CH2—CH2R + RC=0-», etc., , o, This addition appears to proceed most cleanly with “terminal” olefins (having, =CH2 or =CF2 groups) and with compounds in which the C=C bond is in, conjugation with a cyano, keto, or similar group. With substrates of the latter, sort, the RC=0 radical tends to attack the /3-carbon, yielding a resonancestabilized radical adduct; for example,, , RC=0 H-C=C—C=0, R—C—C—C—C=0 ^ R—C—C—C=C—O:, O, , O, , However, if the /3-carbon is sterically hindered, attack at the a-carbon may occur, also. Thus, the addition of butyraldehyde to the acid, Me2C=CH—COOH, yields a mixture of both possible addition products, with, over a attack by a margin of, , 3, , attack predominating, , to 1,170^, , PrCHO -f Me2C=CH—COOH, PrC, , /3, , Bz202, , -», , CMe2—CH2—COOH + Me2CH—CH—COOH, , I, °, , C-Pr, 75 per cent, , O, 25 per cent, A similar ratio of products results from addition of the same aldehyde to the, O, , ester Me2C=CH, , COOEt, but if, however, mesityl oxide, Me2C=CH-LMe, arpredominates—«» a, , it, , Position rT’J, mt'"e M^C=CH-CN, attack is exclusively at the 0, P ■ mon This trend suggests that an a-cyano group is more effective in stabilizing, •«»., , <»«). (*) Patrick, ibid., 17,, , homolytic additions of aldehydes to olefins, s^e Wailing Refi 1^,, , ^ **
Page 765 :
Homolytic Cyclizations. Diradicals, , -, , 749, , nary conditions. It does, however, react with a number of cyclic dieneoid systems, in the light to give six-membered ring (“transannular ) peroxides such as CXV,, CXVI, and CXVII. Sometimes, as is the case with the formation of ascaridole,, , ascaridole, Ph, , Ph, , CXVI, the addition does not take place readily unless a photosensitizer is added;, the latter is a colored substance (for example, eosin or methylene blue) that ab¬, sorbs radiant energy, then transfers it to the substrate molecule during a col¬, lision. Although we cannot be sure, it appears that the light absorbed excites the, dienoid structure to a diradical state, CXVI 11, which, in turn, reacts rapidly, , CXVIII, with the oxygen diradical. Since the overall process involves consumption of, diradicals with no regeneration, these are not chain reactions. These additions, bring to mind the Diels-Alder reaction (p. 536), which is not, however, photochemically induced and may therefore be assumed not to require preliminary, conversion of the substrates to their diradical forms., A number of additional photochemical cyclizations are known which are, of why molecular oxygen assumes a diradical structure rather than structure PXTV . u- u •, , Sii mr LTrvr.acquain,anw wi,h *, Inorganic Compounds, Elsevier Publishing^,'incl)?1 40»/, some objection among workers in nhntrw'hf>m- * \, > ^^0, pp. 403 411. There is, in which there is significant interaction hetw ^, 3PP ym? the term “diradical” to species, Simons, Quart. Revs., XIII, 10 (1959), VCen W° unPaired electrons; see, for example,, Naturwissenschaften, 32, 157^0944) M^Mouret’, 1859 ^93<^* ^ Schenck and Ziegler,, 1584 (1926); note that the f^, ^ompt. rend., 182, 1440, reversible., 1 Peroxide CXVII (but not CXV and CXVI) h
Page 766 :
750, , Free-radical Reactions, , likewise presumed to proceed through preliminary excitation of one or both, reactants to a diradical. Four of these are listed below:, 173(a), , +, , HCPh, , II, , hv, , (20), , HCPh, , 173(b), , PhCH, II, HC-C-Ph, II, O, , hv, , (21), PhC, II, , II, O, , CMe2, , +, , II, , II, , o, I>h, , Ph—C-Me, , CPh, , hv, , ->-, , O, , Me, , Me, , CHMe, , 173(c), , lTe, , (22), , c)—, Me, , (23), , For cyclizations (20) and (22), we cannot as yet say whether it is the carbonyl, compound or the olefin which, in the initial step, is converted to a diradical., Note that the cyclization product from acetophenone and 2-methyl-2-butene, (reaction 22) is derived from the diradical intermediate CXIX. The alternative, product, CXXI (which is not obtained), would be derived from diradical CXX,, which is presumably less stable than CXIX; for in the former, the odd electron, center on the right is stabilized by two a-methyl groups, whereas in CXX, it is, stabilized by only one a-methyl group. In the same way, we may explain why, cyclization (23) yields the observed “head-to-head” dimer, rather than the, alternative “head-to-tail ’ dimer, CXXII., , ., , In contrast to the 02 molecule, the methylene diradical, -CH2,^ is exceedm (a) Schonberg and Mustafa, J. Chem. Soc., 1948, 2126. (b) Mustafa Chem^Revs., 51 1, (1952). W Buchi, Inman, and Lipinsky, J. Am. Chem. Soc, 76, 4327 (1954). (J) Greene, M.srock, and Wolfe, ihid., 77, 3852 0955), , “°pny, , as to whether the methylene, , -Ha- U. .Hat these e.ecteons aee unpaid
Page 767 :
Homolytic Cyclizations. Diradicals, , 751, , -, , Ph, , I, , Me-C-, , CMe2, , I, , I, , O — CHMe, CXIX, , ingly reactive. This species is generally prepared by the photochemical decom¬, position of ketene (CH2=C=0)*™ or diazomethane (CH2N2),*''(a) or by the, thermal decomposition of the latter.*™ By use of the Paneth technique (p. 683),, methylene produced in the gas phase may be made to remove mirrors of ele¬, mental selenium or tellurium, converting these elements to the polymeric sub¬, stances (CH2Se)r and (CH2Te)I./75(o), Like the CBr2 molecule (p. 535), methylene adds readily to C=C double, bonds, forming cyclopropane derivatives, and, once again, the additions appear, to be stereospecific., , Me, , \, , H, , /, , 177(a), , Me, \, , CH2, , Me, /, , /C=cx, , 177(b), , Me, , CH2, , Me, , —v, , /G~G\, , H, , Me, , Me, , H, , H, , Another, and a more unusual, reaction of the methylene diradical is its attack on, C, , H bonds, converting them to C—CH3 groups. Just how such “homologa¬, , tions” occur is not yet clear. Although it has been suggested that CH2 somehow, “adds directly” into the G—H bond in a single step,*™ the two-step process,, , \, —C—H, /, , + .CH2, , \, , -c-, , + H—CH2, , fast, , \, , -c—ch3, , cannot, in the opinion of this author, be excluded, especially since ethane is, generally formed (possibly from dimerization of methyl radicals) as a side, product in such reactions.*75(c), It is interesting that the selectivity of attack by methylene on C—H bonds, varies with the manner in which this diradical is generated. When methylene,, (ti) Pearson, Purcell, and Saiph, J. Chem, , (c), , Soc, , ano (u\ d, , i, , ,, , p., , ., , Taylor, J. Am. Chem. Soc., 63, 1956 (1949), Frev’and K' f, ^ ?Urt?n Davis> Gordon> and, irsD^^ipi, . ’, y, w treY and Kistiakowsky,, 79. 6373 ('19571, Kice and Glasebrook, J. Am. Chem. Soc 55 4329 (iniit. J oioi /fATL, U, , (1956)., , (, , ibid.,, , { ) Doenn8> Laughlin, Buttery, and Chaudhari,, , ibid.,, , 78, 3224
Page 768 :
752, , Free-radical Reactions, , produced from photolysis of ketene, reacts with propane, the ratio of rc-butane to, isobutane in the resulting product is about 7 to 4, whereas if attack were staH3CCH2CH3, , +, , CH2-> H3GCH2CH2CH3, , +, , H3CCHCH3,75<c>, , I, 63 percent, tistically random, this ratio should be, , 6, , ch3, 37 percent, , to 2. Under these conditions, then,, , secondary C—H linkages are attacked somewhat more easily than are primary, C, , H linkages. More specifically, attack at a secondary hydrogen is about, , 1.7 times as rapid as attack at a primary hydrogen, a figure which is in agree¬, ment with that obtained from similar experiments using different paraffin, hydrocarbons.m(o) On the other hand, attack by methylene produced photochemically from diazomethane is much less discriminate.777(c)’77S(6) For example,, its reaction with /z-pentane gives the three possible hexanes in the proportions, shown :/77(c), Me(CH2)3Me, , Me(CH2)4Me + Me2CH(CH2)3Me +, 49 percent, 34 percent, MeCH2CHMeCH2Me, 17 percent, , This is in very close agreement with the ratio 6:4:2 that would result from ran¬, dom attack. Such differences in selectivity are seemingly at variance with the, elementary principle that the mode of reaction of a given intermediate is, independent of the manner in which it is produced. However, we are dealing, here with photochemical reactions, and it may well be that the methylene, molecules produced from diazomethane have more excess energy than those, produced from ketene, and are thus less selective in the sites that they attack.779, Aside from molecular oxygen, the diradicals that we have considered are, short lived, and evidence for their existence, although convincing, has been in¬, direct. However, a number of diradicals related to triphenylmethyl are known;, some of these are stable almost indefinitely and their diradical character may be, confirmed by magnetic measurements (p. 673). Treatment of hexachloride, CXXIII with zinc, for example, yields diradical CXXIV, solutions of which are, ns (a) Knox and Trotman-Dickenson, Chem. and Industry, 1957, 731. (b) Frey, J. Am., Chem. Soc., 80, 5005 (1958)., i» The imine diradical, -N—H, is formally similar to methylene. The preparation of this, species, by passage of an electric discharge through gaseous HN3 has been reported:, HN3-> H—N- + N2, (Rice and Freamo. J. Am. Chem. Sac., 75, 548 (1953)). This condenses at liquid air tempera¬, tures to a blue solid, which decomposes at -125” to ammonium az.de, NH.N,.
Page 769 :
Homolytic Cyclizations. Diradicals, , -, , 753, , paramagnetic./80(o) The quinoid structure, CXXV, for this compound, in which, , Cl Cl, , all electrons are paired, is not stable, for in such a structure, the two central, benzene rings must lie in or near a common plane, a feature prohibited by steric, interference between the bulky orMo-chlorine atoms. A second stable diradical, is hydrocarbon CXXVI ;7SO(6) no less than one hundred structures may be, drawn for this compound, in which each of the two unpaired electrons may be, “placed” on one of ten possible carbon atoms. However, low-energy quinoid, structures, in which all electrons are paired, cannot be drawn. It should be noted,, however, that the paramagnetic susceptibilities of solutions of CXXIV and, , CXXVI are far less than would be observed if all solute were in the diradical, form. Some type of association, the exact nature of which is not yet clear, is, apparently occurring, the association product presumably being diamagnetic., Quinoid structures for hydrocarbons CXXVII, , and, , CXXIX may be, , CPh., CXXVII, , CXXVIII, , PhoC, , CPh., CXXIX, , (#) Muller, Neuhoff, and Tietz Ber 70 oc\^/x /i noox — A ~~_, Muller-Rodloff, Ann., 517, 134 (1935)’, ’, ’, 939)>' 74> 80V (1941). (b) Muller and
Page 770 :
754, , Free-radical Reactions, , drawn; yet, both of these have been found to be paramagnetic, not only in solu¬, tion, but also as solids.'5* Moreover, solutions of CXXVII and CXXIX in, benzene are found to become more strongly paramagnetic as the temperature is, raised. Two explanations for this effect come to mind. Association of the diradical, may again be occurring, with the degree of association falling as the tempera¬, ture is raised. Or, it may be that the hydrocarbons exist in two monomeric, forms in solution, a paramagnetic diradical form and a diamagnetic quinoidal, form (for example, CXXVIII). If this be the case, the increase in the concentra¬, tion of the diradical with temperature indicates that the diradical is of somewhat, higher energy than the quinoidal form, although the difference is probably less, than 1 kcal. With hydrocarbon CXXX, the degree of conversion to diradical, CXXXI is so small that it cannot be detected magnetochemically,'S0(6) but, is easily detectable by the techniques of electron-spin resonance spectroscopy, (p. 684).18S Assuming once again a quinoid-diradical equilibrium, it has been, , Ph2c^ZH(Z)=CPh2, , Ph2c-4^VO^Ph:!, , CXXX, , CXXXI, , estimated that about 4 percent of this hydrocarbon exists in the diradical form,, CXXXI, at room temperature. It is to be emphasized once again that CXXX, and CXXXI, , (or CXXVII and CXXVIII) cannot be regarded simply, , O, , O, I, -N, , Me, Me-, , I, N-, , Me, , Me, -Me, , Me-, , H—N, , H, CXXXI I, , N—^, N—H, H, , O, , -N, , N-, , Me, , -N, H—N, , H, , N—\, N-H, H, , porphryrindine, , H2SC>4', , o -, , "h2cT, , bianthrone, , CXXXI 11, 181, 182, , Me, , Vn-n^, , V-N=N-^, N, , O, , as, , Muller and Pfanz, Ber., 74, X°S1, 1075, Hutchison, Kowalsky, Pastor, and Wheland, J., , Chm. Phys.,, , 20 ,485 (1952)., 4U, 1
Page 772 :
756, , Free-radical Reactions, , Me, I, , Et— C —CH,, , I, Me, , CXL, , CXXXIX, , The fact that aryl groups, but not alkyl groups, may shift in rearrangements, of free radicals suggests that such rearrangements proceed through a bridged, activated complex or intermediate such as CXLII, analogous to the phenoniumion intermediate CXLII I which has been shown to intervene in a number of, +, , >, , CXLIII, , CXLII, , heterolytic rearrangements (p. 575). However, such an aryl shift generally does, not occur unless there is considerable crowding at C#. Thus, no migration of, aryl groups is observed in reactions of radicals of the type Ar2CHCH2, even if, one of the aryls isjfr-anisyl (which might be expected to be a very effective bridg¬, ing group).185 It appears then that the formation of bridged radical intermediates, such as CXLII is by no means as easy as the formation of a bridged carbomum, ion such as CXLIII. This should not be surprising if we note that a rearrange¬, ment through CXLII is, in effect, a radical displacement at Ci of the benzene, ring, whereas a rearrangement through CXLIII is an electrophilic displacement, at this carbon. As has been pointed out (p. 691), radical displacements on car¬, bon are rare and presumably require considerable activation energies, whereas, electrophilic displacements on aromatic carbon comprise one of the most, •, , #, , 187, , familiar classes of organic reactions., Nevertheless, for those few cases in which aryl migration is facile, the ques¬, tion arises as to whether the unrearranged radical may exist at all; for it is, not inconceivable that formation of the alkyl radical and shift of the aryl group, I*' In this regard it may be noted that anchimeric assistance to radical formation by a, migrating ^ ^up (analogous ,o anchimeric assistance to carbomum ton format,on, p. 577), has not yet bLn observed. Thus, the rate of a transformation of the type, Ar, CMe2—CHMe—N=N—CHMe, , Ar, , Ar, , CMe2 —* 2CMe2-CHMe + N2, , Jverbergir and Gainer, J. Am. Ch'm. JW., 80, 4561 (1958)).
Page 774 :
758, , Free-radical Reactions, , Ph3C—O—O—CPh3, , ., , Ph2C—OPh, ^ Ph3C—O* -> Ph2C—OPh ->, , Ph3C—OOH, , Ph2C—OPh, , Such i earrangements are of particular interest when two diflercnt aryj, , groups are bound to the migration origin. By comparing the yields of the possible, , rearrangement products, we may estimate the relative migratory aptitudes of a, , number of aryl groups in homolytic rearrangements. (Conformational effectsj, , which sometimes complicate migratory aptitude studies, do not enter the picture, , here.) Thus, by carrying out decompositions of peroxides of the type ArCPh2—, O2, , CMe3 in cumene (Me2CHPh),790 it has been shown that the phenyl and, , p-tolyl groups have very nearly the same migratory aptitude, but the a-naphthy], , and /?-biphenylyl groups migrate about six times as readily as phenyl. Further, the, decomposition of the nitro-substituted peroxide CXLVIII in benzene yields over, , three times as muchjfr-nitrophenol (resulting from migration of the/?-nitrophenyl, , group) as phenol (resulting from phenyl migration). This suggests that in¬, , corporation of a jfr-nitro group increases, by a factor of about 6, the migratory, , aptitude of the phenyl group in homolytic rearrangements/97 contrary to what, , is observed in carbonium-ion rearrangements. It appears then that the migratory, , NOz-^y CPh2-0-0H, CXLVIII, , + PhOH, 3, , 1, (+further products), , aptitude of a substituted aryl group is related to the effectiveness of the substit¬, uent in stabilizing the bridged transition state (or the bridged intermediate), , in the rearrangement. As shown below, the ^-nitro group (CXLIX) and the, , /;-phenyl group (CL) (as well as most other ortho and para substituents) would be, , expected to stabilize such a bridged radical by aiding in the delocalization of the, unpaired electron.79*, , Bartlett and Cotman, ./. Am. Ch,m. Soc., 72, 3095 (1950). This reaction is complicawl, by formation of rather large amounts of additional products among them p-mtrobenzophenone and, benzophenone, was isolated., I1U />-nitrotriphenyl carbinol., --- No, ", *, 19f, , The decomposition of optically active PhCHMe-O^O-CMe, in the presence of
Page 777 :
761, , Exercises for Chapter 16, , lytic ^-elimination reaction, proceeding in the direction indicated largely be¬, cause of the relief of steric strain accompanying the opening of the four-membered ring. The rearrangement of radical CLIV is, in essence, an intramolecular, addition reaction, resulting from interaction on the ‘ underside of the ling, between the odd electron at C2 and the 7r-electron lobe at C6. This transforma¬, tion, sometimes called a homoallylic rearrangement, is closely analogous to leairangements, , observed, , during, , heterolytic, , additions, , to, , norbornadiene;, , for, , example,, , 196(a), , and similar also to certain rearrangements observed during solvolyses of norbornenyl derivatives., , 196(b), , EXERCISES FOR CHAPTER, , 16, , 1. Consider the following sequence for the homolytic halogenation of a hydrocarbon RH:, hv, , T2 ~^ 2Z* (initiation), kt, X-, , R. -f, 4- HZ, Hi’, + RH —R—*, , (propagation), R- + Z2 —> RZ + Zk<, 2R- —* R2 (termination), Using the steady-state approximation with respect to radical intermediates Z- and, R , derive an expression for the rate of the reaction, </(RZ)/dt., , rearrange, that is, for displacements of the types,, and, , Al", , +H—Ar'^Ar—H, , O, R, , + -Ar', , O, , C~°' + ArOAr' -* R—C—OAr + -OAr', , and Schreiber, ibid., 72, 5795, , TmO). ^, , S°°'', , 78, 592, , (J, , 956^, , W Winstein, Walborsky,
Page 778 :
762, , Free-radical Reactions, , 2. The mixture of monochlorohexanes obtained from the vapor phase chlorination of, 3-methylpentane at 450° has the following composition:, , ClCH2CH2CHMe, , CH3CHCI—CHMe, , Et, , Et2CHCH2Cl, , Et2CClMe, , 14 percent, , 14 percent, , Et, , 28 percent, , 44 percent, , Predict the composition of the mixture of monochlorohexanes obtained in the, chlorination of 2-methylpentane under the same conditions., 3. Consider the sequence for decomposition of a peroxide ROOR in a solvent SH:, k1, , ROOR —> 2RORO- -f SH —^ ROH + Ski, , S- + ROOR —> SOR + ROki, , 2S- -> S2, ki, , 2RO-, , nonactive products, , *, , RO- + S-, , (possible termination steps), , ROS, , Assume that the concentration of SH remains constant. Taking into account both the, unimolecular (spontaneous) decomposition and the induced decomposition, show, that the rate expression may take the following forms, depending upon the mode of, chain termination:, (a) rate = -—--- = h(ROOR) + A'(ROOR)**; if termination is by reaction, at, , between two S- radicals., (b) rate = h(ROOR) + A:"(ROOR)if termination is by reaction between two, RO- radicals., (c) rate = Ar'"(ROOR); if termination is by a reaction between RO- and S-., (d) Show that the rate law in (a) is applicable also to the decomposition of a diacyl, peroxide via the following path:, O, II, R—C—O—, , ki, , 2RCOO-, , kt, , RCOO-, , R- + C02, O, , O, ki, , R- +, , R—C, , O, , R—C—OR + RCOO-, , ki, , 2R- —> R2 (termination), 4, , Predict the major products resulting from each of the following reactions. Justify your, guess in each case:, , (a) Me2PH, , +, , O, , Bz202, , CH.
Page 782 :
766, , Free-radical Reactions, , (y) Me3CO—C—O—O—C—OCMe3->, heat, , o, , o, benzene, heat, , 6. Consider the homolytic addition of a halogen, X2, to a C=C double bond:, hv, , X2, , ki, , 2X’ (initiation), ', , \, *• +, , c=c, , /, , ', , /*., , ->, , V-, , I, , •, , c- c—, I, , X, , X, , i, , I, , + Xi, , kt, , -» —C—C-b, |, , ki, , |, , X'-> (propagation), etc., , I, , (a) Show that if chain termination involves two haloalkyl radicals, the rate law will, be of the form,, rate =, where, , k(Xd*fl*, , I is the intensity of the radiation bringing about dissociation of X2. Show, , that this is true whether or not the addition step is reversible., (b) Show that if chain termination involves two, , rate =, , k(, , C—C, , ], , X- radicals, the rate law becomes, (X2), , (c) Show that if termination involves a reaction between, , law becomes, rate =, , X• and —C—C—, the rate, , y\ _ /v4 (Xt)I*, V/ v, , k(, , C—C, , J, , (Assume that the rate of initiation is much less than the rate of propagation, is, that the kinetic chains are long.), 7. Suggest mechanisms for each of the following transformations:, hv, , (a) Ph3CCPh3 + CH2N2 —»> Ph3CCH2CPh3, , OH, , OH, , that
Page 785 :
Exercises for Chapter 16, , (c), , Azonitrile CLV decomposes about 75 times as rapidly as does azonitrile CLVI., CN, , CN, , CN, , |, Me, , CN, , I, , I, , Me3C—C—N=N—C, , CMe 3, , I, , EtCHMe—C—N=N—C—CHMeEt, , I, , I, , Me, , Me, , Me, CLVI, , CLV, , (d), , 769, , “Chlorophenylation” of MeCH=CHCOOMe with PhN2Cl and CuS04 (the, Meerwein reaction) yields ester CLVI I in which the /3-carbon has become, phenylated, whereas in the chlorophenylation of PhCH=CH, , COOMe, the, , a-carbon is phenylated., MeCHPh—CHClCOOMe, CLVII, com(e) Homolytic addition of HBr to 3-bromocyclohexene gives a trans dibromo, pound, whereas homolytic addition of HBr to 1-bromocyclohexene gives a cis, dibromo compound., , (0, , Homolytic additions to carbon-carbon triple bonds tend to be slower than, additions to double bonds under comparable conditions., , (g) The reaction of thiophenol with methyl acrylate gives a higher yield of simple, addition product in the presence of ;-BuOK than in the presence of /-BuOOH., , (h), , The “Wurster salt” obtained by oxidation of/>-phenylene diamine is more stable, in moderately acidic solutions than in either strongly basic or strongly acidic, solutions, whereas the radical obtained from oxidation of 4,4'-dihydroxydiphenyT, amine is more stable in strongly basic or strongly acid solutions than in neutral, , 0), , The organometallic compounds Ph3PbPbPh3 and Ph3GeGePh3 are not measur¬, ably dissociated into radicals at ordinary temperatures, even though Pb—Pb, and Ge, , (j), , Ge bonds are, in general, much weaker than carbon-carbon bonds., , Alkyl nitrates, which readily yield alkoxy radicals when heated, are nevertheless, poor initiators for free-radical chain reactions., , 00, , The apparent hydrogen isotope effect in the homolytic bromination of PhCH2D, may be made to increase by removing HBr from the reaction mixture as it is, formed., , (1), , In the fluorination of C13CCHC12, the chlorocarbons C12C=CC12 and C2C16 are, formed, along with the expected product C13CCFC12., , (m), , Molecular oxygen may accelerate free-radical brominations but almost in¬, variably inhibits free-radical chlorinations., , (n) A mixture of tetralin and cumene undergoes autoxidation much more slowly than, does either pure hydrocarbon., , (o) The (small) yields of ethane and ethylene formed in the decomposition of, proptony perox.de in isooctane are not affected by addition of iodine or quinone, to the solution., n, (P), , The ratio of as to Irans addition products resulting from the addition of X,0, to, cyclohexene ts very nearly the same as that resulting from the addition of NO.C1, to this olefin., 2, , (q) Hie extent of rearrangement in the decarbonylation of PhCMe2CH2CHO is, greater when the reaction is carried out in a dilute solution of chlorobenzene, than when carried out in the absence of solvent.
Page 786 :
770, , -, , Free-radical Reactions, , (r) Polymerization of styrene in the presence of n-butyraldehyde yields polymers of, much higher molecular weight than polymerization of styrene in the presence of a, corresponding amount of n-butyl mercaptan., (s) The ratio of isomerization to addition in the photochemical reaction of bromine, with cis-1,2-dichloroethylene is independent of the intensity of incident radiation, and is not affected by addition of traces of 02., (t) When two methyl radicals in the vapor phase collide, they are much more likely, to dimerize than are two H- radicals., (u) Benzyl iodide undergoes photochemical exchange with radioactive iodine more, rapidly in CCI4 than in, the reaction.), , CCI2, , CC1CC1==CC12. (Neither solvent is consumed in, , (v) The electrolysis of the deuterated acid CD3CH2COOH yields at the anode not, only D3CCH2CH2CD3, but also CD2, CH2. The yield of this olefin increases as, the intensity of the current is increased.
Page 787 :
Author Index, Abeles, R. H., 549, Abell, P. I., 733, Adam, J., 685, Adams, K. H., 602, Adamson, D. W., 629, Addison, L. M., 45, Adkins, H., 336, 339, Ainsworth, J., 73, Akawie, R. J., 152, 636, Albin, J., 529, Alder, K., 537, Alderman, D. M., 741, Alexander, E. R., 442, 501,, 646, Alfrey, T., 730, Allen, P. W., 69, 345, Allinger, J., 579, Allred, A. L., 569, Altscher, S., 281, Altschul, L. H., 332, Amis, E. S., 186, Anbar, M., 345, Anderson, A. G., 416, Andrews, L. J., 174, 442,, 443, 537, Angyal, S. J., 133, Archer, B. L., 334, 622, Arcus, C. L., 355, Arganbright, R. P., 499,, 573, 743, Arimoto, F. S., 690, Armstrong, R., 596, 706, Arnold, R. T., 151, 352,, 355, 628, 639, Arroya, J., 627, Asami, R., 619, Aschner, T. C., 545, Ash, A. B., 697, Ashdown, A. A., 334, Asinger, F., 703, Asperger, S., 479, Aston, J. G., 239, Augood, D. R., 720, Avakian, S., 133, Aycock, B. F., 733, Bachmann, W. E., 608,, 627, 673, 681, 727, Bacon, R. G. R., 690, 730, Baddeley, G., 238, 447, 451, Bader, A. R., 734, Badger, G. M., 83, Baird, R., 579, 639, 757, Baker J., W., 49, 396, 433,, 438, Baker, W., 417, Balfe, M. P., 267, Ballentine, A. R., 457, Ballinger, P., 208, 284, 526, Baltzer, O., 298, Baltzley, R., 281, 654, Bamford, W. R., 684. 706, Barnes, R., 529, , Barrett, E., 543, Barsky, G., 164, Bartlett, P. D., 72, 137, 145,, 164, 176, 177, 206, 279,, 280, 284, 294, 304, 366,, 371, 373, 374, 419, 522,, 534, 543, 548, 568, 570,, 572, 586, 595, 598, 611,, 622, 633, 687, 695, 714,, 718, 719, 758, Barton, D. H. R., 81, 240,, 501, 503, 505, 638, Barton, W. H., 743, Basolo, F., 102, Bastiansen, O., 413, Bateman, L., 161, 186, 256,, 701,706, Bates, R. G., 103, Batten, T. J., 707, 714, Baxendale, J. H., 690, Bayles, J. W., 232, 304, Beal, P. F., 628, 755, Beel, J. A., 138, Beesley, R. M., 568, Bell, E. R., 710, 713, Bell, F., 355, Bell, R. P.,110,111,113,227,, 232, 320, 372, 380 , 540, Bender, M. L., 316, 320, 331, Benfey, O. T., 211,256, 274, Benjamin, B. M., 523, 604, Benkeser, R. A., 462, 723, Bent, H. E., 676, Berenbaum, M. B., 726, Berger, A., 629, Bergmann, E., 402, Bergstrom, C. G., 588, Beringer, F. M., 704, Bergkvist, T., Ill, Berliner, E., 176, 332, 445,, 456, Bernies, H. L., 234, 474, Bernstein, H. I., 82, 371, Berson, J. A., 537, Berthelot, M., 319, Berthier, G., 64, Berthoud, A., 524, Beste, G. W., 256, Betts, R. L., 330, Bhatnager, S. S., 674, Bickel, A. F., 692, Biechler, S. S., 329, Bigeleisen, J., 192, 677, Bigelow, L., 445, 521, 698, Bird, M. L., 258, 312, Birkenbach, L., 444, Bjerrum, N., 202, Blacet, F. E., 689, Blackadder, D. A., 658, Blackall, E. L., 427, Blackwood, R. K., 296, BlaJes, A. T., 503, Blair, F. T., 627, , Blatt, A. H., 619, Blomquist, A. T., 716, Blumenthal, E., 345, Bockemuller, W., 354, 522, Boekelheide, V., 414, Boeker, G. F., 154, Boeseken, J., 538, Bohme, H., 561, Bohrer, J. J., 730, Bolland, J. L., 707, 709, Bonhoeffer, K. F., 391, 547, Bonner, W. A., 423, 437,, 579, 587, Boozer, C. E., 285, 294, 684,, 708, Bordwell, F. G., 131, 218,, 494, 499, 500, 501, 564,, 618, 742, Boschan, R., 566, Bottini, A. T., 463, Boudreaux, E. A., 457, Boulet, E., 457, Bowden, K., 675, Bown, D. E., 299, Boyd, G. V., 515, Boyer, J. H., 626, Boyles, FI. R., 403, Brader, W. H., 495, Bradley, W., 339, 610, Brady, J. D., 119, Braithwaite, D., 689, Branch, G. E. K., 325, Brand, J. C. D., 98, 446,, 734,739, Brass, K., 118, Braude, E. A., 192, Bredig, M. A., 680, Breslow, D., 392, 419, Breslow, R., 305, 397, Brewster, J. H., 565, 640, Brierglieb, G., 211, Briggs, E. R., 691, Brindell, G. D., 742, 760, Broadbent, H. S., 695, Broche, A., 390, Brockway, L. O., 29, 46, 217, Brjzfnsted, J. N., 110, 111,, 163, 185, 292, 542, Brooks, J. W., 713, Brotherton, R. J., 462, Brower, K. R., 716, Brown, B. R., 349, 382, Brown, C. J., 660, Brown, D. A., 184, 333, 660, Brown, F., 595, Brown, H. C., 119, 136, 180,, 200, 226, 231, 234, 240,, 259, 277, 278, 281, 434,, 435, 437, 448, 449, 450,, 474, 476, 477, 481, 483,, 549, 685, 697, 699, 700,, 702, 703, , 771
Page 788 :
772, , Author Index, , Brown, J. F., Ill, 139, 542, Brown, R. D., 723, Brown, W. G., 49, 335, Browne, C., 293, Brownstein, S., 656, Brugger, YV., 591, Bruson, H. A., 528, Brutschy, F. J., 545, Bruylants, A., 697, Buchanan, G. H., 164, Buchi, G., 750, Buchner, E., 729, Buck, J. S., 395, 396, Buckler, S. A., 705, Buckles, R. E., 271, 565, Buckley, R. P., 639, 734, Bunnett, J. F., 452, 453,, 455, 459, 462, Bunton, C. A., 261, 339,, 340, 342, 426, 630, 656,, 658, Burger, A., 629, Burl ant, W., 452, Burries, C. T., 716, Burske, N. W., 381, Burton, H., 451, 751, Burwell, R. L., 345, Buselli, A. J., 716, Bushwell, A. M., 296, Buting, W. E., 462, Butler, E. T., 635, Buttery, R. G., 751, Butz, L. W., 536, Buu-Hoi, Ng. Ph., 700, Buzbee, L. R;, 739, Bywater, S., 726, Calingaert, G., 739, Calvert, J., 689, Calvin, M., 380, 681, Campbell, A., 355, 625, 640, Campbell, B. K., 741, Campbell, K. N., 741, Cannell, L. G., 419, 592, Carboni, R. A., 598, Cardon, S. Z., 231, Cardwell, H. M. E., 382, Carlsmith, L. A., 461, Cartmell, E., 2, Cason, J., 328, 355, Cass, W. E., 414, 714, Catchpole, A. G., 532, Chambers, V. C., 601, Chanley, J., 140, Chao, T. H., 685, 703, Chapman, A. W., 618, Charlton, J. C., 586, Chatt, J., 118, Chattaway, F. D., 524, Chaudhari, D. K. R., 751, Christiansen, J. A., 171, Chu, T. L., 755, Chupka, W. A., 36, Church, M. G., 170, 262, Claisen, L., 134, Clark, M. T., 635, , Clarke, J. T., 714, Cleveland, F. F., 76, Cline, J. E., 676, Close, VV. J., 136, Clunie, J. C., 540, Cobb, R. L., 136, Coburn, W. C., 256, Cohen, M. D., 658, Cohen, S. G., 342, 570, 688,, 724,727, Cohn, M., 374, 431, 541, Cole, R. H., 81, Collins, C. J., 523, 579, 604,, 610, Comyns, A. E., 339, 419, Conant, J. B., 97, 164, 284,, 370, 543, Conley, J. B., 354, Conroy, H., 647, Convery, R. J., 721, Cook, C. D., 88, 681, Cook, N. C., 517, 733, Cookson, R. C., 240, Coombs, E., Ill, 391, Cooper, G. D., 218, 564, 729, Cooper, H. R., 708, Cooper, K. A., 263, 488, 489, Cope, A. C., 355, 503, 599,, 600, Coppinger, G. M., 286, Cormack, J. F., 459, Cotman, J. D., Jr., 758, Cottrell, T. L., 36, Coulson, C. A., 2, 13, 28, 48,, 58,214,529,588, Cowdrey, W. A., 270, 295,, 446, 563, 729, Cox, J. C„ 182, 419, 599, Cram, D. J., 472, 490, 492,, 502, 503, 551, 575, 579,, 581, 587, 611, 644, Crawford, R. J., 647, Crew, M. C., 605, Criegee, R., 539, 633, 706, Cristol, S. J., 345, 401, 479,, 495, 498, 499, 573, 742,, 743, 760, Croce, L. J., 174, Cromer, D. T., 51, Cross, R. P-, 543, Crossley, M. L., 457, Culbertson, G. R., 675, Cullis, C. F., 539, Cunneen, J. I., 701, 741, Curtin, D. Y„ 493, 551,605,, 607, 610, 631, 647, 755, Cutforth, H. G., 680, 681, Cymerman, J., 81, Dainton, F. S., 730, D’Aleio, G. F., 730, Danilov, S., 637, Dannley, R. L., 724, Darby, R. E., 344, Darwent, B. de B., 541, Datta, S. C., 319, , Dauben, VV. G., 240, 354,, 551, Davenport, D. A., 446, Davies, A. G., 267, 339, 344, Davies, D. S., 445, 729, Davies, W. C., 725, Davis, P., 691, Davis, T. W., 751, Dawson, H. M., 163, 374,, 375, 651, Day, A. R., 545, Day, J. N. E., 314, 319, Dean, P. M., 749, De la Mare, P. B. D., 234,, 254, 261, 275, 284, 288,, 440, 442, 510, 526, 530, Demerseman, P., 700, Denbigh, K. G., 161, Denney, D. B., 402, 450, 581,, 632, 633, 721, Denney, D. G., 633, Dennison, D. M., 16, Deno, N. C., 99, 107, 305,, 446, De Puy, C. H., 482, 490, Derbyshire, D. H., 442, 735, Dermer, O. C., 727, de Salas, E., 385, de Sousa, A., 320, De Tar, D. F., 457, 685,, 722, 760, de Vries, J. L., 419, Dewar, M. J. S., 428, 439,, 653, 706, De Wolfe, R. H., 286, 291, Deyrup, A. J., 96, 99, 103,, 304, 618, Dhar, M. L„ 472, 475, 481, Dickens, J. W., 214, Dickerman, S. C., 729, 730, lickinson, R. G., 737, liedrichsen, J., 644, lietrich, W., 633, )illon, R. L., 366, 380, 387, )illon, R. T., 494, 519, , )iPPy> J-, , F. G., 209, Jjerassi, C., 700, )obres, R. M., 674, )odson, R. M., 151, 352, )oering, W. E., 269, 280,, 305, 369, 372, 399, 415,, 15, 751, ierty, D. G., 332, iohue, J., 568, fman, E., 630, trovsky, I., 345, 584, 586, ighty, M. A., 267, /ding, A. L., 320, 3ps, R. D., 623, mmond, A. Y., 539, >oux, M., 320, raisse, C., 749, ican, J. F., 602, litz, J. D., 46, 416, ikle, F. B., 320
Page 789 :
Author Index, Dunn, G. E., 403, Durham, L. J., 711, Dyatkina, M. E., 2, 13, 28, Eaborn, C., 436, Earl, J. C., 417, Eberly, K., 335, Ecke, G. G., 517, 733, 739, Eddy, R. W., 272, Edgerton, P. J., 456, Edmison, M. T., 727, Edward, J. T., 329, Edwards, F. G., 718, Edwards, J. O., 260, Edwards, W. G. H., 354, Ehrenson, S. J., 284, Eidinoff, M. L., 380, Eiland, P. F., 46, Eisner, J. R., 759, Eitel, A., 547, Elderfield, R. C., 220, Eldred, N. R., 180, 259,, 278, Eleuterio, H. S., 648, Elhafez, F. A. A., 492, 502,, 551, Eliel, E. L., 242, Elmer, O. C., 151, 352, Elmore, G. V., 320, Emmons, W. D., 384, 740,, 741, England, B. D., 290, Ericks, K., 419, Errera, J., 30, Esson, W., 168, Evans, A. G., 304, 716, Evans, D. P., Ill, 391, Evans, M. G., 537, 690, 691, Evans, O. P., 382, Evans, W. V., 402, 689, Everard, K. B., 75, Ewald, L., 673, Eyring, H., 38, 159, 178,, 179, 184, Fahey, R. C., 285, Fainberg, A. H., 302, 582, Fair brother, F., 448, Fairclough, R. A., 348, Fankuchen, I., 413, Farkas, A., 682, 711, Farkas, E., 643, Farmer, E. H., 394, 504,, 531, 533, 706, Farnham, N., 631, 755, Fateley, W. G., 415, Faulkner, I. J., 139, Fava, A., 681, Fawcett, E. S., 346, Felletschin, G., 640, Fenton, G. W., 480, Fenton, S. W., 599, Ferguson, J. W., 608, Ferguson, L. N., 13, Fieser, L. F., 680, Fife, W., 646, , Fillet, P., 708, Finkelstein, M., 453, 700, Firestone, R. A., 647, Fischer, E. O., 416, Fischer, O., 652, Fisher, G. S., 711, 746, 760, Fix, D. D., 479, Fletcher, R. S., 474, Flett, M. S., 84, Flory, P. J., 695, 730, Flygare, H., 706, Flynn, J., 479, Fodor, G., 601, Fonken, G. S., 551, Fono, A., 710, 757, Forbes, G. S., 738, Ford, M. C., 700, Fordham, J. W. L., 710, Fournier, A., 599, Fowden, L., 234, Fowles, G. W. A., 2, Foy, M., 546, Fraenkel, G. K., 684, 755, Frampton, V. L., 166, Francis, A. W., 521, Franck, J., 719, Frankland, E., 489, Franzus, B., 345, Freamo, M., 752, Fredenhagen, H., 547, Freeman, J. P., 741, Freudenburg, K., 264, 570, Frey, H. M., 751, 752, Friedlander, H. Z., 151,238,, 744, Friedman, B. S., 741, Friedman, L., 319, Friedmann, H. B., 320, Friedrich, R., 520, Friess, S. L., 159, 631, 755, Froemsdorf, D. H., 482, Frost, A. A., 110, 159, 166,, 167, 178, 713, Fugassi, P., 543, Funk, H., 117, Fuson, R. C., 571, 705, Gainer, 1*1., 756, Gardner, G. K., 755, Gardner, H. J., 707, Garbisch, E. W., Jr., 453, Garner, A. Y., 399, 534, Garrett, A. B., 99, 322, Gassmann, A. G., 322, Gazith, M., 704, Gebhardt, A. I., 602, Geissman, T. A., 152, 548,, 646, Gelissen, H., 716, Gelles, E., Ill, 252, Gent, W. L. G., 61, George, P., 707, Gerard, M., 749, Gerstein, M., 240, Gettler, J. D., 174, Gilbert. R., 390, , 773, , Gildenhorn, H. L., 624, Gillespie, R. J., 98, 99, 420,, 425, 446, 451, Gillette, R. H., 84, Gilman, H., 133, 400, 402,, 461, 673, Gindler, M., 140, Ginger, R. D., 328, Gintis, D., 277, Gippen, M., 724, Girard, C., 646, Glasebrook, A. L., 751, Glasstone, S., 159, 178, Glazer, T., 304, 653, Gleave, J. L., 251, 488, Glick, C. F., 569, Gnanapragsam, N. S., 443, Goddard, D. R., 419, Goering, H., 500,733,735,742, Gold, V., 194, 258, 419, 423, Goldblatt, L. A., 760, Goldey, R. N., 401, Goldman, A., 705, Goldschmidt, H./653, Goldschmidt, S., 680, 681, Goldstein, B., 581, Gollub, M. C., 351, Golumbic, C., 570, Gomberg, M., 303, 673, Goodman, L., 611, Gordon, A., 751, Gordon, J. J., 382, Gordon, M., 238, 416, Gordy, W., 40, 60, 76, Goring, J. H., 118, Gorvin, J. H., 330, Goubeau, J., 444, Gould, E. S., 52, 260, Goulden, J. D. S., 425, Graham, H., 326, Granick, S., 678, Grayson, M., 448, Greene, F. D., 490, 638, 687,, 701, 750, Greenstreet, C. H., 261, Greenwood, F. L., 701, Gregg, R. A., 741, Greiner, R. VV., 569, Grelicki, C., 680, Griffing, M. E., 739, Grim, A. G., 254, Grimmel, H. W., 88, Grison, P. E., 419, Griswold, P. H., 118, Groebel, P., 394, Groll, H. P. A., 737, Grovenstein, E., Jr., 445,, 495, Groves, L. G., 65, Grundemeier, W., 658, Grunwald, E., 101,107, 256, 258, 271, 302, 564, 575, Guggenheim, E. A., 111,542, Guldberg, C. M., 319, Gutbezahl, B., 101, 107, Gwynn, D., 646
Page 790 :
774, , Author Index, , Haeussermann, C., 461, Haflinger, O., 200, Haggerty, C. J., 690, Hagman, S. M., 563, Haines, R. M., 254, Halberstadt, F. S., 428, Hall, H. K., 232, Halverson, F., 81, Hamer, J., 626, Hamilton, G. E., 708, Hammett, L. P., 95, 96, 98,, 99, 103, 153, 171, 190,, 200, 220, 237, 256, 274,, 304, 325, 330, 374, 618, Hammick, D. L., 349, 660, Hammond, G. S., 174, 194,, 231, 344, 376, 378, 438,, 658, 684, 708, 714, Hannum, C., 518, 733, Hantzsch, A., 304, Happe, VV., 641, Harned, H. S., 101, 650, Harris, E. E., 161, Hart, H. S., 648, Hartman, R. J., 322, Hartogs, J. C., 655, Hass, H. B., 696, Haszeldine, R. N., 83, 355,, 744, Hatch, M., 644, Hause, N. L., 498, Hauser, C. R., 337, 392, 419,, 472,479,621,622,623,641, Hawes, B. W. V., 194, 423, Haworth, R. D., 635, Hawthorne, M. F., 384, Hay, A. S., 440, Head, A. J., 501, Heald, K„ 274, Hearne, G. W., 737, Heath, D. F., 16, Hebbelynck, N. F., 700, Heck, R., 569, 579, 639, 757, Hedges, R. M., 438, Heine, H. W., 569, Heine, R. F., 529, Heintzeler, M., 355, Heinz, W. E., 353, Heilman, H. M., 530, Helmreich, W., 741, Henderson, R. B., 293, 569, Hendley, E. C., 635, Hendricks, S. B., 51, Henglein, A., 684, Henne, A. L., 518, Hepp, E., 652, Herbert, J. B. M., 345, Herbrandson, H. F., 294, Hermans, P. H., 716, Herrmann, R. A., 691, 703, Herz, VV., 599, Herzberg, G., 34, 42, 76, Hess, H. V., 271, Hewett, I. V., 742, Hey, D. H., 480, 688, 714,, 722, 725, , Heyne, G., 448, 449, Hiatt, R. R., 719, Hickenbottom, VV. J., 653, Higginson, W. C. E., 113,, 691, Hill, D. G., 479, Himel, G. M., 233, Hinds, VV. H., 458, Hine, J., 106, 208, 250, 282,, 284, 371, 381, 397, 398,, 495, 635, 646, Hine, M., 106, 381, Hinkley, D. F., 490, 518, Hinshelwood, C. N., 318,, 322, 447, 658, Hippie, J. A., 36, Hiron, F., 565, Hirschfelder, J. O., 180, Hirshon, J. M., 684, 755, Hirzel, H., 628, Hitz, F., 711, Hlyinsky, A., 760, Hodgdon, R. B., Jr., 649, Hodges, R., 354, Hodnett, E. M., 479, Hoegger, E. F., 498, Hofeditz, W., 683, Hoffman, F. W., 354, 522, Hoffman, R. A., 727, Hoffmann, A. K., 372, 399, Hofmann, A. W., 480, 653, Holleman, A. F., 437, 438,, 456, 655, Holmes, H. L., 536, Holmquist, H. E., 345, Holness, N. J., 241, Holst, K. A., 329, Holt, P. F., 680, Horeld, G., 728, Horn, E., 683, Hornhardt, H., 644, Houssa, A. J. H., 295, Houston, A., 640, Howe, K. L., 500, 572, Howlett, K. E., 505, Hoyle, G. K. E., 524, Hsu, S. K„ 373, 374, 386, Huang, R. L., 747, Huckel, W., 21, 295, 414,, 497, 593, 749, Hudson, R. F., 184, 333, 334, UggCtt’, uggins, S’, M. L., 36, 45, ughes, B. P., 680, ughes, E. D., 147,160,161,, 170, 173, 175, 182, 184,, 234, 251, 252, 256, 258,, 261, 262, 263, 267, 268,, 270, 272, 290, 295, 326,, 340, 385, 398, 409, 426,, 427, 428, 440, 446, 473,, 475, 476, 478, 481, 487,, 497, 530, 565, 584, 586,, <;<n hqs 655. 656. 660,, 706, Huisgen, R., 577, 728, , Hunig, S., 225, Hurd, C. D., 645, 649, Hurwitz, M. J., 755, Hussey, R. E., 284, Hutchison, C. A., Jr., 754, Hutton, R. F., 149, 549, 628, Hyde, J., 327, Ibbotson, K., 456, Ichikawa, K., 549, Ide, VV. S., 395, 396, Iffland, D. C., 296, Ingberman, A. K., 729, Ingold, C. K., 100, 120,, 147, 160, 170, 175,, 206, 208, 228, 234,, 251, 252, 254, 256,, 261, 262, 263, 267,, 270, 272, 275, 279,, 314, 315, 319, 326,, 366, 374, 385, 386,, 398, 412, 419, 421,, 427, 428, 433, 438,, 446, 452, 472, 473,, 476, 478, 480, 481,, 487, 489, 497, 505,, 535, 546, 548, 563,, 568, 584, 593, 595,, 660, Ingraham, L. L., 36,, 292, 353, 562, 575, Inman, C. G., 750, Ipatieff, V. N., 741, Ivin, K. J., 692, 730, Izard, E. F., 524, , 141,, 184,, 250,, 258,, 268,, 295,, 340,, 393,, 426,, 439,, 475,, 482,, 530,, 565,, 658,, 271,, , acobson, P., 645, affe, H. H., 186, 220, 222,, 225, ager, A., 736, agow, R. H., 285, ames, J. C., 98, ander, J., 83, aquiss, M. T., 713, arrett, H. S., 677, aruzelski, J. J., 305, effrey, G. H., 419, enkins, A. D., 684, ensen, E. V., 450, ohannesen, R. B., 231, ohansson, H., 563, ohnson, A. D., 684, ohnson, H. VV., 647, ohnson, M., 457, ohnson, VV. S., 353, ohnston, F., 517, olley, J. E., 689, ones, G. T., 655, ones, H. VV., 258, 312, 564, ones, J. R-, 304, ones, L. VV., 624, ones, R. G., 402, ones, R. N., 81, ones, VV. J., 528, ones, VV. M., 537, ordan, VV. E., 490
Page 791 :
Author Index, Joubert, J. M., 413, Jungk, H., 136, 448, 449, Kadesch, R. G., 71, Kahler, YV. H., 371, Kahn, N. A., 706, Kainer, H., 596, 677, Kalm, M. J., 355, Kalnin, P., 392, Kantor, S. W., 622, 641, Kaplan, J. F., 233, Kappelmeier, P., 380, Karle, I. L., 42, Karle, J., 29, 42, 73, Kaschanov, L. J., 116, Kaspar, R., 633, Kassinger, R., 529, Katellar, J. A. A., 214, Katz, L., 72, Kaufman, H. S., 413, Kaufmann, H., 163, Kaye, S., 452, 518, Keefer, R. M., 442, 443, 537, Kellert, M. D., 701, Kellom, D. B., 493, Kemp, K. C., 328, Kenner, J., 456, 629, Kenyon, J., 265, 267, 269,, 335, 339, 342, 343, 344,, 355, 625, 640, Kepner, R. E., 290, Kerr, E. R., 739, Ketley, A. D., 440, Khan, N. A., 740, Kharasch, M. S., 140, 420,, 518, 531, 535, 546, 685,, 687, 688, 690, 692, 697,, 699, 702, 703, 704, 710,, 733, 744, 746, 747, 755,, 757, Kice, J. L., 633, Kidd, H. V., 141, Kilby, D. C., 445, Kilpatrick, Martin, 100, 163, 292, Kilpatrick, Mary, 163, 292, Kimball, G. E., 38, 150, Kincaid, J. F., 645, Kindler, K., 209, 318, Kirby, R. H., 400, Kirkwood, J. G., 202, 552, Kirner, W. R., 284, Kistiakowski, G. B., 38, 692,, 751, Klages, F., 39, Klein, H. C., 452, Kleinberg, J., 690, Klemchuk, P. P., 450, 721, Klemperer, W., 569, Klevins, H. B., 87, Kline, M. W., 640, Kloetzel, M. C., 536, Kluiber, R. W., 646, Knell, M., 131, Knox, L. H., 280, 308, 415,, 622, 752, , Koch, K., 705, Koch, S. D., Jr., 698, Kochi, J. R., 729, Koelsch, C. F., 677, Koerner, O., 116, Kohler, E. P., 742, Kohnstam, G., 284, Kolka, A. J., 739, Kolthoff, I. M., 690, Kondakov, I. L., 118, Kooyman, E. C., 697, 746, Kornblum, N., 296, 759, Kovacs, O., 601, Kowalsky, A., 754, Kraus, C. A., 97, 102, 371, Kreevoy, M. L., 252, 272, Kresge, A. J., 714, Krieble, R. H., 329, Krieger, K. A., 545, Kritchevsky, J., 531, Kuderna, B. M., 747, Kuderna, J., 685, Kuffner, F., 354, Kuhara, M., 618, Kuhn, L. P., 84, 341, Kuhn, R., 677, Kuhn, W., 552, Kuivila, H. G., 99, 436, Kung, F. E., 549, 599, Kutz, W. M., 336, 339, Kwart, A., 611, 695, Lacher, J. R., 744, Lachman, A., 412, 636, Ladbury, J., 539, La Flamme, P., 751, Lagrave, R., 638, Laidler, K. J., 159, 160, 178,, 184, 714, Lambourne, L. J., 444, La Mer, V. K., 179, 472, Lampert, B. B., 618, Landis, P. S., 500, 501, Lane, C. E., 628, Langer, A., 36, Langford, P. B., 398, Lapkin, M., 726, Lapporte, S., 757, Lapworth, A., 208, 543, Larsen, D. W., 735, 742, Lauer, W. M., 423, Laughlin, R. G., 751, Lautsch, W., 683, Lawrence, C. D., 734, Leavitt, F., 734, Lee, C. C., 597, Lee, D. E., 282, 495, Leermakers, J. A., 683, Le Fevre, R. W. J., 57, Leffek, K., 254, Leffler, J. E., 402, 632, 672,, 718, ’, Le Goff, E., 402, Legutke, G., 497, Leighton, P. A., 525, Leisten, J. A., 98, 326, 341, , 775, , Leittke, W., 81, Le Maistre, J. W., 472, Lennard-Jones, J. E., 23, Le Roux, I. J., 453, Letort, M., 708, Letsinger, R. L., 389, Levene, P. A., 295, 563, Levine, S., 704, Levitt, A., 455, Levitz, M., 280, Levy, J. B., 515, 688, 740, Levy, L. K., 369, Levy, M., 717, 718, Lewis, D. YV., 757, Lewis, E. S., 280, 285, 286,, 294, 371, 458, Lewis, F. M., 688, 725, Lewis, G. N., 11, 115, 305,, 677, Lewis, I. C., 230, Lewis, J., 45, Lewis, S. N., 494, Lewis, T. A., 630, Lichtin, N. N., 304, Lindegren, C. R., 562, Linnett, J. YV., 16, 214, Linstead, D., 352, Lipkin, D., 305, 677, Lipkinsky, E. S., 750, Lippincott, E. R., 415, Lippincott, YV. T., 689, Lipscomb, YV. N., 72, Liu, Y. C., 755, Livingston, R., 42, Llewellyn, D. R., 630, Lobunez, YV., 720, Lock, G., 547, Loening, K. L., 322, Loftfield, R. B., 158, 643, Lohman, L., 641, Lohmann, R. H., 256, Lombardino, J. G., 692, 726, Long, F. A., 103, 190, 293,, 319, 320, 515, 722, Longuet-Higgins,, H., C.,, 214, 372, 414, 585, Lord, R. C., 83, Lossing, F. P., 504, 713, Lovgren, T., 547, Lowry, T. M., 139, 542, Luborski, F. E., 100, Lucas, H. J., 169, 516, 519,, 567, 574, Luder, YV. F., 93, Luning, O., 134, Lutz, R. E., 293, 490, LuValle, J. E., 705, Lynn, K. R., 602, Lyons, B. J., 344, Maccoll, A., 505, Mackie, J., 234, Magel, T. T., 305, Mamalis, P., 372, Mangold, R., 640, Manske, R. H. F., 208
Page 792 :
776, , Author Index, , Marchand, B., 539, Mark, H., 413, 677, 730, Maron, S. H., 381, Marsden, R. J. B., 65, Marshall, H., 256, 562, 588, Mathes, W., 303, Martin, E. L., 136, Martin, J. C., 72, 598, Martin, R. H., 700, Martin, R. J. L., 268, Martinsen, H., 420, Martius, C. A., 653, Marvel, C. S., 233, 502, 563,, 675, Mason, S. F., 660, Masson, G. R., 689, Masterman, S., 269, 295,, 339, 487, Matheson, M. S., 688, 695,, 725, Mathews, J. H., 652, Mathur, K. M., 674, Mauguin, C., 128, 621, Maw, G. A., 481, 482, Mayer, F., 88, Mayerle, E. A., 337, Maynert, E. W., 705, Mayo, F. R., 140, 517, 531,, 691, 706, 718, 730, 733,, 741, Mazur, R. H., 589, McBee, E. T., 696, McCabe, L., 569, McCarron, F. M., 389, McCarty, J., 503, 587, McCollum, J. D., 145, 548, McCormack, W. B., 688, McDaniel, D. H., 200, McDevit, W. F., 320, McElhill, E. A., 218, 708, McEwen, W., 136, 690, McGary, C. W., 226, 434, McGregor, A., 636, McKay, F. C., 459, McKenzie, A., 149, McKettrick, D. S., 325, McLafferty, F. W., 698, McNamara, J. H., 500, MacNulty, B. J., 487, Meacock, S. C. R., 329, Meadows, G. W., 541, Medalia, A. I., 690, Medard, L., 153, 420, Meek, J. S., 401, Meerwein, H., 591,594, 595,, 729, Mehta, T. N., 394, Meigh, D. F., 268, 339, Meisenheimer, J., 454, 619, Meislich, E. K., 607, Melander, L., 423, Melville, H. W., 708, 746, Merrifield, R. E., 83, Mesrobian, R. B., 688, 710, Meyer, K. H., 377, 380, Mhala, M. M., 658, , Michael, A., 520, 525, Michaelis, L., 154, 674, 678,, 679, 681, 705, Michel, R. H., 220, Milas, N. A., 710, Millen, D. J., 420, 425, Miller, A. A., 611, 706, 730, Miller, A. D., 569, 743, Miller, J. G., 330, 720, Miller, M. L., 179, 472, Miller, R. J., 345, Miller, W. T., 698, Millet, H., 651, Mills, R. H., 355, Mislow, K., 530, Misrock, S. L., 750, Modic, F. J., 438, Moeller, T., 93, Moelwyn-Hughes, E. A.,, 262, 457, Moffitt, W. E., 416, 433,, 529, 588, Mok, S. F., 272, 584, Mole, T., 428, Moller, E., Ill, Moller, F., 649, Moore, R. F., 709, Moore, W. J., 101, 159, 168, Moreland, W. T., 203, Morgan, P., 355, Mori, T., 563, Moritani, I., 477, 481, 483, Morris, A. L., 706, Morris, J. C., 188, Morse, B. K., 269, Mortensen, E. M., 65, Morton, I. D., 527, Moseley, R. B., 299, Mosettig, E., 629, Mosher, H. S., 402, 545, 711, Mosher, W. A., 599, Mosset, M., 524, Moureau, C., 749, Mudrak, A., 501, Muhr, G., 175, Mulcahy, M. F. R., 708, Muller, E., 753, 754, Muller-Rodloff, I., 753, Mulliken, R. S., 40, 49, 748, Mumm, O., 644, 649, Murphy, G. W., 182, 649, Musgrave, W. K. R., 529, Muskat, I. E., 530, Mustafa, A., 750, Muus, J., 348, Nace, H. R., 501, Nador, K., 601, Nagel, S. C., 623, Nakagawa, M., 483, Nakaten, H., 728, Neber, P. W., 644, Nelson, A. F., 738, Nelson, J. A., 416, Nelson, W. E., 635, Neuhoff, H., 753, , Nevell, T. P., 595, Neville, O. K., 635, Newman, M. S., 73, 99, 318,, 322, 324, 325, 472, 624,, 628, 755, Nichols, P. L., Jr., 292, Niclause, M., 708, Nicolaides, N., 755, Niederlander, K., 117, Nilsson, H., 567, Nobel, J. A., 320, Noland, W. E., 423, Noller, C. R., 402, Nolin, B., 81, Norris, J. F., 332, 413, Norris, J. H., 254, Norris, W. P., 495, Norton, H. M., 316, Noyce, D. S., 551, 601, Noyes, R. M., 691, 692, 703,, 704, 705, 735, 737, Noyes, W. A., Jr., 622, 689,, 738, Nozaki, K., 524, 714, Nudenburg, W., 685, 690,, 710, 755, 757, Nuzak, K., 738, Oae, S., 274, Ochs, C., 677, 681, O’Connor, G. L., 501, Offenbacher, M., 681, Ogg, R. A., 524, Ogston, A. G., 572, Okamoto, Y.., 226, 281, 434,, 477, 481, Olander, A., 117, Oldroyd, D. M., 760, Ollis, W. D., 417, Ollson, H., 181, Olson, A. C., 142, 633, Olson, A. R., 327, 345, 651, Opotsky, V., 397, Orgel, L. E., 416, Orloff, H. D., 739, Orton, K. J. P., 650, O’Shaughnessy, M. T., 717, Oumonov, A. T., 637, Overberger, C. G., 692. 725,, 726, 756, Overhults, W. C., 401, Owen, B. B., 701, Page, M., 81, Palmer, K. J., 217, Paneth, F. A., 683, Park, G. S., 690, Park, J. D., 744, Partington, J. R., 57, Partridge, S. M., 343, Passaglia, E., 711, Pastor, R. C., 754, Patai, S., 273, 584, Patel, C. S., 251, Patrick, T. M., 747, Paul, I. I., 547
Page 793 :
Author Index, Paul, M. A., 103, 190, 237,, 515, Paul, R., 400, Pauling, L., 18, 28, 40, 46,, 50, 58, 748, Pauson, P., 416, Paviak, S. C., 638, Pearl, I. A., 546, Pearlson, YV: H., 166, Pearson, R. G., 102, 110,, 159, 167, 178, 337, 366,, 380, 381, 499, 500, 713, Pearson, T. G., 683, 751, Pearson, YV. V., 402, 689, Pease, YV. F., 695, Pederson, K. J., 346, 348,, 350, Pegues, E. E., 273, Pellon, J. J., 714, 716, Penney, YV. G., 48, Pepinsky, R., 46, Perez-Ossorio, R., 385, Perren, E. A., 393, Peterson, R. C., 453, 700, Petrenko-Kritchenko, P., 397, Petropoulos, C. C., 632, Pettit, H., 325, 414, Pfanz, H., 754, Pfeiffer, P., 490, Pfluger, H. L., 374, Phillips, A. P., 654, Phillips, H., 265, 269, 343,, 578, Pilz, H., 706, Pinckard, J. H., 87, Pines, N., 741, Pitt, B. N., 131, Pittman, V. P., 269, Pitts, J. N., Jr., 691, Pitzer, K. S., 2, 13, 239, 240, Platt, J. R., 87, Pockel, I., 595, Pocker, Y., 261, 272, 273,, • 584, Podall, H., 277, Pode, J. S. F., 538, Pohmer, L., 462, Polanyi, M., 145, 683, 716, Pollack, M. A., 649, Poole, V. D., 417, Poplett, R., 267, Porter, G. B., 638, 689, Poshkus, A. C., 757, Post, B., 52, Potter, H., 742, Praill, P. F. G., 451, Prelling, E. R. A., 398, Prelog, V., 546, 549, 551,, 599, 601, Pressman, D., 494, Prestt, B. M., 284, Price, C. C., 220, 721, Pritchard, FI. O., 293, Prue, J. E., 351, Pruett, R. L., 529, Pruitt, K. M., 453, , Pryde, D. R., 651, Pudovik, A., 530, Purcell, R. H., 751, Purchase, M., 320, Purlee, E. L., 515, Quayle, O. R., 316, Quinn, M. G., 456, 706, Raaen, H. P., 610, 690, Rabinovitsch, M., 440, Rabinovvitch, E., 719, 738, Rainsford, A. E., 472, Raisin, C. G., 100, Raley, J. H., 162, 713, 733, Rammelt, P. P., 738, Ramsperger, H. C., 724, 725, Raney, D. C., 402, Rasmussen, R. S., 82, Rassack, R. C., 133, Rath, C., 298, Rauhut, M. M., 459, Raulins, R., 644, Rausch, M. D., 690, Read, J., 522, Rebbert, R. E., 714, Reber, R. K., 154, Reboul, E., 518, Reed, R. I., 160, 426, Reeder, J. A., 760, Reeve, YV., 292, Reiche, A., 711, Reichstein, T., 131, Reid, E. E., 328, 380, Reinheimer, J. D., 453, Reinmuth, O., 744, Relyea, D. I., 500, 742, 760, Rembaum, A., 717, 718, Remers, YV. A., 701, Renfrow, YV. B., Jr., 337,, 621, 623, Reynolds, R. D., 537, 644, Rhoades, S. J., 644, Rice, F. O., 680, 683, 751,, 752, Rice, O. K., 50, 724, Richter, J., 639, Ridd, J. H., 175, Rideal, E. K., 707, Ridge, M. J., 707, Rieche, A., 705, Rieke, C. A., 49, Riesz, R. YV., 515, Rietz, O., 374, Riibcr, C. N., 533, Ritchie, M., 695, Rittenhouse, J. R., 720, Ro, R. S., 242, Robb, J. C., 746, Roberts, I., 150, Roberts, J. D., 172,203,2, 274, 400, 461, 463, 5, 589, 591, 596, 597, 5, 601, 607, 635, 692, 7\1, Robertson, A., 707, 710,', , 777, , Robertson, J. PI., 42, Robertson, P. YV., 444, 524,, 527, 528, 529, Robertson, R., 339, Robinson, C. A., 582, Robinson, R. A., 102, Rodebush, YV. H., 296, Roger, R., 636, Roitt, I. M., 724, Rondesvedt, C. S., 730, Rose, J. B., 497, 593, Rosen, R., 97, Rosenmund, K. YV., 653, Ross, S. D., 453, 562, 700, Ross, YV. C. J., 571, Rossini, F. D., 38, Rothen, A., 295, Rothrock, H. S., 295, Rothstein, E., 451, Rowe, J. L., 692, Rubin, YV., 537, Rudesill, J. T., 344, Ruf, H., 254, Rumpf, P., 678, Rundle, R. E., 118, 585, Russell, G. A., 697, 699, 700,, 706, 730, Russo, C., 354, Rust, F. F., 162, 695, 697,, 713, 714, 733, Rutherford, A., 98, Ruzicka, L., 591, Rydon, H. N., 372, Rytina, A. YV., 536, Sadie, A., 292, Saegebarth, K. A., 538, Sage, M., 746,, Saigh, G. S., 751, Saint Gilles, L. P. de, 319, Salama, A., 254, 526, Salomon, G., 570, Samuel, D., 345, Sanderson, R. T., 40, Sandin, R. B., 438, 440, Sandy, A. C., 337, Sanford, J. K., 627, Satchell, D. P. N., 190, 419,, 446, Sauerland, H. D., 593, Saunders, YV. H., 597, Saunders, YV. H., Jr., 479,, 482, 627, Saville, B., 334, Saville, R. YV., 451, Sayigh, A., 280, Saytzeff, A., 480, Schaad, L. J., 714, Schaeffer, H. J., 523, 604, Schaeffer, YV. D., 587, Schaffel, G. S., 543, Schenck, G. O., 749, Schenkel, H., 349, 353, Schenkel-Rudin, M., 353, Schenker, K., 599, Schildknecht, C. E., 730
Page 794 :
778, , Author Index, , Schlenk, W., 402, Schlenk, W., Jr., 402, Schlesinger, H. I., 231, Schlitt, R., 646, Schmerling, L. R., 449, 645, Schmid, H., 143, 175, 644, Schmid, K., 143, 644, Schmidt, E., 518, Schmidt, G. S., 711, Schmidt, M. T., 337, Schmitz, H., 736, Schneider, A., 342, Schnurr, W., 653, Schoenthaler, A. E., 495, Schomaker, V., 43, 568, 737, Schonberg, A., 750, Schorigin, P., 97, Schramm, R. M., 426, Schreiber, K., 581, 597, 761, Schriescheim, A., 305, Schroeder, W., 723, Schroeter, G., 628, Schubert, M. P., 678, 679, Schubert, W. M., 353, Schuler, F. W., 649, Schumacher, H. J., 736, 738, Schwabe, K., 738, Schwartz, J. R., 658, Schwarzenbach, G., 107,, 377, 681, Schwechten, H. W., 677, Schwemer, W. C., 166, Scott, A. D., 184, 267, 295, Scott, C. B., 256, 299, Scott, W. D., 531, 533, Scatchard, G., 184, Seffe, R. J., 744, Sehon, A. H., 745, Sell, K., 561, Seitz, H., 651, Selwood, P. W., 674, 680,, 681, Semenow, D., 461, Sen, J. N., 684, 708, Seubold, F. H., 692, 714,, 755, Shadan, A., 345, Shapiro, U. G., 173, Sharp, D. W., 304, Sharpe, A. G., 355, Shatavsky, M., 598, 761, Shavel, J., Jr., 741, Shaw, F. R., 433, Shenk, W. J., Jr., 705, Shepherd, N., 304, Sherk, K. W., 627, Sherman, R. H., 371, Sherrill, M. S., 524, Shine, H. J., 174, 658, Shiner, V., 285, 286, 476, Shryne, T. M., 295, Sickman, D. V., 724, Sidgwick, N. V., 388, Siegel, M., 220, Siegel, S., 588, Simmons, H. E., Jr., 461,599, , Simons, J. H., 166, Simonsen, J. L., 298, Sims, L. L., 733, Sindler, B., 569, Sivertz, G., 741, Skell, P. S., 399, 479, 500,, 534, 746, 751, Slater, J. C., 18, Sloan, G. J., 677, Small, G., 216, Smid, J., 718, Smiley, R. A., 296, Smoot, C. R., 448, Smith, D. H., 658, Smith, G. F., 443, Smith, H. A., 325, Smith, J. F., 595, Smith, L., 189, 567, Smith, P. A. S., 623, Smith, R. P., 65, Smith, W. E., 390, 638, Smith, W. V., 60, Smyth, C. P., 57, Smyth, I. F. B., 355, Snow, A. I., 585, Sobotka, H., 140, Soffer, L. M., 714, Sogo, P. B., 681, Sommelet, M., 132, 641, Soper, F. G., 442, 443, 650,, 651, Speck, J. F., 351, Speers, L., 630, Spencer, C. F., 599, Spielman, M. A., 136, 337,, 627, Spitzer, W. C., 238, Spivey, E., 163, 375, Sprecher, M., 208, 685, Springall, H. D., 414, Sprung, M., 648, Stadnikov, G. L., 116, Stannett, V., 688, Staskun, B., 619, Staudinger, H., 533, Steacie, E. W. R., 672, 692,, 718, Stein, G., 537, Stein, L., 182, Steinberg, M., 717, Steinberger, H. R., 608, Steinberger, R., 351, Steinfink, H., 51, Steinmetz, H., 692, 735, Stempel, G. H., Jr., 543, Stephen, H., 134, 619, Stern, E., 163, 192, 395, Stevens, I. D. R., 734, 739, Stevens, T. E., 740, Stevens, T. S., 640, 641, 643, Stevenson, D. P., 43, 504, Stewart, R., 193, Stirling, C. J. N., 688, Stockmeyer, W. H., 714, Stokes, R. H., 102, Storey, R., 298, , Stork, G., 291, Strachen, A. N., 689, Strauss, F., 533, Streitwieser, A., Jr., 250, 253, 269, 279, 283, 285,, 504, 561, 576, 577, 587, Struve, W. S., 627, Style, D. W. G., 683, Sudborough, J. J., 324, Sugden, S., 65, Sulzberger, R., 107, Sundralingham, A., 706, Surgenor, D. M., 710, Sutton, L. E., 58, 68, 69, 75, Suzuki, S., 285, Swain, C. G., Ill, 112, 139,, 166, 252, 256, 272, 273,, 279, 299, 300, 375, 401,, 403, 440, 540, 542, 572,, 695, 714, Swallow, A. J., 689, Swern, D., 535, 720, Swift, E., 371, Synerholm, M. E., 355, Szabo, A. L., 145, Szwarc, M., 36, 713, 717,, 718, 734, 745, Tada, R., 619, Taft, R. W., Jr., 200, 223,, 224, 227, 229, 235, 329,, 446, 515, 592, Taher, N. A., 170, Takeuchi, C., 714, Takezaki, Y., 714, Talat-Erben, W., 726, Tanabe, K., 398, Tanner, D., 587, Tappe, W., 497, Tarbell, D. D., 137, 522,, 611, 645, Taube, H., 729, Taylor, H. A., 751, Taylor, T. I., 628, Taylor, W., 564, Tazuma, J., 416, Tchelitchev, S., 400, Teitelbaum, C., 759, Tengler, E., 118, Ten Have, P., 709, Thewlis, J., 42, Thiessen, P. A., 116, Thomas, C. H., 284, Thomas, D. S., 705, Thomas, VV. M., 717, Thompson, A. L., 570, Thompson, T., 641, Thomson, T., 366, Thorpe, J. F., 531, 568, Tickner, A. W., 713, Tietze, E., 134, 753, Tilles, H., 354, Timm, E. W„ 322, Tokura, N., 619, Tolman, R. C., 319, Tommila, E., 318
Page 795 :
Author Index, Trambarulo, R. F., 60, TrefTers, H. P., 99, 153, 325, Treibs, W., 64, Trifan, D. S., 597, Troeger, J., 134, Trombe, F., 678, Trotman-Dickenson, A. F.,, 717, 718, 752, Truce, W. E., 529, Truter, M. R., 419, Tsuruta, T., 443, Turkevich, J., 681, Turner, M. K., 324, Turnquest, B. W., 331, Tutton, R. C., 746, Ugai, M., 397, Ulich, H., 448, 449, Ultree, A. J., 531, Urban, R. S., 636, Urbanek, L., 596, Urech, H. J., 601, Urey, H. C., 374, 541, 635, Urmanczy, A., 452, Urry, W. H., 687, 692, 744,, 747, 755, 759, Vander Werf, C. A., 274, Van Eenam, D. N., 641, Van Emster, K., 594, 730, Van Helden, R., 697, Vaslow, F., 332, Vaughan, W. E., 162, 461,, 495, 695, 697, 713, 714,, 733, Vejvoda, E., 138, Vennesland, B., 351, Venus-Danilova, E., 637, Verhoek, F. H., 348, Vernon, C. A., 288, 440, Vick, K., 414, Viebel, S., 426, Voegtli, W., 704, Vogl, O., 730, Vogt, A., 680, Von Braun, J., 478, 563, Von Euler, H., 547, Von Kiss, A., 452, Von Kummer, U., 619, Von Pechmann, H., 298, Von Richter, V., 459, Von Schilling, V., 216, Vorlander, D., 394, Voss, J. C. S., 451, Waage, P., 319, Walborsky, H. M., 761, Walden, P., 203, Walling, C., 36, 506,, 672, 690, 691, 705,, 714, 716, 722, 730,, 733, 741, 744, 747, Wallis, E. S., 623, 624,, 648, Walsh, A. D., 206, Walter, J., 38, , 649,, 708,, 731,, 628,, , Walters, W. D., 391, Walz, F., 695, Wang, C. H., 688, 724, 727, Wannowius, H., 539, Ward, A. M., 184, Warder, R. B., 316, Ware, J. C., 627, Warford, E. W. T., 428, Warhurst, E., 537, Warkentin, J., 403, Warnhoff, E. W., 81, Wassermann, A., 163, 537, Waters, W. A., 442, 538,, 539, 692, 700, 709, 711,, 714, 724, 735, Watson, H. E., 59, Watt, I. C., 708, Webb, R. L., 218, 290, Weber, P., 696, Weedon, B. C. L., 689, Weil, I., 188, Weingarten, H. I., 601, Weinheimer, P. H., 641, Weinstock, J., 494, 499, 500,, 757, Weis, C., 685, Weiss, J., 729, Weissberger, A., 81, 86, 159,, 705, Weissman, S. I., 685, Weitz, E., 677, Weizmann, M., 744, Welch, K. N., 391, Weller, S. W., 71, Wendler, N. L., 545, Wepster, B. M., 215, Wertyporoch, E., 448, Wertz, J. E., 684, West, J. P., 449, Westheimer, F. H., 164, 200,, 202, 211, 351, 420, 426,, 543, 549, 635, Wheeler, O. H., 483, 549, Whelan, W. P., Jr., 280, Wheland, G. W., 21, 23, 38,, 42, 61, 70, 86, 97, 236,, 238, 366, 370, 376, 412,, 414, 619, 658, 673, 674,, 748, 754, White, A. M., 267, 344, White, T. R., 414, White, W. N., 291, 646, Whiteway, S. G., 689, Whitmore, F. C., 295, 517,, 584, 733, wiDerg, K. B., 149, 192, 1<, 295, 396, 529, 538, 6i, 635, 685, Widmark, E. P. M., 346, Wieland, H., 680, 681, 7C, Wigner, E., 682, Wilder, P., 355, Wilds, A. L., 688, Wilhelm, M., 546, Wilkins, R. G., 118, Wilkinson, G., 46, , 779, , Williams, A. L., 713, Williams, E. D., 545, Williams, G., 274, 520, Williams, G. H., 688, 720,, 725, Williams, H. L., 710, Williams, J. L., 438, 502, Williams, R. A., 482, Williams, R. J., 501, Williams, R. V., 652, Willis, J. B., 81, Wilson, C. L., 100, 373, 374,, 385, 386, 595, 689, Wilson, I. S., 433, Wilson, J. W., 701, Wilson, W. J., 442, Winning, W. I. H., 695, Winstein, S., 169, 241, 258,, 269, 271, 288, 290, 293,, 302, 494, 561, 562, 564,, 565, 566, 567, 574, 575,, 577, 579, 581, 582, 588,, 597, 598, 611, 639, 755,, 757, 761, Winston, A., 355, Wistar, R., 419, Wittig, G., 462, 640, 641, Wittwer, C., 377, Wohl, A., 700, Wolfe, J. R., Jr., 750, Wolff, H., 254, 623, Wolfstirn, K. B., 741, Wollschitt, H., 303, Wong, R., 448, Wood, G. W., 600, Wood, J. L., 339, 342, Woods, G. F., 366, Woods, W. G., 598, Woodward, R. B., 216, 545, Woodworth, R. C., 751, Woolf, L. I., 481, 482, Wooster, C. B., 681, Wulf, O. R., 30, Wurster, G. F., 711, Yancey, J. A., 591, Yeddenapalli, L. M., 443, Yoffe, A. D., 345, Young, D. P., 335, Young, R. W., 546, Young, W. G., 164, 286, 288,, 290, 291, 296, 400, 494,, 519, 738, Young, W. Y., 680, Zahler, R. E., 452, Zechmeister, L., 87, Zeigler, K., 303, 673, 677,, 707, Zeiss, H. H., 269, Ziegenbein, W., 64, Zimmermann, P., 619, Zirkle, C. L., 571, Zook, H. D., 401, 438, Zucker, L., 190, 374, Zuffanti, S., 93
Page 796 :
Subject Index, Acetoacetic acid, decarboxylation of, 346,, 350, Acetoacetic ester, acidity of, 365, alkylations of, 298, eno'lization of, 377, Acetone, halogenations of, 162,175, 372ff, 382, Acetyl peroxide, decomposition of, 717f, Acetylacetone, acidity of, 365, cleavage of, 338, Acetylene, molecular orbitals in, 19, Acid anhydrides, conversion to amides, 361, (Ex. 6), Acid-base catalysis, 11 Off, 195 (Ex. 4), mechanisms for, 188f, of mutarotation of sugars, 139, 542, of semicarbazone formation, 543, 555, (Ex. 4), specific vs. general, 110, 189, of tautomerization, 373, 385, Acidity, 122 (Ex. 1, 3), of alicyclic dicarboxylic acids, 242, Br0nsted system of, 93, of C=C—C=0 systems, 117, of C—H bonds, 365ff, of diketones, (3-keto esters, and aliphatic, nitro compounds, 365f, effect of structure on, 200ff, 244 (Ex. 4,, 11), 367, 404 (Ex. 1), leveling effect in, 96, Lewis system of, 115ff, 125 (Ex. 11), of phenols, 215, quantitative evaluation of, 100, 103ff, 122, (Ex. 2), and resonance, 215, of sulfones, 369, Acidity scales, Hammett (h0 scale), 103ff. See also H0, function, and reaction rates, 190, in water-ethanol mixtures (Grunwald, scale), 107ff, 124 (Ex. 7), 192, 198 (Ex., 12), , •, , •, , 107, , in water-hydrazine mixtures, 1U/, Acids, Br0nsted, 94, conjugate, 94, Lewis, 115, very weak, 97, 369ff, Acrylonitrile, additions to, 528, Activated complex, definition of,, i'°, Activation, Arrhenius energy of, 179, entropy of, 179, 181 ff, 196 (Ex. 8), free energy of, 130, 179, volume of, 716, ., im, Activities and activity coefficients, 101, Activity coefficients, degenerate, 101, evaluation of, 108ff, 124 (Ex. 7), Acyl azides, formation of, 623, rearrangement of, 624, , 780, , Acyl chlorides, alcoholysis of, 332, hydrolysis of, 334, Acyl iodides, exchange with iodine, 705, Acylations, aromatic (Friedel-Crafts reac¬, tions), 450, Acylium ions in esterification and hydrolysis., 325, Addition compounds, acid-base, 118, 231, stabilities of, 231, Addition reactions, 1,2- vs. 1,4-, 530, homolytic vs. heterolytic, 732, nucleophilic vs. electrophilic, 121, 125 (Ex., 11), 527, Additions to C=C double bonds, of aryl halides, 729, of CBr2, 535, m-additions, 533ff, 743, to conjugated dienes, 530, of halogens, 137, 150, 520ff, 553 (Ex. 2),, 736ff, halide catalysis of, 524, intermediates in, 521, kinetics of, 524, stereospecificity, 523, homolytic, 730ff, of hydrogen halides, 140, 516ff, 527, 732fl, ^kinetics, 517, orientation in, 518, 733, rearrangement during, 517, of hypochlorous acid, HOC1, 525, of methylene, CH2, 751, of N20<, 739, nucleophilic, 527f, of polyhalomethanes, 743f, of thiols, 741, of water, 514, 553 (Ex. 1), Additions to C=C triple bonds, 520, Additions to C=0 double bonds, 539ff, of alcohols, 541, of hydroxylamine, phenylhydrazine, and, semicarbazide, 543, 555 (Ex. 4), stereochemistry of, 549, of water, 540, Adducts, see Addition compounds, Alcohols, configuration of, 552, dehydes, addition to C=C double bonds, 74/, autoxidation of, 708, hydration of, 540, rearrangement in acid, 637, dol condensation, 389ff, 406 (Ex. 3, 4), kyl aryl ethers, rearrangement of, 648, ■Alkylanilinium salts, rearrangement ol,, , 652, kylation, aromatic (Friedel-Crafts reac¬, tion), 447, . , ,, lophanates, formation from alcohols, 130, lylic rearrangements, 286, bimolecular, 290, in ester hydrolysis, 343
Page 800 :
784, , -, , Subject Index, , E\ and E2 mechanisms for elimination, reactions, 473, 478, Eclipsing effects, in elimination reactions,, 489, Eigen function (^-function), 23, Electromeric effect, 217, Electron, energy levels of, 2, s, p, d, and /, 10, wave-mechanical picture of, 4ff, Electron diffraction, 42, Electron-spin resonance spectroscopy, 684, Electronegativities, 39ff, from bond energies, 40, 53 (Ex. 6, 7), Pauling, table of, 41, Electrophiles, 115, Electrophilic substitution reactions, see, Substitution, electrophilic, Elimination reactions, 472ff, alpha, 398, 535, bimolecular mechanism (E2) for, 478, branching, effect of, 486, in cyclic systems, 497ff, intramolecular, 500, orientation in, 480, 507 (Ex. 1), solvent polarity, effect of, 488, stereochemistry of, 489, unimolecular mechanism (.El) for, 473, in vapor phase, 504f, Endoxocyclohexyl halides, hydrolysis, 598, Energy of activation, 130, 179, Enolization, 365ff, acid- and base-catalyzed, 573, equilibria in, 376, 405 (Ex. 2), solvent, effect of, 380, structure and rate in, 380, Epichlorohydrin, ring closure of, 169, Epoxides, formation from olefins, 534, as intermediates in hydrolysis of chlorohydrins, 567, ring opening in, 291 f, 311 (Ex. 9), Equatorial bonds, 240, Ester interchange, 321, Esterification, 314ff, acid-catalyzed, 318f, via acylium ions, 325, with alkyl-oxygen cleavage, 340, via carbonium ions, 340, mechanisms for, 315f, Esters, elimination reactions of, 502, hydrolysis of, 314ff, 356 (Ex. 2), acid catalysis of, 318f, via acylium ions, 325, with alkyl-oxygen cleavage, 339, 342, allylic rearrangement during, 343, via carbonium ions, 339, mechanisms for, 315f, of inorganic acids, hydrolysis of, 34b, Ethylene, molecular orbitals in, 1), Ethylene oxide, formation from epichlorohydrin, 167, ring opening, 163, .71, Ethyleneimonium-ion intermediates, 5/1, Exclusion principle, Pauli, 8, , Favorskii rearrangement, 157 (Ex. 13), 642, Fenton’s reagent, 690, Ferrocene, bond order in, 46, reactions of, 416, Fischer indole synthesis, 156 (Ex. 9), Fluorinations, 445, 698, Fluoroolefins, additions to, 529, Formal charge, 13, 31 (Ex. 6), effect on acidity, 205, Four-center reactions, 646, Free radicals, configuration of, 685ff, coupling of, 691, definition of, 672, detection by electron spin resonance, spectroscopy, 684, detection by magnetic measurements,, 153, 673, detection by ortho-para hydrogen conver¬, sion, 682, detection by Paneth technique, 683, displacements by, 69Iff, disproportionation of, 692, electrolytic formation of, 689, formation from metal salts and peroxides,, 690, initiators, 687f, measurement of concentrations of, 684, from photochemical dissociations, 689, reactions of, 672ff, rearrangements of, 692, 755ff, Free-radical reactions, additions, 730ff, stereochemistry of, 734, 740, 742, arylations, 720ff, autoxidations, 705ff, decomposition of azo compounds, 724ff, decomposition of peroxides, 71 Off, halogenations, 696ff, kinetic treatment of, 694, 761 (Ex. 1),, 762 (Ex. 3), 766 (Ex. 6), rearrangements, 755ff, Sandmeyer reaction, 729, Freezing-point lowering studies in concen¬, trated sulfuric acid, 98ff, 123 (Ex. 4),, 153, 305, Frequency, carbonyl, 84, characteristic, in infrared spectroscopy,, hydrogen bonding, effect on, 83, 90, (Ex. 10), isotopic shifts, 81, 91 (Ex. 12), Friedel-Crafts reactions, 136, 447ff, kinetics of, 448, rearrangements during, 4491, Fries rearrangement, 152 (Ex. 2), 653f, intermolecularity of, 653, ion pairs in, 654, ortho vs. para migration, 654, “Front, rroni strain,”, sir<uu, 232, i'll, Furfuryl chloride, reaction with CM , 1-^, Glucose, mutarotation of, 542, Gomberg-Bachmann reaction,, , Ut
Page 801 :
Subject Index, , -, , 785, , Hydride transfers, in solvolyses, 594, 597, 600, Graphite, bond order in, 46, structure of, 46, ..., uncertainty in heat of sublimation of,, 36, 53 (Ex. 4, 5), Grignard reagents, addition reactions of, 138, 400, electrolysis of, 689, reduction by, 403, 545, /3-Haloacids, decarboxylative debromination of, 495, Haloforms, see Trihalomethanes, Halogen atom carriers, 700, Halogen exchange reactions, .235, 266, 273, Halogenation, aromatic, 440ff, free radical, 695ff, 761 (Ex. 1), Halogenonium ions, 440, 443, Hammett equation, 220ff, 246 (Ex. 6, 8, 9), applied to decompositions of benzoyl, peroxides, 717, applied to substitutions on benzene ring,, 226, 434ff, 466 (Ex. 5), limitations of, 223, Heisenburg uncertainty principle, 3, Hemiacetals, formation of, 541, hydrolysis of, 542, Hess’s law, 35, 53 (Ex. 2), Heterolytic reactions, 121, possible occurrence in vapor phase, 505, Hexamethylacetone, rearrangement of, 638, Hexaphenylethanes, autoxidation of, 707, dissociation of, 233, 673f, H0 function of Hammett, 104ff, 190, in Beckmann rearrangement, 618, in hydration of olefins, 515, in hydrolysis of /3-lactones, 327, Hoesch reaction, 134, Hofmann elimination, 156 (Ex. 10), 480, Hofmann-Martius rearrangement, 653, Hofmann rearrangement, 62If, intermediates in, 128, 135, 151, stereochemistry of, 622, Hofmann rule (for elimination reactions), 480, Homoallylic rearrangement, 761, Homologation reactions (using CH2 di¬, radicals), 751, Homolytic reactions, 672ff. See also Freeradical reactions, Hunsdiecker reaction, 353, Hybridization, of atomic orbitals, 16ff, and resonance, 22, Hydrazoic acid, reactions of, 624, Hydride transfers, in Cannizzaro reaction, 546, in deamination reactions, 592, in the Oppenauer oxidation, 545, in reactions of coenzymes, 549, inideedU5C45°nS Whh aluminum isopropoxm reductions with Grignard reagents, 545, , transannular, 549, 600, Hydrocarbons, dipole moments of, 63ff, polar, 64, Hydrogen bonding, 11, 28ff, and acidity, 30, 209ff, 244 (Ex. 3), in alcohols, 29, conditions for, 30, intramolecular vs. intermolecular, 30, 32, (Ex. 12), and spectra, 30, 83, and volatility, 29, 32 (Ex. 11, 12), Hydrogen exchange, and acidity, 371, in the aldol condensation, 391, in aromatic systems, 419, in elimination reactions, 479, in ketones, 373, Hydronium-ion catalysis, 11 If, H ydroperoxides, decomposition of, 71 Off, formation from hydrocarbons, 706, Hydroxyl radicals, formation from Fenton’s, reagent, 690, 2-Hydroxypyridine as an acid-base catalyst,, 139, Hyperconjugation, 49ff, in addition reactions, 518 /, in aromatic substitution, 431 ^, C—C vs. C—H, 50, and dipole moments, 64, and electronic spectra, 87, in elimination reactions, 476, in enolization, 384 *, and interatomic distances, 49 ', steric hindrance to, 238 f, Hypochlorous acid (HOC1), addition to, G=C double bonds, 525, /-Effects, see Inductive effects, i-Factor (j/-factor) in cryoscopy, 98ff, 123, (Ex. 4), Imidazole, as catalyst in ester hydrolysis,, 332, Induction period, 693, Inductive effects, 200ff, 243 (Ex. 1, 2), of alkyl groups, 205, and aromatic substitution, 432, 437, of aryl groups, 207, in elimination reactions, 482, quantitative evaluation of (Taft treat¬, ment), 227, table of, 207, Inductomeric effect, 208, Inhibitors, free-radical, 693, Initiators, free-radical, 687, 694, Intermediates, physical detection of, 152, stable and metastable, 130, testing possible, 134, vs. transition states, 128, trapping of, 137f, Internal return, 288, 583, Inversion vs. retention of configuration in, reactions, 148, \
Page 802 :
786, , -, , Subject Index, , Iodine, exchange reactions of, 703, Iodine monohalides (IC1 and IBr), addi¬, tion to C=C double bonds, 524, Iodobenzene, exchange with iodine, 704, Ion pairs, 102, intimate vs. solvent-separated, 5801T, Ionic bonding, 1 Off, Ionic strength, and reaction rate, 185, 196, (Ex. 7), Iron(II) salts, as catalysts in autoxidations, 709, in production of free radicals, 690, Isobornyl chloride, rearrangement of, 596, Isobutylene oxide, from decomposition of, di-i-butyl peroxide, 714, Isopropylidenemalonic ester, hydrolysis of,, 326, Isotope effects, 146, 192ff, in elimination reactions, 479, in iodination, 445, in nitration, 423, in nucleophilic substitution, 285, in sulfonation, 447, Isotopic labeling, 143ff, “/ strain,” 240, Kenyon-Phillips cycles, 265, Ketenes, as intermediates in the Wolff re¬, arrangement, 627, Keto-enol equilibria, 376ff, 405 (Ex. 2), analysis of mixtures, 337, Ketones, enolization of, 376, halogenations of, 372ff, 382, racemization of, 373, radicals from reduction of, 690, rearrangement of, in acid, 637, Ketyls, 681, Kinetic vs. thermodynamic control, 172, in additions to conjugated olefins, 394, in Michael reaction, 394, Kinetics, 159ff, ambiguities in interpreting data, 187,, 197 (Ex. 10), consecutive reactions, 168, isotope effects in, 192, mechanistic implications from, 174, 197, (Ex. 11), parallel reactions, 171, reversible reactions, 167, Kirkwood-Westheimer treatment of acid, strengths, 202, Kolbe electrolysis, 689, Kolbe reaction, 387, Leveling effect, 96, Lewis structures, llff, 31 (Ex. 4, 7, 9), Linear free-energy relationships, BrjzSnsted catalysis law, 113, 227, Hammett equation, 220ff, Swain treatment (for nucleophilic sub¬, stitution), 299f, Taft treatment (for aliphatic systems),, 227, Winstein-Grunwald treatment (for, solvolyses), 302, , Lossen rearrangement, 623, Lyonium and lyate ions, 95, Magnesium, univalent, 690, Mannich reaction, 410 (Ex. 7), Markownikoff’s rule, 516, Mass-law effect, 256, 572, Meerwein chloroarylation reaction, 729, Meerwein-Pondorff reduction, 545, Menthyl and neomenthyl derivatives, elimination reactions of, 497, solvolysis of, 593, Mercury salts, as catalysts in nucleophilic, substitution, 274, Mesitoic acid, esterification of, 325, ionization in concentrated H2S04, 153, in the Schmidt rearrangement, 625, Mesomeric effect, 217, />-Methoxy-//-nitrobenzoyl peroxide, de¬, composition of, 632f, Methyl radicals, from acetyl peroxide, 717, attack of solvents by, 717, from azomethane, 724, Methylene, CH2, addition to olefins, 751, reaction with saturated hydrocarbons,, 752, Michael reaction, 392f, Microscopic reversibility, principle of, 319,, 515, Migratory aptitudes, in homolytic rearrangements, 758, in the pinacol rearrangement, 607f, Modes of vibration, normal, 79, Molecularity vs. reaction order, 164, Moment, bond, 61, 88 (Ex. 3), table of, 62, dipole, see Dipole moments, group, 62, 88 (Ex. 4, 5), addition of, 63, 68, table of, 62, Mustard gas, reactions of, 572, phthalyne intermediate, 461, aer rearrangement, 644, ghboring-group reactions, 271, 561 ff,, 613 (Ex. 2), n additions, 611, n bicyclic systems, 594ff, n cyclic systems, 588, n eliminations, 611, vidence for, 561, leighboring acyloxy, 271, 566, 600, leighboring alkoxide, 569, leighboring alkyl, 584, leighboring /3-aryl, 575, leighboring 5-aryl, 579, leighboring bromine, 574, 600, leighboring carboxyl, 270, 562ff, leighboring chlorine, 525, leighboring hydrogen, 591, leighboring hydroxide, 569, leighboring iodine, 575
Page 803 :
Subject Index, Neighboring-group reactions, neighboring, nitrogen, 562, 570, neighboring oxygen, 567ff, neighboring sulfur, 562, 572, stereochemical requirements for, 573, Neopentyl compounds, rearrangements of, 585, substitution reactions of, 132, Neutron diffraction, 42, Nitration, aromatic, kinetics, 169, 420ff, 464 (Ex. 2, 3), mixtures from, 172, 195 (Ex. 3), via nitrosation, 426, nitrous acid in, 425ff, 465 (Ex. 3), orientation in, 429, 464 (Ex. 1, 5), structure and reactivity in, 428, Nitrites, in aromatic nitration, 425f, substitution reactions with, 297, Nitro compounds, aliphatic, acidity of, 365, conversion to acz-nitro forms, 366, zV-Nitroanilines, rearrangements of, 655, Nitrogen mustards, hydrolysis of, 570, Nitrohalobenzenes, reactions of, 238, 452ff, Nitronium ion, 169, 419ff, formation of, 422, Raman spectrum of, 420, salts of, 419, Nitrosation, aromatic, 426, .Af-Nitrosoacetanilides, decomposition of,, 727, A-Nitrosoanilines, rearrangement of, 652, Nitrosonium ion, NO+, 425, Norbornadiene, addition reactions of, 761, Norbornene, addition reactions of, 611, 742, Norbornenyl derivatives, solvolysis of, 598,, 761, Norbornyl derivatives, solvolysis of, 586ff, Nucleophiles, 116, ambident, 296, Nucleophilic substitution reactions, see Sub¬, stitutions, nucleophilic, Nucleophilicity of reagents, 259, 306 (Ex. 1), quantitative evaluation of, 301, Octet rule, 12, Olefins, additions to, see Additions to C=, double bonds, Oppenauer oxidation, 545, Orbitals, atomic, 10, IT,, , 19, , Order of a reaction, definition, 161, , £ dependence on solvent, 184, determination of, 165, Organocobalt compounds, free radicals, from, 688, Organolithium compounds, addition reactions of, 400, displacement reactions of, 402, Organometallic compounds, condensation.; and additions with, 400,, decomposition in the vapor phase, 683, , 787, , Organosodium compounds, condensations, with, 402, ortho-para hydrogen conversion, 682, orlho:para ratio in aromatic substitution,, 436ff, Orton rearrangement, 650, Oximes, formation of, 543, rearrangement of, 618, Oxygen, as a diradical, 693, 748, as an inhibitor, 693, transannular addition reactions of, 749, Paneth (mirror removal) technique, 682,, 751, Paramagnetic resonance spectroscopy, 684, Paramagnetism of free radicals, 673, Pauli exclusion principle, 8, Pentaphenylpyrrole, free radical from, 678, Periodic table and electron quantum num¬, bers, 9, Perkin reaction, 391, Permanganate, oxidation of olefins with,, 538, Peroxide effect, 140, 519, 733ff, Peroxides, diacyl, decomposition of, 714f, dialkyl, decomposition of, 712f, as radical initiators, 687, reactions with phenols, 649, rearrangements of, 629f, 757, transannular, 749, Peroxy acids, additions to olefins, 534, formation from aldehydes, 708, reaction with ketones, 630, Peroxy esters, decomposition of, 720, Peroxylauric acid, decomposition of, 720, Phenols, acidity of, 214, Phenonium-ion intermediates, 575ff, Phenoxides, carbanion character in, 386, halogenation of, 443, reaction with chloroform (ReimerTiemann reaction), 387, 3-Phenyl-2-butyl tosylate, solvolysis of,, o dl ' , , ’ '-“““ignncnt oi, 00.J, 6-Phenylethyl bromide, dehydrobromination of, 478, , Fhenvtelyoxylic «,ers. addition reactions, Phenylhydrazones, formation of, 543, , S“o\"8yfWlaneS', , °f' 436, , Photosensitization, 749, Pi bonds, 15, 19, 31 (Ex. 8), Pi complexes, 118, , ^hydrogen-bridged transition states,, as intermediates in aromatic, rearrangements, 653, in olefin hydrations, 515, vs. sigma complexes, 119
Page 804 :
788, , -, , Subject Index, , Pi-electron systems, 20, 31 (Ex. 8), in aromatic compounds, 414, excitation of, 85, in pyridine and pyrrole, 415, thickness of, 53, Picramide, basicity of, 237, Pinacol rearrangement, 601 ff, conformational effects in, 605, migratory aptitudes of substituents in, 607, stereochemical requirements for, 603, Polarization, 25, 58, and covalent bonding, 26, , electron and atom, 59, orientation, 59, Ponderal effects, 235, 275, Porphyrindene, 754, Propagation, in free-radical reactions, 694, /3-Propiolactone, hydrolysis of, 327, Propionyl peroxide, decomposition of, 719, Prototropy, see Tautomerism, “Push-pull” reactions, 299, Pyridine, aromaticity and basicity of, 414, a-Pyridone, reactions of, 298, Pyrrole, basicity of, 414, carbanion character in conjugate base,, 386, 7r-electron system in, 414, Quantum numbers for electrons, 5ff, 31, (Ex. 1)., , relationship to periodic table, 9, Quinaldic acid, decarboxylation of, 349, Quinuclidine, basicity of, 215, nucleophilicity of, 278, R effects, see Resonance effects, Radicals, free, see Free radicals, Radii, constant energy, 44, 54 (Ex. 8), covalent, 42ff, 54 (Ex. 11), table of, 44, van der Waals, 50, Reaction constants, Hammett treatment, 222, Taft treatment, 230, Reaction coordinate, 129, Reaction mechanism, definition of, 127, Rearrangements of 1,2-diketones, see, Benzilic acid rearrangement, Reimer-Tiemann reaction, 387, 399, Reissert reaction, intermediates in, 135f, Resonance, 21 ff, and acid strengths, 215, 370, and aromatic substitution, 429, 439, conditions for, 27, and dipole moments, 65ff, effects, associated with substituents, 218,, 245 (Ex. 5), and interatomic distance, 47, stabilization of free radicals by, 675, steric inhibition of, 22, 70, 90 (Ex. 11),, 152, 236ff, 378, 455, , Resonance energies, 23, 54 (Ex. 9), of benzene derivatives, 37, as criteria for aromaticity, 413, “extra ionic,” 40, Retarders, in free-radical reactions, 693, Rotating-sector” method for studying, photochemical reactions, 695, Rotation, intramolecular, 72, in carbonium ions, 605, in free radicals, 737, Rule of six, Newman, 323, Salicyl phosphate, hydrolysis of, 140, Salt effects, in acetolysis, 583, in kinetics, 185ff, in nucleophilic substitution, 256ff, primary vs. secondary, 184, ^ “special,” 583, Sandmeyer reaction, 729, Saponification of esters, 315ff, 356 (Ex. 1), catalysis by imidazole, 332, mechanisms for, 315, polar effects in, 318, steric hindrance to, 317, tracer studies, 145, 316, Saytzeff rule (for elimination reactions),, 480, Schmidt rearrangement, 624ff, Schomaker-Stevenson equation, 43, 54, (Ex. 7), Schroedinger wave equation, 4, Screening effect, 39, Semicarbazones, formation of, 164, 195, ^ (Ex. 2), 543, Semidines, from rearrangement of hydrazo, compounds, 657, 659, Semipinacolic deamination, 602, 605, inversion at migration terminus during,, 604, 613 (Ex. 3), Semipolar double bond, 13, Semiquinones, 679, Sextet, 7r-electron, 414ff, Shells and subshells, atomic, 9, 1,2-Shifts, 584ff, Sigma bonds, 15, ,, Sigma complexes, 119, Six, Newman rule of, 323, Silver salts, as catalysts in nucleophilic substitutions,, 274, 297, 307 (Ex. 5), as catalysts in the Wolff rearrangement,, 628, decarboxylation of, 353, •Sjvl and Sn2 mechanisms, 25Iff, Sn2' mechanism, 290, iSV* mechanism, 294, S.vi' reactions, 296, Solvent cage, in free-radical reactions, 718,, 728, Solvent polarity, effect on elimination reactions, 488, and reaction order, 184, and reaction rate, 182f, 254, 196 (Ex. 6, 7), Solvolytic displacement, 252, Swain treatment of, 299
Page 805 :
Subject Index, Sommelet-Hauser rearrangement, 641 f, Sommelet reaction, 132, _, Spectra, molecular, 76ff, 90 (Ex. 9-14), electronic, 85, electron-spin resonance, 684, and hydrogen bonding, 30, 83, 90, (Ex. 10), microwave and radiofrequency, 76, Raman, 76, rotational, 78, 92 (Ex. 14), use in determining acid strengths, 369, vibrational, 77, 79ff, Spin, electronic, 7, Steady-state approximation, 170, 694, Steric assistance, 234, to dissociation of hexaphenylethanes, 676, to unimolecular eliminations, 475, 477, to unimolecular nucleophilic substitution,, 279, 586, Steric hindrance, 180, 230fT, in aromatic substitution, 457, in bimolecular nucleophilic substitution,, 275, in ester saponification, 317, in esterification, 322, in keto-enol equilibria, 378, Steric strains, 23Iff, Stevens rearrangement, 640, Substituent constants, Hammett’s, 221, 226, for substitution on aromatic rings, 435, Taft’s, 230, Substitution reactions, classification of, 120ff, 125 (Ex. 11), homolytic vs. heterolytic, 121 f, polar vs. free-radical, 121, Substitutions, electrophilic, in aromatic systems, 418ff, orientation in, 429ff, structure and reactivity in, 428, nucleophilic, 250ff, with allylic rearrangement, 286ff, in aromatic systems, 452ff, attacking reagents, 258, borderline cases, 312 (Ex. 12), catalyzed by Ag+ and Hg2+, 274, competition with elimination. 263, ., 473f, 485ff, 507 (Ex. 2), effects of solvent, 254, internal (Sni), 294, internal, with rearrangement (Sui'),, isotope effects in, 285, kinetics of, 254, leaving groups, 261, in nonpolar solvents, 272ff, quantitative correlation of rates, (Swain treatment), 299, steric effects, 274ff, stereochemistry of, 147f, 263ff, 309, (Ex. 7), structure-reactivity relationships in,, 281 ff, Succinimido radical, 701, Sulfonation, aromatic, 445f, 466 (Ex. 4), , 789, , Sulfones, ., .no, (8-chloro, elimination reactions of, 4)), conversion to selenides, 155 (Ex. 8), Sulfonium ions, acidity of, 372, elimination reactions of, 474, , O, Sulfonyl (—S—) group, electron attraction, , A, , by, 218, 369, Sulfuric acid, freezing-point lowering stud¬, ies in, 97ff, 123 (Ex. 4), Sulfuryl chloride, chlorination with, 702, Swain equation (for correlation of nucleo¬, philic substitution rates), 299, Sydnones, 417, T effect (tautomeric effect), 217, Taft treatment of inductive effects, 227ff, Tautomerism, 366ff, acid- and base-catalyzed, 373, mechanisms for, 384, Telomerization, 745, Termination step in free-radical reactions,, 694, Tetrahedral carbon atom, 17, 31 (Ex. 7), Tetramethyl lead, methyl radicals from,, 683, Tetramethyl-/>-phenylenediamine as an, antioxidant, 709, Tetramethylglucose, mutarotation of, 139,, 542, Tetraphenylhydrazine, radical from, 678, Tetrazoles, formation in Schmidt reaction,, 627, Thiazole-2-carboxylic acid, decarboxylation, of, 349, Thiols, addition to C=C double bonds, 529,, 741, Thionyl chloride, reaction with alcohols,, 295, Tiglic acid, addition of hydrogen iodide to,, 519, Titanium(III) salts, in the production of, free radicals, 691, o-Tolylmagnesium bromide, reaction with, benzil, 636, Transition state, definition of, 129, 178, Transition-state theory, 178ff, Trialkylamine oxides, conversion to ole¬, fins, 501, 504, Triarylmethyl ions, 273, 305, Triarylmethyl peroxides and hydroperox¬, ides, rearrangements of, 757, Triarylmethyl radicals, 672ff, Trifluoroiodomethane, addition reactions, of, 744, Trifluoromethyl group, electron attraction, by, 218, Trihalomethanes, acidity of, 371, hydrolysis of, 381, 397ff, Trimethylaluminum dimer, 585, 1 rimtroanisole, reactions of, 454
