RONALD BRESLOW

second edition

ORGANIC
REACTION
MECHANISMS
An Introduction

X
;?-;-.
y^y

Ou ^V-^iV^ 4V -^^
r r 1
Wr^-i-^K^
yffl

t %;
819
k-xr
mm
T ') V
/>€^n*^p ^^^^t^^i
ORGANIC REACTION MECHANISMS

Second Edition
THE ORGANIC CHEMISTRY
MONOGRAPH SERIES
Ronald Breslow, editor
Columbia University

editorial advisory board: P. D. Bartlett, Harvard University

David Curtin, University of Illinois
W. S. Johnson, Stanford University
J. D. Roberts, California Institute of
Technology
R. B. Woodward, Harvard University

ORGANIC REACTION MECHANISMS, SECOND EDITION

Ronald Breslow
Columbia University

THE MOLECULES OF NATURE

James Hendrickson
Brandeis University

MODERN SYNTHETIC REACTIONS

Herbert House
Massachusetts Institute of Technology

PEPTIDESAND AMINO ACIDS

Kenneth D. Kopple
Illinois Institute of Technology

INTRODUCTION TO STEREOCHEMISTRY
Kurt Mislow
Princeton University

PRINCIPLES OF MODERN
HETEROCYCLIC CHEMISTRY
Leo A. Paquette
The Ohio State University

STRUCTURE DETERMINATION
Peter Yates
University of Toronto
 Second Edition

ORGANIC REACTION
MECHANISMS
AN INTRODUCTION

ROMLD BRESLOW Columbia university

W. A. BENJAMIN, INC. New York

19 6 9
ORGANIC REACTION MECHANISMS: AN INTRODUCTION
Second Edition

Copyright © 1969 by W. A. Benjamin, Inc.

All rights reserved
Standard Book Numbers: 8053-1252-8 (Cloth)
8053-1253-6 (Paper)

Library of Congress Catalog Card Number 73-80663

Manufactured in the United States of America
12345K32109

The manuscript was put into production on January 13, 1969;

this volume was published on September 5, 1969

W. A. BENJAMIN, INC.. New York, New York 10016

 EDITOR'S FOREWORD

undergraduate education in chemistry is in the midst of a ma-

jor revolution. Sophisticated material, including extensive treat-
ments of current research problems, is increasingly being intro-
duced into college chemistry courses. In organic chemistry, this
trend is apparent in the new "elementary"
textbooks. However,
it has become clear that a single no matter how sophisticated,
text,

is not the best medium for presenting glimpses of advanced mate-

rial in addition to the necessary basic chemistry. A spirit of critical
evaluation of the evidence is essential in an advanced presentation,
while "basic" material must apparently be presented in a relatively
dogmatic fashion.
Accordingly, we have monographs in-
instituted a series of short
tended as supplements to a first-year organic they may, of
text;
course, be used either concurrently or subsequently. It is our hope
that teachers of beginning organic chemistry courses will supple-
ment the usual text with one or more of these intermediate level
monographs and that they may find use in secondary courses as
well. In general, the books are designed to be read independently
by the interested student and to lead him into the current research
literature. It is hoped that they will serve their intended educa-

tional purpose and will help the student to recognize organic chem-
istry as the vital and exciting field it is.
VI
m EDITOR'S FOREWORD

We welcome any comments or suggestions about the series.

Ronald Breslow
New York, New York
December, 1964
 -PREFACE TO THE FIRST EDITION

it is generally agreed that some study of organic reaction

mechanisms should be part of undergraduate organic chemical
training. There are several good reasons for this trend: (1) an
elementary grasp of how reactions occur is a great help in remem-
bering the factual data which constitute the bulk of an elementary
organic chemistry course; (2) an understanding of reaction mech-
anisms serves as a guide in the design of synthetic sequences; (3)
the investigation of organic reaction mechanisms is one of the most
active areas of current chemical research, and students should at
least be made aware of the existence of this branch of the science;

(4) a study of the type of evidence which is used to establish

reaction mechanisms introduces the student to a rigorous and
stimulating style of thought and induces him to be critical in evaluat-
ing the significance of scientific evidence.
Most modern elementary textbooks introduce students to the
ideas of reaction mechanisms, and they use mechanistic outlines
in describing reactions. However, only in advanced monographs
such as Gould's Mechanism and Structure in Organic Chemistry
or Hine's Physical Organic Chemistry are unified outlines of the
field of organic reaction mechanisms presented, together with a
critical examination of the evidence for proposed mechanisms.
In the course of teaching undergraduates the author has long felt

vii
Viii m PREFACE TO THE FIRST EDITION

the need of a brief introductory book in this area; the present

volume is an attempt to fill this need. It is hoped that the book
can be read independently by undergraduates who are enrolled in
a good course in elementary organic chemistry. It should also
be useful in intermediate level courses and for advanced students
whose training in this area of organic chemistry was not up to
current standards.
In order to solve the problem of providing an introductory
treatment and at the same time indicating the current state of re-
search on mechanisms, I have adopted the device of associating a
special topic with each chapter of the book. While the chapters
are general introductions, each special topic deals with an area of
current research interest in some detail. General references follow
the chapters, but the special topics have footnotes to lead the
reader directly to the current research literature. Tables of
quantitative data are provided in both chapters and special topics
when they seem useful; organic chemistry is still far from being a
mathematical science, but the inclusion of a discussion of partial
+
rate factors and the a -p relationship in aromatic substitution, for
instance, indicates to the student one direction in which the science
will certainly grow.
The first chapter on bonding and the second one on methods for
determining reaction mechanisms establish the framework for the
discussion which follows; each of these chapters is followed by a
special topic which further illustrates the chapter material in terms
1
of current research. The remaining five chapters, with their spe-
cial topics, take up specific classes of reactions and discuss their
mechanisms. The criteria used to select these classes of reactions
are ( 1 ) the reactions are important in synthetic organic chemistry,
and (2) a fair amount is known about their mechanisms. Even
so, some important reaction types have been neglected (e.g.,
catalytic hydrogenation, electrophilic aliphatic substitution, photo-
chemical reactions) because of the desire to keep the and price size
of this book attractive to its potential readers. However, it is
hoped that the choice of topics made will indicate both the scope
and depth of current mechanistic theories.
A number of people have read this manuscript at various stages
of its preparation, and it is a pleasure to be able to acknowledge

1
There are six remaining chapters in the second edition.
PREFACE TO THE FIRST EDITION ix

their helpful comments. Professors Paul Bartlett, Esther Breslow,

David Curtin, William S. Johnson, David Lemal, Andrew Streit--

wieser, and Cheves Walling and Dr. Marjorie Caserio all read the
manuscript and made important suggestions at a sacrifice of their
own valuable time; my students, Drs. John Brown, Sheila Garratt,
Roger Hill, and Edward Robson and Messrs. Lawrence Altman,
David Chipman, and Edmond Gabbay not only read and com-
mented on the manuscript, but also tolerated the diversion of my
attention away from their research problems. Finally, I wish to
thank Mrs. Helga Testa for her outstanding services as typist.

Ronald Breslow
New York, New York
April 1965
 PREFACE TO THE SECOND EDITION

Although less than four years has elapsed since publication of

the first edition of this book, progress in the field of reaction
mechanisms has made a second edition essential. In this new
edition we have brought the previously covered topics up to date
and added new material on ribonuclease, the Br0nsted relation-
ships, hydroboration, and other areas.
The most striking advance since the first edition appeared is the
emergence of the ideas of Woodward and Hoffmann on orbital
symmetry and its role in determining reaction rates and stereo-
chemistry. These ideas are now briefly introduced in Special
Topic 1 and 4 and form the subject of a new Special Topic 8. A
new Chapter 8 on Photochemistry has also been added in recog-
nition of the developing mechanistic patterns in this branch of
chemistry.
Helpful comments from a number of users of the book have
guided the revisions. I would particularly like to acknowledge the
contributions of Prof.Thomas Spencer of Dartmouth College and
Prof. Robert Bergman of California Institute of Technology for
critical reviews of the first edition, and Prof. Nicholas Turro of
Columbia University for his review of the new Chapter and Special
Topic 8.

Ronald Breslow
New York, New York
June 1969
 CONTENTS

EDITOR S FOREWORD V
PREFACE TO THE FIRST EDITION vii
PREFACE TO THE SECOND EDITION xi

1. BONDING IN ORGANIC COMPOUNDS 1

1-1 Simple Molecules 1

1-2 Conjugated Molecules 6

1-3 Organic Acids and Bases 14
General References 20
Special Topic 1. Aromaticity and the 4n + 2 Rule 22

2. REACTION MECHANISMS AND REACTION RATES 36

2-1 Classes of Organic Reactions 36
2-2 Reaction Mechanisms 38
2-3 Kinetics 40
2-4 Reaction Rate Theory 44
2-5 Catalysis 51
General References 56
Special Topic 2. Catalysis by Enzymes 57

xiii
xiv CONTENTS

3. NUCLEOPHILIC ALIPHATIC SUBSTITUTION 67

3-1 Reaction Mechanisms 67

3-2 Reactivity in Nucleophilic Substitution 82
General References 91

Special Topic 3. Carbonium Ion Rearrangements 93

4. IONIC ELIMINATION AND ADDITION REACTIONS 109

4-1 Elimination Reactions 110

4-2 Addition to Carbon-Carbon Double Bonds 121
General References 128

Special Topic 4. Three- and Four-Center Addition Reactions 129

5. AROMATIC SUBSTITUTION 147

5-1 Electrophilic Substitution 148

5-2 Nucleophilic Substitution 158
General References 163
Special Topic 5. The Role of tt Complexes in Aromatic Substitution 165

6. REACTIONS OF CARBONYL COMPOUNDS 174

6-1 Enols and Enolates 174

6-2 Addition to Ketones and Aldehydes 183
6-3 Reactions of Carboxylic Acids and Their De-
rivatives 192
General References 202

Special Topic 6. Oxidation-Reduction Reactions 203

7. REACTIONS INVOLVING FREE RADICALS 214

7-1 Stable Organic Radicals 214

7-2 Bond Dissociation Energies 217
7-3 Radical Chain Reactions 220
General References 233

Special Topic 7. Polar Effects in Free Radical Reactions 235

CONTENTS XV

8. PHOTOCHEMISTRY 243

8-1 Excited States 243

8-2 Energy Transfer 247
8-3 Photochemical Reductions 247
8-4 Photochemical Cycloadditions 250
8-5 Photoisomerizations 254
General References 255

Special Topic 8. Orbital Symmetry Relationships in Thermal

and Photochemical Rearrangements 256

INDEX 269
 BONDING IN ORGANIC COMPOUNDS

the primary concern of this book is with chemical reactions

and their mechanisms. However, much of the evidence on
reaction mechanisms is derived by considering the relative reac-
tivities of similar compounds; changes in a molecule which speed

up or slow down one of its reactions tell us much about the

mechanism of that reaction. In order to understand the relation-
ship between structure and reactivity it is first necessary to review
certain aspects of valence theory.

1-1 Simple Molecules

Methane. Bonding between two atoms can be described by

saying that putting an electron pair in the space between two
positive nuclei "cements" them together. In these terms methane
can be symbolized as is shown below, the line simply representing
an electron-pair bond.
H

H— C-H
I
2 ORGANIC REACTION MECHANISMS

On a higher level of sophistication the bonds are described in

terms of atomic orbitals and their overlap. Thus carbon has two
electrons in its inner Is shell which are not involved in bonding,
and its accommodate up to eight electrons
outer valence shell can
and three 2p orbitals, two electrons in each. These are
in the ,2s
shown below (the x, y, and z axes are arbitrary in direction but
must be mutually perpendicular).

Carbon atomic orbitals.

The 2s orbital has an inner sphere and an outer shell,

two parts,
and the p have two parts, or lobes, projecting on
orbitals also
opposite sides of the carbon nucleus. In methane these four atomic
orbitals are hybridized into four tetrahedral orbitals, which are
3
called sp orbitals. These have somewhat the form of a p orbital,
but the mathematical parts of valence theory indicate that adding
the 25 % of s character should build up one lobe of a p orbital and
diminish the other; the quantitative treatment furthermore indi-
cates that these four orbitals will be tetrahedrally oriented. Bond-
ing occurs by overlap of each of these tetrahedral orbitals with a
hydrogen Is orbital.
In methane the four orbitals are completely equivalent, but in
less symmetrical molecules hybridization may occur so as to put

more s character in some orbitals and more p character in others.

109°28'

An sp orbital. Tetrahedral set of four sp orbitals.

BONDING IN ORGANIC COMPOUNDS . 5

An important consideration is that a carbon 2s orbital is 5.3

electron volts (122 kcal/mole) lower in energy than is a 2p or-
bital; accordingly, hybrid orbitals with more s character will at-

tract electrons more strongly. For instance, in methyl chloride the

very electronegative chlorine atom attracts the electrons of the
—
C CI bond. The methyl group can release these electrons some-
what by using a hybrid orbital for bonding with less than 25% s
character, so that the orbital is less electron-attracting, but then
the extra s character shows up in the three C —H bonds.

ch aracter
r l e>> s

More s character

This results in a decreased electron density at the protons

compared with that in methane, since the more electro-negative
carbon orbital (with extra s character) pulls the bonding electrons
closer.
Geometrical factors can also cause a departure from perfect
tetrahedral hybridization. For example,
in cyclopropane the ring
angles of 60° are much smaller than the 109° between tetrahedral
orbitals; on the other hand, simple p orbitals make an angle of 90°,
which is much closer to the ring angle. Accordingly, in cyclopro-

More than 25% s character " Pv More than 75% p character

Cyclopropane
4 ORGANIC REACTION MECHANISMS

pane the ring bonds are made using hybrid orbitals with more p
character; the bonds to hydrogen thus have more than 25% s
character, so they are shorter than the C H bonds in methane. —
It is apparent that very few saturated molecules will have

perfect tetrahedral hybridization, but for most purposes this is

ignored and bonding by saturated carbon is stated to involve the
use of sp 3 orbitals.
Ethylene. One could imagine a description of the ethylene
double bond in terms of overlap of two tetrahedral orbitals from
each carbon. In contrast to this "bent-bond" approach is a
description in which each carbon atom hybridizes the s orbital with
only two of the p orbitals. Such an sp 2 hybridized carbon will
have three hybrid orbitals in a plane (the xy plane if p x and p 7
were used) at 120° angles, with the remaining p orbital (p z )

The bent-bond description of ethylene.

perpendicular to the plane. Then the double bond between

carbons will consist of one ordinary single bond from overlap of an
2
sp orbital from each carbon (a o- bond, or straight-line bond, in
which the major electron density is along the internuclear line) and
one special bond formed by sideways overlap of the pz orbitals (a tt
bond, in which the major electron density is above and below the
internuclear line).

-'c > \
Side view.

The <r bonds of ethylene. The molecule lies The it bond of ethylene. This bond is produced
in the plane of the paper and the p, orbitals by the combination of the p* orbitals of each
are perpendicular to it. carbon.
BONDING IN ORGANIC COMPOUNDS 5

One difference between these two modes of description is that the

bent-bond model predicts an H C H angle of 109°28' while ——
the <r tt double-bond model predicts 120°. The actual value is
117°, intermediate but close to the predictions of the usual o- tt

description.
The p orbital can be exactly described by a mathematical func-
tion, and the square of the function gives the electron densities
which we have been picturing. The function itself has a positive
sign in one lobe of the p orbital and a negative sign in the other
lobe (and a value of zero at the nucleus). Two parallel p orbitals
may thus be arranged in two possible orientations, in one of which
lobes of the same mathematical sign are together, while in the other
the functions are of opposite sign.

-» . . I
A
W (p

bonding
' +Pj)
W (P '~ P2)

antibonding

7i orbitals of ethylene

The first combination (formed by adding the mathematical func-

tions for the two p orbitals and multiplying the sum by 1/V2» a
normalizing factor) forms the stable bonding ?r orbital into which

the two electrons of the ?r bond are placed. The second combina-
tion is antibonding and leads to a high-energy molecular orbital;
this is vacant in normal ethylene but would be occupied in an
electronically excited state, as is discussed in Chapter 8.

Ethylene has no dipole moment. Even though the C —H bond is

undoubtedly somewhat polar the geometry of the molecule is such

that the four C —H dipoles cancel; the same geometrical principle
that C —H dipoles will cancel can be shown for saturated hydrocar-
bons, even quite complex ones. However, propylene has a dipole
moment of 0.34 Debye units, and this reflects a slight drift of
electrons from the methyl group toward the unsaturated carbon.
0.34 Debye units.

CH^ „
c =c
6 ORGANIC REACTION MECHANISMS

This suggests that one carbon atom is more electronegative than

another one, which is at first surprising. However, the important
point is the electronegativity of the orbitals, since the electrons will
3
drift away from the sp orbital of saturated carbon and into the sp 2
orbital of the unsaturated carbon. A 2s orbital is of lower energy
than a 2p orbital; consequently electrons will be more stable in the
orbital with more s character.
Acetylene. The C— C triplebond could again be described
in terms of bent bonds, but the more usual description involves a <j
bond and two ?r bonds. Thus the carbon 2s orbital is hybridized

The two sp hybrid orbitals at an The tr bonds of acetylene,

acetylenic carbon, each with one
large and one small lobe.

The I- bonds of acetylene. The molecule is really

cylindrical!/ symmetrical.

with one of the 2p orbitals; the result is two sp hybrid orbitals

pointing in opposite directions along the same line. These two are
used to construct the a bonds to hydrogen and carbon, while the
remaining two 2p orbitals on each carbon are used for bonds. -n-

1-2 Conjugated Molecules

Benzene. The idea of the electron pair bond between ad-

jacent atoms accounts poorly for the structure of benzene. Here
the carbons are again sp 2 hybridized, all carbons and hydrogens

lying in a plane (the xy plane). The remaining p orbitals (p z

orbitals) are all parallel to each other and perpendicular to the
molecular plane. One could perhaps imagine preferential w bond-
ing of a p orbital with only one of its neighbors if the bond distances
were not equal, but since all carbon-carbon distances are 1.39 A,
BONDING IN ORGANIC COMPOUNDS 7

orbitals should pair equally well on either side. This is indicated

in valence bond theory by retaining the idea of the electron pair
bond but writing benzene as a resonance hybrid of two "canonical"
structures. These latter simply indicate pairing schemes for the
orbitals,and the resonance arrow indicates that the real structure
involves both ways of pairing.

The atomic p t orbitals in benzene. A resonance hybrid of two different pairing schemes.

Some other pairing schemes can be imagined, involving overlap

across the ring. These will contribute less to the structure of
benzene, although they are taken into account in quantitative
valence bond calculations.

A set of higher energy canonical structures (resonance

forms) which contribute only slightly to the structure
of benzene.

Such resonance structures imply no change of geometry; they

simply represent ways of pairing orbitals in the molecule which has
the flat symmetrical structure of benzene.
The fact that benzene does not have three isolated double
?r electron system, means that it is
bonds, but instead a delocalized
considerably more stable than might have been otherwise ex-
pected. The amount of this stability is indicated by heats of
hydrogenation. When an ordinary double bond is hydrogenated
28.6 kcal/mole of heat is evolved under certain standard condi-

tions. benzene had three ordinary double bonds it should thus

If

evolve 85.8 kcal/mole on reduction to cyclohexane; in fact only

49.8 kcal/mole are given off. This means that benzene contains
36 kcal/mole less energy than might have been expected if it were
8 ORGANIC REACTION MECHANISMS

not conjugated (this 36 kcal/mole is called its "resonance en-

ergy"). Valence bond theory accounts for this extra stability.

Resonance interaction between canonical (hypothetical) struc-
tures is stabilizing. The more resonance forms (stable pairing
schemes) can be written for any molecule the more stabilized the
actual compound will be; unstable resonance forms contribute little
to either the structure or the stability. If we count only those
structures without longbonds or ionic bonds (the stable structures
are often called "Kekule forms"), we can account for the reso-
nance energies (amounts of extra stability) of a number of
aromatic systems (Table 1-1). Here we have drawn only one
Kekule form for each compound. This is a common practice
which we shall follow throughout the book to save space; the
actual structure should of course be symbolized by drawing
all the important resonance forms, interconnected by resonance

arrows.

TABLE 1-1

Resonance Energies versus Number of Resonance Forms

Resonance Energy
Molecule Number of Kekule Forms . in Kcal/Mole

2 36

Benzene

3 61

Naphthalene

83.5

Anthracene

91.3

Phenanthrene
BONDING IN ORGANIC COMPOUNDS

There is an alternative way of describing molecules, by the

molecular orbital approach to molecular quantum mechanics.
This is quite different from valence bond theory in that it
abandons entirely the idea of an electron-pair bond. Although
simple molecules such as methane can also be described in molecu-
lar orbital terms mathematically, there is little useful difference in
the qualitative picture which emerges. However, for conjugated
systems molecular orbital theory has some striking advantages.
We shall apply it in a simple and semiempirical fashion; it should
be emphasized that intense research in theoretical chemistry today
is concerned with serious calculations of molecular properties by

the application of molecular orbital methods.

Let us assume that the single bonds in benzene, carbon-to-
carbon as well as carbon-to-hydrogen, are adequately described
already in terms of the electron-pair bond. However, instead of
placing the six ?r electrons in atomic orbitals and only then consid-
ering pairing and exchange interaction, let us first combine the p z

atomic orbitals into molecular orbitals, encompassing all the car-

bon atoms. Then we place the ?r electrons directly in these new
orbitals. Three such orbitals will be needed to accommodate the

Less stable
Energy =*+ ft

+ _ Y = + 2p -p -p3-2p -p
^(P 2 P3 P 5 "P6) 3
7
!

i
2(p 6 I 2 4 5

Occupied benzene ~ molecular orbitals.

10 m ORGANIC REACTION MECHANISMS

six 7T electrons, two in each, and by quantitative application of the

theory (and unfortunately not otherwise) the shape and energy of
these three orbitals can be derived.
There are also three unstable molecular orbitals which ordinarily
play no role. However, when benzene absorbs ultraviolet light an
electron is excited from a stable orbital into an unstable one (cf.
Chapter 8 ) With quantitative molecular orbital calculations it is
.

possible to predict the energy required for this process, and thus to
predict the ultraviolet spectrum of the compound.
The energies of the three occupied benzene orbitals are given in
terms of two quantities, a and (3: a, the coulomb integral, is the
energy of an ordinary isolated carbon 2p orbital, while (3, the
resonance integral, is a negative energy whose value cannot be
conveniently computed; /? is therefore evaluated from experimental
data such as heats of combustion or hydrogenation. In the lowest
energy benzene molecular orbital, E = a + 2/3, an electron is more
stable by 2/3 (a negative energy), because it is delocalized over all

six atoms, than it would be if it were isolated in a single carbon 2p

orbital (with energy = a).
This orbital, *i, has p z orbitals lined up with positive lobes
all six

above and negative lobes below the plane of the ring, so each p—p
interaction is bonding. In orbital # 3 the p z orbitals on carbons 6,
1, and 2 are lined up with positive lobes up, but 3, 4, and 5 have

positive lobes down. The result is that the region of the molecule

between atoms 6, 1, and 2 (and also between atoms 3, 4, and 5) is

a bonding region, where electrons are stable, but between atoms 2
and 3 (and 5 and 6) there is an antibonding region. This is like
the antibond in the excited-state orbital of ethylene, and corre-
sponds to a region in which electrons are of higher energy. In
fact, precisely halfway between carbons 2 and 3 there is a node
BONDING IN ORGANIC COMPOUNDS 11

where the w molecular orbital has a value of zero (in passing from
a positive function at carbon 2 to a negative function at carbon 3),
and at the node the electron density is zero. The mathematical
form of ^ 3 corresponds to this picture: p 6 p u and p 2 have positive ,

signs, while p 3 p 4 and p 5 have negative signs.

, , Electron density
tends to be greater in the center of each bonding region, at carbons
1 and 4, as is shown by the fact that
p x and p 4 have twice the
weight of the other individual orbitals in the expression for ^3 .

The 1/V6, W4, and \/\/\2 are normalizing factors which

indicate that * l5 ^ 2 , and ^3 are weighted averages of the p z orbitals
rather than simple sums.
The * 2 is similar in having two nodes, but these come at
orbital
carbons and 4 rather than between adjacent atoms. Since the
1

function has a value of zero at these two atoms, p x and p 4 do not

appear in the expression for ^ 2 Benzene has six-fold symmetry, .

and it may seem strange that carbons 1 and 4 have been

at first

singled out as different from the others. However, the extra

electron density at carbons 1 and 4 in >^ 3 makes up for the zero
density in # 2 and with six electrons placed two each in ^i, >£ 2 and
, ,

^3 the total electron density is six-fold symmetric. The orbitals

#2 and ^3 have the same energy, <x + /3, and both are less stable
than #i, since nodes and antibonding regions raise the energy of an
orbital.
With two electrons in each of these three orbitals, the total n
electron energy is 6a + 8/3.

An ordinary double bond also has its electrons delocalized, over

two carbons; the energy of each such electron in an ordinary
double bond is computed to be a + (3. Thus, if benzene had three
ordinary double bonds the total ?r electron energy would be
6a +6/?. A real benzene molecule is therefore predicted to be 2(3
more stable ([6a + 8/3] — [6a 4- 6/?]) than the hypothetical un-
conjugated cyclohexatriene. The 2/? of stabilization is often
equated with the — 36 kcal/mole derived from heats of hydrogena-
tion, so that (3 is then set equal to — 18 kcal/mole. Although a
number of gross oversimplifications are involved in this theory, it

is reasonably successful at correlating the resonance energies of

various aromatic systems (Table 1-2).
Butadiene. In this open-chain molecule conjugation effects
play only a small role. Thus, while it used to be thought that the
short C2 — 3 bond (1.483 A compared with 1.526 A for the
12 ORGANIC REACTION MECHANISMS

—*4 —»-|
r1.483A r!.526A

CH
r CH 2 CH3
1 J#— H

Butadiene Propane

TABLE 1-2
Resonance Energies Calculated by Simple Molecular Orbital Theory

Calculated Resonance Energy

Kcal/Mole Measured Resonance
Compound /3's (B = 18Kcal/Mole) Energy in Kcal/Mole

2.00 36.0 36.0

Benzene

3.68 61.4 61.0

Naphthalene

5.31 95.6 83.5

Anthracene

5.45 98.1 91.3

Phenanthrene

7.28 131.0 117.7

Triphenylene
BONDING IN ORGANIC COMPOUNDS 13

C —C bonds of propane) was due to double bond character,

single
it is now clear that most of the shortening is due to hybridization.

The C 2 —
3 bond is essentially a single bond, but a single bond
2
involving (more or less) sp hybrid orbitals; with more s charac-
ter in the orbitals such a bond is shorter than that made by satu-
rated carbons.
Measured heats of hydrogenation indicate only 3.5 kcal/mole
of resonance energy for butadiene, and some of this apparent
"resonance energy" is probably also ascribable to other effects.

Accordingly, only one pairing scheme has major significance in

butadiene; resonance forms with long bonds do not contribute
much to the structure.

H H H H
\ C-C / < >
\
Cz=C
/
CH
/2
% CH 2 CH 2
/ N
CH 2
Minor resonance forms
H CH 2 H ^ CH 2
\ C— C ^ 4 >
x
/ c=c
/
CH
2
/ \H CH 2
// \
H

When a proton is added to butadiene, however, the situation

changes. The resulting carbonium ion has two almost equivalent
resonance forms, and its structure is a hybrid of the two.

© e
CH 2 =CH — CH =CH + 2 H > CH 3 — CH - CH =CH 2

CH 3 — CH = CH— CH,

Many other reactions of butadiene also proceed through interme-

diates which are strongly resonance-stabilized. Thus, even though
butadiene itself does not show major conjugation effects, such
effects may become important in its reactions. In the above case,
for instance, the formation of a strongly stabilized (by conjuga-
tion) carbonium ion from a weakly stabilized diene makes the
reaction favorable. Such effects are most strikingly seen when one
considers the relative strengths of organic acids and bases.
14 ORGANIC REACTION MECHANISMS

1-3 Organic Acids and Bases

In water solution a proton acid ionizes by proton transfer to the

solvent. An equilibrium constant can be measured for this
process, and the negative logarithm of this equilibrium constant is
called the pK a of the acid. It is convenient to think of the pK a
as the pH of a solution in which the acid is 50% ionized or neu-
tralized; thus weak acids will have a high pK a , since they will
require a high pH (strong alkali) to lose their protons.

HA + H 0±^H 0© + A e
2 3

[A"]
pKa = pH when =^ = 1

It is not possible to measure pK a 's in water if they are much

outside the range of to 14, since water itself is both an acid and a
base. An acid so strong that its pK a is below will simply transfer
all its protons to water, forming the new acid H 3
+
, while if

the acid is so weak that its pK a liesabove 14 a base will not re-
move its proton, but instead will attack water to form OH". How-
ever, it is possible to extrapolate outside this region using other
data. In Table 1-3 are listed pK a 's for some compounds of
particular interest.
Since the ionization of a strong acid yields a weak base (called
the conjugate base of the acid which formed it), and vice versa,
Table 1-3 can also be interpreted in terms of base strengths. Thus
it shows us that the methyl anion is an extremely strong base while

the nitrate ion is not very basic, and it also indicates that aniline is a
weaker base than methylamine. is

The position of an equilibrium depends on the relative stabili-

ties of the species on both sides of the equation, and the exact

nature of this dependence can be described in mathematical

form:

AG° = - 2.3RT log K

Here AG° is the energy change, called the standard free energy
change, on going from starting materials to products; K is the
equilibrium constant of the reaction and T is the absolute tem-
perature, while R is the gas constant. For our purposes it will be
enough to note that if one substitutes the correct numbers this
BONDING IN ORGANIC COMPOUNDS 15

TABLE 1-3

Approximate pK n 's of Some Proton Acids

Compound C— H IONIZATIONS PKn

CH 3 — Methane 58

C 6 H 5 CH»-H Toluene 37

C 6 H 5C = C-H Phenylacetylene 21

O Acetone 20
CH 3 CCH — H
||

H— CH N0 2 2 Nitromethane 10.2

H— CN Hydrogen cyanide 9.14

O Acetylacetone 9.0

(CH C)>CH —
||

H — CH(N0 2) 2 Dinitromethane 3.6

N— H IONIZATIONS
NH, — Ammonia 36

C 6 H 5 NH —H Aniline 27

CH3NH, —H Methylammonium ion 10.64

Pyridinium ion 5.17

®
C 6 H 5 NH 2 -H Anilinium ion 4.58

0— H IONIZATIONS
(CH 3 )3 CO- H /- Butyl alcohol 20

CH3O —H Methanol 16

H2 Water 15.7*

C6 H5 —H Phenol 9.95

CH 3 C0 -H
2 Acetic acid 4.76

HC0 2 -H Formic acid 3.77

CF 3 C0 2 —H Trifluoroacetic acid -0.25

H — ONO L> Nitric acid - 1.37

* Corrected from 14 since H.£) concentration is 55 moles/liter.

16 ORGANIC REACTION MECHANISMS

equation indicates that, at room temperature, a lowering of one

unit in the pKa of an acid results if its ionization is made 1.4
kcal/mole more favorable. In interpreting the pK a 's of acids,
then, we consider the relative stabilities of the acid and of the
conjugate base.
Toluene is much more acidic than methane, and this is almost
entirely due to conjugation in the benzyl anion.

H HH HH HH H H H
\CG/ \ C/ \ C/ \C / \C/

Of course the first two resonance forms shown are ordinary

benzene resonance forms, and such resonance would stabilize
toluene as well. The last three forms, however, play a role only
in the anion, so it is they alone which stabilize the anion relative to
toluene, and thus contribute to the extra acidity. If all the extra
acidity of toluene compared with methane is due to this resonance
interaction, and if the pK a 's are correct, the 21 units difference
corresponds to 29 kcal/mole (1.4 X 21) of extra resonance
stabilization in the benzyl anion. This effect of a phenyl group is
also seen in the acidity of aniline compared with ammonia, and in
the pK a 's of phenol and of the anilinium ion, but the effect is much
smaller. In the phenoxide ion, for instance, the negative charge is
more likely to stay on the electronegative oxygen; consequently the
resonance forms in which the charge is found in the benzene ring
become less important.
Acetone and nitromethane are also acidified because their anions
are conjugated systems whose true structures can be represented as
hybrids.

o o o o
II I II© I©
C > N N
CH 3
/ \©CH
<

2 CH 3
C
/ \ CH
2
0/\©
O
CH-2
<- >

CH 2
f \O ©
BONDING IN ORGANIC COMPOUNDS 17

Of course in these cases it is not only this resonance interaction

which matters, but also the fact that in the major resonance form
the negative charge is on oxygen rather than on carbon. Acetyl-
acetone and dinitromethane illustrate the strong effects observed
when two such activating groups are present. A resonance
explanation can also be offered for the acidity of acetic acid.
Even though acetic acid itself has the same number of resonance
forms as does the acetate ion, resonance in the ion is more
stabilizing since thetwo forms are of equivalent energy, which is
not the case for acetic acid. Resonance forms which involve the

e
O o
/
-\
CH 3 - -c
Vo
< -> CH 3 -
o
o©
CH 3 -
/ -* CH 3 -
/
\ O-H

development of charge separation generally do not contribute as

much.
Other factors also play a role in determining acidity. Solvation
effects can be quite important, as several examples in Table 1-3
illustrate. Although methanol and water are of comparable acid-
ity, f-butyl alcohol is a weaker acid. This probably reflects the fact
that it is difficult to solvate the anion of /-butyl alcohol since the
oxygen is so closely surrounded by methyl groups; water can sur-
round the charged oxygens in methoxide or hydroxide ions more
easily, and stabilize them by hydrogen bonding and orientation of

solvent dipoles. The fact that formic acid is more acidic than is

acetic acid has also been shown to be due mostly to better solvation
of the formate ion; one side of the carboxylate group is shielded in
acetate ion by the methyl.
Occasionally a substituent can assist in solvation and thus make
a compound more acidic. Salicyclic acid has a pKa of 2.98 while
benzoic acid has a pK a of 4.20. Comparison with other com-
pounds shows that most of the by the ohydroxyl
acidification
18 ORGANIC REACTION MECHANISMS

Salicylate ion.

group is the result of internal hydrogen bonding of the carboxy-

late ion, the OH group thus acting like a solvent molecule.
Polar groups can favor anion formation by inductive electron-
withdrawing effects. Thus trifluoroacetic acid is more acidic than
is acetic acid, and due to electron shifts in o- bonds;
this is partly

there is a general drift of electrons toward the fluorine atoms,

making the oxygen atoms slightly electron-deficient and thus better
able to accommodate a negative charge. In addition to this factor
there is a very important interaction which has been called a field
effect. A negative charge can be solvated by bringing some
positively charged particle into the vicinity, or by orienting a
solvent dipole so that its positive end is nearby. The field effect
is an intramolecular example of the same thing; the carbon-fluorine
dipole is oriented so that its positive end is near the carboxylate
ion, and this dipole-charge interaction helps stabilize the ion. Such
field effects of polar groups are quite common; another example
may be found in the nitrophenols. Since the pK a of p-nitrophenol
is 7.14 and that of ra-nitrophenol is 8.35, both are more acidic than
phenol itself (9.95). In the p-nitrophenoxide ion a resonance
form is important in which the negative charge is placed in the
nitro group, but in m-nitrophenol this is not possible and the major
stabilization of the anion comes from interaction of the negative
charge with the dipole. Such an interaction also plays some role in
p-nitrophenoxide ion.
Inductive effects and field effects also play a role in some of the
systems which we have discussed only in terms of resonance; thus
obvious polar effects are expected in the carbonyl and nitro
compounds in our list. Part of the lowered basicity of aniline
compared with methylamine (increased acidity of the anilinium ion
compared with the methylammonium ion) is due to the fact that
phenyl groups have an electron-withdrawing inductive effect
compared with methyls. This is due to orbital electronegativity
BONDING IN ORGANIC COMPOUNDS 19

Oe O

Conjugative stabilization of p-nitrophenoxide ion.

O
lon-dipole interaction in p- and m-nitrophenoxide ions.

3
(an sp 2 orbital from phenyl compared with an sp orbital from
methyl in the o- bond) . Orbital electronegativity can also show up
more directly in determining acidity. Acetylenes are acidic be-
cause the —
C H o- bond involves an sp orbital from carbon; this sp
orbital accommodates an electron pair, and thus a negative charge,
quite well because of the lower energy of the s orbital. A similar
effect shows up in HCN, along with an inductive effect of nitrogen.
Hydrogens on double bonds are also acidified by orbital electro-
negativity, and the fact that pyridine is a weaker base than
methylamine is largely due to the presence of the unshared electron
2
pair of pyridine in an sp orbital.
Although by implication the factors which affect base strength
have been included in our discussion of acids, a few special points
should be mentioned. Amides are much less basic than amines.
Protonation on the nitrogen of an amide would interfere with the
amide resonance.
©
O NH 2 O NH 2
% /
c < >
\
c
#
I I

R R

Amide resonance
20 ORGANIC REACTION MECHANISMS

However, because of this resonance an amide actually protonates

on oxygen in strong acid; the protonated amide has even more
resonance stabilization, since now there are two resonance forms
with no charge separation.
It should be apparent that simple resonance arguments are not
adequate to let us predict whether this process, which does not
interfere with the amide resonance, will be easier or harder than the
protonation of a simple amine, since the two reactions are not
The same situation is met in guanidine. Protona-
really related.
tion on an amino group would interfere with the guanidine reso-
nance, but protonation actually occurs on the imino nitrogen,
leading to a strongly stabilized cation. Because of this, guanidine
turns out to be a very strong base.

G G
NH NH NH
U I I

> c
../ \,.
H2N NH 2
©/
H 2N
\..
NH 2 H 2N
../ \©
NH 2
Resonance in guanidine
©
nh 2 :nh 2 :nh,
II I I

c < > c < > c

../
H 2N
\..
NH 2 H2N
../ X©
NH 2
©/
H2 N
\..
NH 2
Resonance in guanidinium ion

General References

L. N. Ferguson, The Modern Structural Theory of Organic Chemistry

(Prentice-Hall,Englewood Cliffs, New Jersey, 1963). The most
comprehensive discussion of the material in Chapter 1

L, Pauling, The Nature of the Chemical Bond (3rd ed., Cornell Uni-
versity Press, Ithaca, New York, 1960). A good general introduc-
tion.

G. W. Wheland, Resonance in Organic Chemistry (John Wiley & Sons,

New York, 1955). A general book on bonding in organic com-
pounds. Particularly useful for material on resonance energies and
on acid-base strengths. The last chapter is an introduction to the
mathematics of valence bond and molecular orbital calculations.
BONDING IN ORGANIC COMPOUNDS 21

C. A. Coulson, Valence (Oxford University Press, London, 1961). A

rather mathematical treatment, but many sections will be compre-
hensible and interesting even to the student with little mathematical
training.

H. B. Gray, Electrons and Chemical Bonding (W. A. Benjamin, New

York, 1964). A good general introduction, chiefly concerned with
inorganic compounds.

J. D. Roberts, Notes on Molecular Orbital Calculations (W. A. Benja-

min, New York, 1961 ) . The best simple introduction to m.o. calcu-
lations for organic chemists.

A. Streitwieser, Molecular Orbital Theory for Organic Chemists

Jr.,

(John Wiley & New York, 1961 ) An extensive treatment. A

Sons, .

fair amount of background is needed to understand some sections,

but the descriptive material which outlines the results of m.o. calcu-
lations can be understood by any chemist.

H. A. Bent, "An Appraisal of Valence-Bond Structures and Hybridiza-

tion in Compounds of the First-Row Elements," Chemical Reviews,
61, 275 (1961). A discussion of hybridization, with emphasis on
the ways in which bond distances and angles depend on the exact
hybridization involved.

M. Dewar, Hyperconjugation (Ronald Press, New York, 1962). A

critical attackon the conclusions of simple molecular orbital theory,
in which it is argued that many of the effects which have been as-
cribed to "conjugation" are due instead to the dependence of single-
bond properties on hybridization.
 Special Topic

AROMATICITY AND THE 4n + 2 RULE 1

benzene is considered to be an "aromatic" compound because

its cyclic conjugation leads to great stability (for reasons we have
discussed in both resonance and m.o. terms). "Aromaticity" in
this sense of the word thus means due to cyclic
special stability
conjugation. Other definitions of aromaticity are sometimes used
which focus attention on some of the other special properties of
benzene and its relatives, but we shall restrict ourselves to
considering it in terms of conjugative stabilization (not total
stability, but that part of it which is due to cyclic conjugation).
The various substituted derivatives of benzene, such as phe-

1. For good reviews on this topic, cf. (a) A. Streitwieser, Jr., Molecu-
lar Orbital Theory for Organic Chemists (John Wiley & Sons, New
York, 1961), Chapter 10; (b) M. F. Vol'pin, "Non-benzenoid
Aromatic "Compounds and the Concept of Aromaticity," Russian
Chemical Reviews, 1960, p. 129; (c) D. Lloyd, Carbocyclic Non-
Benzenoid Aromatic Compounds (Elsevier Publishing Co., New
York, 1966).

22
AROMATICITY AND THE 4n + 2 RULE 23

nol, are aromatic compounds as well since they contain the benzene
system, but other aromatic systems are also known. Naphthalene,
anthracene, and phenanthrene derivatives contain new aromatic
systems, but these are still related to benzene since they simply
contain several benzene rings fused together. Other aromatic
compounds contain nitrogen analogs of benzene, such as pyridine
or pyrimidine rings, either singly or fused to other aromatic rings,
but again these represent only a slight extension of the benzene
concept.
A found in the aromatic furan,
slightly different situation is
pyrrole, and thiophene rings, and such extensions as the oxazole,
imidazole, and thiazole systems.

u Furan
u H
Pyrrole
u
Thiophene

N—
CI
N —
o —
o
N

H
Oxazole Imidazole Thiazole

2
The resonance energies (conjugative stabilizations) found from
heats of combustion are furan, 16 kcal/mole; pyrrole, 21 kcal/
mole; and thiophene, 28 kcal/mole. Thus all three systems are
strongly stabilized by conjugation, although not by as much as is

benzene. In resonance terms this stabilization is due to the

contribution of a series of charge-separated forms.

o — o — p — u -P e

Resonance forms for the five-membered heterocycles.

'The fact that pyrrole is more aromatic than is furan can be

ascribed to a greater resonance contribution when the positive
2. G. Wheland, Resonance in Organic Chemistry (John Wiley & Sons,
New York, 1955), p. 98.
24 ORGANIC REACTION MECHANISMS

charge is on nitrogen than when it is on the more electronegative

oxygen. The even greater stabilization in thiophene is usually 3
attributed to the fact that unlike carbon, oxygen, nitrogen, etc.,
sulfur is not totally restricted to having eight electrons in its

valence shell, since more electrons may be placed in the vacant 3d

orbitals. An extra resonance form for thiophene in which the
sulfur has expanded its valence shell to ten electrons is shown
below.

o
Closely related to these heterocycles is cyclopentadienyl anion,
another aromatic system.

> etc.

It is not really convenient to determine the resonance energy of

this anion by hydrogenation or combustion, but a good indication is

found 4
in the fact that cyclopentadiene has a pK a of about 20. At
1.4 kcal/mole for each pK unit, the difference in acidity between
cyclopentadiene and methane corresponds to a large resonance
stabilization of the cyclopentadienyl anion.
Another fundamental aromatic system is the cycloheptatrienyl
5
cation, often called the "tropylium" ion.

etc.

3. G. Cilento, "The Expansion of the Sulfur Outer Shell," Chemical

Reviews, 60, 147 (1960).
4. Ref. la, p. 414.
5. W. E. Doering and L. H. Knox, "The Cycloheptatrienylium
(Tropylium) Ion," Journal of the American Chemical Society, 76,
3203 (1954).
AROMATICITY AND THE 4n + 2 RULE 25

Here too the most convenient indication of aromaticity is a pK. A

carbonium ion may react with water in a reversible equilibrium to
form the corresponding alcohol.

R + 2H 2
+
?=± ROH + H 3
+

[Rl
pK R+ - pH when g^gj = 1

Again a pK may be defined for this reaction as the negative log of

the equilibrium constant. This is commonly 6 called a pK R + to

indicate that it involves a carbonium ion, but again it is simply
equal to the pH when the carbonium ion is half neutralized, so it

is very similar to a pK a The pK R

. + of tropylium ion is + 4.7,
5

showing that carbonium ion is so stable that even in water

this
solution it only becomes converted to the alcohol when the pH is
raised above 4.7, i.e., when the solution approaches neutrality.
An unconjugated carbonium ion would be very reactive toward
water. Even triphenylmethyl cation, in which the positive charge
is conjugated with three phenyl rings, has 7 a pK R + of — 6.63, so it
is much less stable than tropylium ion and exists only in very
strong acid solutions.

etc.

Triphenylmethyl cation, pK R + = —6.6

might seem from these examples that any cyclic system will be
It

aromatic if the electrons are delocalized so that a number of

resonance forms contribute to the structure. Interestingly, this is

not the case. The cycloheptatrienyl anion is extremely unstable

and has only recently been prepared. 8 The same instability is
apparent in the cyclopentadienyl cation; 1 it has not even been

6. Ref. la, p. 363.

7. N. Deno, J. Jaruzelski, and A. Schriesheim, "Carbonium Ions. I.,"

Journal of the American Chemical Society, 77, 3044 ( 1955)

8. H. Dauben and M. Rifi, "Cycloheptatrienide (Tropenide) Anion,"

Journal of the American Chemical Society, 85, 3041 (1963 )

26 ORGANIC REACTION MECHANISMS

possible to prepare this, although a number of derivatives are

known. The pK R + of pentaphenylcyclopentadienyl cation
9
is

approximately — 16, indicating instability so extreme that the

cation can exist only in acids stronger than concentrated H 2 S0 4 !

£.6^5^^^ ^^^-^C-ffth

Ct^t C6 H5

Nonaromatic systems Pentaphenylcyclopentadienyl cation,

pK R + -16

What can possibly explain this situation? A five-membered

anion or a seven-membered cation is aromatic, but when the
charges are reversed very unstable systems are formed. An early
10
suggestion called attention to the fact that all of the aromatic
systems contain six ir electrons (considering the fused ring aromatic
compounds as simple extensions of the monocyclic systems).
Thus the idea was put forward that an "aromatic sextet" of n
electrons had special properties. This idea would also explain
why cyclobctatetraene is not aromatic. Even though the com-
pound is not planar 11 it seemed surprising that it did not prefer to
flatten in order to gain aromatic stability; but it contains eight ?r

Cyclooctatetraene, nonaromatic and nonplanar.

9. R. Breslow and H. Chang, "The Rearrangement of Pentaphenyl-

cyclopentadienyl Cation," Journal of the American Chemical So-
ciety,83, 3727 (1961); R. Breslow, H. W. Chang, R. Hill, and
E. Wasserman, "Stable Triplet States of Some Cyclopentadienyl
Cations," Journal of the American Chemical Society, 89, 2135
(1967).
10. J. Armit and R. Robinson, "Polynuclear Heterocyclic Aromatic
Types," Journal of the Chemical Society, 127, 1604 ( 1925)
11. W. Person, G. Pimentel, and K. Pitzer, "The Structure of
Cyclooctatetraene," Journal of the American Chemical Society,
14,3431 (1952).
AROMATICITY AND THE 4n + 2 RULE 27

electrons, so according to the sextet postulate it would not be

expected to be aromatic even were planar. if it

This idea also accounts for the overwhelming evidence now

12
available that cyclobutadiene is a very unstable substance, which
can have a transient existence at best. On inspection it would
seem that this compound has two Kekule structures and that it
should be stabilized by being a hybrid of the two, just as benzene
was.

Resonance forms for cyclobutadiene, a nonaromatic system

However, it has four ir electrons instead of the magic sextet.

The requirement of a sextet of electrons for aromaticity cannot
be explained by valence bond theory. However, if we examine
the way in which cyclobutadiene is described in molecular orbital
terms it becomes possible to understand why the compound is so
different from benzene in its properties. The four it electrons of
cyclobutadiene are placed in the m.o.'s shown. 13

All four m.o.'s are pictured, although since four ?r electrons are
present the orbital of highest energy will not be occupied. Two
electrons go into ^ 1} and the other two go, one each, into # 2 and
^3 . Thus, the total n electron energy for four electrons is 4 a + 4/3;
for two isolated double bonds it would also be 4 a + 4/?, so
cyclobutadiene has no resonance energy. This certainly explains

12. M. P. Cava and M. J. Mitchell, Cyclobutadiene and Related Com-

pounds (Academic Press, New York, 1967).
13. Ref. la, p. 62.
28 u ORGANIC REACTION MECHANISMS

why it is not aromatic, although other factors must be responsible

for the observed high instability.
Unfortunately, this result depends in entirety on the fact that
quantitative m.o. calculations do predict the energies indicated for
these orbitals, and a fair background in
. quantum mechanics is
required to make these calculations comprehensible. When such
calculations are done on a variety of conjugated systems an
interesting pattern emerges. The monocyclic four ir electron
systems, such as cyclobutadiene and cyclopentadienyl cation, are
predicted to have little or no extra stability because of conjugation,
while the six tt electron systems such as cyclopentadienyl anion,
benzene, and tropylium ion are predicted to be strongly stabilized.
Thus the magic effect associated with a sextet emerges as a
prediction of m.o. theory. Strikingly, this is not the only magic
number for which aromaticity is predicted. The calculations
predict that the cyclopropenyl cation, a two ir electron cyclic con-
jugated system, should also be aromatic while its four tt electron
14
relative, the cyclopropenyl anion, should not.

H H H
Cyclopropenyl cation, a two n electron aromatic system

H H H
Cyclopropenyl anion, a four tt electron nonaromat/c system

This result can also be understood easily once we see the

stabilities predicted for the cyclopropenyl m.o.'s. Three such

14. J. Roberts, A. Streitwieser, and C. Regan, "Molecular Orbital

Some Small-Ring Hydrocarbons and
Calculations of Properties of
Free Radicals," Journal of the American Chemical Society, 74,
4579(1952).
AROMATICITY AND THE 4n + 2 RULE 29

m.o.'s are formed by combining the three atomic p z orbitals of the

ring carbons.
In the cyclopropenyl cation, two electrons are placed in the

The three m.o.'s of the cyclopropenyl system

lowest m.o., so the total * electron energy is 2a + 4/3. If the system

were unconjugated its it energy would be 2 a + 2/3 (two electrons in
an ordinary double bond) so resonance stabilizes it by 2/?, the same
as the resonance energy of benzene. In the cyclopropenyl anion,
two more electrons will be added to the next two m.o.'s; because
they happen to have the same energy each of these m.o.'s will
accept one electron and will be half -filled as in cyclobutadiene.
The total ?r electron energy for the anion is thus 4 a + 2(3. If the

system were unconjugated it would have a ?r electron energy of

4a + 2/3 (2a + 2(3 for the double bond, la each for the two
electrons in an isolated p orbital); thus, the cyclopropenyl anion
has no extra resonance energy, and it is not aromatic.
These predictions have been confirmed. A number of very
stable derivatives of the cyclopropenyl cation are known, and the
15
parent compound has recently been prepared. Thus triphenyl-
cyclopropenyl cation 16
has a pK R + of +3.1 while tripropylcyclo-
propenyl cation 17 has a pK R + of + 7.2, and is thus stable even in

15. R. Breslow, J. T. Groves, and G. Ryan, "Cyclopropenyl Cation,"

Journal of the American Chemical Society, 89, 5048 (1967);
D. G. Farnum, G. Mehta, and R. G. Silberman, "Ester De-
carbonylation as a Route to Cyclopropenium Ion and Its Mono-
and Dimethyl Derivatives," Journal of the American Chemical
Society, 89, 5048 (1967).
16. R. Breslow and H. Chang, "Triarylcyclopropenium Ions," Journal
of the American Chemical Society, 83, 2367 ( 1961 )
17. R. Breslow, H. Hover, and H. Chang, "The Synthesis and Stability
of Some Cyclopropenyl Cations with Alkyl Substituents," Journal
of the American Chemical Society, 84, 3 1 68 ( 1962)
30 ORGANIC REACTION MECHANISMS

neutral water. This is the most

hydrocarbon cation known. stable
In further agreement with these predictions,
it is found that the

cyclopropenyl anion system is not aromatic; it is so unstable that no

derivatives have yet been isolated, although they have been

CeH 5 CeH 5 CH3CH2CH2 CH2CH2CH3

/
+
C < > etc. c < -> etc.

1 I

C6H5 CH^CH^CHj

Triphenylcyclopropenyl cation, Tripropylcyclopropenyl cation,

pK R + =+ 3.1 pK R + = + 7.2
detected 18 as transient intermediates. In fact, the experimental
data strongly indicate that cyclopropenyl anion is antiaromatic , i.e.,

destabilized by conjugation. 18
The situation in the cyclopropenyl system can be contrasted with
the calculated energies of the corresponding open-chain analogs,
allyl cation and allyl anion. Three molecular tt orbitals can also be
constructed in this system: a stable bonding orbital an unstable
ty l9

antibonding orbital SJ> 3 , and a nonbonding orbital *2 whose energy

is the same as that of a simple pz orbital.

Energy Energy - a Energy V2P

V + V2P2 V2=^(P,-P3) = V2(P,-V2p + p3 )

^V2 (P '
Pa) 2

Allyl molecular orbitals

The allyl cation has two electrons in #1, and a total ?r electron

energy of 2a + 2\J2$. The cation is thus less stabilized than

cyclopropenyl cation, with tt energy 2a + 4/?. In the allyl anion

18. R. Breslow, J. Brown, and J. Gajewski, "Antiaromaticity of

Cyclopropenyl Anions," Journal of the American Chemical So-
ciety, 89, 4383 (1967).
AROMATICITY AND THE 4n + 2 RULE 31

two more electrons are placed in # 2 and the total n energy is

,

4a + 2\/2f3. This anion is more stabilized than is the cyclo-

propenyl anion, which has ?r energy 4 a + 2/3.
This relationship between the cyclic and open-chain cations and
anions can also be seen by considering orbital symmetries. The
allyl cation has its two electrons in ty u a molecular orbital with all

three p z orbitals aligned in the positive direction. Thus in the

(conceptual) conversion of the allyl to the cyclopropenyl cation,
by joining carbon 1 to 3, the interaction between pi and p 3 will be
bonding, so the cyclic system will be more stabilized. On the
other hand, in #2 orbitals p x and p 3 have opposite symmetry: p x
has its positive lobe up while p 3 has its positive lobe down. When
allyl anion is conceptually converted to cyclopropenyl anion, the
two electrons in ^i develop a bonding interaction as above, but the
two electrons in * 2 develop an antibonding interaction. Since the
coefficients of p x and p 3 are larger in V 2 (1/V2) than in ^i (1/2),
the antibonding effect dominates, and cyclopropenyl anion is less

stable than allyl anion. Of symmetry discussion has

course, this
given us only a qualitative idea of the situation, compared with the
direct quantitative use of the molecular orbital energies. However,
we will see that this kind of qualitative treatment can have wide
utility.

When m.o. calculations of this type are done for a variety of

possible aromaticcompounds, including systems not yet known, it
isfound that not only 2 and 6, but also 10, 14, 18, etc. electrons
may form monocyclic aromatic systems. These numbers fit the
form An + 2, where n = 0, 1, 2, and this generalization has
. . . ,

been called the "4rc + 2 rule" for aromaticity, or Huckel's rule

after its proponent. 19
Considering the case where n = 2, i.e., ten it electron systems, it
seems that one obvious possibility would be cyclodecapentaene. In
a ten-membered ring with five cis double bonds, delocalized to
form an aromatic system, the internal angles would be 144°
compared with the sp 2 optimal angle of 120°; this involves some
strain, although probably not a prohibitive amount. If two trans
double bonds are used the angle strain disappears, but at the
expense of severe hindrance between the two hydrogen atoms in
the middle of the ring. Neither system has yet (1968) been

19. E. Huckel, "Quantentheoretische Beitrage zum Benzolproblem,"

Zeitschrift fur Physik, 70, 204 ( 193 1 )
32 ORGANIC REACTION MECHANISMS

Two possible cyclodecapentaenes.

prepared.
20
On the other hand, two simple ten
tt electron com-

pounds are now known, both of which are apparently aromatic.

21
One is the cyclooctatetraenyl dianion; only the aromaticity of the
system explains why it is relatively easy to place two negative
charges in the same ring.

2K

etc.

\ /
Cyclooctatetraenyl dianion, a 10 tt electron aromatic system

22
The other system is the cyclononatetraenyl anion, again a stable
delocalized system.

Cyclononatetraenyl anion

Vogel has prepared some ten tt electron systems with bridging

23
groups. While these are not monocyclic, the ?r electrons are

20. For detection of an isomer, see E. van Tamelen and T. Burkoth,

"Cyclodecapentaene," Journal of the American Chemical Society,
89, 151 (1967).
21. T. Katz, "The Cyclooctatetraenyl Dianion," Journal of the Ameri-
can Chemical Society, 82, 3784 (1960).
22. T. Katz and P. Garratt, "Cyclononatetraenyl Anion," Journal of
the American Chemical Society, 85, 2852 (1963); E. La Lancette
and R. Benson, "Cyclononatetraenide: An Aromatic 1 0-^-Electron
System," Journal of the American Chemical Society, 85, 2853
(1963).
23. E. Vogel et ah, Angewandte Chemie, International Edition, 3,
228,642, 643 (1964).
AROMATICITY AND THE 4n + 2 RULE 33

restricted to the periphery of the system, and the compounds are

aromatic by spectroscopic and chemical criteria.

It should be noted that naphthalene has 10 tt electrons, as does

azulene. These aromatic compounds are not directly relevant to
the 4n + 2 rule, which is rigorously derived only for monocyclic
systems. However, the same type of molecular orbital calculation
which led to the 4n 4- 2 rule can be applied to individual polycyclic
systems, and their aromaticity can also be accounted for in molecu-
1
lar orbital terms.
One example of a 14 it electron system is the unusual compound

Naphthalene

24
I. is a polycyclic compound, but Huckel's rule still applies
This
n electron system is monocyclic, i.e., is confined to the
since the
14-membered ring around the edge of the molecule.

24. V. Boekelheide and J. Phillips, "2,7-Diacetoxy-rra/w-15,16-

34 ORGANIC REACTION MECHANISMS

25
Another example of a 14 tt a mono-
electron system is II,

cycliccompound with six double bonds and one triple bond.

Of course the triple bond has four ?r electrons, but only two of
them are part of the delocalized -k system; the other two form a
localized -n bond in the plane of the ring. Accordingly, this
compound also fits Huckel's rule, and its properties indicate that
it is aromatic. This is one of a number of large ring polyenes and

polyenynes (i.e., compounds with both double and triple bonds)

which Sondheimer and his associates have synthesized. Among
others, the 18-membered ring with 9 double bonds, which Sond-

heimer has named [18]annulene, and the 30-membered ring

[30]annulene have been made 26 both ; fit Hiickel's Rule. Good
evidence is available that [18]annulene is aromatic, but [30]an-
nulene is quite unstable. This high reactivity can be accounted
27
for on a theoretical basis, but it suggests that there may be a

A Novel Aromatic System with

Dimethyl- 1 5, 16-Dihydropyrene.
Methyl Groups Internal to the -n -Electron Cloud," Journal of the
American Chemical Society, 85, 1545 (1963).
25. L. Jackman, F. Sondheimer, Y. Amiel, D. Ben-Efraim, Y. Gaoni,
R. Wolovsky, and A. Bothner-By, "The Nuclear Magnetic Res-
onance of Annulenes and Dehydro-annulenes," Journal of the
American Chemical Society, 84, 4307 (1962).
26. F. Sondheimer, R. Wolovsky, and Y. Amiel, "Unsaturated
Macrocyclic Compounds. XXIII," Journal of the American Chem-
cal Society, 84, 274 (1962).
27. Ref. la, p. 287.
AROMATICITY AND THE 4n + 2 RULE 35

H H

H H
Cyclooctadecanonaene, an 1 877 electron aromatic system.

practical limit to our ability to extend Hiickel's rule to even larger

values of n.
 REACTION MECHANISMS AND REACTION
RATES
2-1 Classes of Organic Reactions

organic reactions are conveniently classified under four head-

ings:
1. Substitutions (sometimes called displacements)
2. Additions
3. Eliminations
4. Rearrangements

Sometimes a complex over-all reaction may fall into more than

one of these categories, but the individual steps which make up the
complex reaction can always be placed in one of the categories
listed.

The names are self-descriptive. Thus a typical substitution

reaction is the conversion of benzene to bromobenzene, substitu-
bromine for a hydrogen. Addition of bromine to a double
tion of a
bond, or elimination ofHBr to form a double bond, exemplify the
next two categories, while enolization of a ketone is a simple
example of a rearrangement (rearrangements include all cases in
which a change occurs only in the position of the atoms in a
molecule). These classifications describe reactions, but they
imply almost nothing about mechanisms.

36
REACTION MECHANISMS AND REACTION RATES 37

Reactions are also classified as heterolytic or homolytic reac-

tions. This refers to whether bonds are broken unequally, both
electrons remaining with one of the atoms, or equally with one
electron staying on each of the atoms.

x |
: Y X |J Y

Heterolytic bond cleavage Homolytic bond cleavage

Heterolytic reactions are sometimes called ionic reactions, while

homolytic processes are involved in free radical reactions. Most of
the organic reactions with which we will be concerned are ionic,

although free radical reactions will be discussed in Chapter 7.

This classification of reactions implies a bit more knowledge of the

mechanisms, and there are certain cases in which it is not immedi-
ately possible to classify a reaction as homolytic or heterolytic.
Thus in the pyrolysis of cyclobutane to ethylene two bonds are
broken, and one could imagine either a homolytic or a heterolytic
mechanism.

CH 2 I
• CH 2 CH 2 CH 2 CH 2 L CH 2
L-h > II +11 <

—
I I I
I

CH 2 '
I
CH 2 CH 2 CH 2 CH 2 i
\ CH 2
Heterolytic cleavage Homolytic cleavage

Because the product gives no clue to how the reaction went, and
because there is no way of labeling electrons to see which one goes
where, reactions of this type have sometimes been called "no-
mechanism reactions." This was overly pessimistic, as we shall
see in Special Topic 4.
Another classification, which is used only for the ionic reactions,
depends on a particular way of describing reagents. According to
the Lewis definition, a base is any species with an electron pair
which may be shared, while a Lewis acid is any species which can
bond to such an electron pair. Accordingly a Lewis acid is

electron-seeking, while a Lewis base is nucleus-seeking. Since the

terms "acid" and "base" have particular connotations, it is com-
mon which attack carbon as "nucleophilic" or
to refer to reagents
"electrophilic," depending on whether the reagent is an electron
38 ORGANIC REACTION MECHANISMS

donor or an electron acceptor in the reaction being considered.

Reactions are then classified according to the type of reagent
involved. For instance, a nucleophilic addition to a carbonyl
group occurs as one of the steps in oxime formation (see Chap-
ter 6).

O O
II

CH 3 — C— CH 3
I

CH3-C-CH3 CH3-C-CH3
C .. >
el
NH 2 >
I!

N
NH 2 I \
I
OH OH
OH

Here attention is by the use

called to the shifts of electron pairs
of curved arrows. These arrows are always used to indicate the
motion of electrons, it being understood that the atoms will follow
along. The notation is quite useful in helping to keep track of
electron pairs and in calling attention to the electrophilic or
nucleophilic character of any step, and we shall use it repeatedly
throughout the book.

2-2 Reaction Mechanisms

For any real reaction much more detail is needed to describe the
mechanism than is implied in these classifications. The mecha-
nism is literally the detailed pathway by which the reaction occurs,
and to know it would involve knowing the exact position of every
atom which plays a role, in solvent molecules as well as in reacting
molecules, at all times during the reaction. We would also have to
know the nature of the interactions or bonds between these atoms,
the energy of the system at all times, and the rate at which various
changes occur during the reaction. This is more than is known for
any reaction thus far, so we must select a more limited goal. For
our purposes we shall consider that we have made a good start
toward knowing the mechanism of a reaction if we know of all
intermediate compounds that are formed, and if we can specify in
general terms how each single step of the over-all reaction occurs.
This general description will involve as much as we can say about
which atoms are becoming attached or unattached to which others,
REACTION MECHANISMS AND REACTION RATES 39

how and what sort of bonding is found as the

readily this occurs,
reaction step progresses. It will be apparent throughout the book
that even this limited goal is successfully achieved in very few
cases. The study of reaction mechanisms is one of the most active
areas of current research, and much remains to be done.
Mechanisms are generally established by excluding the reason-
mechanism stands up to
able alternatives and by showing that the
every test which the scientist can devise. The following are
important criteria in such tests.

1 . The mechanism must account for the products. This seems

trivial, but it is an important point if we include the requirement
that the mechanism must account for the stereochemistry of the
reaction and for any isotopic labeling results. Thus, when bro-
mine is added to cyclohexene the product is trans-1,2-
dibromocyclohexane, and any mechanism proposed for this reaction
must account for the trans addition.

H Br

When chlorobenzene is treated with potassium amide the product is

aniline, but if the starting material is labeled with radioactive

carbon it has been shown that the amino group of aniline is not
found only on the same carbon atom, but is on a neighboring
carbon in about 50% of the aniline molecules. This result is
explained by the mechanism currently in favor for this reaction
(Chapter 5).

2. If intermediates are postulated in a mechanism it is desirable

that they be detected by chemical or physical means, but in any
case a real intermediate must lead to the correct products if it is
introduced into the reaction. When this does not occur the
40 ORGANIC REACTION MECHANISMS

mechanism can be excluded. Thus when r-butyl bromide is

allowed to stand in ethanol it produces a mixture of ^-butyl ethyl

ether and the olefin isobutylene. However, the ether is not
converted to the olefin under these conditions (and vice versa);
therefore the ether cannot be an intermediate in the mechanism of
formation of the olefin.

CH3 CH3 CH3

I
EtOH I
I

CH 3 — C — Br CH 3 -C + CH3-C— OEt +HBr

I

CH3
\ CH2 I

CH3

3. The mechanism must account for the effect of a change in

reaction conditions. This includes effects on the nature of the
products and on the reaction rates caused by changes in the
medium or the temperature, or produced by added catalysts. It

also means mechanism must explain the relative reactivi-

that the
ties of related compounds. This is one of the most difficult tests to

apply, since there is almost no end to the number of ways in which

a reaction can be modified. This criterion will be amply illus-
trated throughout the book.
4. The mechanism must account for the "kinetics" of the
reaction. This is really a special case of requirement 3, but it is

an important one.

2-3 Kinetics

In general the rate of a reaction will depend in some way on the

concentrations of reactants and catalysts. The study of this kind
of dependence is part of the science of kinetics. For instance, in
the reaction of hydroxide ion with methyl iodide in aqueous
solution, it has been found that the rate of the reaction is propor-
tional to the concentration of each reagent (doubling either con-
centration doubles the rate, doubling both quadruples the rate,
etc. ) . This can be described by a simple rate expression:
OH + CH3I > CH3OH + f
d[CH 3 OH]
Rate see -L_i 1 = k[OH-][CH 8 I]
dt
The rate of formation of methanol, which may be expressed in
differential form as shown, is proportional to the product of the
concentrations of OH" and CH 3 I; the proportionality constant k
REACTION MECHANISMS AND REACTION RATES 41

is called a rate constant. (It should be noted that k for this re-

action is not a pure number, but that it has the units of reciprocal
time and reciprocal concentration so that the units balance on both
sides of the equation.
The rate expression describes the kinetic order of the reaction.
This reaction is of second order, since the right-hand side of the
equation contains the product of two concentrations. The reaction
-
is of first order in OH", since [OH ] appears to the first power.
It should be emphasized that kinetic expressions are experi-
mentally derived: Within the limits of experimental error the rate
of formation of methanol appears to be proportional to the OH"
and CH 3I concentrations. The kinetic order in this case is inter-
preted in terms of a bimolecular mechanism.

HO^*> CH 3
-£- I ^ HO CH 3 + T

According to this mechanism OH" simply displaces I" from the

methyl group.
The great power of kinetics is derived from an important rule:
The kinetic order of any single step of a reaction is the same as the
molecularity of that step. This is reasonable, of course, since the
rate of collision between molecules A and B should depend on the
concentration of each. This rule means that any suggested mecha-
nism implies a kinetic prediction, which can be tested. Thus the
bimolecular reaction of OH" with CH 3I must give second-order
kinetics. The observed second-order kinetics is support for a
simple bimolecular mechanism, but other mechanisms could also
show such kinetics.
An illustration of this is seen in the bromination of cyclohexene.
Simple second-order kinetics is observed

Cyclohexene + Br2 -» dibromocyclohexane

Rate = k [cyclohexene] [Br 2 ]

This would be consistent with a one-step bimolecular reaction, the

Br 2 across the double bond. However, such a
direct addition of
reactionwould lead to c/s-dibromocyclohexane, whereas the trans-
dibromide is actually formed. From this and other evidence
(Chapter 4) it is clear that a two-step mechanism is involved:
42 ORGANIC REACTION MECHANISMS

Br <s> Br H
\
C=C +
/
Br 2
Slow
>
\ / V/
C — C Br I
Fast
>
\|
C -C
|/

H H H V H/ H Br

The second step is fast, so the first slow step is the "bottleneck"
through which the reaction must pass. This first step is bimolecu-
lar, with second-order kinetics; the kinetics of the over-all reaction
is identical with that of the rate-determining step, since the rate of
the reaction is the same as the rate of this slowest step.
Let us consider another reaction, the bromination of acetone in
aqueous solution at constant pH (see Chapter 6). Provided there
is any appreciable amount of bromine present the rate of this

reaction is found to be independent of bromine concentration and

first order in acetone concentration.

CH -CO-CH +
3 3 Br 2 -> CH 3 -CO-CH 2 Br + HBr
d(bromoacetone)
-=- = k( acetone)

This experimental observation is explained by the following

mechanism.

O OH
kx J
CH 3 -C — CH 3
II

> CH 3 — C=CH 2
OH O
k2 ||

CH 3 — C = CH 2 CH 3 — C— CH 2Br
I

+ Br 2 -h HBr

Each simple reaction step follows the rule that the order is the
same as the molecularity, so the rate of formation of acetone enol
(step 1) will be first order in acetone, while the rate at which the
enol is brominated will be first order in the enol as well as in
bromine. The mechanism will fit the observed kinetics provided it

is assumed that the second reaction is very fast compared to the

one.
first Then all the enol will be immediately converted to
bromoacetone no matter what the bromine concentration is (unless
it is vanishingly small); consequently the rate of the over-all
REACTION MECHANISMS AND REACTION RATES 43

reaction will depend only on the rate at which enol is formed. The
slow step, enolization in this case, is the rate-determining step of
the reaction; the subsequent faster step will not affect the kinetics of
the over-all reaction since everything which gets through the
bottleneck of the slow step is immediately converted to product.
At extremely low bromine concentrations the kinetics are of a
different form.

d(bromoacetone)
-j
= k (acetone) (Br 2 ) ( 1

This is because enolization of acetone can be a reversible reaction

if all the enol is not immediately trapped by bromine. Now the
mechanism involves a reversible equilibrium between acetone and
its enol, while bromination of the enol is rate-determining. The
rate of reaction should thus be proportional to the enol and bro-
mine concentrations.

d(bromoacetone)
—j- = k'(enol) (Br 2 )

However, at very low bromine concentration we are suggesting

that acetone and its enol can exist in a true equilibrium, with an
equilibrium constant K, so

(enol)
K= - r-; (enol) = K(acetone)
(acetone)

If we substitute this we get equation (1) above, provided we

identify k as k'K. This is k is just an
clearly all right, since
experimentally determined proportionality constant, which may be
considered to be the product of two other constants if we wish.
In addition to illustrating how the observed kinetics may change
if concentrations are made drastically different, this last example

points out a serious problem in interpreting reaction kinetics in

terms of mechanisms. The kinetics were first order in acetone,
but the enol was the species actually involved in the rate-
determining step. Because the enol is in equilibrium with the
acetone its concentration is proportional to that of the ketone, so
the enol is said to be kinetically equivalent to the ketone. The
observed rate law at very low bromine concentration could be
44 ORGANIC REACTION MECHANISMS

interpreted as if either the acetone or the enol were reacting with

bromine, and only other evidence shows that it is the enol which is
involved. It is usually the case that more than one mechanism
can be written which fits the observed kinetics of a reaction,
especially when fast equilibria make several species kinetically
equivalent.

2-4 Reaction Rate Theory

In our treatment so far we have written rate constants, k's, but

have not discussed them further. Now we shall consider what
makes one k larger than another, i.e., one reaction faster than
another related one. Such information will furnish insight into the
detailed mechanisms of individual steps in a reaction sequence.
Collision theory. It is possible to calculate how frequently
two molecules will collide under given conditions of concentration
and temperature, and it is found that in most cases only a very
small fraction of such collisions results in successful reaction (often
15
only 1 in 10 collisions). To account for this it is suggested that
reaction will occur only when the two particles collide with energy
greater than a certain minimum amount. This minimum required
energy, the activation energy, needed to help break the bonds
is

bonds being broken is lost before

since in general the stability of the
the new stable bonds are completely formed. The situation can be

X
"
~7
E„

Energy
,1

A
\ A

Progress of reaction

FIGURE 2-1
A collision-theory energy diagram.
REACTION MECHANISMS AND REACTION RATES - 45

pictured in terms of an energy profile for the reaction (Figure

2-1).
Thus which the starting materials (A)
in a simple reaction in
are less stable than the products (B) a point X will be reached
during the reaction which is of higher energy than either, and which
is sometimes called the "activated complex."

For instance, in the displacement reaction between OH" and

methyl iodide the activated complex occurs part-way through the
reaction.

HO + CH 3 — I
->
5-
H— O— - CH 3 — -
6-
l -> HO— CH + 3 I

A X B

The complex, X, is not an intermediate compound but is simply

an arrangement of atoms whose energy is higher than that of

any other along the reaction path.
The difference between the average energy of the starting
materials and the energy of this activated complex is E a the acti- ,

vation energy, and only those molecules of starting material which

have this amount of energy, E a more than the average will be
,

able to react on collision. For the reverse reaction, from B to A,

the activation energy is even larger but the activated complex is

the same, as must be since there is only one "highest point"

it

along the path which connects A and B.

It can be shown that the fraction of molecules with an energy E
E/RT
or greater is e" where e is 2.71 ...
, the. base of natural ,

logarithms. Substituting appropriate values we find that 10% of

the molecules will have at least 1.4 kcal/mole more energy than
average at room temperature, 1% will have at least 2.8 kcal/mole
of excess energy, etc., each additional 1.4 kcal/mole causing a de-
crease in the fraction by a factor of 10. With this the rate of a
reaction can be expressed as follows

Rate = Ze"
Ea/RT

where Z is the number of collisions in a unit time (for instance, per

second) and the rest of the expression shows what fraction of the
collisions are successful.
Some reactions obey this rate equation quite well. However, in
most cases it can be shown that some collisions with the required
46 ORGANIC REACTION MECHANISMS

amount of energy do not lead to reaction, because the molecules

collide in the wrong way. In only a certain fraction P of the
collisions will the atoms be lined up correctly for reaction even if
there is sufficient energy, so this probability factor is usually
included in the equation as well, the final form being

Rate = PZe"
Ba/RT

If the temperature changed the collision frequency Z will

is

change slightly. The major

effect of an increase in temperature is
to increase the fraction of molecules which have enough energy to
react, so it is possible to determine the activation energy of a
reaction by studying the temperature effect on the rate. The
value of E a tells us something about the strengths or energies of the
various bonds in the activated complex, while the value of P tells

us something about how sensitive the reaction is to the precise

alignment of the reacting molecules. Thus by the application of
collision theory and a study of the temperature dependence of the
reaction rate it is possible to get information about the detailed
mechanism of a single reaction step.
Transition state theory. There is another approach to
the theory of reaction rates that is of much more use in discuss-
ing reactions in solution. Again it depends on the idea that the
reacting species must surmount some energy barrier in order to
react, but it discards the specific consideration of collisions, and the
energy barrier is thus not formally concerned with the probability
of successful collision. Instead the activation energy is considered
to regulate the position of a particular equilibrium.
Consider a collision between two species which together have
less than the required activation energy. By collision theory they
would begin to react and form a complex which is on the way to the

activated complex, but since the energy is not sufficient the

reaction would not continue. Instead this complex would break up
to starting materials again. If we think of this occurring repeat-
edly we can see that it is possible to consider an equilibrium be-
tween the starting materials and the complex; the position of such
an equilibrium can be described by an equilibrium constant K.

A + B^A- •
B
Equilibrium between starting materials and a complex
REACTION MECHANISMS AND REACTION RATES 47

All equilibrium constants can be related to the relative stabilities

of starting materials and products; the relationship can be ex-
pressed in terms of the difference in their energies.

AG = -2.3 RT log K

Here AG° is the difference in standard free energies of products

and starting materials: AG° = G° (products) — G° (starting ma-
terials). This free energy difference is composed of two parts:

AG° = AH° - TAS°

The standard enthalpy difference, AH°, is a measure of the

difference in bond energies, solvation energies, etc., between
starting materials and products. The TaS° term multiplies the
absolute temperature by a standard entropy difference, AS°.
The entropy is a measure of the freedom of motion of the
system, and any process which causes a decrease in freedom will
cause a decrease in entropy. This factor comes into the determi-
nation of an equilibrium constant because it is relatively improb-
able that several freely moving particles will come together in a
particular way, thus losing all their freedom to move about
independently. Processes which require such loss of freedom are
less favorable. All other things being equal, when two particles
are bound in a complex there is an entropy decrease; therefore
complexes will only be stable if the AH° term makes up for the
TaS° term.
Returning to our reaction, it is postulated that the starting
materials are in equilibrium with all complexes which occur before
the activated complex, and also with the activated complex itself.
Thus the concentration of the activated complex is also governed
by an equilibrium constant, called K* to distinguish it as the
constant for this particularly important equilibrium. The impor-
tant postulate is made in this theory that all activated complexes go
on to products at exactly the same rate. This can be made
reasonable by the application of the theory of statistical mechanics,
but it is still a remarkable postulate. If it is true, then any rate

depends simply on the value of a particular equilibrium constant,

K*, for the reaction.
The universal rate constant derived from statistical mechanics
48 ORGANIC REACTION MECHANISMS

is /cT/h, where k is Boltzmann's constant, T is the absolute

temperature, and h is Planck's constant; at room temperature,
kT/Ii = 6 x 10 12 per second. All rates can be expressed in the
following form:

kT
Rate = —r— [activated complex]

However, the concentration of activated complex is proportional

to the concentration of starting materials.

[activated complex]

[starting materials]

Accordingly, for the rate of the reaction the following expression

can be written:

kT
Rate =—r— K* [starting materials]

This means that the rate constant is equal to (>T/h)K*. The rate
constant at any particular temperature is thus proportional to
the equilibrium constant K*. The value of an equilibrium
constant is determined simply by the free energy difference between
starting materials and products. Thus the rate constant of a
reaction depends on the standard free energy difference between
starting materials and activated complex; this is written AG*, and
read "free energy of activation."

AG* = - 2.3 RT log K*

Relative rates can be discussed in terms just like those we have

already used to discuss equilibria, such as relative pK a 's of acids.
An energy diagram can be drawn for a reaction path (Figure
2-2), and in form it is almost identical to that we have already
used for collision theory.
All complexes to the left of the activated complex are considered
to be in equilibrium with the starting materials, while all complexes
REACTION MECHANISMS AND REACTION RATES 49

Transition state

Activated complex

Free
energy, F

Starting materials

Products

Reaction coordinate

FIGURE 2-2
A transition-state theory energy diagram.

to the right are considered to be in equilibrium with the products.

Thus the activated complex occurs at what is called the transition
state (the place where a transition occurs from the starting-material
side to the product side) quite commonly the activated complex is
;

actually called the transition state.

It is complex series of reactions on
a simple matter to represent a
a single energy diagram. For instance, when benzaldehyde is
treated with potassium cyanide in ethanol, benzoin is formed by the
following mechanism (cf. Chapter 6):

H H
ki ki
C=0+ C=N
I 1
>
C6 H5 <
C 6 H 5C — <
C 6 H 5 -C OH
k_i 1 k_2 1

c C
III III

N N
II
50 ORGANIC REACTION MECHANISMS

H OH O
k3 k4
11+ C 6 H 5C = —C
I

C6 H5 C-C 6 H5 >

C =N H

III

O ) HO O OH
II

— C-C H +C=N
I

C6 H5 -C C-C 6^5
H 6 C 6 H 5C 6 5

^C = N H H

IV Benzoin

Here have been written, two equilibria and three

five reactions

no reaction is truly irre-

irreversible reactions (although, of course,
versible, for all practical purposes the reversibility of all steps
after the slowest one can usually be ignored). Each of these indi-
vidual reactions will have its own transition state. The highest
point of energy in the entire sequence will occur at the transition
state for the rate-determining step, which is in this case step 3,
with rate constant k 3 . Thus the rate of formation of benzoin will
depend only on the difference in free energy between the starting
materials, benzaldehyde and cyanide, and the activated complex
at the transition state of step 3. This is shown in Figure 2-3,
in which the AG* for the overall sequence is shown rather than
that for any given step.

AG* y^ ^s II

III IV
Starting materials
Free
energy

Products

Reaction coordinate

FIGURE 2-3
An energy diagram for the benzoin condensation.
REACTION MECHANISMS AND REACTION RATES 57

This means that the rate of the reaction can simply be described
in terms of an equilibrium between starting materials and activated
complex, all other equilibria or reaction steps being ignored.

OH O e
» 1
:l
C 6 H 5 CHO + CN < CeHs C C C6H5

C
1

=N H
1

arting materials Activated complex

Since the rate will depend only on the concentration of activated

complex this equilibrium tells us immediately that the kinetics will
be of the following form (as is found experimentally).

OH
Rate= k[c 6 H 5C 0] [c 6 H 5 CHo] — k' [c 6 H 5 CHo] [ CN
"

CN

Note that the first expression is kinetically equivalent to the

second. Furthermore, in considering what effect a change of tem-
perature, pH, on the reaction rate we must con-
solvent, etc., has
sider its effect only on this equilibrium, and ignore the effect
on individual steps. This makes the interpretation simpler, but it
points out once again that kinetics will not give us information
about fast equilibria which occur before the rate-determining step.

2-5 Catalysis

A catalyst is a substance which accelerates a reaction without

being itself changed
in the over-all process. It must do this, of

course, eitherby lowering the transition state energy of the uncata-

lyzed process in some way or by making a new path possible which
has a low activation energy. Many catalysts function in ways
which are still not fully understood; in particular, heterogeneous
catalysts (those not in solution) such as the platinum catalyst in
hydrogenations are in this category. On the other hand, definite
mechanisms can be suggested for the action of most catalysts in
solution; the operation of cyanide ion in the benzoin mechanism
52 ORGANIC REACTION MECHANISMS

above is a good example of

this. Since by far the most important
cases of homogeneous catalysis in organic chemistry involve the
action of acids and bases, they will be discussed in a little detail.
General acid and general base catalysis. A solution of
acetic acid in water contains two acidic species, H 3 and HOAc
+

(ignoring H 2 0, a very weak acid). The concentration of each

of these can be measured independently, since the pH is a measure
only of the H 3
+
whereas a titration would indicate the total acid
,

present. Some reaction rates depend only on the pH; such a re-
action whose rate is increased at a lower pH is said to be specific
acid catalyzed, since only the solvent-related acid H 3
+
plays a
catalytic role. By contrast, some reactions may show general acid
catalysis; the rate will increase if the HOAc concentration is in-

creased even if the pH is .Jield constant.

When the bromination of acetone is conducted in water solution
of varying pH's and acetic acid concentrations, it is found that the
reaction is catalyzed by both acidic species, and a simple rate law
is observed.

Rate = ki [acetone] [H 3
+
] + k 2 [acetone] [HOAc]

The reaction is studied with enough bromine present to guarantee

that enolization is rate-determining; therefore bromine does not
appear in the rate expression.
If we consider what is involved in the enolization of acetone it is

easy to see why an acid catalyst would help. A proton must be

removed from carbon and another proton added to oxygen; if only
water is present the solvent must perform these functions.

f\
H— OH H OH
O
/
CH 3 -C T CH 2 CH 3 -C
H. CH 2
H 2 0:-J H

OH
REACTION MECHANISMS AND REACTION RATES 53

However, it would clearly be better if the proton could be donated

by a stronger acid than water, and removed by a stronger base.
Thus the function of an acid in catalyzing this enolization can be
written very simply.

f\ Q
H— A H A
/
O O
CH 3 —
11 I

-C >

cCH I
2 CH 2
H H
I©
:J B

Here we have suggested not only that the proton donor may be
any acidic species, but also that the proton acceptor may be any
base. This is also found to be the case. For instance, in a
solution of sodium acetate both OH" and OAc catalyze the
bromination reaction, and their contribution to the rate can also be
described with a simple kinetic expression.

Rate = k 3 [acetone][OH~] + k 4 [acetone][OAc]

If both acids and bases are present, as in a buffer solution of

acetic acid and sodium acetate in water, the complete kinetic ex-
pression is the following.
Rate = k [acetone] + ki [acetone] [H 3
+
] +
k 2 [acetone] [HO Ac] + k 3 [acetone] [OH] +
k4 [acetone] [O Ac"] + k 5 [acetone] [HO Ac] [O Ac"]

The first term represents enolization accomplished simply by water.

(Water does not appear in the kinetic expression since it is the

solvent whose concentration is kept constant; only concentrations
which can be experimentally varied are shown, since the kinetics
are based on the observed variation of rate with concentration.)
The second term represents enolization in which H 3 + is the
proton donor and H 2 the base, etc. The sixth term
54 ORGANIC REACTION MECHANISMS

indicates that acetate ion can remove the proton from carbon
while acetic acid is it on the oxygen, and this raises an
putting
interesting point. Why is there no term for a reaction in which
OAc" and H 3
+
cooperate, or HOAc and OH"? The answer is
that these terms are already hidden in the expression we have
written. In water, acetic acid is in equilibrium with H 3
+
and
OAc~.

_V
HOAc <__ H3 + OAc

@
[h 3 O ][oAc
J
= K [HOAc]

.'. k[H 3 O®][oAc ] = k' [HOAc]

This equilibrium shows that the third term in the kinetic expres-
sion does not necessarily represent catalysis by acetic acid, but it

may actually indicate catalysis by H 3

+
and OAc", since the
equilibrium makes them kinetically equivalent. Similarly, the fifth
-
term, which indicates catalysis by OAc and water, may really
indicate catalysis by HOAc and OH", since this combination is
again kinetically equivalent. Other evidence suggests that both
types of catalysis contribute to the terms in the observed kinetic
expression.
This example indicates again that it is usually easier to measure
kinetics than to interpretthem unambiguously. Even so, it seems
must be accounted for in terms of
clear that acid or base catalysis
definite molecular mechanisms, and not in terms of mysterious
catalytic "forces." In later chapters, we shall encounter other
examples of acid and base catalysis.

The Br0nsted catalysis law. In the previous example we

have seen that a proton can be removed from acetone by OH",
OAc", or H
2 0. Of course, one might expect that reactions in-
volving strongly basic OH" should be faster than those with the
less basic OAc" or H
2 0, but this expectation is only intuitive.

"Basicity" is an equilibrium concept, involving starting materials

and products, while rates of reactions involve transition state
energies.
REACTION MECHANISMS AND REACTION RATES 55
*

I
AG 1

Acetone + B —»- Activated complex —m- Enolate + BH

+

"
AG' 1

With a stronger base AG° will become more negative, since the
conversion from B to BH becomes more favorable, but there is
+

no obvious relationship between AG° and AG*. Interestingly, it

is frequently found experimentally that with bases of similar struc-

ture (e.g., carboxylate anions) but differing basicities the changes

in AG° and AG* are related:

The change in AG* from one base to another is proportional

to the change in AG°, or A(AG*) = kA(AG°).

This is reasonable, since in the activated complex a proton has

been partially transferred to the base, and we might expect to see
some fraction of the effect of basicity which shows up fully when
the transfer is complete.
The foregoing linear free energy relationship can be expressed
in more standard form by the Br0nsted relation:

Kb — GbK.b

This expression indicates that the rate constant k B for a base

catalyzed reaction is proportional to the dissociation constant KB
for the base, raised to a power {3. The exponent p is called the
Br0nsted coefficient; it varies in magnitude between and 1, and
its value reflects the extent to which proton transfer has occurred
in the transition state and thus the extent to which the strength
of the base can affect the activation energy. GB is a simple pro-
portionality constant.
In the case of acetone enolization, a series of carboxylate anions
has been examined as catalysts and the Br0nsted law is followed,
with /? = 0.88. This value so close to 1 indicates that almost the
full effect of the basicity is already felt at the transition state; an in-
crease of 10 in the strength of the base leads to an increase of 7.5
88
(or 10° ) in the rate. In cases where the transition state occurs
earlier (3 is smaller in value. There is also a Br0nsted relationship
for acid-catalyzed reactions:

KA = GAKA *
56 ORGANIC REACTION MECHANISMS

and it also is generally found to hold reasonably well among a

series of acids with similar steric requirements but differing acid
strengths.

General References

A. Frost and R. Pearson, Kinetics and Mechanism (2nd ed., John

Wiley & Sons, New York, 1961). The best general reference for
this chapter. Covers kinetics, rate theory, and catalysis; in Chapter
11, several reaction mechanisms are discussed in detail.

R. P. Bell, The Proton in Chemistry (Cornell University Press, Ithaca,

New York, 1959). Particularly useful for material on acid and
base catalysis.

S. L. Friess, E. S. Lewis, and A. Weissberger, eds., Investigation of

Rates and Mechanisms of Reactions (2nd ed., in Technique of
Organic Chemistry, Vol. VIII, Interscience Publishers, New York,
1961). An exhaustive treatment of the methods used to examine
reaction mechanisms. Discusses both the experimental details and
the mechanistic implications of such techniques as isotopic labeling,
kinetics, detection of intermediates, etc.
 Special Topic

CATALYSIS BY ENZYMES

essentially all biological reactions are catalyzed by special

substances called enzymes. 1 These catalysts are proteins, which
sometimes are associated with inorganic ions or small organic
molecules; enzymes are so large that it has been possible to de-
termine the three-dimensional structure of only a few of them,
although the study of protein structures is now a very active field
2
of research. In general, each reaction requires a different kind of
enzyme, so that there are almost as many types of catalyst mole-
Thus there are groups of
cules as there are biological reactions.
enzymes which catalyze various oxidation reactions, other groups
which catalyze hydrolyses, others which catalyze carbon-carbon
bond formation, etc. This means that in detailed mechanism each
enzyme-catalyzed reaction is a special case, and a difficult one

1. For more extensive treatments of this topic see (a) S. Bernhard,

The Structure and Function of Enzymes (W. A. Benjamin, Inc.,
New York, 1968); (b) T. Bruice and S. Benkovic, Bioorganic
Mechanisms (W. A. Benjamin, Inc., New York, 1966).
2. The current state of this field can be followed in Annual Review
of Biochemistry Advances in Enzymology, or Advances in Protein
,

Chemistry.

57
58 ORGANIC REACTION MECHANISMS

considering the enormous size and complexity of the catalyst.

Even so, the general principles by which enzymes work can be
simply understood in terms of reaction-rate theory.
There are several respects in which catalysis by these sub-
stances differs from that due to simple catalytic species such as
H 3
+
,CN", etc. One of the most striking is that enzymes are
generally much better catalysts than are simple molecules. For
instance, at 25 °C the hydrolytic enzyme a-chymotrypsin 3 (molecu-
lar weight -~ 25,000) will catalyze the hydrolysis of an amide to a
carboxylic acid and an amine; as a catalyst it is about one million
timesmore effective 4 than either H 8 +
or OH", although the
enzyme operates in neutral solution. The second important
feature is that enzymes are very selective in their action.
Chymotrypsin, for instance, will not hydrolyze amides in general,
but will operate only on amides of certain amino acids. The
enzymes is so great that most
selectivity of will work with only one
enantiomer of a dl pair in an optically active substrate. This
specificity is related to the requirement that the substrate have just
the right shape to fit into a particular place in the protein, for the
third special feature of enzyme-catalyzed reactions is that the
enzyme binds the substrate in a complex before it catalyzes any
reaction.
Thus enzyme-catalyzed reactions can be written in a very
5
general way, in which the enzyme is symbolized by E, the sub-
strate by S^ and the enzyme-substrate complex by ES.

K k
E+ S^±ES^E + products
General scheme for enzyme-catalyzed reactions

This complexing can be detected by spectroscopic means, but it

was first deduced from a study of the kinetics of enzyme-catalyzed

3. P. Desnuelle, "Chymotrypsin, " in P. Boyer, H. Lardy, and K.

Myrback, eds., The Enzymes, Vol. 4 (Academic Press, New York,
1960). Chapter 5.
4. T. Bruice, "Intramolecular Catalysis and the Mechanism of
Chymotrypsin Action," Brookhaven Symposia in Biology, 15,
52 (1962).
5. Cf. H. Mahler and E. Cordes, Basic Biological Chemistry (Harper
& Row, New York, 1968).
CATALYSIS BY ENZYMES 59

reactions.
6
A typical graph of rate versus substrate concentration
is shown in Figure 2-4. At low concentrations of substrate the
rate follows a normal law,

Rate = k (enzyme) (substrate)

but at high substrate concentration the rate depends only on the

total enzyme concentration

Rate = k (total enzyme)

and is independent of substrate concentration.

[S]

FIGURE 2-4
Typical plot of rate versus substrate concentration for an enzyme-catalyzed
reaction.

This result comes about because the rate-determining step is the

further reaction of the enzyme-substrate complex, and the over-
all rate of the reaction will thus depend simply on the concentration
of ES and on the rate constant for its further reaction.

Rate = k (ES)

Now enzyme-substrate complexes are generally quite stable; with a

reasonably high substrate concentration it is possible to push the

6. L. Michaelis and M. Menten, "Die Kinetic der Invertinwirkung,"

Biochemische Zeitschrift, 49, 333 ( 1 9 1 3 )
60 ORGANIC REACTION MECHANISMS

equilibrium to the right so that essentially all the enzyme is tied up

in the complex. This is the maximum concentration the complex
can have, so it is not surprising that higher concentrations of
substrate do not increase the rate further. A mathematical treat-
ment 5 of this general mechanism, in which the rate reaches a
maximum value as all the enzyme becomes complexed, predicts
the exact shape of the plot of rate versus concentration, and the fit

with experimental results is very good.

A number of interactions are responsible for the formation of
such complexes. 7 These include ( 1 ) attractions between charged
groups on the protein and the substrate, (2) hydrogen bonds
between suitable groups, and (3) "hydrophobic interaction," the
tendency of nonpolar groups such as alkyl chains to come together
rather than be surrounded by water molecules (in a sense the
hydrocarbon sections of a protein "extract" nonpolar substrates
out of the water). The total of such binding forces can be
considerable. Thus in the complexing between chymotrypsin and
benzoyltyrosylamide, a typical substrate,

O CH 2 O
II
H |
||

C 6 H 5 C— N — CH — C-NH 2

Benzoyltyrosylamide, a substrate for chymotrypsin

the equilibrium constant 8 indicates a AG of — 2 kcal/mole on

going to the complex; such an energy decrease indicates that
complexing is a favorable process. However, from the tempera-
ture effect on the equilibrium it can be calculated that the AH° for
complex formation is —11 kcal/mole; the binding forces cause

7. (a) R. Lumry, "Some Aspects of the Thermodynamics and Mech-

anism of Enzymic Catalysis," and (b) K. Linderstr0m-Lang and
J. Schellman, "Protein Structure and Enzyme Activity," in P.

Boyer, H. Lardy, and K. Myrback, eds., The Enzymes, Vol. 1

(Academic Press, New York, 1959). Chapters 4 and 10.
8. Ref.7a,p. 178.
CATALYSIS BY ENZYMES 61

this large decrease in energy. The — TaS° term is + 9 kcal/mole,

showing that complex formation is very unfavorable in entropy
terms. This is reasonable, since there is considerable loss of
freedom when the two independent particles, substrate and protein,
are tied into one. Thus the binding forces make it possible for a
stable complex to form in spite of the unfavorable entropy
change.
Such enzyme-substrate complexes are formed in all cases, al-
though of course the exact energetics of the process depends on the
particular system being considered. Once in the complex, the
substrate is acted on by catalytic groups within the protein, usually
side-chain groups ofamino acids which are part of the protein.
Again the detailed mechanism depends on the particular reaction
being considered. In the case of chymotrypsin, the following
mechanism seems likely.
9
A general base, the imidazole ring of a
removes the proton from a serine hydroxyl group as the
histidine,
latter attacks the amide carbonyl. The departing amine may be
simultaneously protonated by a general acid group, although direct
evidence on this point is not yet available. In a second step water
attacks the new ester carbonyl group, perhaps assisted by a general
base, while the departing serine oxygen is simultaneously pro-
tonated by imidazolium ion (Figure 2-5).

O H — A-
(*> O
II

— NH —
/%^n:rf^ h——
-C II

/ OCR
hn HN NH
1
1
«- 1 1 /
1 1 .1 1 1
CH 2
j
^-»2
-R'NH
1 1

histidine histidine serine C

serine
1 /-
Enzyme

Jh 2
RCOoH * enzyme

FIGURE 2-5
A general scheme for amide hydrolysis by a-chymotrypsin.

9. M. Bender et ah, series of papers on the mechanism of action of

a-chymotrypsin, Journal of the American Chemical Society, 86,
3704, 5330 (1964) and references therein. H. Neurath, "Protein
Digesting Enzymes," Scientific American, December 1964, p. 68.
 62 ORGANIC REACTION MECHANISMS

The enzyme ribonuclease operates in quite a similar way. 1 This

enzyme is a small protein, made up of 124 amino acids, which
catalyzes the hydrolysis of ribonucleic acids. As with chymotryp-
sin the hydrolysis is a two-step process: The leaving group is first
displaced by an alcoholic hydroxyl group, and water comes into
the second step in the hydrolysis of the intermediate. In the case
of ribonuclease, however, the nucleophilic hydroxyl is on the sub-
strate, ribonucleic acid, rather than on the enzyme as in the chymo-
trypsin reaction.

Pyrimidine

O
O
\ /
O
H-B
H2

e
<?
=o
OH

o P

o o

a — h
Cyclic phosphate OH
+ Nucleic acid
Fragment II
CH 2 Purine or A: I O
O^ Pyrimidine

CH 2 Purine or
O Pyrimidine
O OH

Ribonucleic acid
O OH
Nucleic acid
Fragment I

In the conversion of ribonucleic acid to the cyclic phosphate a

proton must be removed from one oxygen atom, and a proton must
be added to another oxygen atom. We have shown a base B re- :

moving the first proton and an acid AH + adding the other; in the
reverse reaction, in which Fragment I adds to the cyclic phosphate
to produce the nucleic acid, the law of microscopic reversibility
dictates that we must run back over the same path, so the new base
A: remove the proton from Fragment I while the new acid
will
BH +
proton donor. Although the detailed mechanisms
will act as
of enzyme-catalyzed reactions have not been completely elucidated
in any case, strong evidence indicates that B: is the imidazole of
 CATALYSIS BY ENZYMES 63

histidine-119 (the polypeptide chain is numbered from the end

with a free a-amino group), and AH +
is a protonated imidazole of
histidine-12. Although one might at first expect the two histidines
to have the same pK a the difference in energy between B: and
,

BH + is affected by the environment (neighboring charges, for in-

stance) so His- 12 and His-119 may exist as shown, with 12 pro-
tonated and 119 as the free base. In the reverse reaction A: is

thus again an imidazole ring, while BH +

, the proton donor for the
reverse reaction, is again an imidazolium ion. In step 2, the hy-
drolysis of the cyclic phosphate, this same mechanism is involved,
but water replaces Fragment I, so A : is removing the proton from
H 2 in the transition state.

Lysine (V Histidine(12)- Histidine (1 19) Valine (124)

H2 N CH CO NH CH CO NH CH CO NH CH C0 2 H

I I I I

R CH 2 CH 2 R

-NH

Ribonuclease

Both chymotrypsin and ribonuclease accomplish hydrolysis by a

two-step sequence, in which a leaving group is set free before a
water molecule is tied down. It is easy to see that this is catalyti-
cally useful when we realize that the free energy of activation AG*
involves both an enthalpy and an entropy term; the latter, AS*,
measures the loss of degrees of freedom in the transition state com-
pared with starting materials — the more freedom lost the larger
AG*, and the slower the reaction. With a two-step hydrolysis
mechanism either step could be rate determining. If the first one
is, then H 2 is not yet tied down in the transition state for over-all
reaction, so its loss of freedom does not contribute to AS*; if the
second step is 2 down, but the
rate-determining, then the H is tied
leaving group is already fully free, so its gain in degrees of freedom
(compared with starting materials) balances the AS from the
water. Either situation is better than a one-step process in which
water is tied down before the leaving group is fully released.
The other point of interest is that imidazole was used as the base
64 ORGANIC REACTION MECHANISMS

by both enzymes, and imidazolium ion as an acid by ribonuclease.

This feature is also catalytically useful. Imidazole, with a pK a
near 7, is the strongest base which can exist free at neutrality; if it

were more basic, it would be already protonated by the medium

at pH 7. Similarly, imidazolium ion is the strongest acid available
at neutrality, since an acid with a lower pK a would be dissociated
at pH 7. By the Br0nsted relationships the rate of an acid- or
base-catalyzed reaction depends on the acid or base strength of the
catalyst, so imidazole is optimal for these functions.
Clearly an important part of the catalytic efficiency of any
enzyme must be due to ideal rigid placement of such groups so that
they can perform their functions without any further restriction of
freedom (loss of entropy). We shall not consider this point
further, but shall examine only one additional question: How does
enzyme-substrate complex formation contribute to the catalytic
10
efficiency of enzymes?
In transition-state theory the rate of a reaction depends on the
change of free energy in going from starting materials to activated
complex. We have emphasized that the free energy change de-
pends on two factors: AH*, which reflects changes in potential
energy, and AS*, which reflects changes in freedom. For a catalyst
these two will ordinarily oppose each other. Thus if we wish to
stabilize a developing negative charge in the transition state for a
reaction, we may do this with a general acid which donates a pro-
ton as the charge develops. This would lower AH*, and thus help
the reaction; but it would also decrease AS*, since the general acid
would be tied down next to the substrate at the transition state and
freedom would be lost. Since AG* = AH* — TaS*, the entropy
loss works against the enthalpy decrease. If we require several

independent catalyst molecules, e.g., a general base, a general acid,

a nucleophile, etc., the entropy loss becomes very large; for some
possible catalysts, this entropy loss may be more than the possible
enthalpy gain, so catalysis (rate increase) does not occur.
The entropy problem can be partly solved if all the catalytic
groups are placed in the same molecule to start with. This is the
case for an enzyme; now the translational freedom of only one
molecule is lost, as contrasted with the situation when the catalytic

10. Cf. F. Westheimer, "Mechanisms Related to Enzyme Catalysis,"

Advances in Enzymology, 24, 455 (1962).
CATALYSIS BY ENZYMES 65

groups are in However, there is still an

independent molecules.
important entropy decrease when an enzyme and its substrate go to
the activated complex. The two must be tied together, with a loss
of translational freedom, and they must be fixed in precisely the
right orientation, so that some freedom of rotation within the
molecules is lost as well. The enzyme balances this entropy loss
with an extra enthalpy decrease, the enthalpy decrease involved in
binding to form the complex.
The situation is pictured in Figure 2-6. With ordinary cata-

T>^
i
"
—%
T- TAS*
^ x
~tt\
AG*
i

\
*H*
N- -Krp-
\
-TAS*

» AHofb
t

*
1 1 1
Uncatalyzed Ordinary "Enzyme" Real
catalysts without enzyme
binding including
binding

FIGURE 2-6
The effect of various types of catalysts on AH* and T-^S*.

lysts the rate can be increased because of an enthalpy advantage,

which is, however, opposed by the entropy disadvantage discussed
above. one imagined an "enzyme" which could not bind
If

substrate there would still be a big entropy advantage in combining

all the catalytic groups in one molecule (for simplicity we have

assumed that APT does not change since the same catalytic groups
have been used). Finally the real enzyme has the extra feature
that the activated complex is bound to the enzyme, so the enthalpy
of binding is subtracted as well. As we mentioned earlier, for
chymotrypsin with a particular substrate this enthalpy of binding
was —11 kcal/mole. Since each 1.4 kcal/mole corresponds to a
factor of 10 in the rate, the binding should result in a rate increase
8
of 10 , or 100,000,000.
66 ORGANIC REACTION MECHANISMS

Several questions may be raised about the preceding analysis.

First of all, the measured enthalpy of binding is that for the
substrate, while we are really interested in the enthalpy of bind-
ing of the activated complex. This may not be the same, and
11
there are even suggestions that it may be greater, but the mag-
nitude is probably comparable to that measured for the sub-
strate. Second, it might be objected that both the substrate and
the activated complex are bound, so the energy difference between
them is unaffected by binding. However, the enzyme is ordinarily
present in only catalytic amounts; therefore the true starting
materials for any measured reaction(when concentrations are such
that the enzyme is not largely complexed by either starting ma-
terials or products) are the free substrate and enzyme, not the

ES complex. The great virtue of transition-state theory is that it

focuses attention on the difference in energy between starting

materials and activated complex, and makes it clear that various
intermediates between the two are not relevant. In this sense the
fact that an enzyme-substrate complex forms is not strictly rele-

vant, since this is an intermediate. Only the argument above that

binding of the activated complex is probably similar makes enzyme-
substrate binding forces of interest.

11. Ref. 7a, p. 222.

 3
NUCLEOPHILIC ALIPHATIC
SUBSTITUTION

of all classes of reactions, substitution at saturated carbon

atoms by nucleophilic reagents has been studied most thoroughly.
In this chapter we shall first consider the evidence for the cur-
rently accepted reaction mechanisms; the last section of the chap-
ter is concerned with the relationship between structure and
reactivity.

3-1 Reaction Mechanisms

S x 2 reactions. The reaction of methyl bromide with so-

dium hydroxide solution, to afford methanol, is first order in methyl
bromide and first order in hydroxide ion. Accordingly, both
which has
species are present in the transition state of the reaction,
a simple mechanism.
direct-displacement This mechanism is
described by Ingold by the symbol S N 2, meaning substitution
nucleophilic bimolecular. It should be noted that the 2 stands for
bimolecular, not for second order; the order of a reaction is a real

67
68 ORGANIC REACTION MECHANISMS

experimental quantity, while molecularity refers to the number of

species involved (undergoing a change in covalency) in the
rate-determining step of a particular mechanism being considered.

HO
©^ CH —A 3 Br >CH 3 — OH + Br
e

Rate = k[CH 3 Br][OH"]

It is found experimentally that second-order kinetics, implying an

S x 2 mechanism, are found for a variety of displacement reactions.
When the carbon atom on which displacement occurs is asym-
metrically substituted it is possible to investigate the stereochem-
istry of the reaction, and in all cases examined it is found that
displacement occurs with inversion, so the nucleophile begins
bonding to the back of the carbon atom while the leaving group is
departing from the front. Accordingly the process can be visual-
ized as is shown below.

R2 R2 R2
Ri Ri

VC-X Ri
i :

Y > Y
V C— /
iX > Y-r
\/
+ X

R3
/ I

R3
\ R3

Starting materials Transition state Products

An S N 2 displacement

This inversion of configuration during S N 2 reactions is not due

simply to electrostatic repulsion between an attacking hydroxide
ion and a leaving halide ion, for instance; even displacements in
which there should be electrostatic attraction between the nucleo-
phile and the leaving group proceed with inversion. A good
example is found in the reaction of D-a-phenethyltrimethyl-
ammonium ion with acetate ion by an S N 2 mechanism. In the
transition state there should be attraction between the attacking
acetate group, which still has some negative charge, and the
departing trimethylamine, which still has some positive charge.
Even so, the reaction proceeds with complete inversion.
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 69

y O C6 H 5 CH 3

CH,-c' + ^C-N-CH 3
>

O CH 3 CH 3

CH 3 -C
/ C6 H5
©/
CH 3
\O I

C N— CH 3 >

H
/ \CH \ CH 3 3

O
CH 3 -C^ C6 H5 CH 3

O— C;> ^ + :N— CH 3

\ CH 3
H \ CH 3

In molecular orbital terms the S N 2 displacement can be described

as follows. The nucleophile approaches the sp 3 orbital which is
being used by carbon to bond the displaceable group, and the
nucleophile begins to overlap with the smaller lobe of this orbital.
As bonding proceeds the carbon rehybridizes so that eventually
both the nucleophile and the leaving group are bonded to a carbon
p orbital, one to each lobe. The process continues as carbon goes
from this sp 2 hybrid state on to the product, in which it is again sp 3
hybridized. The fact that inversion is always observed simply
reflects the better bonding in this type of transition state. The
other possibility is not as good, overlap of both nucleophile and
electrophile with the same lobe of an sp 3 orbital in a front-
side displacement.

^CHi x 1 —— p roducts

sp 3

SN1 reactions. In the S N 1 mechanism (substitution nucleo-

philic unimolecular) substitution occurs in two steps. First the
leaving group ionizes, leaving a carbonium ion, and then the
carbonium ion reacts with the nuceophile. Carbonium ions are so
70 m ORGANIC REACTION MECHANISMS

unstable that they will react very rapidly with any nucleophile
(except for a few exceptional carbonium ions, such as tropylium
ion); general the first step, ionization, will be rate-
thus in
determining. Accordingly the kinetics should be first-order, the
rate depending on the concentration of substrate but not on that
of the nucleophile.
slow
R-X >R + X" +

fast
R + Y" >R-
Rate = k [RX]

Such kinetic behavior is often observed for nucleophilic substi-

tutions when an intermediate carbonium ion can be stabilized by
resonance or inductive effects.For instance, benzhydryl chloride
reacts with fluoride ion by an S N 1 mechanism in liquid S0 2
solution, for the rate of the reaction does not depend on the
concentration of fluoride ion (in the concentration range studied).

CI F
_ S0 2
—C H +
I I

C 6 H 6CH 6 5 F
solution
> C6H6 — CH— 6 H5 + CI

Rate = k [benzhydryl chloride]

CI N©
S0 2 e
— CH— —CH — C H
I

C6H5 6 H5 + I
C6H5 6 5 CI

CI © NEt 3

so 2 ©
C6H5 — CH — C H 6 5 + Et3 N > C6H5 — CH— C H
I

6 5 CI

The reaction rate does not depend on the concentration of the

nucleophile, nor does it depend on the nature of the nucleophile.
If benzhydryl chloride is reacted with pyridine or with
triethylamine instead of with fluoride ion the rate of reaction is the
same, although the products are of course different; again the
kinetics is found to be first-order, independent of nucleophile con-
centration.
 NUCLEOPHILIC ALIPHATIC SUBSTITUTION 71

The kinetic picture is not quite so simple as we have indicated

for these substitutions. As the reaction proceeds it is found that
the chloride ion which produced can also act as nucleophile, so it
is

begins to compete for the carbonium ion. Accordingly, the rate at

which substitution occurs can be decreased by the addition of
chloride ion, which traps some of the reactive intermediate and
returns it to starting material. This is an example of the common
ion effect, and it is characteristic of an S N 1 reaction; we can take it
into account by indicating a reversible ionization in our mecha-
nism.

CI , H

:
6 H5 — CH-C H 6 5 C6H5 — C-C H
I

+ CI
e
° > Product
4 6 5

k_i © Nucleophile

The fraction of intermediate ion which goes on to products

is simply the ratio of two rates, involving rate constants and
concentrations.
Fraction of carbonium ion converted to products =
k 2 ( nucleophile)
k_i(Cl") + k 2 ( nucleophile)

Thus the rate of substitution, taking the common ion effect into
consideration, is really not independent of the concentration of
nucleophile

ko( nucleophile)
Rate = kiCR-Cl)
k_!(Cl") + k 2 ( nucleophile)

Experimentally, first-order kinetic behavior can still be found by

examining the early part of the reaction before much common ion
has been formed, but the common ion effect can also be of use in
detecting an S N 1 mechanism. One example of this is discussed
below.
Solvolysis. A simple kinetic test to decide between the S x l
and the S N 2 mechanisms fails if the nucleophile is a solvent
molecule. The determination experimentally of whether or not
the nucleophile is involved in the rate-determining step requires
that we be able to vary the concentration of the nucleophile, which
72 u ORGANIC REACTION MECHANISMS

is not possible for the solvent. For instance, the hydrolysis of

f-butyl bromide in water is independent of the concentration of
hydroxide ion.

H 2

t - Bu - Br + OH > tBu - OH + Br ©
Rate = k [t - Bu - Br]

This shows that the mechanism does not involve an S x 2 displace-

ment by hydroxide ion, but it does not rule out S x 2 displace-
ment by water.

S x 2? + OH"
t - BuBr + H 2 > t - BuOH 2 + Br >t - BuOH + H 2
fast

It might be thought that this problem could be solved by going

to mixed solvents, so the concentration of the water could indeed
be varied. When aqueous acetone is the solvent it is found that
the rate of hydrolysis of r-butyl bromide increases if the solvent
is changed from 10 to 30% water, but the increase is 40-fold
for this 3 -fold change in concentration. This is mostly because
the reaction is faster in the more polar solvent, but it is not pos-
sible to decide whether the due to the
full 40-fold increase is

increased polarity of the medium. This polarity change might

contribute a factor of 13 to the rate, and the remaining factor
of 3 could be present because water appears in the kinetic expres-
sion as a nucleophile. Accordingly, it is not possible to use kinetics
in this way to determine whether the solvent is acting as a
nucleophile in an S N 2 process or whether instead an S x l mecha-
nism is involved.
In some cases the common ion effect can be used to solve this
problem. Thus when benzhydryl chloride is hydrolyzed in aque-
ous acetone the rate is unaffected by base, but this does not remove
the ambiguity we have just discussed, the choice between an S N 1
mechanism and an S x 2 displacement by water. However, it is
found that while the addition of many salts increases the rate,

since they make the medium more polar, the addition of lithium
chloride markedly slows the hydrolysis. The common ion
effect is unambiguous evidence that this hydrolysis has an S x l
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 73

mechanism. However, the common ion effect will only be seen if

the carbonium ion intermediate is stable enough that it can survive
in a sea of solvent molecules until it can be trapped by the common
ion. Unfortunately, the f -butyl cation is relatively reactive; there
is a small decrease in the rate of hydrolysis of /-butyl bromide
caused by added bromide ion, but the effect is so slight that it really
is not good evidence for an S x l mechanism.
Perhaps the most convincing evidence for an S x l mechanism in

a solvolysis reaction is the stereochemistry of the reaction. As we

have seen, S N 2 reactions proceed with inversion of configuration
at the carbon being substituted. A carbonium ion is flat, the
carbon being sp 2 hybridized and using the three hybrid orbitals for
single bonding; the remaining p orbital is empty. Accordingly,
once a carbonium ion is formed it is equally likely to be attacked on
either side (i.e., on either lobe of the p orbital); therefore, an S x l
mechanism should lead to racemization rather than preferential
inversion or retention of configuration.

R>
R
\ c-x -—
»
~J®\
H£-jC "Hi
R
^ Y
>
R

^C-Y + Y-C ^
R

R"
r"V V\R'
R"
R" R' R"

This test has been applied to solvolysis. When optically active

a-phenylethylchloride is submitted to S x 2 conditions, i.e., treat-
ment with sodium methoxide or sodium ethoxide,
the very reactive
one does observe second-order kinetics and clean inversion, as ex-
pected.

CI OR
ROH
— CH — CH C H — CH — CH
I |

C6H5 3 +NaOR > 6 5 3 with inversion

Rate = k[R— Cl][OR"]

S x 2 reaction with inversion
Under solvolysis conditions, e.g., in aqueous acetone, the
solvolysis indeed has an S N-1 mechanism.

<[' 80% OH
C 6 H 5 -CH -CH + 3 H2
^ 7<Te
e
)
C6H5 -CH -CH 3

98% Racemized

S x l reaction with racemization

 74 u ORGANIC REACTION MECHANISMS

However, the "real life" result is almost always more complicated

than simple theories predict, and it is important to note that
racemization is not complete but accompanied by 2% of net
is

inversion. Such partial inversion can be more extensive in other

systems. Thus solvolysis of a-phenylethyl chloride in acetic acid
yields the acetate which is only 85% racemized; the remaining
15% has an inverted configuration, showing that the acetic acid
has added from the side opposite to the leaving group.

CI O O CCH 3

C6 H5
I

-CH -CH + 3
II

CH 3COH-^ C 6 H 5 — CH — CH + HO
I

85% Racemized

15% Inverted

Ion pairs in S N 1 reactions. Such observations, racemiza-

tion accompanied by some inversion, are very common. They
might be explained by the suggestion that there are two simulta-
neous independent mechanisms, S N 1 reaction and some S N 2 dis-
placement by solvent at the same time, but several facts make this

explanation unattractive. same racemization

For instance, the
with partial inversion is observed in the solvolysis of some tertiary
alkyl halides or esters, although other evidence (vide infra) shows
that such compounds are very unreactive in ordinary S N 2 dis-
placements because of steric hindrance at the carbon atom.

CH 3 CH 3 CH 3 CH 3
CH 3 OH I

— CH -C— CH
I I |

CH 3 -CH -CH -C — CH
2 2 3 - > CH 3 -CH -CH 2 2 -CH 3

OCH 3
40% Racemized

60% Inverted

Instead of invoking S x 2 displacement, inversion is explained by

considering that the leaving group is still near the carbonium ion
when reaction occurs. Thus it is suggested that the ionization of
 NUCLEOPHILIC ALIPHATIC SUBSTITUTION 75

an alkyl halide leads first to an ion pair, in which the halide ion
helps solvate the carbonium ion on one side. Solvation on the
other side by an ordinary solvent molecule is also expected. If this

solvated species collapses to product, there will be inversion,

while racemization can occur if the halide ion is first replaced by a
normal solvent species, leading to a symmetrically solvated ion.

R R R
\ H2 l

+
—
_ /
C— X >H 2 0:— C- X > HO— C. (Inversion)
I'-'/
R" R"
S \ R'
VR"
R'

R R R

H2 0:— C — *.OH
I.
2 > HO— / +
\
C— OH (Racemization)
R" R' R" R"

The solvated ions can also be represented in molecular orbital

terms if "solvation" is considered to involve weak overlap of
orbitals.

The reader will have noticed that when a single solvent

molecule is specifically written in coordination with one lobe of the
p orbital, the picture is almost identical with that previously drawn
for the S x 2 transition state. The difference is that in the Sx2
ca^ejhe_entejing and leaving groups are so strongly coordinated
with the central carbon atom that there is no chance for the
leaving grou p to be replaced by another nucleophile before col- j
lapse^to products; thus a symmetrical species cannot be formed,
and inversion occurs. This strong coordination also means that
the central carbon atom has no appreciable positive charge. In
the solvated carbonium ion, the intermediate has enough life-
time that the leaving group can be exchanged for another solvent
molecule, and the coordination is sufficiently weak that in
many respects the carbon can be considered to have a full positive
 —
76 ORGANIC REACTION MECHANISMS

charge. However, when the solvent molecules are specifically

considered in this picture of the S N 1 mechanism, it is apparent
that the S x l and S x 2 mechanisms are very closely related. The
mechanisms of real reactions may well be intermediate between
them, in some cases.
Reactions of allylic halides. Allylic halides react readily
by an S N 1 mechanism, since the intermediate carbonium ion is

resonance stabilized.

©
CH == CH--CH 2 -Cl —>RCH= CH- ©CH 2 *-> RCH- CH = CH 2

H2 0/ OH
1

R- -CHr= CH- -CH 2 OH + R -CH

1

— CH = CH 2

As would be expected from mechanism, the intermediate

this

carbonium ion can react and a mixture

at either positive center,
of allylic isomers is generally obtained. Of course the same car-
bonium ion could be formed by starting with the isomeric chlo-
ride, so it is possible to test this mechanism by examining the
composition of the product mixture obtained from each starting
material. When this is done the usual observation is made
real life results are more complicated than simple theories pre-
dict. Thus, in comparing the reactions of crotyl chloride,
CH CH=CH—CH —CI, with those of its allylic isomer a-methyl-
3 2

allyl chloride, CH — CHC1— CH=CH 2 one finds that on solvolysis

3 ,

in ethanol each leads to a mixture of products, but the composi-

tion of the mixture is not quite the same for each.

OEt

= CH — CH OEt +
f

CH 3CH 2 CH 3CH — CH =CH

CH 3 — CH=CH-CH 2 -CI-
E'OH
> 92%
~ -
8%
78°C
CI
EtOH
I

CH 3 — CH — CH =CH 2 > 82% 18%

78°C
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 77

This result can be explained by proposing that simultaneous S N 1

and S N 2 reactions with solvent are taking place. Alternatively,
as the chloride ionizes, solvation may take place first on the car-
bon atom from which the halogen is leaving; this solvated cation
could either collapse to a product of the same allylic structure, or
go on to a more symmetrically solvated allylic cation which can
lead to both isomers.
A very important observation on the solvolysis of allylic

halides adds strong support to the idea that a tight ion pair of
carbonium ion and leaving group is the first intermediate. Thus
when a,2-dimethylallyl chloride (I) is solvolyzed in acetic acid it

undergoes rearrangement to the allylic isomer (II) simultaneous

with acetolysis. This rearrangement is a first-order process,
independent of added chloride ion, so it does not involve some
sort of allylic displacement reaction. More important, when
radioactive chloride ion is added to the solution the rearranged
product has only a fraction of the expected radioactivity. This
shows that most of the chlorine atoms in the product are directly
derived from the chlorine of the starting material, and not from
free chloride ions. The only attractive explanation of these facts
is and that the tight ion pair
that the starting material ionizes
formed rearranges and collapses to the isomeric chloride before it
can react with external chloride ions.

CI -HOAc
CH 3 CH 3 j
XH 2

HOAc \ //
Cl-C — CH =CH
I !
+
2
->
C-CH ->

CH 3 CH,
Solvent
I

AcOH — q"

!
+
CH 3 CH 2 CH 3
C=CH >
V= C CH
/
CH 2 CI

CH
/ i

CH
/
3 !
3

Solvent II

This phenomenon, collapse of the initially formed ion pair which

is detected because the structure has rearranged, has been ob-
 78 ORGANIC REACTION MECHANISMS

served in the solvolysis of many allylic compounds; it is usually

called "internal return." In all these cases it can be shown that
the particular chloride ion (or other leaving group ion) formed
by dissociation is preferentially recaptured in the rearranged
product, supporting the idea that this particular ion is still closely
associated with the carbonium ion.
Although these allylic rearrangements were first-order processes
which did not involve any external nucleophile kinetically, it
might be wondered whether an S N 2-like reaction could occur with
rearrangement in allylic compounds. Such a process, called the
S N 2' reaction, has indeed been observed in some cases, although it
is not common. The first example was found in the reaction of
a-ethylallyl chloride with the sodium enolate of malonic ester.
The kinetics are strictly second order, but 23% of the product
was of rearranged structure, so an S N 2' reaction accompanies the
normal S N 2 process.

\:oEt
CH 3 — CH 2 -CH — CH=CH + 2 CH No®
i CO*
o
CH 3 — CH 2 — CH — CH=CH 2 £0 Et 2

CH 4- CH 3CH 2 — CH = CH — CH — CH 2

Et0 2 C C02 Et C0 Et2

77% 23%

Rate = k[R — CI] [enolate ion]

A study of the stereochemistry of the S N 2' reaction was per-

formed by treating the substituted cyclohexenyl dichlorobenzoate
(III) The kinetics were cleanly second-order,
with piperidine.
and the product was entirely the rearranged isomer (IV) result-
ing from S N 2' rather than S N 2 attack. Interestingly, it was
found that the entering group comes in entirely from the side of
the ring that holds the leaving group.
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 79

CHv

,M .

The chemistry of allylic compounds shows that the nucleophilic

substitution reactions of an alkyl halide are strongly affected by the
presence of an adjacent double bond. Special effects are found for
many other neighboring groups as well, and these will be consid-
ered in the next section.
Neighboring group participation. If optically active
2-bromopropionic acid is sodium
treated with dilute methanolic
methoxide, substitution of bromine by a methoxyl group takes
place in a reaction with first-order kinetics (i.e., with no depend-
ence on the concentration of methoxide ion). It is found that this
substitution has occurred with complete retention of configuration
at carbon. Thus one has an apparent S N 1 reaction in which there
is not the usual racemization with partial inversion, but instead
from that expected
retention of stereochemistry, a result different
for either the S N 1 or the S N 2 mechanisms.
Such results are best
explained by invoking a double displacement mechanism. First
there is attack on carbon by the neighboring carboxylate ion to
form an a-lactone, and then displacement occurs by the methoxide
ion to form the final products. If both reactions occur with

inversion the result will be over-all retention of configuration, as

observed.

> A%
CH3-CH— C =
r\
n> Rr
CH3-CH
r

L e OCH
C

3
on
O
>CH 3 -CH-C
1

OCH3

The kinetics show that the first step is rate-determining, methoxide

ion playing a role only after the transition state and thus not appear-
ing in the rate expression. The a-lactone is highly strained, and is

more reactive than an ordinary ester would be.

80 ORGANIC REACTION MECHANISMS

Groups other than carboxylate ion can participate in neighbor-

ing group reactions. A well-known example is the reaction of
ethylene chlorohydrin with sodium ethoxide, to form ethylene
oxide.

HO O \ O
I

CH 2 — CH 2 + NaOEt > CH 2 — CH 2
I 1
— -> CH2
/ \ CH + CI
2

CI ^ CI

In this case internal displacement leads to a stable product

(although with more vigorous treatment the oxide may be opened
by a second displacement). Not only does the reaction proceed
only with internal displacement, rather than with S N 2 substitution
by ethoxide, but the reaction of ethylene chlorohydrin with sodium
ethoxide is 5100 times faster than a comparable S x 2 reaction,
displacement on ethyl chloride by sodium ethoxide. This large
preference for intramolecular reaction, even at the expense of
forming a strained due to probability factors. In collision-
ring, is
theory terms, the nucleophile is permanently held next to the

carbon it must attack, so reaction can occur whenever the species

picks up enough energy. In transition-state terms, an ordinary S N 2
reaction involves a loss of entropy when the nucleophile and
substrate are tied down in the transition state. In an internal
reaction there is no need to tie down a second molecule, so the
entropy of activation is much more favorable. This can even
make up for the unfavorable enthalpy associated with making a
strained ring.
An illustration of such factors is found in the internal displace-
ment reactions of aminoalkyl halides.

CH 2 — Br CH 2
(CH 2 ) n NH 2 > (CH 2 L NH 2 Br

As Table 3-1 shows, formation of the strainless five-membered

ring is fastest. The also strainless six-membered ring is formed
more slowly, since on the average the two reacting groups are
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 81

TABLE 3-1
Rates of Cyclization of Aminoalkyl Bromides in H 2 0, 25° C
rate — k (substrate)
Substrate k (per second)
Br— CH — CH — NH 2 6 X 10 4

Br— CH — CH — CH — NH
2 2
0.08 X 10 4

Br—CH —CH — CH— CH— NH

2 2 2
5000 X 10 4

Br— CH — CH,—CH — CH,— CH— NH 2 80 X 10 4

Br— CHo— CH.,— CH,-CH.,— CH— CH„— NH, 0.1 X 10 4

further apart and have less probability of reacting. Thus as the

ring to be formed is made larger the advantage of internal reac-
tion becomes less, since it becomes less probable that the two ends
of the chain will collide. Both the three-membered and four-
membered rings are strained and formed more slowly, but proba-
bility factors make the three-membered ring formation faster

than the four even though the smaller ring is more strained.
Mustard gas, CI— CH 2 CH 2 — —S—
CH 2 CH 2 CI, is a very — —
reactive alkylating agent toward all nucleophiles, and this property
is responsible for its vesicant action on the skin, since it alkylates

proteins. The rate of hydrolysis is independent of hydroxide ion,

supporting a two-step process, and the hydrolysis is slowed by

chloride ion. These data indicate participation of the neighbor-

ing sulfur.

CICH2CH 2 — S-CH
"
2 7_ CI— CH 2CH 2 — sC I
CI
' CH 2
I

CH 2 — CI H2

CI— CH 2 — CH SCH CH OH
2 2 2

Further evidence for the formation of a cyclic sulfonium ion in a

related case is the observation that both V and VI are transformed

to the same chloride VII on treatment with HC1. This is expected

if the isomeric starting alcohols are transformed to a common in-
termediate, the cyclic sulfonium ion.
82 ORGANIC REACTION MECHANISMS

CH 3CH 2 SCH— CH 2OH

1

CH 3
\x HCI CH 3
©<>CH * ©
V CH 3CH 2 -SCI
X CI
CH 2
HCI

CH 3 CH 2 SCH 2CH —OH

CH 3CH 2 — S— CH 2CH — CH 3

CH 3
VI VII

Evidence of the types citedis available now for neighboring group

participation by alkoxy groups, ester groups, halogen atoms,

phenyl groups, and a variety of others. Some of this evidence
will be further discussed in Special Topic 3.

3-2 Reactivity in Nucleophilic Substitution

S N 2 reactions

The Structure of the Alky I Group. The relative reactivity

of methyl chloride versus ethyl chloride in an S x 2 displacement
depends on the nature of the nucleophile, solvent, and reaction
conditions. However, Streitwieser has analyzed the mass of data
available in the literature and has come up with a set of "average"
relative reactivities for various alkyl derivatives in S N 2 reactions.
These are listed in Table 3-2.

TABLE 3- 2
A verage Relative Rates c >/ Alkyl Systems in S N 2 Reactions
Alkyl Group Relative Rate
Methyl 30
Ethyl 1
'*
n-Propyl 0.4
rc-Butyl 0.4
Isopropyl 0.025
Isobutyl 0.03
r-Butyl Nil
Neopentyl 0.00001
Allyl 40
Benzyl 120
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 83

Steric hindrance clearly plays a major role, so methyl com-

pounds are the most reactive of the simple alkyl derivatives.
The largest steric effect comes when alkyl groups are substituted
directly at the carbon to be attacked. Thus there is a large effect
on going from methyl to ethyl to isopropyl to J-butyl, but only a
minor effect on going from ethyl to ^-propyl. One exception to
this generalization is found when the carbon next to the reaction

site is highly branched, for then it can severely hinder approach to

its neighbor; this effect is seen in the very low reactivity of

neopentyl compounds.
CH;

\ CH 2
„ c

CHa'y \
CH 3

Steric shielding in a neopentyl derivative.

Electronic effects also play a role, as the high reactivity of allyl

and benzyl derivatives shows. At the transition state for the S N 2

2
reaction the carbon being substituted is sp hybridized, and uses its
other p orbital to bond the entering and leaving groups. In the
allyl and benzyl compounds this p orbital can also be conjugated

with the rest of the -k electron system, explaining the greater sta-
bility of these transition states and thus the higher reaction rates.

Transition state for an S 2 substitution in an allyl compound.

High reactivity due to electronic is also found in a-haloke-

effects
tones; for instance, chloroacetone 33,000 times as reactive as is
is

^-propyl chloride towards potassium iodide (in acetone at 50°).

84 1 ORGAN IC REACTION MECHANISMS

CH 3 - — C CH 2 CH 3 CH 2 CH 2

II fca
o
Relative 1 33,000 1.0

is lowered by some
In this case the energy of the transition state
bonding between the nucleophile and the carbonyl group.
Very interesting steric effects are found in some cyclic com-
pounds. In Table 3-3 are listed the rate constants for S N 2
reaction of iodide ion with some cycloalkyl bromides.

TABLE 3-3

SN2 Reactivities of Some Cycloalkyl Bromides toward 1~ in Acetone

at 70°C
Alkyl Group k (liters/mole second) X 10 7
Cyclopropyl < 0.01
Cyclobutyl 0.98
Cyclopentyl 208

It is apparent that the three- and four-membered ring compounds

are much less reactive than the cyclopentyl. Since the transition
state for an S N 2 reaction has sp 2 hybridized carbon while the
3
starting material has sp hybridized carbon, the bond angles at
carbon should go from 109°28 / in the starting material to 120°
in the transition state. In cyclopropyl bromide the bond angle is
only 60°. This is 49°28' smaller than sp 3 hybridization re-
quires, so the molecule has considerable strain (called I-strain,
or "internal" strain); in the transition state the strain would be
2
even worse, the angle being 60° less than sp hybridization re-
quires. Accordingly I-strain raises the energy of the transition
statemore than that of the ground state. The activation energy
thus becomes larger, and therefore the rate is slower. The same
effect, to a lesser degree, is found for cyclobutyl bromide. This
type of strain effect is only important for the smaller rings; in five,

six, and larger sized rings more subtle effects due to steric hin-
drance by neighboring hydrogens can be detected.
The Nature of the Nucleophile. The relative reactivities
of different nucleophiles will depend on substrate, conditions, etc.
Again a table of "average" nucleophilicities has been assembled by
Streitwieser, and selected values are shown in Table 3-4.
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 85

At first one might have expected that this order would reflect the
basicity of the nucleophiles. In the transition state the nucleophile
is beginning to form a bond to carbon; the stronger that bond is the
lower the energy of the transition state should be, and thus the
faster the reaction (although basicity measures the tendency to
bond to and carbon "basicity" need not parallel
hydrogen,
hydrogen basicity) However, it is clear from a glance at the table
.

that such basic species as phenoxide or acetate ions are less

nucleophilic than a nonbasic ion such as I". Several factors are
involved in determining nucleophilicity. First of all, when the
attacking atom is the same, e.g., oxygen, basicity does play a role,

TABLE 3-4
Relative Nucleophilicities

Nucleophile Relative Rate

QH 6
S- 470,000
I- 3,700-
EtO" 1,000
Br 500
CH 6 5
0" 400
ci- 80
CH COO- ;!
20
Pyridine 20
N0 3
-
1

so one finds the order: ethoxide > phenoxide > acetate > nitrate.
Secondly, larger atoms are more nucleophilic than smaller ones.
This shows up in the enormous reactivity of thiophenoxide ion
compared to phenoxide, even though the latter is more basic.
The relative reactivities of the halide ions also reflect this factor.
The size effect has been partly attributed to polarizability: the
electrons of a large atom are less firmly held by the nucleus, so they
may more easily move in response to some demand. In the
transition state for S N 2 reaction the nucleophile is still at a large
distance from carbon; the transition state will be stabilized if the
electrons of a polarizable atom can shift toward the carbon so as
to allow effective bonding even at this great distance.

Attack by a polarizable nucleophile, X.

86 ORGANIC REACTION MECHANISMS

Larger atoms also tend to hydrogen bond poorly with solvent

hydroxy Is, so fewer bonds to the solvent need be broken during the
substitution reaction. This factor means that relative nucleo-
philicities may be very different in nonhydroxylic solvents, and it

has been observed that I" is less nucleophilic than Br" in acetone
solution. Currently, active research programs are aimed at sort-
ing out the way in which the factors of basicity, polarizability
solvation, and perhaps other effects all combine in determining the
reactivity of a particular nucleophile.
The Nature of the Leaving Group. Since the leaving group
is breaking bond to carbon we would expect that good leaving
its

groups would be weak bases. Strong bases such as OH" are never
the leaving groups in displacement reactions (although H 2 may
be a leaving group if a hydroxyl is protonated by acid before
displacement). A set of relative average reactivities, compiled
by Streitwieser, is shown in Table 3-5.

TABLE 3-5
Relative Displacement Rates for Leaving Groups
Leaving Group kx/kBr (average)
OSO„C 6 H 5
6
I 3
Br 1.0
OH + 2 1

S(CH :! )2 0.5
CI 0.02
ON0 2 0.01
F 0.0001

There is considerable variation in these values, alkyl iodides

being from 1.2 to 36 times as reactive as alkyl bromides, depend-
ing on the exact reaction studied, but they furnish a good general
index to reactivity. Particularly striking is the low reactivity of
alkyl fluorides compared with the other halides, apparently because
the C—F bond is quite strong; the order of reactivity I >
Br > CI
> F parallels the carbon-halogen bond strengths. The low reac-
tivity of shows that a poor nucleophile does not
alkyl nitrates
necessarily make a good leaving group.
The Nature of the Solvent. Since the S N 2 reaction must
involve ionic species as starting materials, products, or both, it

must generally be conducted in relatively polar media. The exact

 NUCLEOPHILIC ALIPHATIC SUBSTITUTION 87

effect of the polarity of the solvent on the reaction rate depends on

the charges on starting materials and transition states, however.
Thus if the reaction involves a neutral nucleophile and a neutral
substrate, such as the displacement on methyl iodide by trimethyl-
amine, the transition state will be more ionic than the starting
materials; consequently the reaction will be faster in more polar
solvents.

CH 3v CH,
\n:+
-— CH3-I CH 3 —\6© N CH 3 —5©
>(CH 3 4 N
© ©
ch 3 >
I
) I

CH 3 CH 3
An S .2 reaction favored by polar solvents

On the other hand, displacement on trimethylsulfonium ion by

ethoxide ion is favored by less polar solvents; since the starting

material is more ionic than the transition state, polar solvents

lower its energy more.

© ®/CH 3 50 5©/CH 3
CH 3 CH 2 + CH 3 — S^ > CH3CH0O-— CH 3 -— SCX >

CH 3 CH 3
CH3CH2OCH3 + CH3SCH3

An S\2 reaction favored by less-polar solvents

Displacement on a neutral molecule R—X by an anion Y~ in-

volves no net change in charge, so the effect of solvent polarity is

small. However, the transition state has charge dispersed over

several atoms while the starting materials have charge localized.
Polar solvents stabilize localized charges more effectively, so in
many cases such displacement reactions are somewhat slower in
more polar solvents.

SN1 reactions
The Structure of the Alkyl Group. S N 1 reactions only occur
when the intermediate carbonium ion is stabilized. This generally
means that the positive charge is distributed over several atoms
rather than being concentrated on one. The distribution of charge
can occur (1) by resonance, as in the allyl cation; (2) by an
88 ORGANIC REACTION MECHANISMS

inductive effect in which single bond electrons shift toward the

positive carbon, as in 7-butyl cation (hyperconjugation is often
invoked for this case as well, but the relative importance of this
effect and of the inductive effect is still a matter of controversy);
and (3) by electron sharing from a neighboring group, as in the
cation derived from mustard gas.
The rate of ionization to form a carbonium ion depends on the
relative energies of starting material and transition state, of course,
so the stability of the product carbonium ion is really not relevant.
However, it seems likely that the transition state will closely
resemble the carbonium ion. The transition state has a structure
intermediate between that of starting material and that of product;
it furthermore is the highest energy point along the reaction
coordinate. As Hammond has pointed out, it thus seems reason-
able that the very unstable transition state should look very much

Transition state

R-X

Rxn..

FIGURE 3-1
Energy diagram for the formation of a carbonium ion, in which the
transition state strongly resembles the ion.

like the unstable carbonium ion, rather than the stable starting
material, so the energy diagram for ionization is that shown in
Figure 3-1. For this reason it is common to discuss reaction rates
as if the carbonium ion were the transition state, although this is
not strictly true.

Some indication of the importance of structural factors can be

obtained from relative solvolysis rates. In Table 3-6 are listed
the relative rates of solvolysis of a series of alkyl bromides in
water at 50°C. The precise magnitude of the differences between
them depends to some extent on the nature of the solvent (vide
infra).
Such a table underestimates the effect of methyl groups in
facilitating SN 1 reactions, as we go to the more substituted
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 89

carbonium ions. The solvolysis of the primary bromides prob-

ably goes by an S N 2 displacement by the solvent; consequently the
S N 1 rate for methyl bromide is much smaller than that shown in
the table.
Since benzyl chloride solvolyzes about as readily as does isopro-
pyl chloride, the conjugative effect of one phenyl group stabilizes

TABLE 3-6
Relative Solvolysis Rates of Alkyl Bromides, in H 0,
2 50°C
Compound Relative Rate
Methyl bromide 1.05
Ethyl bromide 1.00
Isopropyl bromide 11.6
/-Butyl bromide 1,200,000

the transition state about as much as the inductive (or hyperconju-

gative) effect of two methyls. Of course, as successive phenyls
are added the rates increase.

C 6H 5 CH 2 —CI (C 6 H 5 ) 2 CH—CI (C 6 H 5 ) 3 C—CI

1.0 2000 30,000,000
Relative solvolysis rates in 40% ethanol, 60% ether solution

If a benzyl cation is substituted with electron-donating groups its

stability may increase. For instance, p-methoxybenzyl chloride

solvolyzes at 10,000 times the rate of benzyl chloride in 67%
aqueous acetone, but ra-methoxybenzyl chloride has only 2A the
rate of benzyl chloride. This is because the positive charge can
be stabilized by the p-methoxy group but not the ra-methoxyl.
The latter is destabilizing, probably because of a field effect from
the carbon-oxygen dipole.

CH 2

Since the carbon atom of a carbonium ion should be sp 2

hybridized (electrons will enter orbitals with as much s character as
90 ORGANIC REACTION MECHANISMS

possible because the s orbital is of lower energy) factors which

affect the geometry at that carbon can affect S N 1 reactivity. Thus

both cyclopropyl and cyclobutyl halides solvolyze slowly, because
of the I-strain effect which was encountered in the S N 2 reaction.

CH 3

— CI CH 3
\
CH 3 C CI

CH
/
Very slow 0.62 8.9

Relative rates of solvolysis in 80% EtOH

Furthermore, halogens at "bridgeheads" of bridged ring systems

ionize very slowly. In a compound such as 1-bromobi-
cycloheptane the resulting carbonium ion cannot flatten, so
the carbonium ion is quite unstable. This bromide is also inert in
the S N 2 reaction; backside displacement is not possible.

The Effect of Solvent. Ionization of a neutral compound to

an ion pair is strongly favored by polar solvents. The impor-
tance of solvent polarity seems to depend on the exact S N 1 reaction
being considered, but the effect of solvent on the solvolysis rate of
/-butyl chloride gives an idea of the magnitudes involved.
Solvolysis of /-butyl chloride in water at 25 °C 300,000 times is

faster than solvolysis in ethanol, while mixtures of the two sol-

vents give intermediate rates. In formic acid the rate is 4% of
that in water, while in acetone (containing traces of water)
hydrolysis is even slower than the reaction in ethanol. The
dielectric constant of the solvent is important in determing this

reactivity order, but it also depends on the ability of the solvent to

interact specifically with the two ions formed. For instance, the
anion can be stabilized by hydrogen bonding with a solvent
hydroxyl group, while the carbonium ion can be stabilized by
specific interaction with an electron pair of a solvent molecule
coordinated with it.
NUCLEOPHILIC ALIPHATIC SUBSTITUTION 91

R— CI —H 2
=-> H2
® ©
R-— OH + Cl~— H—
2
/

Major accelerating effects are also observed when the medium

contains a species such as silver ion which can coordinate strongly
with a leaving halide ion.
The Nature of the Nucleophile. Since in the S N 1 reaction
the nucleophile attacks after the rate-determining step, it cannot
affect the rate of reaction. However, if the intermediate car-
bonium ion has a choice of several nucleophiles, then the nature of
the product is determined by their relative nucleophilicities. The
same general order of found here as was discussed for
reactivity is

the S N 2 reaction. Furthermore, many carbonium ions have the

possibility of eliminating a proton to form an olefin, and the
nucleophilicity of the medium will determine the relative amounts
of elimination and substitution. This point will be discussed
further in Chapter 4.

General References

A. Streitwieser, Jr., Solvolytic Displacement Reactions (McGraw-Hill

Book Co., New York, 1962). See also Chemical Reviews, 56, 571
(1956). In spite of the title this is a good general discussion of
nucleophilic aliphatic substitution reactions.

J. Hine, Physical Organic Chemistry (2nd ed., McGraw-Hill Book Co.,

New York, 1962), Chapters 6 and 7. Mechanisms and rates of S x
reactions are extensively and critically treated.

C. A. Bunton, Nucleophilic Substitution at a Saturated Carbon Atom

(Elsevier Publishing Co., New York, 1963) . One of a series of short
monographs; contains many recent references.

E. Eliel, "Substitution at Saturated Carbon Atoms," in M. Newman,

ed., Steric Effects in Organic Chemistry (John Wiley & Sons, New
York, 1956). A survey of nucleophilic, electrophilic, and free radi-
cal mechanisms with emphasis on stereochemistry.

C. K. Ingold, Structure and Mechanism in Organic Chemistry (Cornell

University Press, Ithaca, New York, 1953), Chapter 7. A good
account of the early work in this field.
92 ORGANIC REACTION MECHANISMS

R. Dewolfe and W. Young, "Substitution and Rearrangement Reac-

tions of Allylic Compounds," Chemical Reviews, 56, 753 ( 1956) . A
good review, but a little out-of-date.

B. Capon, "Neighboring Group Participation," Quarterly Reviews,

18, 45 (1964) . Contains many recent references.

J. Bunnett, "Nucleophilic Reactivity," Annual Review of Physical

Chemistry, 14, 271 (1963). A
summary of all the factors which de-
termine nucleophilic reactivity, and the evidence on their importance.

A. Parker, "The Effects of Solvation on the Properties of Anions in

Dipolar Aprotic Solvents," Quarterly Reviews, 16, 163 (1962). In-
cludes material on relative nucleophilicities when hydrogen-bonding
effects are absent.

F. Jensen and B. Rickborn, Electrophilic Substitution of Organomer-

curials (McGraw-Hill, New York, 1968). An account of an area
we have neglected in this book for lack of space.
 3
Special Topic

CARBONIUM ION REARRANGEMENTS 1

in the preceding chapter we have furnished examples of neigh-

boring group participation by oxygen, nitrogen, sulfur, etc. There
is currently much interest incarbon as a neighboring participating
atom; this interest is associated with attempts to understand the
intimatemechanisms of carbonium ion rearrangements. We shall
examine the evidence that alkyl, phenyl, and vinyl groups can, by
migration, directly assist in the ionization step of an S N 1 reaction.
First, however, we must survey the types of carbonium ion rear-
rangements in which such neighboring group assistance can be pos-
tulated.
Carbonium ion intermediates, as in S N 1 substitutions (and also
El eliminations and electrophilic additions to double bonds, Chap-

1. This topic is reviewed by (a) Y. Pocker, "Wagner-Meerwein

and Pinacolic Rearrangements in Acyclic and Cyclic Systems,"
(b) J. Berson, "Carbonium Ion Rearrangements in Bridged Bicy-
clic Systems," (c) R. Breslow, "Rearrangements in Small Ring

Compounds," and (d) J. King and P. de Mayo, "Terpenoid Re-

arrangements," all in P. de Mayo, ed., Molecular Rearrangements
(Interscience Publishers, New York, 1963). Other chapters also
contain relevant material.

93
 94 ORGANIC REACTION MECHANISMS

ter 4), frequently undergo rearrangements. For example, in the

2
nitrous acid deamination of neopentylamine, a methyl group
migrates with its electron pair so as to form a rearranged more
stable carbonium ion. As we shall discuss further below, it is not
really clear that the primary carbonium ion is formed first, since
rearrangement may be simultaneous with loss of nitrogen.

CH 3 CH 3 CH 3
HN0 2 © ©
— C — CH + N
I

— NH
I I

CH 3 — C — CH 2 2 > CH 3 — C— CH2N2 CH 3 2 2

1 H3 CH 3 CH 3

CH 3 CH 3 CH 3 CH 3 CH 3

C-CH 2 > C— CH CH 2 3 + C = CH
/ /\ /
CH 3 CH 3 OH CH 3

Many examples of such alkyl migrations are known, another fami-

3
liar case being the conversion of pinacol to pinacolone.

CH CH 3 CH 3
CH —
IIC— CHC
3

H
+
—C
CH 3
I I

C— CH
3

II
OH OH
3 > CH 3
I

OH
©
3 >

CH 3 CH 3
©
CH 3 — I

C-CH CH 3 — C— C— CH 3
I

II
OH CH 3
3 >

II

O CH
I

The alkyl group which migrates may also be part of a ring, in

which case ring expansions and contractions result. 4

M. Freund and F. Lenze, "Ein Versuch zur Darstellung des

letzten unbekannten Amylalkohols," Chemische Berichte, 24, 2150
(1891).
Cf. (a) J. Hine, Physical Organic Chemistry (2nd ed., McGraw-
Hill Book Co., New
York, 1962), Chapter 14; (b) C. Ingold,
Structure and Mechanism in Organic Chemistry (Cornell Univer-
sity Press, Ithaca, New York, 1953) Chapter 9. ,

Cf. ref. and also C. Gutsche and D. Redmore, Carbocyclic

lc,
Ring Expansion Reactions (Academic Press, New York, 1968).
CARBONIUM ION REARRANGEMENTS 95

CH2OH

CH2CI

H 2o H2o
^ +
OH

Thus solvolysis of cyclopropylcarbinyl chloride affords a mixture

of the unrearranged alcohol and the cyclobutanol from ring ex-
pansion of the intermediate carbonium ion, while cyclobutyl chlo-
ride yields the same mixture, the cyclopropyl carbinol coming
from ring contraction. Some ring opening also occurs.
Many other cases of ring expansions and contractions were
discovered in the course of investigating the chemistry of ter-
5
penes. As an example, camphene hydrochloride (I) is equili-
brated with isobornyl chloride (II) on treatment with Lewis acids,
which can help remove chloride ion and promote carbonium ion
formation.

_
CH 3 ch 3 _ CH 3 _ CH 3

ch 3 ch 3 CH3 CH 3
ch 3
ch 3 CH3 CH3

Again the change is represented as if rearrangement follows

the formation of the carbonium ion, but again one should
consider the possibility that ionization and rearrangement are
6
concerted. Phenyl groups may also migrate to positive carbon, as
in the analog of the neopentyl rearrangement.

C6H3 C 6 H.
I
© ©
— C— CH — CH C«H
.

C6H5 2 C 2 5

C6H5 C6H5

5. Cf. refs. Id and lb.

6. Cf. D. Cram, "Intramolecular Rearrangements," in M. Newman,
ed., St eric Effects in Organic Chemistry (John Wiley & Sons, New
York, 1956), Chapter 5.
96 ORGANIC REACTION MECHANISMS

Carbonium ions can undergo ring closure with double bonds. A

7
simple example occurs in the hydrolysis of 5-chloro-2-methyl-
2-pentene (V), affording almost entirely cyclopropyldimethyl
carbinol. Other examples are the acetolysis of ejto-norbornenyl
"brosylate" (p-bromobenzenesulfonate) to afford chiefly the
8
cyclized product (VI), and the methanolysis of cholesteryl chlo-
ride (VII) to yield i-cholesteryl methyl ether. 9

CH,
H2
CHCH2CH2CI

CH 3

CH3OH

VII OCH3

Among other types of carbonium ion rearrangements, there are

10
many examples of hydride shifts to neighboring carbonium ions.
A rather unusual one
11
occurs in the reaction of cyclooctene oxide

7. Ref. lc,p. 260.

8. Ref. 3a, p. 323; ref. lb, p. 192.
9. N. Wendler, "Rearrangements in Steroids," in P. de Mayo, ed.,
Molecular Rearrangements (Interscience Publishers, New York,
1964), Chapter 16, p. 1075.
10. Ref. 3a, p. 330.
11. A. Cope, S. Fenton, and C. Spencer, "Molecular Rearrangement
of Cyclooctene Oxide on Solvolysis," Journal of the American
Chemical Society, 74, 5884 (1952).
CARBONIUM ION REARRANGEMENTS 97

with formic acid, followed by hydrolysis. A mixture is obtained of

the expected rra«5-l,2-cyclooctanediol and of c/s-l,4-cy-
clooctanediol. The latter compound evidently arises from a
transannular shift of hydride (hydrogen with its electron pair) to
the carbonium ion first formed.

hco 2 h H2O
diol

OH H OH

I Hydri,de shift

*S-dio,

OH O OH
I

0= c
I

Although the distance seems large in this drawing, models show

that the migrating hydrogen can actually be quite close to the
positive carbon.

Rearrangements are also common in which not a carbonium ion

but a related species is involved. Thus the Beckmann rearrange-
ment 12,13 of oximes involves migration to a positive nitrogen,
rather than carbon.

12. L. G. Donaruma and W. Heldt, "The Beckmann Rearrangement,"

in A. C. Cope, ed., Organic Reactions, Vol. 11 (John Wiley &
Sons, New York, 1960), p. 1.
13. Cf. P. Smith, "Rearrangements Involving Migration to an Electron-
deficient Nitrogen or Oxygen," in P. de Mayo, ed., Molecular
Rearrangements (Interscience Publishers, New York, 1963),
Chapter 8.
98 ORGANIC REACTION MECHANISMS

R R' R R'
\C / HA (or Lewis acid)
> /
\ /
C >

5+
N ^* N
\ OH \*+
OH 2

R' R' R'

I

+ I

C=0
I

C C H2
< > >
II
+7
III I

N N NH
I I I

R R R

Migration to positive oxygen is also known, as in the rearrange-

ment of /ra/w-9-decalyl perbenzoate. 13 The benzoate ion remains
intimately associated with the cation at all times; the externally
added benzoate anion fails to compete — this failure to incorporate
external anions is the usual evidence for internal return in an ion
pair —and interestingly the two oxygens of the benzoate ion do not
even equilibrate.

O 18
O 18
O 18

O' 8
e,11 Q
II II

,OC-C 6 H 5 ^ OC— 6 H5 ©OC — 6 H5 |l

OC-QH 5

For clarity the reaction is written as if the unrearranged ion

were first formed, although it seems very likely that rearrangement
accompanies ionization. Finally, rearrangements involving car-
benes and nitrenes are related to these processes. Examples are
14
the Wolff rearrangement of diazoketones (A); the similar
1415
Curtius degradation via acyl azides (B); and the Hofmann
degradation 1416 (C).

14. Ref. 13, p. 528.

15. P. Smith, "The Curtius Reaction," in R. Adams, ed., Organic
Reactions, Vol. 3 (John Wiky & Sons, New York, 1946), p. 337.
16. E. Wallis and J. Lane, "The Hofmann Reaction," in R. Adams,
ed., Organic Reactions, Vol. 3 (John Wiley & Sons, New York,
1946), p. 267.
 CARBONIUM ION REARRANGEMENTS 99

o o
II © ^© A II ••
— CH=:C=0
(A) R-C-CH— N=N > R— C-CH > R

H 0\2

RCH2CO2H

o „ o
Q ^® A ••
R— C-N— N = N
II II

(B) > R— C— N > R-N = C=0

(C)
OO
II

R-C — NH
OH"
>R— C— N
II ©
R— C—
O
II ..

Br Br
R— N=C=0
Anchimeric Assistance

We have written most of these rearrangements as if ionization

occurred first. However, in many cases it is known that the
migration actually assists ionization; both rates and stereochemis-
try support this idea. This process, neighboring group assistance
by a carbon atom, has been called 17 "anchimeric assistance."
Thus in the Beckmann rearrangement it is clear that an unrear-
ranged nitrogen cation is not formed first since either group should
then migrate with equal likelihood; the experimental result is that
1 '-
the group migrates which is trans to the leaving group.

N N
r\®
V- OH 2 R
/

If the nitrogen cation were strongly associated with the leaving

group, as a tight ion pair, there might still be selective stereo-

chemistry. However, this seems very unlikely because the reac-

tion occurs under conditions which are undoubtedly too mild to
generate an unstabilized nitrogen cation. With simultaneous mi-

17. S. Winstein, C. Lindegren, H. Marshall, and L. Ingraham, "Par-

ticipation in Solvolysis of Some Primary Benzenesulfonates,"
Journal of the American Chemical Society, 75, 147 ( 1953)
100 ORGANIC REACTION MECHANISMS

gration a much better cation is produced; the ionization occurs

with neighboring group participation by the phenyl.
As another example, /?,/?,/?-triphenylethyl chloride solvolyzes in
formic acid at a rate 60,000 times that of neopentyl chloride. 18
Phenyls withdraw electrons inductively, so they would destabilize
the already unstable primary cation or a transition state which
resembled the unrearranged cation, but if migration is simultaneous
with ionization the rate can be understood.

C6H5 C6 H C6H5
\e
—C— CH
I

C6 H 5 2 ,
C— CH C 2 6H5
I

C6H5 CI C6H 5 C6H5

Open versus Bridged Ions 19

In the two cases just considered it was apparent that the

transition state for ionization could not simply resemble the unrear-
ranged carbonium ion or nitrogen cation. Instead rearrangement
occurred simultaneously with ionization, and was suggested that
it

the actual product was directly the rearranged carbonium ion

OAc
IX'

18. Ref. 3b, p. 514; cf. S. Winstein, B. Morse, F. Grunwald, K.

Schreiber, and J. Corse," "Driving Forces in the Wagner-Meerwein
Rearrangement," Journal of the American Chemical Society, 74,
1113 (1952).
19. For a discussion of the history of this question with a collection
of key reprints see P. Bartlett, Nonclassical Ions (W. A. Benja-
min, Inc., New York, 1965).
CARBONIUM ION REARRANGEMENTS 101

(nitrogen cation). However, some even more unusual situations

are known. For instance, the acetolysis (solvolysis in acetic acid)
of optically active threo-2-phenyl-3-butyl toluenesulfonate (VIII)
affords the racemic threo acetate (IX and IX'). 6
The formation of only the threo acetate shows that the toluene-
sulfonate group is replaced with retention of configuration, as in
other examples of neighboring group participation. The forma-
tion of completely racemized product is expected from the ion X,
which has a plane of symmetry. An ion of this kind has been
called
20
a "phenonium" ion. Further evidence that X is the direct
product of ionization comes from the finding 21 that 2-p-anisyl-
3-butyl toluenesulfonate (XI) is 80 times as reactive as is the
phenyl compound. Thus the transition state for ionization at least
partially resembles the phenonium ion.

OCH 3 OCH

> etc.

CH 3 — CH— CH— CH 3

OTs
XI

Of course, anchimeric assistance by must compete

a phenyl ring
with simple nucleophilic displacement by solvent. Thus if the
solvent is very nucleophilic, phenyl participation may not compete
22
well, while in a poorly nucleophilic solvent such as trifluoroacetic
23
acid the path involving phenyl participation may be much faster.

20. D. Cram, "Phenonium Sulfonate Ion-pairs. .", Journal of the

. .

American Chemical Society, 74, 2129 (1952).

21. S. Winstein, M. Brown, K. Schreiber, and A. Schlesinger, "Neigh-
boring Carbon and Hydrogen. IX," Journal of the American
Chemical Society, 74, 1140 (1952).
22. For a discussion, with many references, of the case against phe-
nonium ions see H. C. Brown and C. Kim, "Structural Effects in
Solvolytic Reactions. III. "
—
Journal of the American Chemical
Society, 90, 2082 (1968).
23. J. Nordlander and W. Deadman, "Trifluoroacetolysis of 2-Phenyl-

ethyl p-Toluenesulfonate. Evidence for Phenonium Ion," Tetra-

hedron Letters, 4409 (1967).
102 ORGANIC REACTION MECHANISMS

The bridged species has even been trapped in a special case,

24
solvolysis of XII to afford XIII. Treatment of XIII with

+
H

CH3OH

CH 2 CH 2 Br CH 2 — CH 2 CH 2CH 2OCH 3
XII

acidic methanol converts it to the normal open-chain product.

Such phenonium ions are presumably part of the reaction path
in any carbonium ion migration of a phenyl group, but it is in-

teresting that in solvolysis of XI the bridged ion is more stable

than either open carbonium ion. This follows from the fast sol-
volysis of XI, which shows that bridging stabilizes the ion. How-
ever, strictly speaking it is not proved that a symmetrical ion is

OCH3 ©O— CH 3 OCH

©
CH 3 — CH — CH — CH 3 CH 3CH— CHCH 3 CH 3CH — CHCHj
Less stable More stable Less stable

the best; the lowest energy species could be an unsymmetrical ion,

with the anisyl group strongly bonded to one carbon and only weakly
bonded to the other.
When alkyl groups migrate in carbonium ion rearrangements,
an intermediate bridged ion can also be written, often called a
"nonclassical" carbonium ion. The structure of this intermediate
has been written with dotted bonds to carbon; it can also be repre-
sented in molecular orbital terms. The combination of a (hybrid)
atomic orbital from the methyl carbon with an atomic orbital from

24. R. Baird and S. Winstein, "Isolation and Behavior of spiro[2,5]

Octa-l,4-diene-3-one," Journal of the American Chemical Society,
79, 4238 (1957).
CARBONIUM ION REARRANGEMENTS 103

CH 3
CH 3 / _ \ CH 3
— c — c-—z^-c © / ® \
c— ^zzr-c-c-
© I

I I I I II

CH 3
/ \

/ \
/ \

i © \

two carbons leads to a new molecular orbital which can

the other
accommodate the two electrons (of the erstwhile carbon-methyl
bond). The lack of similar alkyl migrations in radicals or in
carbanions is explained by the fact that this m.o. has room for only
two electrons, not the three or four which would be present in a
25
similar intermediate in radical or anion rearrangements. Much
19,26
recent research and discussion has centered on the question of
whether such a bridged carbonium ion can be more stable than
either simple ion to which it is related.
The question is best illustrated further by examining a particu-
lar case, the solvolysis of 2,2,1-bicycloheptyl bromobenzenesulfo-
nate (XIV) in acetic acid at 25 °C. 27
Although the starting
material is optically active, migration of the two carbon bridge
causes racemization (since structure XVI halfway through the
migration has a plane of symmetry, and interconverts XV and
XV which are mirror images).

25. H. Zimmerman and A. Zweig, "Carbanion Rearrangements,"

Journal of the American Chemical Society, 83, 1196 (1961).
26. See, for instance, H. C. Brown, "The Norbornyl Cation — Classical
or Non-classical," Chemistry in Britain, 199 (1966).
27. Ref. lb, p. 123, and references therein; cf. also G. D. Sargent,
"Bridged, Non-classical Carbonium Ions," Quarterly Reviews, 20,
301 (1966).
104 m ORGANIC REACTION MECHANISMS

xvir

The observed result is that completely racemic acetate is pro-

duced in this solvolysis, i.e., an equal mixture of XVII and
XVIF. One explanation of this result would be the scheme
shown, in which a simple carbonium ion XV is first formed, this

then rearranges to the ion XV (its mirror image), and the

resulting racemic mixture of carbonium ions then reacts with
acetic acid exclusively on the unhindered bottom of the
ring to yield the racemic mixture of products. The other
possibility is that the bridged ion XVI is more stable than
the open ion XV, and that solvolysis leads directly to XVI.
Reaction of XVI with acetic acid would then produce the racemic
mixture of acetates, and the classical ions XV and XV would
never have been involved at all.

HOAc
,
> XVII + XVII'

2® 2®

Other evidence had long been interpreted to indicate that this

and that the solvolysis does not
latter possibility is the correct one,
involve classical ions. For instance, acetolysis of XIV (called the
CARBONIUM ION REARRANGEMENTS 105

exo isomer) 350 times as fast as acetolysis of XVIII, its epimer.

is

Both could lead to the same classical ion, XV, but only XIV can

-Classical ion XV

ionize with simultaneous participation of the migrating group at

the back of the carbon, leading directly to XVI, a bridged ion.
This direct conversion of XIV to the better cation was considered
to be the explanation of the large acceleration.
Although formation of the nonclassical ion may contribute to
the difference in solvolysis rates of XIV and XVIII, Brown has
produced effective arguments 26,28 that there must be major steric
effects as well. Recently Schleyer 29 has pointed out a particular
kind of steric effect which differs in the transition states for sol-
volysis of XIV and XVIII and which could account for at least
some of the observed rate difference. Accordingly, in spite of a
large amount of work on this system, the importance of the non-
classical ion XVI is still from clear.
far
On the other hand, it seems quite clear that solvolysis of cyclo-
propylcarbinyl chloride (XIX) leads to a "nonclassical" carbo-
nium ion. 30 Thus this compound is 40 times as reactive as is
2-methylallyl chloride (XX), although the latter compound can
lead to a highly stabilized allylic cation. The cyclopropylcarbinyl
cation, if it were a simple saturated primary cation, would be very
unstable; the high reactivity of the system can be explained if the
product of solvolysis is a bridged ion instead.

28. H. C. Brown and K. Takeuchi, "The Characteristics of a Highly

Stabilized, Classical Norbornyl Cation," Journal of the American
Chemical Society, 90, 2691 (1968).
29. P. Schleyer, "Torsional Effects in Polycyclic Systems," Journal of
the American Chemical Society, 89, 701 (1967).
30. Ref. lc, p. 259.
106 ORGANIC REACTION MECHANISMS

CH 2 —CI CH 2 OH HO CH 2
II

Carbonium CH
ion
+
CH 2
/
—CH2OH
48% 47% 5%

CH 2 — CI CH 2

CH
/ %CH CH 3
/ % CH
3 2 2

XX

Although single-bond electrons are participating in the inter-

mediate ion, they come from the bent bonds of a cyclopropane and
are not typical o- electrons. A study of structural and substituent
31

effects on reaction rates shows that the transition state involves

overlap of the carbonium ion p orbital with both adjacent C —
single bonds.

H H

"
6+c fc 1 i." r ~. 1
:

8+CH 2
X CH 2
8+,

CH,
V
-CH 2

Here the rearrangement leads to a new structure, so it might be

wondered whether one does not simply have ordinary anchimeric
assistance by a ring-expanding migration.

CH 2 — CI Directly

31. P. Schleyer and G. van Dine, "Substituent Effects on Cyclopropyl-

carbinyl Solvolysis Rates," Journal of the American Chemical So-
ciety, 88, 2321 (1966).
CARBONIUM ION REARRANGEMENTS 107

Part of the evidence against this is the fact that since almost half
the product is unrearranged alcohol, the product of ionization is

not a classical cyclobutyl cation.

A nonclassical ion is also formed in the solvolysis of cyclobutyl

chloride, whose rate is also abnormally fast. Although this ion is

apparently not identical with the cyclopropylcarbinyl cation, the

two ions rapidly interconvert: Precisely the same product mixture
isformed starting from either cyclopropylcarbinyl chloride or cy-
30
clobutyl chloride.
Ordinary double bonds can also participate directly in the
ionization step in certain compounds. One of the most striking

a- oso -^_y-cH 2 3

H2

IM
XX /

R- S0 2

XXI XXII

examples of this is found 32 in the solvolysis of 7-am/-norbornenyl

toluenesulfonate (XXI). The rate is 10,000,000 times that for
the syn isomer XXII, showing that the double bond participates in
the ionization step. When the product carbonium ion is trapped
by reaction with borohydride ion 33 a large amount of the strained
hydrocarbon XXIII is formed, together with norbornene (XXIV).
With water as the nucleophile only 7-<2A?//-norbornenol is formed. 32

32. Ref. lb, p. 196.

33. H. Brown and H. Bell, "The Reaction of 7-Norbornadienyl and
7-Dehydronorbornyl Derivatives with Borohydride under Solvo-
lytic Conditions —
Evidence for the Tricyclic Nature of the Corre-
sponding Cations," Journal of the American Chemical Society, 85,
2324 (1963).
108 ORGANIC REACTION MECHANISMS

Although we have written the above carbonium ion as "non-

33
it has been argued,
classical," and disputed, 34 that the system
involves simply anchimeric assistance by a double bond during
ionization.

^
OTs

©
XXV

Since the equilibrating "classical" ions (XXV) are cyclopro-

pylcarbinyl cations, which as we have noted above are them-
selves "nonclassical," the distinction between the two inter-
pretations is not a major one.
We have emphasized cases in which there is good evidence for
neighboring group participation by carbon, but it should be pointed
out that such situations are relatively rare. When ionization of
a compound leads directly to a stable classical cation, then
anchimeric assistance is not observed. Accordingly, nonclassical
ions are not of frequent occurrence in organic chemistry. How-
ever, the subtlety of some of the questions being asked continues
to stimulate research in this area.

34. S. Winstein, A. Lewin, and K. Pande, "The Non-Classical 7-Nor-

bornenyl Cation," Journal of the American Chemical Society, 85,
2324 (1963).
 4
ONIC ELIMINATION AND ADDITION
REACTIONS

olefins are generally synthesized by elimination reactions

from saturated compounds; conversely, the most characteristic re-
actions of olefins are additions to the double bond. Thus there
would seem to be a logical connection between these two classes of
reactions. The connection is actually much more fundamental: a
reaction and its reverse must occur over the same path, although
in opposite directions. Since there is only one lowest energy path
between A and B, it is traveled both for A -» B and for B -» A,
as shown in Figure 4-1. This very important rule is called the
It means that the elucida-
principle of microscopic reversibility.
tion of themechanism of an elimination reaction would simultane-
ously furnish the mechanism of the reverse, addition, reaction,
provided both occur under the same conditions. This last is an
important limitation. For instance, the addition of HBr to an
olefin, under acidic conditions, is not the reverse of the base-
catalyzed elimination of HBr to form the olefin; in the presence
of base a new pathway (vide infra) is made lower in energy.

109
110 ORGANIC REACTION MECHANISMS

FIGURE 4-1
The common path and common transition state for a forward and re-
same conditions.
verse reaction run under the

However, there are other reactions for which the principle is quite
useful; alcohols can be dehydrated with acid, and olefins can be
hydrated to alcohols with acid, so that under the proper conditions
an equilibrium is established. In this case the mechanism of the
hydration step must be exactly the reverse of that of the dehydra-
tion.

4-1 Elimination Reactions

In a ^-elimination reaction two groups are lost from neighboring

atoms, with the resultant formation of a double bond. For
convenience we shall symbolize such reactions as if a carbon-

C— >=<
I I

^-elimination

carbon double bond were formed, although analogous processes

may be written involving the formation of carbon-oxygen, carbon-
nitrogen, etc., double bonds. Most studies of mechanisms of
^-elimination reactions have concentrated on the formation of a
IONIC ELIMINATION AND ADDITION REACTIONS HI

carbon-carbon double bond by the elimination of HX, where X is

a good leaving group such as a halide ion. Accordingly we shall
discuss only such HX eliminations.
Other reactions are known which may be classified as
a-eliminations (two groups from the same carbon), and 1,3-
eliminations and even 1 ,4-eliminations are known. Of these,
only the a-eliminations, to form carbenes, will be briefly discussed.

CI CI
I OH" |
~
CI— C— H > CI— C: via CCI 3

CI
An a -elimination

Br Br
Zn CH 2
CH 2 — CH 2 — CH 2 >
/— \ + ZnBr 2
CH 2 CH 2
A 1,3-elimination

H Br
HO"
— CH=CH— CH
I

CH 2 2 > CH 2 =CH — CH=CH 2

A 1,4-elimination

^-Eliminations

The El Mechanism. If a carbonium ion has a ^-hydrogen,

this may be lost to yield an olefin. Thus in many of the S N 1
reactions discussed in Chapter 3 olefin formation is an important
side reaction. The elimination of HX
by ionization of X, and
subsequent loss of a proton from the resulting carbonium ion, has
been called the El mechanism (Elimination Unimolecular). It

H H
Slow Fast \ / ©
— C— X — c — C©—r^-> £=C
I I
| |

C >
+ BH
II B
I I
'
/ \
The El mechanism
112 ORGANIC REACTION MECHANISMS

would be expected that such a mechanism could occur only when

the intermediate carbonium ion is stable. Thus the same factors
—
which favor S N 1 reactions formation of a stabilized carbonium
ion in a highly polar medium — will operate here. The El
elimination and the S N 1 reaction often occur simultaneously;
substitution is favored by the presence of good nucleophiles, while
strong bases will tend to remove the proton from the carbonium
ion and favor elimination.
The evidence for an El mechanism is, first of all, kinetic.
Since the rate-determining step does not involve the base the
reaction rate will be independent of base concentration. How-
ever, most El eliminations accompany the solvolysis of /-alkyl
halides in the absence of added base; the solvent molecules must
act as proton acceptors, and they cannot be detected kinetically.
For this reason a more subtle mechanistic criterion is used. If a
reaction follows an El mechanism it involves formation of a
carbonium ion, and the behavior of this carbonium ion should be
independent of the nature of the leaving group. Thus in the
competition between S N 1 and El reactions a mixture of olefin and
alcohol (if the solvent is water) is formed, and the composition of
this mixture should be the same whether the starting material is

f -butyl chloride, ?-butyl bromide, ^-butyl iodide, etc., since the same

carbonium ion is formed from each one. In general this is found to

be the case for reactions of f-alkyl halides in quite polar media
(although in less polar solvents the leaving group may stay
associated with the carbonium ion and exert some influence on the
course of further reaction).

CH 3 CH 3 CH
^
2

EtOH El
CH 3 -C — X
I

>CH 3 -C © > CH 3 —
I
25° I I

CH 3 CH 3 CH 3
X = Br,a,l f SMe 2 1
Sn1
/O
CH 3
— C— OEt
I

CH 3
I

CH 3
81%
IONIC ELIMINATION AND ADDITION REACTIONS 113

Another criterion which may be applied is the composition of

the olefin mixture, in cases where a mixture is possible. Thus
solvolysis of £-amyl halides yields, in addition to substitution
products, a mixture of two olefins, 2-methylbutene-l and
2-methylbutene-2. The stable more substituted olefin is preferred
(so it is said that El elimination follows the Saytzeff rule), and the
observation is made that the proportion of the two olefins is

independent of the leaving group in reasonably polar media; again

this means that a common intermediate, the carbonium ion, is

involved in the eliminations.

CH 3 CH 3
EtOH 37%
CH 3 — C— CH
I

2 — CH 3 > CH 3 — C— CH,CH
I

I
25° © El
X S„l

63%
CH 3 CH,

CH 3
— C=CHCH
I

3 + CH 3 — CCH,CH
II

82% 18%

The E2 Mechanism. Most commonly, HX eliminations are

carried out in the presence of strong base. For instance, when
ethyl bromide is held at 60 °C in ethanol nothing happens, since
the ethyl cation is not stable enough for El elimination to occur.
However, in the presence of sodium ethoxide, a strong base, elimi-
nation of HBr occurs and ethylene is formed (together with con-

siderable diethyl ether from an S N 2 reaction). Since it is found

that the rate of olefin formation is proportional to both the ethyl
bromide and the ethoxide ion concentrations, a bimolecular mech-
anism is written, the E2 mechanism (Elimination Bimolecular).

e
OEt
H
|

CHr^CH 2 > CH 2 =< :h 2 + EtOH + Br

©
Cor "
Rate=k C 2 H 5 Br OEt
114 ORGANIC REACTION MECHANISMS

In this mechanism it is suggested that removal of the proton and

loss of thebromide ion occur simultaneously, but of course this is

not required by the observed kinetics. The three schemes shown
below would also show second-order kinetics, and must be ex-
cluded on other grounds. The first one is easily excluded since
other experience with carbonium ions suggests that if the ethyl

p qst © slow
CH.3— CH 2 — Br
,

CH.3— CH 2 Br —7^ CH 2 =CH 2

OEt

H fi
OEt, Slow
I

CH 2 — CH — 2
r< CH2 _CH — 2 Br > CH 2 = CH 2

Fast

H
OEt
u © Fast
I

CH 2 — CH 2 — Br CH 2 — CH 2 — Br > CH 2 =CH 2
Slow

cation were formed would react even in the absence of ethoxide

it

ion, while in fact ethyl bromide is stable in the absence of base.

The second is a more serious possibility, but it is excluded by the
finding that no deuterium is incorporated in an alkyl bromide which
is recovered from partial reaction in EtOD. If the proton is

reversibly removed it must be replaced by deuterons from the

solvent. The third mechanism is not excluded by this finding,
since no equilibrium is suggested. It can be ruled out by the

observation that base-catalyzed eliminations are run under condi-

tions much too mild to permit formation of an unstabilized free
carbanion at an appreciable rate. One may exclude any mecha-
nism containing a step which would be slower than the observed
over-all reaction rate. Simultaneous elimination appears to be the
only reasonable mechanism.
Nevertheless, one might wonder whether everything really
happens at precisely the same time. May not proton removal run
slightly ahead of halide ion loss, so that a little negative charge
builds up on carbon? Conversely, is it not possible that in some
cases ionization of the leaving group may run ahead of proton loss,
so that a little carbonium ion character is developed? The answer
to these questions is "Yes." Cases are known in which E2
eliminations cover the whole range, from "almost carbanion"
IONIC ELIMINATION AND ADDITION REACTIONS 115

processes to "almost carbonium ion" processes. Evidence for this

on the carbanion side comes most readily from a study of substitu-
ent effects, as in the ^-substituted /?-phenethyl derivatives. In
Table 4-1 are listed the second-order rate constants for some
substituted phenethyl bromides in an E2 elimination with sodium
ethoxide in ethanol at 30 °C. It can be seen that groups which

could stabilize a carbanion by conjugation, such as the p-nitro

group, strongly accelerate the reaction; groups which should de-
stabilize a carbanion, such as the p-methyl group, slow the elimina-
tion compared with the unsubstituted case.

CH — —» N =zr X=CH
e / \=/ Qq/

CH 3 ( )= CH worse than H-^< )= CH

Substituent effects of this type have been put on a quantitative

basis by Hammett. For each substituent, such as a p-methoxyl
group, a substituent constant a can be assigned. Thus crp M eo is .

— 0.268. Such a negative o- indicates that the p-methoxy group

TABLE 4-1
NaOEt
C6 H4 — CH — CH —
2 2 Br > p— R— C 6 H4 — CH=CH 2

EtOH

Rate = k 2 {alkyl bromide) (OEf)

p
_ k? X 10 5 (liters/mole second)

CH 3 o- 16

CH3- 23

H- 42

Cl- 191

CH3CO - 1720

N0 - 2 75,200
116 ORGANIC REACTION MECHANISMS

donates electrons to a benzene ring. The p-methyl group, a

weaker electron donor, has a a of only — 0.170, while the electron-
attracting p-cyano group has a positive <r, + 0.660, and p-nitro is
even more positive, + 0.778. Each reaction considered in this
treatment has a reaction constant p which measures the sensitivity
of that reaction to substituent effects. Then the effect of the
substituent on the rate constant is expressed with a simple equation.

LOg (Ksubstituted/kunsubstituted) = O"

p

This, the Hammett equation, is a very general expression of the

effect of aromatic ring substituents on side-chain reactivity. It can
be applied to equilibria as well as rates. Thus the effect of ring
substituents on the ionization constants of benzoic acids (the
standard reaction, for which p is defined as 1.0) or of phenols
(for which P is found to be 2.113, larger than 1.0 since the charge
in a phenol anion is closer to the substituent than it is in a benzoate
anion) can be correlated using the same set of o- constants for the
various substituents. Since rates and acidities are covered by this
equation, it bears some relation to the Br0nsted catalysis law dis-
cussed in Chapter 2, and like the Br0nsted law the Hammett equa-
tion describes a linear free energy relationship. Taft has defined
a similar equation governing substituent effects in aliphatic sys-
tems, and in Chapter 5 we will see that the Hammett equation
with slightly modified a constants can be applied to aromatic sub-
stitution processes. However, for our present purposes it is
enough to note that a constant P can be determined for a reaction
by varying substituents and observing the effect of this on the rate.
It is found that p for phenethyl elimination reactions depends on the

particular leaving group.

For instance, with sodium ethoxide in ethanol at 30 °C, p-
is 23 times as reactive as is unsubstituted
cyanophenylethyl iodide
phenylethyl iodide.

Log 23 = 1.36= (+ 0.660) P

therefore P= + 2.07

The rates with other substituents also give this same value of p
when I" is the leaving group. In Table 4-2 are listed some
relative rates and reaction constants for E2 eliminations of
/?-phenethyl compounds with various leaving groups; the re-
action conditions are again sodium ethoxide in ethanol at 30°C.
IONIC ELIMINATION AND ADDITION REACTIONS 777

The fact that p is positive for all these reactions means that they
all develop some carbanion character, but at the transition state

TABLE 4-2
E2 Elimination of R — C H,— CH—CH X
6 2

Relative rate
X (when R = H)
-I 26,600 + 2.07
-Br 4,100 + 2.14
— Toluenesulfonate 392 + 2.27
-CI 68 + 2.61
-F 1 + 3.12

NMe, — +3.77

this is more pronounced for the elimination of HF than for the

elimination of HI. This is revealed by the fact that the HF
elimination has a larger meaning that its transition state can be
p,

stabilized to a greater extent by groups which would stabilize a

carbanion. Considering the reaction rates as well, these data agree
with what one might expect for the different leaving groups.
Thus, as the ethoxide ion begins to remove a proton from a
phenethyl iodide some carbanion character develops, but iodide is

such a good leaving group that it quickly begins to depart and the
new double bond starts forming. Consequently only a moderate
anionic charge ever develops on the carbon; this means that the
substituent effects are limited in size. It also means that the
reaction goes readily, since a full carbanion of this type would be of
quite high energy and the transition state energy will be raised to
the extent that anionic character must be developed. When
fluoride ion is the leaving group more negative charge must
develop before this ion can be ejected; this makes the reaction
slower, and also more sensitive to substituent effects.
These data illustrate the fact that apparently very similar E2
eliminations can involve the development of varying amounts of
carbanion character. This effect is sometimes invoked to explain
the difference between "Saytzeff" and "Hofmann" orientation in
elimination reactions, although steric hindrance effects are also
involved to a major extent. E2 eliminations with neutral sub-
strates, such as alkyl halides, alkyl toluenesulfonates, etc., ordinar-
ily lead to the more substituted olefin. These cases are said to
 118 ORGANIC REACTION MECHANISMS

Br
NaOEt
CH 3 — CH.2— C — CH
I

3 > CH 3 — CH =C—CH 3 CH 3 - CH 2 — C = CH 2
25°
I

CH 3
I
+ |

CH 3 CH 3
72% 28%
follow the Saytzeff rule. On the other hand, E2 elimination in a
quaternary ammonium hydroxide (Hofmann elimination) occurs
with removal of the most acidic hydrogen. In simple cases this
leads to formation of the least substituted olefin, since primary
carbanions are more stable than secondary or tertiary carbanions
(alkyl groups are electron donating relative to hydrogen atoms).
This orientation is said to follow the Hofmann rule.

CH
CH 2
© OH~ r
— CH — N — CH CH CH -,
I

CH 3 2 2 2 3 > |_CH 2 =CH 2 + EtN(Pr) 2

J +
CH 2 96%
I
"
[CH 3 CH = CH + 2 PrNEt 2
]
CH 2
|
4%
CH 3 Hofmann elimination

The p values listed in Table 4-2 show that the transition state
for Hofmann elimination has a considerable amount of carbanion
character; given a choice the reaction which involves a better car-
banion will occur. The transition state for HBr elimination has less

carbanion character and more double-bond character, so in this

case the reaction which forms the better olefin (alkyl groups sta-
bilize olefins) will occur.
The stereochemistry of E2 eliminations supports the picture of
more or less concerted elimination. Whenever possible the two
groups eliminated assume a toms-coplanar position in the transi-

C :B

cJ
\
c
I

&: <n
Trans-coplanar Elimination
IONIC ELIMINATION AND ADDITION REACTIONS 119

H ^X
C— Z C
R"
/ \ R
R'" N R
,

C/'s-coplanar Elimination

tion state. The statement that ''trans" elimination occurs refers to

the positions of these groups, and not to the geometry of the
resulting olefin. A few cases of c/s-coplanar elimination have
been found in compounds for which the rra/75-coplanar geometry is

impossible.
ElcB Eliminations. In the cases so far discussed we have
seen examples in which the leaving group ionized first (El) or in
which loss of the proton and the leaving group were more or less

concerted (E2). The third possibility, ionization of the proton

before loss of the leaving group, has been called the ElcB
mechanism (elimination unimolecular conjugate base). It should
be emphasized that kinetically such a process would be second
order, since the formation of the carbanion would involve base as
well as substrate, but the rate-determining step, loss of the leaving
group, would be unimolecular.

H X X e
1 1
b:
x ©c-c- 1
-x
= c/
c-c— i
c
1 1
'
Fast / | Slow

The ElcB mechanism

Such a process can only occur if the carbanion is strongly

stabilized,and if ihe leaving group is sufficiently poor that it will
not be lost from the developing anion by E2 elimination. Several
cases are now known in which it can be demonstrated that
deuterium exchange, via the carbanion, is faster than elimination.
This is observed in the base-catalyzed elimination of DF from
(labeled) l,l-dichloro-2,2,2-trifluoroethane.

OH,© 6 CCI slow

CDCI2— CF 3 2 — CF 3
CCI 2 = CF + 2 F

H2 I
fast

CHCI2 — CF 3
120 ORGANIC REACTION MECHANISMS

However, the demonstration that a carbanion can be formed is

not the same as proof that it is an intermediate in the elimination.
Thermal Eliminations. In contrast to these ionic elimina-
tions, a number of processes are known in which elimination occurs
by a unimolecular process without the attack of external reagents.
Two synthetically useful examples are the formation of olefins by
pyrolysis of esters or, at lower temperatures, xanthates.
CH 3

p= o c

f o
V\ V
c^-c
500 C / CH3C0 2 H
/I l
x /

SCH 3

H TO
XN / 300°C
CH 3 S C SH -»- CH 3 SH + C
'\ l
x II

Concerted mechanisms are written for these processes, supported

by the fact that the reactions are carried out in the gas phase, where
solvation of intermediate ions would be impossible, and by the fact
that the reactions involve cis elimination. The measured entropy
of activation, AS*, is negative for these processes, as expected for a
cyclic transition state in which some of the initial freedom of
rotation has been lost.

<x-Eliminations
When chloroform is treated with strong base it loses HC1, form-
ing dichlorocarbene. The carbene is not stable, but it has been
trapped in various ways, e.g., by addition to olefins present during
the carbene-forming reaction.

e ©
HCCI3 + tBuO K :cci 2 + tBuOH + KG

CI

CI
IONIC ELIMINATION AND ADDITION REACTIONS 221

Carbene formation can also be detected in other ways when

chloroform is treated with aqueous base, and under these condi-

tions fast hydrogen exchange occurs with the solvent, detectable if
D2 is the solvent.This base-catalyzed exchange shows that the
trichloromethyl anion is formed reversibly; it is usually considered
to be evidence that the elimination involves the anion. (How-
ever, as was mentioned above, demonstration that a carbanion can

OH G .

HCCb ? © ccb — > :cci 2 + a

Fast Sl<

be formed is not really equivalent to demonstrating that it is an

intermediate in the elimination.) The alternative concerted
elimination mechanism has been found for the hydrolysis of
chlorodifluoromethane.

OH H^CFs— CI > H2O+ :CF 2 + CI

The change mechanism occurs because of two factors: the

in
trichloromethyl anion is more stable than the chlorodifluoromethyl

anion, since chlorine can help stabilize the charge by use of its 3d
orbitals, and difluorocarbene is more stable than dichlorocarbene

CL CI CL
\ v-
/ < >
\/ C
CI

< > etc.

I I

CI CI

since fluorine has 2p bonding with the vacant carbene

orbitals for tt

2p orbital, while chlorine must form the less stable 3p 2p tt bond. —

©
00
C < > F = ©C < > F— ©
F F F®
The addition of carbenes to olefins will be further discussed in
Special Topic 4.

4-2 Addition to Carbon-Carbon Double Bonds

Almost all olefins will react with electrophilic reagents such as

Br 2 while only compounds in which the double bond is activated
,

by a carbonyl group or similar anion-stabilizing function will re-

122 ORGANIC REACTION MECHANISMS

act with nucleophiles. Accordingly, electrophilic addition will be

treated first.

Acid-catalyzed hydration. As discussed before, the

mechanism of acid-catalyzed hydration of an olefin must be pre-
cisely the reverse of the acid-catalyzed dehydration of the alcohol
under the same conditions. This reversible process may be
pictured as follows.

®OH 2 H OH H
H2
\ / \l
c — c c — —c c — 1/
H2 / +H / \
The reverse will be recognized as an El elimination reaction,
involving formation of the carbonium ion.
If protonation of the olefin were rapidly reversible, deuterium
exchange should be observed between the olefin protons and
deuterium in the solvent. However, 2-methyl-2-butene which is
recovered after 50% hydration in D2 contains no deuterium.

CH 3 CH £ CH 3v D OD CH 3 CH 3 CH
D3
© a

CH 3 H CH 3 D CH a

This shows that the carbonium ion is not reversibly formed.

Thus the rate-determining step carbonium ion formation, the
is

latter rapidly going on to product alcohol. For other reasons there

has been a suggestion that carbonium ion formation might involve
two steps reversible addition of a proton to the olefin to form a
: -n-

complex, followed by slow collapse of this intermediate to the

Fast
Slow
H,0'
/ \ / \ \
ONIC ELIMINATION AND ADDITION REACTIONS 123

carbonium ion. The n complex could be pictured with overlap of

the hydrogen Is orbital and lobes of the carbon 2p orbitals.
Evidence for -n complexing in electrophilic aromatic substitution
will be discussed in Special Topic 5.
Addition of HX. Of course, other nucleophiles may attack
the carbonium ion as well, and the addition of HBr will have a
mechanism similar to that outlined above. The stereochemistry
of such additions is not yet clear. It is reported that addition of
HBr to 1,2-dimethylcyclohexene occurs predominantly trans, while

under other conditions the addition of DBr to acenaphthene has

been found to involve cis stereochemistry.

CH 3 CH 3

HBr Trans-addition

////
CHa CH 3

C/'s-addition
+ DBr

In nonpolar solvents HX will add to an olefin to form an ion pair

of carbonium ion and anion. If this pair collapses rapidly to
product, the result will be cis addition.
Addition of halogen. When bromine is added to ethylene
in polar media the product is ethylene dibromide, but in the
presence of chloride ion, of nitrate ion, or of other nucleophiles
mixed adducts are obtained. Furthermore, in cases where the

CH 2 =CH 2 + Br 2 -> BrCH 2 CH 2 Br

CI
CH 2 =CH 2 + Br 2 * BrCH 2 CH 2 CI

NQ 3

CH 2 =CH 2 + Br 2 BrCH 2 CH 2 ONO;

stereochemistry can be detected it is found that these additions

occur trans. Since trans addition occurs even in noncyclic olefins
124 ORGANIC REACTION MECHANISMS

Br H
CH 3
\ C == c + Br 2 — Q
1

meso-2, 3-dibromobutane

\CH 3
CHa^ 1

Br

Br
H

/
c == c
\CH
+ Br 2 — A 1

1
> CH3

dl-2, 3-dibromobutane

CH 3 3
CH 3 Br

,^c c

IT R"

the formation of a free carbonium ion intermediate is excluded, and

instead the participation of a bromonium ion is suggested. The
bromonium ion can then be attacked by any nucleophile in the
system, not merely by bromide ion, so the formation of mixed
adducts is explained. More important, the stereochemistry is thus
explained since displacement on the carbon of the bromonium ion
should occur with inversion, leading to over-all trans addition, and
since cis- and frajw-olefins would form different bromonium ions.
If a classical carbonium ion had been formed, rotation about the
carbon-carbon single bond would interconvert the ions derived
from cis- and trans-butQUQ.

H H Br H

CH»
/ \ CH h--
CH,
/ TT \
3 CH 3
III

CH
/ 3
c=c
\H — «--?^r\ CH3 H

Attack on the bromonium ion by a nucleophile could occur at

either carbon. Usually the nucleophile attacks the more substi-
tuted carbon, showing that the substitution has much of the
IONIC ELIMINATION AND ADDITION REACTIONS 125

character of an S x l reaction favoring the better carbonium ion.

5©
CH 3 Br CH ;

\5©/
C
\ CH 2 Br
CH 2
CH 3 CH 3 OH
OH 2

However, with /-butylethylene, attack at the secondary carbon is

so hindered that substitution occurs at the primary carbon.

CH 3 CH 3 Br
Br 2
CH 3 — C— CH=CH
I

2 CH 3 —C I I

CH — CH OCH 2 3

CH 3 OH
CH 3 CH 3

It must not be concluded that bromonium (chloronium,

iodonium) ions are always formed in halogen additions. If the
classical carbonium ion can be strongly stabilized it may be the
preferred intermediate. This is revealed in the c/s-addition of
chlorine to acenaphthene, in contrast to the usual trans additions.
In this case a classical carbonium ion is probably formed paired
with a chloride ion, and collapse of the ion pair occurs with
over- all cis addition.

Hydroboration. H. C. Brown has developed a variety of

synthetic procedures which start with the addition of diborane,
B 2 H 6 to olefins. The reagent acts as if it were a source of BH 3
,
,

a Lewis acid with six-electron boron which undergoes electrophilic

attack on a double bond. Internal hydride transfer occurs in the
intermediate (I) so the overall result is cis addition of BH 3 to
the olefin. With an excess of olefin the alkylborane formed (e.g.,
II) may react as did BH 3 to form di- or fn-alkylboranes.
126 ORGANIC REACTION MECHANISMS

— H

The alkylboranes may be converted to various other products.

One of the most useful procedures is oxidation with alkaline hydro-
gen peroxide, producing boric acid and an alcohol.

NaOH
BR 3 + 3H 2 2 3R0H + B(OH) 3

The borane coordinates with H 2 2, and the intermediate under-

goes a reaction related to the perbenzoate rearrangement dis-

cussed in Special Topic 3.

\_/ -O~
/
+ H2
hydrolysis
•- ROH + HOB
/
R B
\ \
v
XOH
[

The migrating group in such rearrangements retains its configura-

tion, and the subsequent hydrolysis of the borate ester involves
cleavage of the B —
O bond. Thus in the product alcohol the OH
group has the same configuration as the original boron atom in
the borane; e.g., oxidation of II leads to III; it should be noted
that the over-all addition of H 2 to a double bond by this sequence
is not only stereospecific but also leads to an orientation the re-

verse of that from acid-catalyzed hydration.

H2 2

NaOH

Nucleophilic addition to olefins. Just as electrophilic

addition to olefins may be considered the reverse of El elimina-
IONIC ELIMINATION AND ADDITION REACTIONS 127

tion, nucleophilic additions to olefins are the reverse of elimina-

tions involving carbanions. The nucleophilic addition of HY to

a double bond may be symbolized as follows.

+
\ / I
e/ H

Y Y H

In general such a process will require two special features: the

nucleophile must be a good one, and the resulting carbanion must
be strongly stabilized. It will be seen that these are similar to the
factors cited for the operation of the ElcB elimination mechanism:
a stabilized carbanion, and a poor leaving group (which will
probably therefore be a good nucleophile). The two mechanisms
are connected by the principle of microscopic reversibility; if an
elimination reaction involves an ElcB mechanism, then its reverse
under the same conditions must involve nucleophilic addition via
the carbanion.
As was discussed earlier, some elimination reactions which
appear to have "carbanion character" actually involve extremes
of the E2 elimination mechanism, in which proton removal runs
ahead of departure of the leaving group. Similarly, the micro-
scopic reverse of such an elimination would involve attack by the
nucleophile but then the beginning of proton donation before
carbanion development was complete.

OH 8 ~OH ©OH
H

\ c= / c —» \ c —i-l/ c * \ c — 1/
Y e t y

Thus while a number of nucleophilic additions to activated double

bonds are usually written as if the free carbanion were involved,
the possibility must be kept in mind that proton addition could
occur before complete carbanion formation.
A few examples of such reactions will be encountered in Chapter
6.
128 ORGANIC REACTION MECHANISMS

CN CN CN CN CN
/ KCN / /
C H CH — CH
e
—C
| I

C H CH=C
6 5 C 6 H 5CH >
6 5

\ H2 \ co uQ \ co
_
cor 2 2

50
/
/
H
CN / CN
C6H5 CH—
co 2
Rate is found =k [Olefin] [CN ]

General References

D. Banthorpe, Elimination Reactions (Elsevier Publishing Co., New

York, 1963). A monograph in which a wide range of mechanisms
are critically examined.

D. J. Cram, "Olefin-Forming Elimination Reactions," in M. Newman,

ed., Steric Effects in Organic Chemistry (John Wiley & Sons, New
York, 1956). A good survey of mechanisms with some emphasis
on stereochemistry.

J. F. Bunnett, "The Mechanism of Bimolecular /^-Elimination Reac-

tions," Angewandte Chemie, International Edition, 1, 225 (1962).
A review article, in English, with many references.

J. Hine, Physical Organic Chemistry (McGraw-Hill Book Co., New

York, 1962). Chapters 8 and 9 discuss both eliminations and
additions. Chapter 22 covers a-eliminations.

W. Saunders, "Elimination Reactions in Solution," in S. Patai, ed.,

The Chemistry of Alkenes (John Wiley & Sons, New York, 1964).
A critical treatment at a rather advanced level.
 Special Topic

THREE- AND FOUR-CENTER ADDITION

REACTIONS

there are a number of reactions in which a reagent adds to a

double bond by simultaneous attack on both unsaturated carbons,
or where this is at least a formal possibility. For instance, in the
reaction of an olefin with osmium tetroxide to form a cyclic osmate
ester there is no needto postulate any sort of charged intermediate.
The over-all change can be pictured as involving a simple cyclic
flow of electrons, with a simultaneous valence change of the

129
130 ORGANIC REACTION MECHANISMS

1
osmium atom. The Diels- Alder reaction can also be written with
a cyclic flow of electrons, without the necessity to postulate
transient intermediate ions.

o=c c=o
I
I

CH CH
CH 2 CH 2
ch==ch/
The addition of a carbene to an olefin is an example of a

three-center process, in which again no unstable intermediate is

formally required.

CI CI CI CI
,
c
/ \/
H
/ \\
H. /
• • v /
>
+ y--V

The question we will examine is the following: in such reactions,

is addition to the double bond really simultaneous, or are the two
bonds to carbon formed one at a time?

1,1-Addition Reactions

The addition of a carbene to an olefin is considered a 1,1 -ad-

dition, since both new bonds come to the same carbon (of the
carbene). Similarly, formation of a bromonium ion is a 1,1 -addi-
tion, as is epoxidation by peracids, although in these cases another
fragment is simultaneously eliminated.

1. (a) M. "The Diels- Alder Reaction with Maleic Anhy-

Kloetzel,
Adams, ed., Organic Reactions, Vol. 4 (John Wiley &
dride," in R.
Sons, New York, 1948), p. 1; (b) H. Holmes, "The Diels-Alder
Reaction: Ethylenic and Acetylenic Dienophiles," in R. Adams,
ed., Organic Reactions, Vol. 4 (John Wiley & Sons, New York,
1948), p. 60; (c) L. Butz and A. Rytina, "The Diels-Alder Reac-
tion: Quinones and Other Cyclenones," in R. Adams, ed., Organic
Reactions, Vol. 5 (John Wiley & Sons, New York, 1949), p. 136;
Needleman and M. Chang Kuo, "Diels-Alder Synthesis with
(d) S.
Heteroatomic Compounds," Chemical Reviews, 62, 405 (1962).
THREE- AND FOUR-CENTER ADDITION REACTIONS 131

XX
\ /
XX
\/
c c

\f\/ — \/\/
Br e
c
Br* Br
f ^ \ C / \C /
C^= C — ' / \
C4H5 O

o) H

Q OH
\/ \ / /
y x
\
These processes have been written as if the addition were indeed
simultaneous; the evidence in favor of this is the finding that the
addition is stereospecific. As has already been discussed for
bromonium ion formation, reactions with c7s-2-butene and with
trans-2-butent would lead to the same mixture of products if

addition were stepwise. It is actually found that the isomeric

olefins yield isomeric products in bromonium ion reactions, in
epoxidation, 2 and in
3
some carbene additions.

CH 3
H.
\ c= /
/ \
H

CH 3
~
C 6 H 5 C0 3 H

CH 3
H o
\ /\
/«-%
/
H

CH 3

2. E. Gould, Mechanism and Structure in Organic Chemistry (Henry

Holt & Co., New York,
1959) p. 534. ,

3. (a) P. Skell and A. Garner, "The Stereochemistry of Carbene-

Olefin Reactions," Journal of the American Chemical Society, 78,
3409 (1956); (b) R. Woodworm
and P. Skell, "Methylene, CH,.
Stereospecific Reaction with cis- and fra/7s-2-Butene," Journal of
the American Chemical Society, 81, 3383 (1959).
132 ORGANIC REACTION MECHANISMS

Not all carbenes undergo stereospecific addition to olefins.

There are two possible electronic states 4 for any carbene, the
singlet state and the triplet state; apparently only the singlet gives
stereospecific addition. Considering CH 2 itself, in the singlet state
(in which all electrons are paired) the carbon is approximately sp 2
hybridized. Two of the sp orbitals are used for single bonding to
2

hydrogen while the third contains the unshared pair of electrons

5
(actually spectroscopic evidence suggests that this orbital has
more s character than the other two). The remaining p orbital is

vacant. Thus the singlet state resembles a carbonium ion, except

that a proton is missing. In triplet :CH 2 the carbon is (approxi-
mately) sp-hydridized, each of the sp hybrid orbitals being used for
the C —H single bonds; the two unshared electrons on carbon are
placed in the p y and p z orbitals left over. By Hund's rules 7 one of
these electrons will go into each p orbital, so that they stay as far
apart as possible, and they will also unpair spins so as to give a
"triplet state," i.e., a state with two unpaired spins.

Vacant

Singlet methylene Triplet methylene

In the triplet methylene the two unshared electrons are in p

an sp 2 hybrid orbital, which
orbitals, while in the singlet they are in

4. For a good review, cf. (a) J. Hine, Physical Organic Chemistry

(2nd ed., McGraw-Hill Book Co., New York, 1962), Chapter 24;
(b) J. Hine, Divalent Carbon (Ronald Press, New York, 1964);
(c) W. Kirmse, Carbene Chemistry (Academic Press, New York,
1964).
5. Ref. 4 (a), p. 493.
6. Cf. ref. 5, but also E. Wasserman, A. Trozzolo, and W. Yager,
"ESR Hyperfine of Randomly Oriented Triplets: Structure of
Substituted Methylenes," Journal of Chemical Physics, 40, 2408
(1964), who conclude that triplet carbenes are not completely
linear.
7. C. Coulson, Valence (2nd ed., Oxford University Press, London,
1961), p. 36.
THREE- AND FOUR-CENTER ADDITION REACTIONS 133

has more s Thus one might

character and thus lower energy.
have expected the singlet state to be more
However, the stable.
decrease in electron repulsion from separating the two electrons in
different p orbitals and from unpairing their spins apparently more
than makes up for the higher energy of a p orbital, so that the
triplet state of :CH 2 is more stable than the singlet state.

When diazomethane is decomposed by heat or light, singlet-

state methylene is produced (unless a special type of photosensi-
Chapter 8). If the decomposition is carried out
tizer is used, cf.
in the presence of an olefin, stereospecific addition of this singlet
methylene occurs. 8 Even though the triplet state of :CH 2 is more
stable, it is relatively difficult to interconvert singlet and triplet

ch 3k CH.3

c=c
y

CH 3 CH.3

© Light
CH 2 - N = N
or
>
GCC -> H H

Heat

states; addition occurs before there is time for the formed

initially

singlet to be converted to the more stable triplet.However, in

the gas phase in high dilution there is enough time for decay from
9
the singlet to the triplet states ; furthermore, using certain types of
10
photosensitizers it is possible to produce triplet-state methylene

directly from diazomethane. Then it is found that addition of this

triplet-state methylene is no longer stereospecific.

CH 3
Light, benzophenone
as photosensitizer Cis-2-butene
CH 2™2
2N H-CT-H > H

Gas phase CH 3

in N2
Light „ .

ch 2 n 2 -^(tJ^c;

8. Ref. 3b.
9. F. Anet, R. Bader, and A. van der Auwera, "Chemical Evidence
for a Triplet Ground
State for Methylene," Journal of the Ameri-
can Chemical Society, 82, 3217 (1960).
10. K. Kopecky, G. Hammond, and P. Leermakers, "The Triplet
State of Methylene in Solution," Journal of the American Chemi-
cal Society, 84, 1015 (1962).
134 ORGANIC REACTION MECHANISMS

The difference in behavior between these two electronic states

of a carbene is not surprising. The singlet adds in one step, but
addition of the triplet to a double bond is a two-step process; the
first which must then be converted
step leads to a triplet diradical,
(slowly) to the singlet state before ring closure can occur. As
expected from this mechanism, olefins which are particularly

T
h 2ct+
n
c— c
|CH
\
>-c— c
2

T
— CH 2 ^
V
-c-c
f
— yy ru 2
A"

reactive toward free radical additions are very reactive toward the
11
triplet states of carbenes, although they do not exhibit any such
preference for the singlet carbenes.
When methylene iodide is treated with zinc (as a zinc-copper
alloy) in the presence of an olefin, CH 2 is added to the double
bond. 12 At one time this was considered to be a carbene addition,
but it is now clear that the intermediate is a "carbenoid," i.e., a
carbene-like fragment bonded to something else, in this case zinc
iodide. It is also clear that a number of other "carbenes" produced
in solution by elimination reactions, even ":CC1 2 " from the alkaline

/
Zn. \=S\
/ CH 2
CH 2 2 I + Zn >CH 2 \ / \ C /+
\ C Znl 2
I
/ \
13
hydrolysis of chloroform, are not completely free. Such car-
benoids exist in the singlet state and give one-step stereospecific
addition.

11. R. Etter, H. Skovronek, and P. Skell, "Diphenylmethylene, a

Diradical Species," Journal of the American Chemical Society, 81,
1008 (1959).
12. F. Blanchard and H. Simmons, "Cyclopropane Synthesis from
Methylene Iodide, Zinc-Copper Couple, and Olefins. II," Journal
of the American Chemical Society 86, 1337 (1964). ,

13. Cf. W. Miller and D. Whalen, "Trichloromethyllithium, an Elec-

trophilic Reagent," Journal of the American Chemical Society,
86, 2089 (1964), and references therein.
 THREE- AND FOUR-CENTER ADDITION REACTIONS 135

1,2-Addition Reactions 14

On being heated to 200°C, tetrafluoroethylene dimerizes to oc-

tafluorocyclobutane. This is considered a 1,2-cycloaddition re-
action, in this case a symmetrical one. Fluorinated olefins also
will add 1,2 to other materials, e.g., butadienes, and similar
cycloadditions are known for ketenes, allenes, and activated olefins
such as acrylonitrile.

(1) 200°C CF 2
— CF 2

CF 2 =CF 2 ^CF 2 — CF 2

CH 3 CH 3 CHCH 3 CH 3
(2)
CH 2 =
I

C— C=CH
I

2 -r-CCI 2 = CF 2
100°C
>CH =C-C
2
\^ |
2

:cf 2

CCI
(3)

o+ CH 2 =C = Q
100°C • S

(4) 400°C # /
CH 2 =C=CH S +
% / 15%
CN CN
(5) 250°C /
CH 2 = CHCN >
+
CN CN

In all these except the first reaction there is in principle a choice

of orientation, and the actual direction of addition is that shown.
Addition cannot be a simple, completely concerted process, since
such a mechanism would not explain the preferred orientations.
Considering reaction (5), for instance, the dimerization of acrylo-
it seems that both electrostatic and steric effects would
nitrile,

favor the formation of 1,3-dicyanocyclobutane. The product

14. J. Roberts and C. Sharts, "Cyclobutane Derivatives from Thermal

Cycloaddition Reactions," in A. C. Cope, ed., Organic Reactions,
Vol. 12 (John Wiley & Sons, New York, 1962), p. 1.
136 ORGANIC REACTION MECHANISMS

CH 2 CH CN
— f-

NC— CH=CH 2

actually formed is 1,2-dicyanocyclobutane. For such reasons it

has long been considered that in these cycloadditions one bond is

formed more rapidly than the other. In this case a diradical

intermediate may be postulated, since it is known that radicals are
stabilized by conjugation with cyano groups.

CN CN CN
CH 2 =CH CH 2 — CH- CH 2 — CH
+
CH 2 = CH CH 2 — CH; CH 2 _CH

CN CN CN

An ionic intermediate would not explain this orientation since a

cyano group does not stabilize adjacent positive charge, although

either a diradical or a zwitterionic intermediate structure could
account for the ketene orientation in reaction (3). One of the

./ /
CH 2
or

or CHi

questions one might ask about such intermediates is the following:

Do they represent really free diradicals, perhaps with unpaired
electron spins, or do they simply involve a very long bond in
which the paired electrons are so far apart that they behave to
some extent as if they were in free radicals? The answer to this
15
question is not known for all cases, but recent experiments related

15. P. Bartlett et ah, "Cycle-addition," Journal of the American Chemi-

cal Society, 86, 616, 622, 628 (1964), Journal of Organic Chem-
istry, 32, 1290 (1967); cf. also S. Proskow, H. Simmons, and T.
Cairns, "Stereochemistry of the Cycloaddition Reaction ," . . .

Journal of the American Chemical Society, 85, 2341 (1963), for

a different result.
THREE- AND FOUR-CENTER ADDITION REACTIONS 137

to reaction (2) show that for this case a real diradical is involved,
not simply a long bond.
Reaction of l,l-dichloro-2,2-difluoroethylene with cis,cis-2,A-
hexadiene yields two stereoisomeric products, I and II. These are
formed because the intermediate diradical has time to rotate about
the new single bond before it collapses to product; if formation of
one bond simply ran slightly ahead of the other, such isomerization
would not be possible. Thus even though in the starting material
the two hydrogen atoms were cis to each other, in product I they
are trans because of this opportunity for single-bond rotation in the
intermediate.

s_ /
j" ^~^CH = CHCH 3 +
™< r
?"^H
cHCHs

J*"""
CF 2 — CCI 2 CF 2 — CCI 2

1,3-Addition Reactions

A number of reagents give 1,3-addition to olefins. The

reaction of olefins with Os0 4 has already been mentioned. Ozone
also undergoes 1,3-addition to double bonds, although the "initial"
ozonide then rearranges to a more stable ozonide. 10
The concept of 1,3-additions as a related class of reactions was

16. P. Bailey, "The Reaction of Ozone with Organic Compounds,"

Chemical Reviews, 58, 925 (1958). For evidence that some
ozonolyses have more complex mechanisms see R. Murray, R.
Youssefyeh, and P. Story, "Ozonolyis ," Journal of the Amer-
. . .

ican Chemical Society, 89, 2429 (1967) and references therein.

138 m ORGANIC REACTION MECHANISMS

CH 3 CH 3
©
.O O
/ \Q-~s
c=c + o T o° c—— c
H
/ \H *
© > H
/\ l\
H >

O CH 3 CH 3
©y V°
X
o o® o-o
II II 1,3-addition to / \
CH
i
+ CH
i
c=o
> CH V
i
xo xXH i

CH 3 CH 3 CH 3 CH 3

first advanced in 1938. 17 However, activity in discovering new

reactions of this type and in investigating mechanisms has been
quite recent, and is chiefly due to the efforts of Huisgen and his
18
collaborators. 1,3-Dipolar systems may be symbolized generally
as III or IV; diazomethane is an example of system III, while
ozone is an example of system IV.

aO a© a©
1 1
1
1

1 1
1

D. < * b© b: <- > b ©

III

c u II

in IV

CH 2© CH2© o©
i
o©
1 1
i

N
1

< * N© o <- >

o©
II
III

N© N o© o

17. L. Smith, "Systems Capable of Undergoing 1,3-Additions," Chemi-

cal Reviews, 23, 193 (1938).
18. (a) R. Huisgen, "1,3-Dipolar Cycloadditions," Angewandte
Chemie, International Edition, 2, 563 (1963); (b) R. Huisgen,
"Kinetics and Mechanism of 1,3-Dipolar Cycloadditions," Ange-
wandte Chemie, International Edition, 2, 633 (1963).
THREE- AND FOUR-CENTER ADDITION REACTIONS 139

Most of the possible systems III and IV are known, where a, b,

and c are carbon, nitrogen, or oxygen. Many of them will un-
18
dergo 1,3-additions to double bonds, and the evidence so far sug-
gests that these additions are more or less concerted (i.e., the
second bond at least begins to form before the first is complete)
The primary evidence for this is stereospecificity in all addi-
tions so far examined. Thus diazomethane affords different
products with the esters of dimethylfumaric and of dimethylmaleic
acids, and unstable diphenylnitrilamine (V) affords stereospecific
adducts with the stilbenes if it is generated in their presence.

N
CH 3 CH 3
CH 2 — N = N
©
+
\ C=C / > CH 3 . ,CH 3
X |

7
CHaOsC C0 CH2 3
C0 CK 2 3 C0 2 CH,

CH 3 CO»CH 3
©
CH 2 — N=N + C =C

C6H

C6H5 —C- -C^.c 6 H5

H H

Such stereospecificity shows that a free open-chain intermediate is

not formed.
140 ORGANIC REACTION MECHANISMS

Nonetheless, 1,3-dipolar addition is probably not a completely

concerted process. The best evidence for this comes from the
effect of substituents on the rates of some addition reactions. 18
Simple olefins are much less reactive than are olefins substituted
with conjugating groups; for additions by benzonitrile oxide in
ether at 20 °C, for instance, the following relative rates are found.

CH 2 =CH- alkyl CH 2 = CH — phenyl CH 2 =CH —C0 2 R

3.9 14 100

© C6H5 —
// \
w cs O
C6 H CEEN-O + CH =CHR
5 2 >
/
CH 2 — CHR
|

If addition had been completely concerted one might have ex-

pected no particular help from conjugating groups, while with an
intermediate radical or anion being partially developed in the
reacting olefin the acceleration can be accounted for.

N N

c h —c
6 5
s \9- <— c6h5 — c ^ \9©
CH 2 CH— R CH 2 CH-

Addition to the olefin results in a particular orientation, as was

shown above. This might also be taken as an indication that the
transition state has some charge-separated character, but other
evidence 18 suggests that steric effects are mostly involved, addi-
tion to the olefin occurring so as to separate the bulky substituent
groups as much as possible. In general, 1,3-dipolar additions are
not accelerated by polar solvents; therefore in the transition state
there is no more charge separation than in the starting materials.

1,4-Addition Reactions

In the Diels-Alder reaction 1 a conjugated diene reacts across its

1 ,4-positions with an olefin; the latter is called the dienophile in this

reaction. Almost any olefin may serve as a dienophile undersome
conditions, but olefins substituted by electron-withdrawing conju-
 THREE- AND FOUR-CENTER ADDITION REACTIONS 141

gating groups are usually more reactive. The reaction of methyl

trans-cvoiondXQ with cyclopentadiene is a typical Diels-Alder reac-
tion,

o "h C
II
Acetone

H
/\C0 CH 2 3

It will be noted that in the two products stereospecific addition has

occurred across the crotonic ester bond, i.e., the methyl and
carbomethoxyl groups are still trans. Thus addition to the double
bond has been cis; many cases have been examined, and this result
is always found. Accordingly, we can again say that the addition
must be more or less concerted, and no free diradical intermediate
capable of rotation about the new single bond is involved.
In contrast to this evidence that the two bonds are simultane-
ously formed, the orientation in unsymmetrical Diels-Alder addi-
tions can be most easily understood in terms of open-chain interme-
20
diates. Thus the reaction of 2-methoxybutadiene with acrolein
yields the adduct expected if an intermediate zwitterion or diradical
had been formed, since the methoxyl and carbonyl groups would

CH 3 CH3O. CH 3 Q CHO
not

X
CHO CHO
CH3O.

U etc.
CH =0
or
CH3O

etc.
CH =0

19. J. G. Martin and R. K. Hill, "Stereochemistry of the Diels-Alder

Reaction," Chemical Reviews, 62, 537 (1962).
20. Cf. R. Woodward and T. Katz, "The Mechanism of the Diels-
Alder Reaction," Tetrahedron, 5, 70 (1959).
142 ORGANIC REACTION MECHANISMS

stabilize Although this is still a matter of

such intermediates.
21
controversy, seems most likely that no such intermediate is
it

involved, but that the two single bonds are forming at the same
time, with one running ahead of the other in most cases. The
above reaction can be represented with one
transition state for the
short bond and one long bond.
In a long bond the two electrons
may both be near one carbon or they may be more or less localized
with one on each carbon (though still with spins paired) In some .

cases the transition state will thus appear to be dipolar, while for
other reactions substituent effects will suggest diradical character.
With this concept of a "long bond" both the stereospecificity and
the substituent effects can be explained.

CH 3 (X /^ CH 3 C> /. CH,0

CHO \© ©^CHO \ O^CHO

It is striking that the Diels-Alder reaction seems to be con-
certed, in the sense that both new bonds are forming at essentially
the same time, while at least some of the 1,2 additions we dis-
cussed earlier occur one bond at a time. Since nonconcerted
reactions require the formation of unstable intermediates, we
might wonder why any cycloaddition would be nonconcerted. Is
there any reason why two ethylene molecules cannot simply come
together in a single-step reaction to form cyclobutane? Of course,
such a process has a considerable entropy disadvantage since
four atoms must be aligned, but this entropy problem is also found

in the Diels-Alder process which is in fact concerted.

The distinction in mechanisms seems to derive chiefly from the
fact that the Diels-Alder reaction involves the cyclic flow of six
electrons, while dimerization of ethylene involves four electrons.
In the transition state for a concerted process these electrons be-
come delocalized over the system before they are localized again
in the new bonds, and transition states with 4n + 2 cyclically de-
localized electrons are much better than those with 4n cyclically
delocalized electrons. This sounds like the distinction between
aromatic and antiaromatic systems discussed in Special Topic 1,

21. J. Berson and A. Remanick, "The Mechanism of the Diels-Alder

Reaction," Journal of the American Chemical Society, 83, 4947
(1961). Provides a good critical discussion.
THREE- AND FOUR-CENTER ADDITION REACTIONS 143

and it has the same origins in molecular quantum mechanics. A

consideration of orbital symmetries also shows what the problem
is in olefin dimerization (or other 2 + 2 cycloadditions, involving
two electrons from each component).
It is necessary to consider a molecular orbital description of
the two new single bonds in cyclobutane formed by dimerization
22
of ethylene.

9
>-l B
-;
v;
c c
v.-
c
f c

4 3 4
o 3
I II
l/
2 (0, + 2 + 03 + 04) 1/2(0, - 02 - 03 + 04)

The bonds are formed by overlap of lobes of the sp 3 hybrid

o-

orbitals; as with p orbitals, the function corresponding to one lobe

may have either a positive or negative sign, and bonding occurs
when lobes of two overlapping sp 3 orbitals have the same sign.
The four sp 3 orbitals on carbons 1, 2, 3, and 4 may be combined
in two ways, I and II, so as to have bonding between 1 and 4, and
2 and 3. The ( + ) or ( — ) signs have no intrinsic significance; a
structure like I with all ( — ) signs would be equivalent to I,
3
signifying simply that all four sp lobes have the same sign. Four
electrons are involved in the two new bonds; two electrons go into
m.o. I and two into m.o. II. Each of these is a molecular orbital,
involving all four atoms. Thus orbital I has not only the desired
1-4 and 2-3 bonding interactions, but also some extra 1-2 and
3-4 7r-like bonding. However, in II the 1-2 and 3-4 interactions
are antibonding, and cancel the extra 7r-like bonds in I. The net
result with two electrons in I and two electrons in II is equivalent
to the normal cyclobutane description with localized bonds.

22. R. Hoffmann and R. B. Woodward, "Selection Rules for Con-

certed Cycloaddition Reactions," Journal of the American Chemi-
cal Society, 87, 2046 (1965).
144 ORGANIC REACTION MECHANISMS

Since for most purposes the two-electron localized bond descrip-

tion of o- bonds is equivalent to a full m.o. description, we have
not used the latter up until now. In the present case, however, we
are concerned with the transition state in a hypothetical concerted
cycloaddition of two ethylenes. The two ethylenes can themselves
be lined up in two ways, considering orbital symmetries.

m IV

Arrangement III would go smoothly over

to I of the cyclobutane
as the reaction proceeded, but arrangement IV would not go to
II but instead to a very high energy structure in which the two new
o- interactions are antibonding. In fact, the arrangement which
could go to II is V, in which the two ethylenes themselves have not
bonds but antibonds. Thus one of the two cyclobutane ground-
state m.o.'s is derived from a strongly excited state of the two
ethylenes, and vice versa. For the transition state of a concerted
addition to be of low energy the electrons at the transition state
must occupy orbitals intermediate between ground-state orbitals
of the starting material and ground-state orbitals of the product.
In this case no such intermediate orbitals are possible, but in the
2 + 4 Diels-Alder reaction they are.
This latter point is most easily seen if we consider molecular
symmetry. Ethylene has a plane of symmetry perpendicular to
the carbon-carbon bond axis, and relative to this symmetry plane
the two p z orbitals may be either symmetric (same sign and magni-
tude, leading to the tt bond) or antisymmetric (same magnitude
but opposite sign, leading to the antibond).
-n- These are sym-
bolized S and A.
THREE- AND FOUR-CENTER ADDITION REACTIONS 145

When two ethylenes conceitedly close to a cyclobutane, this same

molecular symmetry plane is always maintained, and any wave
function must therefore be either S or A with respect to the plane.
In cyclobutane, for instance, orbital I is S and orbital II is A
with respect to the plane (which is perpendicular to both the
1-2 and 3-4 single bonds). However,
two starting ethylene
in the
arrangements, III and IV There is no way in which
are both S.
an S orbital in the starting material and an A orbital in the
product could be smoothly interconverted through some inter-
mediate function. Either the intermediate wave function changes
its sign or it doesn't on crossing a symmetry plane, so it must be

either S or A.

A / '%

In the Diels-Alder reaction between an olefin and a diene the

situation is different. Butadiene has two occupied -n m.o.'s, which
are S and A with respect to a symmetry plane. (See next page.)
The product cyclohexene has three new bonds to accommodate the
six electrons involved in the change: a new tt bond and two new
o- bonds. The ?r bond is S relative to the symmetry plane (as is
146 ORGANIC REACTION MECHANISMS

am*
7)
c
| >; +

c
W
- c -
fi
\
© Q £+%

^,(S) Y 2
(A)

bond of the ethylene molecule), while the two

the reacting o- bonds
can be combined into an S and an A molecular orbital.

The six electrons of the starting materials (olefin and diene) are
thus in two S and one A orbitals with respect to the symmetry
plane, and so are the six electrons in the product. Orbitals retain
theirsymmetries throughout the reaction so the transition state
can resemble both starting materials and product. It can have
some of the bonding characteristics of both, and thus be stabilized.
This type of argument will be pursued in more detail in Special
Topic 8.
 AROMATIC SUBSTITUTION

as was discussed in Special Topic 1, there are many types of

aromatic systems. Nonetheless, the chemistry of benzene and its

simple derivatives has been studied in the most detail, so we shall

concern ourselves with reactions on a benzene ring. Nitration of
benzene and conversion of chlorobenzene to aniline illustrate the

two major first is an

classes of aromatic substitution reactions; the
electrophilic substitutionby nitronium ion, while the second is a
nucleophilic substitution by the amide anion. As we will see, in

©
+ H

+ CI
e

147
148 ORGANIC REACTION MECHANISMS

neither case does the actual mechanism involve a one-step dis-

placement, however. Free radical substitution reactions are also
known, but they will not be considered here.

5-1 Electrophilic Substitution

Three questions are usually of concern in aromatic substitution

reactions: ( 1 ) What is the attacking species; (2) how does it carry
out the substitution; and (3) how is the reaction influenced by
other groups on the benzene ring? We shall consider the first two
questions for several important substitution reactions, and shall
deal with the third question at the end of this section.
'Nitration. Benzene may be nitrated by a mixture of nitric
and sulfuric Physical evidence clearly shows that the
acids.
nitronium ion, N0 2
+
is formed under these conditions.
, The
earliest type of evidence was the finding from freezing-point
studies that one molecule of nitric acid forms four particles in
sulfuric acid.

HN0 + 3 2 H0SO4 > N0 2 + H3O0 + 2 HSO 4

More recently, spectroscopic studies have demonstrated the pres-
ence of N0 2 in such solutions. A variety of evidence
+
indicates
that N0 is the actual nitrating agent with the
2
+
ordinary
H S0
2 4 —HNO3 mixture. Thus, it is possible to prepare authentic
salts of nitronium ion, such as 4 N0 BF
2
+
and such salts will nitrate ,

aromatic compounds. Furthermore the relative reactivities of

different aromatic compounds toward authentic nitronium salts are
the same as their relative reactivities toward nitric-sulfuric acid
mixture, suggesting that the same species is attacking in both
cases.
A number of mechanisms could be imagined for attack by
nitronium ion. Direct displacement of hydrogen is perhaps the
most obvious, but it has been ruled out by examination of the
isotope effect. Studies of benzene and its derivatives containing
deuterium or tritium show that such hydrogen isotopes are replaced
as easily as is protium —ordinary hydrogen. However, whenever
a proton is lost in the rate-determining step of a reaction it is found
that there is an isotope effect, the heavier deuterium or tritium
being lost less readily. The absence of an isotope effect here
AROMATIC SUBSTITUTION 149

NO,
©

© N0 2

©Slow ©
+ N0 2 NO. + BH
Fast

NO;
©
shows that the proton is not lost in the rate-determining step of
nitration, and helps establish the two-step mechanism. The re-
moval of proton by some base in the system (e.g., HS0 4 ) must
be a fast step; if it were rate-determining we would again expect a
large isotope effect.
In Special Topic 5, the mode will be discussed
of attack of N0 2
+

in more detail. There be seen that under some

it will also
conditions nitrations do not involve the free nitronium ion.
Sulfonation. Treatment of benzene with fuming sulfuric
acid affords benzenesulfonic acid. In this case S0 3 (or HS0 3 +
)

attacks the aromatic ring in the first step, but the attack is

apparently reversible. This is revealed by the fact that there is a

large isotope effect in this case, so proton removal is rate-
determining. From this evidence alone, proton loss might be
simultaneous with attack by S0 3 , but the two-step mechanism
seems more likely.

H SO3H
^
O*
Fast Slow
+ so 3 ;
«

Halogenation. Many aromatic compounds can be bromi-

nated with solutions of bromine in acetic acid, and kinetic studies
support a mechanism involving attack by molecular bromine.
150 ORGANIC REACTION MECHANISMS

n<
->Br— Br
Slow
>
y^yT^ Br
Fast
>
j

More detailed studies of aromatic halogenation have been per-

formed in aqueous solution (although preparative reactions are
usually not performed in this solvent).
When chlorine is dissolved in water a number of equilibria are
set up involving such species as hypochlorous acid (HOC1),
-
hypochlorous acidium ion (H 2 OCl + ), and hypochlorite ion (CIO ).

The relative concentrations of these species depends on the pH

and the chloride ion concentration.

Cl 2 + H,0 TZl H2OCI0 + cie

H 2 oci e ^z^ hoci + he

HOC1" >
C1O0 +H0
The three species underlined have all been identified as respon-
sible for aromatic chlorination in some cases, by a study of the
effect of pH
and chloride ion concentration on the rate of chlorina-
tion of somesubstrate. Of these, Cl 2 is apparently the most
reactive, and it will attack unactivated benzene rings. Hypochlo-
rous acid is quite unreactive, since the aromatic compound must
displace a hydroxide ion from the chlorine atom, but it has been
identified as the species which chlorinates phenoxide ion, a very
reactive aromatic species.

o©

7-tl-
ci-!-oh b:J
T
CI

Under unusual conditions Cl + can also apparently be formed, and it

is the most reactive chlorinating species of all.

AROMATIC SUBSTITUTION 151

Friedel-C rafts reactions. Treatment of an aromatic com-

pound with an alkylating agent leads to substitution, the so-called
Friedel-Crafts alkylation. The most common alkylating agents
are mixtures of alkyl halides, such as methyl bromide, and strong
Lewis acids such as aluminum bromide. The reaction may be
considered a nucleophilic substitution on the methyl bromide by
the aromatic ring, for which there are two obvious possible
mechanisms. Either the aluminum bromide removes the bromide
ion first, leaving a methyl cation to attack the ring, or there is an
S N 2 displacement on the methyl group and the bromide ion is

removed simultaneously,

CH 3 Br + AlBn Crf+A.Br®
or

© H CH,
CH 3 — Br --AIBr 3
CH 3

Since the methyl cation is so unstable, we would be surprised if the

first mechanism were correct. The direct displacement mecha-
nism has been proved for this reaction by the finding that methyl
iodide gives a different mixture of substitution products on toluene
than does methyl bromide. If the methyl cation were the attacking
species it should behave in the same way no matter how it is

formed.

+ CH 3 X + AIBr 3

CH 3
ch 3 i : 49% 11% 40%
CH 3 Br I 54% 17% 29%

With alkyl groups which afford more stable cations, such as

?-butyl, it is the carbonium ion itself that attacks the aromatic
152 ORGANIC REACTION MECHANISMS

ring. Carbonium ion rearrangements in the alkyl group are fre-

quently observed.
Direct evidence for the proposed reaction intermediates in
aromatic substitution has been obtained by Olah using NMR
spectroscopy. A
mixture of mesitylene with an alkyl halide and
Lewis acid at low temperature yielded the substitution intermediate
I. This went on to the final product at a higher temperature.

+ CH 3 CH 2 F + BF 3

CH 3 CH 3 ' X "CHs CH 3
'

H CH.CH, CH 2 CH 3

BF 4

Friedel-Crafts acylation of aromatic rings involves reaction

with a Lewis acid and an acid chloride or other acylating agent. In
this case there is evidence that the halide ion is lost first, the
attacking species being an acylium ion.

O
// © ©
R-C-CI + AICI 3 ^ZZ! R— C=0 < >R— C=Q AICI 4

Substituent effects. Since the transition state in electro-

philic aromatic substitution involves the creation of positive
charge on the aromatic ring, substitution is facilitated by groups
which can help stabilize such a charge. These groups affect the
rate at which different aromatic compounds can be attacked, and
they also affect the position of attack in a given compound such as
toluene. In simple qualitative terms a group such as the methyl in
AROMATIC SUBSTITUTION 153

toluene is said to be activating, since toluene is more reactive than

benzene, and ortho-para directing, since preferential substitution
occurs in these positions, as the above Friedel-Crafts alkylation
showed.
These two different kinds of reactivity have been collected in

the concept of a "partial rate factor." Instead of asking "How

reactive is toluene compared with benzene, and what is the
percentage of ortho, meta, and para substitution" for some reac-
tion, we ask "How reactive is the para position of toluene
compared with a position on benzene?" This relative reactivity
would be the partial rate factor. Thus for bromination in acetic
acid the partial rate factor for the position in toluene para to the
methyl — — /? f
Me
is 2420. The partial rate factor for meta substi-
tution — — f
Me
is 5.5.
Simple resonance arguments make it clear why the para position
in toluene is over 2,000 times as reactive as a position in benzene.
The transition state resembles the charged intermediate, and in
one of the resonance forms there is a stable tertiary carbonium ion
instead of a secondary one.

For meta substitution all three resonance forms are secondary

carbonium ions, so meta substitution is much less favorable.

Each meta position is only 5.5/2420 times as reactive as the para

position, although a meta position in toluene is still 5.5 times as
reactive as a benzene position. This latter increase must reflect a
small inductive stabilization of the transition state by methyl even
when it is not directly located on a charged carbon.
254 ORGANIC REACTION MECHANISMS

The partial rate factor for bromination of toluene in an ortho

position Here we would again expect strong activation
is 600.
because one of the resonance forms is a tertiary carbonium ion.
The fact that an ortho position is only Va as reactive as the para
position is due to steric hindrance (although there is 50% as much
ortho as para product since there are two ortho positions).

Such hindrance varies with the size of the attacking group. For
chlorination of toluene in acetic acid the partial rate factors are:
l Me M
Ot \ 617; ra f , 5; p ( \ 820. Thus here an ortho position is

almost as reactive as the para, reflecting less hindrance in the

transition state for ortho attack. At first sight it might be thought
that this simply indicates that a chlorine atom is smaller than a
bromine, but something more is involved. The partial rate factor
for para chlorination is smaller than for bromination, although
both charged intermediates look much the same. This is because
the charged intermediates come after the transition states; in the
transition state for aromatic substitution only part of the positive
charge will yet be developed on the benzene ring. The smaller
Me
/? f for chlorination suggests that it has an earlier transition state,

in which less charge has yet developed. Thus there would in any
case be less crowding in the transition state for ortho substitution,
since the attacking chlorine molecule is further away.
It has so far not been possible to arrive at a good quantitative
treatment for ortho substitution, because steric hindrance comes in
as well as carbonium ion stability. However, the relative reac-
tivities meta and para positions of most compounds in most
of
electrophilic substitutions can be successfully correlated with two
sets of constants in a modified Hammett Chapter 4).
equation (cf.
+
Each substituent is assigned a substituent constant, called er (sigma
plus) and each reaction is assigned a reaction constant, called p
+
(rho). The substituent constant a measures the ability of a
particular substituent, such as a p-methoxyl group, to stabilize the
+
charged intermediate; is related to a of Chapter 4, but for sub-
o-

stituents which have strong resonance interaction with a positive

AROMATIC SUBSTITUTION > 755

+
charge the value of o- is different from that of <r. The reaction
constant measures the susceptibility of the particular reaction to
such stabilization (i.e., the extent to which the actual transition
state resembles the charged intermediate). Then the partial rate
factor for substitution para to a group R in a bromination, for
instance, obeys the following relationship,

log Pf = <T p.R '

pbromination

while a similar relationship governs reactivity at the meta posi-

tions

log nif = (T
+
m --R '
pbromination

+
In Table 5-1 are listed values of o- for some common substituents;
in Table 5-2 are listed the P values for several typical aromatic
substitutions.

TABLE 5-1
Substituent Constants for A romatic Subst 'tut ion
+ +
o- meta o- para

CH 0-3 0.047 - 0.778

CH - 3 - 0.066 -0.311

H-
Cl- 0.399 0.114

Br- 0.405 0.150

+
Me 3 N - 0.359 0.408

H0 C- 2 0.322 0.421

-C=N 0.562 0.659

-N0 2 0.674 0.790

+
Negative o- constants indicate activation, while a group with a
+
positive o- para, for instance, is deactivating to the para position.
Thus going down the table, the methoxy group is strongly
activating to the para position (and hence to the ortho position as
well) while it is mildly deactivating to the meta position. A
methoxy group stabilizes an adjacent positive charge by a reso-
nance interaction in para substitution; the intermediate in meta
256 ORGANIC REACTION MECHANISMS

substitution is slightly destabilized since its positive charge is near

the positive end of a carbon-oxygen dipole.

©OCH 3

+
The o- constants for methyl show that it is activating to both the
meta and para positions, as was discussed above. The meta
activation is and the para activation is less effective
very slight,

than for methoxy. Hydrogen is the standard, whose o- + constants are

zero by definition. Chlorine and bromine deactivate both the
meta and the para (and ortho) positions; since the meta position is
most strongly deactivated, substitution occurs mostly ortho and
para to the halogens. This can be understood in terms of an
inductive effect which strongly withdraws electrons from the ring,
and at the same time a possibility for some resonance stabilization
of the intermediate in para (and ortho) substitution.

The last four groups listed deactivate the para position more
than they deactivate the meta. They have either a positively
charged atom or the positive end of a dipole attached to the ring,
and they strongly destabilize the positively charged intermediates
AROMATIC SUBSTITUTION - J 57

in aromatic substitution. The effect is larger for para (and ortho)

substitution because one resonance form of the intermediate places
the positive charge right next to the deactivating group, for
instance the nitro group.
The values of P listed in Table 5-2 show how sensitive typical
aromatic substitution reactions are to the effects of substituents. A
large negative P indicates that the reaction is strongly affected,
while the Friedel-Crafts alkylation with ethyl bromide has a small
negative p, so it is rather insensitive to substituent effects. This is

more apparent if we calculate the partial rate factors for bromina-

tion and for Friedel-Crafts ethylation para to the methoxy in
anisole.
Me +
log /?f° — cr p_0Me

For bromination,
Me
logPf° (bromination) = (- 0.778) (- 12.1) = 9.414
thus
OMe
/? f (bromination) = 2,600,000,000

The para position in anisole is over two billion times as reactive

toward bromine as a simple benzene position would be. However,
for ethylation,

togP™
e
(ethylation) = (- 0.778) (- 2.4) = 1.867
thus
0Me
pf (ethylation) = 74

Although there is still activation, the effect is much smaller. The

transition state for Friedel-Crafts ethylation occurs very early, and
not much charge has yet developed in the ring.
Similar calculations can be done for the other reactions, and
TABLE 5-2

Reaction Constants for Aromatic Substitution

Reaction p

Br, bromination, acetic acid, 25 °C —12.1

Cl 2 chlorination, acetic acid, 25°C —10.0
Friedel-Crafts acetylation — 9.1
HNO., nitration in nitromethane, 25°C — 6.0
Friedel-Crafts ethylation (alkylation) — 2.4
158 ORGANIC REACTION MECHANISMS

using other substituents. In general the agreement with experi-

ment is very good, although it is not always perfect. However, a
very important substitution reaction is missing from the table,
nitration using nitric and sulfuric acids. This reaction does not fit
+
the & P relationship at all; the very interesting reason for this is

discussed in Special Topic 5.

5-2 Nucleophilic Substitution

Chlorobenzene is rather inert under the usual conditions for

nucleophilic displacement reactions. Apparently S \2 displace- x

ment an unsaturated carbon is quite an unfavorable process,

at
while S N 1 reactions do not occur because of the instability of the
resulting carbonium ion, a phenyl cation. Although nucleophilic
substitution reactions of aromatic compounds are thus not as
common as are electrophilic substitutions, they can occur under
certain conditions.
Aromatic diazonium compounds. When aniline is treated
with nitrous acid it is converted to the benzenediazonium cation.
This can be decomposed by heating in water, phenol being formed
by a nucleophilic substitution reaction. In the presence of a
high concentration of chloride ions the product is instead largely
chlorobenzene, but the rate of decomposition of the diazonium salt
AROMATIC SUBSTITUTION 159

is unaffected. This evidence shows that the rate-determining step

does not involve the nucleophile, and supports an S N 1 mechanism.
The very unstable phenyl cation is possible here only because the
formation of N2 is a very favorable process.
Addition-elimination on activated aromatic rings. Just
as electrophilic substitution involves the formation of a positively
charged intermediate, nucleophilic substitution may occur through
a negatively charged intermediate. Such a process is facile only if
the negative charge can be stabilized by appropriate substituents.
Thus 2,4-dinitrochlorobenzene reacts with methoxide ion in
methanol solution via such an intermediate anion; conjugation
with the nitro groups makes this occur under conditions in which
chlorobenzene is perfectly inert.

N0 2
@
+ OCH,

OCH CH 3 CI CH3O CI

Fast ^XV^^NOz /\^N0 2

+ CI
u NO,
- u
/ ®\
o o
It is interesting that 2,4-dinitrofluorobenzene is much more
reactive than is the chloro compound. As was seen in Chapter 3,
fluoride ion is a poorer leaving group than is chloride ion, so the
greater reactivity of the fluoro compound here is at first sight
surprising. However, the rate-determining step in these substitu-
tions is addition of the anion. Loss of the halide ion is a
subsequent fast step, even with the slower fluoride ion. Addition
to the ring is favored (by inductive and field effects) by the
electron-withdrawing halogen, so the first step is faster with the
fluoro compound.
J60 ORGANIC REACTION MECHANISMS

A number of other groups may activate the ring in this way, e.g.,
carbonyl or sulfonyl groups. In addition, this type of mechanism is

important in the chemistry of heterocyclic aromatic systems

when the negative charge can be placed on electronegative
atoms. For instance, 2-bromopyridine reacts with ammonia to
afford 2-aminopyridine.

+ NH 3 > LBr
X Br N ^NH 2

Substitution via benzyne. When /?-bromoanisole is treated

with potassium amide in liquid ammonia solution, it is converted
to a mixture of m- and p-anisidines.

Furthermore, treatment under these conditions of chlorobenzene

labeled with carbon- 14 affords aniline in which the label is essen-
tially equally distributed between the carbon bearing the amino
group and a carbon ortho to it. These results are best rationalized
by postulating that substitution proceeds through a benzyne, a
cyclic acetylene.
AROMATIC SUBSTITUTION 161

For simplicity the elimination of HC1 is written as a two-step

process, involving the carbanion, although this is not really
certain.
Benzyne and substituted benzynes are now well known reactive
intermediates, which are most conveniently generated in ways
other than that above. For example, treatment of o-bromo-
fluorobenzene with lithium in the presence of anthracene generates
benzyne, which is trapped by the anthracene to yield the interesting
compound triptycene.

+ Li

Br

It is not surprising that benzyne is quite unstable and has only been
detected as a reactive intermediate. Acetylenes ordinarily are
linear, since they form o- bonds with sp hybrid orbitals and use the
other p orbitals for -n bonding. In benzyne this is not possible, and
the new 'Sr bond" which has been formed in the plane of the ring
involves overlap of hybrid orbitals.

This strain decreases as the ring becomes larger; cyclooctyne is an

isolable, if somewhat unstable, compound.
162 ORGANIC REACTION MECHANISMS

CH 2 CH 2

CH 2 C

CH 2 C

CH 2 CH 2

Direct displacement. With weaker bases than NH 2 ~,

direct substitution on aromatic halides may compete with ben-
zyne formation. The hydrolysis of chlorobenzene with strong so-
dium hydroxide solution at a high temperature yields phenol;
studies with radioactive carbon show that this largely, but not
exclusively, involves benzyne as an intermediate.

4M NaOH
340° C

The fact that more radioactivity is found in C-l of the phenol

shows that a direct substitution process occurs along with the
benzyne mechanism. Apparently the direct displacement mecha-
AROMATIC SUBSTITUTION 163

nism resembles a simple S x 2 reaction; it becomes more favorable

when the leaving group is better.
However, this S N 2-like reaction is obviously still quite dif-

ficult, requiring conditions much more vigorous than those used

for displacements on simple alkyl halides.

General References

L. Stock, Aromatic Substitution Reactions (Prentice-Hall, Englewood

Cliffs, New Jersey, 1968). A good modern introduction.

G. Olah, Fried el-Crafts and Related Reactions (Interscience Publishers,

New York, 1963). A four-volume compendium which discusses in
detail not only Friedel-Crafts acylation and alkylation, but also
other aromatic substitution reactions, e.g., nitration and halogena-
tion.

J. Hine, Physical Organic Chemistry (McGraw-Hill Book Co., 1962),

Chapters 16 and 17. Discusses-both electrophilic and nucleophilic
substitutions.

L. M. Stock and H. C. Brown, "A Quantitative Treatment of Directive

Effects in Aromatic Substitution," Advances in Physical Organic
Chemistry, 1, 35 (1963). A review of the evidence for the
a-'p correlation.

J. F. Bunnett, "Mechanism and Reactivity in Aromatic Nucleophilic

Substitution Reactions," Quarterly Reviews {London), 12, 1 (1958).
A good review of the S x l, bimolecular, and benzyne mechanisms.

H. Heaney, "Benzyne and Related Intermediates," Chemical Reviews,

62,81 (1962).

G. Wittig, "Small Rings with Carbon-Carbon Triple Bonds," Ange-

wandte Chemie, International Edition, 1, 415 (1962).

K. Wiberg, "The Deuterium Isotope Effect," Chemical Reviews, 55,

713 (1955). For more information on how isotope effects can be
used to elucidate mechanisms.

L. Melander, Isotope Effects on Reaction Rates (Ronald Press Co.,

New York, 1960). Chapter 6 deals with aromatic substitution.
164 ORGANIC REACTION MECHANISMS

H. Zollinger, "Hydrogen Isotope Effects in Aromatic Substitution Re-

actions," in V. Gold, ed., Advances in Physical Organic Chemistry,
Vol. 2 (Academic Press, New York, 1964), p. 163.
 Special Topic

THE ROLE OF tt COMPLEXES IN

AROMATIC SUBSTITUTION

when benzene is added to a mixture of HF and BF 3 a reaction

occurs with formation of the highly colored o- complex. 1

+ HF + BF 3 > < > etc.

A cr complex

A proton has been added to one of the benzene ring carbons, and in
the resulting species the proton and the ring are thus "complexed"
by the formation of a new o- bond. In the last chapter such o-

1. J. Hine, Physical Organic Chemistry (2nd ed., McGraw-Hill Book

Co., New York, 1962), p. 348.

165
166 ORGANIC REACTION MECHANISMS

complexes were simply called "charged intermediates." Solutions

of these complexes are electrically conducting, since the complexes
are actually salts. Furthermore, when DF is used instead of HF
the deuterium exchanges with the benzene protons, since in the o-
complex the deuterium newly attached to the ring carbon is now
equivalent to the proton already on that benzene carbon, so that
either may be lost when the complex is decomposed. These
properties, along with spectroscopic studies, support the structure
assigned.
By contrast, a different sort of interaction occurs when HC1 is
1
dissolved in benzene. Studies of the solubility of HC1 show that it
forms a 1 complex with benzene, but the solution has no color.
: 1

Furthermore, it does not conduct electricity, so a salt has not been

formed. Finally, when DC1 is used no deuterium exchange with

the ring protons occurs. This second type of complex is called a -n
complex. Apparently the proton of HC1 has a weak interaction
with the 7T electrons of the benzene ring without being transferred
away from the chlorine atom.

+ HO > -H > H — CI

A 7r complex

The bonding in tt complexes can be described in either molecular

orbital or valence bond terms. In a molecular orbital picture the
hydrogen Is orbital has a little overlap with the two carbon p
orbitals of a n bond, while retaining most of its overlap with the

chlorine orbital.
ROLE OF tv COMPLEXES IN AROMATIC SUBSTITUTION 167

In valence bond terms the ?r complex is a hybrid of three principal

resonance forms; the neutral one is most stable and contributes
most to the structure.

c. ci ci
I

H < > H < > H

C =C C
/ ©
C
©
C
\>r
The valence bond picture shows that there is some polar character
to these complexes, and they are often called "charge-transfer
complexes." 2
A variety of electrophiles can give such u complexes (charge-
transfer complexes) with aromatic compounds. Silver ion forms
a complex with benzene and other aromatic rings (and also with
simple olefins), halogens such as Cl 2 form tt complexes with aro-
matic compounds under mild conditions in which substitution does
not occur (e.g., low temperature), and polynitro compounds such
as picric acid form crystalline complexes with many aromatic
hydrocarbons. 2

OH
A9 . ^ °2N J ,N0 2

C >CI— CI
/

The exact structure of a -n complex depends on the nature of the

donor (the molecule, such as benzene, which partially donates
electrons and acquires a little positive charge) and the acceptor.
An x-ray diffraction study of the complex of benzene with silver
perchlorate shows that the silver ion lies above one particular
carbon-carbon bond, 2 while in the benzene-bromine complex the
bromine lies above the center of the benzene ring 2 (the bromine

2. For a good general discussion, cf. L. Ferguson, The Modern

Structural Theory of Organic Chemistry (Prentice-Hall, Englewood
Cliffs, New Jersey, 1963), p. 103.
168 ORGANIC REACTION MECHANISMS

molecule lies on the bromine atom

axis of the ring, the nearest
interacting equally with p orbitals). In a complex of picric
all six

acid with an aromatic hydrocarbon the two lie face-to-face. 2

Overlap of the two systems is stabilizing since there is then some
-n-

charge transfer from the donor molecule to the picric acid (with
three electron-attracting nitro groups).
Since electrophiles can form -k complexes with aromatic com-
pounds, and many carry out substitution on
electrophiles also
benzene one might wonder whether ?r complexes play a role in
rings,
substitution reactions. For instance, the following scheme can be
pictured for the bromination of benzene.

Br

H Br

+ Br 2 ^Zl [
-j) >Br-Br

+ HBr

ft Complex

This sequence is probably correct. However, the first step,

formation of a -n complex, is rapid and reversible and the transition
state occurs during the conversion of the -n- complex to theo- com-

plex. Thus in discussing the effects of substituents on aromatic

substitution reactions there is no need to take account of the inter-
mediate formation of a ?r complex, for the transition state is what
matters and it strongly resembles the o- complex. The strongest
evidence for this is found in Table 5-3, in which the relative rates
of bromination in acetic acid solution are compared with the equi-
librium constants for o-complex formation (HBF 4 ) and ^-com-
1
plex formation (HC1) for a series of methylbenzenes.
It is quite apparent from this table that both the rate of
halogenation and the stability of a a complex are strongly increased
by methyl substituents, while the stability of a t complex is
ROLE OF 7T COMPLEXES IN AROMATIC SUBSTITUTION 169

TABLE 5-3

Rates of Bromination of some Methylbenzenes Compared with the

Equilibrium Constants for a Complex and n Complex Formations
Relative Relative Relative
Bromination a Complex 7r Complex
Benzene Derivative Rate Stability Stability

Benzene 1 1 1

Toluene 605 7 1.51

1 ,4-Dimethylbenzene 2500 11 1.65
1 ,2-Dimethylbenzene 5300 12 1.85
1 ,3-Dimethylbenzene 514,000 290 2.06
1 ,2,4-Trimethylbenzene 1,520,000 700 2.23
1,2,3-Trimethylbenzene 1,670,000 770 2.40
1 ,2,3,4-Tetramethylbenzene 11,000,000 4400 2.68
1,3,5-Trimethylbenzene 189,000,000 145,000 2.60
1 ,2,3,5-Tetramethylbenzene 420,000,000 178,000 2.74
Pentamethylbenzene 810,000,000 322,000 —
affected much less. This difference is to be expected. In a -n

complex only a little charge is transferred to the ring, since the

principal resonance form is the neutral one. Furthermore, this
charge is placed in -n- orbitals delocalized over all six benzene
carbons, and it can be stabilized more or less equally by methyl
groups in any position. Thus there is little difference in tt

basicity among the three isomeric xylenes (dimethylbenzenes), or

among the three trimethyl isomers or the two tetramethyls.
On the other hand, the relative stabilities of o- complexes are

strongly affected by both the number and the position of methyl

substituents. (1,3-dimethylbenzene) is almost 30
/rz-Xylene
times as basic as the other xylenes, since both methyls are on
charged carbons.

BF,®

As expected, mesitylene (1,3,5-trimethylbenzene) is even more

basic, since the third charged carbon now bears a methyl group,
170 ORGANIC REACTION MECHANISMS

although the 500-fold effect of this last group seems rather large.
This trimethylbenzene is even more basic than is 1,2,3,4-
tetramethylbenzene; in the o- complex from the latter only two of the
positive carbons carry methyl groups. Relative reactivities to-
ward bromine also follow this order, although the tt complex
stabilities are reversed.
Table 5-3 clearly indicates that the transition state for halo-
genation resembles the o- complex, and not the tt complex. In fact,
halogenation is even more sensitive to substituent effects than o-

complexing is. This surprising result may reflect subtle differ-

ences in the reaction conditions, since it is not reasonable that the
transition state for bromination could have more positive charge
on the benzene ring than is found in a a complex.
As we saw in Chapter 5 (Table 5-2), of all the common
substitution reactions bromination is the most sensitive to substitu-
ent effects. However, the other reactions listed in Table 5-2 also
show relatively large increases in rate for methylated benzenes;
consequently their transition states also resemble the a complex
more closely than the ?r complex. This was also implied in our use
of partial rate factors. The concept of a partial rate factor
involves the idea that in the transition state the electrophilic
reagent is attacking a particular carbon atom of the benzene ring,
on the way to a o- complex. If the transition state were instead on
the way to a tt complex this treatment would not have worked. In
fact, it fails for nitration by nitronium ion. As is discussed below,
this failure seems to be strong evidence that in this case the
rate-determining step is formation of the tt complex.
The nitration of a series of aromatic hydrocarbons has been
3
studied, using nitronium fluoborate in tetramethylene sulfone as
solvent. Relative to benzene, toluene has a reactivity of 1.67,
ra-xylene a reactivity of 1.65, and mesitylene a reactivity of only
2.71. These very small effects of the methyl groups are very
similar to their effects on the stability of tt complexes, suggesting that
the transition state in this case looks like a tt complex, but these data
alone are not enough to establish the point. The rate-determining
step might be formation of a a complex, but the transition state

3. G. Olah, S. Kuhn, and S. Flood, "Aromatic Substitution. VIII,"

Journal of the American Chemical Society, 83, 4571 (1961); see
also C. Ritchie and H. Win, "Some Nitronium Tetrafluoroborate
Nitrations," Journal of Organic Chemistry, 29, 3093 (1964).
ROLE OF tv COMPLEXES IN AROMATIC SUBSTITUTION 171

might occur so early that very little charge had yet developed on

the benzene ring. However, an attempt to calculate partial rate

factors indicates that something is wrong with this picture. Nitra-
tion of toluene leads to only 2.8% meta substitution, with 65.4%
ortho and 31.8% para. Thus the partial rate factor for the meta
Me
position of toluene, /?z f can be calculated from these data:
,

Me
m f
= (1.67) (0.028) (3) = 0.14

The factor of 3 was included to take account of the fact that

benzene has six equivalent positions while toluene has only two
meta positions; the partial rate factor is the reactivity of a meta
position compared to a single benzene position
The conclusion of this calculation is that a methyl group
meta position, reducing
deactivates the its reactivity to 14% of
that of a benzene carbon. This is completely contradictory to all
+
evidence from other reactions. As was discussed in Chapter 5, o-

for ra-methyl is — 0.066, indicating that it is an activating group,

although a weak one. It seems that the partial rate factor treat-
ment is not valid for this nitration. 3
The situation becomes clear if we consider the possibility that
the rate-determining step in this case is formation of a u complex.

Then nitronium ion exhibits low selectivity in forming the u com-

plex, since rr complexes of toluene are only slightly stronger than
772 ORGANIC REACTION MECHANISMS

those of benzene. On the other hand, when the ?r complex collapses

it may go to three different a complexes, leading respectively to
ortho, meta, and para substitution, and it tends to collapse mostly
to the better o- complexes. Since in the rate-determining step it is

not yet clear which carbon will finally be attacked, it is not

surprising that the partial rate-factor treatment does not work.
1
These conclusions have been attacked by Tolgyesi on the basis
of work which seemed to show that reaction occurs while the
reagents were being mixed. If the nitronium fluoborate reacted

before the solution became homogeneous, then the reaction would

seem to be nonselective among substrates for purely mechanical
reasons (scavenging of all aromatic molecules in the vicinity of a

drop of nitrating solution). However, Olah has since shown 5

that this does not occur, so at the present time the original interpre-
tation — rate-determining formation of a ?r complex —seems the
most likely.

Similar studies have been conducted with nitronium salts in

other solvents, 6 and the results are the same. The same selec-
tivity is also observed 6 in nitration with nitric-sulfuric acid mix-
tures, in which the active species is nitronium ion. However,
nitration with nitric acid in nitromethane and other organic sol-
6
vents probably does not involve nitronium ion, since the
toluene/benzene reactivity is 26, a high relative reactivity
consistent with rate-determining a complex formation. The 3.1%
meta substitution which is observed leads to a m Me
f of 2.4; this
activation is as expected for rate-determining o- complex forma-
tion.
It is interesting that the rate of nitration of toluene with nitric
acid in nitromethane is independent of the toluene concentration. 5
This and other kinetic evidence had been taken to mean that the

4. W. Tolgyesi, "Relative Reactivity of Toluene-Benzene in Nitronium

Tetrafluoroborate Nitration," Canadian Journal of Chemistry, 43,
343 (1965).
5. G. Olah and N. Overchuk, "Remarks on the Nitronium Salt Nitra-
tion of Toluene and Benzene," Canadian Journal of Chemistry, 43,
3279 (1965).
6. G. Olah, S. Kuhn, S. Flood, and J. Evans, "Aromatic Substitution.
XIII," Journal of the American Chemical Society, 84, 3687 (1962).
7. C. Ingold, Structure and Mechanism in Organic Chemistry (Cornell

University Press, Ithaca, New York, 1953), p. 275.

ROLE OF 7T COMPLEXES IN AROMATIC SUBSTITUTION 173

7
rate-determining step was formation of nitronium ion, but as we
have seen nitronium ion is probably not involved since in other

solvents it gives toluene/benzene selectivities of 1 .7 or so, not the

observed 26. The true nitrating agent in this system (N 2 5 ?) has
not yet been identified.
It would be expected that ?r complex formation could be rate
determining whenever the attacking reagent is so electrophilic that
the 7T complex cannot dissociate once it is formed. This has also
been found 8 9 to be true in some Friedel-Crafts alkylations. Thus
when toluene is alkylated with benzyl chloride and aluminum
chloride in nitromethane only 4.5% meta substitution occurs,
indicating high position selectivity. However, toluene is only 3.2
times as reactive as benzene, indicating low substrate selectivity;
this combination of facts, which would lead to a m Me
f of 0.43,
shows again that the rate-determining step is -n complex formation,
and that the product distribution is determined later.

Rate- Fast
C 6 H 5 CH 2 CI + + AICI-
Determining

©CH.,C 6 H.

Fast
CH C 6 H
2 5
+ Products

CH,C 6 H,

+
In such cases the partial rate-factor treatment, and the o-
P relation-
ship which was applied to it, are not valid. However, for most
aromatic substitutions ?r complex formation is just a very probable
first reversible step, and or complex formation is rate-determining.

8. G. Olah, S. Kuhn and S. Flood, "Aromatic Substitution. X,"

Journal of the American Chemical Society, 84, 1688 (1962).
9. G. Olah and N. Overchuk, "Aromatic Substitution. XXV," Journal
of the American Chemical Society, 87, 5786 (1965).
 6
REACTIONS OF CARBONYL COMPOUNDS

a carbonyl group conveys two characteristic types of reactivity

to a molecule. Through enol or enolate ion formation a carbonyl
group activates adjacent —
C H bonds toward substitution, while
there are also a number of characteristic addition and substitution
reactions which occur at the carbonyl group itself. Activation of
—
an adjacent C H is characteristic of all types of carbonyl com-
— —
pounds ketones, aldehydes, esters, amides, etc. and it will be
considered first. However, since ketones and aldehydes undergo
reactions at the carbonyl group which are rather different from
those of esters and other carboxyl derivatives, these will be
considered separately in the second and third parts of this chapter.

6-1 Enols and Enolates

As was discussed in Chapter 2, the bromination of acetone

occurs by enolization, followed by bromination of the enol. The
kinetic expression, when the reaction is conducted in an aqueous
acetic acid/sodium acetate buffer, indicates that both acids and
bases catalyze the process. With an acid it seems clear that the
enol is formed, but for the process catalyzed by OH", for instance,

174
REACTIONS OF CARBONYL COMPOUNDS 175

the reactive intermediate is probably the enolate ion instead.

o©
Br,
CH 3 —
I

©
CH 3 — C— CH 3 + OH,0 CH 3 CCH 2 Br + Br

\CH,
The distinction is real, but it is difficult to make experimentally,
and we shall often not distinguish between enols and enolate ions in
discussing base-catalyzed processes.
Quite a complete study has been done on phenyl sec-butyl
ketone (I), which is optically active. The rate of acid-catalyzed
iodination is identical with the rate of racemization under the same
conditions, and the rates of base-catatyzed bromination, deutera-
tion, and racemization have also been found to be the same. All
involve enol (or enolate ion) formation as the rate-determining
step.

CH CH,
3

Slow
^'C-CC 6 H5

CH 3 I

Br.
CH 3CH; OH CH 3 CH< O
II

c=c
I

cx — CC H 6 5

CH,
/
CH, CeHj

DoO

CH 3
— CH 2 CH CH
3 2

C— CC 6 H5 + C — CC H 6 5

1
CH 2
D CH 3

Acetone contains, at equilibrium, only 0.00025% enol. As

Table 6-1 shows, however, other ketones may be much more
enolic. In acetylacetone the enol is stabilized by an additional
resonance form.
176 ORGANIC REACTION MECHANISMS

TABLE 6-1

Per Cent Enol at Equilibrium ( 'n the Pure Liquid)

Compound % Enol

Acetone 0.00025
Cyclohexanone 0.020
Acetylacetone 80
Acetoacetic ester 7.5
Biacetyl 0.0056
1 ,2-Cyclohexanedione 100

O O </ \
11 II

CH 3 CCH 2CO Is —>CH -C

3
1 II

C--CH 3 * >CH 3 -C
ii i

CCH 3
% /
This stabilization becomes less effective in the enol of acetoacetic

ester, for the ester group has less electron-withdrawing ability.

B
/P
° y° 9
II

CH 3 CCH 2 — Cf < >

II

CH 3 CCH 2 -C >

\
OEt
^©OEt

OH O OH—^O
©
— II

CH 3 —
I

C— OEt
I
I

CH 3 < > C=OEt

\CH/ %CH /
v

C D

©OH— ®0
—C II I

CH 3 C_OEt
/ ^
CH

This effect can be described simply by saying that since the ethoxy
group already feeds electrons into the carbonyl, the latter has less
tendency to accept electrons from the enol group. Alternatively,
REACTIONS OF CARBONYL COMPOUNDS 177

in resonance terms we note that the starting ketone has two

resonance forms for the ester group, A and B, while the product
enol has only forms C, D, and E. Forms C and D are the ordinary
ester resonance forms without any special interaction with the enol,
and form E is the only one in which the enol interacts with the ester
carbonyl group. This special interaction thus plays only a minor
role in the structure of the enol; consequently the enol is less

stabilized than it was for acetylacetone.

Comparing acetone with cyclohexanone, it is seen that the cyclic
ketone is much more enolic. This is true chiefly because the
enolization of acetone involves loss of freedom of rotation when
the carbon-carbon single bond becomes a rigid double bond.
Cyclohexanone is already rigid, so rotational freedom is not lost on
enolization.

This is also partly involved in the difference between biacetyl and

1,2-cyclohexanedione, but another big factor comes from the con-
formations of the two starting materials. In biacetyl the two
carbonyl groups are oriented so as to minimize dipole-dipole re-
pulsion, but in the cyclic compound they are held so that this
repulsion is quite strong. Enolization helps to relieve this dipole
opposition; therefore cyclohexanedione is completely enolic even
though the enol oxygen is not conjugated with the second carbonyl
group.

ffl
CH 3 C

C CH S

i
o
178 ORGANIC REACTION MECHANISMS

We have not referred to the effect of hydrogen bonding on the

stabilities of these enols, although for the dicarbonyl compounds
the enol was in all cases written with an internal hydrogen bond.
This is undoubtedly a stabilizing factor as well, and is another one
of the reasons that diketones are more enolic than monoke-
tones. It is not required for enolization, however, for 1,3-
cyclohexanedione derivatives are almost completely enolized
even though they cannot form internal hydrogen bonds; in such
cases the usual factors of conjugation in the enol and restricted
rotation in the ketone are enough.

A-
CH 3

Since, as we have seen, the rate-determining step for many

reactions involving enols is the enolization itself, relative rates of
enolization are of more interest than the equilibrium values just
discussed. The best data available are for the rates of enolate ion
formation, so they will be of use chiefly in considering base-
catalyzed processes. In Table 6-2 there are listed, for some
representativecompounds, both the rate constant for removal of a
proton by water and the equilibrium constant K a for this reaction.
Kl
k
TABLE 6-2 R—H+±RO+H O® s K =-^
a
k>
• 52° C
K
Compound ki (per second) Ka
10" 10 -20
(1) CH3COCH3 4.7 X 10
(2) CH COCH Cl
3 2 5.5 X 10" 8 3 x 10 17

(3) CH COCHCl
3 2
7.3 X 10" 7 10- 15

(4) CH COCH COCH

3 2 3
1.7 X 10- 2 1.0 x 10- 9

(5) CH COCH C0 C H 1.2 X 10" 3 2.1 X 10- 11

(6)
3

CH C0 C H
3 2 2
2

5
2 2 5

— 3 X 10" 25

(7) CH C0 C H
2( 5)2
2.5 X lO- 5 5 X 10" 14

(8) CH (CHO)
2
2 2

2
— l x 10- 5

(9) CH CONH
3 2
3 X 10" 14 10" 25

(10) CH CN 3
7 X 10- 14 10" 25

(11) CH (CN)
2 2
1.5 X 10" 2 6.5 X 10- 12

(12) CH S0 CH
3 2 3
3 X 10- 12 10 -23
(13) CH N0 3 2
4.3 X 10- 8 6.1 X 10- 11
REACTIONS OF CARBONYL COMPOUNDS 179

Of course when a stronger base than water is used the rates will be
faster, but the relationship of rates for different compounds in

general will not be affected.

Examining the first three entries in the table, we see that the rate
of enolate ion formation increases as the equilibrium stability of the
enolate increases. The greater acidity of chlorinated acetones is

due to the inductive effect of the chlorines; this effect carries over
to the rate of ionization as well, since the transition state strongly
resembles the enolate ion.

O 5© O
'CI I! I

CI cx
CH; / > /
CH £
C-H CH 3 C
i

5© OH, HsO©

The relatively large acidity of acetylacetone (4), pK a = 9, is also

reflected in a high rate of ionization, while the related acetoacetic
ester (5) is less acidic and less rapid in ionizing. Resonance forms
similar to those written for the enols explain both the activating
effect of a second carbonyl group and the lower effectiveness of
this carbonyl if it is part of an ester group.

O eo o o
C /C-CH 3
< > CH 3 -C C-CH 3

CH 3 CH CH
,0
o o
CH — C — OEt CH -C C— OEt
\
i
3 3

//
CH CH

o
©
= OEt
1 I

CH 3 -C C
% CH /
180 ORGANIC REACTION MECHANISMS

This weaker activation by an ester group is reflected as well in the

fact that ethyl acetate (6) is a weaker acid than is acetone ( 1 ) , and
malonic ester (7) is both less acidic and slower to ionize than
related keto compounds (4,5). In a sense, ester resonance is lost
in the enolate ion.
An aldehyde group is even more activating than a ketonic
carbonyl, as the high acidity of malondialdehyde (8), pK a = 5,
shows. This is to be expected since a methyl group is electron-
feeding relative to a hydrogen; consequently the ketone carbonyl
is a little less electronegative than an aldehyde carbonyl.

O O
II II

—C— H versus —C < CH 3

The carboxamide group is somewhat less activating (9) than an

ester group, since electron donation into the carbonyl by the amino
group is quite pronounced. A cyano group is also activating
(10,11) and is somewhat more effective than an ester group.

N N© N

c
# c
/ c
^
©CH
/ CH
s CH
/
V < >

\
< >

\
\ N
^ N
S Ne

The last two entries in the table show that both sulfonyl (12)

O O
© ©^ ®/
CH =N
CH.2— N

oe
V < > 2

o©
© '©
CH 2 — I

ft)
S^=i CH 3 < > CH 2 =S^ — CHi

O© O©
REACTIONS OF CARBONYL COMPOUNDS 181

and nitro (13) groups are also acidifying. For the nitro group an
ordinary resonance form can be drawn, but for sulfonyl it is
necessary to place more than eight electrons around sulfur.
This is possible since sulfur, a second-row element, has 3d orbitals
available of relatively low energy and can accommodate more
than eight electrons: two inits 3s orbital, six in its three 3p orbitals,

and two or more distributed among its five 3d orbitals. Actually

in the sulfone group itself there is already some double bonding

involving the use of these d orbitals, so the doubly charged sulfur

we have written is not correct.
We have repeatedly emphasized the parallel between rate of
ionization and stability of the product enolate ion, pointing out that
the transition state resembles the product. This is not always the
case, however. For instance, nitromethane (13) is more acidic,
pK a = 10.2, than is malonic ester (7), pK a = 13.3, but nitro-
methane ionizes more slowly. However, among more closely
related compounds quite good, and the type of
the parallel is

structural arguments we have been using can be employed in

predicting either relative rates or relative acidities.
This is another example of the Br0nsted relationship described
in Chapter 2. In order for the rate constants of ionization ki to
run strictly parallel to equilibrium acidity constants K a, rate con-
stants of reprotonation k 2 would have to be uniform. For acetone
(1) k 2 (=ki/K a ) is 4.7 X 10 10 liters/mole/sec, but it is only
7
5.8 x 10 liters/mole/sec for acetoacetic ester (5). Thus the full
effect of the equilibrium acidity is not seen in the rate. The theo-
retical upper limit for the bimolecular rate constant of a diffusion
controlled process, in which reaction occurs at every collision, is

about 5 x 10 10 liters/mole/sec for two ions in water. Thus ace-

tone enolate is protonated by H 3
+
on almost every encounter.
By contrast, the anion of nitromethane (13) has k 2 of only
6.8 x 10 2 liters/mole/sec. This is of course the rate constant for
protonation on the carbon; protonation at oxygen is much faster.

We have not mentioned one special structural which can effect

be important: the change in geometry which accompanies enoliza-
tion. For instance, it has been found that nitrocyclopropane is a
very weak acid compared with ordinary nitro compounds. This
2
reflects the fact that in the anion the ring carbon must become sp'
hybridized so that the remaining p orbital can be conjugated with
the nitro group. The optimal angles for an sp'~ carbon are 120°,
182 ORGANIC REACTION MECHANISMS

3
while for an sp carbon they are 109°28 / . Since the actual
cyclopropane angle is 60°, the angle strain is greater for the anion
than for the starting nitro compound, so I-strain suppresses ioniza-
tion (cf. similar effects discussed in Chapter 3).

O
CH 2

CH 2
vv ©r/

H
>
CH;

CH 2
©/
v
O

Strain 109°28 ' - 60°C = 49°28' Strain = 120° - 60° = 60'

A different example is found in the fact that compound I is not

particularly acidic, although Here the geometri-
it is a /3-diketone.
cal problem is that all four groups on a double bond must be in
the same plane, but this would not be possible for the enolate of I
(this is best seen with molecular models).

h A A
i
"N
H
1
X / x\
si
/-
/ \
I ± HM <T
^N
X
oe_

Enolate ions can be alkylated. Thus the malonic ester anion

reacts with methyl iodide to form a new carbon-carbon bond, and
this type of reaction is of considerable use in synthesis.

^0

C— OR CQ 2R

CH + CH 3 — I > CH 3 — CH /
C— OR C0 2 R

O
The reaction is simply an S N 2 displacement on methyl iodide by
REACTIONS OF CARBONYL COMPOUNDS 183

the enolate ion. It will be noted that the enolate ion reacts at its

carbon, although most of the negative charge is distributed over

the two oxygen atoms. This mode of reaction occurs because in
this case the transition state for reaction is fairly far along toward

product, and reaction at carbon leads to the more stable product

and hence to the better transition state. However, when enolate
ions are allowed to react with strong acids they protonate in a very
exothermic reaction, and the transition state comes very early.
Accordingly the fastest protonation occurs on oxygen to yield the
enol, and only more slowly does this isomerize to the stable keto
form. Some very exothermic alkylations have also been found to
lead to attack on oxygen, forming enol ethers. The factors which
lead to one or the other mode of attack are still under study.
One of the most important synthetic reactions of enolate ions is

O OR
\
C=C
/ RX
>
\
C=C
/ attack at center with

/ \ / \ most negative charge

O
> r q q attack to form more stable product

their addition to carbonyl groups, as in the aldol condensation.

This will be discussed at the end of the next section.
g

6-2 Addition to Ketones and Aldehydes

Reaction at the carbonyl group is the other characteristic process

for ketones These reactions may be classed as ( 1
and aldehydes.
simple additions or (2) additions followed by dehydration or other
changes.
Simple addition reactions. When ketones and aldehydes
are dissolved in water they establish equilibrium with their hy-
drates, g^ra-diols. This equilibrium is not of much chemical
interest, but it does furnish an indication of the ease with which
additions can occur across various carbonyl groups.

O OH
11

R— C — R' + ,
HO
— :
->
R
I

—C— /

OH
184 ORGANIC REACTION MECHANISMS

In Table 6-3 are listed the extent of hydration of a few carbonyl

compounds in dilute aqueous solution.
Formaldehyde is almost completely hydrated; with the succes-
sive addition of methyl groups one arrives at acetone, negligibly

TABLE 6-3

Compound % Hydrate at Equilibrium

CH 2 99.99
CH,CHO 58
CH,COCH 3

C H CHO
6 5 Slight
CCl,CHO 100
CF COCF
3 3 100

hydrated. This trend is the result of the stabilizing effect of

methyls on double bonds; in this case the electron-feeding methyl
groups stabilize the C=0 dipole.

O 50
5©

CH 3 CH 3

If the carbonyl group is stabilized by conjugation, as in benzal-

dehyde, there is also very little hydrate formation. Because of this
same effect esters, amides, acids, etc., are in general not hydrated to
a detectable extent. However, hydrate formation can be pro-
moted by destabilizing the carbonyl form; in chloral (trichloro-
acetaldehyde) and in hexafluoroacetone the carbon-halogen di-
poles interact unfavorably with the carbonyl group.

Cl
-/ C - C \ '7 C -C
\
/ H / C- > F
F
/ \
F F

The mechanism by which acetaldehyde reaches its hydration

REACTIONS OF CARBONYL COMPOUNDS 185

equilibrium has been studied. The reaction is not instantaneous,

but is catalyzed by both general acids and general bases. For the
acid-catalyzed process the following mechanism seems reasonable;

H H

CH — C
3
/ + HA ;=! CH — C3
/ Slow

h 2o

H H

CH —
3
/ r-
A
Fast
> CH 3
— C — OH +
I

HA

O— OH
H 2 0®
for the base-catalyzed process it is suggested that the catalyst
removes a proton from the attacking water molecule.

H H H
/ Slow / Fast
CH —
/
CH 3 — C > CH 3 -C > 3

O
OH OH

BH

Although these mechanisms are reasonable and fit the observed

catalysis they can hardly be considered as proved. If they are
correct, then the principle of microscopic reversibility tells us that
the dehydration processes must go back over the same paths.
Similar catalysis is found for the hydration of acetone. Although
at equilibrium there is a negligible amount of hydrate, the process
can be detected by observing O 18
exchange.

O 16
OH O 18

CH 3 — C
II

— CH 3 + H2 18
zzl CH 3
— C— CH
I

3 ^ CH 3
II

-C-CH 3

18
H
186 ORGANIC REACTION MECHANISMS

With alcohols, addition to an aldehyde results in hemiacetal

formation. The equilibrium extent of this process varies in a
manner similar to that for hydration, but it becomes quite impor-
tant when the alcohol and the aldehyde are in the same molecule, as
in glucose and other sugars.

CH 2 OH CH OH 2 CH2OH
Jr-Q H /r-OH
(oh X U J\PH /CH = 0"
:>\j__/ '
oh
OH
( -
Hoy-/
OH OH OH
a Open

Glucose

The equilibrium lies almost entirely to the two closed forms,

called, respectively, a and /?,D-glucose. The reason that the
equilibrium in hemiacetal formation is so much more favorable
here is a matter of entropy. In ordinary hemiacetal formation two
molecules, the alcohol and the aldehyde, must be tied down, and
there is freedom of motion; in cyclic
a considerable loss of
hemiacetal formation only one molecule is involved and much less

loss of freedom occurs since only some rotational freedom becomes

restricted. The rate of ring opening and closing can be studied by
following the optical rotation of the solution, since the interconver-
sion of a and fi forms involves opening and then closing in the
other stereochemical sense. This interconversion, mutarotation, is

catalyzed by both acids and bases in mechanisms apparently very

similar to those written for hydration.
Mercaptans add readily to carbonyl groups, and the simple
hemithioacetals and hemithioketals are converted further to thio-
acetals and thioketals, as discussed in the following section. Fur-
thermore, bisulfite ion and cyanide ion can add reversibly to many
carbonyl groups.

OH SR"

R— C— R' + R"-SH * R — C— R' R— C—

I

I I

SR" SR"
REACTIONS OF CARBONYL COMPOUNDS 187

H H O
7"
CHaC + HSO 3 m^CH -C-S^-O 3

O OH O
H H

C 6 H 5C + HCN ;
)
C6H5 — C — OH
%o c
I

III

The equilibrium depends on the structure of the carbonyl com-

pound as it did for hydrate formation, adduct formation being less
favorable when the carbonyl carries electron-feeding groups.
In the aldol condensation an enolate ion adds to a carbonyl
group. Thus treatment of acetaldehyde with mild base produces
/?-hydroxybutyraldehyde, which has the trivial name "aldol."
The mechanism currently favored is clearly related to that of the
other addition reactions discussed.

H O
Slow
>CH 3 — C— CH — #
I

I
\H
Fast

H O
CH 3 — C— CH -C 2

OH
\H
When acetaldehyde is present at moderate concentrations the
addition step is rate-determining, as the kinetic expression shows.

Rate = k (OH ) (CH3CHO) 2

However, when the acetaldehyde concentration is increased con-

siderably every enolate ion gets trapped, as in acetone bromina-
tion, and the rate expression becomes first order in acetaldehyde.
288 ORGANIC REACTION MECHANISMS

Rate = k (OH) (CH 3 CHO)

In the acid-catalyzed aldol condensation an enol, not an enolate

ion, is the species which adds to carbonyl.
A process which formally resembles the aldol condensation
occurs when benzaldehyde is treated with alkaline cyanide solution.
The product is benzoin, and the benzoin condensation involves a
specific role for the cyanide ion.

O OH
CN~
2 C 6 H 5 CHO C 6 H 5 C-C
II I

— 6 H5

Benzaldehyde cannot of course form an enolate ion, and the

hydrogen of the aldehyde group is not acidic. However, when
mandelonitrile is formed by addition of HCN across the carbonyl
group then hydrogen becomes acidified by the cyano group, and an
aldol-like condensation can occur. The full mechanism of this
process was presented in Chapter 2.

OH OH OH
Base /
—C —H —
I

— CG
I

C6H5 > C6 H5 < > C6H5

I
\ %

Complex addition reactions. Under this heading we shall

consider reactions in which the addition is followed by dehydration
or substitution. For when aldehydes
instance, are treated with
acidic alcohol, acetals may be formed. The process involves
hemiacetal formation followed by a substitution involving a
carbonium ion; this is a reversible process whose equilibrium can

O OH ^OH
— ffx 2
H 3 o© -H20
ch 3 — c— h
II I I

-
CH3CH+CH3OH ;
)
CH
I
I

OCH3 OCH-
REACTIONS OF CARBONYL COMPOUNDS 189

CH 3

\®
rwOH
ch,3 ° H
OH OCH 3
©
— C— H
I
I

CH 3
— C— H ;
CH 3 ^Z=± CH 3 — C-H
I I
I

OCH 3 OCH 3 OCH 3

be displaced by removal of the water formed. The intermediate

carbonium ion is strongly stabilized by conjugation with the
unshared electron pairs on oxygen. Thioacetal formation in-
volves a similar mechanism.
The aldol condensation can also be complex. For instance,
acid-catalyzed condensation of benzaldehyde with acetophenone
leads chiefly to the unsaturated ketone; both the addition and the
dehydration involve enolization.

O ©OH
I'
* "
^ OH—
(
I

C 6 H 5 CCH 3 + H^^ZZ! C6H5 C— CH ^Zl 3 C 6 H5-C=CH 2

•

©;Z^ I ©
C6H 5 CH=0 + uHW >
c 6 H 5 CH =OH
OH OH \ 1 ©OH
\
C 6 H CH5
I

— CH =C I

CeHo^ ; C 6 H CH5
— CH»— C — C H II

6 5

OH
©OH, C'OH ©OH
V*| 2
C 6 H 5 CH
I

— CH=C— C H 6 5 > C 6 H 5 CH=CH — C-C II

6 H:

O
II

C6H5 CH=CH— C— 6 H5

The reaction is written out in some detail, so that it can be seen

that the acid plays several roles. First of all, it catalyzes enoliza-
tion; it cannot change the equilibrium amount of enol present, of
course, since the ketone and the enol have the same number of
protons. Second, the acid helps make
the benzaldehyde more
reactive to addition, and it also makes the hydroxyl a better
leaving group by protonating it. Under some conditions the
addition step (numbered 1 in the mechanism) is rate-determining,
190 ORGANIC REACTION MECHANISMS

with dehydration being rapid, while in other cases the early steps
are all rapidly reversible with dehydration (numbered 2) being
rate-controlling. However, the kinetics does not change, since
the same rate expression is expected for each situation.

Rate = k(C 6 H 5 COCH 3 )(C 6H 5 CHO)(H +

)

It will be seen that this rate expression contains the same atoms as
are found in the transition state for either step 1 or step 2. This is

another example of the problem of kinetic equivalence which was

discussed in Chapter 2.

Some important complex addition reactions involve the forma-

tion of Schiff bases from the reaction of ketones and aldehydes with
amines.

R"
/
O
If

C — R'+R"-NH 2 ;=± R— C — R'fH 2

With ordinary amines in aqueous solution this equilibrium favors

the starting materials, but when the amine component is hydroxyla-
mine or a hydrazine derivative then the imine is stabilized by extra
conjugation. With such compounds the reaction goes essentially

O O
© .. II e II

•OH OH NHCNH 2 NHCNH.2

N
/ N
/ N
/
N
//

<-->
R— C — R' R— C — R'
11

R — C—
II

R'
1

r — c— R < >
1
,

©
Oxime Semicarbazone

to completion and carbonyl derivatives —oximes, hydrazones,

semicarbazones —may be isolated.
Oxime formation is a two-step process, addition to the carbonyl
followed by dehydration, and in discussing the rate of oxime
formation we must know which step is rate-controlling. At a pH
REACTIONS OF CARBONYL COMPOUNDS 797

above 4 or 5 apparently addition is rapid and dehydration is slow;

thus when hydroxylamine is added to furfural the characteristic
ultravioletspectrum of furfural rapidly changes to that of the
adduct, and this only slowly goes on by dehydration to form the
oxime.

IO ACHO + H 2 NOH >_

I
O A CH
/ Slow
>

\
NHOH
CI CH = NOH
The initial reactions are just nucleophilic additions of free hy-
droxylamine, while the dehydrations are acid-catalyzed.

CC— CH O
— C— CH
I
OH
— C — CH
I
H
CH 3 — 3 CH 3 3 CH 3
—
S

4
©NH 2
<
NH <

:nh 2 I \
|
OH OH
OH
JbH,
C Slow
— C— CH
|

CH 3 — C— CH 3 CH 3 3
<
CH 3 — C— CH S

I II II

C7,
:nh ©nh n
\ \ \
OH OH OH

Accordingly, it would be expected that oxime formation should be

faster in strong acid. Actually the rate does increase on going from
pH 7 down to pH 5, but then it drops off again as the pH is further
decreased (Figure 6-1).
What happens is as follows : As the pH is lowered the dehydra-
tion step gets faster and faster, but the first addition step slows
down. The addition requires free hydroxylamine, and in strong
acid the hydroxylamine is mostly protonated so that the concentra-

tion of free hydroxylamine becomes very small. Thus below pH

5 the addition step becomes rate-determining. The addition is of
192 ORGANIC REACTION MECHANISMS

Rate

2 3 4 5 6 7 8

pH

FIGURE 6-1
The rate of oxime formation from acetone as a function of pH.

course slower still in stronger acid because of the low concentration

of free hydroxylamine, so the over-all rate falls off at low pH.
For this reason oximes and other carbonyl derivatives are often
made with an acetic acid catalyst, which produces more or less

the optimum pH.

6-3 Reactions of Carboxylic Acids and Their Derivatives

In this section we shall deal first with interconversions among

acids, esters, amides, acid chlorides, etc., and then we shall take up
more specialized topics such as condensation and decarboxylation.
Although perhaps the esterification of a carboxylic acid should be
considered first, the principle of microscopic reversibility tells us
that the mechanism of esterification is just the reverse of that for
ester hydrolysis under the same conditions. Accordingly we shall
deal with ester hydrolysis, since this is the process to which most
study has been given.

The hydrolysis of esters. When an ester is treated with

aqueous acid it hydrolyzes. The reverse process is Fisher esteri-
and the equilibrium may be displaced in either direction
fication,
depending on whether water is present in excess (for hydrolysis) or
whether it is removed from the reaction in some way (as is often
done in esterifications).
REACTIONS INVOLVING FREE RADICALS 793

h
if 8
RC— OR' +H 0;ZZ! 2 RC + R'OH

OH

Esters may also be hydrolyzed with base. This process, called

saponification ("soap-making"), is generally not reversible since the
carboxylic acid forms a salt which displaces the equilibrium.

O O
II
e
RC— OR' + OKT > RC
S + R'OH

In principle the cleavage of an ester could occur in two different

places: between the acyl group and the oxygen, or between the
alkyl group and the oxygen. Acyl-oxygen fission is most common,
although we shall see some cases in which alkyl-oxygen fission
occurs.

Alkyl-oxygen
O fission

R-C^OJ-R'
Acyl-oxygen
fission

The most convincing evidence for acyl-oxygen fission in ordinary

comes from studies with O 18 Thus hydrolysis of
ester hydrolysis .

methyl succinate with H 2

18
yields ordinary methanol, the heavy
oxygen being found in the succinic acid.
For simplicity the heavy oxygen is written in the OH group of the
product carboxyl, although of course rapid proton exchange makes

O O
C— CH CH — C — OCH +H —
II II

H0 2 2 2 3 2
18
> H0 C 2 CH 2 CH C
2

18
H

+ CH3OH
294 ORGANIC REACTION MECHANISMS

the two carboxyl oxygens experimentally equivalent.

Similar experiments have been done for a number of other
simple esters. One of the most interesting was the early demon-
stration that Fisher esterification of benzoic acid with labeled
methanol resulted in the formation of labeled methyl benzoate; by
the principle of microscopic reversibility this demonstrates that the
reverse reaction, hydrolysis, also involves acyl-oxygen fission.

O + O
C 6 H 5C
//
+ CH 3 18
H
H
> C HC
6
S 5 + H2
\ \
OH Q CH18
3

Saponification of ordinary esters also involves acyl-oxygen fission.

For instance, saponification of labeled ethyl propionate affords
labeled ethanol.

O O
II

CH 3CH 2 C— 18
CH 2 CH 3 + OH
.
CH CH C
3
/ 2 + CH 3 CH 2 18
H

\
We have emphasized that acyl-oxygen fission is involved in the
hydrolysis of "ordinary" esters, so we should mention those cases
in which alkyl-oxygen fission occurs instead. Since alkyl-oxygen

CH 3
Optically active

I
CH 3
Racemic
REACTIONS OF CARBONYL COMPOUNDS 195

fission in an ester is simply nucleophilic substitution on the alkyl

group, the structural features which make this process especially
easy will be factors of the sort already discussed in Chapter 3. For
instance, optically active p-methoxybenzhydryl phthalate (I) is
hydrolyzed to optically inactive benzhydrol even in ION NaOH.
The racemization shows that S N 1 substitution is occurring on the
alkyl group, and it competes with the ordinary saponification
mechanism even in this very strong alkali.
The S N 1 process is highly favorable because an excellent
carbonium ion is formed, but a number of other esters undergo this
type of hydrolysis in neutral solution. As an example, a-phenethyl
phthalate hydrolyzes by an S x l mechanism in water, although with
base the ordinary saponification mechanism takes over.

CH —OH

Racemic

with retention of
optical activity

Saponification

Since S N 1 processes are often facilitated by protonation of the

leaving group, it might be expected that alkyl-oxygen fission could
also occur in some acid-catalyzed ester hydrolyses. This is so
when a reasonably stable carbonium ion may be formed, r-butyl
acetate hydrolyzing by an S N 1 process.

CH 3 O
?
CH — — OCCH3
3
?i
+H2O 18
,©
+ CH 3 -C—
I

18
+
II

HOCCH3
3 C H

CH 3 CH,

In general, only good S N 1 reactions can compete with the

ordinary acyl-oxygen fission. S N 2 reactions are too slow, so
196 ORGANIC REACTION MECHANISMS

alkyl-oxygen fission is not seen with methyl esters, for instance.

An however involved in the formation of dimethyl
S N 2 reaction is

ether by the reaction of methoxide ion with methyl benzoate.

+ CH 3 OCH ;

In this case the much faster acyl-oxygen fission reaction is not

detected since it simply regenerates starting material.
Returning to the ordinary acyl-oxygen fission reaction of esters,
one could imagine two possible mechanisms. In one there is direct
displacement of the alkoxide group by hydroxide ion (in saponifi-
cation), while in the other there is addition of the hydroxide to
carbonyl to form a tetrahedral intermediate which then loses
alkoxide. Labeling experiments with O have been used to show
18

that a mechanism involving a tetrahedral intermediate is the

correct one.

Q18 O 18

_
— C — OCH CH
II I

C6H5 C— OCH CH + 2 3 OH > C6H5 2 3 >

OH I

O .O

C 6 H 5C
II

+ OCH2CH3 > C6H5 C

# + CH3CH2OH

OH cP

The experiment was performed by running the hydrolysis only

part way, and then examining the recovered ester. If the tetrahe-
dral intermediate (I) is formed reversibly, and if protons rapidly
come on and off hydroxyl groups (as they do), then one might
expect O 18
to be lost from the ester. This was the result actually
observed.

~
O 18 18
H OEt

C 6 H 5 -C
I

— OEt ;=* C6H5

I

-C— OEt C HC
6
/ +^0
,0
5
18
H
UU
.

OH
%O
I
REACTIONS OF CARBONYL COMPOUNDS 797

The amount of O 18
exchange observed depends on the relative
rate at which the tetrahedral intermediate (I) loses hydroxide ion
versus ethoxide ion. In this particular case hydrolysis is 1 1 times
as fast as oxygen exchange in the recovered ester, so that most of
the tetrahedral intermediate molecules lose ethoxide ion.
The same O 18
exchange was observed for the acid hydrolysis of
ethyl benzoate, showing that it also involves the formation of a

tetrahedral intermediate.

O u+ ®OH
H
— oEt
II ||

c 6 h 5c —oEt + h o ;=z 2 c 6 h 5c ;

OH OH
C6H5 — C— OEt <
>
C6H5 -C — OEt 7=1
0% OH
OH ®OH
H //
C6H5 —C— I

OEt JZZ» C 6 H 5 C + EtOH l

I © \ OH
OH
O
//
C 6 H 5C + EtOH

OH

The mechanism is written out in great detail, showing each indi-

vidual protonation step, to illustrate the fact that it is completely
symmetrical. Run backwards it is an acid-catalyzed esterifi-

cation, and it is apparent that the hydrolysis and the esterification

have exactly the same mechanisms. The oxygen exchange occurs
in the reversible formation of the tetrahedral intermediate.

O 18 18
H O
||
Several
C 6 H 5 COEt <
;
C 6H5 — C— OEt \ \ C 6 H 5 COEt
Steps
H i

+ OH +
18
H2
H2
198 ORGANIC REACTION MECHANISMS

The hydrolysis of other carboxyl derivatives. Amides may

be hydrolyzed by either acid or base, although acid hydrolysis
is usually more rapid. In the basic hydrolysis a tetrahedral
intermediate has been demonstrated by O 18
exchange, and the
exchange is actually faster than hydrolysis, NH 2
~
being such a poor
leaving group.

° 18
^ "u - oe °
Q
— NH — NH +Q
U
11 J II

C 6 H 5C 2 ;=! C 6 H 5C 2 ^Zll C 6 H 5C 18
H

OH NH 2

I
o o
/ + WNH
//
C HC
6 5

\ H
2 > C 6 H 5C
V NH 3

The acid hydrolysis probably also proceeds through a tetrahedral

intermediate, but in this case no O 18
exchange is found in the re-
covered amide since a much better leaving group, neutral ammo-
nia, is involved in the forward reaction.

O ®OH OH
II II

— C — NH
I

C 6 H 5C — NH 2 > C 6 H 5C > C6H5 2

>

NH 2 OH 2

OH ®OH O
C6H5 -C— NH
I
©
3 > C H
#
6 5 + NH 3 > C HC
6
#+
5
NH
©
4

I \ \
OH OH OH
Some oxygen exchange has even been found in the neutral
hydrolysis of benzoyl chloride and of benzoic anhydride, showing
compounds also form a tetrahedral intermediate. How-
that these
ever,most of the intermediate goes on to product, since chloride
ion and benzoate ion are such good leaving groups.

O OH
I! I

C6H5 C— CI + H 2 0^=t C 6 H 5C — CI >

OH
REACTIONS OF CARBONYL COMPOUNDS 199

®OH O
c 6 h 5c + cr > C 6 H 5C

OH OH
O O OH O
II II
II

— C— OCC H
I

C6H5 C— OCC H 6 5 ;z=? C6 H5 6 5

>

OH

2 C6H C 5

OH
It should be noted in passing that we can explain why the
postulated tetrahedral intermediates in these cases give only a
small amount of exchange, but we cannot rule out the possibility
that a tetrahedral intermediate is not formed at all in the cases in
which exchange is not observed. Even the observed exchange
reactions could be side processes unrelated to the hydrolysis
mechanism. The carbonyl addition mechanism seems an attrac-
tive idea, but it may not be correct in all cases.
The Claisen condensation. When ethyl acetate is treated
with sodium ethoxide it is transformed into acetoacetic ester.

This condensation reaction between two molecules of an ester is

called the Claisen condensation.

o o o o o
II NaOEt || || OEt" |
||

2CH 3 COEt > CH CCH,COEt

3 > CH 3 C COEt
% CH /
Since a /3-ketoester is quite acidic, the product actually formed is

the enolate salt. The reaction involves attack by the enolate ion
of one ethyl acetate molecule on the carbonyl of a second; the
resulting tetrahedral species then loses ethoxide ion.

o eo
= C/
II ©
U
CH 3 C + OEt ,
- CH,
\ \
OEt OEt
200 ORGANIC REACTION MECHANISMS

oe O
CH 3 — c— OEt , ! CH --C
3
1

— CH 2 C
II

OEt

/ OEt OEt
/
=rC
/
CH 2

r °

o o
CH.3— C
II

— CH COEt
2
II

As shown, the process is reversible and the reaction is successful

only because the product is converted to a stable enolate, so that
the equilibrium is displaced.
Decarboxylation of acids. The loss of a carboxyl group
is similar to the loss of a proton.

o O
// II

R— i- c -> > R
K c H®
II

O—H o

A H R© H®

In both processes the electron pair which originally bound the

carboxyl or proton to R must stay behind; decarboxylation will
occur easily only when the resulting species is stabilized.
An obvious example is found in the decarboxylation of /?-

ketoacids such as acetoacetic acid. Here kinetics shows that both

the free acid and the carboxylate ion can decarboxylate.

Rate = kx (CH 8 COCH,COOH) + k2 (CH 3 COCH2 COO)

From the carboxylate ion an enolate results, while the free acid
affords the enoi of acetone.

v^ll
o o o
CH.3 —C C— O > CHs-C + C0 2 > CH 3 CCH,
CH 2 CH 2
REACTIONS OF CARBONYL COMPOUNDS 201

OH o
II

CH S
,c=o CH 3 — + CO, CH CCH3 5

CH 5
^ CH
S

These are processes in which the carboxyl group is lost first, and
then the proton adds in its place. In principle there are two other
possibilities as well; a proton could simultaneously replace a
carboxyl group, or it could come in before the carboxyl is lost.

There is no known example of the first reaction, which would be a

bimolecular electrophilic substitution, although in organometallic
chemistry bimolecular electrophilic reactions are common.

O
II
©
R-H+C+H
II

O-H O
H©

On the other hand, a number of examples are known in which

CO, H

protonation precedes decarboxylation. The decarboxylation of

anthracene-9-carboxylic acid is such a case; it will be seen that this
is really electrophilic substitution on an aromatic ring.
In Chapter 7 we shall also come across an example of free radical
decarboxylation; in Special Topic 6 we shall mention an example
of oxidative decarboxylation.
202 ORGANIC REACTION MECHANISMS

General References

C. Gutsche, The Chemistry of Carbonyl Compounds (Prentice-Hall,

Englewood Cliffs, New Emphasizes synthetic meth-
Jersey, 1967).
ods rather than evidence on mechanisms.

J. Hine, Physical Organic Chemistry (2nd ed., McGraw-Hill Book Co.,

New York, 1962), Chapters 10, 11, 12, and 13. The best general
reference for this material.

M. Newman, "Additions to Unsaturated Functions," in M. Newman,

ed., Steric Effects in Organic Chemistry (John Wiley & Sons, New
York, 1956). A good discussion of mechanisms which emphasizes
steric hindrance effects on reaction rates of ketones, esters, etc.

M. Bender, "Mechanisms of Catalysis of Nucleophilic Reactions of

Carboxylic Acid Derivatives," Chemical Reviews, 60, 53 (1960).
There good review of the mechanisms of hydrolysis of
is a esters in
most of it is concerned with catalysis, including
this article; catalysis
by enzymes and models for enzymes.

R. Bell, The Proton in Chemistry (Cornell University Press, Ithaca,

New York, 1959). The later chapters discuss enolization, and
provide useful rate data.

D. Cram, Fundamentals of Carbanion Chemistry (Academic Press,

New York, 1965). Some of the carbanions discussed are enolate
ions.
 203

6
Special Topic

OXIDATION-REDUCTION REACTIONS 1

oxidations and reductions occur quite generally throughout

organic chemistry, but it seems appropriate to examine them
formally at this point since so many of the interesting cases
involve carbonyl compounds.
It is actually rather difficult to define oxidation in a satisfactory
way. It is often considered to be electron removal, but this

concept is not useful when atoms as well as electrons are changing

place. The following example will illustrate the dilemma.

OH OH
OH"
CH 2 =CH + 2 Br 2 > CH 2 — CH 2 >
|

CH 2 — CH
I

2 + 2Br~
I I

Br Br

1 (a). The best general review available in this field is R. Stewart,

Oxidation Mechanisms (W. A. Benjamin, York, 1964). (b). New
Review articles at an advanced level are found in Oxidation in
Organic Chemistry, K. Wiberg, ed. (Academic Press, New York,
1965).
204 ORGANIC REACTION MECHANISMS

The over-all process is clearly an oxidation-reduction, for the

originalbromine molecule has been reduced to two bromide ions.
The second reaction written is just nucleophilic displacement, so it
is presumably in the first step that the oxidation-reduction occurs.

This is sensible if we adopt the arbitrary idea that in a carbon-

bromine bond the electrons already "belong" to the bromine, for
then the first reaction is indeed an electron transfer. However,
this definition does depend on quite an arbitrary decision, and it is

apparent that this particular oxidation reaction is a perfectly

normal ionic addition to a double bond.
We have already discussed a number of oxidation mechanisms
in Special Topic 4. Thus, ozonization of a double bond,
epoxidation of an olefin with a peracid, and addition of osmium
tetroxide to an olefin are oxidation reactions as well as being cyclo-
addition reactions. A similar cyclic mechanism has been demon-
strated for the hydroxylation of olefins by alkaline permanga-
1S
nate; labeling experiments show that both oxygens of the re-
sulting glycol come from the permanganate ion. 2

^0
n r^^°\ Mno 18
H
+ Mn0 18
4 >
\ / 2
-\ / _-0 18 H
+
MnV

The intermediate cyclic manganate ester cannot be isolated. It

rapidly hydrolyzes to the glycol in strong base, although at lower

pH it may be further oxidized by more permanganate ion and the
3
yield of glycol is lowered.
It must not be thought that organic oxidations and reductions
never involve simple electron transfer. Oxidizing agents such as
ferricyanide ion can remove one electron from the hydroquinone
dianion to form semiquinone, while alkali metals can add one
electron to a variety of unsaturated systems.

2. K. Wiberg and K. Saegebarth, "The Mechanisms of Permanganate

Oxidation. IV. Hydroxylation of Olefins and Related Reactions,"
Journal of the American Chemical Society, 79, 2822 ( 1957)
3. Ref. l,p. 62.
OXIDATION-REDUCTION REACTIONS 205

.0

+ Fe'"(CN)6 > + Fe»(CN) 6

O K

-> C6H5 — C—I

6 H5
©
The free radicals in the cases shown are stabilized by conjugation.
4
Furthermore, in Kolbe electrolysis an electron is simply removed
from a carboxylate ion, although the resulting radical then under-
goes further changes.

O o
//
CH 3 C -> CH
V-n 3 —C — > CH 3 -
h CO,
o o-

2 CH 3
-
CH 3 — CH 3

However, since most work on mechanisms has concerned two

general groups of reactions —
hydride transfers, and oxidations by
inorganic compounds —
we shall discuss examples of these in more
detail.

Hydride Transfer Reactions

The reduction of a carbonyl group by lithium aluminum hydride

0° OH
I

— C— R
I

R— C— R > R > R— C—
/ I
I

/ H H

to.
H— B^H BH 3

4. Ref. l,p. 128.

206 m ORGANIC REACTION MECHANISMS

or sodium borohydride involves transfer of a hydride ion. 5

Hydride ions can also be donated by organic molecules. Thus,
treatment of benzaldehyde with strong base causes an oxidation-
reduction, the Cannizzaro reaction.

//
C6 H5 —C + OH ^Zl C6H5 -C-OH
\ /'
H / H

O H

— C^ —C—
I

C6H5 +
,

C6H5 C 6 H 5C02 + C 6 H 5CH 2OH

X !

oh or

As required by this mechanism the reaction is second order in

benzaldehyde and first order in hydroxide ion; when the reaction is

run in D2 as solvent the benzyl alcohol formed still has only

hydrogen on carbon, as is expected. With furfural, the reaction
becomes fourth order kinetically, second order in both aldehyde
and hydroxide. G Here the carbonyl is less reactive, and a stronger
hydride donor is required.

U^W OH
OH
C—
I

'0
OH

IXc— 'I

CH

C0 ©
+
2
X CH 2 OH
O
II

5. N. Gay lord, Reduction with Complex Metal Hydrides (Inter-

science Publishers, New York, 1956)
6. J. Hine, Physical Organic Chemistry (2nd ed., McGraw-Hill Book

Co., New York, 1962), p. 267.

OXIDATION-REDUCTION REACTIONS 207

Any alkoxide ion can be a hydride donor

if there is an adjacent

hydrogen. However, aluminum alkoxides prove to be particu-

larly effective in reducing other carbonyl compounds. The use of
a compound such as aluminum isopropoxide to catalyze the
reduction of benzaldehyde by isopropanol is an example of the
Meerwein-Ponndorf reaction, if it is considered to be a reduction of
benzaldehyde, or the Oppenauer reaction if it is considered to be an
oxidation of isopropanol. The reaction probably involves a
7
cyclic transition state with transfer of a hydride ion.

OH O
AI(OiPr) 3 ||

— CH— CH
I

C 6 H 5 CHO +CH 3 3 <

)
C 6 H 5 CH 2 OH + CH3CCH3

Al

OO
Al

( O CH 3

^1/ — C—H
II

—
I

C<^. C C6H5 C
C 6H5

H
I
^/
H
\ I

H CH
/ \
CH
CH 3 3

A similar cyclic transition state is involved in the abnormal

Grignard reaction, a reduction which occurs with some hindered
8
ketones.

CH 3 O CH 3
CH 3 — CH — CH MgBr
2 +
\ CH — C— CH II

CH 3
\
CH 3

CH 3 OH CH,

CH— C-CH + CH 3 — CH =CH 2

CH
/ I

H
\ CH 3 + Normal adduct
3

7. Ref. l,p. 19.

8. M. Kharasch and O. Reinmuth, Grignard Reactions of Non-
metallic Substances (Prentice-Hall, New York, 1954), p. 147.
208 ORGANIC REACTION MECHANISMS

Br
I

Mg
/
O
CH 2
—C—H +
|

> II

CH
I

CH 3

These have been examples of hydride transfer within a com-

plex; other cases areknown in which true intramolecular hydride
transfer occurs. One case which has been studied is the rearrange-
6
ment of phenylglyoxal to mandelic acid.

O OH
— CH — C0 OU
11

C6H5 — C-CHO+OH^ > C6H5

I

When this reaction is run in D 2 no carbon-bound deuterium is

found in the product (the mandelate anion has an enolizable

hydrogen, but exchange next to a carboxylate ion is quite slow).
Thus the mechanism involves an intramolecular hydride shift, as

shown.

O (O OH
C6 5
II
Q_
H C— CH = 0+OH"t=^ C6H5
I
^0
—NlC — C^MD
O OH OH H

C H C —C
/ I

> C6 H CH — C0
I

U 2
6 5 5

H
^O
v

Such a reaction is very similar to the benzilic acid rearrangement,

9
in which a phenyl migrates when benzil is treated with base.
In fact, the rearrangement of phenylglyoxal to mandelic acid

9. S. Selman and J. Eastham, "Benzilic Acid and Related Rearrange-

ments," Quarterly Reviews, 14, 221 (1960).
OXIDATION-REDUCTION REACTIONS 209

o o
HO—
^1 IK
—
C 6 H 5 C— CC 6 H 5 + OH > C 6 H £

Benzil
C6H f

O O OH
I
II I

HOC-C— C 6 H5 > (C 6 H 5 ) 2 CH- CO^

I

C6H5

could actually have involved a phenyl shift; this has been excluded
by labeling experiments using C14 .

a OH
OH° U Oxidize *
II

C 6 H 5 C*-CHO > C6H5 —CH— C0 H 2 >C 6 H C0 2 H

5 + C0 2

So far we have been examining cases in which hydride ions are

transferred, to carbonyl groups, from molecules which are espe-
cially good hydride donors. There are also cases known in which
an especially good hydride acceptor may pull hydride ion from an
otherwise unreactive molecule. The most important examples of
this type of process are found in carbonium ion chemistry. A
carbonium ion can usually abstract hydride from a hydrocarbon,
10
particularly if this leads to a better carbonium ion.

R0+H-R'^R-H + R'0
For instance, tropylium ion may be synthesized by treating cy-
11
cloheptatriene with triphenylmethyl cation. The aromatic
tropylium ion is a much better carbonium ion, so the reaction goes
to completion.

10. N. Deno, H. Peterson, and G. Saines, "The Hydride-Transfer Re-

action," Chemical Reviews, 60, 7 (1960).
11. H. Dauben, F. Gadecki, K. Harmon, and D. Pearson, "Synthesis
ofTropenium (Cycloheptatrienylium) Salts by Hydride Exchange,"
Journal of the American Chemical Society, 79, 4557 ( 1957)
 ORGANIC REACTION MECHANISMS

H H

G
(C 6 H 5 ) 3 CH + (/ \ BF 4

Oxidation by Inorganic Compounds

We have already mentioned the oxidation of olefins by ozone,

Mn0 4 Os04
~, etc. One of the most useful oxidizing agents for the
,

conversion of alcohols to ketones is chromic acid, commonly used

in acetic acid solution (although another useful combination is
pyridine and Cr0 3 ). Westheimer and his research group have
studied the oxidation of isopropyl alcohol to acetone by chromic
acid solutions. The mechanism involved is apparently as fol-
12
lows.

O OH
CH 3 CH 3 O^Cr
/ cJI

\ CH— OH+HCr0 \ C A ® \ OH
4 + 2H ^Zt >

/ CH 3
/ XH
CH 3
* — : b
CH 3 OH
\C=0+BH
© /
+ Oz=Cr
/ \
CH 3 OH

A chromate ester is reversibly formed, and in the rate-determining

IV
step there is an elimination with loss of a Cr species. Through a
series of other rapid reactions this is converted to Cr
m , the stable
final oxidation state.
Evidence that a proton is being removed in the rate-
determining step of the oxidation, as the mechanism indicates, is
found in the kinetic isotope effect observed when deuterioiso-

propanol, (CH 3 ) 2 CDOH, is oxidized. The deuterio compound is

6.6 times slower at 25 °C than normal isopropanol; such large

12. F. Westheimer, "The Mechanisms of Chromic Acid Oxidations,

Chemical Reviews, 45, 419 (1949); cf. also ref. l,p.33.
OXIDATION-REDUCTION REACTIONS 211

13
isotope effects always indicate that the bond to hydrogen is being
broken in the transition state. Another piece of evidence in favor
12
of this mechanism is the observation that diisopropyl chromate
willundergo base-catalyzed elimination with pyridine, a process
very similar to the elimination written for the normal reaction.

CH 3 O^Cr —O CH 3 O
\ A
C A
II

O
\ CH/ >
N
CH3CCH3 + Cr IV Species

CH 3 H„ CH 3

3
Aldehydes may also be oxidized with chromic acid, and the
mechanism is apparently quite similar to that already discussed."

OH OH
^— :b /
A-
RCHO + H^)-HCro ^=? 4
w
R
I

— O-rH > R— C + Cr IV ,etc.

OO O
C*i OH

An was observed for the

isotope effect on the reaction rate of 4.3
oxidation of C 6 H 5 CDO.
Again this 4-fold slower reaction for the
deuterium compounds shows that the proton is being removed in
the rate-determining step. Many saturated hydrocarbons can also
be oxidized by chromic acid, as well as by permanganate ion. 15 In
these cases apparently a hydrogen atom is removed to form a free
radical, which undergoes further oxidation by the inorganic spe-
cies.

When 1,2-diols are treated with periodic acid they are cleaved

13. K. Wiberg, "The Deuterium Isotope Effect," Chemical Reviews,

55,713 (1955).
14. Ref. l,p. 48.
15. Ref. l,p. 50.
212 ORGANIC REACTION MECHANISMS

16
oxidatively. Thus ethylene glycol affords formaldehyde and
iodic acid, and the mechanism involves formation of a cyclic
periodate.

CH 2 -CH 2 + |O > CH2-7— CH, CH 2 CH 2

I
I
4
\*S
O
k > II

O
II

OH OH O ) O

c/HV
O 18
labeling experiments show that the carbonyl oxygens of the
products come from the original glycol oxygens. One of the pieces
mechanism is the fact that glycols
of evidence in favor of the cyclic
for which such a cyclic ester would be impossible, such as trans-
9,10-decalindiol, are unaffected by periodate.

OH

OH
a-Diketones are also cleaved by periodate, and again a cyclic
mechanism seems probable.
Labeling experiments show that the two new oxygens in the
product acetic acid molecules in fact come from the periodate, as
this mechanism suggests.

OO OH
II OH
CH3— C — C — CH — C-r-C — CH
II II

3 CH3 3 2CH 3 C0 2 H +
\^ K
HO OH > O O) > |v Species
\ /
I —OH
A
^lr- OH
/-^
o o o o

Lead tetraacetate can also cleave glycols, although it is less

16. Ref. 1, Chapter 7.

OXIDATION-REDUCTION REACTIONS 213

16
selective than periodate. There are some data in favor of a
cyclic mechanism for this reaction as well, but the cleavage of
/rarcs-9,10-decalindiol shows that a noncyclic mechanism is also
possible.

-C-^O^i -c=o
K Pb(OAc) 2 >

— C—
+Pb(OAc) 2

—C^O\J

+ Pb(OAc) 2 + AcOH

Thus it is apparent that for many of these inorganic oxidations a

standard mechanism is involved: is formed
an inorganic ester
between the oxidizing agent and the substrate, and then an ordinary
elimination reaction or cyclo-decomposition process leads to prod-
ucts. However, when the substrate is a hydrocarbon rather than
an alcohol or ketone then free radical chain processes become more
important. Such processes are discussed in the next chapter.
 REACTIONS INVOLVING FREE RADICALS

a free radical any species which has an unpaired electron.

is

Under this definitionone would include not only such molecules as

NO and N0 2 (which have an odd number of electrons so that one
must be unpaired) but also atoms such as I- and Na-. 2 is also

included since it has two unpaired electrons and is thus a biradical.

A few radicals, such as the nitrogen oxides above, are stable
species; most free radicals are not stable, but they may still play an
important role as reactive intermediates. Typically, reactions at
high temperatures, including reactions in flames, involve free
radical intermediates. Furthermore, many organic reactions at
ordinary temperatures go by free radical mechanisms; this is

particularly true of reactions in the gas phase or in nonpolar

solvents. Before we consider such reactions in detail we will

mention briefly the factors which can lead to the stable existence of
organic radicals.

7-1 Stable Organic Radicals

Hexaphenylethane is in equilibrium with the triphenylmethyl

radical in solution.

214
REACTIONS INVOLVING FREE RADICALS 215

C6H5 C6H5
\ K equil
— C— C— c h
|

C H5ff 6 5;
>
2 (C 6 H 5 ) 3 C-

C6H 5 C6H5

Although this equilibrium is quite unfavorable (a 0.1M solution of

hexaphenylethane is 2% dissociated at 20°C), the triphenylmethyl
radical is easily detected in solutions of hexaphenylethane, and the
tris-p-nitro derivative exists almost completely dissociated.
A major part of the stabilization of these radicals comes from
extra resonance forms which are not available in the dimers.

The participation of the nitro groups in delocalizing the odd

electron explains the greater stability of tris-p-nitrophenylmethyl
radical. Dissociation is also favored by steric crowding in the
dimer, the hexaphenylethane. The three bulky phenyl groups on
one carbon are jammed into the three on the other carbon of the
ethane.
A number of other stable free radicals are known. Compounds
like diphenylpicrylhydrazyl (I) are stable chiefly because of exten-
216 ORGANIC REACTION MECHANISMS

sive derealization of the odd electron, while a species such as the

anthracene anion radical (II) fails to dimerize both because of
conjugative stabilization in the radical and because of electrostatic
repulsion of one anion molecule by another.

etc.

etc.

A free radical has a net magneticmoment. All electrons have

spin, and a spinning charge generates a magnetic field, but in
ordinary molecules the electrons are paired and the magnetic field
from those electrons with spin "up" is exactly cancelled by the field
from electrons with spin "down." Since a radical has an odd
number of electrons there will be one net spin uncompensated, so a
free radical generates a magnetic field. Thus free radicals are
attracted by other magnetic fields, and one way of showing that a
solution contains free radicals is to show that it is attracted by a
magnet. Unfortunately, this method is not very sensitive, and
only relatively large concentrations of radicals can be detected.
Much more sensitive is the method of electron spin resonance
spectroscopy, often called e.s.r. (or e.p.r., electron paramagnetic
resonance) for short. This depends on the fact that a free radical
in a strong magnetic field will preferentially orient its odd electron
spin in the more stable direction (as a compass needle aligns in the
earth's field) but with the absorption of light energy (with the
usual magnetic fields, microwave frequencies are used) the spin
REACTIONS INVOLVING FREE RADICALS 217

can turn over to the unstable orientation, as illustrated in Figure

7-1.

N N

energy

N
electron

applied field

Stable orientation Unstable orientation

FIGURE 7-1
Illustration of the principle of e.s.r.

This is a very sensitive method of detecting radicals, even at

7
concentrations of 10' M.
Using e.s.r. it has been possible to detect the intermediate
radicals in a free radical polymerization. Furthermore, even very
unstable radicals such as CH 3
-
can be detected using e.s.r.; con-
veniently such reactive radicals are trapped at — 200 °C in a frozen
and be destroyed. Using e.s.r.,
solution so that they cannot react
many which were formerly only postulated intermediates
radicals
have now been directly observed.

7-2 Bond Dissociation Energies

For radicals which are too unstable to exist at equilibrium,

information on relativestabilities can be obtained from bond

dissociation energies (the energy required to break a covalent

bond and form two radicals). Table 7-1 contains values for
some molecules of interest; these are determined in a few cases by
direct spectroscopic measurements, but in most instances by more
complex indirect means.
It is apparent that the relative stabilities of different radicals
cannot be derived from this table without the exercise of some
care. Thus F 2 has a low dissociation energy compared with Cl 2 ,
218 ORGANIC REACTION MECHANISMS

TABLE 7-1

Some Bond Dissociation Energies (Kcal/Mole at 25°C)

Dissociation Dissociation
Bond Energy Bond Energy
CH,— 102 H— 104
CH,CH,— 98 CI— CI 58
(CH,),CH— 94 Br— Br 46
(CH,),C— 90 F— 37
CoHnCH,>— 78 I— 36
CcH 5 — 102 HO— OH 52
CC1.3— 90 OoN—N0 2 13
O= CH— 78 H,N— NH 2 60

HO— 120 CH.,— CH 3 84

HOO— 90 CH — OH
3 90
F— 135 CH,— CI 82
CI— 103 CH.,— Br 67
Br— 87 CH,— 53
I— 71
CH,S— 89

suggesting that the fluorine radical (atom) is more stable than the
chlorine atom. On the other hand, HF has a higher dissociation
energy than does HC1, suggesting that a chlorine atom is more

stable than a fluorine. The reason for this dilemma is that we

have ignored the properties of the bond which is being broken; the
single bond in HF is much more stable than the single bond in F 2 ,

while the HC1/C1 —

CI difference is smaller. Bond dissociation
energies can be used as an indication of radical stabilities only if a
very similar bond is being broken in all the cases compared. This
is approximately true for the first five hydrocarbons in the table,
and to a lesser extent for the next three cases as well in which a
C —H bond is being cleaved. Thus the values in the table show
that tertiary alkyl radicals are more stable than secondary, which
are more stable than primary. The stabilizing effect of alkyl
substituents is often ascribed to hyperconjugation.

CH 3 CH 2 H CH 3
I II I

C- < > r < > C etc.

CH
/ \ CH
3 3 CH3
/ \ CH,
CH 3
/ % CH H
2
REACTIONS INVOLVING FREE RADICALS 219

The benzyl radical is strongly stabilized by conjugation of the

odd electron with the benzene ring; trichloromethyl radical and
formyl radical are also stabilized by conjugation.

.'CI' CI ©Cl
w CI
•
\©/ -> v, j) etc.

1
1

CI CI

:o: O© •'o
II / •

/ H
->

H
/ c:

H-0-0° i
\
>
' H
n -Ot-^
©*

The same sort of stabilization is found in hydroperoxy radical,

HOO-, compared with hydroxyl radical.
Bond dissociation energies can be used to predict whether
various simple radical reactions are exothermic or not. Thus we
see that hydroxyl radical can attack methane in an exothermic
process, but that a bromine atom cannot.

HO- + CH 4 -> H,0 + CH3 - AH = - 18 kcal/mole

Br- + CH 4 -> HBr + CH3 -

AH = + 15 kcal/mole

The over-all energy change is just the difference of two bond

Breaking a methyl-hydrogen bond requires
dissociation energies.
102 kcal/mole, but forming a hydroxyl-hydrogen bond releases
120, so the over-all process releases 18 kcal/mole.
Since the reaction of a bromine atom with methane is endo-
thermic, the reverse reaction — on HBr by a methyl radi-
attack
cal — is exothermic. There are actually two ways in which CH,^
could attack HBr, to form methane or to form methyl bromide.

CH 3
-
+ HBr -> CH 4 + Br- AH° = -15 kcal/mole
CH 3
-
+ HBr-> CH 3 Br + H- AH = +20 kcal/mole
220 ORGANIC REACTION MECHANISMS

The difference between a bond dissociation energy of 67 kcal for

methyl bromide and 87 kcal for HBr shows that the second
reaction is endothermic by 20 kcal/mole and the first reaction is
thus the favored one. This is very important, for reactions which
are strongly endothermic cannot play a major role in radical chain
reactions. Chain reactions are the most important feature of free
radical chemistry.

7-3 Radical Chain Reactions

Polymerization. The free radical polymerization of sty-

rene, initiated by benzoyl peroxide, is a typical chain reaction.

o o o
II II Heat ||

C 6 H 5 CO— OCC 6 H 5 > 2 C 6 H 5 C— O-

O O
II

C 6 H 5C-0 + CH 2 =CHC 6 H 5 >

II

C 6 H 5C — OCH CH— 2 6 H5

R. + CH 2 = CHC H 6 5 > RCH 2 CHC 6 H 5

2 R- > R—
The first step is called the chain initiation step. Here a special
substance used which has a relatively low bond dissociation
is

energy; cleavage occurs not only because a weak oxygen-oxygen

bond is being broken (cf. hydrogen peroxide in Table 7-1) but
also because the product radicals are resonance-stabilized.

In step 2 this radical adds to the styrene double bond. This is a

favorable step because the product radical is also resonance
stabilized, and because a relatively strong carbon-oxygen single
REACTIONS INVOLVING FREE RADICALS 221

bond (cf. CH 3 —OH in Table 7-1) is being formed while a

relatively weak carbon-carbon n bond is being broken.
Step 3 is the chain-propagating step. It occurs over and over
again, the radical becoming longer with each addition to a styrene
molecule. This step is favorable because a new carbon-carbon
single bond is formed while a carbon-carbon double bond is
broken; the single bond is about 20 kcal/mole stronger than a

7T bond.
Eventually two of the growing radical chains collide and react.
Then chain termination, step 4, occurs.

O C6H5 C6H5 C6H5 C6H5 O

II I I
I I II

C 6 H 5 CO(CH 2 CH) n CH 2 CH • + • CHCH (CHCH

2 2) m OC€ 6 H;

CeHs C6H5
X CH>CH CH 7

Termination has been written as coupling of the two radicals, but

it can also occur by hydrogen transfer.

CeHs CsHb CeHs C6H5

CH 2 CH- + -CHCHf >

X CH CH 2 2 + CH =CH —

There is a high probability that two such radicals will react if

they collide, and in fact it has been found that two methyl radicals
will couple to form ethane on almost every collision. Thus it will

be possible to obtain long-chain polymers only if the chain propa-

gation step is also very fast. Typically, the time between chain
initiationand chain termination is of the order of one second.
During this time all of the chain-propagating steps occur.
We have already emphasized the fact that reactions which are
strongly endothermic cannot play an important role as chain-
propagating steps. This is because a reaction must have a low
activation energy in order to be fast, and the activation energy for
any step can never be less than the over-all energy change for that
step. This is illustrated in the energy diagram of Figure 7-2.
222 ORGANIC REACTION MECHANISMS

/ Product
>sJl AG*

Starting material

Reaction

FIGURE 7-2
Energy diagram for an endothermic reaction.

It is apparent that the transition state must be at least as high in

energy as the products are, since the transition state is the highest
energy point on the path. When we recall that at room tempera-
ture each 1.4 kcal/mole of extra activation energy slows a
reaction by a factor of 10, it is easy to see that a reaction which is

endothermic by 20 kcal/mole, for instance, will be very slow

indeed, too slow to play a role as a chain-propagating step.
Although even an exothermic process could in principle have a
large activation energy, the exothermic addition of a radical to
another styrene unit is actually relatively fast, and long chains are
produced.
Many other olefins can be polymerized by radical chain
processes, including ethylene, vinyl acetate, vinyl chloride, acrylo-
nitrile, etc. In general a reactive monomer will have a terminal
double bond and addition occurs at the unhindered CH 2 to yield
the more stable substituted radical.
Substitution. Molecular fluorine is enormously reactive
towards hydrocarbons. The careful reaction of F2 with methane
yields methyl fluoride, along with further fluorination products.
For this reaction the chain propagation steps are the following:

F- + CH 4 -> HF + CH 8 -
AH = -33 kcal/mole

CH 3 + F 2 -* CH 8 F + F-
•
AH° = -60 kcal/mole
Both steps are very exothermic because the hydrogen-fluorine
and carbon-fluorine bond are strong while the F 2 bond is weak.
REACTIONS INVOLVING FREE RADICALS 223

The initiation step is probably a direct reaction of methane with

fluorine, since no other initiators are required.

CH + F
4 2 -> CHo- + HF + F- AH° = +4 kcal/mole

Chain termination comes by combination of two methyl radicals,

by combination of methyl with fluorine radicals, and perhaps by
recombination of fluorine atoms to form F 2 .

Molecular chlorine is much less reactive, although it can give a

chain process in which both steps are exothermic.

CI + CH 4 -» HC1 + CH 3
• • AH° = -1 kcal/mole

CH,- + Cl 2 -> CH3CI + CI- AH° = -24 kcal/mole

These energies can be obtained from Table 7-1 by a combination
of the appropriate bond dissociation energies. For instance, for
the reaction of methyl radical with Cl 2 the energy change is just the
difference between the 58 kcal of the Cl 2 bond and the 82 kcal of
the methyl chloride bond. Again using the data in this table, it is
apparent that a direct reaction between Cl 2 and methane is prob-
ably not the initiating step, since it would be endothermic by 57
kcal. Accordingly, free radical chlorinations are generally initi-

ated with light, although other initiators may also be used.

light
Cl 2 >2C1-
The 58 kcal/mole required to break this bond is furnished by the
lightwhich is absorbed. The energy of a quantum of light
depends on the wavelength, of course, and it can be derived from
a simple formula:

E (in kcal/mole) = —28,635

y-.
-

N
A (in mfx)

Thus blue light at 400 m/x has an energy of 71.6 kcal/mole, while
ultraviolet light at 200 mp can furnish 143.2 kcal/mole. This
energy is enough to break any covalent bond, if it is absorbed;
therefore ultraviolet irradiation is often used to initiate radical
chain reactions.
Free radical bromination is even less easy than chlorination.
Since the HBr bond has a strength of only 87 kcal, Br- can
attack only fairly weak C —H bonds. We have already seen that
224 u ORGANIC REACTION MECHANISMS

attack on methane would be endothermic by 15 kcal; on the other

hand, the hydrogen of toluene can be abstracted in an exothermic
process (cf. Table 7-1).

Br- + H—CH C H 2 6 5 -> HBr + C 6 H 5 CH 2 - AH° = -9 kcal

C 6 H 5 CH 2 + -
Br 2 -> C 6 H 5 CH 2 Br + Br- AH° = -4 kcal

Brominations of this type are usually initiated either with light or

by the thermal decomposition of a peroxide or other unstable
molecule.
Since an allyl radical is about as stable as a benzyl radical, it

would be expected that hydrogens next to a double bond could also

be replaced in a chain bromination. This is true, but olefins can
also add bromine to the double bond in a polar reaction. To
prevent this polar addition, iV-bromosuccinimide is usually used to
effect allylic bromination by a free radical process.

// //
CH 2 — CH 2 —
\
N— Br +CH =CH— CH;
I
2 NH
CH 2 —C/ CH 2 —
+ \
CH 2 = CH — CH 2 Br

Although it was once thought that this involved a special pathway,

it is now clear that the reaction of 7V-bromosuccinimide with HBr
formed in side reactions simply furnishes Br 2 in low concentrations.
Under these conditions addition of bromine to the double bond is
slow, and free radical bromination occurs instead.

// //
CH 2 — CH 2 —
N— Br + HBr
\ NH + Br,

CH 2 — CH 2 -C
\
Radical bromination
REACTIONS INVOLVING FREE RADICALS 225

Free radical halogenation with iodine is generally not possible,

since the 7 1 kcal bond energy of HI is lower than the energy of any
C—H bond.
The combustion of hydrocarbons involves a free radical chain
substitution by 2 followed by further complex changes.
, The
first step may be illustrated by the autoxidation (i.e., reaction with

molecular oxygen) of benzaldehyde.

H
/ —C
R- + C 6 H 5C > RH -hC 6 H 5 v

o o
o-o-
/
C6H5 —C + 2
> C6H5 —C
o o
O— OH

00
O— O- H
/ / /
C6H5 -C V + C 6 H 5C >C 6 H5 —C
% % %o
+ C 6 H 5 C-

A radical formed in an initiating step removes the very reactive

(cf. Table 7-1) aldehyde hydrogen, and the next two steps are
then chain-propagating. The product of this chain is perbenzoic

benzoic acid.

C 6 H 5C
OO
acid, but this reacts with

//////
\
O— OH
+
benzaldehyde

C 6 H 5C
\
H
in an ionic process to afford

> 2 C 6 H 5C
\
O

OH

Generally the step in which a radical adds to 2 is quite favorable

energetically, but in the second propagation step a relatively weak
(90 kcal) ROO—H bond is being formed. Accordingly autoxi-
dation at ordinary temperatures is facile only with relatively reac-
226 ORGANIC REACTION MECHANISMS

tive C —H bonds, such as those in aldehydes or the C —H bond

next to an ether oxygen. Allylic hydrogens are also susceptible
to autoxidation; this is part of the reason that unsaturated fatty
acid esters turn rancid in air.

Addition. The most famous free radical addition process is

the "abnormal" addition of HBr to olefins. With propylene, for

instance, the chain-propagating steps are as follows:

Br-+ CH 2 =CH— CH 3 Br — CH — CH — CH
2 3

AH°=-9kcal/mole
Br— CH 2 — CH— CH 3 +HBr > Br— CH 2 — CH 2 — CH 3 +Br-

AH°=-7kcal/mole

The over-all result is "anti-MarkownikofT" addition, i.e., the

reverse of the orientation in ionic addition to an olefin; the
orientation is bromine adds to
as expected for a radical chain since
the less hindered end of the bond to form at the same time the more
stable radical. Chain initiation may be effected with light or by the
thermal decomposition of initiators such as peroxides. Indeed, the
peroxides present in most olefin samples, because of autoxidation,
are usually sufficient to initiate these chains; olefins must be
carefully purified if the radical process is to be suppressed.

heat or
ROOH >RO + OH
light

RO +HBr^ROH + Br-
Chain termination occurs by combination of bromine atoms with
themselves or with alkyl radicals, or by coupling or disproportion-
ation of alkyl radicals.
It is interesting that polymer is never formed under these
conditions. In principle the alkyl radical could add to another
olefin molecule, as in olefin polymerization, but in practice with any
reasonable concentration of HBr the latter traps all alkyl radicals
formed before they can attack a second olefin molecule.
Neither HI nor HF has so far (1969) been added to an olefin in
a free radical chain process. To understand this we must simply
REACTIONS INVOLVING FREE RADICALS 227

consider the energy changes which would be associated with each

propagating step of the chain. Thus for the HI addition to
ethylene

I+ CH, = CH : -> I— CH>— CH,- AH° = + 7 kcal/mole

.

I_CH,— CH L + HI -» I—CHo— CHo— H + I-

>-

AH° = -27 kcal/mole

The first step is unfavorable because the carbon-iodine bond is so

weak (with substituted olefins in which a highly stabilized radical is
formed this step might become favorable). The second step is
fine, but with one slow step the chain propagation cannot compete
with termination. For HF the slow step is the second one.

F— CHo— CH,- + HF -> F—CH 2

—CU —H + F- 2

AH = +37 kcal/mole

This value is obtained from Table 7-1 simply by considering the

bond strengths in ethane and in HF.
In general, HC1 add to olefins in a radical chain
also fails to
process. Addition of a chlorine atom to ethylene is exothermic by
26 kcal/mole, but the reaction of the product radical with HC1 is

unfavorable by 5 kcal/mole, as the bond energies of ethane and

of HC1 in Table 7-1 show. Slow addition of HC1 in a radical

o o o
II
II

C 6 H 5 CO— OCC 6 H 5
II

> 2 C6H5 C— O.
O O
// II

C 6 H 5C + HCI > C 6 H 5C +CI.

\ \
O- OH
CI-+ CH 2 = CH 2 CI— CH 2 — CH2 -

CI— CH 2 CH 2 - + CH 2 =CH 2 > CI(CH 2 CH 2 ) 2

-
etc.

CI— (CH CH 2 )„-2 + HCI CKCHzOUnH + CI- n=2to>10

228 ORGANIC REACTION MECHANISMS

process is possible with some olefins, but the slowness of the second
step means that polymerization now competes with simple addi-
tion. Thus short polymers (called "telomers") are obtained when
ethylene is heated under pressure with aqueous HC1, using benzoyl
peroxide as initiator.

Thiols of all sorts can be added to double bonds in free radical

chain reactions. For instance, thiophenol reacts with styrene to
afford a simple adduct.

C 6 H 5 S- + CH 2 = CH— 6 H5 C6H5 S— CH 2 CH-C 6 H 5

C 6 H5S — CH — CH — C H
2 6 5 + C 6 H 5 SH > C6H5SCH2CH2C6H5
+
C 6 H 5 S-

Both steps are exothermic, and polymers are not formed since the
hydrogen transfer step is so rapid.
Carbon-carbon bonds can also be formed in some free radical
addition reactions. For instance, carbon tetrachloride can be
added to propylene in 80% yield.

•CCI3 + CH 2 =CH — CH 3 CI3C— CH 2 — CH— CH 3

CI

CI3C — CH — CH — CH +
2 3 CCI4 — > CI3C — CH CH — CH
2
I

3 +-CCI 3

The reaction by the thermal decomposition of diben-

is initiated
zoylperoxide; the benzoate radicals attack CC1 4 to produce a CC1 3 -

and start the chain. The first propagation step is favorable

because a carbon-carbon single bond replaces a double bond.
This overcomes the fact that the radical formed is less stable than
the trichloromethyl radical consumed (cf. chloroform and propane
in Table 7-1). no change in the types
In the second step there is

of bonds present, but a conversion of the alkyl radical to a more

stable trichloromethyl radical, so it is favorable as well.
Aldehydes can also be added to some olefins. Thus reaction of
acetaldehyde with diethyl maleate produces the adduct in 78%
yield.
REACTIONS INVOLVING FREE RADICALS 229

O H

O
CH 3 —C O
\
CH-CH 2 + CH C-
3
/
/ \
C2H5O2C C0 2 C 2H5

Even alcohols can be added, ethanol reacting with ethylene to

afford a 10% yield of the simple adduct.

OH OH
/ CH 3 CH 2 OH
CH 3 — CH- -CH 2 - -CH 2
1

CH 3 CH +CH =CH 2 2 >

-

OH OH
1 /
CH CH3 -CHo— CH + 3 CH 3 CH-

In many of these cases telomers, short polymers, are formed as

well.
Several features will be noted which are common to all these
mechanisms. After initiation steps, which may occur in various
ways depending on the reaction considered, the two steps of chain
propagation consist of (1) addition to a double bond, and (2) free
radical substitution. By contrast, polymerization consists of a
series of addition reactions, and halogenation of hydrocarbons
consisted of a chain of alternating substitution reactions. The
other significant point is that in general addition to double bonds
occurs at their least hindered end, and radical substitution in-
volves attack on unhindered atoms like hydrogen or halogen on
the outside of a molecule, rather than attack on a carbon atom
inside a molecule. The requirement that chain-propagating steps
230 ORGANIC REACTION MECHANISMS

have a fairly rapid rate means not only that they should be
exothermic, but also that there should be no serious steric barrier to
be overcome.
Free radical decompositions and rearrangements. So
far we have considered only the ways in which free radi-
cals may attack other molecules. In some cases intramolecular
reactions of radicals may also be important. For instance, we
have frequently referred to the use of benzoyl peroxide as a radical
and have suggested that it functions by furnishing benzo-
initiator,

ate radicals. In fact, this is only part of the story. Decomposition

of benzoyl peroxide by heating in cyclohexane affords both benzoic
acid and benzene, together with other products derived from the
benzoate and phenyl radicals.

o o O
C6H5
II

C— O— O— CC
II

6H5
A ) 9 r H C
I'

o
II

f u f~
C6H3V-
1

1
DU
KM •>

> C 6 H 5 C0 2 H + R-

II

C6H5C > C6H5 -

+ co 2

\
o

C6H5 - + RH > C6H6 +R.

Carbon dioxide is also found, as expected since the benzoate
radical decarboxylates to form a phenyl radical.
Similarly acetyl peroxide decomposes, at 60-1 00°C, to yield
methyl radicals almost exclusively; decomposition of the interme-
diate acetate radical occurs before it can be trapped.

CH3CO — O— O— COCH3 CH3CO— O CH3+CO2

On standing at room temperature, trimethylacetaldehyde de-

composes to carbon monoxide and isobutane.
REACTIONS INVOLVING FREE RADICALS 231

CH 3 CH 3

CH 3 — C— CHO
I

> CH 3 — C— HI

+ CO
I I

CH 3 CH 3

This is a free radical chain process, initiated presumably by traces

of peroxides in the aldehyde.

CH 3 O CH 3

3
#
CH — C — C.
I

> CH 3 —O +
I

CO
I
I

CH 3 CH 3

O CH 3 O

II
CH 3 CH 3 CH 3

CH 3 —
II
C +
I

CH 3
HC
II

— C~CH
I

CH 3
3 CH 3 -C — H-hCH 3
I

CH 3
— C-C-
I

CH 3
//

Although this is a particularly easy decarbonylation, many acyl

radicals lose COon heating, and such decarbonylations represent
attractive ways to generate alkyl radicals. For instance, heating
trimethylacetaldehyde with 1-hexene leads to 2,2-dimethyloctane
in good yield.

CH 3 CH 3
= CHC — C — CHo— CH —
I I

CH 3 — C- + CHo 4 H9 > CH 3 4 H9
I
I

CH CH /
/
3 3

CH 3

/ CH 3 -C-CHO
CH 3

CH 3 CH O CH 3

/
3

CH 3 -C— CH -CH — C4H9 +CH -C — C-

I

2 2 3
I

> CH 3
— C- I

+ CO
I I I

CH 3 CH 3
CH 3

Fragmentation also occurs when di-r-butyl peroxide is heated,

the r-butoxy radical undergoing some cleavage to acetone and a
methyl radical.
232 ORGANIC REACTION MECHANISMS

CH 3 CH 3 CH 3 CH 3
Heat \
— C— O— O— C— CH
|

— C — O*
I I

CH 3 3
> CH 3 > CH 3
-
+ C =C
CH 3 CH 3 CH 3 CH 3

Thus in the presence of solvents which can donate hydrogen to a

radical, both /-butyl alcohol and methane are formed.
Finally, some mention should be made of the rearrangements of
free radicals. These are relatively rare. Generally one does not
observe shifts of alkyl groups or of hydrogen atoms in free
radicals, in contrast to the situation with carbonium ions, but the
migration of phenyl groups (unknown)
is common. In the
migration of a methyl group, an intermediate configuration would
have to be that shown below.

CH 3
CH 2 — CH,

This is just like the intermediate geometry in a carbonium ion

rearrangement, but for the radical there are three electrons spread
over the three carbon atoms while in a carbonium ion rearrange-
ment there are only two. The situation turns out to be very
similar to that discussed in Special Topic 1, where it was seen that
the cyclopropenyl cation, with two -k electrons spread over three
carbons, was much more stable than a cyclopropenyl radical
with three ?r electrons. Detailed molecular orbital calculations
on the intermediates for carbonium ion and radical rearrange-
ments show here as well that the three electron case is much less
favorable. On the other hand, phenyl migration can occur by
bridged radical formation, without the need for "dotted line" in-
termediates.

CH 2 — CH 2 CH 2 — CH 2 CH 2 — CH,
This exact rearrangement has been detected, using C 14 as a tracer,
but only 4% of the product was derived from phenyl migration.
 REACTIONS INVOLVING FREE RADICALS 233

o
*
C6H3CH2CH2CHO
R-
>C H CH CH C-
6 5
/ —165°
*
£ C H 6 5 2
*
CH CH 2 - — >
*
-CH 2 — CH 2 C 6 H 5
2 2

RCHO

C 6 H 3 CH 2 CH 3 CH 3 CH 2 C 6 H 5
96% 4%

However, the decarbonylation of /2-phenylisovaleraldehyde affords

a mixture of hydrocarbons including as much as 80% of the
product from phenyl migration.

CH 3 CH 3 CH 3
Heat
— C-CH CHO — CH
I I

C6 H5 2
>C 6 H 5 CCH 3 +C 6H5 CH 2 + CO
Initiator I

CH 3 CH 3 CH 3

20% 80%

General References

W. Pryor, Free Radicals (McGraw-Hill, New York, 1966). The best

general introduction, with a good list of references.

C. Walling, Free Radicals in Solution (John Wiley & Sons, New York,
1957). The classic book in this field.

C. Walling and E. Huyser, "Free Radical Additions to Olefins to Form

Carbon-Carbon Bonds" and F. W. Stacey and J. F. Harris, "For-
mation of Carbon-Hetero Atom Bonds by Free Radical Chain Addi-
tions to Carbon-Carbon Multiple Bonds," in A. Cope, ed., Organic
Reactions, Vol. XIII (John Wiley & Sons, New York, 1963).
Guides to more recent work.

J. A. McMillan, "Electron Paramagnetic Resonance of Free Radicals,"

Journal of Chemical Education, 38, 438 ( 961 ) 1

B. A. Bohm and P. I. Abell, "Stereochemistry of Free Radical Addi-

tions to Olefins," Chemical Reviews, 62, 599 (1962). The stereo-
chemistry of free radical reactions is still under active investigation.
234 ORGANIC REACTION MECHANISMS

O. Dermer and M. Edmison, "Radical Substitution in Aromatic

Nuclei," Chemical Reviews, 57, 77 (1957). An interesting topic
which has not been covered in this book.

H. Lankamp, W. Th. Nauta, and C. MacLean, "A New Interpretation

of the Monomer-Dimer Equilibrium of Triphenylmethyl and Alkyl
Substituted-Diphenylmethyl Radicals in Solution," Tetrahedron Let-
ters, 249 (1968). Modern tools show that "hexaphenylethane" has
another structure!
 Special Topic

POLAR EFFECTS IN FREE RADICAL

REACTIONS

throughout the last chapter we have emphasized the fact

that chain-propagating steps must be very rapid. Even so, it is
apparent that these reactions do not in general occur with every
collision, for radical reactions can often be quite selective. In
considering the relative rate at which two competing free radical
processes might occur, e.g., addition at one or the other end of the
styrene double bond, we have explained the preference for addi-
tion at the unsubstituted end in terms of steric factors and in terms
of product stability, it being implied that the transition state for

R. + CH 2 = CH— 6 H5 > R — CH — CH — C«H

2 5

^v CH -CH-
I

C6H5

235
236 ORGANIC REACTION MECHANISMS

addition at least partly resembles the product. There is now in-

creasing evidence that polar effects also often play a role in de-
termining the stability of such a transition state.

One of the clearest examples of this is found in free radical

bromination of the methyl group in substituted toluenes. In Table
7-2 are listed relative rates, determined with Br 2 in benzene
1
solution at SCTC.

TABLE 7-2
Relative Rates of Free Radical Bromination of Substituted Toluenes 1

Substituent Relative Rate <T+

None 1.00
p-ocn. 9.00 - 0.778
p-CU, 2.42 - 0.256
p-t-Buty\ 2.47 -0.311
p-C\ 0.73 + 0.405
m-Br 0.22 + 0.114
p-CN 0.12 + 0.659

It is not at all surprising that a p-methoxy group activates toluene

towards hydrogen abstraction, for the product radical has addi-
tional stability because of resonance forms which involve the
oxygen.

-h Br. > HBr

"jt>
:och 3 :och 3 :och 3 ©ooch 3

Similarly a p-methyl stabilizes the product radical, and thus also

stabilizes the transition state to the extent that this resembles the
product. However, a p-cyano group deactivates the toluene, al-

though it can also conjugate with the radical.

1. R. Pearson and J. Martin, "The Mechanism of Benzylic Bromina-

tion," Journal of the American Chemical Society, 85, 354 (1963).
POLAR EFFECTS IN FREE RADICAL REACTIONS 237

CH 2 CH 2

Furthermore, a m-bromo substituent is also strongly deactivating.

This result cannot be explained in terms of stability of the product
radical, which should be unaffected by this unconjugated group.
It seems apparent that the reaction is favored by groups which
+
can stabilize a positive charge (i.e., those with a negative o- ) and
slowed by those which destabilize a positive charge. This polar
effect can be explained 2 if one considers the structure of the

transition state for hydrogen abstraction.

-H— Br CH 2 — H -Br CH 2 - H— Br CH 2" H Br°

I II III

As the bromineatom begins to bond to the hydrogen, a species is

formed which three electrons are distributed over two partial
in
bonds, shown in dotted lines. This species can also be repre-
sented as a hybrid of three resonance forms: one in which there is a
carbon-hydrogen "long bond" (I), one in which there is instead a
bromine-hydrogen "long bond" (II), and the third ionic form
(III). Here it is recognized that since bromine is more electro-
negative than carbon there will be a drift of electrons from left to
right, so carbon will acquire some positive character while bromine

gets some negative charge. To the extent that the transition state
resembles resonance form II, all conjugating substituents should
help; to the extent that form III plays a role, then polar effects will
be seen.

2 (a) C. Walling, Free Radicals in Solution (John Wiley & Sons, New
.

York, 1957) (b). W. Pryor, Free Radicals (McGraw-Hill Book

Co., New York, 1966). Cf. the index under "Polar Effects."
238 ORGANIC REACTION MECHANISMS

A similar polar effect is seen 3 in the chlorination of rt-butyl

chloride (initiated by light). The relative reactivities of the
hydrogens on the four carbons are as indicated below.

5.6 17 5.6 1

CH 3 CH 2 CH 2 CH 2 CI

As expected, a primary hydrogen is less reactive than an ordinary

secondary hydrogen, but among the three secondary carbons there
is a striking decrease in reactivity on going from left to right.
This reflects the inductive effect of the chlorine substituent, which
destabilizes the transition state which has positively charged
carbon.
Chlorination of propionic acid with Cl 2 (and a light initiator)
4
also reveals the operation of a polar effect, the hydrogens on a
primary carbon being more reactive even though the secondary
radical would also be conjugated with the carbonyl group. Both
the per cent substitution on the two positions and the relative
reactivity of a hydrogen on each carbon are indicated; the latter is
corrected for the fact that there are three methyl hydrogens to only
two methylene hydrogens.

% 70 30
CH 3 —CH 2 C0 2 H
1 : 0.64 reactivity

5
Of course the polar effect is and it can be
only one factor,
4
overcome in some cases. For instance, chlorination of isobutyryl
chloride gives 20% substitution next to the carbonyl, but when this
result is corrected for the fact that there are six primary hydrogens
compared with one tertiary one, it is seen that the tertiary hydrogen
isreally more reactive.

3. J. Hine, Physical Organic Chemistry (2nd ed., McGraw-Hill Book

Co., New York, 1 962) , p. 459.
4. Ref. 2, p. 364.
5. For an W. Thaler, "Photobromina-
interesting additional effect, cf.
Unusual Orienting Effect in the Bromina-
tion of Alkyl Halides, an
tion of Alkyl Bromides," Journal of the American Chemical So-
ciety, 85,2607 (1963).
POLAR EFFECTS IN FREE RADICAL REACTIONS 239

% 80 20
(CH,),- CH— COC1
1 : 1.5 reactivity

In this case the greater stability of a tertiary conjugated radical is

the dominant factor.

The importance of a polar effect should also depend on the
electronegativity of the attacking radical. Methyl radicals can be
generated by the decomposition of diacetyl peroxide; when this is

done in isobutyryl chloride labeled with deuterium, found that

it is

the tertiary hydrogen is 12.4 times as reactive as any one of the

primary hydrogens.

o o O
II II
A II

CH3CO— OCCH3 > 2 CH3C > CO2 + CH3

\
O"
CH3 •CHo
\ \
CH 3 -
+ CDCOCI > CH 4 + CDCOCI
/ /
/ CH 3 CH 3

^ CH 3
\
CH D 3 + C- — COCI
CH3

Experimentally the relative reactivity is determined from the

amount of deuterium which is found in the methane (after correc-
tion for a measured isotope effect) . Polar effects play no role here
since the attacking methyl radical has essentially the same electro-
negativity as the product radical.
One of the most interesting manifestations of polar effects in
radical reactions is found in copolymerization.
7
When a radical
polymerization chain is initiated in the presence of two different
monomers, for instance acrylonitrile and acrylamide, it is fre-

6. C. Price and H. Morita, "The Reaction of Methyl Radicals with

Isobutyryl and a-Deuteroisobutyryl Chloride," Journal of the Amer-
ican Chemical Society, 75, 3686 ( 1 953 )

7. Ref. 2(a), Chapter 4.

240 ORGANIC REACTION MECHANISMS

quently found that a mixed polymer is formed, a so-called

copolymer, in which the two monomers are randomly incorpo-
rated.

CN CONH 2

R- + CH 2
= CH / +CH 2 = CH / >

CN CN CONH 2

/
2
/
(CH CH— CH — CH~ 2 CH 2 CH
/— —
.
)
etc.

This is monomer has added to the

not surprising, for whenever a
chain a new radical end is formed which can now attack either
monomer, and in this case their reactivities are quite similar so that
they become incorporated more or less randomly. On the other
hand, when two monomers of greatly different reactivity are used,
such as the very reactive acrylonitrile and the much less reactive
allyl alcohol, the reactive monomer will be incorporated first; the

unreactive one will only seldom enter the polymer until the
reactive monomer is used up.

CN CH2OH

R. + CH =CH
/ _:
+CH =CH
/ 2 >
2

CN CN CN CN
R— CH — CH— CH — CH — CH — CH — CH CH —etc.
I I I I

2 2 2 2

These are the results expected without a consideration of polar

effects, and they are the results which are often observed.

However, when styrene is copolymerized with maleic anhydride,

7
it is found that the reactivity of a monomer toward the growing
POLAR EFFECTS IN FREE RADICAL REACTIONS 241

C6H5

R- + CH 2 =CH + CH=CH
I I

c c

o
s\ o/ Xo
C6H5 C6H5

— <CH>CH— CH — CH CH
IICH— c
CH 2 CH
IIc ) etc.

o
c
• \o / Vo o
c
• \ o/ N o
radical chain depends on the nature of the end of the chain. If a
styrene unit has just been added, then there is a strong tendency for

add to maleic anhydride, and when the anhydride has

the chain to
been added there is a strong tendency for the new end to attack
styrene.

C6H5 < h c 6 H.

— CH = CH — CH — CH =
II
c
CH-

c
+ CH 2
c
||
>— CH
c
CH 2

o
s\ o/%o o
/\o/%o
C6H5 C 6 H.

— CH — CH — CH
— CH
II CH- CH 2 =CH
I
CH 2

000 000
I

c c c
s\/%
<
c
•\
>

/
r C6H5

— CH CH
•

2
/
CH — CH ©

000
I I

c c
^ \ /\
The is that the polymer has more or
result less regular alternation
of the two monomers.
242 ORGANIC REACTION MECHANISMS

The due to polar factors in the transition state

alternating effect is

for addition. When the end of the chain carries a maleyl radical,
addition to styrene involves a transition state in which a polar
form can make an appreciable contribution.
In the transition state the new bond is still quite long and the styryl

double bond is not yet completely broken. Drift of electrons to the

left occurs because the benzene ring can help stabilize a positive
charge, and the carbonyl group is strongly electron-attracting.
This extra ionic resonance form could not play a stabilizing role if

maleic anhydride were being attacked instead, since maleic anhy-

dride could not tolerate the partial positive charge. However,
when a chain with a styryl end attacks maleic anhydride an ionic
form contributes to stabilizing the transition state.

C 6 H. C 6 H.

CH 2 -CH CH =CH — CH.2— CH —CH CH

I
< I I
I

>
//
c
\o /v o o /\a\ o (

C 6 H.

CH 2 —CH CH - CH
I I

It will be noted that this type of resonance form resembles the

structures drawn in Special Topic 5 to explain the stability of

charge-transfer complexes.
The alternating effect quite commonly found when copoly-
is

merization involves two monomers with strongly different polar

character. Such examples show that a simple treatment of
reactivity in terms of bond energies and steric factors, as was done
in Chapter 7, may have to be supplemented by consideration of the
polar effects which can play a role in many radical reactions.
 8
PHOTOCHEMISTRY

although a solution of benzophenone in isopropyl alcohol is

perfectly stable under ordinary conditions, irradiation with ultra-

violet light causes the formation of acetone and benzpinacol.

OH OH

hv
2C 6 H 5 COC 6 H 5 + CH 3 CHOH CH 3
* (C 6 H 5 ) 2 C C(C 6 H S ) 2 + CH 3 COCH 3 (1

benzpinacol

In considering the mechanisms of such photochemical reactions

we will be concerned with two questions
1. What kind of new species is formed from the interactions of
light with an organic molecule?
2. What is the sequence of subsequent changes which this species
undergoes?

8-1 Excited States

As we pointed out in Chapter 7, ultraviolet light at 200 m/x has

an energy of 143.2 kcal/mole, while even blue light at 400 m^ has
71.6 kcal/mole. These very large amounts of energy can break

243
244 ORGANIC REACTION MECHANISMS

bonds and cause extensive changes in a molecule if the light is

absorbed. In order that light be absorbed by a molecule there must
be a quantum mechanically-permitted excited state of the mole-
cule of the correct energy, and the absorption of light by exciting
the molecule to this state must be "allowed." This latter require-
ment indicates that certain transitions between states are much less
probable than others; we will see some of the most striking ex-
amples of this in considering the properties of triplet excited states.
In addition to this direct mechanism, molecules may often be raised
by energy transfer from another excited molecule.
With the kinds of energies available in ultraviolet light we are
dealing with electronic excitation of molecules; the energies re-
quired to produce vibrational excitation are much smaller. In
most cases we will be concerned with excited -k electrons. For
example, in ethylene absorption of light at 162 m/* leads to an
excited state which has one of the tt electrons promoted to the tt
antibonding state, tt*.

it— Electronic states of ethylene

The electron spins remain paired, so this excited state is a sing-

let, Si. The groundwhich must have spins paired, is S
state, .

However, in an excited state with each orbital half occupied spins

may also be unpaired, leading to the triplet state Ti. In more com-
plex molecules with many occupied and unoccupied tt orbitals
there will be a number of possible higher energy singlet and triplet
excited states.
Light of the correct energy can excite a molecule from its S
state to the Si or higher singlet state. In almost all cases,
direct excitation of S to a triplet state such as T x is quite im-
probable, and light of the correct wavelength for the S -> Ti
change is only weakly absorbed. This is one of the selection rules
alluded to earlier. Spinning electrons have angular momentum,
and the change from S to Ti would involve a change in angular
momentum as one of the spins is reversed. Conservation of total
angular momentum is a fundamental law of nature, and during
the 10~ 15 seconds required for the electronic transition it is
PHOTOCHEMISTRY 245

quite improbable that some other interaction can occur so as to

compensate for the change momentum on going
in spin angular
from S to T
and allow the total momentum to stay constant.
x

Light can only be absorbed for an S — » Ti transition when such

an improbable compensating interaction occurs.
Of course all the electronic states of a molecule have vibra-
tionally-excited levels associated with them. Excitation could oc-
cur from S to various vibrational states of Si, so absorption spectra
generally consist not of a single line but of a series of lines corre-
sponding to energy differences between each of the populated vibra-
tional states of S and the various vibrational levels of Si. In most
spectra this series of lines coalesces into a broad absorption band.
According to the Franck-Condon principle atoms don't have time
15
to move during the 10" seconds involved in absorption of light,
so the excited molecule initially has the same geometry as in the
12
ground state; but in 10" seconds or so after absorption of light
the molecules in upper vibrational states of Si cascade down to
the lowest levels of Si and adopt the new equilibrium geometry of
Si by dissipating their small amounts of excess vibrational energy
to the environment. Even molecules in higher electronic singlet
states (S 2 , etc.) usually cascade rapidly down to the lowest levels
of Si. Accordingly, we can often refer to "the excited singlet
state" —meaning Si in vibrational equilibrium with its surround-
ings —even though excitation at first may have been to a higher
energy singlet.

One could imagine a further cascade of Si down to S passing ,

successively down the vibrational levels of S to the ground state

with dissipation of the energy as heat. This process, an example
of internal conversion (radiationless conversion between different
same total spin), happens to some degree
electronic states with the
for all molecules. However, the energy difference between Si and
S is often quite large, and the probability of transferring large
amounts of energy to the environment in quantum mechanically
allowed ways is limited, so Si has an appreciable lifetime.
Si may emit the energy as a photon and drop to (some vibra-
tional level of) S This process, fluorescence, generally occurs
.

within 10" 9 to 10" 6 seconds. During this time the excited molecule
may instead undergo internal conversion, as discussed above, or
it may undergo chemical reactions. The latter must be quite
rapid since the lifetime of Si before fluorescence is so short, and
246 ORGANIC REACTION MECHANISMS

even diffusion-controlled bimolecular reactions would require

10" 10 seconds or so. Of course, unimolecular chemical reactions
of Si or facile bimolecular reactions with other materials in high
concentration, such as solvent, are possible in the time available.
The other process of importance for Si is intersystem crossing,
conversion of a singlet to a triplet state (or vice versa). In some
molecules Si can convert to T x before fluorescence occurs. The
difference in spin angular momentum is not a prohibitive problem
since during the lifetime of Si there is a finite opportunity to com-
pensate for it with changes in angular momentum elsewhere in
the system; energy must also be dissipated since triplet states are
almost always lower in energy than the corresponding singlets,
but Si may first be converted to a vibrationally-excited Ti of
similar energy with subsequent cascade to the lowest levels of Ti.
The efficiency of intersystem crossing varies with molecular
structure in a way which is not completely understood. Benzo-
phenone, the molecule we will be considering in detail, under-
goes the process with almost unit efficiency, so the excited singlets
are all converted to the triplet. The possible processes we have
discussed are illustrated in Fig. 8-1.

FIGURE 8-1
Excitation, fluorescence, phosphorescence, and radiationless transitions.
Light of energy h v a absorbed promoting the molecule from its ground
is

state S to an excited singlet S2 Internal conversion (IC) of this to the

.

lowest vibrational state of St may be followed by fluorescence in which

Sx -» S + h v f internal conversion of S to S without radiation, or inter-
,
t

system crossing (ISC) of Sj to Tt The latter may phosphoresce (losing

.

energy h v p ) or undergo a radiationless intersystem crossing to S Wavy .

lines indicate radiationless processes while solid arrows show transitions

involving absorption or emission of photons.
 PHOTOCHEMISTRY 247

The Ti may not convert to Si again since Tj is of

triplet state
lower energy. can go to S by emission of a photon, but this
It

process is quite slow because of the spin angular momentum prob-

lem, so triplet lifetimes before such phosphorescence may range
from 10" 3 seconds to 10 seconds or even longer. The triplet state
may undergo intersystem crossing to S the energy being dissi- ,

pated as heat, but this is again slower than the corresponding

Si -* S process because of the change in spin. The triplet state
Ti may also enter chemical reactions; the time available is much
longer than for Si, so photochemistry involving triplet states is
not restricted to very fast processes and is more extensive than
for singlets.

8-2 Energy Transfer

There one other process available to Si and T x to which we

is

have not referred. These excited species may transfer their

energy to molecules in the environment not just in small bits but
all at once. The acceptor molecule is thereby raised to an elec-
tronically excited state of its own. This can be detected first of all
in the quenching of the fluorescence or phosphorescence of molecule
A by the addition of acceptor B. Secondly, the excited molecule
B may fluoresce or phosphoresce itself, or undergo chemical
changes. Considered from the standpoint of B, A is a photo-
sensitizer; it absorbs the light and transfers the energy to B.
This is particularly important in the case of triplet energy trans-
fer. The triplet state of A may transfer both

A(Ti) +B(So)^A(S ) +B(Tx)

energy and spin angular momentum to B, so total angular mo-

mentum is conserved. Such transfer occurs at a diffusion-
controlled rate provided it is energetically downhill (the S ->Ti
difference is greater for A than for B). Since Ti is not available
for many molecules by direct processes (So + hy—>Ti is "for-
bidden," and Si —> Ti does not always occur), photosensitization
is quite important in photochemistry.

8-3 Photochemical Reductions

Let us now return to a consideration of Reaction ( 1 ), the photo-

chemical reduction-dimerization of benzophenone by isopropyl
248 ORGANIC REACTION MECHANISMS

alcohol. The solution is irradiated with ultraviolet light with a

wavelength distribution from 300-350 m^. In this region iso-
propyl alcohol does not absorb light, but benzophenone undergoes
an n -> tt* singlet transition with A max near 350 m/x. The tt* orbital
is the. antibonding tt orbital of the carbonyl group, although it is

conjugated with the phenyl groups and more stable than ?r* of
aliphatic ketones. The
which is promoted to the tt*
electron
orbital is not one of the ?r electrons of benzophenone, but is one
of the so-called n electrons of the unshared pairs on the carbonyl
oxygen. The ?r electrons are of lower energy than the unshared
pairs, so tt -^ tt* excitation in benzophenone requires more energy
and occurs at 250 m/x. The energy of (delocalized) ir electrons
depends on structure, and there are some other ketones for which
the tt -> tt* excitation rather than n —» tt* leads to the lowest singlet
state Si.

In benzophenone the state with one of the oxygen n electrons in

a tt* orbital, Si, undergoes very rapid intersystem crossing to Ti
10
(the lifetime of Si is ca 10" seconds), and the triplet is actually
the species which reacts with isopropyl alcohol. One piece of
evidence that it is Ti, not Si, which attacks the isopropyl alcohol
is the effect of quenchers on the quantum yield.
The quantum yield of a reaction, 3>, is the ratio of product mole-
cules produced to photons absorbed. If each photon absorbed
leads to the formation of a product molecule, then <S> will be 1.0,
while dissipation of some of the light energy by fluorescence,
phosphorescence, or radiationless cascades to S will reduce the
quantum yield. In the case of chain reactions, such as the free
radical chlorination initiated by light which was discussed in Chap-
ter 7, quantum yields can be much larger than 1. The quantum
yield of Reaction(1) varies with the concentration of benzo-
phenone, and at high concentrations it is close to 1.0, so this
process can be quite efficient. However, the addition of a small
concentration of naphthalene to the solution results in a drastic
PHOTOCHEMISTRY 249

drop in the quantum yield, although its spectrum and that of the
benzophenone show that the naphthalene is not absorbing any of
the light. Naphthalene is instead acting as a quencher of an ex-
cited state of benzophenone, the energy transfer causing a return
of benzophenone to S with simultaneous excitation of the naph-
thalene. The naphthalene must be quenching a triplet state, not
a singlet, since only the Ti state of benzophenone would have a
lifetime long enough to undergo collisions with very low con-
centrations of naphthalene. Thus the singlet states must be going
to Ti before chemical reaction occurs. By the technique of flash
photolysis benzophenone triplet has also been directly observed
to be the species which attacks isopropyl alcohol.

(0^0 = t (C,H,),C — p t

Two resonance forms of benzophenone n -> n* triplet, T,

The n -»
tt* triplet has a remaining odd electron in an oxygen

n which makes it resemble an oxygen free radical. Like

orbital
other oxygen radicals it can remove the a-hydrogen of isopropyl
alcohol leading to two stabilized radicals.

(C 4 H 5 ) 2 C pt+ (CH 3 ) 2 CHOH — (C 6 H 5 ) 2 C OH + (CH 3 ) 2 C OH

T, I II

The radical I dimerizes to benzpinacol, while II can dispropor-

tionate to acetone and isopropyl alcohol. However, this mechanism
would lead to a maximum quantum yield 3> for benzpinacol forma-
tion of only 0.5, since two such reactions are needed to form the
two molecules of I for one dimer. At high concentrations of ben-
zophenone, <£ goes to 1 .0, so some other process must also be in-
volved. Apparently II can react with benzophenone to produce
another molecule of I, so that two molecules of I are produced per
photon absorbed.

(CH 3 ) 2 COH + (C 6 H 5 ) 2 C = -* (CH 3 ) 2 C =O+ (C 6 H 5 ) 2 6—OH

II I

The reaction occurs because I is more resonance stabilized than II.

Although few reaction mechanisms have been examined in as
much detail as that of Reaction ( 1 ) ,
photoreductions are known
with a number of other ketones and hydrogen donor combinations.
250 ORGANIC REACTION MECHANISMS

8-4 Photochemical Cycloadditions

When benzophenone is irradiated in solution with isobutylene,

the benzophenone n^-ir* triplet is again formed, and to some
extent it removes allylic hydrogen atoms from the olefin to give
photoreduction. However, the major product is an oxetane,
from two-step cycloaddition of the triplet to the olefin. Since
the intermediate III has an unpaired electron on a tertiary carbon,

= —
(QH S)2 C O (C 6 H 5 ) 2 C O
|

(c6 h 5)2 c — qt+ (ch 3 ) 2 c ch 2

I
+ .

I I

(CH 3 ) 2 C CH 2 CH 2 C(CH 3 ) 2

90% oxetanes
10%

(C 6 H 5) 2 C O

CH2 d(CH3)2
III

there is some preference (90%) for cycloaddition to occur in the

direction shown, but the isomeric oxetane is formed in 10% yield.
Intermediate III when first formed is still a triplet, and formation
of the second bond requires spin inversion.
One might expect that cycloaddition of benzophenone triplet

to a conjugated diene would be even easier than to a simple

olefin, since the intermediate radical would have allylic resonance

stabilization. However, irradiation of benzophenone with buta-

diene (again at the wavelengths used the light is absorbed only by
the benzophenone) leads not to oxetane formation but to dimeri-
zation of the butadiene.
CH = CH 2

CH 2 =CH CH=CH +(C H C==02 6 5)2

!i
— CH 2

CH 2 —
C

C
|
^ C^ ==Q ^
.

///H

IV

= CH ^CH 2

CH 2
>H CH CH CH = CH 2

II I

CH CH 2
CH 2 C//
= CH
-

/CH = 2
CH 2
VI
PHOTOCHEMISTRY 251

When benzophenone triplet collides with butadiene, a triplet energy

transfer occurs, so the ketone is acting as a photosensitizer. The
triplet state of butadiene is of lower energy than that of benzophe-
none (relative to their ground states), so triplet energy transfer is

possible here but not for a simple olefin (which would require a
higher energy sensitizer); when energy transfer is energetically al-
lowed, it is always faster than chemical bond formation. Butadiene
triplet cannot be formed without benzophenone since S + h„ —> T x
is strongly forbidden. The Si state of butadiene can be formed with
light of higher energy than 300 m/x, but even if such high energy
light is used, intersystem crossing is quite inefficient in the diene,
and the excited singlets would not tend to form much triplet. In
a real sense the triplet state of butadiene is available for reaction
only through the use of triplet photosensitizers such as benzo-
phenone.
The triplet butadiene now adds to ordinary butadiene by two-
step mechanisms to give the three dimers. However, the remark-
able observation has been made that the relative amounts of IV,
V, and VI formed depends on the exact photosensitizer used. It
is straightforward to determine the triplet energies of various
photosensitizers (by measuring the wavelength of their phospho-
rescence), and it turns out that sensitizers with a triplet energy
greater than 60 kcal/mole — including benzophenone with ET of
69 kcal/mole —produce
type product VI, with the remainder a
only a few per cent of the Diels-Alder
4 ratio of IV and V. As
1 :

ET of the photosensitizer is decreased to 53 kcal/mole, the 1:4

ratio of IV and V stays constant, but over 40% of the product is

C=CH 2
benzophenone^
y Q CHj
CH,= C^ trip ' et
)CH 2 — <^
\
j

H H

s-trans , .

fast

CH 2 . H CH *. .H
1

C C
benzophenone

J. triplet

CH ^
2
XH ICH 2
^ \H
252 u ORGANIC REACTION MECHANISMS

now VI. Below ET of 53 kcal/mole the quantum yield falls off

rapidly since excitation of the diene now requires more energy
than the sensitizer has.
Butadiene exists in the s-trans conformation in solution, with a
few per cent of s-cis as well. When butadiene accepts triplet
energy and undergoes a tt — » tt-* excitation to T^ it forms a triplet
with defined stereochemistry. Although the particular resonance
forms we have written indicate that rotation between carbons 2
and 3 is restricted in the triplets, a consideration of the molecular
orbitals of butadiene (Figure 8-2) will also make this clear.

f-3
V4 = .371 p, - .600 p 2 + .600 p 3 - .371 p<

C Y 3
= .600 p,
- .371 p2 - .371 p 3 + .600 p 4

c c C C ^ 2
= .600 p, + .371 p2 - .371 p 3 - .600 p 4

FIGURE 8-2
Molecular tt orbitals of butadiene.

The lowest tt —> tt* transition involves removing an electron from

<k 2 , which is antibonding between carbons 2 and 3, and placing it
 PHOTOCHEMISTRY 253

into #3 —
which has a 2 3 bond. The result (together with the
,

2 — 3 bonding of the electrons in ^i) is a strong 2 3 tt bond which —

prevents rotation during the lifetime of the butadiene triplet.

The energy change involved can be in going to the triplet state

and trans conformations of butadiene, and other

different for the cis
evidence indicates that E T is 60 kcal/mole for trans and 53 kcal/
mole for cis. A high-energy sensitizer can transfer energy to either
conformation; the trans triplet is thus formed predominantly, be-
cause the diene is mostly trans. A sensitizer with ET below 60
kcal/mole but above 53 kcal/mole will preferentially form the cis-
butadiene triplet, since energy transfer will only occur readily to
s-a's-butadiene.
Attack of a /ra/w-butadiene triplet on s-frafts-butadiene leads
to an intermediate hexadienyl diradical which still does not have
free rotation about the erstwhile 2 3 bonds, so it cannot close to —
VI. The c/s-butadiene triplet can give VI. Of course, in both
cases a spin must invert before the second bond can form.

C=C -I- c — c —^ c = c c = c

CH/
/ \ H CH,'
/ \ H
, /
CH/
\ H H
/ \CH 2

IV and V

,CH 2 H .CH 2s

/ X
Q=C/
XH H
- II I
2
^VI
C
CH 2 CH 2 I CH
/
H CH 2
CH 2 t

Generally triplet state photosensitizers are used in this way to

produce other triplets, but there one important exception. The
is

2 molecule exists in its ground state as a triplet, and its lowest

excited state is a singlet. Direct conversion of 2 to this excited
singlet state with light is again a forbidden process, but singlet
oxygen can be produced using a triplet photosensitizer since total

spin is thus conserved during energy transfer. This singlet excited-

state 2 can add in one step to dienes, as in a Diels-Alder reaction.
Singlet 2 has also been generated by chemical means, and it
254 ORGANIC REACTION MECHANISMS

f .Q P. I + 'A .* I »~ tp p I + A It

To T', S, S'

behaves identically with that generated with light and sensitizer.

Sensitizer O
+
i

o,
hv I P

8-5 Photoisomerizations

If a dilute solution of fra^s- 1,3 -pentadiene is irradiated in the

presence of benzophenone, it is converted to a mixture of 55%
trans and 45% cis pentadiene (dimerization is of course suppressed
by dilution). This same mixture is also formed starting with the
Dure cis diene. The two dienes can be interconverted through the

CH 3v H H.

" ^=^ =\
/ \\
:
c
benzophenone q^
/ \C
C CH 2 CH 2

trans -1,3- pentadiene cis - 1 , 3 - pentadiene

triplet of the diene, since in this triplet the 1 —2 (and 3 —4) ?r

bonds are weakened. This is seen by considering the molecular

orbitals of butadiene in Figure 8-2; an electron is removed from
#2 (1 —2 bonding)and placed in # 3 (1 2 antibonding). The —
photo stationary state comes when the rate of excitation of the cis
diene equals its rate of formation from excited states. The com-
position of the photostationary mixture thus depends on a number
of rate constants, including those for energy transfer from the
photosensitizer to the dienes. For cis pentadiene E T is lower
than for trans, so with low energy sensitizers the cis diene is
selectively excited, and the photostationary composition has a
higher proportion of trans diene than the 55% with benzophenone.
Direct irradiation of trans, cis, ^ran5 -2,4,6-octatriene leads to a
,

different sort of photostationary equilibrium, with its valence

isomer /nmy-5,6-dimethylcyclohexadiene. This reaction is com-
pletely stereospecific in both directions as shown, for very interest-
ing reasons to be discussed in Special Topic 8.
PHOTOCHEMISTRY 255

CH 3 hv

CH 3

A number of more complex photochemical rearrangements have

been discovered and studied, and more are being found each day.
However, an extensive consideration of the field is beyond the
scope of this book, and this chapter has been devoted instead to
illustrating fundamental principles which are common to all photo-
chemical processes.

General References

N. J. Turro, Molecular Photochemistry (W. A. Benjamin, New York,

1965). The best introduction to this topic.
J. G. Calvert and J. N. Pitts, Photochemistry (John Wiley & Sons,
York, 1966). A more advanced, and encyclopedic, treatment.
New
R. O. Kan, Organic Photochemistry (McGraw-Hill Book Co., New
York, 1966). Surveys a variety of photochemical reactions at an
introductory level.
D. C. Neckers, Mechanistic Organic Photochemistry (Reinhold, New
York, 1967). Similar to the books by Turro and Kan.
W. A. Noyes, G. S. Hammond, and J. N. Pitts, Advances in Photo-
chemistry (John Wiley & Sons, New York). A series, started in
1963, in which recent advances are discussed.
 8
Special Topic

- ORBITAL SYMMETRY
RELATIONSHIPS IN THERMAL AND
PHOTOCHEMICAL REARRANGEMENTS 1

on heating, cyclobutene derivatives undergo a stereospecific re-

arrangement to butadiene derivatives. Thus c/s-3,4-dimethylcyclo-
butene affords only cis, tomy-l,4-dimethylbutadiene while the trans

1. A simple introduction to some of the ideas involved in this treat-

ment is presented by J. Vollmer and K. "Woodward-Hoff-
Servis,
mann Rules: Edu-
Electrocyclic Reactions," Journal of Chemical
cation, 45, 214 (1968) and by M. Orchin and H. Jaffe, The
Importance of Antibonding Orbitals (Houghton-Mifflin, Boston,
Massachusetts, 1967). A more difficult treatment is that by
R. Hoffmann and R. B. Woodward, "The Conservation of Orbital
Symmetry," Accounts of Chemical Research, 1, 17 (1968).

256
ORBITAL SYMMETRY 257

(i)

(2)

dimethylcyclobutene goes to trans, rra/w-dimethylbutadiene. 2

By contrast, on irradiation trans, frww-dimethylbutadiene closes
3
to c/s-3,4-dimethylcyclobutene. In Chapter 8 we referred to a
related case, the stereospecific photochemical interconversions of
a cyclohexadiene and a hexatriene. In this special topic we will
consider the reasons for such stereospecificity. Thermal reactions
of various kinds will be considered first.

Cyclobutenes

Let us consider an orbital symmetry correlation diagram of the

type introduced in Special Topic 4. Thus both cyclobutene and
5-c/s-butadiene (the conformation which is formed first) have a

plane of symmetry perpendicular to the molecular plane. If in a

c/s-disubstituted on the two saturated
cyclobutene the groups
carbons rotate in opposite directions (the right one clockwise and
the left one counterclockwise), then the plane of symmetry is
maintained throughout the conversion to butadiene. For such
a transition state to be stable it should be intermediate between

2. R. E. Winter, "The Preparation and Isomerization of cis- and

frart5-3,4-Dimethylcyclobutene," Tetrahedron Letters, 1207 (1965).
3. R. Srinivasan, "Mechanism of the Photochemical Valence Tau-
tomerization of 1,3-Butadienes," Journal of the American Chemi-
cal Society, 90, 4498 (1968).
 ORGANIC REACTION MECHANISMS

the electronic ground states of starting material and product; as

we pointed out previously, this will not be possible if the starting
m.o.'s and product m.o.'s have different symmetries. Relative
to this plane the o- single bond of cyclobutene is S and the -n- bond
as well. These two orbitals cannot smoothly convert to the ground
state orbitals of butadiene, however, by a disrotatory process
(rotation of the saturated carbons in opposite directions) which
maintains the symmetry plane: Relative to this plane #1 of buta-
diene is S, but ^2 is A (cf. Figure 8-2 for the butadiene orbitals).
The excited-state orbital *3 is S with respect to the plane between
carbons 2 and 3; thus in the transition state for disrotatory ring
opening one of the electron pairs can be in an S orbital which re-
sembles both o- and <fr u but the other pair is in an S orbital which
resembles tt and ^3 (Figure 8-3 ) . The orbital #3 is of high energy.
Alternatively, the second electron pair could go into an A
transition-state orbital, so it resembled the more stable tf 2 , but
then it would also have some of the character of tt* in cyclobutene

and thus again be of high energy. It is because of these problems

that Reactions (1) and (2) occur by a conrotatory path.

ii A

cyclobutene butadiene

FIGURE 8-3
Orbital symmetry correlations for disrotatory interconversions of cyclo-
butene and butadiene.
ORBITAL SYMMETRY 259

a* A -^^^ ^^^ S Y 4

K* S
^""^ A Y 3

n A ^^^ ^__^S V 2

a S
^"^ "~^^~-A V,

cyclobutene butadiene

FIGURE 8-4
Orbital symmetry correlations for conrotatory interconversion of cyclo-
butene and butadiene.

The stereochemical results indicated in Reactions (1) and (2)

show that the saturated cyclobutene carbons rotate in the same
direction, so the change is called conrotatory instead of disrota-
tory.

The transition state for a conrotatory change does not have the
plane of symmetry we have been considering. Consequently its
wave functions are neither S nor A with respect to such a plane,
and it is possible for them to resemble to some extent the ground-
state wave functions of both starting material and product.
Although the transition state for conrotatory ring opening does
not have the plane of symmetry common to cyclobutene and buta-
260 ORGANIC REACTION MECHANISMS

diene, it does share another symmetry element with them. All

three have a two-fold rotational axis of symmetry: Rotation of
cyclobutene by 180° around a line passing through the centers
of the 1 —2 and —4 bonds
3 leads to an identical atomic arrange-
ment.
Under symmetry operation the o- bond is S, but the ?r bond
this

has the signs reversed, so it is A. On the other hand, o-* is A both

under this rotation and under the previously discussed reflection
through a plane. The symmetries of the cyclobutene and buta-
diene orbitals under this rotation are shown in Figure 8-4. With
respect to this rotational axis both ground states have an S and an
A orbital. The transition state in a conrotatory change also can
have one S and one A orbital with respect to this axis, so it can
correlate with the ground states on both extremes and need not be
of high energy.

s'*^?*-". rotation or reflection .Lr-^ass**-

Such symmetry correlation arguments are not always satisfying,

and they are not strictly applicable when unsymmetrical molecules
are involved. In particular, the reader will have noticed that while
cyclobutene may have both the plane and the axis of symmetry,
cw-S^-dimethylcyclobutene has only the plane, while trans-3,4-
dimethylcyclobutene has only the rotational axis. The perturba-
tion of molecular symmetry when a hydrogen is replaced by methyl
is not likely to cause drastic changes in chemistry, and in fact Re-
actions (1) and (2) both involved conrotatory changes, so some
more general argument than strict symmetry correlations must be

invoked.
The interconversion of butadiene and cyclobutene is an example
of an electrocyclic process. In such a process interconversion oc-
curs between one isomer with n ir electrons and the other with
n — 2 7T electrons and two electrons in a new o- bond. In the
course of the change each product orbital is being formed by trans-
formation of a particular starting material orbital; the transforma-
by a smooth continuous increase of electron density at
tion occurs
some atoms and decrease at others. The new ?r orbitals in the
cyclic isomer are derived from -n- orbitals of the open-chain isomer
 ORBITAL SYMMETRY 261

in a straightforward way: The lowest energy tt orbital of starting

material and product interconvert (each is nodeless), the next
highest tt orbitals interconvert (each has one node in the center of
the tt system), etc. Thus the ir orbitals of the cyclic isomer are
derived from the lowest energy -n orbitals of the open-chain isomer,
and the a orbital of the cyclic isomer is derived from the highest ?r

orbital of the open-chain isomer.

As we have seen in Figure 8-2, the n orbital of cyclobutene is re-
lated to #! of butadiene, since only #j bonding between carbons
is

2 and 3. The orbital ^ 2 has an antibond where the ?r bond of cy-

clobutene is to develop, so it cannot contribute to w, but with con-
rotation # 2 can lead to the bond. Disrotation is not allowed
o-

since # 2 would then form a tr* antibond. Even with unsymmetri-

cal derivatives, ^i and # 2 preserve their essential character of being
bonding and antibonding respectively between carbons 2 and 3, so
the argument is general.

«f

Finally, it might be noted that disrotation leads to the kind of

upper lobe-upper lobe overlap which is similar to normal ?r bond-
ing; as we saw in Special Topics 1 and 4, this kind of bonding is
unfavorable when 4 electrons are cyclically delocalized, but it is
favorable when 2, 6, and other 4rc + 2 numbers of electrons are
involved. Accordingly, we might expect the stereochemical situa-
tion to change in a six-electron case.

Cyclohexadienes

When trans, cis, /rarcs-l,6-dimethylhexa-l,3,5-triene is heated,

4
it cyclizes to c/s-l^-dimethylcyclohexa-S^-diene (reaction 3).

4. E. N. Marvell, G. Caple, and B. Schatz, "Thermal Valence Isom-

erizations . ," Tetrahedron Letters, 385 (1965).
. . This reaction
is also reported by E. Vogel, W. Grimme, and E. Dinne, ibid., 391
(1965).
262 ORGANIC REACTION MECHANISMS

CH 3

CH 3

This is the result of a disrotatory closure, the opposite of the cyclo-

butene <=± butadiene pathway. This stereochemistry can be under-
stood by simply noting that with six electrons to be cyclically
delocalized, normal overlap between the 1 and 6 carbons of a hexa-
triene will stabilizing, and such overlap occurs with disrotation.
be
Alternatively, we may consider the symmetries, bonds, and anti-
bonds of the three occupied hexatriene molecular ?r orbitals. As

f §

is the general pattern in polyenes the lowest orbital has no anti-

bonds, the second has one antibond in the middle, and the third
has two symmetrically-placed antibonds. Carbons 1 and 6 form a
a bond on cyclization, while carbons 2 through 5 form the diene
system in the cyclic isomer. This diene system will have a ^ 1 which
ORBITAL SYMMETRY 263

is all bonding and a ^2

with one antibond in the middle, as in bu-
tadiene, so $1 —>
#1 and 2 —> #2- Again it is seen that the new
<J>

a bond must be derived from the highest occupied ?r orbital of the

starting triene, <£ 3 and disrotation
, is needed to form this o- bond.
Finally, symmetry correlation diagrams like those in Figures 8-3

and 8-4 can be constructed which also show that in this case dis-
rotation is favored. This is left to the reader.
It should be mentioned in passing that the electrocyclization of
hexatriene derivatives is intellectually related to the Cope rear-
rangement of 1,5-hexadienes. 5

CH 2
* ^CD 2
- CH f \ Dj

C
^V j£cD, - CH C'd 2

In this case a a, not a w, bond is broken in forming the new bond,

but again 6?r electrons are delocalized in the transition state. Such
processes can be very rapid if all the atoms are held in the right
position by other bonds.

Cyclopropyl Cation

Cyclopropyl cation undergoes a ring opening to allyl cation;

again orbital symmetry considerations predict a stereospecific proc-
6
ess. disrotation.

The argument is quite straightforward since only two electrons are

involved, in the o- orbital of cyclopropyl cation and the completely
bonding lowest -n orbital (Special Topic 1) of allyl cation.

5. S. J. Rhoads, "Rearrangements Proceeding through 'No Mecha-

nism' Pathways," in Molecular Rearrangements, P. de Mayo, ed.
(John Wiley & Sons, New York, 1963), Vol. I, p. 655.
6. R. B. Woodward and R. Hoffmann, "Stereochemistry of Electro-
cyclic Reactions," Journal of the American Chemical Society, 87,
395 (1965).
264 ORGANIC REACTION MECHANISMS

Solvolysis of cyclopropyl tosylates does in fact lead to products

7
derived from allyl cation, but the free cyclopropyl cation is not
involved as an intermediate. Instead, rearrangement is concerted
with ionization, as in many of the solvolyses discussed in Special
Topic 3. Under these circumstances disrotatory opening still

occurs, but in addition the direction of rotation is such as to move

the electrons of the o- bond being broken to a side opposite that of
the leaving group.

The reason for this preference is related to that involved in the

general rule (Chapter 3) that S N 2 reactions go by backside dis-
placement. A 8
particularly striking example is found in the rela-
tive acetolysis rates of the two bicyclohexyl tosylates (I) and (II).
At 100°, I reacts 2,500,000 times as rapidly as does II.

OTs
OTs
OTs

25,000 1.0

Relative acetolysis rates

Since ring opening should accompany ionization, the two hydro-

gens in I move outward leading to acceptable geometry in the allyl
cation product (III). If II opened in the same stereospecific fash-
ion, it would lead to an extremely strained cyclohexenyl cation IV.

C. H. De Puy, L. G. Schnack, and J. W. Hausser, "The Solvolysis

of Cyclopropyl Tosylates," Journal of the American Chemical So-
ciety, 88, 3343 (1966).
U. Schollkopf, K. Fellenberger, M. Patsch, P. Schleyer, T. Su, and
G. van Dine, "Acetolyse von Endo- und Exo-Bicyclo-(n,l,0)-alkyl-
tosylaten," Tetrahedron Letters, 3639 (1967).
ORBITAL SYMMETRY 265

disrotation

backside displacement

disrotation
II — CH,
backside displacement

Photochemical Electrocyclic Reactions

In Chapter 8 we described the photostationary interconversion of

a hexatriene and a cyclohexadiene. Comparison of that case with
the thermal reaction (3) just discussed shows that the photochemi-
cal process is conrotatory while the thermal reaction is disrotatory.
Furthermore, the thermal opening of a cyclobutene to a butadiene
in reactions (1) and (2) was conrotatory, while the photochemi-
cal closure of the butadiene to a cyclobutene was disrotatory. The
difference between the two- kinds of electrocyclizations is general.
Let us consider the simple case of the photocyclization of a
butadiene to cyclobutene. The photoexcited butadiene (Si or T x )

has two electrons in # x and one each in & 2 and *J> 3 . As closure
occurs this pattern will persist, with one fully-occupied and two
singly-occupied orbitals, and the system will begin to resemble an
excited state of the product. At some reasonably late stage energy
is and the ground-state product is formed. An allowed stereo-
lost
electronic pathway is one which permits the excited starting mate-
rial to transform smoothly into the lowest excited state of the
product.
As Figure 8-5 shows, singly-occupied ?r and ?r* of excited cyclo-
butene can be derived from # 3 and ^ 2 of excited butadiene (the
266 ORGANIC REACTION MECHANISMS

°*-c+ -c

-- + + Y*
I

±±±± Yi j|^ ^|{.-c+ +c-

Butadiene (T,) Cyclobutene (T,)

FIGURE 8-5
Correlation of states in a photochemical cyclobutene-butadiene intercon-
version.

triplet is shown, but the same is of course true for Si) The a bond .

now comes from * 1? and disrotation would be required to get

bonding in \^i between carbons 1 and 4.
The general pattern in photochemical electrocyclic conversions
between lowest excited states is that the two singly-occupied or-
bitals on each side of the reaction correlate (e.g., \J> 2 and \J> 3 with

7T and 7T* ) and that the o- bond correlates not with the highest oc-

cupied orbital but with the highest fully-occupied orbital. Thus

a diagram similar to Figure 8-5 can be constructed for the hexa-
triene-cyclohexadiene interconversion. In the lowest excited state
of hexatriene, one of the electrons in <£ 3 (cf. the next-to-last sec-
tion for the hexatriene orbitals) is promoted to <£ 4 . The new o-

bond must be constructed from 3> 2 , the highest fully-occupied or-

bital, and this requires conrotation as is experimentally observed.
The o- orbital in a photochemical electrocyclic reaction is corre-
lated with the 7T open chain isomer immediately below
orbital of the
the one used in a thermal reaction. Thus thermal and photochemi-
cal processes must always have different rotatory preferences, since
the relative symmetries of the two terminal atoms in a polyene
alternate in it orbitals of increasing energy. This theorem follows
from the presence of an additional node in each higher ?r orbital
ORBITAL SYMMETRY 267

and is easily confirmed by a glance at the hexatriene or butadiene

orbitals.
The ideas involved in orbital correlation are of wide generality,
and they have been applied to a variety of processes in addition to
those discussed here. As with all such unifying concepts they have
not only rationalized old observations but also stimulated a flood
of new experiments.
 NDEX

Acetal formation, 188-189 Amides, basicity, 19

Acetic acid, acidity, 15 hydrolysis, 198
Acetone, acidity, 15 Anchimeric assistance, 99
bromination, 42-44, 52-54 see also Neighboring group
Acetylacetone, acidity, 1 participation
Acetylene, acidity, 19 Aniline, acidity, 15
bonding, 6 basicity, 15, 16
Acid, general, 52, 185 Annulenes, 34
specific, 52 Antiaromaticity, 30
Acids and bases, structure and Antibond, 10, 31
reactivity, 14-20 Antibonding orbital, 5
Activated complex, 45 "Aromatic sextet," 26
Activation energy, 44 Aromaticity, definition, 22
Acyl-oxygen fission, 193-194 Autoxidation, 225
1,3-Addition, 137-140
Aldol condensation, 187-189 Base, general, 52
Alkyl-oxygen fission, 194-196 Beckmann rearrangement, 97
Allyl anion and cation, molecular Benzene, bonding, 6-11
orbitals, 30 resonance energy, 8
Allylic halides, reactions, 76 Benzhydryl chloride, S N 1 reac-
Alternating effect, 242 tions, 70
268
 NDEX 269

Benzilic acid rearrangement, 208 Chlorination, of butyl chloride,

Benzoin, mechanism of forma- 238
tion, 49, 188 of isobutyryl chloride, 238
Benzophenone, 243, 247 of propionic acid, 238
Benzyne, 160 Chromic acid oxidation, 210-212
Bond dissociation energies, 217- a-Chymotrypsin, 58
220 Claisen condensation, 199-200
Boranes, 126 Classes of organic reactions, 36
Bridged ions, 100-108 Collision theory, 44
"Bridgehead" reactivity, 90 Common ion effect, 71
Bromination, of benzene, 168 Conrotation, 258
radical, 236 Cope rearrangement, 263
Bromonium 124
ion, Copolymerization, 239-242
Ar-Bromosuccinimide, 224 Coulomb integral, 10
Br0nsted catalysis law, 54, 64, Crotyl chloride, solvolysis, 76
181 Curtius degradation, 98
Butadiene, molecular orbitals, Cyclization, ring size effects on
146, 252 rate, 81
photocyclization, 265 Cyclobutadiene, molecular or-
photodimerization, 250 bitals, 27
structure and stability, 1 Cyclobutane, molecular orbitals,
f-Butyl alcohol, acidity, 15 143
/-Butoxy radical, fragmentation, Cyclobutene, ring opening, 256
231-232 Cyclodecapentaenes, 31
f-Butyl bromide, hydrolysis, 72 Cycloheptatrienyl anion, 25
Cycloheptatrienyl cation, 24
Cannizzaro reaction, 206 Cyclohexadiene, isomerizations,
Carbenes, addition to olefins, 261
130-134 Cyclohexene, bromination, 41
formation, 120 molecular orbitals, 146
singlet and triplet states, 132 Cyclononatetraenyl anion, 32
Carbenoids, 134 Cyclooctatetraene, 26
Carbonium ion, as reaction in- Cyclooctatetraenyl dianion, 32
termediate, 69-78, 87-91, Cyclopentadiene, acidity, 24
93 ff., 111-113, 122-124, Cyclopentadienyl cation, 25
149-159, 165, 209 Cyclopropane, bonding, 3
Catalysis,51-56 Cyclopropenyl anion, molecular
Chain initiation, 220 orbital theory, 28
Chain propagation, 221 Cyclopropenyl cation, and deriva-
Chain reactions, 220 tives, 29
Chain termination, 221 molecular orbital theory, 28
Charge-transfer complex, 167 Cyclopropyl cation, opening,
see also tt complex 263 ff.
270 INDEX

Cyclopropylcarbinyl chloride, Esters, hydrolysis, 192-197

solvolysis, 105 Ethylene, bonding, 4

d orbitals, in sulfones, 181 Field effect, 18

in thiophene, 24 Fisher esterification, 194
Decarbonylation free radical, Flash photolysis, 249
230-231 Fluorescence, 245
Decarboxylation, 200-201 Formic acid, acidity, 17
free radical, 230 Franck-Condon principle, 245
Diazomethane, addition to olefins,
Free energy, and equilibrium, 14
138 Friedel-Crafts reactions, 151, 173
decomposition, 133 Furan, resonance energy, 23
Diazonium salts, aromatic, 158
Diels-Alder reaction, 140-146
Glucose, mutarotation, 186
Diffusion controlled rate, 181,
Glycols, oxidation, 211-213
246
a,a-Dimethylallyl chloride, solvol-
Grignard reaction, abnormal,
ysis, 77
207-208
Guanidine, basicity, 20
Dinitromethane, acidity, 15
Diradical, in cycloaddition, 134,
136-137 Halogenation, of aromatic com-
Disrotation, 258 pounds, 149
free radical, 222-225
El reactions, 111-113 see also Bromination, Chlorin-
ElcB reactions, 119 ation
E2 reactions, 113-119 Haloketones, reactivity, 84
rates, 115-117 Hammett equation, 116, 154
stereochemistry, 118 Hammond's postulate, 88
Electrocyclic reactions, 260 HBr, free radical addition to
Electron spin resonance spectros- olefins, 226
copy, 216-217 Heterolytic and homolytic re-
Electrophilic, definition, 37 actions, 37
a-Elimination, 120-121 Hexaphenylethane, 214, 234
Energy transfer, 244, 247 Hexatriene, molecular orbitals,
Enolate ions, rates and equilibria, 262
178 Hofmann degradation, 98
Enolization, equilibrium, 176 Hofmann rule, 117
Enols and enolates, 174-183 Huckel's rule, 31
Enthalpy, 47 Hund's rules, 132
Entropy effects, 47, 80, 120 Hybrid orbitals, effect of s char-
Enzyme-substrate complex, 58 acter on electronegativity, 3
Epoxide formation, 80, 131 Hydration, of carbonyl com-
E.s.r. {see Electron spin reson- pounds, 183-185
ance spectroscopy) of olefins, 122
INDEX 271

Hydride shift, 96 a-Methylallyl chloride, solvolysis,

Hydride transfer reactions, 205- 76
210 Microscopic reversibility, princi-
Hydroboration, 125 ple of, 109, 194
Hydrogen bonding, in enols, 178 Molecular orbital theory, 9
in salicylate ion, 18 Mustard gas, 81
Hyperconjugation, 88 Mutarotation, 186

I-strain, in carbanions, 181 Neighboring group participation,

in S x l reactions, 90 79
in S N 2 reactions, 84 Neopentyl derivatives, in S N 2 re-
Inductive effects, 18 actions, 83
Internal conversion, 245 Neopentylamine, deamination, 94
Internal return, 78 Nitration of aromatic compounds,
Intersystem crossing, 246 148-149, 170-173
Ion pairs, 74 Nitromethane, acidity, 15
Isotope effect, 149 Nitronium ion, 148
Nitrophenols, acidity, 18
Kinetic equivalence, 43, 51, 190 Node, 10
Kinetic order, 41 No-mechanism reactions, 37
Kinetics,40-44 Nonclassical carbonium ion, 102
of enzyme reactions, 59 Norbornyl cation, 103-105
Kolbe electrolysis, 205 Normalizing factor, 5
Nucleophilic, definition, 37
a-Lactone, 79 Nucleophilic aromatic substitu-
Law of mass action, 41 tion, 158-163
Lead tetra-acetate, 212-213 Nucleophilicity, 84-86
Leaving groups, reactivity, 86
Light, energy of varying wave- Oppenauer oxidation, 207
lengths, 223 Orbital electronegativity, and
Linear free energy relationship, acidity, 19
55, 116 and dipole moment, 5
Long bond, in Diels-Alder re- Orbital symmetry, 31, 143 ft*.,

action, 142 256 ff.

in radical transition states, Osmium tetroxide, addition to
237 olefins, 129
Oxetanes, 250
Mass law effect {see Common ion Oxime formation, 190-192
effect) Oxygen, singlet, 253
Mechanism of reaction, criteria Ozone reaction with olefins, 137—
in establishing, 39-40 138
definition, 38
Meerwein-Ponndorf reaction, 207 Paramagnetic resonance {see
Methane, bonding, 1 Electron spin resonance)
272 NDEX

Partial rate factor, 153 Schiff bases,190

Pentadienes, photoisomerization, Semiquinone, 204
254 a bond, in ethylene, 4
Periodic acid, 211-212 o- complex, 165
+
Permanganate, reaction with o- , in radical brominations, 236
olefins, 204 table, 155
Phenol, acidity, 15 S x l reactions, 69
Phenonium ion, 101 of esters, 194-195
Phenylglyoxal, rearrangement, stereochemistry, 73-76
208 S x 2 reactions, 67
Phosphorescence, 247 of chlorobenzene, 162
Photosensitizers, 133, 247 of esters, 196
Photostationary state, 254 stereochemistry, 68
7T bond in, ethylene, 4 S x 2' reaction, and stereochemis-
7r complex, 122, 166 try, 78

7T* state, 244 Solvation, of carbonium ions, 75

Pinacol, rearrangement, 94 effects on acidity, 17
pK 'sa of acids (Table), 15 and nucleophilicity, 86
pK R+ , definition, 25 Solvent, effect in S x l reactions, 90
Polarizability, and nucleophilic- in S x 2 reactions, 86
ity, 85 Solvolysis, 71
Polymerization, 220-222 Substituent effects in aromatic
Propylene, dipole moment, 5 substitution, 152-158
Pyridine, basicity, 19 Sulfonation, 149
Pyrrole, resonance energy, 23
Telomers, 229
Quantum yield, 248 Thermal additions, 135, 142-146
eliminations, 120
Rate-determining step, 42 Thiophene, resonance energy, 23
Rearrangements, free radical, Toluene, acidity, 16
232-233
103, Transition state, 49
Resonance energy and molecular Transition state theory, 46-51
orbital theory (Table), 12 Trifluoroacetic acid, acidity, 18
Resonance forms and resonance Triphenylmethyl cation, stability,

energies (Table), 8 25
Resonance integral, 10 Triplet quenching, 249
Ribonuclease, 62 state, 244 ff.

p, in aromatic substitution, 157 Triptycene, 161

in E2 reactions, 116 Tropylium ion, 24, 209

Salicylic acid, acidity, 17 Valence bond theory, 7

Saponification, 193
Saytzeff rule, 113, 117 Wolff rearrangement, 98
ORGANIC REACTION MECHANISMS
AN INTRODUCTION
second edition

Ronald Breslow/Columbia University

The scope of the second edition of ORGANIC REACTION MECHANISMS is unique: it is at

once an introduction for advanced students, a valuable supplementary text for the under-
graduate who wants a deeper look at organic chemistry, and an excellent overview text
for scientists working outside the field. The book will complement general organic
courses and can be used in both senior level and first-year graduate courses in advanced
organic chemistry and reaction mechanisms.

New topics and areas of increased coverage include:

• The Woodward-Hoffman rules

• Acid-base catalysis and the Brjrinsted equation

• Photochemistry
• Ribonuclease
• Linear free energy relationships

BENJAMIN/ MARUZEN
HGS MOLECULAR STRUCTURE MODELS
ORGANIC CHEMISTRY SET

$3.95 Per Set

Complete in permanent plastic box and with Instruction Booklet. Originally developed in

Tokyo by Maruzen Co., Ltd., these new student molecular model sets have been adapted
for W. A. Benjamin, Inc., by its academic consultants. The new kits are unusually versa-
tile, flexible, and durable.

Pre-formed color-coded parts are scaled to size. Various polyhedra with holes for bonds
represent the different atoms. The use of straight and curved lengths of plastic rod
allows the illustration of single, double, and triple bonds. All these features enable the
student to construct easily and quickly three-dimensional visualizations that are aes-
thetically superior to those possible with other model sets. And since these models stress
location of atoms as well as bond angles and distances, constructions are more meaning-
ful—visually and conceptually— than those built with sets showing only bonds.

With the extra oxygen atoms provided in the Organic Set, models of compounds such
as the following can be made: simple sugars, such as fructose, mannose, and glucose,
showing chair and boat forms and cis-trans isomerism; models of derivatives such as
pentamethylglucose; ascorbic acid; amino acids, such as cysteine, lysine, glycine, ala-

nine, and tryptophan; dipeptides, such as glycylalanine and phenylalanine; and com-
pounds such as nicotine and adrenaline, as well as many others.

W. A. BENJAMIN, INC. NEW YORK

CHEM 1253