Dokumen - Pub Why Chemical Reactions Happen 9780199249732 0199249733

why chemical
reactions happen
: James Keeler
Peter Wothers
Why chemical reactions happen
Why chemical reactions happen
James Keeler
Lecturer in the Department of Chemistry and
Fellow of Selwyn College, Cambridge
and
Peter Wothers
Teaching Fellow in the Department of Chemistry and
Fellow of St Catharine’s College, Cambridge
OXFORD
UNIVERSITY PRESS
Great Clarendon Street, Oxford 0x2 6DP
Oxford University Press is a department of the Universiay of Oxford.
It furthers the University's objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dares Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© James Keeler and Peter Wothers 2003
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2003
Reprinted 2004 (with corrections), 2005, 2006
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
ISBN 13: 978-0-19-924973-2
ISBN 10: 0-19-924973-3
5 79108 6
Typeset by the authors
Printed in Great Britain on acid-free paper by
Antony Rowe Ltd, Chippenham, Wiltshire
Acknowledgements
We are grateful to Michael Rodgers, then commissioning editor for science

texts, for being so enthusiastic about our original proposal for this book and
for being responsible for its adoption by OUP. Michael has since moved on to
higher things, leaving us in the capable hands of Jonathan Crowe, John Gran-
didge, Miranda Vernon and Mike Nugent, who have been very helpful through-
out the writing and production process — we are most grateful to them for their
assistance. James Brimlow and John Kirkpatrick were kind enough to read and
comment on the final draft; we are very grateful to them for their suggestions
and corrections.
It is something of a cliché to offer thanks to our students for what they have
taught us. However, be it a cliché or not, it is true that having to think about,
explain and re-explain this material over a number of years has improved our
own understanding and, we hope, the clarity of our explanations. So in a very
real sense we are indeed grateful to the many students we have had the privilege
to teach.
JK wishes to thank Prof. A. J. Shaka and his colleagues in the Department
of Chemistry, University of California Irvine, for being excellent hosts during
a three-month sabbatical which made it possible for a great deal of progress
to be made on the text. He also wishes to thank Prof. Jeremy Sanders for his
invaluable advice, support and encouragement.
PW wishes to thank Del Jones for his support and encouragement during
the production of this book.
This book has been typeset by the authors using I4TRxX, in the particular
implementation provided by MiK TEX (www.miktex.org); we are grateful
to the many people who have been involved in the implementation of this type-
setting system and who have made it free for others to use. The text is set in
Times with mathematical extensions provided by Y&Y Inc., (Concord, MA).
Most of the diagrams were prepared using Adobe Illustrator (Adobe Systems
Inc., San Jose, CA) or Mathematica (Wolfram Research Inc., Champaign, IL),
and chemical structures were drawn using ChemDraw (CambridgeSoft Corp.,
Cambridge, MA). Molecular orbital calculations were made using HyperChem
(HyperCube Inc., Gainesville, FL).
Our intention is to maintain some pages on the web which will contain
some additional material in the form of appendices, downloadable versions of
many of the diagrams and corrections to the text. The address is
www.oup.com/uk/best.textbooks/chemistry/keeler.
Finally, we hope that you enjoy reading this book and that it helps in your
understanding and exploration of the fascinating world of chemistry. How re-
actions happen, and why they happen, is the most fundamental question in
chemistry — we hope that you find some answers here!
Cambridge, February 2003.

TO OUR PARENTS
AND OUR TEACHERS

Contents
1 What this book is about and who should read it
What makes a reaction go?

2.1 Energy changes
2.2 Entropy and the Second Law of Thermodynamics
2.3 Entropy and heat
2.4 Role of temperature
2.5 Exothermic and endothermic reactions
2.6 How do we know all this is true?
2.7 The Second Law and Gibbs energy
2.8 Gibbs energy in reactions and the position of equilibrium
2.9 Where does the energy come from?
3 Ionic interactions
3.1 Ionic solids
3.2 Electrostatic interactions
3.3 Estimating lattice energies
3.4 Applications
3.5 Ionic or covalent?
Electrons in atoms
4.1 Atomic energy levels
4.2 Photoelectron spectroscopy
4.3 Quantum mechanics
4.4 Representing orbitals
4.5 Radial distribution functions
4.6 Hydrogen orbitals
4.7 Atoms with more than one electron
4.8 Orbital energies
4.9 Electronegativity
5 Electrons in simple molecules

5.1 The energy curve for a diatomic
5.2 Molecular orbitals for H2
5.3 Symmetry labels
5.4 General rules for forming molecular orbitals
5.5 Molecular orbitals for homonuclear diatomics
5.6 Heteronuclear diatomics
vil
6 Electrons in larger molecules 80
6.1 Two-centre, two-electron bonds 81
6.2 Hybrid atomic orbitals . 83
6.3 Using hybrid atomic orbitals to describe diatomics 88
6.4 Simple organic molecules 89
6.5 Reconciling the delocalized and hybrid atomic orbital approaches 93
6.6 Orbital energies: identifying the HOMO and LUMO 95
7 Reactions 97
7.1. The formation of Hz from Ht and H— 97
7.2. Formation of lithium borohydride 98
7.3. Nucleophilic addition to carbonyl 102
7.4 Nucleophilic attack on C=C 106
7.5 Nucleophilic substitution 107
7.6 Nucleophilic attack on acyl chlorides 110
8 Equilibrium 113
8.1 The approach to equilibrium 113
8.2 The equilibrium between two species 114
8.3. General chemical equilibrium 118
8.4 Influencing the position of equilibrium 122
8.5 Equilibrium and rates of reaction 130
9 Rates of reaction 131

9.1 What determines how fast a reaction goes? 131
9.2 Why is there an energy barrier? 133
9.3. Rate laws 135
9.4 The Arrhenius equation 137
9.5 Elementary and complex reactions 139
9.6 Intermediates and the rate-determining step 141
9.7 Reversible reactions 143
9.8 Pre-equilibrium 145
9.9 The apparent activation energy 148
9.10 Looking ahead 150
10 Bonding in extended systems — conjugation 151

10.1 Fourp orbitals in a row 153
10.2 Three p orbitals in a row 157
10.3 Chemical consequences — reactions of carbonyl compounds 163
10.4 Stabilization of negative charge by conjugation 168
11 Substitution and elimination reactions 172

11.1 Nucleophilic substitution revisited — Sj 1 and Sn2 172
11.2 The stabilization of positive charge 174
11.3. Elimination reactions 180
11.4 Addition reactions — elimination in reverse 183
11.5 E2 elimination 185
Vili
12 The effects of the solvent 189
12.1 Different types of solvent 190
12.2 The solvation of ions 194
12.3 Solubilities of salts in water 197
12.4 Acid strengths and the role of the solvent 202
12.5 How the solvent affects the rate of reaction 207
13 Leaving groups 210

13.1 Energy profiles and leaving groups 211
13.2 Leaving group ability 213
13.3 Leaving groups and pK, 215
14 Competing reactions 216

14.1 Reactions of carbonyls with hydroxide 216
14.2 Reactions of enolates 221
14.3 Unsymmetrical enolates 225
14.4 Substitution versus elimination revisited 228
14.5 So why do chemical reactions happen? 233
Index 234
iX
ay
1 What this book is about and
who should read it
When we first come across chemical reactions they are presented to us as

‘things that happen’ — this book is about why and how chemical reactions hap-
pen. The ideas and concepts presented here will give you a set of tools with
which you can begin to understand and make sense of the vast range of chem-
ical reactions which have been studied.
Let’s look at some examples of the kinds of things we are going to be
talking about.
e In solution Ag* ions and Cl7 ions react together to give a precipitate of
solid AgCl:
Ag* (aq) + Cl (aq) —> AgCl(s);
on the other hand Agt and NO, do not form a precipitate but remain in
solution. Why?
e Acyl chlorides react readily with water to give carboxylic acids:
I +
Cc + H2O
2 _ UC HCl
whereas amides are completely unreactive to water;
O
I + 20 Sc + no reaction
Why?
e In solution 2-bromo-2-methylpropane can form the positively charged

ion (CH3)3C*, but bromomethane cannot form CH}:
H3C. CH3 CHs
H3C~
“o—-
Br H3C~
it~CHg + BP
What this book is about and who should read it
e Propene reacts with HBr to give the bromoalkane with the structure A
and not B:
H H
| HBr tr
H.-C HBr % Br. UC.
cov ~ Cc C CH
’ CHs H3C~ CH; aS 3
5 H H
ay A B
Why?
You may already have come across explanations for some of these obser-
vations.
e Areason often given as to why a precipitate of AgCl is formed from Agt

and Cl” ions in solution is that there is a large release of energy when
these ions form a solid lattice.
However, the precipitation of CaCO3 from a solution of Ca?+ and CO;-
ions is endothermic, i.e. energy is taken in. This being the case, our ex-
planation for what is going on in the case of Agt and Cl” cannot be
correct! We will see in Chapter 12 why it is that both of these precip-
itation reactions are possible, even though one is exothermic and one
endothermic.
e A reason often given for why acyl chlorides are more reactive than
amides is that the chlorine substituent is electron withdrawing and so
makes the carbony] carbon ‘more positive’ and so more reactive towards
nucleophiles such a water. The problem with this argument is that nitro-
gen has just about the same electronegativity as chlorine, so we would
expect it to have the same electron withdrawing effect. Clearly some-
thing more complicated is going on here, and in Chapter 10 we will
explain just what this is.
e You will find that the explanation often given for the easy formation of
the (CH3)3CT ion is that the positive charge is stabilized by the electron-
donating methyl groups. The question is, why, and how, do methyl
groups donate electrons and why does this stabilize the cation? We will
answer this in Chapter 11.
e The result of the reaction of propene with HBr may be predicted by

Markovnikov’s rule, which states that the bromine ends up on the carbon
with the most alkyl groups attached to it. This just begs the question as
to why Markovnikov’s rule works — a question we will also answer in
Chapter 11.
What we hope to do in this book is provide you with a logical and soundly
based set of ideas which you can use to explain why these, and many other,
reactions behave in the way they do.
How this book is laid out

It turns out that the most fundamental principle which determines whether or
not a reaction will take place is the Second Law of Thermodynamics, so we will
start out by discussing this very important and fundamental concept. Powerful
though this law is, it does not really help us to understand what is going on in
detail in a reaction — for example, which bonds are broken, which are formed
and, most importantly, why. To understand these aspects of reactions we need
to look in more detail at chemical bonding.
There are two basic interactions which lead to the formation of chemi-
cal bonds: electrostatic interactions between charged species, leading to ionic
bonds, and covalent interactions in which electrons are shared between atoms.
Of these, ionic bonds are perhaps the easiest to understand, so we consider
them first. In Chapter 3 we will see that we can develop a simple picture of the
bonding in ionic compounds which enables us to make some useful chemical
predictions.
To understand how covalent bonds are formed, we need to go into quite a
lot of detail about the behaviour of electrons, first in atoms and then in progres-
sively more complex molecules. These topics are considered in Chapters 4, 5
and 6. Once we have mastered how to describe the bonding in reasonably sized
molecules — which boils down to describing where the electrons are and what
they are doing — we are ready to start considering reactions. It may seem like
a long road to get to Chapter 7, but it is well worth it, as by then you will have
all of the tools necessary to really understand what is going on.
In the subsequent chapters we will go on to look at different kinds of re-
actions, introducing along the way new concepts and ideas such as the idea of
reaction mechanisms and the way in which chemical equilibrium is established
and influenced. The final chapters look at the crucial role the solvent plays
in a chemical reaction and then at what happens when there is more than one
pathway open to a reaction.
How to read this book

This is not a textbook which aims to cover all the material of, say, a typical first-
year university course. Rather, this book is aimed at giving you an overview of
why and how chemical reactions happen. For this reason, the text cuts across
the traditional divisions of the subject and we have freely mixed concepts from
different areas when we have felt it necessary and appropriate.
Our intention is not to give a complete, detailed, description of what is
going on, but rather to give an overview, bringing the key ideas into a clear
perspective. If you continue your studies in chemistry you will no doubt learn
about many of the ideas we introduce here in greater depth. We hope that
having read this book you will have an overview of the subject so that you can
see why you need to know more about particular ideas and theories and how
these will help you to understand what is going on in chemical reactions.
This book will probably be most useful to those who are starting a univer-
sity course in chemistry. It will also be of interest to those who are coming
to the end of pre-university courses (such as A level) and who wish to deepen
and extend their understanding beyond the confines of the syllabus. Teachers
in schools and colleges may also find the text to be a useful source. We develop
our argument from chapter to chapter, so the book is best read through from
the beginning.
Some practical matters

Generally in the text we have used the systematic names for compounds, but
have also given the common names, as you are likely to encounter these much
more at university level. In any case, the molecules we have chosen as exam-
ples are all rather simple and so naming them is not a great problem. When
it comes to drawing molecular structures, we have included all of the carbons
and hydrogens rather than simply drawing a framework as is sometimes done.
Although this makes the structures a little more cluttered, it avoids ambiguity
and for simple structures does not seem to cause a problem. The only exception
we have made to this rule is that benzene rings are drawn as a framework.
As you know, the four single bonds around a carbon atom are arranged so
-
that they point towards the corners of a tetrahedron. The usual way of repre-
-
o
°
“,
senting this is shown in Fig. 1.1. We imagine that two of the bonds lie in the
C plane of the paper — these bonds a represented by simple lines. The third bond
Fig. 1.1 The usual representation of the
is represented by a solid wedge with its point toward the carbon; this bond is
tetrahedral arrangement of groups the one coming out of the page. Finally, the fourth bond, which is going into
around a carbon atom. the page, is represented by a thick, dashed line.
Where we show pictures of atomic and molecular orbitals we have been
at particular pains to make sure that these are realistic. Often the pictures are
based on actual calculations, and so can be relied upon. Even where we have
drawn sketches of orbitals we have made sure that they are realistic. So, for
example, the 2p orbital is not represented as an ‘hourglass’ simply because this
is not its true shape, despite what you will see in many books!
The molecules we are going to talk about are the smallest we can find
which contain the functional groups we are interested in. This makes is possi-
ble to focus on what is really important— the functional group and its electronic
properties — without worrying about other details. We will also use just a few
molecules to illustrate the ideas rather than introducing numerous different ex-
amples. Again, this will allow us to focus in detail on what is going on in these
reactions.
Abbreviations
For quick reference, the abbreviations we are going to use are listed below.
Some if these may be not be familiar to you as yet, but as you read on you will
come across explanations of all of these terms.
2c-2e two-centre, two-electron

AO atomic orbital
HAO hybrid atomic orbital
HOMO _highest occupied molecular orbital
LUMO lowest unoccupied molecular orbital
MO molecular orbital
DMF dimethyl! formamide
DMSO _ dimethyl! sulfoxide
2 What makes a reaction go?
From experience we know that, once started, some reactions simply ‘go’ with
no further help from us. For example, gaseous ammonia reacts with gaseous
hydrogen chloride to give white clouds of solid ammonium chloride:
NH3(g) + HCl(g) —> NHsCl(s).

Another example is magnesium burning in oxygen — once ignited the metal
burns strongly, giving off heat and light:
Mg(s) + 2O2(g) —> MgO(s).

In writing these equations we used the right arrow —> to symbolize that the
reaction goes entirely to the products on the right-hand side.
Some reactions, rather than going entirely to products, come to equilibrium
with reactants and products both present in significant amounts. For example,
the dissociation of ethanoic acid in water is far from complete and at equilib-
rium only about 1% of the ethanoic acid in a 0.1 M solution has dissociated:
CH3COOH(aq) + H20(J) = CH3COO™ (ag) + H307* (aq).
Another example is the dimerization of NO2 to give N2Oa:
2NO2(g) = N20a(g);
100% fi00%
at 25 °C the equilibrium mixture consists of about 70% N20q4. Na B*:
In these equations we use the equilibrium arrows = to indicate that the re- < : :
aS
action can go either way. Indeed, as you may know, at equilibrium the forward
e
ie}
and back reactions have not stopped but are proceeding at equal rates. fo]
There is another class of reactions which simply do not ‘go’ at all, at least
not on their own. For example, ammonium chloride does not spontaneously 0% “—$§ SJ 0%
x oy Zz
dissociate into ammonia and hydrogen chloride, neither does magnesium oxide position of equilibrium
spontaneously break apart to give magnesium and oxygen. Fig. 2.1 Illustration of the concept of
Taking all of these reactions together we see that the real difference be- position of equilibrium for the simple
tween them lies in the position of equilibrium, illustrated in Fig. 2.1. For some equilibrium between A and B; the
percentage of A and B in the equilibrium
reactions the equilibrium lies so completely to the side of the products that to mixture is given by the grey and black
all intents and purposes the reaction is complete. For other reactions, there are lines, respectively. {f the position of
equilibrium is at x the equilibrium mixture
significant amounts of products and reactants present at equilibrium. Finally,
contains almost entirely A — the reaction
for some reactions the equilibrium lies entirely on the side of the reactants, so has not gone to a significant extent. At y
that they appear not to take place at all. the equilibrium mixture contains
significant amounts of A and B. Finally, at
So, the question to answer is not ‘what makes a chemical reaction go?’ but
z the reaction has gone almost entirely to
‘what determines the position of equilibrium?’. We will begin to answer this products.
question in this chapter.
2.1 Energy changes

The reaction between magnesium and oxygen is strongly exothermic, meaning
that heat is given out; we can represent this on the energy diagram shown in
Mg + $0, Fig. 2.2. The products are lower in energy than the reactants, and the energy
difference between them, shown by the arrow, is the amount of energy that
energy given
out as heat
appears as heat.
As the products are lower in energy than the reactants we often say that
energy
‘the products are more stable than the reactants’. In this context, ‘more stable’
means ‘lower in energy’. It would be tempting to conclude that magnesium
MgO
oxide is formed from magnesium and oxygen because magnesium oxide is
more stable (i.e. lower in energy) than magnesium + oxygen. Generalizing
Fig. 2.2 Energy diagram for the reaction
this, we might say that a reaction goes if the products are more stable, that is
between magnesium and oxygen to
produce MgO. The reaction is exothermic lower in energy, than the reactants. This train of thought is very tempting but,
and the heat given out is represented by as we shall see, it is at best incomplete and at worst completely wrong!
the arrow.
The best way to see that this tempting argument is wrong is to find a clear
example in which it fails. We need look no further than the dimerization of
nitrogen dioxide:
2NO2(g) = N204(g).
As we commented on above, this reaction comes to a position of equilibrium
which involves a mixture of reactants and products. The reaction going from
left to right involves making a bond between the two nitrogen atoms — it is
2NO5 the formation of this bond which leads to a reduction in energy and hence the
A
energy
reaction being exothermic. The reverse, in which the bond is broken, therefore
has to be endothermic, that is heat is absorbed; Fig. 2.3 illustrates these points.
Suppose we start with pure NO2 and then allow the system to come to equi-
librium. We can actually do this experiment by first lowering the pressure, as
Fig. 2.3 Energy diagram for the this favours dissociation (dissociation involves increasing the number of moles
dimerization of nitrogen dioxide. The of gas); at sufficiently low pressures it is found that very little N2O« is present.
forward reaction is exothermic, indicated
by the solid arrow. The back reaction is If we then increase the pressure up to, say, one atmosphere, the position of
therefore endothermic, indicated by the equilibrium changes and some N2Ozg is formed; this is illustrated in Fig. 2.4.
dashed arrow. The reaction involved is the exothermic process:
2NO2(g) —> N2Oa(g).

On the other hand, we can approach equilibrium by starting with the pure
product, N204. Experimentally, this can be achieved by increasing the pres-
sure on the system; this favours the dimer and at high enough pressures dimer-
ization is essentially complete. If we then decrease the pressure back to one
atmosphere some of the dimer dissociates to NO? as the equilibrium position
is approached. The reaction involved is the endothermic process:
N204(g) —> 2NO2(g).
We see that the equilibrium position can be approached either starting from
solely reactants or solely products. Approaching equilibrium from one side
involves an exothermic process and approaching from the other side involves
an endothermic process.
2NOz > N20, _: oe N,O, —> f 2NO3

{ L- < ——e
exothermic endothermic
low pressure 1 atmosphere high pressure

>99% NO» 70% NoO, >99% NzO4
Fig. 2.4 Illustration of how the equilibrium between N2O4 and NO» can be altered by altering the
pressure. Since NOg is dark brown and N2QO, is colourless the shift in the position of equilibrium can
easily be detected by observing the brown colour.
Clearly, it cannot be the case that the only reactions which ‘go’ are exother-
mic ones in which the products are lower in energy than the reactants. En-
dothermic reactions, in which the products are higher in energy than the reac-
tants, also take place and in general the approach to chemical equilibrium will
involve both types of reactions. The tempting idea that reactions which ‘go’
are ones in which the products are ‘more stable’ has to be abandoned.
The reaction between ammonia and hydrogen chloride provides us with an-
other example. If we apply sufficient heat, ammonium chloride will dissociate
to ammonia and hydrogen chloride:
h
NH4Cl(s) > NH3(g) + HCl(g).
This is another example of an endothermic reaction which can be made to “go’.
The true driving force for a reaction is not moving to lower-energy products
but is concerned with a quantity known as entropy and is governed by the
Second Law of Thermodynamics; we will discuss both of these in the next
section.

The Second Law of Thermodynamics allows us to predict which processes will
‘go’ and which will not. It is therefore the key to understanding what drives
chemical reactions and what determines the position of equilibrium. The Sec-
ond Law is expressed in terms of entropy changes. Entropy is a property of
matter, just like density or energy, and we need to understand what it is if we
are going to apply the Second Law.
The simplest description of entropy is to say that it is a quantity associated
with randomness or disorder. A gas has greater entropy than a solid, as in the
gas the molecules are free to move around. A liquid has higher entropy than increasing entropy
a solid, but lower entropy than a gas. We can rationalize this by arguing that Fig. 2.5 The entropy of a gas is greater
the molecules in a liquid are more free to move than those in a solid, but not as than that of a liquid which in turn is
free as those in a gas; Fig. 2.5 illustrates these points. greater than that of a solid. This increase
in entropy can be associated with the
You may have come across the statement that ‘entropy always increases’. increasing freedom with which the
The idea is that any process which takes place is always associated with an in- molecules move as we go from a solid to
crease in entropy, that is an increase in chaos or disorder. However, a moment’s a gas.
thought reveals that there must be something wrong with this statement. After
all, if it were true, how could ice form from water? This is a process in which
the entropy decreases, as ice has lower entropy than liquid water. Another
example is the reaction
NH3(g) + HCl(g) —> NH4Cl(s)

in which there is clearly a reduction of ‘entropy as gaseous reactants give rise
to a solid product. Yet we know that this reaction takes place.
In fact the statement ‘entropy always increases’ is almost correct, it just
needs to be expressed more carefully to give the Second Law of Thermody-
namics:
In a spontaneous process, the entropy of the Universe increases.
This statement introduces two important ideas. Firstly, the idea of a sponta-
neous process. Secondly, that it is the entropy of the Universe that we need
to consider, not just the entropy of the chemical system that we are thinking
about. We will look at each of these ideas in turn.
A spontaneous process is one which takes place without continuous inter-
vention from us. A good example of this is water freezing to ice — a process
about which we will have a lot more to say in later sections. If we put water
in a freezer set to —20 °C we know that ice will form; this is an example of a
spontaneous process. The reverse will never happen: at —20 °C ice will never
melt to give water. Another example of a spontaneous process is the mixing
of gases: once two unreactive gases are released into a container they will mix
completely. The reverse, which would involve the separation of the two gases,
is never observed to take place.
We can bring about the reverse of spontaneous processes by intervening.
For example, we can melt ice by applying heat and we can separate gases by
using chromatography or liquefying them followed by distillation. The key
point is, however, that the reverse of a spontaneous process never happens on
its own.
we think of as The Second Law tells us that we have to consider the entropy change of
the Universe, not just the entropy change of the process we are interested in.
This sounds like a tall order, but it turns out not to be as difficult as it at first
seems. We simplify things by separating the Universe into two parts: the sys-
Universe
tem, which is the part we are concentrating on, such as the species involved in
a chemical reaction, and the surroundings, which is everything else (Fig. 2.6).
The entropy change of the Universe can then be found by adding together the
entropy changes of the system and the surroundings:
Fig. 2.6 The Universe is separated into
the part we are interested in — the
system, and the rest — the surroundings.
entropy change of Universe =
For example, the system would be our entropy change of surroundings + entropy change of system
chemical reactants in a beaker and the
surroundings would be the bench, the or, in symbols,
laboratory and everything else in the t
Universe.
ASuniv = ASsurr + A Ssys
where ASuniy, ASsurr and ASsys are the entropy changes of the Universe, the
surroundings and the system, respectively.
We shall see in the next section that in an exothermic process the heat given
out increases the entropy of the surroundings. This increase can compensate
for a decrease in the entropy of the system so that overall the entropy of the
Universe increases. This is why water can freeze to ice and NH3 can react with
HCI. These processes involve a reduction in entropy of the system but they are
sufficiently exothermic that the surroundings increase in entropy by enough to
make the entropy of the Universe increase.
These are rather subtle points, which we will tease out in the next few
sections. The first thing to discuss is how the exchange of heat can lead to an
entropy change and how the temperature affects the size of this entropy change.

Imagine an exothermic reaction which gives out heat to a nearby object. What
effect does this have on the entropy of the object? We know that absorbing heat
increases the energy of the molecules in the object and so causes them to move
around more.
We have seen before that more motion — increasing disorder — is associated
with an increase in entropy. It is therefore not surprising to find that when
an object absorbs heat its entropy increases (Fig. 2.7). A further consequence
of the heat being absorbed is that the temperature of the object increases, so
we can also see that raising the temperature of an object increases its entropy.
Conversely, if heat flows out of an object, its entropy decreases.
What happens, then, if the object we are supplying heat to is very large —
for example the surroundings, which comprise all of the Universe apart from Fig. 2.7 When an object absorbs heat
the system? The temperature of such a large object as the rest of the Universe the molecules move more rapidly and so
its entropy increases.
is not going to be affected by the amounts of heat which typical chemical pro-
cesses produce — put another way, we are not going to heat up the rest of the
Universe significantly by leaving the oven door open! Nevertheless, although
the temperature of the surroundings has not changed, they have absorbed en-
ergy (in the form of heat) and, as described above, this leads to an increase in
the entropy. In summary:
In an exothermic process the system gives out heat to the surroundings

and so the entropy of the surroundings is increased.
In an endothermic process the system takes in heat from the surroundings

and so the entropy of the surroundings is decreased.
We can apply these ideas right away to understand how it is that water can
freeze to ice. To melt ice to water we need to apply heat, so ice — water is an
endothermic process. This makes sense, as in melting ice we are beginning to
break the hydrogen bonds between the molecules. The reverse process, water
— ice, in which bonds are being made, is therefore exothermic.
Ice has lower entropy than liquid water, so on freezing the entropy of the
system decreases. However, the process is exothermic, so the entropy of the
surroundings increases. It must be that the increase in entropy of the surround-
ings is sufficient to compensate for the decrease in entropy of the system so that
IEE entropy of system decreases

4 ons ie | >
= <= heat given out to ‘> entropy of the
the surroundings Universe increases
y
entropy of the surroundings increases 4
Fig. 2.8 When water freezes to ice its entropy decreases; however, the process is exothermic and
ihe heat given out results in an increase in the entropy of the surroundings. For this process to be
spontaneous, the increase in entropy of the surroundings must outweigh the decrease in entropy of
the system so that overall the entropy of the Universe increases.
overall the entropy of the Universe increases. The overall process is illustrated
in Fig. 2.8.
Coming back to the reaction we discussed above

entropy of we see that there is a decrease in the entropy of the system as a solid is formed
NH.(9) system
39. decreases NH,Cl(s) from gases. In solid NH4Cl there are strong interactions between NHj and
CI” ions and so it is not surprising to find that the reaction is exothermic; it
HCI(g) heat
follows that the entropy of the surroundings increases. These entropy changes
entropy of are illustrated in Fig. 2.9. As the reaction is spontaneous it follows from the
surroundings Second Law that the increase in entropy of the surroundings must outweigh the
increases decrease in entropy of the system, giving an overall increase in the entropy of
Fig. 2.9 When gaseous HCl reacts with the Universe.
gaseous NHg, solid NH4Cl is formed; There is one more subtlety to sort out here. Water freezes to ice only if
there is consequently a reduction in
entropy of the system. However, the the temperature is below 0 °C; at higher temperatures the water remains liquid.
reaction is exothermic and the heat given Water solidifying to ice is always exothermic, and as a result the entropy of
out increases the entropy of the the surroundings would always increase. However, it seems that this increase
surroundings. Overall, the entropy
increase of the surroundings outweighs in entropy is only sufficient to compensate for the decrease in entropy of the
the decrease in entropy of the system, so system when the temperature is sufficiently low. We will see in the next sec-
the entropy of the Universe increases
tion just exactly how and why temperature affects the entropy change of the
and the process is spontaneous.
surroundings.
2.4 Role of temperature

The entropy increase resulting from supplying a certain amount of heat to an
object depends on its temperature. We can understand this in the following
way.
Suppose that the object is very cold so that the molecules are not moving
around very much. Supplying some heat will make the molecules move around
more, and so the entropy increases. Now suppose that we apply the same
amount of heat to a much hotter object in which the molecules are already
moving around a great deal. The heat will cause the molecules to move even
more vigorously, but the change is much less significant in the hotter object
than in the cooler one, simply because in the hotter object the molecules are
already moving more vigorously in the first place. We conclude that supplying
heat to a hotter object causes a smaller increase in entropy than supplying the
same amount of heat to a cooler object; this is illustrated in Fig. 2.10.
2.4 Role of temperature 11
~
increase in entropy
, os
On
‘’
warmer
smailer increase in entropy {vo

—<—o=>
—
io
a little hotter
Fig. 2.10 Supplying heat to a cool object, in which the molecules are not moving very much, causes
a larger increase in entropy than does supplying the same amount of heat to a hot object in which the
molecules are already moving vigorously.
For the surroundings, which are very large, the entropy change can be com-
puted very simply from:
heat absorbed by the surroundings

entropy change of the surroundings = - .
temperature of the surroundings
Using symbols we can write this as
(2.1)
where qsurr and Tsurr are the heat absorbed by and absolute temperature of the
surroundings, respectively.
Heat is normally measured in joules and temperature in kelvin, so it follows
from Eq. 2.1 that the units of the entropy change (and of entropy) are J K~!. In
chemistry we often quote molar quantities, that is the entropy change per mole,
in which case the units are J K~! mol7!.
Equation 2.1! can only be used to compute the entropy change of the sur-
roundings. Finding the entropy change of the system requires a different ap-
proach which we will not discuss here but simply note that the absolute en-
tropies (as they are called) of many substances have been determined and ta- Some typical values of absolute
entropies at 298 K:
bles of these can be found in any standard text. Such tables also list the entropy
He(g): 126 J K~! mol7!
change associated with phase changes, such as ice melting to water. With the Ho(g): 131 J K~* mol7!
aid of such data we shall, in the next section, use the Second Law to explain Hg(/): 76 J K71 mol7*
why it is that the temperature has to be below 0 °C for water to freeze. Cu(s): 33 J K71 mol-1
How to make ice

The approach we will adopt is to compute the entropy change of the system
and of the surroundings; adding these together gives us the entropy change of
the Universe, ASyniy. If this is positive, the entropy of the Universe increases
and from the Second Law we know that the process can take place. If A Suniv
is negative, that is the entropy decreases, we know that the process cannot take
place, as it is forbidden by the Second Law.
From tables we can find that the entropy change for ice — water is
22.0 J K~! mo!7!. As expected, this is a positive number since disorder is
increasing. We are interested in the change water — ice, for which the en-
tropy change is simply minus that for ice > water. The entropy change of the
system, AScys, is therefore —22.0 J K~! mol™!.
To compute the entropy change of the surroundings using Eq. 2.1 we need
to know the amount of heat absorbed by the surroundings. This heat can only
come from the system. If the system gives out heat in an exothermic process
(i.e. qsys is negative), the surroundings must be absorbing this heat, making
sur positive. On the other hand, if the process in the system is endothermic
(i.e. Gsys iS positive) the heat that the system absorbs must come from the sur-
roundings. The surroundings are therefore losing heat, making gsurr negative.
surroundings surroundings
In summary, from the point of view of the surroundings, the heat is travelling
a
exothermic endothermic
in the opposite direction to that of the system; Fig. 2.11 illustrates these points.
We can therefore write:
Gsurr = —4sys
sys Negative sys Positive
Asurr POSitive Qsurr Negative where qsurr is the heat absorbed by the surroundings and qgys is the heat ab-
sorbed by the system.
Fig. 2.11 Illustration of the heat flow
between the system (the black circle) and From tables we can easily find that the heat needed to melt ice to wa-
the surroundings. An exothermic process ter (called the heat of fusion) is 6010 J mol7!; as expected, this is posi-
results in the flow of heat into the
surroundings; an endothermic process
tive. We are interested in the process water — ice, which is the reverse of
results in the flow of heat out of the melting, so the heat for this is simply minus that for melting ice to water:
surroundings. Gsys = —6010J mol—!. Therefore, when water freezes to ice, this heat is given
to the surroundings so qsurr = 6010 J mol~!,
We can now compute the entropy change of the surroundings. First, we
will do this at 5 °C, which is 278 K:
Gsurr
A Ssurr =
Tsurr
_ 6010
~ 278
21.6) K7! mol7!.
We already know that ASsys = —22.0J K~! mol7! so we can find the entropy
change of the Universe as follows:
ASuniv = ASsur + ASsys

= 21.6 — 22.0
= -0.4JK7! mol!.
The entropy change of the Universe is negative, meaning that the entropy has
decreased. Such a process is forbidden by the Second Law and so we conclude
that water cannot freeze at 5 °C.
Now, let us repeat the calculation at —5 °C, which is 268 K:
surr
A Ssurr =
SUIT
_ 6010
~~ 268
22.43 K7! mol7!,
2.5 Exothermic and endothermic reactions 13
ASuniv = ASsur + ASsys
= 22.4 — 22.0
= 0.43 K7! mol7!.
At this lower temperature the entropy change of the Universe is positive, mean-
ing that the entropy has increased. So, the process is allowed by the Second
Law and at this lower temperature water can freeze to ice.
It is important to notice that what has changed between 5 °C and —5 °C is
the entropy change of the surroundings. Furthermore, the reason that A Ssur
has changed in size is simply due to the temperature change; the heat given to
the surroundings has remained the same. Due to the way in which we compute
ASsurr (Eq. 2.1), for a given amount of heat the size of the entropy change can
be altered simply by changing the temperature of the surroundings.
It is now clear why we put water in a freezer in order to make ice (Fig. 2.12).
The process of water going to ice is associated with a reduction in the entropy
of the system. For this to be allowed by the Second Law, the reduction in en- water entropy ice
tropy must be compensated for by an increase in the entropy of the surround- IES
on,
decreases
ings. Such an increase comes about when the surroundings absorb heat, as will
be the case for an exothermic process such as water freezing. However, the en-
tropy change of the surroundings depends on their temperature. The lower the surroundings
temperature, the greater the entropy change. So, water will only freeze once entropy increases
the temperature of the surroundings is low enough that the increase in their Fig. 2.12 When water solidifies to give
ice, the entropy of the system decreases.
entropy is large enough to compensate for the decrease in entropy of the water
However, the process is exothermic and
as it freezes. so the entropy of the surroundings will be
So, the reason why we put water in the freezer when we want to make ice raised as they absorb the heat. The
process will only be spontaneous if the
is so that the entropy change of the surroundings (inside the freezer) is large
surroundings are cold enough that their
enough to compensate for the entropy decrease of water freezing. It can be entropy increase outweighs the entropy
hard to explain this to people who think that the freezer “sucks the heat out of decrease of the system, thus making the
entropy of the Universe increase.
the water’!
Why all this fuss about making ice?

You may well be asking yourself why it is that we are discussing making ice in
a book devoted to understanding chemical reactions! The reason for this is that
it is easier to describe how the Second Law applies to a physical process such
as water freezing to ice than it is to apply it to chemical equilibrium. However,
we can use the tools we have developed so far to understand some reactions
which either go entirely to products or do not go at all; these are the subject
of the next section. Later, in Chapter 8, we will return to the discussion of
chemical equilibrium.
2.5 Exothermic and endothermic reactions

Using the Second Law and the idea of separating the entropy change of the
Universe into the entropy change of the system and the surroundings, we can
begin to understand why some chemical reactions go. We have already dis-
cussed the reaction

in which there is a reduction in entropy of the system. However, as the reaction

is exothermic the entropy of the surroundings increases, and this compensates
for the decrease in entropy of the system.
Another example is the precipitation of solid AgCl from solution (illus-
trated in Fig. 2.13):
Ag* (aq) + Cl” (aq), — AgC\(s).
Again, there is a reduction in entropy of the system, as a solid is being
formed. In solid AgCl there are strong interactions between the Agt and Cl
entropy of system ions; however in solution these ions have favourable interactions with the polar
decreases entropy of the water molecules. Which of these two sets of interactions are stronger depends
heat +> Surroundings on the particular ions in question but it turns out for AgCl that the interactions
increases
in the solid are the stronger, so the above reaction is exothermic. As a result
there is an increase in the entropy of the surroundings. Since we know that
the precipitation takes place it must be allowed by the Second Law and so the
increase in entropy of the surroundings must outweigh the decrease in entropy
of the system. We can simply turn this discussion round and use it to explain
AgCl(s) why solid AgCl! does not dissolve in water as such a process would involve a
Fig. 2.13 The precipitation of solid AgCl decrease in the entropy of the Universe and so cannot occur.
by mixing solutions containing Ag*t and Reactions which involve an increase in the entropy of the system do not
CI” ions is a spontaneous process. The necessarily need to be exothermic to be allowed by the Second Law. For ex-
entropy of the system decreases as a
solid is formed from two solutions; ample, dissolving solid ammonium nitrate in water, illustrated in Fig. 2.14,
however, the process is exothermic, involves an increase in the entropy of the system because the ions which were
resulting in an increase in the entropy of held in the crystal lattice are free to move in solution — there is an increase in
the surroundings which must outweigh
the decrease in entropy of the system so disorder. For ammonium nitrate it turns out that the interactions between the
that overall the entropy of the Universe ions in the solid are stronger than those between the ions and solvent water, so
increases.
forming the solution is an endothermic process. This means that the entropy of
the surroundings decreases. We know that ammonium nitrate dissolves, so it
must be that the entropy increase of the system overcomes the entropy decrease
of the surroundings so that the Second Law is obeyed.
entropy of system
increases
i = entropy of the
heat taken in Universe increases
H,0() NH,NO,(s) from surroundings NH,NO,(aq)
4
entropy of the
surroundings decreases
Fig. 2.14 When ammonium nitrate dissolves in water there is an increase in the entropy of the system
as the solution is more ‘random’ than the solid plus liquid which we start with. Dissolving this salt
turns out to be endothermic, which means that heat is absorbed from the surroundings resulting ina
decrease in their entropy. The process is spontaneous at room temperature, so we conclude that the
entropy increase of the system must outweigh the entropy decrease of the surroundings so that the
entropy of the Universe increases.
A contrast to ammonium nitrate dissolving in water is when solid sodium

hydroxide dissolves. As before, the entropy change of the system is positive but
this time the process is exothermic so the surroundings also increase in entropy.
The entropy of the Universe increases unconditionally and so the Second Law
is obeyed.
2.6 How do we know all this is true? 15
As was discussed in Section 2.4 on p. 10, temperature plays an important

role in determining the size of the entropy change of the surroundings. A good
example of the effect of this is in dissociation reactions such as
N204(g) —> 2NO2(g).
Since two moles of gas are produced from one there is an increase in dis-
order and the entropy of the system increases. This reaction is endothermic,
resulting in a decrease in the entropy of the surroundings. The key thing is that
the size of this decrease depends on the temperature of the surroundings, as can
be seen directly from the way in which the entropy change of the surroundings
is calculated: AScurr = surr/Tsurr- Increasing the temperature decreases the
size Of ASsurr simply because of the division by Tsu; Fig. 2.15 illustrates this
point.
For the dissociation reaction, when the temperature is low the entropy de-
crease of the surroundings outweighs the entropy increase of the system and so
the reaction does not take place as it would result in a decrease in the entropy
increasing temperature
of the Universe, in violation of the Second Law. At higher temperatures, the
entropy change of the surroundings is smaller in size and eventually it ceases Fig. 2.15 Illustration of the way in which
to outweigh the entropy increase of the system. Then the dissociation can take the entropy change of the Universe is
affected by an increase in temperature.
place as it is allowed by the Second Law. On the left, the entropy change of the
Another example of a dissociation reaction is solid MgO going to gaseous system is positive (upward pointing
atoms: arrow), but the entropy change of the
surroundings is sufficiently negative that
MgO(s) —> Mg(g) + O(g). it outweighs the entropy increase of the
system and so the entropy of the
This is a rather strange looking reaction, but at high enough temperatures it will Universe decreases. Increasing the
take place. The argument goes as follows: firstly, we note that the entropy of temperature reduces the magnitude of
the entropy change of the surroundings
the system increases as gases are formed from a solid. Secondly, the reaction is
and, as shown on the right, eventually
endothermic as in forming the gaseous atoms strong interactions in the crystal the entropy change of the Universe
lattice are broken up. So, the reaction is analogous to the dissociation of N2O4 becomes positive.
and will only take place when the temperature is high enough. In the case of
MgO this turns out to be a very high temperature as the reaction is much more
endothermic than is the dissociation of N2Qx4.
In summary, to be allowed by the Second Law a process which leads to a
decrease in the entropy of the system must be exothermic. A process which
leads to an increase in the entropy of the system can be either exothermic or
endothermic. The temperature of the surroundings is also important as this
affects the magnitude of the entropy change of the surroundings.
2.6 How do we know all this is true?

In this chapter we have simply asserted the truth of the Second Law of Ther-
modynamics and have also introduced a definition of entropy without much
justification. How do we know that these ideas and definitions are correct, and
where do they come from? To answer these questions is by no means a simple
task — it would involve a detailed study of thermodynamics, and we simply do
not have room to go into this now. So, for now we ask you simply to accept
these assertions.
Like all physical laws, the truth of the Second Law of Thermodynamics
rests On experimental observations and these show that the Second Law is a
very strong law indeed. In fact no process which is in violation of the Sec-
ond Law has ever been observed. We can happily apply the Second Law to
understanding chemical reactions in the knowledge that it will not let us down.
hy
2.7 The Second Law and Gibbs energy

It is not really convenient to have to think about the entropy change of the
system and the surroundings in order to apply the Second Law. A more con-
venient approach is to define a quantity called the Gibbs energy which can be
computed from properties of the system alone. We will show how the change
in Gibbs energy is related simply to the change in the entropy of the Universe
so that it is possible to apply the Second Law once the change in Gibbs energy
is known.
We saw before that we can separate the entropy of the Universe into a part
referring to the surroundings and a part referring to the system:
ASuniv = A Ssurr + ASsys. (2.2)
We also saw that the entropy change of the surroundings can be computed from
the amount of heat absorbed by the surroundings
ASsurr = dure > (2.3)

SuIT
and that this heat absorbed is opposite to that absorbed by the system
Ysurr = —sys:- (2.4)
If we used qsurr from Eq. 2.4 in Eq. 2.3 and then put the resulting expression
for ASsurr into Eq. 2.2 we have a new expression for A Syniy
A Suniv = —4sys + ASsys (2.5)

sys
in which we have assumed that the system and surroundings are at the same
temperature, so Tsyy = Tsun (we will do this from now on). The nice thing
about Eq. 2.5 is that all of the quantities on the right-hand side refer to the
system — we do not have to worry explicitly about the surroundings.
Under conditions of constant pressure, sys 18 identical to the enthalpy
change of the system, given the symbol AH. Chemists virtually always work
under conditions of constant pressure, so the identification of Qsys with the en-
thalpy change is certainly appropriate. Using this, Eq. 2.5 becomes
—AFsys
A Suniv = + ASsys. (2.6)
sys
We now introduce the Gibbs energy, given the symbol G. For a process the
change in the Gibbs energy of the system, AGsys, is defined as
AGgys = A Hgys _~ TsysA Ssys (2.7)

2.7 The Second Law and Gibbs energy 17
where Tsys is the absolute temperature, in kelvin. For reasons which will be-
come clear in a moment, we divide both sides of Eq. 2.7 by —Tsys to give
+ ASsys- (2.8)
—AGsys _ —AHgys
Tsys Tsys
Comparing Eq. 2.8 with Eq. 2.6 we have
—AG
A Suniv = —_—*
sys
In words, —AGsys/Tsys is exactly the same thing as the change in entropy of Remember that Tsys is in kelvin and so
the Universe. A process which is allowed by the Second Law has a positive is always positive.
value for ASuniv and therefore has a positive value for —AGsys/Tsys; hence
AGsys must be negative. A positive value for AGgys implies a negative value
for ASuniy, and so such a process would not be allowed by the Second Law. Processes which are allowed by the
We see that once AGgys is known, we can simply inspect its sign to deter- Second Law of Thermodynamics have
ASuniy > 0 which is the same thing as
mine whether or not the process is allowed by the Second Law. The convenient having AGsys < 0.
thing about this approach is that AGsy; is computed only from properties of the
system — we do not need to worry directly about the surroundings. Indeed, from
now on we will drop the ‘sys’ subscript and simply assume that we are dealing
with the properties of the system.
Making ice — again!

To illustrate the use of the Gibbs energy we will go through the same calcu-
lation as in Section 2.4 on p. 10 about making ice. All we need to do is to
compute AG for the process water — ice using
AG = AH —TAS.
We already know that the heat absorbed by the system in going from wa-
ter to ice is —6010 J mol™!; in fact this value is for the change at constant
pressure and so is equal to AH. We also know that the entropy change is
—22.0J K~! mol7! and so it is easy to compute AG. At 5 °C (278 K) this is
AG = AH —TAS
= —6010 — 278 x (—22.0)
— 106 J mol7!.
Note that the units of the Gibbs energy are joules, or for molar quantities, joules
per mole. We see that at 278 K, AG is positive; this implies that the process
cannot take place as it is in violation of the Second Law.
At —5 °C (268 K) the calculation is
AG = —6010 — 268 x (—22.0) = —114J mol™!.
Now AG is negative and so the process can take place as it is in accord with the
Second Law. The conclusions are as before, but the calculations are somewhat
easier as we only have to think about the system, not the surroundings.
2.8 Gibbs energy in reactions and the position of

equilibrium
In Section 2.1 on p. 6 we spent some time pointing out that energy changes are
not the criterion for deciding whether or not a reaction will ‘go’. We have now
seen that the criterion for whether or not a process can take place is the sign
of ASuniy which is reflected in the sigmof ‘AG. If it is negative, the process is
allowed by the Second Law and so can take place; if it is positive, the process
is not allowed by the Second Law and so cannot take place.
For a chemical reaction AG means ‘Gibbs energy of the products — Gibbs
energy of the reactants’, just in the same way that AZ is defined. So, if AG is
negative, it means that the products have lower Gibbs energy than the reactants.
The driving force for a chemical reaction can therefore be described as going
to products which have lower Gibbs energy than the reactants. Note that it is
the Gibbs energy which has to be lowered, not the energy.
There are two contributions to the Gibbs energy: an enthalpy term, AH,
and an entropy term, AS:
AG = AH -TAS.
A negative value of AH (i.e. an exothermic reaction) and a positive value of

AS both contribute to making AG negative, and hence a reaction which ‘goes’.
If AZ is large and negative, the sign of the entropy term is unimportant as the
—T AS term is swamped by the AH term. This is why it seems that having
a large negative AA (i.e. a strongly exothermic reaction) results in a reaction
which goes to products.
It is for this reason that in much of the discussion that follows we will
be concentrating on energy (enthalpy) changes and trying to identify reaction
pathways which lead to the greatest reduction in energy. The chances are that
such reactions will be favoured because of their negative AH values. The
entropy term plays a much more important role when it comes to reactions
involving ions in solution, a topic which we will consider in Chapter 12.
The position of equilibrium

At the beginning of this chapter we pointed out that the important thing to
understand about a reaction is the position of equilibrium. So far we have in-
troduced the idea that by inspecting the sign of AG we can determine whether
or not a process is allowed by the Second Law. We have not, however, ex-
tended this to the more subtle point of determining the position of equilibrium.
We will do this in Chapter 8, but for the moment we will simply concentrate
on looking at reactions from the point of view of energy changes.
2.9 Where does the energy come from?

The energy change which accompanies a reaction is due to the making and
breaking of chemical bonds: energy is given out when a bond is formed, and
taken in when one is broken. As energy changes are very important in deter-
mining whether or not a reaction will go, we need to understand the factors
which influence the energies involved in making and breaking bonds.
2.9 Where does the energy come from? 19
There are two extremes of bonding: one is when electrons are shared
equally between two atoms in a covalent bond, the other is when two charged
groups are held together by electrostatic interactions (i.e. the attraction between
opposite charges). The vast majority of bonds fall somewhere between these
two extremes, with the electrons in the bond being shared unequally between
the two atoms, thus giving a polar bond.
Later on we will discuss the nature of covalent bonding, but first we will
look at the bonding due to purely electrostatic interactions. As we shall see,
such bonds are particularly strong and so their formation is a powerful driving
force for a reaction.
3. lonic interactions
In the previous chapter we saw how energy changes are crucial in determining
whether or not a reaction will take place. There are two major sources of
energy in chemical reactions: interactions between charged species (ions) and
the making and breaking of covalent bonds. Ionic interactions are probably
the easiest to understand so we will discuss these first before moving on to the
more complex topic of covalent bonding. This chapter will only consider the
interactions between ions in solid materials; ions in solution are of course very
important in chemistry but are more complex to describe and so we will delay
this until Chapter 12.
3.1. lonic solids

That solid sodium chloride is made up of positively charged sodium ions and
negatively charged chloride ions is such a familiar idea that we are perhaps in
danger of taking it for granted, so we will spend a few moments reviewing the
evidence for the existence of ions in such solid materials.
The best evidence comes from X-ray diffraction studies which can both
#
\
a
locate the positions of the nuclei and also map the electron density between
f
them. Such studies show that the sodium and the chlorine atoms are arranged
Sy
eae’, in a regular lattice, as depicted in Fig. 3.1, and that the charge on the chlorine
1 and sodium atoms is close to one unit. Furthermore, it is found that between
adjacent sodium and chlorine nuclei the electron density falls to a very low
value. These observations are consistent with the almost complete transfer of
@na Ocr an electron from sodium to chlorine, leading to the formation of ions with little
Fig. 3.1 Representation of part of the
electron density between them.
lattice of solid NaCl as determined by X-ray diffraction studies also allow us to measure the distances, r(MX),
X-ray diffraction. between the ions in the lattice. If we look at this data for the series of salts
MBr and MC], where M is any of the Group I metals (Li, Na, K, Rb, Cs), we
find that the difference r(MBr) — r(MCl) varies by only a few percent as we
change M. Similarly, if we look at the difference r(RbX) — r(KX), where X is
one of the halogens, there is little variation from halogen to halogen.
These observations lead to the idea that each ion can be assigned an ionic
radius which is more or less independent of the salt in which the ion is present.
So, for example, the Na—Cl distance can be found by adding together the ionic
radii of Na* and Cl-; the same value of the ionic radius of Cl~ can be used
to find the K—-Cl distance in solid KCI. That tables of such ionic radii can be
drawn up suggest that, to a large extent, the ions can be thought of as behaving
independently in the solid.
There are some physical properties which are often taken as being indica-
tive of the presence of ions in a solid. For example, such solids are generally
insulators but, on melting, the electrical conductivity rises sharply; this is at-
3.2 Electrostatic interactions 21
tributed to the ions being immobile in the solid but free to move in the liquid.
The energy of interaction between charged species, such as ions, has a
simple form and we will be able to exploit this in the next section to develop a
straightforward expression for the total energy of the ions in a lattice. We will
then go on to see how this expression can be used to understand the energetics
of the formation of a lattice and rationalize the outcome of some reactions
involving solids.
3.2 Electrostatic interactions

In this section we are going to calculate the energy of interaction between the
ions in a lattice just by considering the ions as charged objects. As you know,
oppositely charged objects attract one another — that is there is a force between
them, we call this the electrostatic or Coulomb interaction. This force is found
to be proportional to the charges involved and inversely proportional to the
square of the distance, r, between the two charges. The force is given by
4142
force = 5
4n €or
where qi and q2 are the two charges (in units of coulombs), and ¢9 is
a fundamental constant called the vacuum permittivity (it has the value
8.854 x 1071? Fm7)),
The energy of interaction of two charges, which is what we are interested
in, is given by
4142
energy = (3.1)
Ameor’
which is properly described as the electrostatic (or Coulomb) potential energy.
Note that it varies inversely with the distance in contrast to the force which
varies as the inverse square of the distance.
The charge on an ion can be thought of as the result of an atom gain-
ing or losing a whole number of electrons, which means that the charge
can be expressed as a multiple of the charge on the electron, e (which is
1.602 x 10~!9C). So fora singly charged negative ion the charge is —e, for a
doubly charged negative ion the charge is —2e and so on. Similarly, singly and
doubly charged positive ions have charges of +e and +2e, respectively.
We need to be careful to distinguish the numerical charge on the ion, which
is an integer, and the actual charge on the ion, which is some multiple of e. The
numerical charge of a positive ion will be written as z+ and that of a negative
ion as z.; z and z_ are always positive numbers, even for negatively charged
ions. The charge on a positive ion is z;e and on a negative ion is —z_e; note
the introduction of the minus sign for the negatively charged species.
Putting these values for the charges qi and q2 into Eq. 3.1, the energy of
interaction of a positive and negative ion with numerical charges z4 and z_,
respectively, is
z4z-e*
energy = — (3.2)
4reor
The expression is negative, indicating that there is a lowering of the energy due
to this favourable interaction between opposite charges.
Ionic interactions
22
Figure 3.2 shows a plot of this energy as a function of the distance between
the charges; the closer the two charges come, the more negative the energy. At
ze -ze@ large values of r the energy goes to zero, but on account of the 1/r dependence,
e e this approach to zero is rather slow. The graph shows us that two oppositely
charged ions will move towards each other, if they are able to, since this will
a Mm
®Cc
o lead to a lowering in the energy. We will now use this expression to discuss the
0 energetics of the formation of an ion pair.
Forming an ion pair

Imagine we have sodium and chlorine atoms in the gas phase. Given what we
know about these two elements, we might reasonably expect the sodium to lose
an electron to form Nat and the chlorine to pick this electron up to form Cl”.
Fig. 3.2 Plot showing how the energy of
interaction of two opposite charges The two oppositely charged ions would then be attracted to one another, and
varies with the distance between them. so might form an NaCl molecule which we imagine to be an ion pair, NatCl-.
The energy is negative as the interaction The whole process can be written:
is favourable, and the closer the ions are
to one another the more negative the
energy becomes. Na(g) +CKg) ts Nat(g) +Cl-(g) —+> NatCl“(g).
E) is the ionization energy of sodium, and has the value 502 kJ mol—!; E> is
the electron affinity of chlorine, which is the energy of the process
Cl(g) +e7 —> CI (g).

The value of the electron affinity is —355 kJ mol7!, so E; +E? is 147 kJ mol~!,
The fact that E, + E2 is positive tells us that when a sodium and chlorine atom
are far apart it is not favourable simply to remove an electron from the sodium
and add it to the chlorine.
However, when the two oppositely charged ions come together there is a
favourable energy of interaction, as shown by Fig. 3.2 — this means that E3
will be negative. The closer the ions come together the more negative F3 will
become, eventually cancelling out the 147 kJ mol~! (which is E} + E2) and so
making the overall process exothermic.
The usual expectation for molecules is that there will be a particular dis-
tance, the equilibrium distance, at which the energy of interaction is a min-
imum. However, this will not be the case for these two ions experiencing a
Coulomb interaction; all that happens to them is that the energy goes on and
on decreasing as the ions approach — there will be no minimum in the energy,
which is simply not a realistic situation.
What is missing from our model is that at short enough separations even
ions of opposite charge will begin to repel one another due to interactions be-
tween their electrons. This repulsive interaction is much shorter range than the
Coulomb interaction and typically varies inversely as some high power of r:
]
repulsion energy « —
rn?
where n is between 5 and 12; it is around 8 for Na and Cl.

So what we have is a balance between the electrostatic interaction which
makes the energy more negative as the ions get closer, and this repulsive inter-
action which makes the energy increase. The result is that there will be some
750 -
repulsive term
| 500+
g 250 y energy of the
2 totalenergy tee separated ions
mR 0 7 | ------ energy of the
5 : 4 10 12 14 separated atoms
5 -250b | r/K ——~+
500 of
f
750 Lb /<t— attractive (Coulomb) term

Fig. 3.3 Plot showing the different contributions to the energy of an NatCI~ ion pair as a function
of the distance between the two ions. The attractive Coulomb term is always negative and becomes
increasingly so as the distance decreases. In contrast, the repulsive term is only significant at much
shorter internuclear distances. Adding together these two contributions results in a total energy curve
which has a minimum at around 2.5 A, shown by the dashed line. Zero on the energy scale is defined
as the energy of the two separated atoms. Since we are considering the interaction of two ions the
energy at very large separations tends to 147 kJ mol—! (shown by the dashed line) which, as was
calculated on p. 22, is the energy needed to form the ions from the atoms.
distance at which the total energy is a minimum; this will be the equilibrium
separation of the two ions. These points are illustrated in Fig. 3.3 which shows
how the attractive Coulomb term, the repulsive term and the total energy vary
with internuclear distance for the NatCl pair.
As noted before, the Coulomb term is always negative and falls as the in-
ternuclear separation decreases. The repulsive term is much shorter range and,
for this ion pair, only makes a significant contribution when the distance is less
than about 3 A; however, once it does get going it rises very steeply. Looking 1 A (Angstrom) = 10-1 m or 100 pm.
at the curve for the total energy we see that at large distances it follows the
attractive Coulomb curve and falls as the distance decreases. Then, once the
repulsive term starts to become significant the fall is slowed and eventually re-
versed leading to the formation of a minimum at around 2.5 A; this corresponds
to the equilibrium separation of the ion pair.
From the graph we can see that forming the ion pair from the atoms is
significantly exothermic and so, although the entropy change is likely to be
unfavourable, we can expect that the reaction will go. Our simple electrostatic
model therefore accounts very nicely for the formation of NaCl molecules in
the gas phase as ion pairs. However, this was not really what we set out to do
— we are trying to work out the energy of a solid lattice — but as we will see in
the next section, this can be done by quite a simple modification of the ion pair
case.
Forming a lattice
In a lattice there are many ion—ion interactions which we need to take into
account if we are going to estimate the energy using this model. However, as
the ions are in a regular array it turns out not to be too difficult to work out
what the total electrostatic energy is. .
Look at the NaCl lattice depicted in Fig. 3.4 and concentrate on the Nat
ion at the very centre, which is shown cross-hatched. Its closest neighbours, at
a distance ro, are six Cl~ ions arranged at the corners of an octahedron. The
energy that these six interactions contribute is just six times that of an ion pair
separated by ro:
Z47-e"
—6 x
47 €0ro
The next nearest neighbours are 12 Nat ions, shown in black; these are at
the opposite comer of a square to the central Nat and so are at distance /2ro
(see Fig. 3.5). As these ions are positive their interaction with the central Nat
leads to an increase in the energy (i.e. it is a repulsive interaction), so the term
we have to add is positive:
2 2
Z4+Z2-—€ Z4+2—€
—6 x ote’ 412 | ~te "= ],
4m €0ro 4x eov 2ro
interaction with six CI™ at ro interaction with 12 Na* at /2ro
The next set of ions are the eight Cl”, and these are across the body diago-
nal from the central Nat and so the separation is /3r9. These ions are opposite
Fig. 3.4 Part of the lattice of solid NaCl.
in charge to the central ion and so contribute a negative term to the energy:
The closest neighbours of the central
Nat (shown cross-hatched) are the six Z4Z-€ 2 Z42-e 2 Z47_-€ 2
CI~ ions shown in grey, next are the 12 —6 x —_ +12 x (| —8 x | ————
Nat ions shown in black and next after 47 €0ro Ar eg 2ro AnepV3r0
that are the eight Cl- ions shown in ~y
white. The distances between all of these interaction with six CI~ at ro —_ interaction with 12 Nat at /2r9 _ interaction with eight CI~ at /3r9
ions and the central Na* can be
expressed in terms of the separation, ro, We can tidy this expression up by taking out the common factor:
of adjacent ions.
2
. Z4Z-€ 12 8
electrostatic energy = ————— | 6 - —= + —- +... (3.3)
4n €or ( J/2 V3 )
where the ... indicate that we can go on adding more terms as we move further
away from the central ion. The series in the bracket in Eq. 3.3 turns out, after
many terms, to converge on a value of 1.748.
This number is called the Madelung constant, M, and is characteristic of the
particular arrangement of ions in the lattice; typical values of M are between
1.6 and 2.5. Using the value of M we can write the total electrostatic (Coulomb)
energy as ‘
MNaz+2_e*
electrostatic energy = — 4
Weor
where we have multiplied by Avogadro’s constant, Na, so as to obtain the
molar energy. Also, we have written the lattice spacing more generally as r,
rather than as the equilibrium separation, ro.
Fig. 3.5 Part of the lattice of NaC! As with the ion pair, we need to include a term to account for the repulsion
showing how the distances can all be between ions; writing this as + B/r” we have our final expression for the total
expressed in terms of the distance, ro,
between adjacent ions. The closest
energy as
distance between Na* ions is across the
2
diagonal of a square of side ro, giving a total energy = _ MNaz+2-6" + Be
distance of /2rg. The distance between 4m eor rn
Na®* ions and Cl ions which are at We do not know the value of the constant B, nor of the exponent, n. It turns
opposite corners of a cube is the length
out that the total energy is not particularly sensitive to the value of n so we
of the body diagonal, which is /3ro.
can safely set it at some reasonable value. A value for B can be found by
recognizing that the energy must be a minimum at the equilibrium separation,
ro (we will not go into the details here). The final expression for the energy of
the lattice at the equilibrium separation is
MNaz4z-e? (: ")
energy at equilibrium = — — (3.4)
4reoro
which 1s known as the Born—Landé equation.
The importance of the lattice

If instead of forming a lattice we form ion pairs, the energy would be given by
Eq. 3.4 with M set to 1. This is because the only interaction is between ion
pairs and so in Eq. 3.3 the quantity in the bracket will just be 1, representing
the single interaction between oppositely charged ions.
For a lattice, M is greater than 1 which means that its formation is more
energetically favourable than the formation of ion pairs. The reason for this is
that in the lattice it is not just two ions which are interacting but many. True,
some of the interactions in the lattice are between ions of like charge and so are
repulsive, but on balance the favourable interactions between opposite charges
win.
The example of forming NatCl~ ion pairs which we discussed on p. 22
illustrates these points nicely:
Na(g)
+ Clg) —!*#25 Nat(g)
+ CI“(g) —2+ NatCl(g).
As explained before, EF) + E2 is 147 kJ mol7!. If we assume that the bond
length in Nat Cl7 is the same as the equilibrium separation in the lattice (2.8 A)
then we can use Eq. 3.4 with M = | to find £3 = —417 kJ mol—!, These
values make the overall process exothermic. However, if we allow a lattice to
form:
Na(g)
+ Clg) —2t#25 Nat(g) +Cl-(g) —*> NaCl(s).
E4 is greater in magnitude than E3 by a factor of M (1.75) which is rather
significant. So, forming the lattice is a much more exothermic process than
forming the ion pairs.
The case of forming MgO shows up this difference even more dramatically:
2 E _
+ O?-(g) —2>
+ O(g) —1t#2s Mg2+(g)
Mg(g) Mg?+0?-(g).
Now E] + E> is 2807 kJ mol~! and using an internuclear separation of 2.1 A
we find E3 is —2266 kJ mol~!. Thus, the process of forming the ion pairs is
endothermic and so unlikely to be favourable. However, if we form a lattice
rather than the ion pairs
O(g) its
Ma(g+ ) + O7-(g) —*> MgO\(s),
Mg?*(g)
we find that E4 is now —3961 kJ mol~!, making the overall process exother-
mic. So, forming ion pairs is endothermic, but forming the lattice is exothermic
~anice illustration of the large extra lowering in energy to be had from forming
the lattice.
We should end on a note of caution here. Species such as NaC] and MgO
are known in the gas phase, but their internuclear separations are significantly
less than the corresponding separations in the crystal. This is taken to imply
that the bonding in these simple diatomics is not entirely ionic but has a signif-
icant covalent component. ‘Ny
‘
3.3 Estimating lattice energies

The Born—Landé equation, Eq. 3.4, is an expression for the energy of interac-
tion between the ions in a crystal lattice. Usually, when we are thinking about
the energetics of solid lattices we talk in terms of the lattice energy which is the
energy needed to take the ions from the lattice into the gas phase where they
are so far apart that they are not interacting. For NaCl the process is
NaCl(s) —> Na*(g) + Cl” (g).
The energy of the ions in the lattice is given by Eq. 3.4 and as the gaseous ions
are not interacting at all their electrostatic energy is zero. Hence the lattice
energy is just minus the energy given by Eq. 3.4:
MN
A ITEQrO _e* ( _ -)
lattice energy = MAZHE€ n
1
. (3.5)
Strictly speaking this expression gives the change in what is called the internal
energy, rather than the enthalpy; what is more the energy is for a process taking
place at 0 K. It is possible to correct the value to give the usual enthalpy change
at 298 K, but the correction is small and not really of any significance for
the calculations we are going to do here. We will simply use Eq. 3.5 as an
expression for the lattice enthalpy, which is the enthalpy change of going from
the solid to the gaseous ions.
Let us use Eq. 3.5 to estimate the lattice energy of NaCl. For this crystal
the value of ro determined by X-ray diffraction is 2.81 A, the appropriate value
of n is 8 and, as already mentioned, the Madelung constant is 1.748. These
data are all we need ta compute the lattice energy as 754 kJ mol~!, a value
which compares very favourably with the accepted value of 773 kJ mol~!. It
is remarkable that this very simple model produces such a good estimate of the
lattice energy.
A similar calculation for MgO (ro = 2.00 A) gives the lattice energy as
4043 kJ mol~!. This value is much larger than that for NaCl on account of the
much stronger interactions between the doubly charged Mg?*+ and O27 ions
in MgO. It is striking how large these lattice energies are, indicating that the
interaction between ions is a very significant contributor to the energy changes
in chemical processes.
3.3 Estimating lattice energies 27
Ionic radii
We mentioned right at the start of this chapter (p. 20) that in a crystal it is pos-
sible to assign a radius to each ion, the value of which is broadly independent
of the lattice in which the ion is found. This is very useful to us as it means that
we can compute the equilibrium separation, ro, needed in Eq. 3.5 as the sum of
the radii of the positive and negative ions, ry and r_, respectively:
ro=rat+r_.
So, if we want to work out the lattice energy of a compound, rather than need-
ing to know the value of ro (which would require X-ray diffraction experiments
on the crystal) we can simply look up the ionic radii from tables. Some typical
values of ionic radii (in units of A) are shown in the table below:
cations anions
z= 1 z= 2 z_=1 z=
Lit 0.68 Mg*t 0.68 F- 1.33 O*% 1.42
Nat 1.00 Ca*+ 0.99 CI- 1.82 S*- 1.84
Kt 133 Sr 1.16 Br~ 1.98 Se*- 1.97
Rb*+ 1.47. Batt 1.34 I- 2.20 Te*- 2.17
There are some general trends evident from this table:
e as we go down a group the radii increase;
e for elements in the same row of the Periodic Table, 2+ ions are smaller
than 1+ ions;
e anions are generally larger than cations;
e for elements in the same row of the Periodic Table, 2— ions are larger
than 1— ions.
It makes sense that 2+ ions are smaller than 1+ ions, as when an electron is
removed the remaining electrons are held more tightly and so are pulled in by
their attraction to the nucleus. Similarly, the increased size of anions is due
to the electrons experiencing less attraction to the nucleus. In the following
chapter we will look in more detail at the interactions which are responsible
for these changes.
We should be aware of the limitations of such tables of ionic radii. Usually
they have been ‘adjusted’ to give, on average, the best estimates of the lattice
energy. Different people have different ideas of the best adjustments to make,
so you will find that the values of ionic radii quoted in one book are often
different to those quoted in another! For a particular lattice the value of ro
determined by using these tabulated ionic radii is best regarded as a reasonable
estimate, and so the value of the lattice energy we determine must be treated
with caution. It will not be more precise than a few percent, at best.
Ionic interactions
28
The Kapustinskii equations

The Russian scientist A. F. Kapustinskii noticed that the ratio M/v, where M
is the Madelung constant and v is the number of ions in the molecular formula,
varied rather little from crystal structure to crystal structure. He proposed that a
good compromise was to take the value for this ratio as 0.874, thus enabling the
value of M to be written as 0.874v. Kapustinskii also proposed that a typical
value of n = 9 should be used in. Eqa3.5. With these two simplifications the
expression for the lattice energy becomes:
0.874vNaz4z-e7 (: ;)
lattice energy =
Anmeo(r4, +r—) a)9
where we have replaced ro by (r+ + r_). Putting in the values of all the con-
stants we find the following rather simple expression for the lattice energy
1070 v z4z-
lattice energy/ kJ mol~! = 3.6
(rz +r) (3.6)
where the ionic radii are in A.

For NaC! there are two ions in the molecular formula, so v = 2; from tables
we find the radius for Nat to be 1.00 A and for CI to be 1.82 A. Putting these
numbers into Eq. 3.6 gives us a value of the lattice energy of 759 kJ mol~!,
which is quite close to the accepted value.
Equation 3.6 makes it clear that the lattice energy is dominated by two
things: the charges on the ions and the ionic radii. A large lattice energy will
result if the charges on the ions are large or the ions small.
The Kapustinskii expression for the lattice energy is rather simple and we
will find it very useful in the following examples. However, we should rec-
ognize that the values it gives will be quite approximate and so we must be
cautious about interpreting the results of our calculations in too detailed a way.
3.4 Applications
In this section we are going to look at two reactions in which trends can be
understood using calculated lattice energies.
Stabilization of high oxidation states by fluorine

For many metals it has been observed that whereas the fluorides of high oxida-
tion states are relatively easy to prepare, it is often the case that the other halides
are more difficult or impossible to form. For example, CuF> 1s known, whereas
attempts to prepare the corresponding copper (II) iodide have not been success-
ful. Similarly, for cobalt in oxidation states III and IV, only the fluorides have
been prepared. The story is similar for manganese where for oxidation state
II all of the halides are known but for oxidation state III only the fluoride has
been prepared. There are numerous other examples of this tendency of high
oxidation states only to be found as the fluoride.
3.4 Applications 29
It is thought that the difficulty in preparing these high oxidation state

halides comes about due to the tendency for the halide to decompose into the
lower oxidation state with the release of the halogen. For the case of an oxida-
tion state III metal the reaction we are talking about is
MX3(s) —> MX2(s) + X(g) (3.7)

where M is the metal and X a halogen.
What we are going to do is to estimate the energy change for this reaction
and see what effect changing the halogen has. The analysis will be done using
the following cycle
A Haecomp 1
MX3(s) MX2(s) + 2X2(g)
| A Fhattice (MX3) 8 Fae MX2) | —Y2DO%2)
_ —AH(M2+-M?3+) _
M**(g) + 3X7 (g) ——— > MP*(g) + 2X(g) + X(g)
Starting on the left the solid MX3 is dissociated into its constituent ions in the
gas phase: the energy needed 1s the lattice enthalpy of MX3. The M?+ ion then
gains an electron from one of the X~ ions to give M?+ and an X atom: the
energy needed is minus the third ionization energy of M, minus the electron
affinity of X, EA(X). Finally, the M*+ and X7~ ions are recombined to form
solid MX», for which the energy change is minus the lattice enthalpy of MX2,
and the X atom recombines to form an X2 molecule, for which the energy
change is minus half the dissociation energy of X2, D(X2).
Overall, the enthalpy change for decomposition, A Hdecomp; 1S
A Haecomp =A Hiatice(MX3) — AH (M?7+ > ' MP*) (3.8)

— EA(X) — A Mattice(MX2) — 2D(X2).
We can find the dissociation energies, ionization energies and electron affini-
ties from tables. The lattice energies can be estimated using the Kapustinskii
formula, Eq. 3.6. For MX3, the number of ions, v, is 4, z+ = 3 and z— = 1, so
the lattice energy is given by
1070x4x3x1 12840
A MatticeMX3)/
lattice ( 3)/ kJ mol! (ry3+ + rx-) = (ry3+ + rx-)
For MX2, v = 3, z4 = 2 and z_ = 1, so the lattice energy is given by
1070x3x2x1l _ 6420
A Aattice(MX2)/ kJ mol! =
(Tyq2+ +rx-) (ryt +rx-)
We will complete the calculation for the specific case of cobalt, for which
the 3+ ion has a radius of 0.58 A, and the 2+ ion a radius of 0.71 A. These data
enable us to compute the required lattice energies and thus we can complete
the calculation for all the halogens as shown in the table below.
Ionic interactions
30
F Cl Br I
ry-/A 1.33 1.82 1.98 2.20
1 AMhatice(CoX3) / kJ mol7! 6723 5350 5016 4619
2 AH(Co** > Co3t+)/kJ mol7! 3238 = 332383238) 3238
3. EA(X7)/kJ mol7! —334 -—355 —331 —301
4 AF hanice(COX2)/kJ mol7! * ~ 3147 2538 =. 2387 = 2206
5 %yD(X2)/kJ mol"! 79 121 112 107
6 AHéecomp / kJ mol7! 592 -192 -—390 -—631
In accordance with Eq. 3.8, the value on line 6 is computed by subtracting

the values on lines 2 — 5 from that on line 1; these different contributions to
AH decomp are visualized in Fig. 3.6. We see clearly from the table that the
decomposition reaction
CoX3(s) —> CoX2(s) + X2(g) (3.9)
is exothermic for all of the halogens except fluorine. This provides a neat
explanation as to why it is possible to prepare CoF3 but not any of the other tri-
halides as these can decompose to the lower oxidation state via an exothermic
reaction.
In this reaction there is certainly an increase in the entropy as a gas is
formed from solid reactants. However, this favourable entropy change is un-
likely to outweigh the large positive enthalpy change in the case of the fluoride.
For the other halogens, both the increase in entropy and the exothermicity im-
ply that the reaction will go to products.
Just why it is that the fluoride is so different to the other halides can be
appreciated by looking at Fig. 3.6. What the diagram shows is that in compar-
8000
6000 |
4000 F
|| | |
/ kJ mol"!
2000 |} | | |
| |
wo IT CT
energy
-4000 <
Fig. 3.6 Visualization of the different contributions to AHyecomp for the decomposition of CoX3 accord-
ing to Eq. 3.9. The bars indicate the size of the different energy terms given in the table above, with the
numbers on the bars corresponding to the lines in the table. The quantities from lines 2 — 5 are plotted
as minus the values shown in the table, as this is how they appear in the calculation of AHdecomp
according to Eq. 3.8. Note how the lattice energies (numbered 1 and 4) and the ionization energy
(numbered 2) are the dominant terms. The dashed lines show how the lattice energies of the chloride,
bromide and iodide follow a steady trend, whereas the values for the fluoride are out of line with the
others, being rather larger than the trend would suggest. It is this which makes the decomposition
reaction for the fluoride endothermic.
3.4 Applications 31
ing one halide with another the lattice energies (lines 1 and 4 in the table) are
the dominant terms (the ionization energy of the cobalt is significant, but is a
constant for this series of compounds). We see that for the iodide, bromide and
chloride there is a steady rise in the lattice energies which is simply due to the
ions getting smaller. However, the increase in the lattice energy for the fluoride
is much greater than the trend would suggest — something we can attribute to
the unusually small size of the fluoride ion compared to the other halogens.
Figure 3.6 also shows us that as the halide ion gets smaller the increase
in the lattice energy of CoX3 is greater than for CoXo; the different slopes of
the dashed lines highlight this observation. It is easy to see why there is this
difference by looking at the expressions for the lattice energies:
12840
A Miattice(COoX3) =
(To93+ + 1x-)
6420
A Ahattice(COX2) =
(rege trx-) |
On account of the larger number in the numerator for CoX3, changing the
radius of the negative ion will lead to a greater change in the lattice energy than
for CoX2. So, when it comes to the fluoride the increase in the lattice energy is
significantly greater for CoF3 than for CoF?; ultimately it is this large increase
which tips the balance making the decomposition reaction endothermic for the
fluoride.
Peroxides and superoxides of the Group I metals

Our final example is concerned with the range of different compounds that are
produced when Group I metals combine with oxygen. Lithium gives the sim-
ple oxide Li2O, sodium gives mainly the peroxide, Na2O2, and metals further
down the group give mainly the superoxide MOz (the anion present in the per-
oxide is O57 and in the superoxide the anion is O, ). The trend is for the metals
with the larger ions to form the superoxide — an observation which we will be
able to rationalize using lattice energies.
Let us consider the reaction in which the superoxide decomposes to the
peroxide plus oxygen gas; this is presumably the reaction by which the super-
oxides of sodium and lithium would decompose, were they to be formed:
2MO2(s) —> M202(s) + O2(g).
We can analyse this reaction using the following cycle:
AR, decomp
2MO2(s) > M202(s) + O2(g)
| 2A MMattice(MO2) 8 Fiase(M202) |
2M+(g)+20;(g) ——> 2Mt(g)+057(g) + O2(g)

The quantity simply denoted x is the enthalpy change on going from the super-
oxide anion to the peroxide anion plus gaseous O2; for the present discussion
we do not need to know its value.
Ionic interactions
32
The enthalpy change of the decomposition reaction, AAgecomp, Can be

computed from:
A Adecomp = 2A Aattice(@MO2) + x — A Mattice(M202). (3.10)
As before, we can use the Kapustinskii expression (Eq. 3.6 on p. 28) to estimate
the lattice energies: hy >
_, 1070x2x1xI1 2140
A Miattice MO2)/ kJ mol” = Oe ros) = tae + ros)
_; 170x3x1x2 6420
A Mattice(M202)/ kJ mol” = rae + ron) = ue + roe)
Substituting these expressions into Eq. 3.10 we have
4280 6420
A Hdecomp =
(rm+ + ror) (rm+ + ro2-)
— re /
2A Miattice (MO) A Matice(M202) 4
The radii of the O57 and O; ions are not that much different, so the quantity in
the square bracket is negative on account of the larger numerator of the second
fraction. Now imagine what happens as the size of the cation Mt is increased;
both lattice energy terms decrease but the one for M2O2 decreases more than
the one for MO? on account of the larger numerator in the second fraction. The
term in the bracket therefore becomes less negative as the radius of the metal
ion increases.
We therefore conclude that the trend is for the decomposition reaction of
the superoxide to become less favourable in energetic terms as the size of the
metal ion increases. This rationalizes the observation that the superoxide is not
found for the early members of Group I, but 1s found for the elements toward
the bottom of the group which have the larger ions.
The argument here is not as detailed as in the previous example. We have
not computed explicit values of the A Hgecomp for the different metals and then
compared them with one another. Rather, recognizing from the first example
that the variation in the lattice energies is usually the crucial factor, we have
simply looked at the way in which these terms will vary as the ionic radii are
changed.
3.5 lonic or covalent?

We mentioned at the start of this chapter that X-ray diffraction studies of crys-
tals such as NaCl reveal that the electron density falls to a very low value
between adjacent ions, and we cited this as evidence for the existence of ions
in such crystals. However, if we look more closely at the electron density in
‘ionic’ solids such as LiF, NaCl and KCI we find two things. Firstly, although
the density between ions of opposite charge does become very low it does not
3.5 Ionic or covalent? 33
Fig. 3.7 Experimental electron density map for LiF determined using X-ray diffraction. The electron
density is indicated by contours and the map shows the density in what would be a horizontal plane
through the schematic structure shown in Fig. 3.1. Note that the electron distribution around the lithium
is not spherical, which is taken as evidence of some covalent character in the bonding. (Reproduced,
with permission, from H. Witte and W. Woelfel, Reviews of Modern Physics, 30, 53, (1958). Copyright
(1958) by the American Physical Society.)
actually fall to zero. Secondly, the distance between the nucleus and the min-
imum in the electron density does not correspond to the tabulated value of the
ionic radius. These two observations cast doubt on whether these solids really
do contain ions in the way we have assumed.
Close inspection of the electron density map, such as the one shown in
Fig. 3.7 for LiF, reveals further problems. The contours of electron density
around the lithium are not circular; if the lithium really was an ion its electron
density would be spherical and so the contours would be circular. This dis-
tortion is taken as evidence that the lithium is not truly an ion but that there
is some sharing of electrons with the fluorine — in other words there is some
covalent contribution to the bonding.
Further evidence for there being some covalent contribution to the bond-
ing in these apparently ‘ionic’ solids comes from comparison of the calculated
and experimentally determined values of the lattice energies (remember that
such energies can be determined experimentally using a Born—Haber cycle).
Whereas compounds such as NaCI and KI show good agreement between the
experimental and calculated values, for compounds of Ag(I) (e.g. AglI) and
Ti(1) (e.g. TIBr) the experimental lattice energy is significantly in excess of the
calculated value. As the calculation only considers the electrostatic contribu-

tion, this discrepancy is taken as indicating the presence of additional covalent
contributions.
In reality what we have is a continuum of types of bonding. At one extreme
there is purely ionic bonding and at the other there is purely covalent bonding.
Compounds such as LiF lie pretty close to the purely ionic end of the scale,
whereas solids such as diamond and gsgphite are purely covalent . Everything
else lies in between and so will have both covalent and ionic contributions to
the bonding; which is dominant is sometimes quite hard to decide!
Unfortunately, all of this casts our lattice energy calculations into further
doubt. However, we can continue to use these and to think about such solids
as being ‘ionic’ provided that we realize that this is just an idealization and so
treat our conclusions with some caution.
We now need to turn to the question of how to describe covalent bonding
which, along with ionic interactions, is the other important source of energy in
chemical reactions. As you know, it is the interactions between charged nuclei
and electrons which are responsible for forming the covalent bonds which hold
molecules together, so the first topic we must address is how to describe the
behaviour of electrons in atoms and molecules. The next three chapters are
devoted to this subject.
4 Electrons in atoms
A chemical bond involves the sharing of electrons, so to understand how bonds

are formed we need to understand where the electrons are and what they are
doing; we will start this process by first talking about electrons in atoms. You
probably know quite a lot about this already and are used to describing the
arrangement of electrons in an atom in terms of which orbitals are occupied.
Each orbital has a distinct energy and a particular shape, and we will see that
when it comes to describing how bonds are formed these two properties are
crucial. The material covered in this chapter is therefore vital to our under-
standing of chemical bonding.
The theory we need to describe the behaviour of electrons in atoms and
molecules is quantum mechanics. We certainly do not have the space in this
book to go into the background and mathematical formulation of this theory.
Rather, we will simply introduce the key ideas and then present the results of
quantum mechanical calculations in a pictorial way. This approach will bring
us quickly and without too much effort to the really important points — the
shapes and energies of the orbitals.
Many of the ideas of quantum mechanics seem rather unusual, even
counter-intuitive. This is because the world which we experience directly is
governed by Newton’s Laws (classical mechanics) which are quite different
from the principles of quantum mechanics which apply to tiny particles such
as electrons. For the moment, we will simply ask you to accept the sometimes
rather strange world of quantum mechanics. If you go on to study this in more
detail you will see where these ideas come from and how the predictions of the
theory have been tested against experiments and found to be correct.
4.1 Atomic energy levels

One of the most important predictions of quantum mechanics is that the energy
of an electron in an atom or molecule cannot take any value but is restricted
blue yellow red
to a specific set of values; the energy is said to be quantized. This prediction
is entirely at odds with our direct experience of the world. For example, when
we ride a bicycle we expect to be able to go at any speed (that is, have any 500 600 700 800
kinetic energy). We do not expect to find that our speed, and hence energy, is wavelength /nm
restricted to particular values — but this is exactly what happens to the electron.
Fig. 4.1 Schematic emission spectrum
Each allowed value of the energy is called an energy level, and we imagine that from atomic sodium, such as would be
a particular electron ‘occupies’ one of these levels. seen by passing the light from a sodium
The most direct evidence for the existence of atomic energy levels comes discharge lamp through a prism or
reflecting it off a diffraction grating. The
from the study of atomic spectra, which involves measuring the wavelength of most striking thing is that the emissions
the light emitted by excited atoms. Such spectra show that light is only emitted occur at a series of sharply defined
at certain wavelengths, and that these wavelengths are highly characteristic of wavelengths. in the case of sodium, the
strongest by far is the yellow line at
the atom being studied. Figure 4.1 shows part of such an emission spectrum 589 nm.
Electrons in atoms
36
that the
from sodium; such spectra are often called line spectra to indicate
emissions are confined to a small number of well-defined wavelengths.
To understand how line spectra provide evidence of quantization of energy
we first need to recall the relationship between the wavelength of light, A, and
its frequency, v:
Av=c
A my
where c is the speed of light. Light can be thought of as being composed of

photons whose energy, E, depends on their frequency in the following way:
E=hv
where h is Planck’s constant. It therefore follows that the energy and wave-
length of a photon are related by
c
E=h-.
+ Xv
energy
photon
So, light of a particular wavelength corresponds to photons of a particular en-
-—O- ergy.
Fig. 4.2 Atomic emission spectra are
Our interpretation of atomic line spectra is that an electron in an excited
produced when an electron (represented atom drops down from one energy level to another, releasing a photon whose
by the circle) drops down from a higher energy matches the difference in energy between the two levels (Fig. 4.2). As
energy level to a lower one. This releases
a photon whose energy matches the
the photons emitted from a particular atom are at sharply defined energies, the
energy difference between the two levels. energies of the levels themselves must also have fixed values. If the electrons
The energy of the photon therefore in an atom were allowed to have any energy, any change in the energy of the
depends on the energy leveis in the atom
and the observation that only photons of electron would be possible and so light would be emitted at all wavelengths.
specific energies are produced is taken The observation of atomic line spectra is thus direct evidence for the existence
to imply the existence of energy levels. of atomic energy levels.
Atomic orbitals
The energy levels which the electrons can occupy in an atom are called atomic
orbitals. The electronic configuration for an atom is generated by filling up
the orbitals, starting from the lowest energy one and moving up in energy until
all the electrons are used up; each orbital can accommodate two electrons, one
with spin ‘up’ and one with spin ‘down’.
The lowest energy orbital is the Is, and this single orbital forms the first
shell, or K shell. This shell can accommodate up to two electrons giving the
possible configurations 1s! and 1s* which correspond to hydrogen and helium,
respectively.
The next in energy is the L shell, which contains the 2s orbital and the three
2p orbitals. Lithium (atomic number, Z = 3) has the configuration 15s*2s! and
neon (Z = 10), in which the L shell is full, has the configuration 15*2s22p°.
The next shell is labelled M@ and contains the 3s orbital, three 3p orbitals
and five 3d orbitals. The elements of the third row of the Periodic Table have
electronic configurations which involve filling this shell.
We will see in the next section that we can use a special kind of spec-
troscopy to find direct evidence for the existence of these orbitals and to mea-
sure their energies. Then, in the following section, we will look more closely
at just exactly what an orbital is and what is meant by the labels 2s, 3p, etc.
4.2 Photoelectron spectroscopy 37
4.2 Photoelectron spectroscopy

Photoelectron spectroscopy (PES) provides a particularly direct way of prob-
ing the energy levels of electrons in atoms and molecules. Such spectra are
recorded by irradiating the sample with light of a fixed frequency high enough
for the photons to ionize electrons from the sample; what we actually measure
are the energies of these ionized electrons. As energy is conserved, it follows
that
energy of photon =
energy of ionized electron + ionization energy of that electron.
So, the higher the energy of the ejected electron, the lower the ionization en-
ergy of the electron which has been ejected; the whole process is illustrated in
Fig. 4.3.
o 3p —-— + (a) Oo “, —O— (b) —-O- 7

oe 3s —O- OO 3s =O— — 9.
22 “a o7
2° | 2p ~o— —° 2p—o— —Q—
= 25 —O— —O— 35 —O— —o—
ad
increasing electron energy
<}
increasing ionization energy
Fig. 4.3 Schematic representation of how a photoelectron spectrum is generated. In box (a) we see
a photon (represented by the curly line) coming in and causing an electron (e~) to be ionized from a
2p orbital. In box (b) a photon of the same energy ionizes a 3s electron which is not so tightly held;
as a result the electron has more energy than is the case in (a) and the corresponding peak therefore
appears at higher electron energy. Note that the scale on the spectrum can be marked in terms of
electron energy or ionization energy, and that these two scales run in opposite directions.
What we see in a typical photoelectron spectrum is a number of peaks at

different electron energies which we interpret as corresponding to the ioniza-
tion of electrons from different energy levels of the atom. It is important to
understand that these peaks are not due to the successive ionization of elec-
trons to give Mt, M2+, M?+ ... ions but are due to the formation of Mt by
the ionization of electrons from different levels.
Itis usual to quote the energies of the electrons and their ionization energies
in units of electron-volts, given the symbol eV. This is a unit of energy, with
1 eV being equal to 1.60 x 1079 J or 96.5 kJ mol7!.
Photoelectron spectra of atoms

The photoelectron spectrum of helium, shownin Fig. 4.4, shows a single peak
at around 25 eV; this corresponds to ionizing an electron from the 1s orbital.
There are no further peaks in the spectrum, which makes sense as there are no
other occupied orbitals.
| | I ] |
50 40 30 20 10
<-
ionization energy / eV
Fig. 4.4 Photoelectron spectrum of helium recottted using X-ray photons as the ionizing radiation.
(Adapted from Fig. 4.1 of ESCA Applied to Free Molecules by K. Siegbahn ef al., p. 24, Copyright
(1969), with permission from Elsevier Science.)
1s 2s 2p
l T PU T I T
880 870 860 50 40 30 20 10
<I ionization energy / eV
Fig. 4.5 Photoelectron spectrum of neon. (Adapted from the same source as Fig. 4.4.)
The heights of the peaks in photoelectron The photoelectron spectrum of neon, Fig. 4.5, has more peaks, which we
spectra cannot be interpreted in a simple
can assign in the following way. The peak with the highest ionization energy
way.
is from the Is electrons; these are the most tightly held. The ionization energy
of these 1s electrons is much higher than that of the 1s electrons in helium
because neon has a much higher nuclear charge and so attracts the Is electrons
more strongly. The two peaks at around 49 eV and 21 eV are due to ionizing
electrons from the 2s and 2p orbitals, respectively. We assign the peaks this
way round as we expect the 2s to be more tightly held than 2p. Indeed, this
spectrum provides direct evidence that the 2s and 2p orbitals have different
energies, the explanation of which we will give later in this chapter.
Finally, in the spectrum of argon, Fig. 4.6, we see peaks due to the 2s and
2p electrons at around 325 eV and 250 eV. These are at much higher ionization
energies than the corresponding electrons in neon due to the increased nuclear
charge of argon; at much lower energies there are peaks due to 3s and 3p.
The peak at around 250 eV is split into two; it turns out that this is due to the
influence of the electron spin, but this is a rather complex matter which we will
not go into here. The peak due to the 1s electrons is at a very much higher
ionization energy (about 3000 eV) and is not shown.
These photoelectron spectra show us in a direct way how the electrons fill
2s 2p 3p
3s
] | Po | I po [ I T
340 330 320 260 250 240 40 30 20 10
<b ionization energy / eV
Fig. 4.6 Photoelectron spectrum of argon. (Adapted from the same source as Fig. 4.4.)
4.3, Quantum mechanics 39
the various orbitals and the way in which the energies of these orbitals vary
from element to element.
4.3. Quantum mechanics

Quantum mechanics is the theory which enables us to predict the energy levels,
and much more besides, of electrons in atoms and molecules. Each energy level
has associated with it a wavefunction which is also predicted by the theory.
These wavefunctions are very important, as using them we can calculate any
property of the system. We will focus on how wavefunctions predict the spatial
distribution of the electrons — something which is crucial to our understanding
of chemical bonding.
What is a wavefunction and what does it tell us?

Before we start thinking about wavefunctions we should just remind ourselves
what a mathematical function is. Let us take as an example the quadratic func-
tion f(x):
T
f(x) =ax? +bx +c. (4.1)
The notation f(x) means that f is a function of the variable x; a, b and c are
just numbers, sometimes called coefficients. If we make a plot of f(x) against
-4 ~ 24
x the appearance of the graph will depend on the values of these coefficients.
Figure 4.7 shows such a graph for the case a = 1, b = —2 andc = —8; with Fig. 4.7 A plot of the quadratic function
given in Eq. 4.1 for the particular values
these values, the function is zero for x = 4 and x = —2. of the coefficients a= 1, b= —-2,c=-8.
Provided a, b and c are known, any value of x can be substituted into the For any value of x the function can be
right-hand side of Eq. 4.1 and, with a calculator, we can evaluate the expression evaluated to give a number; for example
when x = —4, the function evaluates to
to give a number; the function f(x) therefore just returns a number for any 16.0.
value of the variable x. The same idea can be extended to a function of more
variables, for example g(x, y, z) which for any values of the three variables x,
y and z can be evaluated to give us a number.
Quantum mechanics predicts that for the hydrogen atom there is a set of
energy levels each with an associated wavefunction; in fact, these wavefunc-
tions are the familiar orbitals. Not surprisingly, the wavefunctions are three-
dimensional and depend on the distances x, y and z from the nucleus.
What, then, does the wavefunction tell us? One interpretation, called the
Born interpretation, is as follows. Imagine a very small box hovering in space
and centred at a point with coordinates (x, y, z) (Fig. 4.8). The probability
of finding the electron in this small box is proportional to the square of the Xx y
wavefunction at this point. The greater the value of the wavefunction the higher
the probability of finding the electron in the box. Fig. 4.8 The probability of finding the
We have to think about a ‘very small box’ so that we do not need to worry electron in a small box centred at point
(x, ¥, Z) is proportional to the square of
about the wavefunction changing value from one side of the box to another. If the wavefunction at that point. The
you are familiar with calculus you will know that the very small box can be nucleus is at the centre of the coordinate
system and the size of the box has been
described as a volume element.
greatly exaggerated for clarity.
The strange thing about the quantum mechanical description of the electron
is that the theory predicts a probability of the electron being in a small box
at a particular position, rather than saying that the electron is at a particular
position. You may have come across the Bohr model of the atom in which
the electron orbits around the nucleus at a specific distance determined by the
energy. The quantum mechanical picture is quite different to this; there are
no specific orbits, but only a probability.that the electron will be found at a
particular distance from the nucleus.
The square of the wavefunction is properly described as the probability
density. This is because the probability of finding the electron in our small box
is equal to the square of the wavefuncwon'multiplied by the volume of the box:
probability of being in the box = (4.2)

(wavefunction) x volume of the box.
This makes sense, as if we increase the volume of the box we expect the prob-
ability of finding the electron in it to increase. Equation 4.2 is analogous to
the one we would use to calculate the mass of an object of known volume and
density:
mass = density x volume. (4.3)
Comparing Eqs. 4.2 and 4.3 we can see that the square of the wavefunction
takes the same role as the density of the material — hence the description of the
square of the wavefunction as a probability density.
4.4 Representing orbitals

The hydrogen atom wavefunctions (the orbitals) are three-dimensional func-
tions and so we have something of a problem to represent them on paper. It
is very important to understand the shapes of the orbitals as when it comes to
forming bonds these shapes are crucial. So, in this section we will spend some
time discussing different ways of representing the orbitals on paper and the
problems that each of these representations has.
We will start with the wavefunction of the 1s orbital which, like all of the
hydrogen atomic orbitals, is a function of the position in three-dimensional
x
space. It turns out that for atoms it is more convenient to think of the position
Fig. 4.9 In atoms it is convenient to not in terms of x, y and z but in terms of r, the distance from the nucleus (r is
specify the position in space using r, the always positive), and the two angles @ and @ as shown in Fig. 4.9. These two
distance from the nucleus, and the two
angles are like the latitude and the longitude, respectively. The coordinates r,
angles # (which is like the latitude) and ¢
(which is like the longitude). These are @ and @ are called spherical polar coordinates.
called spherical polar coordinates. Expressed in these coordinates, the mathematical form of the wavefunction
of the ls orbital, ¥),(r), is
wavefunction —~
Wis(r) = Ae F";
>
this particular wavefunction does not depend on 6 or ¢. The values of A and

B depend on fundamental physical constants in a way predicted by the theory;
their precise values do not concern us here.
r—
When r = 0 (i.e. at the nucleus) e~” is equal to 1 (as e? = 1) and so the
Fig. 4.10 A plot of the wavefunction for
wavefunction has the value A. As r increases, e~3" decreases from 1 and tails
the 1s orbital against the distance, r, from off towards zero as r gets large enough. These features can be seen in the plot
the nucleus. Recall that r is always of W);(r) as a function of r shown in Fig. 4.10.
positive.
4.4 Representing orbitals 41
The wavefunction falls off as we go away from the nucleus; the same is true
of the square of the wavefunction, so the implication is that as we move away
from the nucleus the probability of finding the electron falls off, eventually
reaching a negligible level. The electron is not localized to a particular region
nor is it going round a particular orbit; rather it is smeared out over space.
Contour plots and shaded plots

Another way of representing the wavefunction is to draw a contour plot of a
cross-section through the wavefunction. Imagine taking the three-dimensional
wavefunction and cutting a slice through it, say in the xy-plane. Then, on a
flat surface, we draw lines connecting points which have the same value of the
wavefunction. This is the same idea as representing the height of the surface
of the Earth using contour lines on a map.
Figure 4.11 (a) shows a contour plot of the 1s wavefunction. We can see
that points with the same value of the wavefunction come out as circles on the
plot. This is what we expect since the value of the wavefunction depends only
on the distance from the nucleus.
(b)
xX —o »Gee
Fig. 4.11 Two different representations of the 1s orbital. Each plot is taken through the xy-plane and
is centred on the nucleus. in (a) contours are drawn connecting points at which the wavefunction has
wavefunction —~
the same value; the distances r;, r2 and rg marked along the dotted line refer to Fig. 4.12. In (b) the
value of the wavefunction is represented using a grey scale; the darker the shade of grey, the larger
the value of the wavefunction.
The way in which this contour plot is related to the plot of the wavefunction
as a function of r is indicated in Fig. 4.12. Here, three distances (r}, r2 and
r3) are indicated, and these same distances are marked on the contour plot. 3 fo a In Vg
The wavefunction clearly decreases in value as we go from r; to r2 to r3, so nucleus
the circular contours with the larger radii correspond to lower values of the
Fig. 4.12 A plot of the 1s wavefunction
wavefunction. showing its value at three distances ry,
Another way of representing this two-dimensional slice is to shade in the rg and rz from the nucleus. The same
plot in proportion to the value of the wavefunction; Fig. 4.11 (b) shows such distances are marked on Fig. 4.11. This
plot also shows that the wavefunction
a representation of the 1s orbital. The darker the shade of grey, the larger only depends on the distance, r, from the
the value of the wavefunction. Such shaded plots often give a more easily nucleus; the direction in which we
appreciated picture of the wavefunction than do contour plots. measure r is unimportant.
Ny
N
~
>
——>
Fig. 4.13 Surface plots of the 1s orbital. The value of the wavefunction at which the surface is drawn
decreases going from left to right, and in the right-hand plot the surface has been chosen so that 90%
of the electron density is within it. The maximum distance along any of the axes is 265 pm from the
nucleus.
Surface plots
The final representation we will consider is a three-dimensional version of the
contour plot, sometimes called a surface plot, examples of these for the ls
orbital are shown in Fig. 4.13. Imagine throwing a net over the function and
then pulling the net tight so that each point on the net is at a position which has
the same value of the wavefunction. The net is.a three-dimensional equivalent
of a line on a contour plot. A computer program is used to make the net look
solid and present us with a three-dimensional view, such as those shown in
Fig. 4.13.
As expected, for the 1s orbital the surface plot turns out to be a sphere,
as the wavefunction depends only on the distance from the nucleus. The size
of the sphere depends on the value of the wavefunction at which we choose
to draw the net. In the case of the 1s wavefunction, the smaller the value we
choose, the larger the sphere will be, as is illustrated in Fig. 4.13. Another way
of choosing the value at which to draw the net is to set it so that some fraction,
say 90%, of the electron density is inside the net; this has been done for the
right-hand plot of Fig. 4.13. Such a plot gives us an idea of the size of the
region of space in which the electron is likely to be found for the majority of
the time.
4.5 Radial distribution functions

It is often useful to know the total probability of finding the electron at a set
distance from the nucleus, rather than the probability of finding the electron ina
small box at a certain position. The way to do this is to imagine a thin spherical
shell of radius r and centred on the nucleus (Fig. 4.14). The total probability
of finding the electron at distance r is equal to the square of the value of the
wavefunction at this distance multiplied by the volume of the shell:
x ya
probability of being in the shell =
Fig. 4.14 Rather than thinking about the (wavefunction)* x volume of the shell.
probability of finding the electron ina
small box it is sometimes more useful to It is important to understand that the volume we are talking about is nor the
think about the probability of finding the
electron in a thin shell of radius r. The
volume inside the shell but of the shell itself.
diagram shows this shell, with a section When we were thinking of the probability of finding the electron in a small
cut Out so that we can see inside. box we argued that the probability was proportional to the volume of the box.
4.5 Radial distribution functions 43
For the same reasons, the probability of finding the electron in the shell is
proportional to the volume of the shell. This brings us to an important point
which is that if we keep the thickness of the shell constant its volume increases
as the radius r increases. We need to take this into account when working out
the probability of finding the electron in the shell.
As we have assumed that the shell is very thin, its volume is simply given
by the surface area of the sphere multiplied by the thickness of the shell. Given
that the surface area of a sphere is 47? we can write
probability of being in the shell =

ae) 2 (4.4)
(wavefunction)~ x [4zr x thickness of the shell | ;
where the quantity in the square brackets is the volume of the shell.
We now define the radial distribution function (RDF) as
RDF = (wavefunction? x 4zr?. (4.5)

Comparing Eqs. 4.4 and 4.5 we see that
probability of being in the shell = RDF x thickness of the shell.
It follows that the RDF tells us the probability of the electron being found in
a thin shell of radius r. It is the most useful function for discussing how the
electron density varies with distance from the nucleus.
Figure 4.15 compares the 1s wavefunction with the corresponding RDF,
the two plots are quite different. The first thing we note is that the RDF goes
to zero at the nucleus (r = 0) even though the wavefunction is not zero at this
point. It is the r? factor in the definition of the RDF (Eq. 4.5) which forces it to
zero at r = OQ. Put another way, the r2 factor is there to account for the volume
of the shell; as r decreases the volume of the shell decreases, falling to zero
when r = 0, making the RDF zero at this point.
(a) r/ Bohr radii (b) r/ Bohr radii

1 2 3 4 5 1 2 3 4 5
1 1 1 1 1 i 1 1
com
2 LL
rs) 1s re O
5 Lo ow fan
®
w
Ss -
0 sp T T 7 0 T T T T T
0 50 100 150 200 250 0 50 100 150 200 250
r/pm r/pm
Fig. 4.15 The wavefunction for the 1s orbital is plotted in (a) as a function of distance from the nucleus,
r,in units of pm (1 pm = 10-12 m); the dashed line is a plot (on a different vertical scale) of the function
r2. Plot (b) shows the radial distribution function (RDF) for the same orbital; as described in the text
the RDF gives the probability of finding the electron in a thin shell at distance r from the nucleus. The
most striking feature of the RDF for 1s is the presence of the maximum at r = 53 pm; this distance is
called the Bohr radius. At the top of both plots the scale for r is shown in units of the Bohr radius.
The second thing to note is that whereas the wavefunction drops off froma
maximum at r = 0, the RDF has a maximum at 53 pm from the nucleus. The
maximum comes about because although the square of the wavefunction falls
off as r increases, the r2 factor (from the volume of the shell) increases (this is
shown as the dashed line in Fig. 4.15 (a)). At small values of r the r? term is
dominant and so the RDF increases with increasing r; at large values of r the
exponential drop-off of the wavefunction dominates, forcing the RDF to zero.
The overall result is the formation of a maximum.
The maximum occurs at r = 53 pm, a radius which is the same as the
lowest energy orbit in the Bohr model of the atom; for this reason this distance
is often called the Bohr radius. From now on we will plot the wavefunctions
and RDFs against the distance expressed in units of the Bohr radius.
It is interesting to contrast this picture which the wavefunction gives us
with the Bohr model. The latter says that the Is electron is in an orbit of radius
53 pm, whereas quantum mechanics simply says that the electron is most likely
to be found at a radius of 53 pm.
4.6 Hydrogen orbitals

In this section we will describe the atomic orbitals which are going to be im-
portant in bonding. To start with we will just consider the orbitals of hydrogen,
and then in subsequent sections we will see how these can be adapted to de-
scribe the orbitals of other atoms.
Using quantum mechanics we can calculate the mathematical form of the
wavefunctions (the orbitals) of hydrogen along with the associated energies. It
turns out that these mathematical forms are specified by a set of three quantum
numbers. These numbers are just integers (1, 2, 3, ...) and — perhaps sur-
prisingly — it turns out that we can say a lot about the shape and energy of an
orbital just from a knowledge of these numbers.
You have already been using quantum numbers, but without perhaps re-
alizing it, as the familiar names of the orbitals, such as ‘1s’ and ‘2p’, give the
values of the quantum numbers, albeit in a somewhat coded way which we will
describe shortly.
The three quantum numbers needed to describe an orbital are:
the principal quantum number, n
the orbital angular momentum quantum number, /
the magnetic quantum number, m,
Each is an integer which can take a range of values; however, we will see that
the values which the quantum numbers can take are sometimes dependent on
one another.
We can draw an analogy between these quantum numbers and the coeffi-
cients a, b and c in the quadratic function of Eq. 4.1 on p. 39. The values of
these coefficients determine the exact form of the graph of f(x) against x, i.e.
where the curve crosses zero and the position of the minimum. In the same
way, knowledge of the values of the quantum numbers enables us to say a lot
4.6 Hydrogen orbitals 45
about the form of the corresponding wavefunction. We will discuss the mean-
ing of each of these quantum numbers in turn.
The principal quantum number

The principal quantum number takes integer values starting from one:
n = 1, 2, 3, ... and its value determines the energy of the orbital, E,,:
—Ry
E, n = n2 .
Ruy is the Rydberg constant whose value is predicted by the theory and has
been confirmed by experiment.
There are two things to note about this expression for the energy. Firstly,
energy
the energies are negative; this is because they are being measured downwards
from the energy of the electron when it is infinitely far away from the nucleus,
i.e. when it has been ionized.
The second point is that as n increases the energies increase (they become {
less negative). So the lowest energy orbital has n = 1, the next has n = 2 and n
so on. Figure 4.16 illustrates how the energies of these orbitals vary with n.
Fig. 4.16 The energies of the hydrogen
In describing an orbital as being 1s, 2p or 3d the number is the value of n, the orbitals depend only on the value of the
principal quantum number. principal quantum number, n, and are
measured downwards from the energy of
the ionized electron, represented by the
The orbital angular momentum quantum number
dashed line. The diagram shows the
The orbital angular momentum quantum number, /, takes values from (n — 1) energies of the orbitals with n = 1 to 6,
down to zero in integer steps. The table summarizes the possible values of / for and is drawn to scale.
some commonly encountered values of the principal quantum number, n.
principal gq. n. (nm) orbital angular momentum q. n. (/)

1 0
2 0, 1
3 0, 1,2
The value of / is usually represented by a letter: / = 0 is denoted by ‘s’,

1 = 1 by ‘p’ and/ = 2 by ‘d’. So, when we talk of a 3d orbital we are referring i 0 1 2 3
to an orbital with n = 3 and/ = 2; likewise a 2s orbital has n = 2 and/ = 0. letter Ss p

Taken together, the values of n and / tel! us about the three-dimensional shape
The table shows the letters used to
of the orbital, which we will discuss later in this section. denote different values of the orbital
angular momentum quantum number, /.
The magnetic quantum number
Finally, there is the magnetic quantum number, m;, which takes values from /
to —/ in integer steps. This means that there are (2/ + 1) separate values of mz;
of these, there are / positive and / negative values plus / = 0.
For example, if ] = 1 the values of m; are 1, 0 and —1; if / = 0 the only
value of m; is zero. The value of m, tells us about the orientation of the orbital
in space.
As the value of the principal quantum number, n, determines the energy,
it is usual to think of orbitals with the same value of n as forming a shell.
The shells with n = 1, 2, 3 are sometimes called the K, L and M shells,
respectively. We will now discuss the orbitals in each of these shells in turn.
The K shell (x = 1)
The K shell has n = 1; for this value of the principal quantum number the only
value of / is zero and hence the only value of m; is zero. So, we have just one
orbital with n = 1, / = 0 and m; = 0 which is the Is orbital. This orbital has
been pictured in detail in Section 4.4 (Figs. 4.11 and 4.13 on p. 41 and p. 42,
respectively).
The 1s wavefunction is sphericak,it‘only depends on r), and we have al-
ready seen in Fig. 4.15 on p. 43 that the RDF has a single maximum at the Bohr
radius.
The L shell (n = 2)
The L shell has n = 2; this means that / can be 0 or 1.
radial The orbital with / = 0 is the 2s orbital, and there is just one of these
Cc
Ks) node corresponding to m; = 0. Like the 1s orbital, the 2s orbital does not depend on
xs)
Cc
=)
the angles 6 and @; it is therefore possible to plot the wavefunction against the
o distance r as is shown in Fig. 4.17.
>
oO
The new feature we see here is that the wavefunction has a radial node;
0 2 4 6 8 10 12 this is a value of r (in this case 2 Bohr radii) at which the wavefunction is zero.
r/ Bohr radii This radial node can also be seen in the contour and shaded plots shown in
Fig. 4.17 Plot of the 2s wavefunction
Fig. 4.18.
against the distance from the nucleus, r. For n = 2 we can also have / = 1, and for this value of / there are three
Note the radial node, which is the point at possible values of m,: 1,0 and —1. These are the three 2p orbitals, depicted as
which the wavefunction goes to zero, at
r = 2 Bohr radii.
surface plots in Fig. 4.19. When compared to Is and 2s, the new feature of the
2p orbitals is that the wavefunctions are no longer spherical but depend on the
angles @ and @.
The 2p, orbital is positive for positive values of the coordinate z, and neg-
ative for negative values of z. When z is zero, the wavefunction is zero for
all values of x and y; the xy-plane is therefore called a nodal plane. Another
way of describing this nodal plane is to look at Fig. 4.9 on p. 40 and note that
1
\
(o>)
/
\
£
4
~~
ine)
radii
Bohr radii
io)
oO
Bohr
Ne)
De)
I
|
~
-
Vy
y
6 4 2 0 2 4 6 —6 “4 2 0 2 4. 6
x / Bohr radii x / Bohr radii
Fig. 4.18 Contour and shaded plots of the 2s wavefunction. In the contour plot positive values of the
wavefunction are indicated by solid lines and negative values by lines with long dashes; the contour
corresponding to the wavefunction being zero (the radial node) is indicated by a line with short dashes.
In the shaded plot no distinction is made between positive and negative values of the wavefunction;
the radial node is clearly visible as the white ring around the intense central core.
rd
aN
REX
ROA ae
RKC
ERO
a!
Fig. 4.19 Surface plots of the three 2p orbitals. The dark grey shading indicates a positive value for
the wavefunction and light grey indicates a negative value. The three orbitals point along the z-, x-,
and y-axes, and so are called 2pz, 2px and 2py, respectively. Each has a positive lobe and a negative
lobe. The 2pz orbital is zero everywhere in the xy-plane; this is called a nodal plane or an angular
node. Similarly, the other two orbitals each have a nodal plane.
the plane in which the 2p, wavefunction is zero is the one which has @ equal
to 90° and @ takes any value. This plane can be called an angular node as it
corresponds to an angle (here @) at which the wavefunction is zero (coinpare
this to a radial node which is a radial distance, r, at which the wavefunction is
Zero).
As the electron density depends on the square of the wavefunction (the or-
bital), the sign of the wavefunction is not of any particular significance. How-
ever, aS we will see in the next chapter, the sign is very important when we
start to allow orbitals to interact with one another, which is what happens when
bonds are formed.
The mathematical form of the 2p, wavefunction is
Wop, = C [cos] |re~™” | (4.6)

where C and D are constants given by quantum mechanics. We can see from
this that if 9 = 90° then cos@ = 0 and so Wp, is zero; this is the origin of the
angular node.
When @ is greater than 0° but less than 90°, cos@ is positive; this corre-
sponds to the positive lobe of the 2p; orbital which points along the +z-axis.
When @ is greater than 90° but less than 180°, cos@ is negative; this corre-
sponds to the negative lobe of the 2p, orbital which points along the —z-axis.
The wavefunction W2p, is a product of two parts: the angular part (given
in the first square bracket in Eq. 4.6) which only depends on the angles 0 and
¢, and the radial part (given in the second square bracket) which only depends
on the distance r. It turns out that all of the orbitals are separable in this way
into a radial and an angular part.
The three 2p orbitals only differ in their angular parts — they all have the
same radial part, which is plotted in Fig. 4.20 along with the radial part of the
2s orbital. Whereas the 2s shows a radial node at r = 2, the 2p has no radial
nodes. Furthermore, the 2s is non-zero at the nucleus‘(r = 0) whereas the
2p is zero at this point; note that having the wavefunction equal to zero at the
nucleus is not counted as a radial node.
Electrons in atoms
48
oo
2p, \
§ |\2s
Sc ue /
QO
fy / 2s \ N
3
g yw 7 2p _ / \ .
a NX
nN 0 / —
0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
r/ Bohr radii r/ Bohr radii
Fig. 4.20 On the eft are plotted the radial parts of the 2s and 2p orbitals; note that whereas the 2s
orbital shows a radial node at r = 2, the 2p orbital has no radial node. At r = 0 the 2s orbital is non-zero
but the 2p is zero (this is not, however, counted as a radial node). The RDFs are shown on the right;
note the presence of two maxima for 2s but only one for 2p. The minimum in the RDF of 2s at 2 Bohr
radii corresponds to the radial node in the wavefunction.
Figure 4.20 also shows the RDFs for 2s and 2p. Again, there are significant
differences. For 2s, the RDF has its principal maximum at 5.2 Bohr radii and
there is a subsidiary maximum at 0.76 Bohr radii; there is also a minimum at
2.0 Bohr radii which corresponds to the radial node in the wavefunction. In
contrast, the RDF for 2p shows just one maximum at 4.0 Bohr radii.
Recall from Section 4.5 on p. 42 that the RDF gives the probability of
finding the electron anywhere in a thin shell. This definition means that the
probability has to be summed over all possible values for the angles @ and
g@. We can therefore talk about the RDF of a 2p orbital even though such
orbitals depend on the angles 6 and ¢, as the way an RDF is calculated involves
summation over all angles.
Comparing Fig. 4.20 with the corresponding plot for 1s (Fig. 4.15 on p. 43)
shows clearly how much larger the 2s and 2p orbitals are when compared to Is.
It is also interesting to note that the principal maximum of the 2s is at a larger
distance from the nucleus than is the maximum for 2p, which might lead us to
conclude (falsely) that if the electron is in the 2s it will be less tightly bound
than if it is in the 2p on the grounds that a 2s electron is “further away’ from
the nucleus. In fact, in hydrogen (and only in hydrogen) the 2s and all three
2p orbitals have exactly the same energy; as was described on p. 45 the energy
depends only on the value of the principal quantum number, n.
Distinct orbitals which have the same energy are said to be degenerate; so
in the L shell there are four degenerate orbitals: 2s, 2p,, 2py and 2p,. This
degeneracy of the 2s and 2p is a result of there only being one electron present
in hydrogen; we will see later on that in atoms with more than one electron
the 2s and 2p are not degenerate. However, even in such an atom the three 2p
orbitals are degenerate.
If we wish to represent the 2p orbitals using contour or shaded plots then
we have to be careful about which cross-section we choose. For example, if
we plotted the xy-plane for the 2p, orbital we would see nothing! Figure 4.21
shows the cross-section taken in the xz-plane (at y = 0) through the 2p, orbital.
The two lobes are clearly visible, as is the nodal plane.
6| 6
|
4 4
>= 2 ©) 5 2
© ©
ofe)
oe oe peu Te4, 74S NS
= Oo
jaa) LM ha AMAAN A jaa)
~N ~2 ereLEELA tA
D Ale ~N —2
-4 eas} -4
Ct
6 ee —6
—§ =4 42> 05F25 45.6 6 -4 2 0 2 4 6

x/ Bohr radii X/ Bohr radii
Fig. 4.21 Contour and shaded plots of the 2p2 wavefunction taken through the xz-plane (at y = 0);
the coding of the contours is as in Fig. 4.18. Note the positive and negative lobes, and the nodal
plane (indicated by the zero contour with the short dashes). In the shaded plot, no distinction is made
between positive and negative parts of the wavefunction.
The 2s orbital has one radial node and no angular nodes, whereas the 2p
orbitals have one angular and no radial nodes. So, all orbitals with n = 2 have
one node of some kind; this is an example of the general rule that the total
number of nodes (angular plus radial) is (n — 1).
The M shell (7 = 3)
For n = 3 there are three values of /: 0, 1 and 2 which correspond to the 3s,
3p and 3d orbitals respectively. As before, there is one 3s orbital and three 3p
orbitals. For the 3d the possible values of m; are 2, 1, 0, —1 and —2 giving
a total of five orbitals. Since the 3s, three 3p and five 3d orbitals all have the
same principal quantum number they all have the same energy — that is they
are degenerate (in the case of the hydrogen atom).
Figure 4.22 compares the radial parts of the 3s, 3p and 3d orbitals; compar-
ing these to the plots shown in Fig. 4.20 on p. 48 shows that the orbitals with
n= 3 are considerably larger than those with n = 2. Looking at both these plots,
~
\
3d
S |I\ y-Fe~.
oO t / = =“
ay
jam
20
Sole
®
\
N\
oe
7 SS====
oo
> _—
ao “~- 73
ad LZ P
ou
0 5 10 15 20 25 0 5 10 15 20 25
r/ Bohr radii r/ Bohr radii
Fig. 4.22 On the jeft are shown plots of the radial parts of the orbitals with n= 3 and on the right the
corresponding RDFs; the wavefunction for 3s goes to a finite value at r = 0, but this is not shown on the
plot. Note that the number of radial nodes for 3s, 3p and 3d’are 2, 1 and0 respectively. The principal
maximum in the RDFs moves in closer to the nucleus as / increases; despite this, in hydrogen all the
orbitals have the same energy.
oe 7]
20F
:E> |
10]
: 0 .
=
a a
S =
—10
—20 : :
—20 —10 0 10 20
Fig. 4.23 Contour and shaded plots for the 3s orbital; the orbital is spherically symmetric and the two
radial nodes are clearly visible in the right-hand plot.
a pattern begins to emerge: the 3s orbital has two radial nodes, the 3p has one
and the 3d has none. If we compare this to the orbitals in the L shell (Fig. 4.20
on p. 48) we see that the number of nodes has increased by one for the s and p.
For the RDFs a pattern is also discernible: the 3s shows three maxima, the
3p shows two and there is just one for the 3d. Note that the position of the
principal maximum moves in towards the nucleus as / increases, just aS was
the case for 2s and 2p.
As before the minima in the RDF correspond to the values of r at which
the radial part of the wavefunction goes to zero. In between two such minima
there must of course be a maximum in the RDF, a feature which we can clearly
see from Fig. 4.22.
Figure 4.23 illustrates clearly the spherical symmetry of the 3s orbital; the
two radial nodes are clearly visible in the shaded plot. Figure 4.24 shows the
form of the 3p, orbital; like 2p, this orbital has an angular node (the xy-plane).
However, in contrast to 2p, 3p also has a radial node, clearly visible as the
white circle on the shaded plot. Inside this radial node the two lobes appear
20 —— 20)
10)
: = 0 ~
8 8 °
N N
—10
—20}. ; 4
—20 —10 0 10 20 —20 -10 0 10 20
Fig. 4.24 Contour and shaded plots for the 3pz orbital. Like a 2p orbital, the 3p basically has a positive
and a negative lobe, separated by a nodal plane. In addition for a 3p orbital there is a radial node
closer in to the nucleus; this node is clearly visible in the shaded plot as a white ring.
4.6 Hydrogen orbitals $1
3dy, 3d,, 3d
Fig. 4.25 Surface plots of the three 3d orbitals whose lobes point between the axes. Each orbital has
two positive and two negative lobes, and has two angular nodes.
Fig. 4.26 Surface plots of the two 3d orbitals whose lobes point a/ong the axes.
again. The other two 3p orbitals are similar to 3p, except that they point along
the x- and y-axes.
Finally we have the five 3d orbitals, which are shown as surface plots in
Figs. 4.25 and 4.26. These five orbitals come in two groups. The first group,
shown in Fig. 4.25, consists of three orbitals each of which lies in a particular
plane and whose lobes point between the axes. For example, take the 3d,,
orbital: this lies in the xy-plane and has two nodal planes (or angular nodes):
the xz-plane and the yz-plane. The other two orbitals in this first group are
similar to 3d, except that they lie in the xz- and yz-planes.
The second group, shown in Fig. 4.26, consists of two orbitals whose lobes
point along the axes. The orbital 3d,2_,2 lies in the xy-plane and has its positive
lobes pointing along the x-axis and its negative lobes along the y-axis. Like
3d,y this orbital has two nodal planes, but in this case they bisect the xz- and
yz-planes.
The orbital denoted 3d,2 points along the z-axis. For this orbital the two
angular nodes are not planes but take the form of two cones about the z-axis
defined by the angles 9 = 54.7° and 125.3° with @ taking any value, as is
shown in Fig. 4.27. Fig. 4.27 Illustration of the two angular
We see that, in accordance with the rule that the total number of nodes is nodes present in the 3d_2 orbital. These
nodes take the form of two cones with
(n — 1), all of the orbitals with n = 3 each have two nodes of some kind. You
angles of 54.7° and 125.3°; the two
should not worry about the names of these d orbitals — they are in fact derived lobes and the ring of the orbital are
from the mathematical forms of these functions when they are expressed in shown in cross section.
Electrons in atoms
52
3dyy
20° c=—_—ae 2om0 7
10 aii ia.
: 8a0 *oe
' «?
£ .
= = j E
> 40 Bie mer
a —20
—20 —10 0 10 20
x / Bohr radii
z/ Bohr radii
-20 —10 0) a 10 7 20 —20 -10 0 10 20

Fig. 4.28 At the top are shown contour and shaded plots of the 3dxy orbital. The ‘clover leaf’ pattern
of the four lobes is clearly visible, as are the two nodal planes (here the xz- and yz-planes). At the
bottom are shown contour plots of the 3d,2_y2 and 3d,2 orbitals. The former also shows the clover
leaf pattern; for the latter the two angular nodes making angles of 54.7° and 125.3° to the z-axis are
clearly shown by the zero-value contour (short dashed lines).
cartesian coordinates (that is, in x, y and z coordinates).

Finally, Fig. 4.28 shows contour and shaded plots of a selection of the 3d
orbitals. The ‘clover leaf’ patterns which the lobes of the 3d,y and 3d,2_ya
orbitals form are clearly visible, as are the two nodal planes. As we have
already noted, for the 3d,2 orbital the two angular nodes take the form of cones
rather than planes, and this can clearly be seen.
The general pattern

If we look back over these orbitals we see that there is a pattern to the way in
which the radial and angular nodes occur. The number of angular nodes (nodal
planes) is / and the number of radial nodes is (n — / — 1); the total number of
radial and angular nodes is therefore (n — 1).
Using these general ideas we can have a good guess at what other orbitals
might look like. For example, the 4s orbital will have three radial and no
angular nodes: it is therefore spherically symmetric like the other s orbitals.
4.7 Atoms with more than one electron 53
Alyz 4fos
Fig. 4.29 Surface plots of two of the 4f orbitals, each of which has three angular nodes. For the orbital
shown on the left these angular nodes are the xy-, xz- and yz-planes; the orbital is thus divided into
eight lobes. For the orbital on the right, cones at angles of 41° and 139° to the z-axis form two of the
angular nodes and the xy-plane forms the third.
A 4f orbital, which has / = 3, will have no radial nodes and three angular
nodes; these can clearly be seen in the surface plots of two of the 4f orbitals
shown in Fig. 4.29.
For the 4f,y, orbital these three angular nodes are the xy-, xz- and yz-
planes; the orbital is thus divided into eight lobes. It is interesting to compare
this orbital with the 3d, (shown in Fig. 4.25) in which the angular nodes are
the xz- and yz-planes. These planes divide the 3d,y into four lobes; for the
4 fxyz orbital the xy-plane is an additional angular node which further divides
these four lobes into eight.
The 4 f,3 orbital, also shown in Fig. 4.29, is reminiscent of the 3d,2 orbital
shown in Fig. 4.26. However, for this 4f orbital the angles of the two cones
are different to those for the 3d,2. The additional angular node, the xy-plane,
effectively ‘cuts’ the 3d,2 orbital in half.
4.7 Atoms with more than one electron

All we have said so far applies to an atom with only one electron, such as hy-
drogen or ions such as Het and Li2+. In a one-electron atom the only interac-
tion we need to consider in the quantum mechanical treatment of the problem
is the electrostatic attraction between the nucleus and the electron (Fig. 4.30
(a)); however, if there are two electrons present (as in He) we also need to con-
sider the electrostatic repulsion between the like charges of the two electrons
(Fig. 4.30 (b)).
Unfortunately, simply going from one electron to two changes the problem
from one which quantum mechanics can deal with exactly to one which cannot
be solved ‘on paper’ (although computer calculations can still solve the prob-
lem to very high accuracy). We want to be able to retain the simple picture
which the hydrogen orbitals provide rather than having to resort to computer
calculations, and this is where the orbital approximation comes in.
The orbital approximation

The idea behind the orbital approximation is to make a multi-electron atom
‘look like’ a one-electron atom. If we achieve this it is then easy to find the
Electrons in atoms
54
(a) (b) (c)
\ ae J oN J °
one-electron atom two-electron atom orbital approximation
Fig. 4.30 In a one-electron atom (a), the only interaction we need to consider is the electrostatic
attraction between the positive nucleus and the nagative electron (represented by the filled circle). Ina
two-electron atom (b), we also need to consider the electrostatic repulsion between the two electrons
(represented by the curly line). The orbital approximation (c), considers that each electron in a multi-
electron atom experiences only an effective nuclear charge, Z,4, whose value takes account of the
effect of electron—electron repulsion. Each electron then effectively forms a one-electron atom as there
is no electron—electron repulsion to consider explicitly.
wavefunctions as the problem is then essentially the same as the hydrogen

atom. Remember that finding the wavefunctions is crucial as once we know
these we can find the probability distribution of the electrons, their energies
and so on.
In the orbital approximation we concentrate on just one of the electrons.
Then, rather than considering the details of its repulsive interactions with all
the other electrons we imagine that the effect of these is simply to alter the
nuclear charge seen by the electron we are interested in from its actual value to
an ‘effective’ value.
For example, in the case of helium the actual nuclear charge is two. How-
ever, as the electrons repel one another they are not as tightly held as would
be expected solely on the basis of a nuclear charge of +2. To account for this
reduced binding of the electrons, we imagine that the electrons are held by an
effective nuclear charge of somewhat less than 2 (Fig. 4.30 (c)).
The orbital approximation allows us to consider each electron as if it were
on its own (a one-electron system) experiencing an effective nuclear charge. It
turns out that in such a situation the orbitals (wavefunctions) for the electron
are of the same form as those in hydrogen, but with their energies and sizes
modified in rather a simple way which just depends on Zegr.
We saw on p. 45 that the energy of the orbitals in hydrogen is a function
of the principal quantum number, n. If the effective charge is Zer, this energy
becomes 2 R
E,=- a Ru
Leg
(4.7)
The Rydberg constant, Ry, is conveniently expressed in electron-volts (eV)
and takes the value 13.6 eV. We see from Eq. 4.7 that increasing the effective
nuclear charge makes the energy more negative, which means that the electron
is more tightly held — this is exactly what we would expect.
We can use computer-calculated values of the orbital energies in multi-
electron atoms to get a feel for the kinds of values that the effective nuclear
charge takes. For example, the energy of the Is orbital in hydrogen is —13.6eV
and so, just as we would expect, the value of Zegr calculated using Eq. 4.7 is
1. In helium, the orbital energy of a 1s electron is —25.0 eV, and using Eq. 4.7
we can work out that this corresponds to Zep = 1.4; the value is considerably
less than the real nuclear charge of 2 on account of the effect of the electron—
electron repulsion.
4.7 Atoms with more than one electron 55
Changing Zegf also affects the size of the orbitals as well as their energies.
Increasing the effective nuclear charge causes the orbitals to contract, which is
hardly surprising as we expect that the increased attraction caused by a higher
nuclear charge will pull the electrons in. It turns out that the effect is inversely
proportional to Zep. So, for example, as is shown in Fig. 4.31 the maximum in
RDF
the RDF for a 1s orbital occurs at (ao /Zeft), where ao is the Bohr radius. All
of the orbitals scale in this simple way.
0 1 2 3 4 5
Electronic configurations r/ Bohr radii
The orbital approximation allows us to imagine that each electron is behav- Fig. 4.31 Illustration of the effect of
ing as if it is in a one-electron atom, and so the wavefunctions of each of the increasing the nuclear charge on the
form of the 1s orbital. The thick line
electrons (the orbitals) are the same as those for hydrogen albeit modified to shows the RDF for a 1s orbital with a
take account of the effective nuclear charge. Each electron can therefore be nuclear charge of 1; the thinner line
assigned to a hydrogen-like orbital. This is the basis for something which you shows the RDF for the same orbital but
with a nuclear charge of 2. Note how the
have already been doing, which is writing configurations for multi-electron maximum shifts from 1 Bohr radius when
atoms. the nuclear charge is 1 to 0.5 Bohr radii
When we write the electronic configuration of helium as 1s? what we imply when the charge is 2.
is that both electrons are in a hydrogen-like 1s orbital. As you know, we can

have up to two electrons in one orbital, provided we oppose the spins, i.e. one
spin is ‘up’ and the other is ‘down’.
For lithium (Z = 3) the configuration is 1s*2s!; we cannot put the third
electron into the 1s orbital as this is already full with two spin-opposed elec-
trons, so it has to go into the next lowest energy orbital, which is the 2s. We can
carry on in this way building up the electronic configuration of any element.
We know that the ground state electronic configuration of lithium is 1s?2s!
and not 1s?2p!, because we have learnt that the 2s orbital is lower in energy
than is the 2p. However, for hydrogen we have already commented (on p. 48)
that 2s and 2p have the same energy — what, then, has changed? The crucial
point here is that, in contrast to hydrogen, in a multi-electron atom the 2s and
2p are no longer degenerate. In the next section we will explore why this is so.
Orbital energies in multi-electron systems

From Eq. 4.7 on p. 54 we see that the energy of an orbital depends on the
principal quantum number and the effective nuclear charge; we should also
recall that this effective nuclear charge is a result of the balance between the
actual nuclear charge and the electron-electron repulsion.
We have already seen that the effective nuclear charge in helium is 1.4, a
value significantly less than the real nuclear charge. For lithium the 2s orbital
energy is —5.3 eV; if we assume that this electron has n = 2 we can use Eq. 4.7
to compute the effective nuclear charge as 1.3. The first thing to notice is that
this is much less than the actual nuclear charge of 3. What this is telling us is
that the two 1s electrons are repelling the outer electron to a very significant
extent.
A convenient way to describe the effect of this repulsion is to use the con-
cept of screening. Let us think about the outer electron in lithium — what does
it see? There is the nucleus with a charge of 3 and this is surrounded by the two
1s electrons, each of which has a single negative charge. If the outer electron
is far enough away, then we might reasonably assume that it would ‘see’ an
effective nuclear charge of +1, i.e. +3 from the nucleus and —1 from each Is
electron giving +1 overall. The idea is illustrated in Fig. 4.372.
We say that the two Is electrons screen the outer electron from the nuclear
charge. If they form a perfect screen, the negative charges of the two 1s elec-
trons will cancel out two positive charges from the nucleus leaving an effective
nuclear charge of +1.
In fact we know that for lithium Zegf is 1.3, and this tells us that the screen-
ing by the two inner 1s electrons is not perfect. This should come as no surprise
as the representation given in Fig. 4.32 does not fit in with the quantum me-
chanical view that the electron is not localized to an orbit, but is smeared out
over space. We can investigate how this might affect the shielding by compar-
ing the RDFs of the 1s, 2s and 2p orbitals; these are shown in Fig. 4.33.
Fig. 4.32 Illustration of the concept of
To make these plots we have assumed that the effective nuclear charge
screening. In (a) the lithium atom is
represented as an outer electron (the experienced by the Is electrons is 2. What this value of Zee does is to contract
filled circle), a nucleus of charge 3 and the 1s orbitals, and this is clearly visible in Fig. 4.33 where the RDF for the Is
the two 1s electrons (the shaded area
forms a compact peak close to the nucleus.
close to the nucleus). If the outer electron
is far enough away, it experiences an Figure 4.33 shows us clearly that much of the probability density of the 2s
effective nuclear charge of 1 as shown in and 2p electrons falls well outside the area occupied by the 1s electrons. It is
(b). We say that the two inner electrons
not surprising, therefore, that the ls electrons do form such a good screen of
screen the outer electron from the
nuclear charge. In practice the screening the nuclear charge, pretty close to the rather crude picture of Fig. 4.32.
is not perfect and so the effective nuclear However, there is some probability of the 2s and 2p electrons being inside
charge will be greater than 1.
the area occupied by 1s. When the electrons are in this region they will expe-
rience a larger nuclear charge than when they are further out. We describe this
effect by saying that the outer electron penetrates the screen formed by the Is
electrons and so experiences a greater nuclear charge. This is the explanation
for why Zegr for the outer electron in lithium is greater than 1.
Figure 4.33 also provides an explanation of why the 2s orbital is lower in
energy than the 2p. Close inspection of the RDFs shows that the 2s has greater
amplitude than the 2p inside the region occupied by Is; the 2s is said to be
more penetrating than the 2p. The effective nuclear charge experienced by the
2s is greater than that for 2p and this is why the former is lower in energy.
The crucial thing about the 2s orbital which makes it more penetrating than
1s
RDF
0 2 4 6 8 10 12
r/ Bohr radii
Fig. 4.33 Plots of the RDFs of the 1s, 2s and 2p orbitals, all to the same scale. For the ts the
effective nuclear charge has been taken as 2 in order to mimic the likely charge experienced by these
electrons in lithium; for the other orbitals the effective nuclear charge has been taken as 1. Most of
electron density of the 2s and 2p falls outside the region occupied by the 1s, and so 2s and 2p are well
shielded from the nucleus. However, there is some probability of the 2s and 2p electrons penetrating
inside the 1s, and this probability is greater for 2s than for 2p. This explains why the 2s Is lower in
energy than the 2p.
4.8 Orbital energies 57
2p is the presence of the small subsidiary maxima close in to the nucleus. The
principal maximum for the 2s is in fact further out than for 2p.
If we compare the RDFs for 3s, 3p and 3d (Fig. 4.22 on p. 49) we see a
similar pattern. The 3s has a subsidiary maximum close in to the nucleus, as
does the 3p, but the latter does not come in as close to the nucleus as the former.
Not surprisingly, therefore, the 3s is more penetrating than the 3p which is in
turn more penetrating than the 3d. The 3s is thus the lowest in energy, followed
by the 3p and then the 3d.
In summary, the energies of the electrons in multi-electron atoms are most
strongly affected by the value of the principal quantum number, n. Within a
given shell (a given value of n) the ordering of the orbitals depends on the
degree of penetration each shows, with those with the lower value of / being
the more penetrating and hence the lowest in energy.
4.8 Orbital energies

We can put all we have learnt about multi-electron atoms together to discuss
the trends in the orbital energies of the elements H to Ne; the data are shown in
Fig. 4.34. Recall that these orbital energies are negative as they are measured
downwards from the energy of the ionized electron which is taken as zero (see
Fig. 4.16 on p. 45). So the lower — meaning more negative — the energy goes
the more tightly the electron is held.
Looking first at the data for the 1s orbital, we see that the energy for helium
is lower than that for hydrogen. We have already discussed why this is so:
the nuclear charge has increased to two and although the electron—electron
repulsion results in an effective nuclear charge of 1.4 this is significantly greater
than that for hydrogen. Thus the 1s orbital is lower in energy in helium than it
is in hydrogen.
Going from helium to lithium we see that the energy of the 1s drops sharply.
We can attribute this to the fact that the third electron in lithium, the 2s, does
not shield the 1s electrons from the nucleus. As we have noted, the 1s electrons
have most of their electron density closer to the nucleus than the 2s and so the
latter is not in a position to provide any shielding. For the remaining elements
H He Li Be B CG N OF Ne
orbital energy / eV
Fig. 4.34 Plot showing the energies of occupied orbitals for the elements H to Ne. Data for the 1s
electron are only shown for the first three elements, as after that the energies fall very steeply. The
lines connecting the data points are of no significance; they just serve to highlight the trends.
the energy of the 1s orbital drops even lower and, as we shall see, such electrons
are so tightly held that they are hardly involved in bonding.
Turning now to the energies of the 2s orbital we see that there is a general
fall as we go across the first row of the Periodic Table. The reason for this
is that the nuclear charge is increasing, resulting in the electron being more
tightly held. Of course, each time the nuclear charge increases an additional
electron is added, but as these electrons are all going into the n = 2 shell they
do not shield one another as well as.fhe Is electrons shield those in the n = 2
shell. This is because the electrons in the n = 2 shell broadly occupy the same
region of space, whereas the |s electrons are well inside the region occupied
by the 2s and 2p electrons.
As we discussed on p. 56 the 2s orbital is lower in energy than the 2p so
the 2s is occupied first. It is only by the time that we get to boron that the 2p
orbital is occupied for the first time. Then, like the 2s, the 2p orbital steadily
falls in energy as we go across the first row of the Periodic Table. This fall in
energy is for exactly the same reason as we gave for the 2s.
Looking at Fig. 4.34 it is striking how the energy separation between the 2s
and 2p increases as we go across the first row. The explanation for this is that
as the 2s is more penetrating than the 2p, the increase in nuclear charge as we
go from one element to the next has a greater effect on the 2s than the 2p. So
while both orbitals drop in energy, the 2s falls more than the 2p. We will see in
the next chapter that this increasing separation of the 2s and 2p has some direct
consequences when it comes to bonding.
As you might expect, the orbital energy of the 3s electron in sodium,
—5.0 eV, is much less negative than for the 2p in neon; the effective nuclear
charge for this electron is 1.8. In sodium the outer electron is now in the n = 3
shell and we argue that this is quite well shielded from the nucleus by the elec-
trons in the n = | and n = 2 shells. As we continue across the second row of
the Periodic Table, we find a pattern very similar to that shown in Fig. 4.34 for
the first row.
4.9 Electronegativity
If we have a chemical bond between two different atoms we expect the bond
to be polarized. For example, in water the O~H bond is polarized towards the
oxygen; we say that the oxygen has a partial negative charge (often written as
6—) and the hydrogen has a partial positive charge (6+).
In a bond, the atom which attracts more of the electron density is described
as being more electronegative. So, we say that oxygen is more electronegative
than hydrogen, on account of the polarity of bonds formed between the two
atoms.
The problem is that there is no measurable quantity which corresponds di-
rectly to electronegativity. Different numerical scales, based on quantities such
as ionization energies, bond strengths and orbital energies, have been proposed;
these scales are usually in broad agreement, showing the same trends. Typi-
cally we find that the electronegativity increases as we go across the first row
of the Periodic Table, e.g. oxygen is more electronegative than nitrogen which
4.9 Electronegativity 59
is in turn more electronegative than carbon.

We will see in the next chapter that the polarity of a bond between two
atoms depends on the relative energies of their atomic orbitals. Generally, the
atom with the lowest energy orbitals will end up with the partial negative charge
— in other words, the more electronegative element is the one with the lower
energy orbitals. Looking at Fig. 4.34 on p. 57 we see that the orbital energies
decrease as we go across the first row, which means that we would expect the
electronegativity to increase, which is exactly what it does.
Electronegativity is a consequence of the energies of the orbitals — the lower
the orbital energy the more electronegative the element. If we are trying to
work out which of two atoms has the lowest energy orbitals it is sometimes
convenient to relate this to the familiar concept of electronegativity and note
that the more electronegative element has the lower energy orbitals. However,
it is important to realize that electronegativity is a consequence of orbital ener-
gies, and not the other way round.
Electrons form the bonds which hold molecules together, and in chemical reac-
tions electrons are rearranged as bonds are made and broken. So, understanding
what the electrons are doing in a molecule is essential for understanding both
chemical structures and reactions.
In this chapter we will introduce the powerful molecular orbital (MO) de-
scription of bonding, first applying it to simple diatomics and then, in the fol-
lowing chapter, extending it to larger molecules. Using the MO approach we
will be able to rationalize many observations about the bonding in molecules,
such as why Hp readily forms but He2 is unknown, why O2 is paramagnetic
(meaning that it is drawn into a magnetic field) but N2 is not and why the bond
in N2 is stronger than that in O2 which in turn is stronger than that in F2. Sim-
pler models of bonding are quite unable to explain these observations in any
satisfactory way — we need the more subtle MO approach.
5.1 The energy curve for a diatomic

Figure 5.1 illustrates how the energy of two hydrogen atoms varies as we bring
them together from a large separation.
When the atoms are far apart there will be no interaction between them and
so the total energy is just the sum of the energies of the two atoms. As we
bring the atoms closer together they start to interact in a favourable way and
the energy falls, eventually reaching a minimum; this distance corresponds to
<4
®
Ol
Cc
Yo
dissociation
energy
. —
internuclear separation
atoms
pushed
equilibrium
separation atoms far apart
together
Fig. 5.1 Illustration of the way in which the energy of interaction of two hydrogen atoms varies with
the distance between them. At large distances there is little interaction and the energy is just the sum
of the energies of the two atoms. As the distance decreases there is a favourable interaction and
eventually the energy reaches a minimum; this is the equilibrium separation. Finally, as the separation
decreases further the atoms start te repel and the energy rises steeply. The energy change between
the equilibrium position and the separated atoms is the bond dissociation energy.
5.2 Molecular orbitals for H 61
the equilibrium bond length of Hp. Finally, if the atoms are brought still closer
together, the energy rises, slowly at first and then more rapidly as the two nuclei
begin to repel one another.
The difference in energy between the bottom of the potential energy curve
and the separated atoms is the dissociation energy; this is the energy needed to
separate the two atoms starting from the equilibrium position. What we have
to do is to explain why it ts that the energy of two hydrogen atoms decreases as
they come together — in other words why a bond is formed; the MO approach
provides the explanation we need.
5.2 Molecular orbitals for Ho

In Chapter 4 we saw that atomic orbitals (AOs) are used to describe the be-
haviour of electrons in atoms; in an analogous way we use molecular orbitals
(MOs) to describe electrons in molecules. As with AOs, we can talk about the
shape and energy of an MO. Also, electrons fill the MOs in molecules in the
same way as AOs are filled in atoms: we start with the lowest energy MO and
work upwards, placing up to two (spin-paired) electrons in each orbital. Our
task is to determine the shapes and energies of these MOs.
In quantum mechanics a common way of solving a complex problem is
to ‘add together’ the solutions to a simpler, but related, problem. We will
use this approach to construct MOs by combining AOs of the atoms which
are forming the bond. The attractive feature of this method is that there is
a smooth transition from the molecule, where the MOs are a combination of
AOs, to the separated atoms, where the AOs are uncombined. Also, we already
know a lot about the shapes and energies of the AOs, and will be able to use
this knowledge to good effect.
This method of constructing MOs is called the linear combination of
atomic orbitals (LCAO) method; the name comes from the fact that the method
simply involves adding together AOs. In the case of H2 we write the MOs as
MO =c] x (AO on atom 1) + c2 x (AO on atom 2)
where c; and cz are just some numerical coefficients whose values are deter-
mined by quantum mechanics.
For hydrogen we need only consider MOs formed from the 1s AOs on the
two atoms. It turns out that when two AOs on separate atoms combine the
result is two MOs. In one MO the two coefficients have the same sign and in
the other they have opposite signs. For a homonuclear diatomic, such as H2,
the coefficients are either +1 or —1, and so the two MOs are:
bonding MO = Is, + 1s2

anti-bonding MO = 1s; — 1s2
where 1s; and 1s7 are 1s AOs on atoms | and 2, respectively. We see that in
the bonding MO the two coefficients have the same sign whereas in the anti-
bonding MO they have the opposite sign. Just why these orbitals are called
bonding and anti-bonding, we will consider next.
bonding MO
A, A IN
tf
NR.
12
anti-bonding MO
A - AL —_ > —\
tf f
12 12 f
AO onatom 1 AO on atom2 12
Fig. 5.2 Illustration of the formation of a bonding and an anti-bonding MO from two 1s orbitals. What
is plotted is a cross-section through the orbitals along the line of the bond; the arrows marked 1 and
2 indicate the location of the two nuclei. The top part of the diagram illustrates the case where both
coefficients are positive; this leads to constructive interference of the two AOs and the formation of
the bonding MO, which has a build-up of electron density between the two atoms. The lower part of
the diagram illustrates the case where the coefficients are of opposite sign; this leads to destructive
interference and the formation of the anti-bonding MO. Note that this MO has a node at the mid-point
between the two atoms.
Bonding and anti-bonding MOs

Figure 5.2 illustrates the formation of a bonding MO and an anti-bonding MO.
When the two coefficients are both positive the two AOs reinforce one another
and form an MO which extends across both atoms; this is the bonding MO.
The overlap between the orbitals is described as being constructive or in phase.
Remember that the square of the wavefunction (the orbital) gives the electron
density, so we conclude that this MO places electron density between the two
nuclei.
In contrast, when the two coefficients have opposite sign, the two AOs
detract from one another in what is called destructive or out of phase overlap.
This leads to the formation of the anti-bonding MO which has a node at the
mid-point along the bond. Figures 5.3 and 5.4 show surface and contour plots
of these two MOs; from these we can see clearly the way in which the bonding
MO envelops the whole molecule, whereas the anti-bonding MO has a nodal
plane between the two atoms.
It turns out that the bonding MO is Jower in energy than the two AOs from
which it is formed; this can be rationalized by noting that the bonding MO con-
centrates electron density in the region between the two nuclei. When the elec-
trons are in this region, they are attracted by both nuclei, and this contributes
to the lowering of the energy of the orbital when compared to the constituent
AOs. The anti-bonding MO is higher in energy than the AOs from which it is
formed. We can rationalize this by noting that in this orbital, electron density
is pushed away from the favourable internuclear region, as evidenced by the
nodal plane between the two atoms.
The relative energies of the AOs and MOs are usually illustrated with an
MO diagram of the type shown in Fig. 5.5. This diagram shows the lowering
of energy of the bonding MO, and the raising of energy of the anti-bonding
5.2. Molecular orbitals for Hz 63
bonding anti-bonding
Fig. 5.3 Surface plots of the bonding and anti-bonding MOs formed from the overlap of two 1s orbitals;
the two nuclei are placed along the z-axis, symmetrically about z = 0, as shown by the dots on the
set of axes at the top of the diagram. Note that the bonding orbital has the same sign throughout
and encompasses the whole molecule. In contrast, the anti-bonding orbital has a nodal plane (the
xy-plane) between the two atoms. The dark grey indicates a positive value of the wavefunction and
the light grey a negative value.
bonding MO anti-bonding MO
Fig. 5.4 Contour plots (taken in the xz-plane) of the bonding and anti-bonding MOs shown in Fig. 5.3.
Positive values of the wavefunction are indicated by solid lines and negative values by lines with long
dashes; the contour corresponding to the wavefunction being zero (the nodal plane) is indicated by a
line with short dashes. The way in which the bonding MO encompasses the whole molecule is clearly
seen, as is the nodal plane which bisects the anti-bonding orbital. The black dots indicate the location
of the nuclei.
MO, when compared to the energies of the AOs. Also shown are ‘cartoons’ of
the MOs which should be compared to Figs. 5.3 and 5.4; these cartoons are not
supposed to be very precise but simply illustrate the relative signs of different
parts of the orbitals.
It is important to realize that this description of the MOs as resulting from
the combination of AOs is just a way of calculating the form of the MOs.
We do not mean to imply that the AOs really do interfere constructively and
destructively to form MOs - rather, this approach is just a way of predicting
and thinking about the MOs.
Just as we can probe the energy levels of atoms using atomic emission spec-
troscopy (Section 4.1 on p. 35) or photoelectron spectroscopy (Section 4.2 on
p. 37), we can also do the same for molecules. Such experiments provide direct
eo
————, anti-bonding MO
e— — , \ —@
energy
ee” bonding MO
AO on MOs AO on
atom 1 atom 2
Fig. 5.5 MO diagram for the Hz molecule; the vertical scale is energy. On the left the horizontal line
shows the energy of the AO from atom 1; the line on the right shows the same for atom 2. In the
middie are shown the energies of the MOs, with the dashed lines showing which AOs are involved in
the formation of the MO. Beside each energy level is a ‘cartoon’ of the relevant AO or MO with the
positive parts of the orbitals shown shaded. The small dumb-beil is there to indicate where the nuclei
and bond are.
experimental evidence for the MO approach and provide a way of measuring

the energies of the MOs.
Predictions from the MO diagram

We can now use the MO diagram (Fig. 5.5) to make some predictions about
the molecules H2, He2 and their ions. What we do is to count up the number
of electrons in the molecule and then place them in the MOs in the usual way.
Occupation of the bonding MO lowers the energy of the molecule relative to the
separated atoms and so favours the formation of a bond. In contrast, occupation
of the anti-bonding MO raises the energy of the molecule relative to the atoms
and so disfavours the formation of the molecule.
Figure 5.6 shows how the MOs are occupied for a series of molecules with
increasing numbers of ‘electrons. The molecular ion H3 has just one electron
which goes into the bonding MO; we therefore predict that the energy will be
lowered when the molecule is formed and so expect Hy to be stable with re-
spect to dissociation into H + Ht. This prediction is borne out by experiments
which have been able to detect and measure the properties of this molecular
ion.
H? has two electrons, both of which go into the bonding MO and so we pre-
dict that the molecule is stable with respect to dissociation into atoms. Whereas
H> has two bonding electrons, Hy has only one, so we might expect the bond
in the latter to be weaker than in the former. Experiment bears this out: the
bond dissociation energy of Hy is 256 kJ mol~! whereas that of Hp is 432
kJ mol™!.
5.3 Symmetry labels 65
Ht H, Het He,
Fig. 5.6 MO diagrams (as in Fig. 5.5) for a series of molecules with increasing numbers of electrons;
the occupation of the orbitals by the electrons is indicated by the arrows. Occupation of the bonding
MO favours the formation of a bond as it lowers the energy relative to the AOs; occupation of the anti-
bonding MO has the opposite effect. These diagrams predict that HZ, Ha and Hes should be stable
with respect to dissociation into atoms or ions whereas Hes should not form.
He? has four electrons, so we have to put two in the bonding MO and two in
the anti-bonding MO. Occupation of the bonding MO favours formation of the
bond, but occupation of the anti-bonding MO disfavours formation of the bond.
Overall, the two effects cancel one another out and so there is no tendency for
the molecule to form as the energy is not lowered as the atoms come together.
Detailed calculations show that the anti-bonding MO is raised in energy by
slightly more than the bonding MO is lowered, so if both are occupied equally
the overall effect is to disfavour the formation of a bond. In agreement with
our prediction, no He2 molecules have ever been observed experimentally.
Finally, Hey represents an interesting case. Here there are three electrons,
two in the bonding and one in the anti-bonding MO. So, overall there is net
bonding as the number of bonding electrons is greater than the number of anti- (a)
bonding electrons. We therefore predict that the molecule He> should be stable
with respect to dissociation into He + He*. Experiment confirms this pre-
diction: He; has been observed and found to have a dissociation energy of
290 kJ mol™!. WU
This first application of the MO approach is very nice indeed — it gives us (b) a
a simple and straightforward explanation of the bonding and occurrence of the
molecules Hz, He2 and their ions. A ‘dot and cross’ picture of these molecules
simply cannot explain what is going on — we need the MO approach.
b*
Fig. 5.7 Illustration of the two symmetry
5.3 Symmetry labels Jabels which are used to describe MOs in
diatomic molecules. The first symmetry
MOs are given labels which tell us about their symmetry with respect to the label refers to traversing a circular path
around the internuclear axis, as shown in
molecule. For a diatomic, the first of these symmetry labels refers to what (a). If the wavefunction does not change
happens to the orbital when we traverse a circular path around the internuclear sign along this path the orbital is given
axis; this is illustrated in Fig. 5.7 (a). The circular path is centred on the bond the label o. If the circular path crosses a
nodal plane the orbital is given the label
and is in a plane perpendicular to the internuclear axis. x. The second label, shown in (b), refers
If, when we make a complete circuit of this path, the orbital does not change to what happens when we start at any
point a, go in a line to the centre of
sign it is given the label o (‘sigma’). Looking at Fig. 5.3 on p. 63 we see that
symmetry of the molecule (marked with a
neither the bonding nor the anti-bonding MO has a sign change along such a dot) and then carry on an equal distance
path, so they are both given the label 7. Sometimes we distinguish the anti- to point b. If the orbital has the same sign
at a and b the symmetry label is g; if the
bonding orbital by adding a superscript star: a.
sign has changed, the label is u. Note
If during a complete revolution we cross one nodal plane (i.e. a plane where that the second symmetry label only
the wavefunction is zero) the symmetry label is 2 (‘pi’). We will not encounter applies to homonuclear diatomics.
orbitals which have other than o or 7 symmetry.
A homonuclear diatomic possesses a kind of symmetry called a centre of

inversion, which is the point mid-way between the two atoms; the MOs can be
classified according to their behaviour with respect to this kind of symmetry.
The process is shown in Fig. 5.7 (b). We start at any point, a, and move in
a straight line towards the centre of inversion; having reached the centre we
carry on in the same direction for the same distance until we reach point b. If
the orbital has the same sign at points.q and b it is given the label g (for gerade,
‘even’ in German). From Fig. 5.3 on p. 63 we can see that the bonding MO has
this g symmetry; this label is added to the o as a subscript: og.
If, on going from point a to b, the orbital changes sign we attach the label
u (for ungerade, ‘odd’ in German). We see from Fig. 5.3 that the anti-bonding
MO has this symmetry, and so it is given the label o, or a.
The o and z labels can still be applied to MOs in heteronuclear diatomics,
but the g/u symmetry labels cannot as heteronuclear diatomics do not possess
a centre of inversion. We sometimes also label the MOs with the AOs from
which they are derived, so for example the bonding MO in H2 can be labelled
1S Og.
5.4 General rules for forming molecular orbitals

We now want to move on to construct the MOs for molecules more complex
than H2, so we need to consider what happens when there are more AOs present
than just the two Is AOs in Hz. It turns out that there are some simple rules
about how MOs are constructed from AOs; we will look at each of these rules
in turn.
Only AOs with similar energies interact to a significant extent

When two AOs interact to form MQs, the extent of the interaction 1s reflected in
the amount by which the bonding MO goes down in energy, or equivalently the
amount by which the anti-bonding MO goes up in energy. So, a large interac-
tion results in the MOs differing significantly in energy from the AOs, whereas
if there is little interaction the energies of the MOs are not much different from
those of the original AOs.
When two AOs interact, it is always the case that the bonding MO is lower
in energy than the lowest energy AO, and that the anti-bonding MO is higher
in energy than the highest energy AO. Figure 5.8 illustrates the effect on the
MO energies of increasing the energy separation of the AOs.
In (a) the AOs are matched exactly and the interaction is strong, as evi-
denced by the large drop in energy of the bonding MO when compared to the
AOs. In (b) the match is not quite as good and we see that the drop in energy
of the bonding MO is less; in (c) the mismatch is even greater and the fall in
energy of the bonding MO is correspondingly smaller. Finally in (d) there is
hardly any interaction and the bonding MO is more or less at the same energy
as the lowest energy AO.
5.4 General rules for forming molecular orbitals 67
—=
(a) (b) (c) (d)

a
decreasing interaction
Fig. 5.8 Illustration of how the interaction between two AOs decreases as the energy match between
them gets poorer; the downward pointing arrow gives the amount by which the bonding MO is lower in
energy than the lower energy AO.
The size of the AOs is important

The two AOs interact in the region where their wavefunctions overlap; the ex-
tent of overlap will depend on the size of the orbitals and how close together
they can come. The size of an AO is determined by the principal quantum num-
ber and the effective nuclear charge, whereas how close two AOs can approach
depends on the bond length.
Generally what is found is that when considering the interaction between
identical orbitals increasing their size decreases the degree of overlap. So,
for example, two 3s orbitals overlap to a lesser extent than two 2s orbitals.
Similarly, we find that the overlap between a 2s and a 3s orbital is less than
between two 2s orbitals.
AOs must be of the correct symmetry to interact

This condition is best illustrated by giving an example of a case where the two
AOs are not of the correct symmetry. Consider the case shown in Fig. 5.9 in
which the overlap along the z-axis (the axis of the bond) of a 2p, orbital with
a ls orbital is depicted; recall that the 2p, orbital points along the x-axis, so it
will point perpendicular to the z-axis. In (a) the orbitals are far apart so that
we can see Clearly the positive and negative lobe of the p orbital and the single
positive lobe of the 1s orbital.
2p, 1s (b) region of

constructive
A overlap
* \
\. region of
‘ad estructive
overlap
Fig. 5.9 Illustration of the symmetry requirement for AOs to overlap. (a) Shows a 2px and a 15 AO
about to overlap along the z-axis; the positive parts of the orbitals are shaded grey. In (b) the overlap
has begun, but we see that there is a region of constructive overlap (where the two orbitals have the
same sign) and a region of destructive overlap (where the orbitals have opposite sign). These two
regions cancel one another out leading to no net overlap. The black dots indicate the location of the
nuclei.
As the orbitals come together, shown in (b), the positive lobe of the 2p
orbital overlaps with the positive 1s orbital; this is the kind of constructive
overlap which leads to the formation of a bonding MO. However, the negative
lobe of the 2p orbital overlaps with the positive Is orbital; this is the kind of
destructive overlap which leads to the formation of an anti-bonding MO. In the
case shown here the constructive and destructive overlaps exactly cancel one
another out, leading to no net overlap.
So, we conclude that the 2p, and Is orbitals do not have the correct sym-
metry to overlap along the z-axis. The same is true for the 2p, and Is orbitals.
However, as is illustrated in Fig. 5.10,the 2p, and 1s orbitals do have the cor-
rect symmetry to overlap.
When n AOs interact the same number of MOs is formed

Fig. 5.10 In contrast to the case shown in We have already seen an example of this when two AOs in H2 gave two MOs.
Fig. 5.9, a 2pz orbital has the correct In Chapters 6 and 10 we will see some examples of the interaction of more
symmetry to overlap with a 1s AO along
the Z-axis. than two AOs.
5.5 Molecular orbitals for homonuclear diatomics

We will now use these rules to construct the MO diagram for the series of
diatomic molecules Liz, Bez, ..., Nez formed by the elements of the first row
of the Periodic Table. The first thing we need to think about is which AOs might
overlap with one another; from our knowledge of the electronic configurations
of these atoms we know that the relevant orbitals are the Is, 2s and 2p. In
(b)
fact we need not consider the Is as they are so contracted towards the nucleus
(due to the high effective nuclear charge they experience) that, as illustrated in
Fig. 5.11, they do not overlap to a significant extent. So, the only orbitals we
need to consider are 2s and 2p.
The 2s orbital is spherical, just like the Is orbital, but as we have seen
Fig. 5.11 Illustration of why the 1s (Fig. 4.18 on p. 46) the 2s has a radial node close in to the nucleus. However,
orbitals are not involved in bonding in this does not really affect the way in which the 2s AO overlaps with other
diatomic molecules formed from
elements of the first row of the Periodic orbitals as the region of overlap is much further out than is this node. So, to all
Table. On account of the high effective practical intents and purposes, we can simply treat the 2s as a spherical orbital.
nuclear charge experienced by the 1s There will be a perfect energy match between the 2s AO on one atom and
electrons the orbitals are highly
contracted and so, for typical bond the 2s on the other, so we expect a strong interaction between these orbitals.
lengths, they do not overlap to a Similarly, the 2p AOs will have a perfect energy match. However, as we saw
significant extent, as shown in (a). In on p. 56 the 2s is lower in energy than the 2p, so the energy match between
contrast, the 2s orbitals are much larger
and so overlap well as shown in (b).
these two orbitals is not perfect and therefore the interaction is not as strong as
between 2s and 2s or between 2p and 2p. Taking advantage of this effect we
will for now ignore any overlap between 2s and 2p when we construct our MO
diagram.
We have already seen how two 1s AOs combine to give a o bonding orbital
(og) and a o anti-bonding orbital (o,,); the same thing happens for the 2s or-
bitals and we will denote the resulting MOs 2s o, and 2s o,. The 2p orbitals
can overlap to give MOs of both o and z symmetry, as described in the next
section.
MOs from the overlap of p orbitals

If we take the internuclear axis to be the z-direction, then two 2p, orbitals can
overlap ‘head on’ to give 0 MOs as shown in Fig. 5.12 (a). The bonding MO
5.5 Molecular orbitals for homonuclear diatomics 69
(a) (b)
|
|
OO 2po, &> apr,
Fig. 5.12 MO diagrams illustrating the formation of « and m type MOs from the overlap of 2p AOs. The
‘head-on’ overlap, shown in (a), leads to o MOs, whereas the ‘side-on’ overlap (b) leads to 7 MOs. The
2p AOs are drawn so that adding them together leads to constructive interference. The dots indicate
the location of the nuclei.
is formed when there is constructive interference between the two lobes which
face toward one another along the z-axis; as with the 2s 0 MO the concentra-
tion of electron density between the two nuclei is largely responsible for the
lowering in energy of the MO.
In Fig. 5.12 (a) the 2p orbitals are drawn so that simply adding them to-
gether gives constructive interference. The other combination we need to con-
sider comes from subtracting the orbitals; this leads to destructive interference
and the formation of the anti-bonding MO in which electron density is pushed
away from the internuclear region.
It is important to realize that the way we choose to represent the 2p AOs is
arbitrary; we could just as well have made the left-hand lobe on both AOs pos-
itive. With this choice, adding together the two AOs would lead to destructive
interference and the formation of the anti-bonding MO, whereas subtracting
the two AOs would lead to the formation of the bonding MO. Either way, we
still end up with a bonding and an anti-bonding MO.
Figure 5.13 shows surface plots of the two MOs and Fig. 5.14 shows con-
tours plots of the same orbitals. The two MOs are both o; with respect to the
centre of inversion the bonding MO has symmetry g and the anti-bonding MO
has symmetry u. We therefore denote the orbitals 2p og and 2p o, to indicate
their symmetries and that they are derived from 2p orbitals.
The 2p, and 2py AOs point in directions perpendicular to the internuclear
axis and, as was discussed on p. 67, these orbitals do not have the correct
symmetry to overlap with the 2s orbitals. A similar line of reasoning, illustrated
in Fig. 5.15, shows us that a 2p, on one atom cannot overlap with a 2p; on the
other.
However, the two 2p, AOs can overlap ‘side on’ in the way illustrated in
Fig. 5.12 (b); the shapes of the resulting orbitals are best appreciated from the
surface and contour plots shown in Figs. 5.16 and 5.17.
Constructive interference leads to a bonding MO in which there is electron
density above and below the internuclear axis; note the contrast to ao bonding
MO which has electron density directly between the two nuclei.
‘7
ZB
Ss
2p o,, bonding MO 2p o,, anti-bonding MO
Fig. 5.13 Surface plots of the 2p og (bonding) MO and the 2p ay (anti-bonding) MO formed from the
head-on overlap of two 2p orbitals. Note how the bonding MO concentrates electron density in the
internuclear region whereas the opposite is true for the anti-bonding MO. The two nuclei are placed
along the z-axis, symmetrically about z = 0.
‘ : : errr
. : H
‘ i :‘
5 :
‘ : !
* . : é ¢
: H VA
-7 \5 :i mS :H age o= SAN yt
‘ a!vos H if{ ‘‘ we
: a NNN
iat \ \ ‘
: H : ' !
‘. so?
ii an)
, ON 7
' Pals
Ha Tre
titan
Oe
7 ! * sae PTS a aft
: 5 : z
i ‘ i Soar
‘
“eee”
2p a, bonding MO 2p o,, anti-bonding MO

Flg. 5.14 Contour plots of the 2p ag (bonding) MO and the 2p ay (anti-bonding) MO. The cross-section
shown is the xz-plane (with z running horizontally); the positions of the two nuclei are indicated by the
small black dots.
This difference in the electron distribution explains why the MO formed

by the side-on overlap is less strongly bonding than the o bonding MO. Side-
on destructive interference leads to an anti-bonding MO in which the electron
density is pushed away from the region above and below the internuclear axis.
Both the bonding and anti-bonding MOs have a nodal plane which contains
the two nuclei. In addition, the anti-bonding MO has a second nodal plane
2p, 2p, which lies between the two nuclei, perpendicular to the internuclear axis.
Inspection of Fig. 5.16 shows that the symmetry is 7 as when traversing a
Fig. 5.15 Illustration of why a 2pz AO and
a 2px AO have no net overlap; the region
circular path perpendicular to the internuclear axis (see Fig. 5.7 (a) on p. 65)
of constructive interference is balanced we cross a nodal plane. The bonding orbital has u symmetry and so is given
out by the region of destructive the full label 2p 7, and the anti-bonding orbital is g so is given the label 2p z,.
interference, just as in Fig. 5.9 on p. 67.
The two 2py AOs can also overlap with one another to give two more 7
MOs, one bonding and one anti-bonding. So, there is a total of two bonding
x orbitals and two anti-bonding z orbitals: one bonding and anti-bonding pair
2p t,, bonding MO 2p Tg anti-bonding MO
Fig. 5.16 Surface plots of the 2p xy (bonding) MO and the 2p zg (anti-bonding) MO formed from the
side-on overlap of two 2px orbitals. Both MOs have a nodal plane containing the two nuclei, and the
anti-bonding MO also has a nodal plane between the two nuclei, perpendicular to the internuclear
axis. Note that for the bonding MO there is a concentration of electron density in the regions above
and below the internuclear axis. The two nuclei are placed along the z-axis, symmetrically about z = 0.
2p r,, bonding MO 2p Ng anti-bonding MO
Fig. 5.17 Contour plots of the 2p zy (bonding) MO and the 2p 7g (anti-bonding) MO; the nodal planes
are clearly visible in both piots. The cross-section shown is the xz-plane (with z running horizontally);
the positions of the two nuclei are indicated by the small black dots.
lie in the xz-plane and the other pair lie in the yz-plane. The two bonding MOs
are degenerate, as are the two anti-bonding MOs.
Two MO diagrams for the Az molecules

If we consider only the 2s-2s and 2p-2p overlap it is quite easy to assemble
the complete MO diagram, which is shown in Fig. 5.18 (a). On the left and
right are shown the AOs, with the 2s lower in energy than the 2p. The three
degenerate 2p orbitals are represented by the three closely spaced lines labelled
2p on the diagram.
The 2s orbitals overlap to form a bonding MO, 2s og, and an anti-bonding
MO, 2s 0,. The 2p orbitals which point along the bond (the z-direction) also
overlap to form o orbitals: 2p a, (bonding) and 2p o, (anti-bonding).
The two 2p, orbitals overlap to give a 77 bonding MO (2p 7) and a 7 anti-
2po, ———. 20 9 ii
2pn,, In, 1o,; _—__.
2p, _ 20, 25 —_—
256, ——. 10, lo,

‘ AOs on MOs AOs on
atom 1 atom 2
2s
2s Oy ‘ ; 16,
AQOs on MOs AQs on

atom 1 atom 2
Fig. 5.18 Shown in (a) is a simple MO diagram for the homonuclear diatomics of the first row, con-
structed by considering only 2s-2s and 2p-2p overlap; this diagram turns out to be appropriate for O2
and Fo. However, for the molecules Lia...., No the MO diagram is of the form shown in (b). The
difference between (b) and (a) is that the s and p orbitals are much closer in energy in (b) so that we
need to include the effect of s-p mixing; this results in a re-ordering of the x and o bonding orbitals
derived mainly from the 2p AOs. The MOs in diagram (a) are labelled according to which AOs each
MO is derived from; diagram (b) is labelled using a different scheme in which orbitals of the same
symmetry are simply distinguished from one another by numbering them 1, 2, ...; the MOs in diagram
(a) are also labelled using this second scheme.
bonding MO (2p zz). The same is true of the two 2py orbitals, so there are two
degenerate 2p m, orbitals, indicated by the two closely spaced lines, and two
degenerate 2p mg orbitals indicated in the same way. Generally o interactions
are stronger than 7, so we have shown the 2p oy as being lower in energy than
the 2p my.
Right at the start of this section (p. 68) we made the assumption that the 2s
and 2p orbitals are far enough apart in energy that we do not need to consider’
the possibility of o overlap between them. Therefore, the MO diagram shown
in Fig. 5.18 (a) is only suitable for cases where this assumption is appropriate.
We saw in Fig 4.34 on p. 57 that the energy separation between the 2s and
2p increases as we go across the first row of the Periodic Table. It turns out
that this separation is sufficient in oxygen and fluorine for the MO diagram of
Fig. 5.18 (a) to be correct. However, for the other elements (Li, ..., N), the
separation of the 2s and 2p is such that we cannot ignore the effect of o overlap
between these AOs.
The most significant effect of these s-p interactions (usually called s-p mix-
ing) is to make the 2p og MO less bonding (i.e. to raise its energy) and the 2s
oy, MO less anti-bonding (i.e. to lower its energy). For the elements Li, ..., N
the shift of the 2p og is large enough to move it above the 2p m, MO, as is
shown in Fig. 5.18 (b).
When there is significant s-p mixing we cannot really label an orbital 2p og
as it is not correct to identify it as being derived solely from 2p; the symmetry
labels o and g are still applicable, though. Under these circumstances we drop
the 2p prefix and simply number the orbitals to distinguish ones with the same
symmetry, so the first og MO is labelled lo, and the second 20. The other
orbitals are labelled using a similar scheme, as is shown in Fig. 5.18 (b); the
same scheme can be used to label the orbitals in MO diagram (a).
Using the MO diagram

We can now use the MO diagrams of Fig. 5.18 to make some predictions about
the homonuclear diatomics of the first row. All we have to do is to slot the
appropriate number of electrons into the MOs and then work out what are the
consequences of the resulting electronic configuration.
For example, lithium has the configuration 1s?2s!, but as we commented
on before, the 1s electrons are not really involved in bonding, so we need only
be concerned with the single 2s electron. The electrons which are involved in
bonding are called the valence electrons, and there are two of these in Lio.
For this molecule the appropriate MO diagram is Fig. 5.18 (b) and so the
two valence electrons go spin paired into the lo, orbital giving the electronic
configuration lo. We can carry on in the same way for the diatomics Liz,
..., N2, giving the results shown in the table below (a star has been added
to indicate the anti-bonding MOs). The table also shows the bond order, the
meaning of which we will come on to shortly.
log | lo, lr, 202 Im, bond order
Liz | tl 1
Be2 | th | tl 0
Bo; th | th | tft |
QIN | tinh 2
No} tH | th | th th |] ty 3
In Liz both electrons are in a bonding orbital and there are no electrons in
anti-bonding orbitals, so we predict that the molecule is stable with respect to
dissociation into atoms. Experiment bears this out: molecular Liz has been
observed in the gas phase and it is found to have a bond dissociation energy of
101 kJ mol~!. The two electrons in a o bonding MO create a o bond between
the two atoms, so the molecule is described as having a single o bond.
In Be» there are two electrons in a bonding MO and two in an anti-bonding
MO; the simple expectation is that the molecule has no net bonding and so will
the
not form (just as was the case for He2). In fact, because of the s-p mixing,
lo, orbital is somewhat less anti-bonding than the log is bonding, so overall
Be? is expected to be a weakly bound species. The molecule has been observed
in the gas phase and is found to have a rather low dissociation energy of 59
kJ mol7!.
B> has six valence electrons; four of these are spin paired in o orbitals and
the last two go into different 17, orbitals with their spins parallel. Remember
that when there are degenerate orbitals the lowest energy arrangement for two
electrons is to put them into separate orbitals with their spins parallel.
There is a special feature of Bz which we can explain using the MO dia-
gram, which is that the molecule is found to be paramagnetic. Paramagnetic
substances are drawn into a magnetic field — an effect which can be measured
experimentally. Paramagnetism is associated with the presence of unpaired
electrons, which, according to the MO diagram, is precisely what we have for
B>.
The bonding effect of having two electrons in the log is roughly cancelled
out by having two electrons in the anti-bonding 1o,, so the bonding in Bz is
mainly due to the two Iz, electrons. A useful way of expressing the degree of
bonding is to calculate the bond order, which is given by
bond order =
Y [number of bonding electrons — number of anti-bonding electrons] .
The factor of one-half is included so that a pair of bonding electrons, which we

think of as comprising a bond, gives a bond order of one.
In the case of Liz the bond order is (2 — 0) = 1, whereas for Be? it is
(2 — 2) = 0. These values fit in with the description of Liz having a single
bond and Be? having (to a first approximation) no bond. Similarly the bond
order in Bz is (2 — 2+ 2) = 1, i.e. a single bond. The table on p. 73 gives
the values of the bond order for each molecule.
The bond order is a rather rough measure of the degree of bonding as it
takes no account of the relative contribution to the bonding of different bond-
ing orbitals (and similarly for anti-bonding MOs). However, there is quite a
good correlation between the bond dissociation energy (a measure of the bond
strength) of the molecules Az and the bond order, as shown in Fig. 5.19.
The next diatomic is C2 which has eight valence electrons. These fill the
log, the lo, and both of the 17, orbitals with all spins paired; the result is
two mz bonds. N2 has 10 valence electrons and these occupy the MOs in such
a way that there are two wz bonds and one o bond; this is consistent with the
simple view that N2 has a triple bond and indeed the bond order is 3. This
molecule has the largest bond order of all of the diatomics of the first row and
So is expected to have the strongest bond, which is indeed the case.
Once we get to oxygen and fluorine the appropriate MO diagram is
Fig. 5.18 (a) which has a different ordering of the orbitals. The electronic
configurations for these last two diatomics are therefore:
7 1000
2 *N,
~ 800F
3
D
5 | 600- °C,
iS °0,
@ | 400F
3
8 °B,
55 200F Fg7 eLi
Qa 2
(a) lL l l
1 2 3
bond order
_
Fig. 5.19 Plot showing the correlation between bond dissociation energy (bond strength) and bond
order for the homonuclear diatomics of the first row. There is a useful correlation between these two
quantities.
log | loy | 20% lx, 1%, bond order
Q|/ NW) N] Nit wit t 2

Fo; th | TL | TL | TY Oth | tL OTN 1
In O all of the bonding MOs have been occupied and so the final two
electrons have to go into the z anti-bonding MOs. As is the case for B2 the last
two electrons are split between two degenerate orbitals with their spins parallel,
making the molecule paramagnetic — which is what is found experimentally.
In F the 7 anti-bonding orbitals are now full, so the net bonding is simply
from the 20, orbital. Finally, the hypothetical molecule Nez has all of the
bonding and anti-bonding orbitals filled and so is not expected to form.
The MO approach is remarkably successful at explaining the properties of
the Az molecules. Using it, we have been able to rationalize the variation in
bond strengths, explain why Bez and Nez are not expected to form, and predict
the observed paramagnetism of Bz and O2. We can be well satisfied with this
outcome!
Figure 5.20 shows a plot of the energies of the occupied MOs for the
homonuclear diatomics of the first period. We see that in general the ener-
gies of the MOs fall steadily across the period; this has the same origin as the
fall in the orbital energies of the atoms, i.e. it is due to the increasing nuclear
charge — see Section 4.8 on p. 57.
For B> and C2 the highest occupied MO is the 17, showing that this lies
lower in energy than the 20¢. In contrast, for Oz and F2 the 20 lies lower than
the I7,,. For Nz these two MOs are rather close in energy and there is some
uncertainty over their ordering.
Li, Be, Bs C, Np 0» Fe
16
1o we. In
:a In,
g
" 26
Fig. 5.20 Plot showing the calculated energies (to scale) of the occupied MOs for the homonuclear
diatomics of the first row; the dashed lines connect similar MOs. Note that there is a general fall of the
energies of the MOs as we go across the period and that for O2 and F2 the 2og lies below the 1zu.
5.6 Heteronuclear diatomics

The bonding in heteronuclear diatomics (ones in which the two atoms are dif-
ferent) can be described in a similar way to that used for homonuclear di-
atomics. There is one significant difference which is that there will no longer
be an exact match between the energies of the AOs of the two atoms; this makes
it more difficult to construct the MO diagram as we have to decide which AOs
will overlap. We will also see that this energy mismatch between AOs leads to
polarized bonds.
On p. 66 and in Fig. 5.8 it was explained that when the energy separation
between two AOs becomes larger the interaction between them decreases. A
further consequence of this energy mismatch is that the contribution of the two
AOs to the MQs is not the same: this is illustrated in Fig. 5.21.
Figure 5.21 (a) shows the situation in which the two AOs have the same
energy. The bonding MO has equal contributions from the two AOs, and the
same is true of the anti-bonding MO except that the orbitals are combined with
opposite coefficients. When the two AOs have different energies what we find
is that the AO which jis closer in energy to the MO is the major contributor to
that MO. This is illustrated in (b) in which the bonding MO is closest in energy
to the AO of atom 2 and so this AO is the major contributor to the MO. In
contrast, the anti-bonding MO is closest in energy to the AO from atom | and
so this is the major contributor to the MO.
As the energy mismatch increases further the relative contributions of the
two AOs to a given MO become even more unequal, as is shown in Fig. 5.21
(c). Now the bonding MO has a much larger contribution from the AO on atom
2 than from that on atom 1, and vice versa for the anti-bonding MO.
If the bonding MO of Fig. 5.21 (c) is occupied the electron distribution will
be uneven across the molecule, simply because the orbital (wavefunction) is
greater in the region of atom 2 (recall that the electron density is proportional to
5.6 Heteronuclear diatomics 77
increasing orbital energy mismatch
Fig. 5.21 {llustration of the effect on the MOs of increasing the energy mismatch between the two
interacting AOs. In (a) the two AOs have identical energies, and so contribute equally to the bonding
and anti-bonding MOs; this is shown in the cartoons of these orbitals where the MO is drawn such that
the size of the constituent AO indicates its relative contribution. !f the energy of the AO on atom 2 is
tower than that of atom 1, as shown in (b), the AO on atom 2 is a more significant contributor to the
bonding MO whereas the AO on atom 1 is the more significant contributor to the anti-bonding MO. As
the energy mismatch between the AOs becomes greater, the effect is more pronounced, as is shown
in (c).
the square of the wavefunction). As a consequence, the bond will be polarized

towards atom 2, with a partial negative charge on atom 2 and a partial positive
charge on atom 1. The bond will have a dipole moment, as shown in Fig. 5.22.
It is clear from this discussion that the relative energies of the AOs are im-
portant as this determines which will overlap most effectively and the polarity
of any resulting bonds. In Section 4.8 on p. 57 it was described how the en-
ergies of the AOs decrease as we go across the first row of the Periodic Table. e-
We can therefore use this observation to work out which of any two atoms has —e’
the lowest energy orbitals and hence draw up an appropriate MO diagram.

In Fig. 5.22 the AO from atom 2 is lower in energy and so the bond is
polarized towards this atom. If we were describing this polarization in terms
<@
of electronegativities we would say that atom 2 is more electronegative than
atom 1 because the bond is polarized towards atom 2. We can now see that
the reason for the greater electronegativity of atom 2 is that its AOs are lower o+ Ce _) 6-
in energy than those of atom 1. This connection between orbital energies and u-+-—
electronegativity has been discussed earlier on p. 58.
Fig. 5.22 The occupation of an MO which
has unequal contributions from the two
Two simple examples: LiH and HF AOs leads to a polarized bond. Here, the
bonding MO is closest in energy to the
The molecules LiH and HF provide a good illustration of the concepts intro- AO from atom 2 and so this AO is the
duced in the previous section. They also illustrate two extremes, as in LiH the major contributor to the MO. As a result,
H has the lower energy AOs (see Fig. 4.34 on p. 57) and so the bond is polar- the electron density is skewed towards
atom 2, which acquires a partial negative
ized towards hydrogen, whereas in HF it is the fluorine which has the lower charge. The bond is therefore polarized
energy AOs resulting in a bond which is polarized the other way round. and has a dipole moment.
In LiH the valence orbitals are the 2s on lithium and the 1s on hydrogen.
In Section 4.8 on p. 57 we saw that the orbital energy of the 1s in hydrogen
is considerably lower than that of the 2s in lithium, so in constructing our MO
diagram we must take this into account.
Figure 5.23 shows an approximate MO diagram for LiH; there are just two
Li valence electrons and these occupy the o bonding MO, labelled lo. Note that
2s
although we can classify this orbital as o (according to the rules on p. 65) we
cannot use the g and u labels as a heteronuclear diatomic does not possess a
1s centre of inversion.
On the basis of this diagram we predict that the molecule will be stable
with respect to dissociation into Li and H atoms, that it has a single o bond and
that the bond is polarized towards the hydrogen. Figure 5.24 shows a contour
Fig. 5.23 MO diagram for LiH; note that plot of the bonding MO in this molecule. From this plot we can clearly see
the 1s orbital on H is lower in energy that most of the electron density is on the hydrogen, which is expected as the
than the 2s orbital on Li (see Fig. 4.34 on
p. 57). There are two valence electrons hydrogen Is AO is lower in energy than the Li 2s.
and these occupy the 10 bonding orbital; In HF the orbitals are the other way round: the fluorine AOs are much lower
the result is a single bond, polarized in energy that those of hydrogen (see Fig. 4.34 on p. 57). We also know that the
towards the hydrogen.
2s lies at lower energy than the 2p, so to a first approximation we will ignore
the 2s and just assume that the only overlap is between the 2p on fluorine and
the ls on hydrogen.
There are, of course, three 2p orbitals, but as we saw in Fig. 5.10 on p. 68
only the 2p. orbital (the one which points along the bond) has the correct sym-
metry to overlap with the s orbital, to give ao MO. The major contributor to
this o bonding orbital (labelled 2c) is the fluorine 2p, as this is closest to it
in energy. The other 2p orbitals have no interaction with the 1s and so remain
non-bonding. The resulting approximate MO diagram is shown in Fig. 5.25.
HF has a total of eight valence electrons: two are in the fluorine 2s (labelled
lo in Fig. 5.25) and the remainder occupy the 20 bonding MO and the two
non-bonding fluorine AOs, 2p, and 2py. We predict, therefore, that there will
Fig. 5.24 Computer generated contour be a single o bond between the H and F, and that there is a concentration of
plot of the « bonding MO of LiH which electron density towards the fluorine in this bond. In addition, there are further
corresponds to the 2a MO shown
electrons localized on fluorine in the 2s, 2p, and 2py AOs. The electron density
schematically in Fig. 5.23. Positive and
negative contours are shown in black and in this molecule is therefore very biased toward the fluorine. We can attribute
grey. respectively; the atoms and Li-H
bond are also shown. The Li AO involved
in this orbital is the 2s which has a radial
node close-in to the nucleus; this can
clearly be seen by the change in colour
of the contours around the Li. As
expected, the electron density for this MO
is mainly on the hydrogen on account of
the lower energy of the hydrogen AO.
-@:
Fig. 5.25 Approximate MO diagram for HF assuming that only the fluorine 2p and hydrogen is AOs
are involved in bonding. The fluorine AOs are much lower in energy than the hydrogen 1s AO
(see
Fig. 4.34 on p. 57). Only the 2pz orbital has the correct symmetry to overlap with the hydrogen
1s,
giving rise to the 2c orbital which is polarized towards fluorine. The fluorine 2s is assumed
to be
too low in energy to overlap with the 1s and so remains non-bonding. The MOs labelled
10 and 17
are really the 2s, 2px and 2py fluorine AOs, but the labels are included to facilitate comparison
with
Fig. 5.26.
5.6 Heteronuclear diatomics 79
-19 eV
10 —42 eV
Fig. 5.26 Surface plots of the four highest energy occupied MOs of HF; the MOs have been calculated
using a computer program. The surface is shown as a ‘net’ so that we can see through it to the balls
which represent the atoms; the line joining them is drawn in to help identify the bond. The fluorine is
shown in black and the hydrogen, which is coming toward us, in white. The energy of each MO is also
given. These MOs should be compared with the approximate picture shown in Fig. 5.25.
this imbalance to the fact that the fluorine orbitals are much lower in energy
than the hydrogen AOs and that there are three non-bonding electron pairs on
the fluorine.
It is interesting to compare our cartoons of the MOs with the computer cal-
culated MOs shown in Fig. 5.26. In the figure, we recognize the MO labelled
lo as arising from the overlap of the fluorine 2s with the hydrogen Is. In draw-
ing up the simple MO diagram we discounted this interaction on the grounds
of the large energy separation between these two AOs, but the more complete
calculation shows that they do have some interaction. Note how much lower in
energy the lo MO is than all the other MOs; this reflects the low energy of the
fluorine 2s.
The next highest MO, 20, is exactly what we expected; it is formed from
the overlap of a 2p orbital on fluorine with the hydrogen Is. At only slightly
higher energies are the two degenerate 12 MOs, which can clearly be identified
as the two 2p orbitals from fluorine which remain non-bonding.
Moving on
So far we have used the molecular orbital approach to describe the bonding
in diatomic molecules — in the next chapter we will extend this to rather more
complex molecules. We will also see how the use of hybrid atomic orbitals
greatly simplifies the description of bonding in these larger molecules and en-
ables us to focus on the parts which are really important when it comes to
thinking about reactions.
6 Electrons in larger molecules
hy an
The MO approach is not limited to describing the bonding in diatomics but can
be applied to any molecule we care to think of. In more complex molecules,
AOs on several atoms can contribute to each MO, and it rapidly becomes rather
difficult to draw up the MO diagram ‘by hand’.
Luckily for us, many computer programs have been developed for calcu-
lating both the shapes and energies of MOs. All we have to do is specify the
types and positions of the atoms, and then the program does the rest. For small
to medium sized molecules, such calculations are well within the capabilities
of desk-top computers.
Figure 6.1 shows the four occupied MOs of methane (CH4) computed (and
displayed) using one of these programs. From their form, we can guess which
AOs are contributing to some of these MOs. For example, MO (a) is clearly
a combination of the 2s AO on carbon and all four 1s AOs on the hydrogens;
MO (b) looks like a combination of one of the 2p carbon AOs with just two
of the hydrogen 1s AOs. The thing to notice is that several AOs contribute to
each MO.
(a)
Fig. 6.1 Surface plots of the four occupied MOs of methane. The surface is shown as a transparent
‘net’ so that we can see through it to the model of the molecule which shows the positions of the atoms
and bonds; the carbon atom is shown in black and the hydrogen atoms in white. Positive parts of the
orbital are shown by a black net and negative by a grey net. The lowest energy MO is (a); (b), (c)
and
(d) are ail degenerate.
6.1 Two-centre, two-electron bonds 81
Energy minimization
Experimentally, it is found that methane has a tetrahedral geometry with a bond
angle of 109.5° and a C-H bond length of 1.089 A; this geometry must there- 1 A (Angstrém) is 10719 m or 100 pm.
fore be the lowest energy arrangement of one carbon and four hydrogen atoms.
The computer programs used to calculate MOs usually have an option for
finding this lowest energy arrangement (or equilibrium geometry, as it is often
called) for a given molecule. All that the program has to do is to compute the
MOs for a given arrangement of atoms, assign the electrons to the MOs and
hence find the total energy of the molecule. Then the atoms are moved around
and the calculation repeated until the minimum energy arrangement is found —
a process called energy minimization.
Computers are very good at this kind of repetitive calculation and clever
procedures have been developed to enable the program to locate the energy
minimum in an efficient way. So, even if we do not know the geometry of our
molecule, all we need to do is to make a reasonable initial guess and then ask
the program to determine the equilibrium geometry.
Working from what we already know

Sophisticated and convenient though these computer programs are, we really
need to describe the bonding in a simpler way which we can use on a day to
day basis without resorting to computer programs. The key point which helps
us here is that for most molecules we already have a pretty good idea of their
shape, so we do not need to work out the equilibrium geometry — all we need
to do is find a description of the bonding for a molecule of known shape.
For example, we know that the four bonds around a saturated carbon are
arranged in a tetrahedral geometry, that ethene (C2Hs) is flat with bond angles
of 120° and that the bond angle in water is 104.5°. We also know that Nz
has a triple bond, and that there is a double bond joining together the two
carbons in ethene. What we need is a simple ‘back of the envelope’ method
for determining the MOs in molecules like these whose geometry we already
know. We can then use these MOs to think about the reactions of the molecules,
which is ultimately what we are aiming at.
6.1 Two-centre, two-electron bonds

When we first learn about the formation of covalent bonds it is common to
use ‘dot and cross’ diagrams such as those for methane and water shown in
Fig. 6.2. The diagram for methane shows how the four electrons from carbon
are each paired up with one electron from hydrogen, resulting in four bonding
pairs of electrons. As their name implies, each pair is associated with a bond
between carbon and hydrogen which we represent by a line joining the two Fig. 6.2
bonding
‘Dot and cross’ pictures of the
in methane and water. The dots
atoms. Such a bond is called a two-centre, two-electron (2c-2e) bond, as it indicate the electrons from carbon (or
involves two atoms and two electrons. For water, there are just two bonded oxygen) and the crosses indicate
pairs, and hence two 2c-2e bonds. The remaining four electrons form two lone electrons from hydrogen. The lines
indicate the presence of two-centre
or non-bonding pairs on the oxygen. two-electron bonds; water also has two
Our problem is how to reconcile this simple (but very useful) picture of lone pairs of electrons. Methane is, of
the bonding in methane with the MO description. The key difference between course, tetrahedral rather than flat.
the two models is that the MO picture is a delocalized one in the sense that
the MOs (and hence the electrons) are spread over several atoms, whereas the
picture in Fig. 6.2 is a localized one in which electrons are either localized
between two atoms forming 2c-2e bonds or are localized on the atoms as lone
pairs.
In the MO picture, two orbitals oveglap to form a bonding MO and an anti-
bonding MO, and if there are just two electrons in the bonding MO then we say
that there is a bond between the two atoms. This overlap of just two orbitals is
essentially the MO picture of a 2c-2e bond as shown in Fig. 6.3.
Suppose that we try to apply this MO picture of 2c-2e bonds to methane.
We have four valence orbitals on carbon (the 2s and the three 2p) so each could
overlap with one of the hydrogen 1s orbitals. This would give us four 2c-2e
bonds each of which can be described by an MO diagram just as in Fig. 6.3.
The problem with this approach is that the carbon AOs do not point in the
Fig. 6.3 The overlap of one orbital from right directions to make good overlap with the hydrogen AOs. Remember that
atom 1 with one from atom 2 gives a
the geometry of the molecule is tetrahedral, but that the three 2p orbitals point
bonding MO and an anti-bonding MO.
Occupation of the bonding MO by two along the x-, y- and z-axes; the 2s orbital is spherical so does not ‘point’ in any
electrons gives the MO description of a particular direction at all. With such a set of orbitals it is hard to see how we
2c-2e bond between atoms 1 and 2.
can sensibly form four 2c-2e bonds.
The way out of this problem is to combine the carbon AOs into new orbitals
called hybrid atomic orbitals (HAOs). These hybrids are designed to point
in the appropriate directions to overlap with the hydrogen AOs. Each HAO
overlaps with one hydrogen 1s AO to give a bonding and an anti-bonding MO;
two electrons occupy the bonding MO thus giving a 2c-2e bond.
How we go about constructing these HAOs is the subject of the next sec-
tion, but before we move on to that we need to give a few words of caution
about how we represent molecules.
Not all lines are bonds

In Fig. 6.2 the lines joining the atoms certainly imply the presence of 2c-2e
bonds, but we have to be careful, as when we draw lines on chemical structures
we do not always mean to imply that a 2c-2e bond is present. Consider, for
example, the two structures shown in Fig. 6.4.
@ Nat
NaCl
Fig. 6.4 On the left is shown part of the lattice of the ionic solid NaCl and
on the right is shown the
structure of gaseous diborane (BoHg). In diborane the two central hydrogens
and the two borons lie
in the same plane, with the two hydrogens joined by the dashed lines lying
below this plane and those
joined by the wedges lying above it. For both structures, the lines joining
the atoms are not 2c-2e
bonds.
6.2 Hybrid atomic orbitals 83
Solid sodium chloride is a giant lattice structure consisting of Nat and Cl7
ions. We often draw it, as in Fig. 6.4, with lines joining the ions together;
however these lines do not represent 2c-2e bonds. Rather, the lines are just
there to guide our eye and indicate the spatial relationship between the ions. In
fact, as we discussed in Chapter 3, it is electrostatic forces between all of the
ions (not just adjacent ones) which are responsible for the binding energy of
this lattice.
Diborane, B2H¢, provides another example; this molecule has 12 valence
electrons, yet the structure shown in Fig. 6.4 has eight lines and if each repre-
sented a 2c-2e bond we would be four electrons short. In fact, as with NaCl, the
lines are there just to help us understand the spatial arrangement of the atoms
— they do not indicate 2c-2e bonds.
So, we have to be careful: not every line drawn on a chemical structure
represents a 2c-2e bond. However, for many simple compounds, especially
those involving carbon and other first row elements, it is often the case that the
lines do represent 2c-2e bonds. It is to the description of the bonding in such
compounds that we now turn our attention.
6.2 Hybrid atomic orbitals

Recall that what we are trying to do is to find an orbital description of the
bonding in methane which is consistent with the view that there are four 2c-
2e bonds. The problem we have is that the carbon AOs do not ‘point’ in the
right directions to overlap with the hydrogen AOs; we will overcome this by
combining the carbon AQOs to form hybrid atomic orbitals, HAOs, which do
point towards the hydrogens.
HAOs are formed by making a linear combination of AOs just as we do
when we form MOs. However, the important distinction is that to form HAOs
we combine AOs on the same atom. This is in contrast to MOs which are
formed by combining AOs on different atoms.
Each HAO is constructed by adding together AOs:
HAO = c; x AO; + c2 x AO2 +3 x AO3+ ...
where Cc}, C2, etc. are simply numerical coefficients which determine the direc-
tional properties of the HAO; AO; is the first AO and so on. As with MOs,
the number of HAOs is equal to the number of AOs being combined. So, for
example, if we combine the 2s and the three 2p AOs we will obtain four HAOs.
The shapes and energies of the HAOs are determined by the values of the
coefficients cj, C2, .... Generally, we will not be able to guess at what these
coefficients are so instead we will look at a few particular cases where the re-
sulting HAOs are useful for describing the bonding in commonly encountered
molecules.
sp° hybrids
If we take the 2s and the three 2p AOs we can form a set of four equivalent
HAOs which point towards the corners of a tetrahedron; these HAOs, depicted
in Fig. 6.5, are called sp? hybrids; the name indicates which AOs were used to
four sp? HAOs
Fig. 6.5 Surface plots of the four sp? HAOs; each hybrid points toward one of the corners of a tetrahe-
dron. The conventional representation of a tetrahedral carbon is also shown to indicate the directions
in which the HAOs point.
construct the hybrids. Often it is said that HAOs are ‘more directional’ than
AOs. We can understand what this means by thinking about the directional
properties of the AOs which are used to construct these HAOs: the 2s does not
point in any particular direction and each 2p orbital points in two directions
(e.g. along x and —x for the 2p,). In contrast, as we can see from Fig. 6.5, the
major lobe of each of the hybrids points in just one direction.
Using these HAOs, the description of the bonding in methane is straight-
forward and is shown in Fig. 6.6. Each sp? HAO overlaps with one of the
hydrogen Is orbitals, forming a bonding MO and an anti-bonding MO; two
anti-bonding |
MOs
I
Fig. 6.6 MO diagram for methane. On the left are shown the four equivalent sp? HAOs on carbon, and
on the right are the four 1s orbitals on hydrogen. One HAO overlaps with one hydrogen orbital to give
a bonding and an anti-bonding MO; two electrons occupy the bonding MO, giving a 2c-2e bond. The
other HAOs overlap, one-on-one, with the other hydrogen orbitals; it is important to understand that
each HAO only overlaps with a single hydrogen orbital.
electrons are placed in each bonding MO thus generating four 2c-2e bonds and
using up all eight valence electrons. The use of HAOs has allowed us to de-
velop a simple MO description of the bonding in methane which matches up
with the 2c-2e description.
Just as the energies of MOs are different from those of the AOs from which
they are formed, the energies of the HAOs are different from the AOs which
form them. The energy of the four equivalent sp> HAOs lies between that of 1 o——
the 2s and 2p but closer to 2p than 2s, as shown in Fig. 6.7.
%
Ammonia and water
2s
The sp* hybrids can also be used to give an approximate description of the
AQs HAOs
bonding in ammonia (NH3) and water (H20). Ammonia is best described as
Fig. 6.7 Illustration of the relative
being trigonal pyramidal, a shape which we can imagine as being derived from
energies of the sp? HAOs and the AOs
tetrahedral methane simply by ‘removing’ one of the hydrogens, as shown in from which they are formed. Note that the
Fig. 6.8. This gives us the clue as to how to describe the bonding in ammonia: three 2p orbitals are degenerate, but are
shown slightly separated on this diagram;
we form four sp> hybrids on the nitrogen and let three of these overlap with the same is true of the four sp? HAOs.
hydrogen 1s AOs to form three 2c-2e bonds. Ammonia has eight valence elec-
trons, so we assign the final pair not used for the N-H bonds to the fourth sp?
hybrid, classifying it as a lone pair.
Of course there is a problem with this description, which is that the H-N-H
bond angle in ammonia is 106.7° — not the same as the tetrahedral angle of
109.5° between the sp? hybrids. The difference is small, however, so our de-
scription is not far from the truth and we will see later on how to refine the
model to produce the correct bond angle.
The same approach can be used for water; we imagine its structure as being Fig. 6.8 The structures of ammonia and
derived from methane with two of the hydrogens removed. As before, we form water can be thought of as being based
four sp* hybrids on the oxygen, and overlap two of these with the hydrogens to on the tetrahedral shape of methane with
one and two of the bonds to hydrogen
form two 2c-2e bonds. The remaining four electrons are assigned to the other being removed, respectively.
two sp® hybrids, forming two lone pairs. Once again, the experimental bond
angle of 104.5° does not quite match the angle between the sp> hybrids, but it
is pretty close.
These sp> hybrids are not only useful for describing the bonding in
methane, but also give us a start in describing the bonding in other molecules.
In these we not only use the hybrids to form bonds but also use them to ac-
commodate lone pairs. We will now turn to how other hybrids can be formed
which point at different angles to the tetrahedral arrangement of sp° hybrids.
Other kinds of hybrid atomic orbitals

The molecule BH3 has a trigonal planar shape in which the three hydrogens lie
in a plane and at the corners of an equilateral triangle with the boron at its centre
(Fig. 6.9). Clearly the sp? hybrids are not going to be useful in describing the H—Bx ~wHH
bonding in this molecule — in fact what we need are the three equivalent sp”
hybrids which are formed from the 2s and two of the 2p AOs. These hybrids are Fig. 6.9 BHg is a trigonal planar
moiecule; all the atoms lie in a plane,
depicted in Fig. 6.10 along with the remaining 2p, orbital which is not involved with the hydrogens at the corners of an
in forming the hybrids (the choice of the remaining orbital as pointing along z equilateral triangle.
is arbitrary).
Each sp” hybrid overlaps with a hydrogen 1s AO to forma bonding and an
anti-bonding MO; a pair of electrons is assigned to each bonding MO, creating
(not involved in hybridization)
Flg. 6.10 Surface plots of the three equivalent sp” hybrids formed from the 2s, 2px and 2py AOs; also
shown is the 2pz AO which is not involved in the hybridization. The three hybrids point at 120° to one
another and all lie in the xy-plane.
three 2c-2e bonds. BHs3 has six valence electrons, so there are just enough
for these three 2c-2e bonds. The remaining out-of-plane 2p, orbital is not
occupied, but we will see later on that this orbital plays an important role in the
chemical reactions of BH3. The complete MO diagram is shown in Fig. 6.11.
We can also form hybrids of 2s with just one 2p (arbitrarily chosen to be
the 2p,) to give two equivalent sp hybrids; these are depicted in Fig. 6.12 along
=.
with the two unhybridized 2p orbitals. The sp hybrids point at 180° to one
weeeaes 2p, i 4
another.
2p = i 1s
A pattern now emerges: the angle between the sp* hybrids is 109.5°, be-
= ; tween the sp* hybrids it is 120° and between the sp hybrids it is 180°. The
” spr’. :
greater the proportion of s-character in the HAOs the larger the angle between
25——" = them. Put the other way, the greater the p-character the smaller the angle be-
AOs HAQOs MOs AOs tween the hybrids. So, by altering the ratio of s to p in our HAOs we can alter
the angle between them.
Fig. 6.11 MO diagram for BH3; on the left When we combined the 2s with the three 2p AOs we described the result-
are shown the AOs of boron combining to ing sp? hybrids as equivalent, which means that each has the same ratio of s-
give three sp? hybrids and leaving the
2pz orbital unhybridized. The three
and p-character; this is why they all have the same angle between them. How-
HAOs overlap one-on-one with the three ever, if we increase the s-character in one of the hybrids and proportionately
hydrogen 1s AOs shown on the right. decrease the amount of s-character in the other three, we expect to decrease
The molecule has six valence electrons
which occupy the three bonding MOs;
the angle between the orbitals which have lower s-character (and hence greater
the 2pz is empty. p-character). There are still four HAOs, but they are no longer equivalent.
This set of inequivalent hybrids is just what we need to describe the bond-
ing in ammonia. The three hybrids with increased p-character are used to
form the bonds to the hydrogens and the fourth hybrid (the one with greater
s-character) accommodates the lone pair. Similarly, to describe the bonding in
2p, AO 2p, AO
(not involved in hybridization) (not involved in hybridization)
Fig. 6.12 Surface plots of the two equivalent sp hybrids formed from the 2s and 2pz AOs; also shown
are the 2px and 2py AOs which are not involved in hybridization. The two hybrids point at 180° to one
another, and the unhybridized 2p orbitals point in perpendicular directions to the two hybrids.
water, we increase the p-character in two of the hybrids until the angle between
them is the required 104.5° and use these to form the 2c-2e bonds. The remain-
ing two hybrids (which have increased s-character) are used to accommodate
the two lone pairs.
The idea that the proportion of s and p AOs in the hybrid atomic orbitals
can be varied in order to account for different bond angles is nicely illustrated
by the inversion of ammonia.
The inversion of ammonia

H H H
Although the equilibrium geometry of ammonia is trigonal pyramidal it turns
out that this molecule constantly undergoes a process called inversion by which His
y|
ee fie
Y ae tty
i
it is pulled ‘inside out’ like an umbrella on a windy day (Fig. 6.13). The H HH H
trigonal trigonal
molecule starts in its lowest energy geometry (a trigonal pyramid) and then
pyramidal planar pyramidal
the NH bond angles start to open out until the molecule becomes planar. One
way of describing what has happened is to say that the three hydrogens have
energy
moved from a plane below the nitrogen into a plane containing the nitrogen.
The process then carries on with the three hydrogens moving to a plane above
the nitrogen as the bond angles return to their equilibrium values. We regain Fig. 6.13 Ammonia undergoes an
the original geometry but with the hydrogens on the other side of the molecule inversion in which it turns ‘inside out’ by
compared to where they started from. passing from the equilibrium trigonal
pyramidal geometry to a planar geometry
We have already described the bonding in ammonia as being based on sp? and then continuing to a trigonal
hybridization, but with the lone pair occupying.a hybrid of greater s-character. pyramidal geometry which is the mirror
The planar form of NH; can best be described as having sp~ hybridization at image of the original. The energy is a
maximum at the planar geometry and at
the nitrogen. These three hybrids overlap with the three hydrogen 1s orbitals, this point is 24 kJ mol above the
and a total of six electrons are assigned to the resulting bonding MOs. The energy at the equilibriurn geometry.
remaining pair of electrons is assigned to the out-of-plane 2p, AO.

So, as the inversion proceeds we have an interesting progression of bond-
ing. To start with the lone pair is in an sp type HAO with more s-character; as
the inversion proceeds the p-character of this HAO increases reaching 100% at
the planar geometry. At the same time, the HAOs involved in bonding to the
hydrogens decrease in p-character as they move from sp? to sp”. Then, as the
trigonal geometry is regained on the other side, the s-character of the lone-pair
orbital increases and that of the HAOsjnvolved in bonding decreases.
Tetrahedral ions
The ionic species NH7 and BH; are both tetrahedral and have eight valence
electrons each. For the case of NHj the electron count is five from nitrogen,
four from the four hydrogens and —1 for the positive charge giving a total of
8. For BH; we count three from boron, four from the four hydrogens and +1
for the negative charge, giving a total of eight.
To describe the bonding for both ions we form four equivalent sp° hybrids
on the central atom and allow these to overlap (one-on-one) with the four hy-
drogen 1s AOs. Placing two electrons in each bonding MO uses up all eight
electrons to form four 2c-2e bonds.
We might ask on which atom the charge is located: from the MO picture it
is clear that the charge cannot be associated with any one atom, but is shared
around all of the bonds. This is an important point which we will return to
when considering the reactions of these two species.
6.3 Using hybrid atomic orbitals to describe diatomics

Hybrid atomic orbitals can be used to describe the bonding in diatomics and
sometimes this results in a more convenient description than the full MO treat-
ment which we gave in Chapter 5. We will describe just one example: N2.
We start out by supposing that the nitrogen is sp hybridized, with the hy-
brids pointing along the direction of the internuclear axis (z). The two sp HAOs
which point towards one another overlap to form o bonding and anti-bonding
MOs; we place two electrons in the bonding MO. The other two HAOs which
point away from the bond are occupied by two lone pairs.
There are two sets of out-of-plane p orbitals: the two 2p, orbitals overlap to
give a bonding MO and a corresponding anti-bonding MO; the two 2py AOs
overlap in the same way. Two electrons are placed in each of the 7 bonding
MOs, using up all 10 valence electrons.
Our description of the bonding in N2 is that there is a o bond and two
m bonds between the nitrogen atoms; in addition each atom has a lone pair,
which is directed away from the internuclear axis. The picture is summarized
in Fig. 6.14.
In fact, this picture is not too dissimilar to the MO description of the bond-
ing in N2 given on p. 74 (see the MO diagram of Fig. 5.18 (b) on p. 72). In
the MO picture the s-p mixing causes the 1a, MO to become less anti-bonding
and the 2a¢ orbital to become less bonding; it turns out that occupation of these
orbitals contributes rather little to the net bonding so they can be described as
On@@ro 6 bonding MO x ne
N e ©N
C)
N——On@ m bonding MO from 2p,

F
sp lone pair y eS C)
eu sp lone pair
es N
m bonding MO from 2p,

N
Fig. 6.14 Cartoons showing a description of the bonding in No using sp hybrids. Two of the hybrids
overlap to form a o bonding MO and the remaining two hybrids become lone pairs. A x bonding MO
is formed from the overlap of the two 2px AOs, and similarly another x bonding MO is formed from
the overlap of two 2py AOs. In these terms, the description of No is of a « bond, two x bonds and
two lone pairs. Note that for the a overlap of the two sp hybrids the lobes of the HAOs are shown with
the signs that will lead to the bonding MO; do not forget, though, that an anti-bonding MO will also be
formed. The same applies to the x overlap which gives a bonding and an anti-bonding MO.
approximately non-bonding. It is the occupation of the log MO (two electrons)

and 27, MOs (four electrons) which is responsible for most of the bonding in
N2; in other words there is a o and two bonds just as in the description given
in Fig. 6.14.
6.4 Simple organic molecules

Hybrid atomic orbitals are a very convenient way of describing the bonding in
simple organic molecules so in the following chapters we will use this approach
very often. Generally, it is clear from the structure what kind of hybrids we
should use for a particular carbon: if the coordination iis tetrahedral we need sp*
hybrids, if the carbon is doubly bonded we need sp” hybrids and if it is triply
bonded we need sp hybrids. For other atoms, such as nitrogen and oxygen, the
choice is sometimes less clear and we will discuss this later on in this section.
First, however, we will look at the description of the bonding in three simple
hydrocarbons.
Ethane, ethene and ethyne

In ethane, C2H¢, both carbons are tetrahedral and so it is clear that we should sp*-sp*
use sp? hybrids to describe the bonding. Of the four hybrids on each carbon, H H
one overlaps with a hybrid on the other carbon and the other three overlap with
"NS, — d
hydrogen Is orbitals. Each overlap is just between two orbitals and results in seis |/ \" 4
a a bonding and ao anti-bonding MO. We put two electrons in each bonding
MO thus forming a 2c-2e bond; the overall picture is summarized in Fig. 6.15. Fig. 6.15 Schematic structure of ethane
There are 14 valence electrons in ethane (four from each carbon and six from in which each line represents a 2c-2e
the hydrogens), and these fill all of the a bonding MOs. bond. Each bond can be described as
being the result of the overlap between
Ethene, C2H4, is flat with 120° H-C-H and H-C-C bond angles; the choice either two sp? hybrids (for the C-C bond)
of sp” hybrids for the carbons is therefore clear. The bonding in this molecule or an sp® hybrid and a 1s AO on
separates into two parts: firstly, a o framework which involves the sp” hybrids hydrogen (for the C~H bonds).
sp?-sp*
H
\ /
" @
Hing Cc wth
eis NY "OF
o framework m system on o framework
~
; ty
Fig. 6.16 The bonding in ethene can be separated into a o framework, involving the overlap of sp*
hybrids on carbon and 1s AOs on hydrogen, and a x system which lies out of the plane of the a
framework. The out-of-plane 2p orbitals are shown with the signs that lead to the 7 bonding MO, but
do not forget that there will also be a w anti-bonding MO formed by the destructive overlap of these
two 2p orbitals.
and the 1s AOs; secondly, the 2 bonding between the two carbons.
The o framework is much the same as that for ethane with the single excep-
tion that sp”, rather than sp?, hybrids are involved. For the C-C bond, two sp?
hybrids overlap to give o bonding and anti-bonding MOs; for the C-H bonds
the overlap is between an sp* hybrid and a hydrogen 1s orbital. Filling all of
these o bonding MOs accounts for 10 electrons.
Pointing out of the plane of the molecule are the two 2p, carbon AOs which
were not involved in the hybrids (see Fig. 6.10 on p. 86). These two orbitals
overlap to give az bonding and a z anti-bonding MO. Assigning two electrons
to the bonding MO accounts, together with those in the o framework, for all
of the 12 valence electrons in ethene (four from each carbon and four from the
hydrogens). Figure 6.16 summarizes the picture of the bonding which we have
developed: the two carbons are joined by ao bond and a z bond, forming the
double bond which we expect for ethene.
This diagram also explains a marked difference between ethane and ethene.
Whereas in ethane there is quite a small energy barrier for rotation about the
C-C bond, in ethene a great deal of energy is needed to rotate about the double
bond — so much that we can safely regard it as being fixed. The easy rotation
about the single bond in ethane comes about because the bond is formed by
aa MO in which the orbitals overlap head-on. Rotation about the bond axis
makes no difference to the overlap, and so does not affect the bonding.
In contrast, forming the z bond in ethene requires that the two 2p AOs lie
in the same plane, as shown in Fig. 6.16. Rotating about the C-C axis will
result in the loss of this overlap, thus breaking the 2 bond; this is why rotation
about this bond needs so much energy.
Finally, we come to ethyne (acetylene), C2H2, which is a linear molecule;
sp hybridization at the carbons is therefore appropriate, giving the bonding
scheme illustrated in Fig. 6.17. As with ethene, we can separate the bonding
into ao framework and a z system. This time each carbon has two out-of-
plane 2p orbitals (see Fig. 6.12 on p. 87) which overlap in two pairs to give two
mz bonding MOs and two z anti-bonding MOs. Occupation of all of the o and
x bonding MOs accounts for all 10 valence electrons.
o framework
m™ system on o framework
Fig. 6.17 The bonding in ethyne (acetylene) is similar to that in ethene except that the carbons are
now sp hybridized and there are two sets of x interactions.
Carbonyl compounds
The simplest carbonyl compound is methanal (formaldehyde), CH20, shown in
Fig. 6.18. The molecule is flat, with bond angles at the carbon of approximately
120°, so the natural choice of hybridization for this atom is sp”. For the oxygen,
the choice is less clear; we know that there is a double bond between the C and
the O, so we need to have at least one unhybridized 2p orbital on the oxygen
which we can use to form a z bond. So the hybridization needs to be sp? or sp;
for the moment we will choose sp”.
The o framework is simple to describe. One of the carbon sp? hybrids
overlaps with a hydrogen Is to give a o bonding and anti-bonding MO; two
electrons are placed in the bonding MO forming the C-H o bond. The other
C-H bond is formed in the same way. The third carbon sp” hybrid overlaps
with one of the oxygen sp? hybrids to give a bonding and an anti-bonding o
MO; two electrons occupy the bonding MO to form the C-O o bond.
As the oxygen AOs are lower in energy than those on carbon, the HAOs on
oxygen will also be lower in energy than those on carbon. Therefore the major
contributor to the o bonding MO will be the oxygen based orbital; we therefore
expect the o bond to be polarized towards the oxygen. The remaining two sp”
hybrids on oxygen are not needed for bonding; they are each occupied by two
electrons and so form two lone pairs.
H
\o — methanal
/ (formaldehyde)
H
\ C—O. @ @
H
=s" Hing ——9

eis H sp’-sp* "OO
° framework m™ system on o framework
and lone pairs
Fig. 6.18 Methanal (formaldehyde) has a carbonyl group, that is a C-O double bond. One description
of the bonding involves sp* hybridization of both the carbon and oxygen. One of the hybrids on oxygen
forms ao bond to the carbon, and the other two are occupied by lone pairs.
The z system is essentially the same as in ethene, with the out-of-plane

(a) 5 (b) ; 2p orbitals from oxygen and carbon overlapping to give a m bonding and a
m anti-bonding MO; two electrons occupy the bonding MO, forming the z
“SO
Hino O Heine O bond between the carbon and the oxygen. The difference between the 2 bonds
in ethene and methanal is that whereas in ethene the two AOs have the same
energy, in methanal the oxygen 2p AO is lower in energy than that from carbon.
As a result, the oxygen AO is the major contributor to the bonding MO and the
Fig. 6.19 Cartoons of (a) the bonding
resulting 2 bond (like the o bond) % pdlarized towards oxygen. The major
and (b) the 7 anti-bonding MOs ina
carbonyl group. The AO on oxygen is contributor to the 2 anti-bonding MO is the carbon, and we will see later on
lower in energy than that on carbon, and that this is crucial in explaining a great deal of the chemistry of the carbony]
so the oxygen AO is the major contributor
to the bonding MO; for the anti-bonding
group. Cartoons of these MOs are shown in Fig. 6.19.
MO, it is the carbon AO which is the We should check that all the electrons are accounted for: methanal has
major contributor. 12 valence electrons (two from the hydrogens, four from carbon and six from
oxygen). In our description there are three pairs in o bonding MOs, one pair
in a x bonding MO and two lone pairs in sp? hybrids on the oxygen; all the
electrons are therefore accounted for.
The description of the bonding in ethanal (acetaldehyde, Fig. 6.20),
CH3CHO, is trivially different from methanal. The methyl carbon is sp? hy-
=
bridized and uses these HAOs to form the o bonds to the carbonyl! carbon and
C= ethanal
a
(acetaldehyde)
the three hydrogens. The description of the bonding in the carbonyl] group is
the same as before.
H3C
We could have chosen the carbonyl oxygen to be sp hybridized, in which
Fig. 6.20 The structure of ethanal case one of the lone pairs will occupy an sp HAO and one will occupy the 2p
(acetaldehyde).
oxygen AO which is at right angles to the one forming the 2 bond. It makes
little practical difference which hybridization state we choose for oxygen as
the key features of the bonding in the carbonyl group are not really affected by
this choice.
The simplest ester is methyl methanoate (methyl formate), shown in
Fig. 6.21. As in methanal the carbonyl carbon and the carbony! oxygen are
sp? hybridized, leading to the same description of the C-O double bond. The
C-O-C bond angle at the ester oxygen is 127 ° so sp? hybridization is probably
the most appropriate choice for this atom. The methyl carbon is sp? hybridized,
as before.
The o framework therefore consists of the C—O bond in the carbonyl! group,
the C-H bond to the carbonyl! carbon and the following additional interactions:
7 oe sp?-sp?
\ H
C=O sp-sph— \ (
Oo Sp— ~~. 7 °
Non op. ay ;
~ \ sp*-is
i H sp*-sp* Ny
H
methyl methanoate
(methyl! formate)
Fig. 6.21 Description of the bonding in methyl methanoate (methyl formate); only the o
framework and
the lone pairs are shown, the x system is the same as in methanal (Fig. 6.18 on
p. 91). The ester
oxygen is sp hybridized, so one of the lone pairs is in the 2p orbital which points
out of the plane in
which the hybrids lie.
6.5 Reconciling the delocalized and hybrid atomic orbital approaches 93
three C—H bonds in the methyl group, each formed by the overlap of an sp? hy-
brid and a hydrogen 1s; a bond between the methyl carbon and the ester oxygen
formed by the overlap of a carbon sp? and an oxygen sp? hybrid; and a bond
between the carbonyl carbon and the ester oxygen, formed from sp* hybrids
on each. The remaining sp* hybrid and the unhybridized 2p AO on the ester
oxygen are occupied by lone pairs. Figure 6.21 summarizes the description.
The description of amides, such as methanamide (Fig. 6.22), is a little more
complex as it turns out that the lone pair on the nitrogen interacts with the car-
H
bonyl z system. We will discuss how this comes about and the consequences
it has for both the structure and reactions of amides later on in Section 10.2 on
\o — methanamide
/ (formamide)
p. 160.
HoN
Nitriles Fig. 6.22 The structure of methanamide
(formamide).
Nitriles, such as ethanenitrile (acetonitrile), CH3CN, shown in Fig. 6.23, have a
triple bond between carbon and nitrogen; as with ethyne (p. 90) both the carbon
and the nitrogen are sp hybridized. The two hybrids which point towards one
another along the C—N axis overlap to form o bonding and anti-bonding MOs;
two electrons occupy the bonding MO to form the C-N o bond. The two pairs
of out-of-plane 2p orbitals overlap to form two z bonding and two z §anti-
bonding MOs. Together, occupation of the o and the two m bonding MOs
forms the triple bond.
The remaining sp hybrid on nitrogen is occupied by two electrons forming
a lone pair; the remaining sp hybrid on carbon is involved in forming a o bond
to the methyl carbon.
H3C—-C==N _ ethanenitrile
SP*SP > HgC @@

H3C 7-C——N:i~—sp
OO
soso
o framework H3C 99
m system on o framework
Fig. 6.23 The C-N o bond in ethanenitrile is formed using sp hybrids on carbon and nitrogen, just
as in ethyne (Fig. 6.17 on p. 91); the second sp hybrid on the nitrogen is occupied by a lone pair of
electrons. The out-of-plane 2p orbitals form the two x bonds.
6.5 Reconciling the delocalized and hybrid atomic orbital

approaches
The MO and HAO approaches take a very different view of the bonding in
a molecule. Whereas the MO picture combines AOs from several atoms, the
HAO picture is strictly a localized one in which bonding is between pairs of
atoms; electrons not involved in bonding (lone pairs) are assigned to orbitals
which are localized on a single atom. The question is, do the two approaches
give us the same picture of the bonding in a molecule?
Electrons in larger molecules
MO3 MO4
view A view A
MO3 MO4
view B view B
Fig. 6.24 Computer calculated surface plots of the occupied MOs of water. MO1 is the lowest in
energy, MO2 the next lowest, and so on for MO3 and MO4; this last orbital is the HOMO. All four
orbitals are shown in a view looking down onto the plane of the molecule (view A); a side view of MO3
and MO4 is also shown in view B.
We will explore this question by considering the case of water whose local-
ized bonding description was given on p. 85. In this description the occupied
orbitals are: (1) a 0 bonding MO between the oxygen and one of the hydro-
gens; (2) an identical MO between the oxygen and the other hydrogen; (3) two
sp? type orbitals on oxygen, neither of which is involved in bonding.
A computer calculation gives the MOs shown in Fig. 6.24; at first sight
these do not appear to have much connection with the localized picture, but on
closer inspection it turns out that there are similarities.
Consider first the two lowest energy MOs, MOI and MO2. If we form the
combination (MOI + MO2) there will be reinforcement on the right-hand side,
where both orbitals are positive, but on the left-hand side the negative part
of MO2 will cancel with the positive MO1; the process is shown in cartoon
form in Fig. 6.25. The result will be an orbital which is mainly between the
oxygen and the right-hand hydrogen; this is the localized orbital between these
two atoms. In a similar way if we form the combination (MO1l — MO2) we
end up with an orbital which is mainly between the oxygen and the left-hand
hydrogen. So, MOI and MO2 can be thought of as representing between them
6.6 Orbital energies: identifying the HOMO and LUMO 95
the two localized o bonds between the oxygen and the two hydrogens.
This leaves the other two orbitals MO3 and MO4 which, in the localized
picture, ought to correspond to our two lone pairs. MO3 is essentially an out-
of-plane 2p orbital which is clearly non-bonding. However, MO4 is a bonding
MO as it has some electron density shared between the hydrogens and the
oxygen. Clearly, then, MO3 and MO4 do not correspond simply to the two sp
hybrids from the HAO description. MO1 M02
In the hybridization approach we made the four sp* hybrids equivalent, Fig. 6.25 Cartoons showing the way in
SO it is inevitable that when we use these HAOs to describe the bonding in which combinations of MO1 and MO2
water the two orbitals which are occupied by the lone pairs will have the same from water (shown in Fig. 6.24) give rise
to localized o bonding orbitals. The
shape and energy. The MO description imposes no such restrictions and so we combination MO1 + MO2 results in
should not be surprised to find that the orbitals which are occupied by the lone cancellation on the left-hand side and so
pairs are not the same. Experiments confirm that this prediction from the MO gives an orbital which is mostly between
the oxygen and the right-hand hydrogen.
calculation is indeed correct. Similarly, the combination MO1 — MO2
The computer generated MOs are certainly a more precise description of gives an orbital which is mostly between
the electronic structure of water than is our simple HAO approach. We could the oxygen and the left-hand hydrogen.
hardly expect it to be otherwise as the HAO approach is purely qualitative

and based on simple approximations such as only permitting the interaction
between pairs of orbitals. Nevertheless the HAO approach does capture the
essence of the bonding in water, namely that there are o bonds between the
oxygen and the hydrogens, together with extra electron density on the oxygen.
If we want to think about the shape of the water molecule or the way in which
it reacts, the HAO picture is a perfectly good starting point.
In summary, although we recognize the deficiencies of the HAO approach,
it does allow us to make a reasonable start at describing how a molecule is
bonded together, and the form of the orbitals involved, without having to resort
to computer calculations. A further advantage of the HAO approach is that it
clearly associates MOs with particular bonds — an idea which fits in well with
the way we think about reactions.
When it comes to describing the bonding in larger molecules we will find
that a useful approach is to use HAOs to describe the o-framework and then
use the fully delocalized MO approach to describe the rest of the bonding.
Typically this means using the MO approach to describe the 7-system. Often
the o-framework is not involved in reactions, so it does not matter too much
if our description of the framework is approximate. The more precise MO
approach is reserved for dealing with the part of the molecule which is actually
involved in the reaction. We will use this approach extensively in Chapters 10
and 11.
6.6 Orbital energies: identifying the HOMO and LUMO

When it comes to chemical reactions it is very important to identify the highest
occupied MO (the HOMO) and the lowest unoccupied MO (the LUMO). This
is because, as we shall see in the next chapter, these orbitals are the ones which
are most directly involved in reactions. Of course, we can always use a com-
puter program to compute the MOs and so identify the HOMO and the LUMO,
but in many cases it is not really necessary to go to this level of sophistication.
For our purposes, it will be sufficient to use the HAO approach to draw up
a picture of the bonding. As we have seen in this chapter, for simple molecules
this approach allows us to identify electron pairs as being involved in 2c-2e
bonds (in o or x bonding MOs) or localized on atoms as lone pairs. All we
then need to do is to have a method of ordering these orbitals in energy so that
we can determine the HOMO and LUMO.
o* anti-bonding MOs Typically the ordering of energies is that shown in Fig. 6.26. The o bonding
energy ————{>
MOs come lowest, followed by 7 boftding MOs; as we commented on above,

iW THE
n* anti-bonding MOs
the sideways-on overlap which leads to MOs is not generally as effective as
the head-on overlap which leads to a MOs.
non-bonding orbitals Not surprisingly, next after the bonding orbitals come the non-bonding or-
bitals, which can be AOs, HAOs or, as we shall see, MOs. Finally, higher still
are the anti-bonding MOs (indicated by a superscript *). For the same reason
nm bonding MOs
that the o bonding MO lies below the z bonding MO, the o” lies above the z*
6 bonding MOs MO.
This order of energies is not immutable, but it is a good starting point. So,
Fig. 6.26 The approximate ordering of for example, we can easily identify the HOMO in methanal and ethanal as a
orbital energies found in molecules;
anti-bonding MOs are indicated by a lone pair on the oxygen and the LUMO as the z* anti-bonding MO between
superscript *. the carbon and the oxygen.
7 Reactions
In Chapter 6 we looked at the molecular orbitals in some simple compounds.

We now want to see how these are involved in chemical reactions. The energies
of the orbitals help us to understand why the reaction occurs and the shapes are
crucial in understanding how the reaction takes place. As a model, we shall first
consider the hypothetical reaction between a single proton, H*, and a hydride
ion, H~, to produce H:
Ht +H” —> Hp.
7.1. The formation of Hz from H* and H—

It might be tempting to say that this reaction occurs because the reactants are
oppositely charged; however there is more to it than this. The reaction does
not lead to an ion pair held together by electrostatic attraction, as shown in
Fig. 7.1 (a), but to the formation of a molecule where the electrons are shared
between the two nuclei, as shown in (b). It is true that to start with, due to their
opposite charges, the ions are attracted to one another but as the separation be-
tween them decreases the electrons are redistributed. Initially both electrons H° + H°——» H—-H
are associated with one nucleus, but in the product they are located in a molec- Fig. 7.1 The reaction between H~ and
ular orbital and shared between the two nuclei. This redistribution is shown H* does not give an ion pair, as shown in
(a), but a molecule of Hz, as shown in
in Fig. 7.2 which compares the electron density of the separate ions and the
(b).
product, H2.
H- He H5
Fig. 7.2 Contour maps showing the electron density associated with (a) the H~ ion and a separate
proton (H+), and (b) with the Hz molecule.
The sharing of the electrons has led to a product lower in energy than the
starting materials — it is an exothermic process. Even though the entropy of the
system must have decreased as the two ions combined to form one molecule,
the entropy of the Universe has increased due to the heat given out to the sur-
roundings during this exothermic process. re Ho ——» H—H
Chemists illustrate this redistribution of electrons by using curly arrows, Fig. 7.3 A curly arrow mechanism for the
Fig. 7.3 shows how we can represent the reaction between Ht and H™ using reaction between a hydride ion and a
this approach. A curly arrow represents the movement of a pair of electrons. proton.
Reactions
98
The arrow starts where the electrons are in the reactant (here in the H™ ion)
and shows where this pair of electrons ends up in the product (in this case in
between the two nuclei in the new bond).
This curly arrow representation of the reaction between H* and H™ is prob-
ably not a realistic description of what actually happens. Rather than the two
electrons being shared, what probably happens in this case is that one electron
is transferred from H~ to H+ to give two H atoms which then come together
to form H2. a >
Nonetheless, hypothetical though the reaction is, it is a good starting point
for understanding reaction mechanisms. In the next section we will use this
approach to discuss a real reaction in which a bond is formed by the sharing of
electrons.
7.2 Formation of lithium borohydride

Lithium hydride, LiH, and borane, BH3, react to form lithium borohydride,
LiBHg:
LiH + BH3 — LiBHg.
This reaction occurs very readily at room temperature when borane (which
actually exists as a dimer, B2He — see Fig. 6.4 on p. 82) is mixed with the
lithium hydride in ethoxyethane (diethyl ether) as a solvent.
In this reaction there is a decrease in the entropy of the system since two
molecules go to one. Just as in the case of Ht + H™ the reaction is exothermic
so there is an increase in the entropy of the surroundings which, it turns out,
more than compensates for the decrease in the entropy of the system. The
reason why the reaction is exothermic is because the energy of the electrons
in the product is lower than in the reactants. We need to look at the molecular
orbitals in the reactants and products to understand why this is so.
The left-hand column of Fig. 7.4 shows the MOs of LiH, together with
their approximate energies, which have been calculated using a computer pro-
gram. We have already drawn a qualitative MO diagram for this molecule,
Fig. 5.23 on p. 78. In doing this we ignored the 1s orbital on lithium, arguing
that it was too low in energy to interact with the hydrogen Is. In Fig. 7.4 MO-
LH1 is essentially this unchanged lithium 1s orbital; MO-LH2 is the bonding
MO formed by overlap of the lithium 2s with the hydrogen 1s (labelled lo in
Fig. 5.23, which is also’ shown as a contour plot in Fig. 5.24 on p. 78). The
unoccupied anti-bonding MO is not shown.
In Chapter 6 we used sp* hybrids to draw up an approximate MO diagram
for BH3 (Fig. 6.11 on p. 86); as for LiH, we ignored the 1s orbital on B. We
identified three bonding MOs and one non-bonding MO, the out-of-plane 2p
orbital. The right-hand column of Fig. 7.4 shows computer calculated MOs for
BH3. As we described in Section 6.5 (p. 93) the MOs calculated in this way
are delocalized and so do not compare directly with the simpler localized HAO
description. However, we can clearly identify three bonding MOs, MO-B2,
MO-B3 and MO-B4, together with the non-bonding MO, MO-BS, which is the
LUMO. MO-B1 is essentially the boron |s orbital which is not involved in the
bonding.
7.2 Formation of lithium borohydride 99
Lithium hydride MOs Lithium borohydride MOs Borane MOs
a deaenans
or a- —_— —
ers x rf f
MO-LH2 (LiH HOMO) MO-P6 (LiBH, HOMO) MO-B5 (BH, LUMO)

-8 eV -11 eV +2 eV if occupied
ees
MO-P5
© -11 eV
~12 eV
ze
him \
e- mice SO HO
a
~ a
N 8 f
Boxe; MO-P3 MO-B2

-19 eV
-19 eV
“e- > - O
oe MO-P2
eae -67 eV
ae if H Pee
«-- eae fe
Li ~ ¥ MO-P1
SH/B -203 ev
Fig. 7.4 The left-hand column shows the occupied MOs of LiH. The energies are given in eV; a more
negative energy means a more tightly held electron. The right-hand column shows the occupied MOs
and the LUMO for BH3. The middle column shows the occupied MOs for the product, LiBH4. The most
important orbitals for each molecule are shown at the very top. In the LiBH, structure the dashed lines
are used to illustrate the arrangement of the atoms rather than to imply the existence of bonds.
Just as we can think of a molecular orbital as being formed from two atomic
orbitals, we can think of the MOs in the products as arising from the interac-
tions between the MOs in the starting materials. The middle column of Fig. 7.4
shows the occupied molecular orbitals in the product, LiBH4.
100 Reactions
Unoccupied MO
(not shown) OC)
Hi
HP,
@ ++ ©
HOMO + LUMO
of LiH of BH
(MO-LH2) D ) (MO-BS)
Li H'!tiuiB—H
HOMO in
LiBH,
(MO-P6)
Fig. 7.5 Schematic energy level diagram for the interaction between the HOMO of LiH and the LUMO
of BH3.
The lowest energy MO, MO-P1, is essentially the unaffected boron 1s AO

which correlates with MO-B1 from BH3. Likewise, MO-P2 is the 1s AO on
lithium which correlates with MO-LHI on LiH. The boron Is orbital is much
lower in energy than the lithium because boron has a larger effective nuclear
charge. These 1s orbitals remain virtually unchanged in the product.
Comparing the MOs in borane and the product we see that MO-B2, MO-B3
and MO-B4 have similar forms and energies to MO-P3, MO-P4 and MO-P5,
respectively. Thus occupation of these MOs makes little contribution to the
energy change of the reaction.
In contrast, the HOMO of LiH (MO-LH2) and the LUMO of BH3 (MO-B5)
are changed significantly on forming the product. MO-LH2 has no compara-
ble orbital in the product and even though the HOMO in the product (MO-P6)
looks rather like the LUMO in borane (MO-BS) their energies are very differ-
ent, with MO-P6 being much lower.
In summary, we can attribute the lowering in energy when LiBHg is formed
to the two electrons from the HOMO of LiH (MO-LH2) moving into the lower
energy HOMO of LiBH4 (MO-P6). This orbital can be thought of as arising
from the interaction of the HOMO of LiH with the LUMO of BHs3, as is shown
schematically in Fig. 7.5. As we will see in the next section, such HOMO-
LUMO interactions are often of particular importance in reactions.
HOMO-LUMO interactions
Usually, it is the interaction between filled orbitals of one reactant with va-
cant orbitals of the other which leads to a favourable lowering of the energy.
The interactions between two vacant orbitals is of no consequence as they are
unoccupied and so there is no change in the energy of the system. The interac-
tions between two filled orbitals have little net effect on the total energy of the
system since the decrease in energy from the newly formed in-phase combina-
tion is essentially cancelled out by the increase in energy of the corresponding
out-of-phase combination. Only the interaction between a filled orbital and an
empty orbital leads to a net lowering of the energy of the electrons, as shown
in Fig. 7.6.
7.2 Formation of lithium borohydride 101
(a)
a (b)
Fig. 7.6 The interaction of two empty orbitals as shown in (a) does not lead to a lowering in energy
since there are no electrons whose energy is being lowered. In the interaction between two filled
orbitals, shown in (b), two electrons are lowered in energy but this is cancelled out by the two electrons
which are raised in energy. Only in (c), the interaction between an empty and a filled orbital, is there a
net lowering in the energy since the orbital which is raised in energy is unoccupied.
Of course there are many filled and vacant orbitals which can interact — the
question is which interaction is the most important? Recall from the discussion
on p. 67 that the strongest interaction is between orbitals of similar energies,
so the best candidates for a favourable interaction will be an unoccupied MO
and an occupied MO which are close in energy.
If the energies of the MOs of the two reactants are broadly similar, then
we can expect the best energy match to be between the LUMO of one with
the HOMO of the other, as shown in Fig. 7.7. There is no guarantee that this
HOMO-LUMO interaction will be the most favourable, but it turns out that it
is often so.
When two species meet, two HOMO-LUMO interactions are possible;
there is an interaction between the HOMO of reactant A with the LUMO from
Hao
reactant B and also between the HOMO from B with the LUMO from A. The HOMO s
very best interaction is usually between the occupied MO which has the highest
+ . HOMO
a te
energy of either reactant with the LUMO from the other reactant. In Fig. 7.7,
the best interaction (shown with the wavy line) is between the HOMO from
reactant A with the LUMO from B since these two MOs are separated by the
smallest energy difference. The interaction (shown with the dotted line) be- MOs for MOs for
tween the HOMO of B and the LUMO of A is less significant because the reactant A reactant B
separation in energy is greater. Fig. 7.7 The strongest interactions are
Frequently we will find that we can understand the outcome of a reaction between the HOMO of one molecule and
the LUMO of another. For the set of MOs
between two molecules by first identifying which molecule has the very highest
shown here, the best interaction, shown
occupied MO and then considering the interaction between this and the LUMO by the wavy line, is between the HOMO
of the second molecule. In the reaction between LiH and BHs3, the very highest of reactant A and the LUMO of reactant
B. Since the HOMO of A is particularly
occupied MO is the HOMO of the LiH (MO-LH2). This is considerably higher high in energy, the energy gap between
in energy than the HOMO of the BH3 (MO-B4) so the main interaction is this and the LUMO in B is particularly
between the HOMO from the LiH and the LUMO from the BH3, MO-B5. smail.
Orbitals and curly arrows

Just as we did for the Ht + H™ reaction, we can use curly arrows to illustrate,
in a concise way, what is going on in the reaction between LiH and BH3. From H H
studying the orbitals we see that a pair of electrons from the o-bonding MO of ui-Lny | Li
©
| 9
LiH (the HOMO - MO-LH2) end up in an orbital of LiBHy4 which contributes Bi,
rie H~ ¥
to the formation of a new B-H bond. Figure 7.8 illustrates this process using H H H
a curly arrow. The arrow starts where the electrons come from — the bond Fig. 7.8 The curly arrow mechanism for
between the Li and H — and finishes where the electrons end up, in a new bond the reaction between borane and lithium
between the H and the B. It tells us that the electrons that were bonding the Li hydride.
102 Reactions
and H are now helping to bond the H to the B. Since the lithium has effectively
lost its share of electrons from the Li-H bond, it ends up with a positive charge.
The BH3 molecule initially had no charge; in forming the product it has gained
one extra proton and the two electrons that were in the Li-H bond. It therefore
now has a negative charge.
In LiBHg, the lithium is mainly attracted to the rest of the molecule elec-
trostatically. Its MOs shown in Fig. 7.4 show little evidence of electrons being
shared between the lithium and the rest of the molecule and a calculation con-
firms that the lithium does indeed bear a substantial positive charge with the
corresponding negative charge being spread out over the boron and hydrogen
atoms.
The HOMO in lithium borohydride is still fairly high in energy. In the next
section we shall see how the interaction between this and a suitable LUMO
allows us to understand the use of LiBHy as a reducing agent.
7.3 Nucleophilic addition to carbonyl

Nucleophilic addition to a carbonyl group is one of the most important reac-
tions in organic chemistry. In this reaction a nucleophile attacks the carbonyl
group, forming a new bond to the carbon and simultaneously breaking the C-O
m bond; Fig. 7.9 shows the reaction of a general nucleophile, Nu. Hydroxide,
cyanide (CN7) and alkyllithium reagents (R-Li) are all examples of nucle-
ophiles which will attack carbonyl groups in this way.
fe) o
Lun,
R R ‘ Nu v'R'
R
Fig. 7.9 The reaction between a carbonyl group and a nucleophile, Nu~. In this case the nucleophile
is negatively charged, but this is not a requirement — neutral species can also act as nucleophiles.
This simultaneous breaking of the bond is the added complication we now

need to consider. In order to look at the changes in the MOs we will consider
O of the simplest possible example of this reaction which is the attack of a hydride
Ho 4 y —- 6, ion on methanal (formaldehyde) shown in Fig. 7.10.
H@ ~H H~ ¥,H
HOMO-LUMO interactions for the carbonyl and nucleophile
Fig. 7.10 The reaction between methanal
and hydride: this is an example of The two reactants are initially attracted towards each other since the carbonyl
nucleophilic addition. carbon has a partial positive charge and the hydride ion is negatively charged.
However, the two reactants do not just remain attracted towards each other
electrostatically; bonds are formed and broken. In other words, electrons are
redistributed.
The reaction starts with a negative charge on hydrogen and ends with the
charge on oxygen. Since oxygen has a greater effective nuclear charge than
hydrogen, having the increased electron density on oxygen is favourable and
so contributes to the lowering of the energy of the product.
We can see how this redistribution of charge takes place by looking at the
HOMO-LUMO interactions. The highest occupied MO in the whole system
is from the H™ ion. The vacant orbital closest to this in energy is the LUMO
7.3. Nucleophilic addition to carbonyl 103
methanal
LUMO
Ata
Pt
scans
“SEES
Fig. 7.11 The in-phase interaction between the methanal x* and the hydride HOMO. In diagram (a)
two orbitals are shown: the occupied hydride MO and the unoccupied methanal 7*. Diagram (b)
shows the orbital arising from the in-phase overlap of the HOMO and LUMO at an intermediate stage
in the reaction. Electron density is building up on the oxygen at the expense of the hydride. The
corresponding MO in the product is shown in diagram (c). This is essentially a p orbital on the oxygen,
i.e, a lone pair; there is little electron density on the hydrogen when compared to (a).
of the methanal, the 7*. Figure 7.11 (a) shows the initial interaction of these
two orbitals, and (b) shows the orbitals at an intermediate stage in the reaction.
The MO formed from the in-phase interaction between the hydride HOMO
and methanal LUMO shows that electron density is being transferred from the
hydride onto the oxygen. This is even more apparent in the product MOs shown
in Fig. 7.11 (c) where there is much more electron density now on the oxygen
and relatively little on the hydrogen.
The hydride HOMO and the methanal LUMO also combine out-of-phase
to give an anti-bonding orbital higher in energy than the methanal LUMO; this
anti-bonding orbital is not occupied.
The hydride HOMO is closest in energy to the CO 2* MO and so has the
strongest interaction with this orbital. However, it also has an interaction with
the CO x MO, and it turns out that it is this interaction which explains how the
CO x bond is broken. Fig. 7.12 shows the in-phase interaction between these
two orbitals.
At an intermediate stage, shown in (b), we see the interaction between the
carbon and oxygen begins to weaken. In the product, shown in (c), there is
methanal
mt MO
Fig. 7.12 The interaction between the methanal x MO and the hydride HOMO. Diagram (a) shows the
two orbitals as the two reactants approach each other. Diagram (b) shows the orbital arising from the
in-phase overlap of these two orbitals at an intermediate stage in the reaction. The bonding between
the C and O is beginning to weaken while that between the C and the approaching H is strengthening.
In the corresponding MO in the product, shown in (c), there is effectively no bonding between the C
and O, only between the C and H.
104 Reactions
electron density between the C and the H but rather little between the C and O.
Just as we expected, a bond is formed between the C and the H and the C-O x
bond is broken. ,
The interaction of these three orbitals — the hydride HOMO, the z and the
m* — gives rise to three new MOs in the product (recall that the number of
orbitals is conserved); Fig. 7.13 shows a sketch of the relative energies of the
MOs. The exact energies of these MOs can only be obtained from a com-
puter calculation, but we can see from the diagram that the total energy of the
electrons has been reduced.
“empty.
MO *
CO x*
COnr
C-H bond
methanal product hydride
MOs MOs AO
Fig. 7.13 Key orbital interactions between the hydride HOMO and carbonyl 7 and 2* MOs.
Curly arrows
Figure 7.14 uses curly arrows to show the important features of the changes
that take place during the reaction between hydride and methanal.
\ we H (o) 7 4
OF CN,, —_—_—_— O~ Cony
H
Fig. 7.14 The curly arrow mechanism for the reaction of methanal with hydride.
The arrows show that the hydride ion attacks the carbonyl carbon and that
a new bond forms between the carbon and the hydrogen. They also show that
the z bond breaks and electron density ends up on the oxygen.
The arrows imply that it is the pair of electrons from the 2 bond which
end up forming the negative charge on the oxygen. However, in Fig. 7.11 it
is implied that the pair of electrons from the hydride orbital end up on the
oxygen. This apparent contradiction can be resolved by noting two things.
Firstly, the arrows in Fig. 7.14 relate to a structure in which the electrons are
localized in 2c-2e bonds or as lone pairs (charges). In contrast, the MOs shown
in Fig. 7.11 are delocalized over all three atoms, and so there is not a one-to-one
correspondence between the two pictures. Secondly, it is not really possible to
say from which reactant MO a pair of electrons in a product comes from. For
example, looking at Fig. 7.13, how can we decide which pair of electrons ends
up in the middle energy MO when ail of the reactant orbitals contribute to this?
7.3 Nucleophilic addition to carbonyl 105
So there really is no contradiction between the MO picture and the curly

arrows. They are just different ways of representing the same thing — the curly
arrows relate to a strictly localized view of bonding whereas the MOs relate to
a delocalized approach. The important point is that the curly arrows give the
same result as the MO approach: in this case a new bond is formed between
the carbon and the attacking hydride nucleophile, the C-O z bond is broken,
and a negative charge ends up on the oxygen.
The geometry of attack

For the hydride to interact with the 7 and x* orbitals, it must approach from
the right direction. If it were to approach the carbonyl in the same plane as the
molecule, there will be no net interaction since the constructive and destruc-
tive contributions will cancel one another out. This is shown schematically in
Fig. 7.15 (a).
(a) (c)
constructive
~
Oa a
ax
destructive
hydride MO
1* of carbonyl
Fig. 7.15 The angle which the attacking nucleophile approaches from is important. In (a) the hydride
approaches in the same plane as the carbonyl — there are equal amounts of constructive and destruc-
tive overlap leading to no net interaction. Similarly there is no net overlap in (b) where the hydride
approaches from directly over the carbonyl group. The best angle of attack is shown in (c).
Similarly, if the hydride were to approach from above and directly in be-
tween the carbon and oxygen, as shown in (b), there would be no net overlap.
Detailed calculations and experimental evidence suggests the best angle of ap-
proach is approximately 107° from the carbonyl, as shown in Figs. 7.15 (c)
and 7.16. This maximizes the degree of constructive overlap and also mini-
mizes the repulsion between the filled 2 orbital and the incoming nucleophile.
— ep
Borohydride reduction of a ketone
Fig. 7.16 The optimum angle of attack by
We will finish this section by describing a reaction which can be carried out in a nucleophile on a carbonyl group.
the laboratory and which effectively results in the addition of H~ to a carbonyl
group. Rather than using H™ directly, we use the reagent sodium borohydride,
NaBHg4, which reacts with aldehydes and ketones to give the corresponding
alcohol. The reaction is, of course, a reduction as it involves the addition of
hydrogen.
The MOs for NaBH, are very similar to those of LiBH4 which are depicted
in Fig. 7.4 on p. 99. The highest energy occupied MO (MO-P6) is one which
is involved in the bonding between the boron and the hydrogens. When boro-
hydride acts as a nucleophile, it will be the electrons in this HOMO which are
primarily involved.
106 Reactions
‘ows
“VI
H OO oH 47 ViCHs
H3C CHs 3
Fig. 7.17 The reaction between orohydride, BH, , and propanone.
Having identified the HOMO of the system, we need to identify the LUMO
of the other reagent. For a ketone such as propanone (acetone) the LUMO is
once again the z*. As before, the interaction with the HOMO results in the z
bond being broken and the electron pair moving to the oxygen; a curly arrow
mechanism for this is shown in Fig. 7.17.
Figure 7.17 shows the electrons from the B—-H bond attacking the carbon
of the carbonyl. This results in the B—H bond breaking and a new C—H bond
@ being formed. At the same time, the C—O z bond is broken, leaving a negative
SNH OH charge on the oxygen just as before.
| _ |
Cc, Cc, The final step is to add acid which reacts with the -O~ to give an alcohol,
tT aN
Vv‘H v'H
HH Hy as shown in Fig. 7.18. The HOMO of the anion is certainly the lone pair of
Fig. 7.18 A curly arrow mechanism for
electrons on the oxygen and the LUMO is the 1s on H+. These interact to
the protonation of the anion formed from form a new bond to H* in much the same way as the reaction discussed in
the reduction of methanal. Section 7.1 on p. 97.
7.4 Nucleophilic attack on C=C

A carbon-carbon double bond has the same 7 orbitals as a carbon-oxygen
double bond so we might expect nucleophiles to attack C=C in the same way
they attack C=O. However, this is not the case — nucleophiles do nor usually
attack alkenes. In fact, alkenes often act as nucleophiles themselves, preferring
to donate electrons rather than accept them.
We can understand these differences by considering the nature of the C=C
bond and the orbitals involved. Firstly, unlike in the carbonyl, the C=C bond
does not have a dipole moment so there is no electrostatic attraction to bring
the alkene and nucleophile together. The next main point to consider is the
energies of the orbitals involved; these are shown in Fig. 7.19.
The z* MO of the C=C bond is higher in energy than the 2r* of the C=O
Fig. 7.19 Due to the energy differences between the atomic orbitals of carbon and
oxygen, the C=C x
and z* MOs are higher in energy than the corresponding orbitals of C=O.
bond. This means that there is a poorer energy match between the C=C x*
and the HOMO of any attacking nucleophile. As we have seen, the poorer the
energy match, the weaker the interaction will be.
The final point is that if a nucleophile did attack a C=C, a new bond will
form between the carbon and the nucleophile and the carbon-carbon double
bond will break leaving the charge on a carbon, as shown in Fig. 7.20. This is
not as favourable as the attack on the carbony! where the negative charge ends
up on the oxygen atom, the reason being that the oxygen orbitals are lower in
energy — put another way the charge is best located on an electronegative atom.
Taking these factors together provides a rationale for why nucleophilic attack
on C=C does not occur whereas attack on C=O occurs readily.
Nu° Nu
\ S)
~oC=Cm CC
U aN
Fig. 7.20 If a nucleophile did attack a C=C » bond, a negative charge would form on the other carbon
atom. Unlike having a negative charge on oxygen, this is not favourable.
7.5 Nucleophilic substitution

We are now going to look at a reaction in which the nucleophile donates elec-
trons not into an empty AO or a z* MO but into a o* MO. This will lead to
the breaking of the o bond, so an atom or group will break away from the
molecule. The reaction is therefore called nucleophilic substitution as opposed
to the addition reaction we have seen so far.
A typical example is where an incoming nucleophile, Nu~, bonds to the
carbon of chloromethane, displacing Cl” in the process:
Nu” + CH3Cl —> CH3Nu4+ Cl".
In chloromethane the only unoccupied MOs will be o* MOs, of which there

are two different types: the C-H and the C-Cl a”. The nucleophile HOMO will
interact most effectively with whichever of these is lower in energy; as we will
see in the next paragraph, it turns out that this lowest energy MO is the C-Cl
o*.
The energy of the o* MO will depend on the strength of the interaction
between the two AOs involved, and this depends primarily on the energy sep-
aration of these orbitals. As the chlorine AOs are lower in energy than those
of the hydrogen, there will be a better match between the AOs involved in the
C-H interaction than in the C-Cl interaction. As a result the o* MO for the
C-Cl bond will be lower in energy than the corresponding orbital for CH; this
is illustrated in Fig. 7.21.
In larger halides such as chloroethane, there are also C-C interactions.
However, these C-C interactions also involve well-matched AOs, so again we
expect that the resulting o* MO will be higher in energy than for C-Cl. In fact,
of the 0* MOs in such molecules, it will always be the case that the MO asso-
ciated with a bond to an electronegative element (such as Cl, N or O) will be
108 Reactions
(a) (b)
Fig. 7.21 Due to the larger difference in energy between the carbon and chlorine AOs than between
the carbon and hydrogen AOs, the C-Cl o* will be lower in energy than the C-H o*.
lower in energy than the 0* MOs from C-C or C-H bonds. This is why much
organic chemistry often centres around the functional groups which contain
atoms other than carbon and hydrogen, leaving the basic C-C / C-H frame-
work intact. The anti-bonding MOs involving the heteroatoms are inevitably
lower than those in the basic framework and hence more readily involved in
reactions.
A contour plot and surface plot of the C-Cl o* MO (the LUMO) from
chloromethane is shown in Fig. 7.22 together with a schematic representation.
1C Cl
y/] \ y
4H
“ig V4
Fig. 7.22 Three representations of the LUMO (the C-Cl o* MO) in chloromethane: a contour plot, a
surface plot and a sketch. Note the node between the C and Cl, and the lobe on C which points away
from the bond.
The reaction between hydroxide ion and bromomethane

We shall look at the reaction between a halogenoalkane, in this case bro-
momethane and hydroxide ion. This is a simple substitution reaction where
the hydroxide takes the place of the bromide ion: .
HO™ + CH3Br —> CH30H+ Br’.

Having identified the LUMO as the C-Br o* MO, we need to identify the
HOMO that this will interact with. Not surprisingly, this is the orbital occu-
pied by one of the lone pairs on the oxygen of the hydroxide ion. The overall
negative charge raises the energy of this orbital further. For simplicity we will
assume that the lone pair occupies an sp? orbital on the oxygen — the details
are not really important for this discussion.
The angle of approach

Bromomethane has a dipole moment with the carbon atom slightly positively A 8+ 8
charged and the bromine atom slightly negatively charged (Fig. 7.23). The neg- H! y C—Br
atively charged hydroxide ion is therefore attracted towards the carbon. The H oe
best interaction of the HOMO of the hydroxide and the LUMO of the bro-
momethane is achieved if the nucleophile attacks the carbon from the opposite Fig. 7.23 Due to the differences in the
electronegativities of carbon and
side to the bromine. This attack is shown enema in ne 7.24, bromine, bromomethane has a dipole
H H moment with the carbon atom being
O@ - @:000
e\
—- @6- 0
slightly positive and the bromine atom
slightly negative.
aM
HOMO of LUMO of new o Mo" HOMO of
hydroxide bromomethane between C and O bromide
Fig. 7.24 The hydroxide ion attacks the bromomethane from directly behind the C—Br bond since this
maximizes the interaction with the LUMO (the C—Br o*) of the bromomethane.
If the nucleophile were to attack the carbon atom from the side, the orbital
overlap would be poor due to the symmetry mismatch (Fig. 7.25).
During the reaction, the HOMO of the hydroxide interacts with the LUMO HOMO of C)
of the bromomethane, the C-Br o*. A new o bond forms between the car- hydroxide So-—
bon and oxygen at the same time as the old C—Br o bond breaks. Eventually
the bromide ion leaves, carrying with it the negative charge. The curly arrow H Of
mechanism for the reaction is shown in Fig. 7.26.
4 ic}
@:00
dG
Cc Bri
H LUMO of
| / bromomethane
See — HO--C--Br| -——» HO-—C,, BP
‘4 =b ¥H Fig. 7.25 If the hydroxide ion attacked the
HH H carbon more from the side, the orbital
overlap is much worse due to the
Fig. 7.26 A curly arrow mechanism for the attack of hydroxide ion on bromomethane. The species in constructive and destructive interactions
the square brackets is a transition state (see Section 9.2 on p.133) which is not an isolable compound with the o* orbital.
but just an arrangement of atoms through which the system passes during the reaction. Note that the
carbon centre has been inverted during the reaction.
The inversion of the carbon centre

During the course of the reaction, the carbon centre undergoes an inversion —
it is basically turned inside-out with the molecule passing through an arrange-
ment in which all three hydrogen atoms lie in a plane perpendicular to the
C-Br bond. This is similar to the inversion of ammonia described in Fig. 6.13
on p. 87.
For the reaction of bromomethane, there is no way that we could tell that the
carbon atom has undergone an inversion. If, however, instead of three hydrogen
atoms, the carbon has three different groups around it, then the inversion can
be detected. This was demonstrated very neatly in a variation on the reaction
involving radioactive iodide attacking 2-iodooctane as shown in Fig. 7.27.
The non-radioactive starting material 2-iodooctane and the product are
not the same even when the radioactive atom is ignored. As can be seen in
Fig. 7.28, they are in fact non-superimposable mirror images of each other.
110 Reactions
io}
H ———_» | r--c--1
i | ———> Ir-C,,
ae i
I £4
Hac = \/CHs
CeHi3 H3C CeHi3 CeHi3
Fig. 7.27 The inversion at the carbon centre can be detected in this experiment involving the substitu-
tion of the iodine atom in 2-iodooctane by a radioagtive iodide ion, denoted with an asterisk.
Compounds which are non-superimposable mirror images are called enan-

tiomers or optical isomers. The term optical isomers arises from the fact that
H H whilst (R)-2-iodooctane and (5)-2-iodooctane behave the same chemically, one
\ /
wo! I*—C,,, isomer will rotate plane-polarized light clockwise, while the other isomer will
H3C'4 ¥°CH3 rotate it anti-clockwise.
CeHi3 CeHy3
In the experiment, the rate of substitution in the reaction was followed
(S)-iodooctane (F)-iodooctane
by monitoring the incorporation of radioactive iodine into the 2-iodooctane;
Fig. 7.28 The starting material and the rate of inversion was followed by measuring the degree to which plane-
product are non-superimposable mirror
polarized light is rotated. The results clearly showed that each time a substitu-
images of each other (ignoring the
different isotopes of iodine). These tion occurred, the molecule underwent an inversion, which is what we would
isomers are differentiated with the labels predict for this mechanism.
Sand A.
7.6 Nucleophilic attack on acyl chlorides

So far we have described reactions in which the HOMO of the nucleophile
interacts with a LUMO which is either a 7* MO (for example, in ketones and
aldehydes) or ao* MO (for example, in an alky] halide). What happens if the
molecule which the nucleophile attacks has both 2* and o* MOs available?
A simple example is the reaction between hydroxide ion and ethanoyl
(acetyl) chloride, whose structure is shown in Fig. 7.29. The reactive carbon
centre has both a double bond to oxygen and a bond to chlorine. With both
these electronegative elements withdrawing electrons, the carbon has a partial
positive charge but which orbital will a nucleophile such as hydroxide attack
Fig. 7.29 Ethanoy! chloride. into? The hydroxide could either attack into the 2* of the C=O giving a tetra-
hedral species as shown in Fig. 7.30 (a) or it could attack into the C-Cl o* to
form ethanoic (acetic) acid as shown in Fig. 7.30 (b).
The final product from the reaction of ethanoy] chloride with aqueous hy-
droxide is in fact the carboxylic acid so it is tempting to conclude that mech-
anism (b) is correct and that the hydroxide attacks into the C-Cl o* MO. It
HO C —_— C is tempting but sadly incorrect! In order to understand what does happen, we
Hsc~ Cl H3C Cl need to consider the orbitals involved.
A computer calculation reveals that the LUMO for ethanoy! chloride is the
(b)
LN O , HO. 20 * MO with the C-Cl o* MO lying somewhat higher in energy. These orbitals
are shown in Fig. 7.31. The initial attack by the hydroxide nucleophile is into
the C~O m1* MO as shown in Fig. 7.30 (a). There are two reasons why the
nucleophile prefers to attack this MO rather than the C-Cl o*.
Fig. 7.30 Two possible routes of attack
on ethanoy! chloride. In (a) the hydroxide
Firstly, there is a better energy match between the HOMO of the hydroxide
attacks into the C—O 2* MO and in (b) it and the C—O z* MO. This is because the C—Cl o* is higher in energy than the
attacks into the C—Cl a* MO. C-O x”.
7.6 Nucleophilic attack on acyl chlorides 111
dor Oe
Fig. 7.31 Surface plots and cartoons of two MOs from ethanoyl chloride: (a) is the LUMO and (b) is
the next highest energy MO. The LUMO is essentially just the C-O 1* MO and the orbital shown in (b)
is essentially the C-Cl o* MO with some minor contributions from the other atoms.
Secondly, it is easier for the nucleophile to approach the 3* MO in the

correct orientation. As we saw in Fig. 7.16 on p. 105, the best direction of
approach for the nucleophile to attack the 7* MO is from above the plane of
the C=O bond. This approach is unhindered in the case of hydroxide attacking
ethanoy] chloride.
In contrast, when a nucleophile attacks a o* MO it must approach directly
from behind, as shown in Fig. 7.24 on p. 109. In the case of hydroxide attack-
ing ethanoyl chloride this approach is hindered by the methyl group and the
oxygen, both of which are in the same plane as the 0* MO.
Overall, then, attack on the 7* MO is the preferred route. However, as
shown in Fig. 7.30 (a), this leads to what is called a tetrahedral intermediate.
What we now need to work out is how this intermediate reacts further to give
what we know to be the product, ethanoic acid, as shown in Fig. 7.32.
(er OH O
3 %,,,_/ I co)
ovo Cl Cc ---- > C + Cl
Hsc” cI —™~
Fig. 7.32 The initial reaction between the hydroxide and ethanoyl chloride produces a tetrahedral
intermediate. This then goes on to form the final products, ethanoic acid and a chloride ion.
The tetrahedral intermediate has no z bond and therefore no 2* MO either,

so the C-Cl o* MO (which has remained almost unchanged after the attack
of the hydroxide) is now the LUMO of the system. The lone pair of electrons
on the negative oxygen is the HOMO in this molecule and, as we commented
before, the negative charge raises the energy of this orbital. The interacting
HOMO and LUMO are in this case in the same molecule; these orbitals in-
teract to form some new orbitals in the product as shown in Fig. 7.33. The
overall process is rather more complex than this figure suggests, as more or-
bitals are involved. A curly arrow mechanism for the whole reaction is shown
in Fig. 7.34.
112 Reactions
/ , 6 ~,
OH OH
Fig. 7.33 The HOMO of the intermediate shown in (a) is essentially just a lone pair on the oxygen. This
orbital interacts with the LUMO, the C-Cl o* shown in (b), to contribute to the MOs of the products.
Diagram (c) shows two MOs in the products: the new C-O bond and the p orbital on the chloride
ion.
(“oH OH cI oO,
cl
g) %,,,,-/ rotate
~=Cc—Ccl —— Cc Oo "ny — o=— CH 3
7 H,c” ~ —= CH "OH
Fig. 7.34 A curly arrow mechanism for the reaction between hydroxide and ethanoyl chloride. The
intermediate has been rotated after the initial attack to show more clearly how the final products are
formed.
This attack on the C-O x” by a nucleophile to give a tetrahedral interme-

diate and then the subsequent reformation of the C—O z bond is a common
reaction which you will meet many other examples of in later chapters. One
example is the reaction between a carboxylic acid, such as ethanoic acid, and
an alcohol, such as methanol, to give an ester as shown in Fig. 7.35.
fe)
iC + CH3OH
I
Cc CH3z + H20
H3C~ OH H3c~ ~O~
Fig. 7.35 The reversible formation of methyl ethanoate (an ester) and water from ethanoic acid and
methanol.
We can also carry out this reaction in reverse, i.e. by adding water to the
ester and allowing the hydrolysis reaction to produce the carboxylic acid and
alcohol. In other words, this reaction is reversible.
Exactly why some reactions seem to be able to go in both directions and
how chemists can try to alter the conditions to favour either the products or the
reactants is the subject of the next chapter.
8 Equilibrium
In Chapter 2 we saw that whether or not a process will ‘go’ is controlled by

the Second Law of Thermodynamics. What this law says is that a process
will only be spontaneous — that is, take place on its own without continuous
intervention from us — if it is associated with an increase in the entropy of the
Universe (Section 2.2 on p. 7). We also saw that another and entirely equivalent
way of applying the Second Law is to express it in terms of the Gibbs energy
(Section 2.7 on p. 16); using this, a spontaneous process is one in which the
Gibbs energy decreases.
Using either of these criteria we were able to explain why it is that water
will freeze at —5 °C but not at +5 °C, since it is only at the lower temperature
that water —> ice is accompanied by an increase in the entropy of the Universe
or, equivalently, a decrease in the Gibbs energy.
Applying the same ideas to chemical reactions is rather more subtle, as we
need to explain what determines the position of equilibrium, an idea introduced
on p. 5 and illustrated in Fig. 2.1. Some reactions go almost entirely to prod-
ucts so the position of equilibrium lies very much towards the products; others
hardly go at all and so the position of equilibrium lies towards the reactants.
Other reactions come to equilibrium with significant amounts of reactants and
products present. Our task in this chapter will be to work out what determines
this position of equilibrium.
8.1 The approach to equilibrium

On p. 6 we discussed the equilibrium between NO2 and N20Oq:
2NO2(g) = N204(g) >

5 | equilibrium
elt
and pointed out that the reaction could come to the equilibrium position (at Cc
o
which significant amounts of both reactants and products are present) either 2
starting from pure reactants (NO2) or from pure products (N2O4). What this 2
oO
implies is that in going from either pure products or pure reactants to the equi- ——_—_>
librium position the Gibbs energy must decrease. In other words, the equilib- composition
rium mixture of reactants and products has lower Gibbs energy than either pure 2NO, NoO,
products or pure reactants. Fig. 8.1 Plot of the Gibbs energy of the
Shown in Fig. 8.1 is a plot of the Gibbs energy of the mixture of reactants reaction mixture as a function of its
composition for the dimerization of NOo.
and products (here NO2 and N2Og) as a function of the composition. Pure The left-hand side of the graph
reactants correspond to no reaction having taken place and so appear on the corresponds to pure reactants and the
left; pure products correspond to complete reaction and so appear on the right. right to pure products. The minimum in
the Gibbs energy corresponds to the
From the graph we see that the Gibbs energy goes to a minimum at some position of equilibrium which can be
point intermediate between pure reactants and pure products; this minimum approached either from pure reactants or
pure products.
corresponds to the equilibrium position.
114 Equilibrium
No matter whether we start with pure products, or pure reactants, or some

mixture of the two, the Gibbs energy falls as we approach equilibrium. For
example, in Fig. 8.2 if we start at composition a, moving to a’ is accompa-
55 equilibrium nied by a decrease in the Gibbs energy. The change from a to a’ is therefore
a spontaneous process which moves us towards the position of equilibrium by
5
Cc
wn
|y increasing the amount of products. Starting at b and moving to b’ is also ac-
2 :
2 i I
companied by a decrease in the Gibbs*enérgy and so will be spontaneous; as
Oa a’ b'b
L ——_—— 2 for the change from a to a’, the change from b to b’ also moves us towards the
Fo composition g equilibrium position but this time by decreasing the amount of products.
5 ae]
8 5 The shape of the graph therefore funnels the composition towards the value
Fig. 8.2 Starting from point a the Gibbs with the minimum Gibbs energy, which is the equilibrium position. From this
energy will fall as we move to point a’; point any change would involve an increase in the Gibbs energy, which is not
such a change will be spontaneous and permitted. So, once at its equilibrium value the composition cannot change.
will take us toward the position of
equilibrium by increasing the amount of Our task is to try to understand why the plot of Gibbs energy as a function
products. In contrast, from point b the of composition has the form shown in Fig. 8.2 and also to identify the factors
direction which involves a reduction in
which influence the position of the minimum.
Gibbs energy is towards point b’; for this
spontaneous change the amount of
products is reduced. Starting from either
aor b the direction of spontaneous 8.2 The equilibrium between two species
change is towards the equilibrium
position, at which the Gibbs energy is a The simplest kind of equilibrium we can consider is when there is just one
minimum.
reactant (A) and one product (B):
A = B.
The equilibrium between isomers (illustrated in Fig. 8.3) is an example of this
kind of reaction.
Let us imagine that A and B are both gases and that we start out with one
mole of pure A sealed in a container such that the pressure is 1 bar (1 bar is
H 10° N m~?, very close to 1 atmosphere pressure). The reaction will come to
2 H
/\ — voy equilibrium by some of the A converting to B, so we can specify the extent to
H2C — CHs Hs“ “CHa which this has happened simply by quoting the percentage of B in the mixture.
cyclopropane propene As the reaction involves no change in the number of moles, the pressure does
not change as A and B interconvert.
F Figure 8.4 shows how the Gibbs energy varies with percentage of B. On
y N= N® —=
— \ N=N
=n" ° the left we have pure A, and so, as there is one mole of A present, the Gibbs
F energy is the molar Gibbs energy of pure A, Gm(A). On the right we have pure
Fig. 8.3 Examples of equilibria involving B, and so the Gibbs energy is just the molar Gibbs energy of pure B, Gm(B).
just two chemical species which are To draw the graph we have arbitrarily chosen G,,(B) to be less than G»,(A).
isomers of one another. The task now is to try to explain why the curve has this shape, and in particular
why it shows a minimum.
Suppose we start with pure A and then allow a very small amount of A to
convert to B; as a consequence there will be a small change in the enthalpy and
in the entropy. We will concentrate on the entropy change which we can think
of as being due to two contributions.
The first contribution comes from the fact that we have converted some A
to B. Given that the two substances are likely to have different molar entropies,
changing some A into B is surely going to result in a change in entropy.
The second contribution comes from the mixing of this small amount of B
into the bulk of A. So far we have not discussed this entropy of mixing, but you
——»
Gibbs energy
equilibrium
0 ———_> 100
t % of B t
pure A pure B
Fig. 8.4 Plot showing an example of how the Gibbs energy for the equilibrium between A and B varies
with the composition, which in this case can be specified by the percentage of B. Arbitrarily, we have
made the molar Gibbs energy of B lower than that of A.
can see that going from pure A to having a little B mixed in with A certainly
gives rise to an increase in the entropy on the grounds that the mixture is more
‘random’ than pure A. We therefore expect this second contribution always to
result in an increase in the entropy.
The first contribution might result in an increase or decrease in the entropy,
but we argue that the entropy increase due to the second contribution — the
one due to mixing — will always be dominant as there is a large increase in
randomness (that is, in entropy) on going from pure A to a mixture with a
small amount of B in it.
Given that AG is defined as (AH — TAS) (p. 16), it follows that this
increase in entropy will lead to a decrease in the Gibbs energy. Therefore, as
we start from the left-hand side of the graph in Fig. 8.4 we expect the Gibbs
energy to fall. Similarly, when starting from the right-hand side of the plot we
can employ the same reasoning to argue that having a small amount of A mixed
in with the bulk of B leads to an increase in the entropy and hence a decrease
in the Gibbs energy.
Having now convinced ourselves why the Gibbs energy falls when we start
from either side of this plot, we now need to work out what happens in between.
A simple argument is to note that as the Gibbs energy falls when we start from
the left or right, for the curve to join up there must be a minimum somewhere
between these two points.
To determine the exact form of the curve we need some more detailed ther-
modynamics, which we do not have time to go into here — so for now you will
simply have to take the form of these graphs on trust.
Locating the position of equilibrium .

The remarkable thing about the plot in Fig. 8.4 is that it turns out that the lo-
cation of the minimum — that is the percentage of B present at equilibrium —
depends only on the difference between the molar Gibbs energies of A and B.
116 Equilibrium
>
>
@
c
@
wn
2
2
Oo
0 ——_ 100 0 > 100
% of B % of B
(c) (d)
+— G,,(A)
& pp
oCc G,,(B)
oO
en)
2
2
OO —_——_S too ~2OsO ——_——_S T00
% of B % of B
Fig. 8.5 Illustration of how the position of equilibrium (indicated by the open arrow) is affected by
changing the relative molar Gibbs energies of A and B. In (a) Gm(A) and Gm(B) are equal, making the
curve symmetrical and so the position of equilibrium lies at 50% B. In (b) Gm(A) is less than Gm(B) and
so the position of equilibrium lies towards A; in (c) Gm(A) has been reduced further and the equilibrium
therefore lies even further towards A. Finally, in (d), Gm(B) is less than Gm(A) and so the position of
equilibrium lies towards B.
To prove that this is so needs a deeper study of the principles of thermodynam-

ics than we have time to go into here, so we will simply have to take this result
on trust.
Figure 8.5 illustrates how the position of the minimum shifts for different
values of Gm(A) and Gm(B). In (a) the molar Gibbs energies of A and B are
equal; the minimum in the Gibbs energy of the mixture is at 50% B. In (b)
the molar Gibbs energy of A is lower than that of B, and so the position of
equilibrium moves to the left, corresponding to the presence of more A than B;
in (c) the molar Gibbs energy of A is lower still, and this shifts the position of
equilibrium even further towards A. Finally, in (d) the molar Gibbs energy of B
is lower than that of A so now the equilibrium lies towards B. We see that the
position of equilibrium lies towards the species with the lowest molar Gibbs
energy.
If you take a piece of string or chain and hold one end in your left hand
and the other in your right, the string falls in a curve rather similar to those in
Fig. 8.5. The height of your left hand represents the molar Gibbs energy of
A, and the height of your right hand represents the molar Gibbs energy of B.
If you hold your two hands level, your will see that the string hangs so as to
make a minimum at a position midway between your hands: this is like plot
(a). If you lower your left hand, the minimum moves towards the left, just as
in plots (b) and (c). Conversely, if you lower your right hand, the minimum
moves towards the right, as in plot (d).
As you know, when a reaction has come to equilibrium, the concentrations
of products and reactants are related by the value of the equilibrium constant.
In the case of the equilibrium between A and B, the equilibrium constant, K, is

given by
_ (Blea
[Aleq
where [A]eq and [B]eq are the equilibrium concentrations of A and B respec-
tively. If at equilibrium more B is present than A, K is greater than 1 and the
equilibrium lies towards the products. Conversely, if the equilibrium lies to-
wards the reactants there will be more A present than B and the equilibrium
constant will be less than 1.
The percentage of B at which the minimum Gibbs energy occurs deter-
mines the value of the equilibrium constant. We have already described how
the location of the minimum is determined solely by the difference in the molar
Gibbs energies of A and B. It therefore follows that the value of the equilibrium
constant must be related to this difference, and some more detailed thermody-
namics shows us that:
Gm(B) — Gm(A) = —RT Ink (8.1)
where R is the gas constant (8.3145 J K—! mol~!) and T is the absolute tem-
perature (in K).
This equation has the correct form, as if Gm(B) is less than Gm(A), the
left-hand side will be negative and so In K will be positive. This means that K
is greater than 1 so that at equilibrium the product B is favoured, as depicted in
Fig. 8.5 (d).
In contrast, if Gm(A) is less than G»(B), the left-hand side of Eq. 8.1
will be positive and so InK will be negative. This means that K will be a
fraction between 0 and 1, favouring the reactant A at equilibrium, as depicted
in Fig. 8.5 (b) and (c).
The surprising thing about Eq. 8.1 is that it tells us that the equilibrium
constant depends only on the difference between the molar Gibbs energies of
Gibbs energy
pure A and pure B. This really is quite a remarkable result and worth discussing
a little more before we move on.
Interpretation 0 —_———> 100

Suppose that Gm(B) is less than Gm(A). This means that if we start with pure
% of B
A and convert all of it to B the process would be accompanied by a decrease Fig. 8.6 The molar Gibbs energy of the
product B is lower than that of the
in Gibbs energy and therefore be spontaneous. This change in Gibbs energy is reactant A, so going frorn pure A to pure
shown as AG, in Fig. 8.6. B involves a decrease in the Gibbs
However, what we have seen is that there are some ratios of A to B for energy (AG). However, for certain ratios
of A to B the Gibbs energy of the mixture
which the mixture has lower Gibbs energy than either pure A or pure B; one is /ower than that of pure B; the lowest
of these mixtures has the minimum Gibbs energy and this is the one which value of the Gibbs energy corresponds to
corresponds to equilibrium. As is shown in Fig. 8.6, going from pure A to the the equilibrium point and moving to this
point from pure A involves the largest
equilibrium mixture involves a larger decrease in the Gibbs energy, AG2, than decrease in the Gibbs energy (AG»).
going to pure B.
118 Equilibrium
8.3 General chemical equilibrium

‘These ideas about the simple A = B equilibrium are just a special case of a
more general result which is that for any reaction the equilibrium constant, K,
is given by
A,G° = —RT Ink. (8.2)
A,G° is the standard Gibbs energy change which can be computed from
A,G° = A,H° —_ TA,S° (8.3)
where A,H° is the standard enthalpy change and A,S° is the standard entropy
change for the reaction. The first thing we need to do is describe what we mean
by a ‘standard change’.
Standard states and standard changes

We first need to introduce the concept of a standard state; this is a particular
state of the substance in question, defined according to the following agreed
convention:
substance standard state

gases the pure gas at a pressure of 1 bar (10° N m7?)
solids the pure solid
liquids the pure liquid
solutions — the solution at unit concentration
The standard state given in the table for a solution only applies to an ideal
solution in which there are no solute—solvent interactions; for real solutions in
which there are such interactions the definition is more complex and we shall
not go into this here.
The standard state is denoted by a superscript ° or sometimes a Plimsoll line
e. It is sometimes thought, erroneously, that the standard state implies a certain
temperature: this is not the case. In fact, when quoting values for standard en-
thalpy, entropy or Gibbs energy changes we must always state the temperature
at which this value applies as these quantities are temperature dependent.
Having introduced the standard state we can now define standard changes
for reactions. Such changes refer to a particular balanced chemical equation.
For example, A,G° for the reaction
Ho(g) + O2(g) —> H20(1) (8.4)

is the change in Gibbs energy when one mole of H2(g) reacts with half a mole
of O2(g) to give one mole of H2O()), all of the species being present in their
standard states and at the stated temperature.
It is very important to understand that the value of A,G° refers to complete
reaction, i.e. the hydrogen and oxygen must be converted completely to wa-
ter. Note too how the values of the stoichiometric coefficients come into the
definition of A,;G°.
8.3. General chemical equilibrium 119
So, if we doubled the stoichiometric coefficients in Eq. 8.4 and wrote it as
2H2(g) + O2(g) —> 2H20(/)

A;,G® would be the change in Gibbs energy when two moles of H2(g) react
with one mole of O2(g) to give two moles of H2O(/). The numerical value of
A,G° for this reaction would be twice the value of that for Eq. 8.4.
A second example is the formation of ammonia:
*N2(g) + 2H2(g) —> NH3(g)

for which A;G° is the change in Gibbs energy when half a mole of No(g) reacts
with 4 moles of H2(g) to give one mole of NH3(g), all in their standard states
and at the stated temperature.
In many ways A,G° is a hypothetical quantity as it refers to the change in
Gibbs energy when the reactants, in their standard states, are converted com-
pletely to products, also in their standard states. If we actually mixed the re-
actants together in their standard states there is no guarantee that the reaction
would go completely to products in their standard states. All that we can be
sure will actually happen is that the reaction will go to its equilibrium position,
which often will not involve complete conversion to products.
For example, at standard pressure and at room temperature not much am-
monia will be formed by mixing Nz and H2. However, this does not stop us
from imagining what the change in Gibbs energy would be if the reaction went
completely from N2 and H2 to NH3.
The standard enthalpy change is defined in the same way as A,G°: for the
stated reaction it is the change in enthalpy when the reaction proceeds com-
pletely from reactants to products, all species being in their standard states.
Similarly, the standard entropy change is the change in entropy for this pro-
cess.
Returning to the simple equilibrium between A and B which was discussed
in Section 8.2 on p. 114, we can see that because we took the pressure to be
1 bar, the difference in the molar Gibbs energies of A and B is in fact the same
thing as A,G° for the reaction; this is shown in Fig. 8.4 on p. 115. So, Eq. 8.1
is just a special case of Eq. 8.2.
Determining standard changes

The usual way of finding A,H° is to use tabulated values of standard en-
thalpies of formation, A¢H°. These are defined in such a way that for any
balanced chemical equation we can find A,H° by adding together the ArH
values of the products and subtracting those of the reactants, taking into ac-
count the stoichiometric coefficients. For example, to find AgH°® for the reac-
tion
SO2(g) + 202(g) — $O3(g)
we can imagine a Hess’ Law cycle in which the reactants are broken down
into their elements (in their standard states) and then these are re-formed to
products:
120 Equilibrium
SOr(g) + 202g) “S SOs(g)

—arH*(802)| ~earH*(02)| Jareecsos)
| elements in their standard states |

hy
A-H° is thus given by
ArH° = ArH°(SO3(g)) — ArH°(SO2(g)) — ArH? (O2(g)).
As you will recall, the standard enthalpy of formation of an element is zero, so
in the above equation ArH °(O2(g)) = 0.
Similarly, A,S° values are found by adding together the standard (absolute)
entropies (S°) of the products and subtracting those of the reactants, taking into
account the stoichiometric coefficients. So, for the above reaction:
A,S° = S°(SO3(g)) — S°(SO2(g)) — %S°(O2(g)).
As with enthalpies of formation, extensive tabulations of standard entropies
are available. Note that the standard absolute entropies of the elements are not
zero.
Having found A,H° and A,S° we then find A,G° using Eq. 8.3
A,;G° = A,H° —_ TA,S°.
An example
Let us use this approach to find A,G° for
SO2(g) + 02(g) —> SO3(g).
From tables we find (all at 298 K and for the gaseous state) AsH°(SO3) =
—396 kJ mol! and ArtH°(SO2) = —297 kJ mol™!; so, remembering that
ArH? for Oz is zero,
A,H® = Af¢H°(SO3(g)) — AgH°(SO2(g)) — AsH°(O2(g))

= —396 — (-297) —')» x 0
= —99 kJ mol~!.
The same tables give us the following standard entropies: S°(SO3) = 257
JK7! mol™!, S°(SOz) = 248 J K~! mol! and S°(O2) = 205 J K~! mol-!:
so
A,S° = S°(SO3(g)) — S°(SO2(g)) — 45°(02(g))
= 257 — 248 — 'y x 205
= —93.5JK7! mot!.
These values allow us to find A,G° at 298 K:
A,G° = A,H° —TA,S°

= —99 x 10° — 298 x (—93.5)
= —71 x 10° J mol™!.
Next we will see what this implies for the value of the equilibrium constant.
8.3 General chemical equilibrium 121
Equilibrium constants
The equilibrium constant and the standard Gibbs energy change for a reaction
are related according to Eq. 8.2
A,G° = ~RTInK
which can be rewritten to give K as
A,G° “reactants -
K = exp RT K=0 favour
negative 0 positive
This relationship tells us that if A;G° is negative, the exponent (the expression A,G°——>
in the brackets) will be positive and so the equilibrium constant will be greater
Fig. 8.7 Graph showing how the sign of
than 1, i.e. the products are favoured. On the other hand, if A,G° is positive, ArG® affects the position of equilibrium.
the exponent will be negative, giving an equilibrium constant of less than 1, If ArG° is positive the equilibrium
which means that the reactants are favoured. These points are illustrated in constant is less then 1 (but still positive)
which means that the reactants are
Fig. 8.7. favoured. If ArG®° is negative the
In the case of the formation of SO3 from SO, and O 2 we found that at equilibrium constant is greater than 1 and
298 K, AyG° = —71 x 10° J mol! and so the products are favoured.
—A,G°
K = exp(
RT
—(—71 x 103)
= exp
8.3145 x 298
= 2.29 x 10!2.
The equilibrium constant is very large, implying that the equilibrium will be
totally in favour of the products in this reaction.
Before we move on there is one point of possible confusion we need to clear
up. In Chapter 2 we showed that a spontaneous process must be accompanied
by a decrease in the Gibbs energy, and we have used this criterion to discuss
chemical equilibrium. Consider the situation shown in Fig. 8.8, in which A,G°
Gibbs energy
is positive, i.e. the pure product, B, is higher in Gibbs energy than the pure
reactant, A. It therefore follows that the conversion of pure A to pure B (at
standard pressure) will not take place as this would be accompanied by an
increase in the Gibbs energy.
100
However, this does not mean that the conversion of some A to B is forbid-
i=)
% of B
den — far from it. We see this clearly in Fig. 8.8 where the conversion of pure A Fig. 8.8 Illustration of how a reaction
to the equilibrium mixture is accompanied by a decrease in the Gibbs energy, which has a positive A;G°® still takes
even though A,G° is positive. place to a certain extent, i.e. comes to a
position of equilibrium in which some
So, if a reaction has a positive A,G° this does not mean that the reaction product is present. Going from pure A to
will not take place; rather, it means that the reaction will come to an equilibrium pure B would involve an increase in the
position which favours the reactants. Similarly, a reaction which has a negative Gibbs energy and so is not allowed.
Nevertheless, going from pure A to the
A,G° will not go entirely to products, but it will go to an equilibrium position equilibrium mixture involves a decrease
which favours the products. in the Gibbs energy and so is allowed. A
In fact, because of the exponential relationship between A,G° and the equi- reaction with a positive A,G° is not
‘forbidden’ but simply comes to a position
librium constant, once A,G° becomes more positive than a certain value the of equilibrium which favours the
equilibrium constant becomes so small that the equilibrium lies entirely to the reactants.
122 Equilibrium
we 8 10°97
wn
products} 25 | 10°.
6
favoured| “
3
10 AG? / kJ mor"
—_—____———
t T T T , ! Ay T T T t 1
50 -40 -30 -20 -10 ) 20 30 8640 50

10°35
6 reactants
10" - favoured
10°9 J IN
Fig. 8.9 Plot showing how the equilibrium constant, K, varies with the standard Gibbs energy change
of the reaction (ArG°) at a temperature of 298 K; note that the vertical scale is logarithmic. If ArG°
is more negative than about —40 kJ mol~! the equilibrium constant is so large that essentially the
conversion to products is complete. Similarly, if A;G° is more positive that about +40 kJ mol—! the
equilibrium constant is so small that essentially no products are formed.
reactants, i.e. the reaction does not proceed to.a significant extent. Similarly,
once A,G° is more negative than a certain value, the equilibrium constant be-
comes so large that to all intents and purposes the equilibrium lies entirely in
favour of the products.
Figure 8.9 shows how K varies with A,G°; note that the vertical scale
is logarithmic. We see from this plot that once A,G° exceeds around
+40 kJ mol~! the equilibrium constant is so small that essentially none of the
reactants have become products. On the other end of the scale, if A,;G° is more
negative than about —40 kJ mol~! the equilibrium constant is so large that to
all practical intents and purposes the reaction has gone entirely to products. Fi-
nally, note that if A;G° = 0 the equilibrium constant is one, implying an equal
balance between products and reactants.
8.4 Influencing the position of equilibrium

For a given reaction the value of the equilibrium constant is fixed once we
specify the temperature and the states of the reactants (solid, liquid, gas, etc.).
This section describes a number of ways in which the actual amount of products
can be increased — in other words, how to increase the yield of the reaction,
something we often want to do for practical reasons.
Temperature
Both the equation relating A,;G° to the equilibrium constant
A,G° = —RTInK (8.2)
and the definition of A,G°
A,G° = A,H° — TA,S° (8.3)

involve temperature explicitly, so we can expect the temperature to have an

influence on the value of the equilibrium constant. It turns out that the values of
A,S° and A,;H° also depend on temperature, although not very strongly. Over
a modest temperature range it is reasonable to assume that they are constant,
which is what we will do from now on.
As both Eq. 8.2 and Eq. 8.3 are expressions for A,G° we can equate their
right-hand sides to give
—RTInK = A,H° —TA,S°.
Dividing both sides by —RT gives
InK =- ArH? a(1\ A,S° 8.5

" R (;) TR (8.9)
This equation tells us that the way in which the equilibrium constant varies
with temperature depends on A,H°.
If the reaction is endothermic (A,H° is positive) the term (—A,H°/RT) is
negative and so increasing the temperature makes it Jess negative. This means
that as T increases both In K and K increase, as is illustrated in Fig. 8.10. In
In kK —~
other words, for an endothermic reaction increasing the temperature shifts the exothermic
equilibrium towards the products.
If the reaction is exothermic (A,H° is negative) the term (—A,;H°/RT) is
positive and so increasing the temperature makes it less positive. As a result endothermic
increasing the temperature makes both In K and K smaller (see Fig. 8.10). In
words, for an exothermic reaction, increasing the temperature shifts the equi- T —,
librium towards the reactants. Fig. 8.10 Graph showing the different
You are probably familiar with these conclusions as they are often de- ways in which In K varies with
temperature (as predicted by Eq. 8.5) for
scribed as being a consequence of Le Chatelier’s Principle, one statement of an exothermic and an endothermic
which is that ‘the equilibrium shifts in order to oppose the change’. So, for reaction. For an endothermic reaction
an endothermic reaction, an increase in the temperature is opposed by absorb- the equilibrium constant increases with
temperature; for an exothermic reaction,
ing heat which means that the reaction must move further to products, i.e. the the opposite is the case. Thus, Eq. 8.5 is
equilibrium constant must increase. Equation 8.5 puts this application of Le an expression of Le Chatelier’s Principle.
Chatelier’s Principle on a quantitative footing.
We will look at two examples of the effect of temperature on the position
of equilibrium.
Dimerization of NO2
For our first example, let us look again at the dimerization of NO2(g):
2NO2(g) = N204(g).
From tables, we find for this reaction that at 298 K, A; H° = —57kJ mol7! and
A-S° = —176J K~! mol~!. These values make sense as a bond is being made,
so we expect the release of energy, and the reduction in entropy is associated
with two moles of gas going to one.
At 298 K we can compute A,G° and then K:
A,G° => A,H° — TA,S°
= —57 x 10° — 298 x (—176)

— 46x 10° J mol7!.
Equilibrium
124
K =exp
(“ar) 7
—~(—4.6 x 103)
= &P \ 33145 x 298
=63 ”—
We see that A,G° is negative and so the equilibrium constant is greater than l,
showing that at equilibrium the products are preferred. However, A,G° is not
that negative and so the equilibrium constant is modest (6.3) which tells us that
although the equilibrium lies on the side of the products, there are still plenty
of reactants present.
As the reaction is exothermic, lowering the temperature will increase the
proportion of products. Let us repeat the same calculation at 273 K:
AG? = —57 x 103 — 273 x (176) = —9.0 x 10? J mol7!

—(—9.0 x 103)
oo 8.3145 x 273
The equilibrium constant is now much larger, meaning that at this lower tem-
perature the proportion of dimer (N2Oq) present is larger than it was at 298 K.
Repeating the calculation at 373 K gives K = 0.062, showing that at this
higher temperature the fraction of dimer present is very small. These calcu-
lations illustrate very nicely how we can shift the equilibrium one way or the
other simply by altering the temperature.
Extraction of metals
Our second example is rather different, and concerns the commercially very
important process of extracting metals from their oxides. Typically, this is done
by heating the oxide with carbon which acts as a reducing agent, thus releasing
the free metal and forming CO or CO. The chemical processes involved are
often fairly complex, but for a simple divalent metal oxide (general formula
MO) we can get a flavour of what is going on by just considering the reaction
MO(s) + C(s) = CO(g) + M(s). (8.6)

The following table gives thermochemical data (at 298 K) for C, CO, the metals
copper, lead and zinc and their oxides.
Cc Co Cu CuO Pb PbO Zn ZnO

ArH? /kJ mol7! -111 —157 —219 —348
S°JK~!mol7! 5.74 198 33.2 42.6 64.8 66.5 25.4 40.3
Using these we can compute A,G° for the reaction of Eq. 8.6 at any tem-
perature (provided we assume that A;H° and A,S° are not temperature de-
pendent). The results for reaction at 1000 °C and 1500 °C are shown in the
following table (you can check to see if you agree with the numbers); similar
data are also shown in Fig. 8.11.
at 1000 °C at 1500 °C
oxide A;H°/kJ mol7! A,S°/J K7! mol7!
A,G° /kJ mol7! A,G° / kJ mol7!
CuO 46 182 —186 —277
PbO 108 190 —133 —229
ZnO 274 177 12.3 -76.2
As expected, for all the metals the reaction has a positive A,S° result-
ing from the generation of one mole of gas on the right-hand side of the
T/°C
equation. The reactions are all endothermic which means that, as A;G° =
1000 1200 1400 1600
A,H° — TA,S°, a negative value of A;G° can only be achieved by a positive 100 + = ;
A,S° (which is what we have here) and a sufficiently high temperature that the
G°/kJ mol"!
OF
—T A,S° term is dominant.
We need to be careful here not to fall into making an error about how tem- -100
perature influences the equilibrium constant. Read casually, the previous para-
graph makes it sound as if it is the sign of A;S° which determines how the
©. -200 — ob
equilibrium will change with temperature. This is not what the paragraph says: -300l Cu
what it says is that the sign of A, S° determines whether A,G° rises or falls with Fig. 8.11 Plot showing how the standard
temperature, but A,G° is not the equilibrium constant. To find the equilibrium Gibbs energy change for the reaction of
constant from A,G° we need to use A;G° = —RT In K, a relationship which Eq. 8.6 varies with temperature for three
different metals; the process will only be
introduces another temperature dependence; so knowing how A,G° varies with successful if A; G° is significantly
temperature is not the whole story. negative. It is clear, therefore, that the
In fact, as we have seen, the effect of temperature on the equilibrium con- extraction of Zn requires a higher
temperature than for Cu or Pb. The
stant is given by Eq. 8.5 values of ArG° have been computed
assuming that A,H® and A,S° are
ink A,H® (1 n independent of temperature, and so the
nK =— = plots are straight lines.
R T R
This shows that it is the sign of A;H° which determines how the equilibrium
constant depends on temperature as it is A; H° which is multiplying the (1/7)
term. So, although it is true that if A,S° is positive, increasing the temperature
will cause A,G° to decrease, to understand the effect of temperature on the
equilibrium constant we need to look at the sign of A;H°. What the table
tells us therefore is that as the reactions are all endothermic, increasing the
temperature will move the equilibrium to the products.
This subtle point about the temperature dependence does not alter the ba-
sic idea that if A,G° is substantially negative the position of equilibrium will
lie almost entirely toward the products. At 1000 °C, A,;G°® is substantially
negative for copper and lead, showing that metal + carbon monoxide will be
formed. However, for zinc A,G° is positive at this temperature, which tells us
that the equilibrium will favour the reactants (ZnO + C) and so zinc metal will
not be extracted.
In order for the reduction to produce zinc we have to raise the temper-
ature higher, and we can see from the table that at 1500 °C, A,G° is now
substantially negative for the ZnO reduction. We therefore conclude that to
obtain copper or lead from their oxides, reduction with carbon at a temperature
of 1000 °C will be sufficient, but to obtain zinc from its. oxide a substantially
higher temperature will be needed. Figure 8.11 presents the results in a slightly
different way, and from this graph we can see that for Zn the switch over from
a positive to a negative A,G° occurs at around 1100 °C. In practice we need to
126 Equilibrium
be at a higher temperature than this to ensure that A;G° is sufficiently negative

for the equilibrium to favour products strongly.
Concentration
The value of the equilibrium constant determines the ratio of the concentrations
of products and reactants, and this has a fixed value at a particular temperature.
However, we can influence the actual concentration of a product by altering
the concentrations of the other species. How this works is best illustrated by
an example, such as the equilibrium shown in Fig. 8.12 which is the reaction
used to form an ester (E) from an alcohol (A) and a carboxylic acid (C).
O fe)
I + CH3OH I + 0
H3C~ OH HaC~ OCH
Cc A E
Fig. 8.12 An ester, E, is formed from the reaction between a carboxylic acid, C, and an alcohol,
A; water is also produced in the reaction. Here the reaction is between ethanoic (acetic) acid and
methanol to give methyl ethanoate (methyl acetate).
The equilibrium constant, K, for this reaction is given by:
[EJeqlH2O0 Jeq
K = ——————_ 8.7
[CleglA leq 7)
where [E]eq means the equilibrium concentration of the ester, and so on. Re-
member that at a particular temperature the value of K is fixed, so the con-
centrations of the four species must adjust themselves so that the ratio on the
right-hand side of Eq. 8.7 is equal to the value of K.
Suppose we let the reaction come to equilibrium, but then by some means
we Start to remove one of the products (say the ester); what will happen? The
moment the concentration of the ester falls the ratio on the right-hand side of
Eq. 8.7 will be too small; to restore it to the correct value we have to either
increase the amounts of ester and water or decrease the amounts of carboxylic
acid and alcohol. Both of these things are achieved if some of the acid and
alcohol react to give more water and ester.
So, by removing the ester we disturb the equilibrium in such a way that
the only way it can be restored is for more ester to be produced; we say that
the reaction has been ‘forced to the right’ by removing the product. For the
particular reaction shown in Fig. 8.12 it turns out that the ester is more volatile
than any of the other species, so it can simply be distilled off constantly forcing
the reaction to the right and so increasing the yield of the ester.
Another way of forcing the reaction to the right is to add more of one of
the reactants. Once again, this reduces the value of the ratio on the right-hand
side of Eq. 8.7 and so to restore the correct value some of the alcohol and
carboxylic acid have to react to form more ester. If the alcohol is a simple one,
such as methanol, we would probably use it as the solvent for the reaction, thus
ensuring that its concentration was very high.
A similar example is in the formation of an imine from the reaction of a
ketone with an amine, Fig. 8.13. In this reaction water is produced on the right-
hand side of this equilibrium and so the yield of the imine can be increased by
removing the water. This can be achieved by adding a solid material known as
a molecular sieve which is a special kind of zeolite clay which has cavities into
which water will bind tightly. This effectively removes the water, so forcing the
reaction to produce more of the imine in an attempt to restore the equilibrium.
H3C H3C CH3

\ \ /
jf ° + CHs3NH, C=N + H2O
H3C H3C
ketone imine
Fig. 8.13 An imine is formed from the reaction of a ketone and an amine.
We can use the same strategy to shift the reaction the other way; suppose,
for example, we want to hydrolyse an ester back to the alcohol and the car-
boxylic acid — this is the reverse reaction shown in Fig. 8.12. To force the
reaction to the left, we need to increase the amount of the species on the right
and this is easily done by making water the solvent. Now its concentration will
be very large, and so the reaction will be shifted to the left, in favour of the
hydrolysed products.
The formation of an acetal from the reaction of a ketone with an alcohol,
shown in Fig. 8.14, is an equilibrium reaction which we sometimes want to
force one way and sometimes the other.
H3C H3C
‘C=O
= ++ CHOH So Oso +
/ ° 1“ Nocu 2
H3C H3C 3
ketone acetal
Fig. 8.14 The formation of an acetal from a ketone and an alcohol is a readily reversible reaction which
we can ‘force’ one way or the other by altering the conditions.
If we wish to form the acetal we use the alcohol as the solvent and add
a dehydrating agent — both choices force the equilibrium to the right. On the
other hand, if we wish to regenerate the ketone from the acetal, we run the
reaction in aqueous solution so that the excess of water will force the reaction
to the left.
The Haber process

Our final example is the formation of ammonia from hydrogen and nitrogen
(the Haber process):
No(g) + 3H2(g) = 2NH3(g)
which is the way in which millions of tonnes of ammonia are synthesized each
year. In order to make the reaction go at a reasonable rate temperatures of
around 400 °C are needed together with a solid iron oxide catalyst. Unfortu-
nately, the reaction is exothermic (A;H° = —92 kJ mol—!) so raising the tem-
perature shifts the equilibrium towards the reactants (Section 8.4 on p. 122); it
is found that at this high temperature the equilibrium constant is about 10-2.
128 Equilibrium
The strategy for increasing the yield of ammonia is simply to remove the
H, and N, in ammonia and then let the nitrogen and hydrogen come back to equilibrium,
thus generating more ammonia. In practice this is done by passing the gases
over the catalyst so that they come to equilibrium, then separating out the am-
monia and finally recycling the unreacted gases back over the catalyst; the
process is illustrated in Fig. 8.15.
unreacted H, and N,, Removing the ammonia is quite easyas ammonia will liquefy at a temper-
reaction chamber ature (around —35 °C) well above the temperature at which either nitrogen or
(catalyst at 400 °C)
hydrogen will liquefy. So, simply by cooling the gases the ammonia can be
separated as a liquid.
iN
Y
Coupling reactions together

Suppose that the reaction we are interested in has a positive A;G°, which
means that the equilibrium constant will be less than 1, and very little of the
cooling products will be formed. One way to force the formation of larger amounts of
chamber
the products is to ‘couple’ the reaction to another reaction which has a negative
A,G° so that, when taken together, the overall A,G° is negative. What we are
doing is driving the reaction with the unfavourable A,G° using a reaction with
a favourable A,G°.
For the two reactions to be coupled in this way they must be able to influ-
liquid NH, ence one another; having the reactions take place in separate beakers on the
bench or simply mixing all the reagents together will not achieve the required
Fig. 8.15 Simplified picture of the
arrangement for the Haber process for coupling. In Nature this coupling is achieved using enzymes which control
synthesizing ammonia. After coming to and direct the chemistry, and indeed very many of the important chemical pro-
equilibrium over the catalyst, only a small
cesses of life are driven by this kind of coupling. We do not have time here
fraction of the hydrogen and nitrogen has
been converted to ammonia. The to go into the details of exactly how enzymes achieve this coupling, but will
ammonia is separated from the other just illustrate the outcome for one very fundamental reaction in living systems
gases by liquefying it, and the unreacted
— the formation of the peptide bond.
nitrogen and hydrogen are returned to
the reaction chamber. In living systems, proteins are synthesized by the polymerization of amino
acids into long chains. The key reaction in forming the chain is for amino acids
to condense together by forming peptide bonds, as shown in Fig. 8.16 for the
formation of a dipeptide.
O O NH> O
R c HoN : al N
le
“CH” OH * CH”
,
OH Ro NBR
bt
CH™ oH * 420
amino acid 1 amino acid 2 dipeptide
Fig. 8.16 Two amino acids condense together to form a dipeptide, and in doing so form the -CONH—
peptide bond, shown in the grey box. Different amino acids have different side groups, R and R’.
This reaction has a rather unfavourable A,G° of +17 kJ mol7! and so at

equilibrium very little of the dipeptide will be formed. To get round this, Nature
couples this reaction with the hydrolysis of adenosine triphosphate (ATP) to
adenosine diphosphate (ADP):
ATP + H20 —> ADP + inorganic phosphate.

This reaction has a very favourable A;G° of —30 kJ mol~!, sufficiently

negative to outweigh the positive A,G° for the formation of the peptide bond.
So, taken together the two reactions have an overall negative A,G° making the
formation of the peptide bond thermodynamically feasible.
Nature frequently uses the hydrolysis of ATP to ADP to drive reactions
which have unfavourable (positive) values of A,G°, and so ATP is often re-
garded as the source of the energy which drives the chemistry of life. We
can now see that by ‘energy’ we really mean Gibbs energy, as this is what
drives reactions forward. So, when biochemists describe ATP as a high energy
molecule, what they mean is a high Gibbs energy molecule.
For all of this to work, the cell has to be able to make ATP, for example
from ADP by running the hydrolysis reaction backwards:
ADP + inorganic phosphate —> ATP + H20O
but this reaction has an unfavourable A,G° of +30 kJ mol7! and so will not
go on its own. As with peptide bond formation, Nature drives the formation of
ATP by coupling it to another reaction, such as the oxidation of glucose.
This oxidation is often described as the fundamental source of ‘energy’ for
many living systems, including ourselves. The overall reaction is
C6H1206 + 602 —> 6CO2 + 6H20
and this has a very large negative A,G° of —2880 kJ mol7!. In Nature the
oxidation of one molecule of glucose is coupled by a very complex series of
enzymatically controlled reactions to the formation of around 38 molecules
of ATP from ADP, and in this way the high energy ATP molecules needed in
many other processes in the cell are generated.
We obtain sugars, such as glucose, from plants which synthesize them from
carbon dioxide and water, releasing oxygen in the process. The formation of
glucose in this way is simply the reverse of the oxidation given above:
6CO?7 + 6H20 — C6H120¢6 + 602.
However, this reaction has a very large positive A,G° so how can it be made to
go? The answer is that plants utilize the energy from light, in a process known
as photosynthesis, to force this reaction. A very complex and subtle series of
processes and reactions are used to harvest the energy from light and utilize
it to form glucose — Nature has perfected this scheme over the millennia of
evolution that have led to the green plants we know today.
The coupling of an unfavourable reaction to a favourable one is absolutely
crucial in the chemistry of life. We can only marvel at the subtle and efficient
way in which Nature is able to achieve this coupling.
130 Equilibrium
8.5 Equilibrium and rates of reaction

Suppose that we have an equilibrium between reactants A and B and products
C and D:
A+tB2=C+D.
Let us start out with just A and B mixed together; they will react at some rate
to generate some of the products C ajd -D. However, the moment these are
formed they will start to react together and, via the reverse reaction, regenerate
the reactants A and B.
The more of the products C and D that are formed, the faster the reverse
rate becomes; in contrast the more of the reactants A and B that are used up the
slower the forward rate becomes. We know that eventually we will reach the
equilibrium point at which the concentrations are all constant — what must have
happened therefore is that the rates of the forward and reverse reactions have
become equal. When this is the case, as fast as A and B are removed by the
forward reaction they are replenished by the reverse reaction in such a way that
their concentrations do not change. Equilibrium is thus a dynamic situation —
it 1S not that the reactions have stopped, it is just that the rate of the forward
and back reactions are equal so that it appears that nothing is happening.
This brings us on to the whole topic of the rates of reactions — something
which we have been carefully ignoring, or at least sidestepping, up to now. The
problem is that even if a reaction is favourable in thermodynamic terms, i.e. has
a large negative A,G°, this does not guarantee that the reaction will actually
take place at a measurable rate.
For example, the oxidation of glucose is accompanied by a very large neg-
ative A,G° of some —2880 kJ mol~!, yet we can go out and buy glucose
powder safe in the knowledge that it will not burst into flames spontaneously.
The reaction is thermodynamically feasible, but kinetically very slow.
Other reactions are both thermodynamically feasible and fast. For example,
the neutralization of acids and bases (essentially the aqueous phase reaction
H30* + OH~ —> 2H20) is both thermodynamically feasible and very fast.
Even reactions in which the position of equilibrium lies close to the reactants
can come to equilibrium quickly. For example, the dissociation of ethanoic
(acetic) acid has an equilibrium constant of only 107° but nevertheless the
equilibrium is established very quickly once the acid is dissolved in water.
Reactions involving fons — particularly if they are oppositely charged as in
the case of the neutralization of acids and alkalis — tend to be quite fast, and
similarly reactions involving simple proton transfer (for example from ethanoic
acid to water) are often rapid. However, reactions involving the breaking and
making of bonds (say to carbon) are likely to proceed much more slowly.
In the next chapter we will look at the factors which control the rates of
reactions and then go on to see what a study of reaction rates can tell us about
the mechanism of a reaction.
9 Rates of reaction
At the end of the previous chapter we commented on the fact that just because a
reaction has a favourable A,G° (i.e. the position of equilibrium lies well to the
products) this does not necessarily mean that it will proceed at a measurable
rate. In this chapter we will look at the factors which determine the rates of
reactions and how we can influence these. This is clearly a matter of great
importance if we actually want to do some practical chemistry, as well as being
a rather fundamental aspect of chemical reactions.
The second part of this chapter will move on to discuss how the study of
reaction rates can give us information about what is going on in a reaction,
particularly when the reaction involves more than one step. This is the experi-
mental study of reaction mechanisms.
9.1 What determines how fast a reaction goes?

In this section we will focus on a particular reaction (one you have seen before
on p. 108) which is the nucleophilic substitution of Br~ by OH”, Fig. 9.1.
Each factor which influences the rate will be considered in turn.
H.H HH
HO° + jo— Br —» HO-—C + Br
\
H H
Fig. 9.1 The nucleophilic substitution of OH for Br-; here CH3Br is transformed into CH3OH.
The molecules must collide

For there to be a reaction it is pretty obvious that the molecules must come
close enough for them to interact — in other words they must collide. So we
expect the rate of reaction to depend on the number of collisions per unit time,
that is the rate of collisions.
In a gas or liquid the molecules are in a constant state of motion as a re-
sult of their thermal energy; they are rushing around in random directions,
colliding with one another and hence changing their direction and speed. We
therefore expect that the rate of collisions will depend on the concentration of
the molecules, the speed at which they are moving and their size.
The rate of collisions between two different species, A and B, will clearly
go up as the concentration of either of them increases, simply because as an A
molecule rushes around the chance that it will encounter a B molecule goes up
as the number of such molecules increases. It does not seem unreasonable that
the rate of A—B collisions is proportional to both the concentration of A and
that of B:
rate of A-B collisions « [A] x [B].
132 Rates of reaction
The second thing which affects the rate of collisions is the speed at which
the molecules are moving: we argue that the faster the molecules move the
more likely they are to collide. As you may know, there is a distribution of
speeds with which the molecules move, but the majority of molecules are
moving at speeds which are not that far from the average. In a gas, this av-
erage speed increases as the temperature increases; this is hardly a surprise as
increasing the temperature of a gas increases its energy which appears in the
form of kinetic energy — the molecules therefore move faster.
The final factor which determines the collision rate is the size of the
molecules: the larger they are the more chance there is of them colliding and
so the higher the collision rate.
For simple molecules in the gas phase we can use gas kinetic theory to
estimate the number of collisions as being around 102’ per second in a vol-
ume of 1 dm? at room temperature and pressure — a very large number by
any standards. In liquids the situation is more complex because we also need
to consider the rate at which the reactant molecules diffuse together; however,
once in proximity the collision rate between two molecules is even greater than
in a gas.
The molecules must collide with the correct orientation

As we discussed on p. 109, for the reaction shown in Fig. 9.1 to take place
the nucleophile (OH™~ ) must approach the carbon from the opposite side to the
Br; this requirement is dictated by the nature of the HOMO and LUMO which
are involved in the reaction. Only a fraction of the total number of collisions
between CH3Br and OH™ will have this correct orientation.
We have also seen that the attack on a carbonyl! group by a nucleophile has
a strongly preferred geometry (Fig. 7.16 on p. 105). Indeed it is common for
reactions to require the reactants to approach in a particular orientation — we
call this the steric requirement of a reaction.
We can also expect that the larger and more complex a molecule becomes,
the smaller the fraction of collisions with the correct orientation will become.
For example, it would not be unusual for only | in 10° collisions to have the
correct geometry.
There must be sufficient energy in the collision to overcome the energy

barrier ‘
For the vast majority of reactions there is an energy barrier, which is a min-
imum amount of energy that two colliding molecules need in order to react;
simply colliding with the correct orientation is not sufficient — the molecules
must bring with them sufficient energy to cross this barrier.
This energy comes mainly from the collision of the two molecules. Each
has a certain kinetic energy as a result of its thermal motion — some molecules
have more than the average energy and some less. So, some collisions will
have higher energy than others, and it is the ones which have an energy which
is high enough to overcome the barrier that can lead to reaction.
Typically, the energy barrier for a reaction is between 10 and 100 kJ mol.
At room temperature the average thermal energy of a molecule is around
4 kJ mol~!, so you can see that only those molecules whose energy is very
9.2 Why is there an energy barrier? 133
much greater than the average are likely to have enough energy to react when
they collide with another molecule. The fraction of molecules with energies
much greater than the average is very small, so that typically only 1 in 10°
molecular collisions has sufficient energy to overcome the barrier.
Our picture of the reacting molecules is of there being many collisions be-
tween the reactants but of these only a minuscule fraction have the correct
orientation and sufficient energy to lead to products. A reactive encounter be-
tween molecules is thus a very rare event.
9.2 Why is there an energy barrier?

The presence of an energy barrier means that even if the products have lower
energy than the reactants, the energy of the reacting molecules must first go up
as they start to rearrange themselves into product molecules. Once this rear-
rangement has proceeded to a certain extent the energy starts to fall, eventually
reaching the energy of the product molecules when they are fully formed.
Figure 9.2 shows a mechanical analogy which is helpful in understanding
this concept of an energy barrier. We start with a block stood on its short side
and then try to push it over so that it lies on its long side. The end position
has lower (gravitational) potential energy than the starting position, but to push
the block over we first need to tilt it on its edge and in doing this the centre
of gravity rises, thus increasing the potential energy. Only when the tilt is
far enough will the block fall over and reach the low-energy position. Like a
chemical reaction, this system has to first go to a state of higher energy before
it can reach the state with the lowest energy.
(a) (b) (c) (d)
Fig. 9.2 Pushing a block over from position (a) to (d) results in a reduction in the (gravitational) potential
energy as the centre of mass (indicated by the dot) is lowered. However, to get the block to topple
over it first has to be tilted on its edge, as shown in (b) and (c). This raises the centre of mass and so
the energy goes up; only when the block is past position (c) does the energy go down. This process
is an analogy for a chemical reaction: the products are lower in energy than the reactants, but to get
between the two the energy must first rise; in other words there is an energy barrier which has to be
overcome.
A simple example of such an energy barrier occurs in inversion of ammo-

nia, described on p. 87 and illustrated in Fig. 9.3. What happens here is that the
molecule is going from a trigonal pyramidal geometry with the hydrogens on
one side to an identical, mirror image, geometry with hydrogens on the other
side; the whole process is rather like an umbrella being blown inside out. We
can measure the progress along this route by giving the H-N-H bond angle;
this starts at 107°, goes to 120° at the planar geometry and then reduces back
to its original value as the hydrogens move to the other side of the molecule
‘and end up in mirror image positions.
We know that the equilibrium geometry of ammonia has an H-N-H bond
angle of 107°, which means that this is the bond angle which gives the lowest
energy; as the bond angle opens out the energy must therefore increase, as
H H H
shown in Fig. 9.3. Our interpretation of this is that the overlap of the orbitals
is less than optimum as we move away from a bond angle of 107°, and so the
Hy \ N
: x
UH energy increases.
H
4 [:
HH
\H
The symmetry of the problem dictates that the highest energy point is when
trigonal trigonal the bond angle is 120°, as this lies midway between the two trigonal pyramidal
pyramidal! planar pyramidal
geometries. The species at this highest energy point is called the transition
state; in this case it is an ammonia molecule with a planar geometry.
For the reaction of Fig. 9.1 we can trace out how the energy varies as the
energy
reaction proceeds from reactants to products just as we did for the inversion
of ammonia. However, for this reaction the progress from reactants to prod-
107° 120° 107° ucts cannot be specified by a single angle since as the reaction proceeds several
—§\|{*—__
H-N-H bond angle bond angles as well as internuclear distances are changing. So, we plot the en-
ergy against what is called the reaction coordinate which is a complex combi-
Fig. 9.3 In the inversion of ammonia the
progress from the equilibrium trigonal nation of distances and angles, the precise definition of which need not concern
pyramidal geometry to its mirror image is us; all we need to know is that this coordinate starts at the reactants and ends
characterized by the H-N-H bond angle.
at the products. The plot of energy against reaction coordinate is known as an
A plot of the energy against this angle
shows a maximum at 120°; this energy profile for the reaction; such a profile for the nucleophilic substitution
corresponds to the planar transition state. reaction is shown in Fig. 9.4.
reactants transition state products

©
H , H
© \ /
HO" + c—Br HO- -C- -Br Ho—c,, + Br®
Hd ~L ‘H
H HH H
Pal
ou
o
jm
oD
——_—__>
reaction coordinate
Fig. 9.4 Energy profile for the reaction shown in Fig. 9.1; the reactants appear on the left and the
products on the right. At the energy maximum is the transition state, a suggested structure for which
is shown; there are partial bonds (shown dashed) to the OH and Br. The products are shown as
having lower energy than the reactants, and so the reaction is exothermic; nevertheless, there is still
an energy barrier to be overcome in going from reactants to products.
9.3 Rate laws 135
At the top of the profile there is a suggestion as to what the transition state
might look like. The carbon atom is five-fold coordinated with partial bonds
to both the incoming OH™ and the departing Br~. Over the course of the
reaction, the hydrogen atoms have to move from one side to the other and so
in the suggested structure we have shown them in the same plane as the carbon
atom. It is important to realize that there is no way we can guess at the precise
structure of the transition state — all we can do is to speculate about its general
form.
It is clear from these pictures why the energy has to rise when the transition
state is formed. The bonding is far from optimum — not least because there are
five groups around the carbon; in addition, there are partially made bonds to
the OH and the Br. This unusual geometry and the presence of partial bonds
is typical for a transition state; it is not surprising, therefore, that its energy is
higher than that of the reactants.
It is very important to realize that the transition state is not a real molecule:
we cannot isolate it or study its physical or chemical properties. Rather, it is
just a transient arrangement of the atoms through which the reaction passes
on its way to products. The transition state has a fleeting existence, perhaps
having a lifetime of 10~!? s or even less.
Reactions with no barrier

The usual situation is for reactions to have a barrier, but there are some that
do not. A typical example is the reaction in the gas phase between oppositely
charged ions, such as:
Lit + FF” — LiF.
In this case we can attribute the lack of a barrier to the fact that the oppositely
charged reactants attract one another and also that in going to the products no
bonds are broken.
A second example involves the gas-phase recombination of radicals:
*CH3 + °CH3 — CH3—CHs3.
We can attribute the lack of a barrier here to the fact that a bond is being made
and none broken. The C-H bonds only have to adjust themselves in a minor
way to accommodate the new bond.
9.3 Rate laws

We described on p. 131 how the rate of collisions between two molecules A
and B is expected to be directly proportional to the concentration of each. So,
if our nucleophilic substitution reaction (Fig. 9.1) really does take place by the
OH™ and the CH3Br colliding in a reactive encounter, we would expect to find
that the rate of reaction is proportional to the concentration of each of these
species.
Experiment confirms this expectation; it is found that
rate of reaction = K[OH” ][CH3Br]. (9.1)
This is called the rate law for the reaction.

[OH"] | :
time time
Fig. 9.5 Plots of the concentration of the reactant OH~ and the product Br- as a function of time for
the reaction of Fig. 9.1. Note how the concentration of the reactant falls towards zero and that of the
product rises. Also shown as dotted lines on the left-hand plot are tangents to the curve; these give
the slope of the curve which is the same thing as the rate. The way in which the absolute value of the
rate falls as time increases can clearly be seen. In drawing these plots it has been assumed that at
time zero the two reactants have the same concentration and that the concentrations of the products
are zero.
In the rate law k is the rate constant for the reaction; its value depends
on all of the factors described in Section 9.1 (p. 131 onwards) apart from the
concentration, i.e. the size of the molecules, the speed they are moving at, and
the fraction of collisions with the correct orientation and sufficient energy to
overcome the barrier. The value of k has to be determined experimentally and
it is usually found that it depends strongly on temperature, as will be discussed
in the next section.
The rate of reaction is how the concentration varies with time:
change in concentration during time interval Ar

rate of reaction =
At
the dimensions of the rate are thus concentration x time~!. The concentration
of any of the reactants or products can be used to specify the rate. However,
if we use a reactant whose concentration decreases with time, the rate will
be a negative quantity; in contrast, if we use a product whose concentration
increases with time, the rate will be positive.
Looking at the rate law, Eq. 9.1, we can see that at the start of the reaction
the rate will have its highest (absolute) value. As the reaction proceeds, the
amount of the reactants OH~ and CH3Br decreases, and so the rate decreases,
eventually reaching zero at infinite times. Figure 9.5 shows plots of how the
concentration of the reactants and products varies with time for this reaction.
The slope of a tangent to the curve gives us the rate at any time: this is
because the slope is the change in concentration over the change in time, which
is precisely the same thing as the rate. In Fig. 9.5 we see that for the plot of
[OH™] the slopes are negative, as expected for a reactant, and that as time
proceeds the slopes are becoming less negative, i.e. the rate of the reaction is
falling.
Order
Another way of writing the rate law of Eq. 9.1 is
rate of reaction = k[OH™ ]![CH3Br]!

9.4 The Arrhenius equation 137
where we have raised the concentrations to the power of 1. Of course this is

rather superfluous as [OH~]! means exactly the same thing as [OH]. How-
ever, the notional presence of this power in the rate law allows us to define
something called the order.
The order is the power to which the concentration is raised in the rate law.
So, in the above rate law the order with respect to OH™ is |, and similarly the
order with respect to CH3Br is also 1. We can also define an overall order,
which is the sum of the orders of each species present in the rate law. So for
this reaction the overall order is 2.
An order of 1 is often called first order and an order of 2 is called second
order. Thus the rate law for the reaction between hydroxide and bromomethane
(methyl bromide) can be described as being first order with respect to hydrox-
ide and bromomethane and second order overall.
9.4 The Arrhenius equation

The rate constant generally depends quite strongly on temperature, and exper-
imental studies have shown that this temperature dependence often obeys the
Arrhenius equation:
—E,
k= Aexp(Fa). (9.2)
In this equation A is the pre-exponential or A factor, E, is the activation energy

and, as usual, T is the absolute temperature and R is the gas constant. This
strong temperature dependence of the rate constant means that when quoting
its value it is essential to state the temperature at which the measurements were
made.
energy
The interpretation of this equation is that the activation energy is the height
of the energy barrier over which the reactants must pass — to be specific, it
is the energy difference between the reactants and the top of the barrier (the
transition state), as shown in Fig. 9.6.
From gas kinetic theory it can be shown that the fraction of collisions with
energy greater than or equal to E, is given by exp(— £,/RT) and, as we dis-
cussed on p. 132, only these collisions are energetic enough to give rise to
products. We therefore interpret the exponential term in the Arrhenius equa-
——\¥\¥—__
reaction coordinate
tion as arising from the fraction of collisions with sufficient energy to pass over
the barrier. Fig. 9.6 The activation energy, Ea, is the
difference in energy between the
If there is no energy barrier all of the collisions have sufficient energy to reactants and the top of the barrier, i.e.
become products; in mathematical terms if E, = 0 the exponential term, which the transition state. For reaction to take
is the fraction of collisions with sufficient energy, becomes 1. Under these place, the molecules must collide with at
least this much energy and from gas
circumstances the Arrhenius equation (Eq. 9.2) tells us that the rate constant is kinetic theory we can show that the
equal to the pre-exponential factor, A. So, we can interpret the value of A as fraction of collisions satisfying this
giving the rate constant in the absence of an energy barrier to reaction. requirement is exp(—Ea / RAT).
From the discussion in Section 9.1 (starting on p. 131) we can see that
the factors which contribute to the value of A are: the fraction of collisions
which have the correct geometry, the size of the molecules and the speed with
which they are moving. As we noted, this last factor depends on tempera-
ture but it turns out that this dependence is much weaker than that due to the
increasing E, k increasing T
A A ee
ann a OL
—— -
a 0 cEa
Fig. 9.7 Plot (a) shows the Arrhenius equation (Eq. 9.2) prediction for the variation of the rate con-
stant, k, with temperature, T. The three lines are computed assuming the same value of the A factor
but different values of the activation energy; the darker the line the larger the activation energy (the
temperature scale does not start at zero). At high enough temperatures, the rate constant will reach
the value of the pre-exponential factor, A; for the curves shown here the temperature is well below this
limit. Plot (b) shows the variation of the rate constant, k, with activation energy, Ea; the three lines are
computed assuming the same value of the A factor, but different temperatures; the darker the line the
higher the temperature. Note how the rate constant responds rapidly to changes in the temperature
and the activation energy.
exp(—E,/RT) factor in the Arrhenius equation. So, to a reasonable approxi-

mation we can assume that the A factor is temperature independent.
Figures 9.7 (a) and (b) show the predictions from the Arrhenius equation
(Eq. 9.2) for how the rate constant varies with temperature and activation en-
ergy, respectively. The plots show how the exponential term makes the depen-
dence of the rate constant on the values of these two parameters rather strong.
We can turn the Arrhenius equation into a straight line plot by taking (nat-
ural) logarithms of both sides:
—E,
k=A exp RT
“t— InA increasing E, —E, 1
hence Ink =InA+ —)}.

R T
So, a plot of Ink against 1/T will be a straight line with slope (—E,/R); the
intercept with the vertical axis, when 1/T goes to zero, will be In A. Figure 9.8
shows examples of this straight line plot; note how increasing the activation
e)
energy results in a more negative slope.
1/T We can interpret the intercept with the vertical axis in the following way:
Fig. 9.8 An Arrhenius plot of In k against when the temperature becomes very high, i.e. when 1/T goes to zero, each
(1/T) which, as explained in the text, is
collision is sufficiently energetic to overcome the barrier and so each leads to
expected to give a straight line. The three
lines are computed assuming the same reaction. The rate constant then has the value A because, as we noted before,
value of the A factor but different values this is the value of the rate constant in the absence of a barrier. For most
of the activation energy; the darker the
reactions this limit in which the rate constant becomes equal to the A factor
line the larger the activation energy. As
the slope of the line is —Ea/FA the greater occurs at such high temperatures that either the reactants will have dissociated
the activation energy the more negative into atoms or other reactions will have taken over. The limit is therefore of
the slope. The intercept with the vertical
theoretical, rather than practical, interest.
axis (when 1/T goes to zero) gives In A;
this extrapolation is shown by dotted Given some experimental data of rate constants at different temperatures
lines. we can determine E, andA by plotting Ink against 1/7, just as in Fig. 9.8. The
9.5 Elementary and complex reactions 139
slope is (— E,/R) and the intercept with the vertical axis is In A. In practice, the
range of temperatures over which data is available is usually rather limited so
a long extrapolation is needed to find In A, making this value rather inaccurate.
9.5 Elementary and complex reactions

There is plenty of experimental evidence which leads us to believe that the
nucleophilic substitution reaction:
OH” + CH3Br —> CH30H + Br-
really does take place as it is written, i.e. the products are formed directly from
a single reactive encounter between an OH™ ion and a CH3Br molecule.
There are many other reactions for which the experimental evidence is that
they take place as written. For example, the two gas-phase reactions:
Cl + 03 —> ClO + O2
ClO + O —> Cl+ 02
which are key reactions in the destruction of atmospheric ozone by halogens,

are thought to proceed as written.
On the other hand, there are reactions which certainly do not take place
as written. For example, the gas-phase reaction between H2 and Brz has the
stoichiometric equation
H>. + Bro —> 2HBr
but there is much experimental evidence that the reaction does not occur by a
single reactive encounter between an H2 and a Br2 molecule.
Similarly, the oxidation of glucose in living systems has the overall stoi-
chiometry
CgH1206 + 602 —> 6CO2 + 6H20
but certainly does not take place by one molecule of glucose reacting directly
with six of oxygen!
What we are describing here is the distinction between an elementary and a
complex reaction. An elementary reaction is one which we believe to take place
in a single encounter and as the chemical equation is written. For example, the
gas phase reaction between a hydrogen atom and Br2 written as
H+Br2 — HBr+ Br
is elementary and so takes place in a reactive encounter between an H anda

Br2 molecule.
Complex reactions take place by a set of elementary reactions, called
a mechanism; typically such a mechanism will involve the generation of
molecules, called intermediates, which do not appear in the stoichiometric
equation.
For example, the decomposition of ozone into oxygen has the stoichiome-
try:
203 —> ‘302
and it is thought that this overall process has a mechanism involving the fol-
lowing three elementary steps involving an oxygen atom as an intermediate:
b; “+ 0) +0
O2 +0
— 0;
O + 03 —> 20.
A great deal of effort has been put into determining the mechanisms of
chemical reactions, 1.e. identifying the intermediates and the contributing ele-
mentary reactions. Determining a reaction mechanism is by no means a simple
task and usually requires the application of many techniques, such as a study
of how the rate of the reaction varies with the concentrations of the species
involved and the detection of intermediates.
Rate laws for elementary and complex reactions

As an elementary reaction takes place as written we can write down the rate
law directly from the chemical equation. For example:
OH™ + CH3Br —> CH30H + Br— rate = k,[OH™ ][CH3Br]

Cl + 03 — ClO + O02 rate = kp[Cl][O3]
H + Br2 —> HBr+ Br rate = k,[H]({Bro].
In contrast, for a complex reaction there is no way that we can write down
the rate law just by looking at the stoichiometric equation; in such cases the
rate law has to be determined by experiment. A good example is the reaction
between H2 and Br2 in the gas phase
H2 + Br2 —> 2HBr
which has the experimental rate law:

‘
rate of formation of HBr =

ka{H2]{Bro]}°/?
[Bro] + k.[HBr]
which is, to say the least, far from obvious. Indeed, the observation of such a
complex rate law is good evidence for the reaction not being elementary.
However, the opposite is not necessarily true: some reactions have simple
experimental rate laws but turn out to have complex mechanisms! For example
the gas-phase reaction between H) and Ip
H2 + Iz —> 2HI
is found to have the rate law
rate of loss of H2 = k¢[H2] [I].

9.6 Intermediates and the rate-determining step 141
Despite the simplicity of this rate law it is thought that the mechanism for
the reaction may involve the intermediate IH» in the following steps (although
there are certainly other possible mechanisms)
In — 2]
21 — Ih
I+ H2 —> IH?
I+ IH. —> 2HI.
9.6 Intermediates and the rate-determining step

We mentioned on p. 139 that complex reactions often involve the formation
of species, called intermediates, which are neither reactants nor products. The
nature and properties of these intermediates often play a crucial role in the
reaction, and as a result you will find that a lot of attention is focused on them.
Let us take as an example the reaction of the nucleophile OH~ with the
acyl chloride CH3COCI (discussed in Section 7.6 on p. 110). The reaction is
shown in Fig. 9.9 and proceeds in two steps via the formation of a tetrahedral
intermediate, I.
(o)
7 step 1 HO ew step 2 7 }
© — 7 —
OH + C Ze Cc + Cl
H3C~ Cl H3C~ ~Cl H3C~ OH
I
Fig. 9.9 The reaction between OH and CH3COCI proceeds in two steps. Step (1) involves nucle-
ophilic attack on the carbonyl carbon to form a tetrahedral intermediate (I); in step (2) this intermediate
collapses expelling CI~ and re-forming the carbonyl group.
It is very important not to confuse intermediates with transition states (in-

troduced on p. 134). A transition state only has a fleeting existence and is
simply an arrangement of the atoms through which the reaction must pass in
order to reach the products; a transition state is located at a maximum in the
energy profile (see Fig. 9.4 on p. 134).
In contrast, an intermediate is as much a real molecule as any other — we
can probe its structure using spectroscopic and physical techniques, and indeed
we may even be able to isolate it. Admittedly, it is often the case that these
intermediates are very reactive and rather short lived, so studying them can be
difficult, but they do not have the fleeting existence of a transition state.
Intermediates — like product and reactant molecules — are found at minima
in the energy profile of the reaction; this is in contrast to transition states which
are found at maxima. These points are illustrated in Fig. 9.10 which shows the
energy profile for the reaction between OH™ and CH3COCI.
The first energy barrier in this reaction is between the reactants (OH™ +
CH3COCI) and the intermediate, I; at the top of the barrier we have the first
transition state in which presumably the C—O z bond is partially broken and
the bond between the OH™ and the carbonyl carbon is partially made. The
activation energy for this first step is E,,;. The second barrier in the reaction is
between the intermediate and the products, and proceeds via the second transi-
tion state in which we imagine that the C—O bond is partially re-formed and
the C-Cl bond is partially broken.
transition
state 1
energy
transition
>
HO.
hy
aN +c?
i
He07 cI
reactants products
———_
reaction coordinate
Fig. 9.10 Energy profile for the reaction shown in Fig. 9.9. The reactants, products and intermediate
all occur at energy minima, whereas the two transition states occur at maxima. The activation energy
for the first step, E, 4, is shown as being much greater than that for the second, E, 9; this means that
the first step is the rate-determining step.
The rate-determining step

In Fig. 9.10 the activation energy for step (2) is shown as being considerably
less than that for step (1). What this means is that, all other things being equal,
the rate of step (2) is much greater than that of step (1). As a consequence, the
moment a molecule of the intermediate is formed in step (1) it almost imme-
diately reacts in step (2) to give the products. The rate at which the products
are formed is thus determined only by the rate of step (1), which is therefore
described as the rate determining step.
Assuming that step (1) is an elementary reaction we can write its rate as:
rate of step (1) = ky[(CH3COCI][OH™J.
It follows that as step (1) is the rate-determining step, the rate of formation of
the products is the same as the rate of step (1):
rate of formation of the products = k;[(CH3COCI][OH

7 ].
We see that although a two-step mechanism is involved, the rate law is rather
simple.
It is important to be careful not to fall into the mistake of describing this
kind of reaction by saying that ‘the first step is slower than the second; so the
former is rate-determining’. What is wrong with this is that during the reaction,
step (2) is in fact proceeding at the same rate as step (1).
If the concentrations of the reactants and the intermediate were equal it
would undoubtedly be the case that the rate of step (2) would be greater than
that of step (1). However, the concentrations are not equal — far from it. We ex-
pect to find that the concentration of the intermediate will be very low because
it is whisked away to products as quickly as it is formed. No matter how low
the activation energy for step (2) becomes, this reaction cannot take place until
9.7 Reversible reactions 143
step (1) produces some of the intermediate; as a consequence, the rate of step
(2) cannot be greater than that of step (1), and in fact they are equal.
9.7 Reversible reactions

So far we have ignored the possibility that a reaction can go backwards as
well as forwards, i.e. that reactions are reversible. If a reaction has a position
of equilibrium which favours the products very strongly then to a very good
approximation we can ignore the reverse reaction, but if the position of equi-
librium is more balanced between reactants and products we must consider the
reverse reaction.
Like the forward reaction, the reverse (or back) reaction has an activation
energy as is illustrated in the energy profiles shown in Fig. 9.11. The activa-
tion energy for the forward reaction, Ef, is the energy difference between the
transition state and the reactants. For the reverse reaction, which takes us from
products back to reactants, the activation energy, E,,,, is the energy difference
between the transition state and the products.
The profile of Fig. 9.11 (a) is for an exothermic reaction, i.e. one in which
the products are lower in energy than the reactants; it follows that the activation
energy of the reverse reaction must be greater than that of the forward reaction.
In contrast, for an endothermic reaction, shown in Fig. 9.11 (b), it is the forward
reaction which has the greater activation energy.
From the diagram it is clear that the energy difference between the products
and reactants, AE, is related to the activation energies in the following way:
AE = Eat _ Ear. (9.3)
The energy difference is not quite the same thing as A,H° for a reaction, but it
is closely related to it.
Inspection of Fig. 9.11 (b) shows that for an endothermic reaction the ac-
tivation energy (of the forward step) cannot be less than AE, i.e. the activa-
tion energy is greater than the endothermicity. For an exothermic reaction,
Fig. 9.11 (a), there is no particular relationship between the activation energy
of the forward reaction and AE.
(a) exothermic (b) endothermic

energy
energy
Eat
products
AE
™ |
reactants
products
—_—_—_ ————_
reaction coordinate reaction coordinate
Fig. 9.11 Energy profiles for (a) an exothermic reaction and (b) an endothermic reaction. In each
case the energy difference between the reactants and products, AE, is equal to the difference in the
activation energies of the forward (Eq») and reverse (Ea,r) reactions.
Relation to the equilibrium constant

Let us suppose that we have a reversible reaction between A and B to give P
and Q:
ke
A+B = P+Q.
ky
We will also suppose that the forwatd and reverse reactions are elementary
(with rate constants kp and k,, respectively) so that, as described on p. 140, we
can write the rates of the two reactions as:
rate of forward reaction = ks[A][B]

rate of reverse reaction = k,[P][Q].
In Section 8.5 on p. 130 we argued that at equilibrium the rates of the

forward and reverse reactions are equal. So, we can write:
k[A]eq[Bleq = k-[Pleq [Qleg
where we have added the subscript ‘eq’ to remind ourselves that this relation-
ship is only true when the concentrations are at their equilibrium values. Some
rearrangement of this equation gives:
kt (PleqlQleg
kr [A]eqlBleq
We recognize the quantity on the right-hand side as the equilibrium constant,
Kegq, and so see that the ratio of the rate constants for the forward and reverse
reactions is equal to the equilibrium constant:
Keg = T-. (9.4)
We have already seen in Section 9.4 on p. 137 that rate constants can be ex-
pressed in terms of an A factor and activation energy using the Arrhenius equa-
tion, Eq. 9.2. If we do this for the forward and reverse reactions we have:
-E —-E
kp = Af exp ( ae) k, = A,exp ( a)
where Af and A, are the A factors for the forward and reverse reactions, re-
spectively. Using these expressions in Eq. 9.4 we have:
Ag exp (=e)
Keg = _E
Ar exp ( a)
_ Af —(Ea — Ear)
= — exp
Ay RT
At (>
= — exp | ——
Ar RT
where, to obtain the last line, we have used Eq. 9.3.
What the final line tells us is that the equilibrium constant varies with tem-
perature in a way determined by the value of AE, the energy change for the
reaction. If AE is positive (i.e. an endothermic reaction), an increase in the
temperature results in an increase in the equilibrium constant. In contrast, for
an exothermic reaction which has a negative AE, increasing the temperature
reduces the equilibrium constant.
If we identify AE with A,H° these conclusions are exactly the same ones
we came to in Section 8.4 on p. 122, albeit using a very different approach.
9.8 Pre-equilibrium
Look again at the energy profile for the CH3COCI + OH™ reaction (Fig. 9.10
on p. 142), and consider the ‘choices’ which are open to the intermediate, I.
It can either go on to products, which is what we assumed would happen, or
it could take the reverse of step (1) and go back to reactants. However, as the
profile is drawn the activation energy for I going back to reactants is larger
than for it going on to products; we therefore expect that the latter will be the
dominant process as it will have a larger rate constant. With hindsight you can
see that it was no accident that the energy profile was drawn in this way!
Figure 9.12 shows two different energy profiles for a two-step reaction.
Profile (a) is just like that in Fig. 9.10; the dominant process for the interme-
diate is for it to go on to products and, as we have seen, this leads to step (1)
being the rate-determining step. In profile (b) the activation energy for the in-
termediate returning to reactants is lower than for it going to products, and this
leads to quite a different result.
In the case of profile (b), once an intermediate is formed its most likely fate
is to return to reactants in the reverse of step (1), which we will call step (—1).
The slower conversion from the intermediate to products via step (2) does take
place but only a small fraction of the intermediates take this route.
(a) (b)
B i
dominant z
oa) dominant
4 _—_—_—_ D —_
3 3
Cc
P P
—- ——_—_—_
Fig. 9.12 Alternative energy profiles for a two-step reaction. In (a) the pathway with the lowest activa-
tion energy open to the intermediate I is to go to products, P.-In (b), the lowest energy pathway is for
the intermediate to return to reactants, R, by the reverse of the first step. Profile (a) leads to kinetics
in which the first step is rate-determining; profile (b) leads to pre-equilibrium kinetics, as described in
the text. The dominant fate of the intermediate, I, is indicated by the arrow.
The picture of what is happening is therefore this: the reactants and the
intermediate are rapidly interconverting via steps (1) and (—1), with a small,
slow, ‘bleed off’ of some of the intermediate to give products via step (2). Un-
der these circumstances we are justified in assuming that the reactants and the
intermediate are in equilibrium; even though some of the intermediate goes to
form products, the equilibrium between reactants and intermediates is quickly
re-established. Step (2) is the rate-linfiting step as its rate constant controls the
rate of formation of the product.
This idea that the intermediate (or intermediates) are in equilibrium with
the reactants is called the pre-equilibrium hypothesis. The ‘pre-’ comes about
because this equilibrium precedes the rate-determining step. This kind of situ-
ation is quite commonly encountered, particularly if the pre-equilibrium steps
involve protonation and deprotonation; let us see an example of how this works
out.
Example: oxidation of methanoic acid with bromine

In aqueous solution, methanoic (formic) acid is oxidized to CO2 by Br ac-
cording to the stoichiometric equation:
HCOOH + Br2 + 2H20 —> CO, + 2Br~ + 2H307+.
This reaction has been investigated experimentally and it is found that the rate
of the reaction is sensitive to the concentration of H3O* in the solution (this
can be varied by adding a strong acid, such as HCl). The rate law is found to
be:
k[Br2][HCOOH]
rate of consumption of Brz = (9.5)
[H307]
This rate law tells us that the rate goes down as the concentration of H30+
goes up, i.e. the reaction is inhibited by H30°. This gives us a clue as to what
might be happening in this reaction: suppose that the bromine actually oxidizes
the anion HCOO™ rather than the methanoic acid itself; this anion would have
to be formed by the usual dissociation of the methanoic acid:
HCOOH + H20 = HCOO™ + H30°%.
Increasing the amount of H30* (for example by adding a strong acid) would
shift this equilibrium to the left (see p. 126) and so decrease the amount of
HCOO™ present; this would then explain why the rate decreases as the con-
centration of H3O* increases.
With this in mind we propose the following mechanism:
ks
HCOOH + H20 — HCOO™ +H307 steps (3) and (—3)
k 4
HCOO™ + Br sy products step (4).
We have numbered the steps (3), (—3) and (4) to avoid confusion with steps (1)
and (2) in the reaction in Fig. 9.9 on p. 141.
The first line is simply the acid dissociation equilibrium. We have written
the rate constants of the forward and reverse reactions as k3 and k_3, respec-
tively, but recall from the discussion on p. 144 that the equilibrium constant is
the ratio of these two rate constants:
Ka =i (9.6)
If we assume that pre-equilibrium applies, then we can find an expression

for [HCOO7] using the definition of the equilibrium constant, Ka:
[H30* ][HCOO7]
K, =>
[HCOOH]
This can be rearranged to give
K, [HC
[HCoo~}
= SsCOOH]
[H307]
The rate of consumption of bromine is just the rate of step (4), so
rate of consumption of Br2 = k4[HCOO™ [Bro]

_ kaKal[HCOOH][Bra]
7 [H30*]
The rate law we have derived is therefore
kerp[ HCOOH] [Br2]
rate of consumption of Br2 =
[H307]
where Keg; iS
and on the last line we have used Eq. 9.6.

The rate law we have predicted has the same form as the experimental
rate law, Eq. 9.5, which at least means that our mechanism is consistent with
the experimental observations and that the pre-equilibrium assumption is valid
for this mechanism. Finally, it shows us that the experimental rate constant
is in fact a composite of three rate constants, or — looked at another way — a
composite of a rate constant and an equilibrium constant.
When can we apply the pre-equilibrium assumption?

For the pre-equilibrium assumption to be valid the rate constants have to be
such that an intermediate, once formed, is much more likely to return to where
it came from than to go on to react further. In other words, the further reactions
of the intermediate must be slow compared to the intermediate returning to the
species from which it was formed.
A common situation in which these conditions hold is when the intermedi-
ate is formed by protonating or deprotonating one of the reactants. Such proton
transfers often have rather large rate constants — they are sometimes described
as being ‘facile’ or ‘easy’ processes.
9.9 The apparent activation energy

For any reaction — be it an elementary reaction or one with a complex mecha-
nism — we can always determine the activation energy by measuring how the
rate constant varies with temperature and then fitting the data to the Arrhe-
nius equation in the way described on p. 138 and illustrated in Fig. 9.8. If the
reaction 1s a complex one proceeding via a number of elementary steps, the
question is what is the relationship between this measured activation energy
(we will call this the apparent activation energy) and the activation energies of
the elementary steps? We will use the two reactions we have been discussing
in this chapter to illustrate this relationship and this will lead us to a rather neat
general conclusion about the size of the apparent activation energy.
The reaction between CH3COCI and OH™ (Fig. 9.9 on p. 141) proceeds by
a two-step mechanism:
CH3COCI + OH “4 I step (1)

1-2; CH;COOH +CI- step (2).
We argued that if the first step is rate-limiting the overall rate just depends on
the rate constant for step (1):
rate of formation of the products = ky [CH3COCI][OH7].
We can write the rate constant k; using the Arrhenius equation (Eq. 9.2 on
p. 137) as:
—Eq)
k,; 1 = A) 1 ex p( RT — ) },
Therefore the apparent activation energy for the overall reaction is exactly the
same as the activation energy for step (1), the rate-limiting step.
Identifying the activation energy for the oxidation of methanoic acid is a
little more complicated; let us remind ourselves of the reaction scheme:
ky
HCOOH + H20 = HCOO™ + H30* steps (3) and (—3)
k_3
_ k
HCOO™ + Br. —> products step (4).
We found that by applying the pre-equilibrium assumption the overall rate law
was:
kerp[HCOOH][Br]
rate of consumption of Brz =
[H30*]
where
9.9 The apparent activation energy 149
We now write each rate constant in the expression for kerr using the Arrhe-
nius law:
A4 exp (=H) A3 exp (ae

kett =
A_3 exp (=H)
A4A3 (Ss + E43 _ E,,-3)
= exp .
A_-3 RT
From this we can identify the apparent activation energy as the energy term in
the exponent:
Ea apparent = Ea4 + Ea3 _ Ea,-3

9.7
= Ea,4 + AE3 O7)
where on the last line we have used the relationship between the energy change
for a reversible reaction and the difference in activation energies between the
forward and reverse reactions: AE = E43 — Ea,—3 (Eq. 9.3 on p. 143).
We see in the case of this mechanism that the apparent activation energy
depends on the activation energies of all three steps or, alternatively, on the
energy change of the reversible reaction and the activation energy of the final
step.
The general result

In fact, there is a simple result which describes both reaction schemes. It is that
the apparent activation energy for a reaction is the energy difference between
the reactants and the highest energy transition state on the reaction pathway:
this is illustrated in Fig. 9.13.
Profile (a) is appropriate for the reaction between CH3COCi and OH™ in
which the first step is rate-determining and so the highest energy transition state
energy
energy
—————“ ——_p-
Fig. 9.13 Two different energy profiles showing how the apparent activation energy is equal to the
energy separation between the reactants and the highest energy transition state along the reaction
pathway. Profile (a) is for the reaction between CH3COCI and OH™; the first step is rate-determining
and so has the highest activation energy. The apparent activation energy is therefore equal to the
activation energy of the first step. Profile (b) is for the reaction between HCOOH and Bro; the second
transition state is the highest energy point and so the apparent activation energy depends on the
activation energies of all of the reactions, as described in the text.
on the reaction profile is the first transition state we encounter. The apparent
activation energy is simply the activation energy of the first step.
Profile (b) is for a reaction in which the first step is reversible, such as the
reaction between HCOOH and Brp which we have been discussing. The second
transition state, between the intermediate and the products, is now the highest
energy point on the pathway and so this determines the apparent activation
energy. As we have discussed, this activation energy can either be expressed in
terms of the activation energies of all three reactions or the energy change of
the first reaction and the activation energy of the second (Eq. 9.7).
9.10 Looking ahead

We mentioned at the start of this chapter that a knowledge of the rates of reac-
tions can be helpful in teasing out the mechanisms of complex reactions. The
oxidation of methanoic acid is a good example of this — we saw that the rather
unusual rate law, in which H3O0* was seen to inhibit the reaction, suggested a
mechanism which we went on to show was consistent with the experimental
rate law.
In the coming chapters we will see a great deal more of how a study of
kinetics helps us to unravel the mechanism. For example, consider the follow-
ing two nucleophilic substitution reactions and their experimentally determined
rate laws:
CH3Br+ Nu” —> CH3Nu + Br- rate = k[CH3Br][Nu]

(CH3)3CBr + Nu~ —> (CH3)3CNu + Br rate = k’[(CH3)3CBr].
Note that the reaction involving bromomethane is first order in both the nucle-
ophile and the alkyl halide, and hence second order overall. In contrast, for
the 2-bromo-2-methylpropane (t-butyl bromide) the concentration of the nu-
cleophile does not appear in the rate law, making it first order overall. How can
this be? We will see in Chapter 11 how changing the alkyl group from methyl
to t-butyl causes a change in mechanism and hence in the rate law. However,
before we can understand that we must look at a further electronic effect —
delocalization and conjugation.
10 Bonding in extended systems
— conjugation
In Chapter 6 we introduced hybrid atomic orbitals and used these as the basis
for a localized description of bonding; such an approach is perfectly adequate
for many molecules, especially those commonly encountered in organic chem-
istry. We then saw in Chapter 7 how we could use a knowledge of orbitals to
help us to understand and rationalize reactions.
In the following chapters we want to move on to some reactions which
involve molecules and intermediates whose bonding cannot be described ade-
quately using only the localized approach of Chapter 6. The kind of molecules
we are talking about typically have extended wz systems and are often charged;
we will see that a proper description of the bonding in such molecules needs
the delocalized MO approach for parts, but not usually all, of the molecule.
For example, let us consider how to describe the bonding in propene and
benzene (Fig. 10.1). For propene we choose sp? hybridization for C), and sp”
hybridization for the other two carbons. The overlap of these hybrids with one
another or with hydrogen 1s AOs gives us the C-C and C-H o bonds, and the
mz bond is formed by the overlap of out-of-plane 2p orbitals on C2 and C3. For
propene, this localized description is entirely adequate.
For benzene we choose all of the carbons to be sp* hybridized and once “a
again form C—C and C-H o bonds by overlapping these HAOs with one an- H3C~ 5 SCH,
other and with the hydrogen 1s AOs. The six out-of-plane 2p orbitals then
overlap in three pairs to form three mz bonds, thus giving alternating single and
double bonds round the ring.
However, the problem with this localized description is that there is much
physical and chemical evidence that all the C-C bonds in benzene are the same
(for example, they are all the same length); a structure in which there are al-
ternating single and double bonds is simply not consistent with these observa-
tions. Benzene is therefore an example of where a delocalized description is
required.
You may have come across the idea of using resonance structures to ex-
plain the delocalized nature of the bonding in benzene. The idea is illustrated Fig. 10.1 The bonding in propene can be
described adequately using 2c-2e bonds,
in Fig. 10.1 where two different structures for benzene are shown which only
but benzene requires a delocalized
differ in the placement of the double bonds. On their own, both structures are approach for the x system. The two
an inadequate representation of the bonding in benzene, but the two taken to- resonance structures for benzene,
gether convey the crucial point that there is 7 bonding between all the adjacent connected by a double headed arrow, are
an attempt to describe this delocalized
carbons. In a sense, the ‘real’ structure of benzene lies in between these two, bonding using structures with localized
and this is what the structure at the bottom of Fig. 10.1 tries to convey — the bonds. The structure at the bottom, in
which a circle is drawn in place of the
circle indicating that the 7 bonds are delocalized round the ring.
separate 7 bonds, is another way of
It is tempting to think that the molecule flips between these resonance struc- representing the delocalized bonding in
tures, but this is a mistaken view. These structures are simply our attempts to benzene.
152 Bonding in extended systems — conjugation
Fig. 10.2 Surface plots of the three highest energy occupied MOs of benzene; positive and negative
parts of the wavefunction are indicated by different shading of the grid lines. The skeleton of the
benzene molecule is also indicated. These three MOs are clearly formed from the overlap of the out-
of-plane 2p orbitals; the lowest energy MO, 17, is reminiscent of the ‘doughnut’ picture often used to
explain the delocalized x bonding in benzene.
make a localized description of the bonding in a molecule which in fact has

delocalized bonding. No one of the structures is an adequate representation
of the actual bonding, but taken together they convey something closer to the
truth.
A useful analogy here is that of an mule, which we get by crossing a horse
with a donkey. To describe an mule as a horse or a donkey would be inaccurate
and it certainly does not interconvert rapidly between the two! Rather we need
to recognize that although an mule has parts that are reminiscent of a horse and
parts that are reminiscent of a donkey, it is something altogether different.
The MO picture of benzene naturally takes a delocalized view of the bond-
ing, and the three highest energy occupied MOs shown in Fig. 10.2 illustrate
this very well. It is clear from looking at these MOs that they are formed from
the overlap of 2p orbitals which point out of the plane of the molecule, and in-
deed you may recognize the 1a MO as being like the ‘doughnut’ picture often
given to explain the equivalence of the C-C bonds in benzene. However, this
doughnut is just one of the six MOs which are formed from the overlap of the
six out of plane 2p orbitals; only three of the MOs are occupied, and these are
the ones shown in Fig. 10.2.
It is not really necessary to use the delocalized MO approach to describe
all of the bonding in benzene. By using sp* hybrids on the carbons we can
generate a o framework of 2c-2e C-H and C-C bonds, just as we described
above. The MO approach is only needed to treat the six out of plane 2p orbitals.
This way of descrtbing benzene is typical of the approach we will adopt.
The majority of the bonding in a molecule can usually be described using 2c-2e
bonds to give the o framework. This leaves a few orbitals, typically pointing
out of the plane of the o framework, whose interaction we need to treat using
the MO approach.
What we will see in the next section is that it is quite easy to construct de-
localized 2 MOs which are formed from 2p orbitals arranged in a row; such an
arrangement is surprisingly common in molecules and so we will find imme-
diate use for these MQOs in describing aspects of structure and reactivity. We
will start with four p orbitals in a row to introduce the idea and then reduce the
number to three; this might seem a bit illogical, but it turns out that it is easier
to see what is going on if we start with four orbitals.
10.1 Four p orbitals in a row 153
10.1. Four p orbitals in a row

A good example of what we mean by a ‘row of p orbitals’ is the case of bu-
tadiene, shown in Fig. 10.3. We can describe the bonding in this molecule by
H
making all of the carbons sp* hybridized, and then using these orbitals to form Zon ee
2c-2e bonds with the hydrogens, and carbon-carbon o bonds between the ad-
H»C
H
jacent carbons. There are 22 valence electrons in butadiene, four from each butadiene
carbon and one from each hydrogen. The o framework uses up 18 of these
electrons, leaving four to go into the out of plane z system which is formed
from four 2p orbitals. "o-®s¢° Sy OHO60C
We know that these four 2p orbitals will overlap to form four MOs (remem- HoH
ber that in Section 5.4 on p. 68 we described how the number of MOs is equal o framework m™ system
to the number of AOs which are being combined). These four MOs will each Fig. 10.3 At the top is shown the
be a linear combination of the four AOs: conventional representation of butadiene,
with two C—C double bonds separated by
MO = c) AQ; + c2AQO2 + c3A03 + cq4AO4 a C-C single bond. If the carbons are
sp@ hybridized the bonding can be
described as a o framework formed from
where AO, is the AO on atom 1, and c, is the coefficient for this AO and so the overlap of these hybrids with one
on. These coefficients are a measure of the contribution that each AO makes to another and with the hydrogen 1s AOs,
together with a delocalized x system
the MO. It is relatively easy to compute the energies of the MOs and the cor-
formed from the overlap of the
responding coefficients; the results are presented in cartoon form in Fig. 10.4. out-of-plane 2p AOs. These orbitals are
an example of four p orbitals in a row.
4n
energy
1x
{
‘in a line’. The coefficient
Fig. 10.4 Cartoons of the four x MOs resulting from four 2p AOs overlapping
n) of each AO is indicated by the size of the AO; note that some of the coefficients are
(or contributio
lowest in-energy and has no nodes
negative so the sign of the AO has been inverted. MO 1m is the
lowest and has one node between atoms 2 and 3. As the energy rises
between the atoms; 27 is next
. The orbital coefficients are
to MOs 37 and 4x the number of nodes increases to 2 and 3, respectively
the molecule, as shown on the right; for 1 there is one
given by a sine wave which can be drawn over
half sine wave, for 27 there are two and so on.
The o framework lies in a nodal plane for each MO; this is why the MOs are
_ designated 7. First look at the energies of the MOs: there are two MOs lower
in energy than the AOs and two higher.- The first two are therefore bonding
MOs and the latter two are anti-bonding.
The coefficients of the AOs fall into a simple pattern as the MOs go up in
energy. In the lowest energy MO, 17, all of the coefficients are positive; there
is therefore constructive overlap betweeh all of the adjacent orbitals and it is
therefore not surprising that this MO is the most strongly bonding.
In 27 there is a nodal plane between atoms 2 and 3; there is thus construc-
tive interference between the AOs on atoms | and 2, and between those on
atoms 3 and 4, but destructive interference between atoms 2 and 3. Since there
are two constructive interactions but only one destructive, the orbital is overall
bonding, but not as strongly as Iz.
MO 3z has two nodal planes (between atoms | and 2, and 3 and 4), so
there is destructive overlap between the AOs on atoms | and 2, and 3 and 4.
There is constructive overlap between the AOs on atoms 2 and 3, but this time
(in contrast to 27) the orbital coefficients are such that the orbital is net anti-
bonding.
Finally, 477 has a nodal plane between each adjacent atom, and so there is
only destructive interference; it is therefore the most anti-bonding MO. The
general pattern is that the number of nodal planes increases as the energy goes
up; this is a feature of both AOs and MOs which we have seen before.
It turns out that for each MO the coefficients follow a sine curve which
can be inscribed over the molecule in a particular way. As the energy goes
up the number of half sine waves increases, as is shown in the diagram. The
coefficients for the jth atom in the kth MO are given by
. jkr
sin
N+1
where N is the number of AOs which are overlapping; N = 4 in the case of
butadiene.
So, for MO 37, which has k = 3, the coefficients are:
cy = sin1x3
( : *) = 40.95 2
ec =sin ( | = ~0.59
c3 = sin(3x3
( = *) = 0.59 4
cy=sin ( =~] = +0.95
which match up with the picture shown in Fig. 10.4 (to compute these you need
to remember that the argument of the sine is in radians, not degrees).
Butadiene
In butadiene we have four electrons to place in the 7 system; we therefore
put two (spin paired) in 1 and two in 27. All of the electrons are therefore
accommodated in bonding MOs.
MO Im gives rise to m bonding across the whole molecule so, In contrast
to the simple picture shown in Fig. 10.3, there is 2 bonding between
C2 and
C3. In fact, because in I the coefficients on C> and C3 are the greatest
(see
10.1 Four p orbitals in a row 155
Fig. 10.4), this MO contributes more to the bonding between C2 and C3 than it
does to the bonding between C, and C2 or C3 and C4.
MO 27 has a node between C2 and C3 and so gives rise to an anti-bonding
interaction between these two atoms. However, because of the smaller orbital
coefficients on these atoms, the anti-bonding interaction does not outweigh the
bonding interaction due to 177. So, overall, the occupation of MOs Iz and 27
leads to some z bonding between C2 and C3.
MO 2z also gives rise to a bonding interaction between C; and C2, and
between C3 and C4. However, the contributions from MOs 1x and 27 do not
add up to the same amount of 7 bonding as is present in the simple 2c-2e 2
bond in ethene. So in butadiene the degree of 7 bonding between C, and C2
and between C3 and Cy is less than in ethene. \ H 1.506
The overall picture for butadiene is of 2 bonding across all four carbons,
but with a larger interaction between atoms 1 and 2, and 3 and 4. The bond ON
HC a CHg
length data give in Fig. 10.5 confirm this description. The C2—C3 bond length
in butadiene is shorter than the C-C single bond in propene, but not as short
1.349 1.467
as the C-—C double bond. Similarly, the nominal double bond in butadiene
between C; and C2 is a little longer than the double bond in propene. These
data point to partial 7 bonding between C2 and C3, and a weakening of the z
Vat
Hoc’ 2 Cc Aa4
bonding between C, and Co, just as we predicted from the MOs. { H
3
The two a bonds which we drew in the simple structure of butadiene are
Fig. 10.5 Experimental bond length data,
described as being conjugated, which means that they form a pattern of alter- in A, for the C-C bonds in propene and
nating single and double bonds; the key feature of such systems is that they butadiene. The C;—Cz bond in butadiene
contain an uninterrupted chain of p orbitals lying out of the plane of the o is longer than the C—C double bond in
propene, whereas the Co—-Cz bond is
framework. In conjugated systems the mz bonding is not localized between shorter than the C—C single bond in
pairs of atoms, but rather is delocalized across the whole conjugated system. A propene. These data provide evidence
proper description of the bonding in such a molecule requires the delocalized for there being partial x bonding between
Co and C3.
MO approach.
In contrast, a molecule such as CH2=CH—CH2—CH=CH1? is not conjugated
as the m bonds are separated by two o bonds; a localized description of the
bonding is perfectly adequate for such a molecule.
In summary, because the two z bonds in butadiene are conjugated the sim-
ple structure given in Fig. 10.3 does not really describe the bonding adequately.
In fact, the four electrons assigned to the two z bonds in the simple structure
are occupying x MOs which are delocalized over the whole molecule and so
give rise to 2 bonding between all the adjacent carbon atoms.
Propenal
Propenal (also called acrolein), whose structure is shown in Fig. 10.6, is closely
related to butadiene; both have two conjugated double bonds with a total of
four a electrons. Just as we did in butadiene, we can model the 7 system in
propenal as four 2p orbitals in a row. The crucial difference, however, is that
the four orbitals no longer have the same energy — the oxygen orbital is lower
in energy than that on the carbon (see Fig. 4.34 on p. 57). The four 2p orbitals propenal
will still give rise to four MOs, but they will not be quite the same as those
Fig. 10.6 The structures of butadiene
shown in Fig. 10.4; the presence of one oxygen AO will alter both the energies and propenal; to go from one to the other
and the orbital coefficients of the MOs. It requires a computer calculation to all we have to do is replace the CHo at
work out the MOs which are shown, as surface plots, in Fig. 10.7. position 1 in butadiene with an oxygen.
Fig. 10.7 Surface plots of the four 7 MOs in propenal; the numbering scheme matches that used in
Fig. 10.8 and that of the corresponding MOs in butadiene shown in Fig. 10.4.
Broadly speaking the four MOs are comparable with those for butadiene
shown in Fig. 10.4 on p. 153. MO Iz: is the lowest in energy and has no nodal
planes between the atoms. For propenal the coefficient of the AO from Cg is
so small that it hardly shows up on this plot.
MO 27 is the next highest in energy; as in butadiene it has constructive
overlap between the AOs on C3 and Cg, but in propenal the coefficient of the
AO from C2 is very small so there is little constructive overlap between this
AO and the one from oxygen. Finally, just as in butadiene, MOs 3m and 4x
have two and three nodal planes, respectively.
Figure 10.8 shows the computed ordering of the MO energies for propenal;
compare these with the corresponding MOs in butadiene shown in Fig. 10.4 on
4n
p. 153. It always turns out that the lowest energy MO will be lower in energy
3x than the lowest energy AO, and the highest energy MO is higher in energy
C=
than the highest energy AO; all of the MO diagrams you have seen conform to
this pattern. The other two MOs lie somewhere in between, and it is not really
o—
possible to guess where they will fall. What we do know is that as the number
2n of nodes in an orbital increases, so will the energy, and this is consistent with
1n
the ordering shown in Fig. 10.8.
AOs MOs
In Section 5.6 on p. 76 we described the MOs formed from the overlap of
two AQOs of unequal energies. What we found was that the two AOs do not
Fig. 10.8 MO diagram for the x system in contribute equally to a given MO, and that the AO which is closest in energy
propenal. As explained in the text, the
lowest energy MO (1:7) lies below the to a particular MO is the major contributor to that MO. This same idea carries
lowest energy AO and similarly the forward to more complex situations, so we expect that the 12 MO in propenal
highest energy MO (4:7) lies above the will have a greater contribution from the oxygen AO than will 47; the surface
highest energy AO.
plots in Fig. 10.7 bear this out.
There are four electrons to accommodate in this 7 system, so the la and
27 MOs are occupied, making 27 the HOMO and 3z the LUMO. The Iz MO
contributes to bonding between C2 and the oxygen, with a smaller contribution
between C2 and C3. The C2—-C3 bond length of 1.484 A, shorter than a typical
C-C single bond, confirms the presence of partial 7 bonding between these
two atoms. MO 2z contributes only to bonding between C3 and C4, making a

mz bond just as in the localized structure.
The way in which this partial double bond character between C> and C3
can be represented using localized 2c-2e structures in shown in Fig. 10.9.
H ins _ H (>
C @ Cc
HecA_Pc He *_*c~
<x
H H
A B
Fig. 10.9 Resonance structures for propenal which illustrate the partial bonding between carbons 2
and 3; resonance structure B contributes less than structure A.
The fact that the Cp—C3 bond length is only slightly shorter than a typical
C-C single bond suggests that resonance structure B only makes a minor con-
tribution. This structure also suggests that there is more electron density on the
oxygen and less on C4.
The removal of electron density from C4 is supported by the observation
that, whereas C=C bonds are not usually attacked by nucleophiles (Section 7.4 H 4H
on p. 106), such attack is sometimes seen for conjugated C=C bonds. For Fig. 10.10 In the reaction between CN—
example, CN™ can either attack the carbonyl carbon directly (like the reactions and propenal the nucleophile can also
attack C4 as well as the carbonyl carbon.
in Section 7.3 on p. 102), or it can attack at C4, as shown in Fig. 10.10.
10.2 Three p orbitals in a row

When three p AOs overlap in a line, three MOs are formed and these can be
constructed in just the same way that we did for four overlapping p orbitals;
Fig. 10.11 shows the MOs in cartoon form. As before, the energy of the MOs
increases as the number of nodes increases, and the orbital coefficients are
given by sine waves which can be inscribed over the molecule.
The one slight difference between the overlap of three as opposed to four
orbitals is that in the former case one of the MOs (277) is non-bonding (it has
approximately the same energy as the AOs). That this is so can also be seen
from the form of the MO; the orbital coefficient on the central atom is zero so
there is neither constructive nor destructive overlap between AOs on adjacent
atoms. The end carbons are too far apart for their AOs to interact significantly.
Understanding the 7 MOs which are formed from the overlap of three p
orbitals turns out to be very useful for rationalizing all sorts of observations,
such as why amide bonds are flat and why the two C-C bonds in the allyl anion
are of equal length.
Allyl anion
The allyl anion can be thought of as arising from the removal of Ht from
propene, as shown in Fig. 10.12; a very strong base is needed to remove a
proton from this molecule. The bonding in the allyl anion can be described
by having sp? hybridization on each carbon atom; these hybrids overlap with
one another or hydrogen 1s orbitals to give a o framework. The remaining
out-of-plane 2p orbitals overlap to form the three MOs pictured in Fig. 10.11.
po SM 3)
ay
energy
——tf a eneeen ences —— Pn
——— in CO
Fig. 10.11 Cartoons of the three s MOs resulting from three 2p AOs overlapping in a line; the diagram
is drawn in the same way as Fig. 10.4. As before, the orbital coefficients are given by sine waves which
can be drawn over the molecule; for the 17 MO there is one half sine wave, for 27 there are two and
so on. As expected, the energies of the MOs increase as the number of nodes increases.
The (valence) electron count in the anion is four from each of the three
H carbons, plus one from each of the five hydrogens plus one more electron for
| H the negative charge; this gives a total of 18 electrons. Of these, 14 are used in
He o7™ oa H propene the o framework and the remaining four go into the 2 system. We thus assign
| | two electrons to MO Iz and two to 27.
H H
Occupation of 12 contributes to the 7 bonding over all three atoms,
strong
whereas occupation of 27 contributes nothing to the 2 bonding. However,
base
having 27 occupied does increase the electron density on the two end carbons
H
preferentially over the centre carbon.
HN oN ge allyl As Iz is symmetrical about the centre, we conclude that the two C-C bonds
l anion must be identical. Indeed, experiment bears this out and the bonds are found to
H H be the same length. It is also interesting to compare this bond length with those
Fig. 10.12 The allyl anion can be formed in propene (Fig. 10.13); we see that the bond length in the anion is intermediate
by the action of a very strong base on between that of the C-C single and double bonds in propene. Since there are
propene.
only two electrons (these in 17) which are contributing to the bonding across
1.341 1.506
three atoms we expect the m bonding to be partial, which is just what the bond
length data indicate.
The other interesting thing about the ally] anion is how the electron density
differs between the carbons. MO Iz has the greatest orbital coefficient, and
hence the greatest electron density, on the central carbon. However, MO 27
puts the electron density on the end carbons and none in the middle. It takes a
more detailed calculation to work out which of these two wins — in fact it turns
out that the end carbons have the greatest z electron density.
Fig. 10.13 Comparison of the C-C bond To explain these properties of the allyl anion using localized structures, we
lengths (in A) of propene and the allyl
have to draw two resonance structures, as shown in Fig. 10.14. These structures
anion. The anion is symmetrical and has
a bond length intermediate between the convey the idea that the charge is delocalized and may be ‘found’ on either of
single and the double bond in propene. the terminal carbons; similarly, the double bond may be located between Cj
and C» or between C2 and C3. This picture is consistent with the MO approach
which, as we have explained, predicts a symmetrical structure with partial 7
bonding and maximum electron density on the end carbons. Sometimes the
anion is represented using a dashed line across all three atoms to indicate the
partial 2 bond; such a structure is also shown in Fig. 10.14.
HoC po CH> —- HC”HoySc, Ho

HHe~ SCH>
Fig. 10.14 On the left, connected by a double-headed arrow, are two resonance structures for the allyl
anion. On the right is an alternative representation in which the dashed line shows the partial x bond
across all three atoms. In this structure the minus signs in brackets (—) indicate the positions where
negative charge can be found, as shown by the resonance structures on the left.
It is interesting to compare the two resonance structures with the form of

the HOMO. Recall that this MO, the 277, has the electron density confined to
the end carbons. This is exactly where the negative charge is located in the
resonance structures. There are some parallels, therefore, between the form of
the HOMO and the charge distribution predicted by the resonance structures.
Carboxylate anion
Closely related to the ally] anion is the carboxylate anion, formed by removing
Ht from the OH group of a carboxylic acid, as shown in Fig. 10.15.
p p 2 20
R—c a8 p—-c¢ =—+ > «SOR-C R—C!
\ Ve \ \
OH O O 6 (-)
Fig. 10.15 A carboxylate anion is formed when a base removes Ht from a carboxylic acid. The 7 sys-
tem in the anion is delocalized and can be described using two resonance structures, or represented
using a dashed line to indicate a partial bond.
The resulting anion can be described in very similar terms to those used
for the allyl anion, the only difference being that instead of the 7 system being
formed from three carbon 2p orbitals it is formed from one on carbon and two
on oxygen. Surface plots of the three 7 MOs are shown in Fig. 10.16 — they
are very reminiscent of the cartoons shown in Fig. 10.11 on p. 158.
Fig. 10.16 Surface plots of the three 7 MOs of the methanoate ion, HCOO-. The molecule is viewed
sideways on with the two oxygens coming toward and the hydrogen going away from us. Note the
similarity between these MOs and the cartoons shown in Fig. 10.11 on p. 158. MO 2z is the HOMO
and 37 is the LUMO.
Of these MOs, only the 1m and 27 are occupied, so the conclusions about
the carboxylate are much the same as for the ally] anion, i.e. there is partial
m bonding across all three atoms, the negative charge 1s greatest on the two
oxygens and the two C—O bonds are equivalent. Experimental data confirm
that these two C-O bonds have the same bond length of 1.242 A. This value
is intermediate between a typical C=O bond length of 1.202 A and a typical
C-OH bond length of 1.343 A, confirming that there is partial 7 bonding across
all three atoms.
As is shown in Fig. 10.15, these features of the bonding in the carboxylate
ion can be represented either using resonance structures, or a dashed line to
represent a partial bond.
Amides
Studies of the structures of amides using experimental techniques such as X-
ray diffraction show us that the bonds to the carbonyl carbon, the nitrogen and
the two hydrogens attached to the nitrogen all lie in the same plane, as shown
in Fig. 10.17.
o sp?
\
oy :
eee Hyvol ZH
Nw
Vere sts
sp?-1s L
Fig. 10.17 Ail of the bonds within the shaded box of this amide structure are found to lie in the same
plane. This being the case, the a framework of methanamide is best described as being formed from
the overlap of sp* hybrids on oxygen, carbon and nitrogen. Two lone pairs occupy the remaining two
sp hybrids on oxygen.
Given the planarity of the amide, to describe the bonding we should choose
sp” hybrids for both the carbonyl carbon and the nitrogen. As we discussed on
p. 92 it does not matter too much whether we choose the oxygen to be sp or
sp? hybridized, so we will choose sp’. With this hybridization scheme the a
framework of methanamide is straightforward to form and is shown Fig. 10.17.
The planarity of the amide group means that the remaining three p orbitals
(one each on carbon, nitrogen and oxygen) are all perpendicular to the plane
and so in the correct geometry to overlap with one another to form 2 MOs.
The z system can be modelled as three p orbitals in a row but, unlike the ally]
anion, the AOs do not have the same energy. From Fig. 4.34 on p. 57 we expect
that the AO from oxygen will be lowest in energy, next will come the AO from
nitrogen and the highest will be the AO from carbon.
Figure 10.18 shows surface plots of these three 2 MOs; there are four elec-
trons in the 7 system, so these occupy MOs 17 and 27. The lowest energy MO
(17r) is bonding across all three atoms. At the level plotted, MO 27 shows vir-
tually no contribution visible from the carbon AO and so this MO contributes
little to the bonding. Finally, as in the allyl anion, 37 has two nodal planes
between the atoms.
As there is experimental evidence that amides invariably adopt this planar
geometry we conclude that this must have the lowest energy. The alternative
arrangement, in which the nitrogen is pyramidal and sp? hybridized with the
lone pair not involved in the 2 system must therefore be higher in energy. The
i
SOA A+ yy Brest
No t , KES Se
SSS er ososseay \ ee foe IY
ate TY
Saat wo >
Lee aS _—<ae
7 < ty &
¥
KAS
0
I
nt LC LH
H N
Fig. 10.18 Surface plots of the three 7 MOs from methanamide.
question is, therefore, why is it that participation of the nitrogen lone pair in
the m system lowers the energy?
Figure 10.19 offers an explanation of this by considering the effect of over-
lapping the nitrogen lone pair with the carbonyl 7 MOs. In (a) we see the
situation in a hypothetical molecule in which there is no interaction between
these orbitals. The AOs from nitrogen are intermediate in energy between those
of carbon and oxygen and so in the diagram we have placed the energy of the
nitrogen AO between that of the two z orbitals.
If we ‘switch on’ the interaction the resulting MOs are as shown in (b). As
noted on p. 156, the highest and lowest energy MOs must come, respectively,
above and below the highest and lowest energy AOs. The 27 MO falls quite
close in energy to the 2p on nitrogen, but we need a computer calculation to
find whether it is above or below the energy of the AO; arbitrarily, we have
shown it at slightly lower energy than the nitrogen 2p.
The pair of electrons which start in the carbonyl 2 MO are lowered in
energy as they end up in the 17 MO. Whether or not the overall energy is low-
ered on going from (a) to (b) also depends on how the energy of the lone pair
changes as it moves into the 27 MO. Since we know that the planar geometry
is favoured and that the feature of this geometry is the participation of the lone
pair in the 7 system, we conclude that (b) must have the lower energy. It could
CO x* CO n* f
1x
co N co N
Fig. 10.19 MO diagrams illustrating the interaction between the nitrogen lone pair and the x MOs of
the carbonyl group in an amide. In (a) is shown the situation before interaction; the nitrogen orbital
appears at an energy intermediate between the x and x* MOs on the grounds that the nitrogen AOs
lie in energy between those of carbon and oxygen. MO diagram (b) shows the case after interac-
tion between the orbitals; as drawn, the interaction results in a lowering of the overall energy of the
electrons.
be that, compared to (a), the lowering in energy of the 12 MO outweighs any

increase in the 27 or that the 277 MO is itself also lowered in energy.
The creation of this delocalized 2 system also has consequences for the
strength of the bonding. If we do not involve a 2p orbital from the nitrogen
in the x system there are just two electrons and these occupy the carbonyl z
bonding orbital which is just bonding the carbon and the oxygen. However,
if the nitrogen is involved, there are four electrons to consider. Two occupy
MO Iz which is bonding across’all three atoms, and two occupy 27 which
contributes little to the bonding. The overall effect is that the C—O bond in the
1.220. O 1.368 amide is weakened when compared to methanal as the two electrons in the z
_N
bond in methanal are effectively spread over three atoms in methanamide; in
other words, the C—O z bond in an amide is partial.
In contrast, the C-N bond is strengthened somewhat because occupation of
the 1z MO leads to some z bonding between the carbon and nitrogen which,
for example, is not present in molecules like methylamine. The partial 2 char-
acter of this bond also explains why it is more difficult to rotate about this
bond than for a simple o bond. The barrier to rotation is not as large as for a
full z bond (see p. 90), such as that in ethene, but it is large enough to have
consequences for the structures of molecules containing amide bonds. As we
discussed on p. 128, amide bonds are formed when amino acids are joined to-
gether to form proteins, and it turns out that the restricted rotation about the
Fig. 10.20 Bond length data, in A, for
C-N bond is a crucial factor in determining the three-dimensional structures of
ethanamide, methylamine and ethanal. proteins.
Most striking is the shortening of the C—N The bond length data given in Fig. 10.20 confirms these predictions. The
bond in the amide when compared to
C-—N bond length in the amide is considerably shorter than the straightforward
that in the amine — something we can
attribute to the partial bonding involving C-N single bond in methylamine. Also, there is some lengthening of the C-O
the nitrogen lone pair and the carbonyl bond in ethanamide when compared to ethanal, although the effect is not as
group.
marked as for the C—N bond.
In terms of localized structures, we can represent the participation of the
nitrogen lone pair in the 7 system by writing a resonance structure in which a
m bond is formed between the carbon and the nitrogen, as shown in Fig. 10.21.
a, Sy
R~ a a7 SS
. A B
Fig. 10.21 Resonance structures for an amide; in structure B the presence of a double bond between
carbon and nitrogen explains the observed planarity of the amide unit and the shortening of the C-N
bond.
We know that the C-N bond only has partial 7 character, so structure B
must be a minor contributor. In this structure, to form the C-N 2 bond we
have to break the C-O bond resulting in a negative charge on the oxygen.
The nitrogen no longer has the lone pair (it was used to form the C_-N x bond)
but shares these electrons with the carbon. Effectively it has lost one electron
and so becomes formally positively charged.
10.3 Chemical consequences - reactions of carbonyl

compounds
As we have seen in Section 7.3 on p. 102, the carbonyl group is susceptible to
nucleophilic attack leading, in the first instance, to a tetrahedral intermediate
as shown in Fig. 10.22. Changing the group X has a large effect on how readily
this reaction occurs and it is the explanation of this effect which we will focus
on in this section.
is)
0 Nu O
Nu I —_—_—_——— of
R~ ~ X R aN x
Fig. 10.22 Carbonyl compounds are attacked by nucleophiles (Nu7) to give, in the first instance, a
tetrahedral intermediate. The group X has a profound influence on how readily this reaction goes.
Amides and aldehydes

Amides and aldehydes show very different reactivity towards water as a nucle-
ophile. As shown in Fig. 10.23, aldehydes are readily attacked by the water to
form a tetrahedral intermediate which leads to what is known as a hydrate. In
contrast, amides show no perceptible reaction with water.
O ® ©
A LN Ig H20, {0 HO
~DP
OH
—_——_—_» _
Cc ow
H R~ ~H R H
R~ ~
hydrate
O
il HO
no reaction
31
Fig. 10.23 Aldehydes react readily with water to give significant amounts of the hydrate; amides do not
react with water. COn
The initial interaction in these reactions is between the HOMO of the nu-
cleophile (in this case an oxygen lone pair) and the LUMO of the carbonyl
compound. In the case of an aldehyde, we saw on p. 91 that the LUMO is the
z* MO formed from the overlap of 2p orbitals on the carbon and oxygen.
In the amide, the nitrogen lone pair is involved with the wz system of the
cont
carbonyl, forming three 2 MOs. We discussed how this interaction leads to the aldehyde
3a MO (the LUMO) being higher in energy than the carbonyl 2* MO, a point nucleophile
illustrated in Fig. 10.19 on p. 161.
Fig. 10.24 On the left are shown the w
Figure 10.24 illustrates the relative energies of the LUMO of the aldehyde, MOs of an aldehyde, in the middle the
the LUMO of the amide and the HOMO on the nucleophile. Recall that as the HOMO of a nucleophile and on the right
the z MOs of an amide. Possible
HOMO is a lone pair on oxygen, we expect it to lie lower in energy than the z*
HOMO/LUMO interactions are shown by
LUMO (see p. 95). We can clearly see from this diagram that the HOMO of the the wavy lines. Due to the interaction of
nucleophile has a much better energy match with the LUMO of the aldehyde the carbonyl x system with the nitrogen
lone pair, the LUMO of the amide (37,
than it does with the LUMO of the amide; we therefore expect the aldehyde to
shown in the shaded box) is higher in
be the more reactive of the two species, which is precisely what is found. energy than the LUMO of the aldehyde
In summary, the presence of a nitrogen substituent on the carbonyl leads to (the CO x* also shown shaded). The
aldehyde is more reactive towards the
reduced reactivity towards nucleophiles — we can attribute this to the raising in
nucleophile than is the amide because
energy of the LUMO which results from the participation of the nitrogen lone the LUMO of the aldehyde is closer in
pair in the carbonyl z system. The effect is often called the conjugative effect. energy to the HOMO of the nucleophile.
Amides as bases
Before moving on to other carbonyl reactions, it is interesting to look at the
behaviour of an amide when it acts as a base, since it turns out that this is also
affected by the participation of the nitrogen lone pair in the x bonding system.
As you will recall, a base is a compound which can accept (‘pick up’) a
proton from another molecule. It is usually an atom with a lone pair which is
involved in bonding to the incomingsH*; for example, in the case of amines
and alcohols the Ht becomes bound to the nitrogen and oxygen, respectively:
RNH2 + Ht —> RNH}

ROH + H+ —> ROH;.
As we saw on p. 106, in these reactions the HOMO is a lone pair and the
LUMO is the 1s orbital on the H*. Amines turn out to be much more basic
than alcohols, meaning that amines are more ready to pick up a proton; an
explanation for this in terms of orbitals is as follows.
In these reactions, the HOMO is a lone pair on nitrogen or oxygen and
these orbitals clearly lie lower in energy than the LUMO of the Ht. However,
the oxygen lone pair is lower in energy than the nitrogen lone pair, and so
the energy separation between the HOMO and LUMO is greatest for oxygen.
There is thus a poorer interaction with the lone pair on oxygen compared to
that on nitrogen, resulting in the alcohol being the weaker base.
Given that this is the case, it is at first sight surprising that when an amide
acts as a base it is the carbonyl oxygen which is protonated first, not the nitro-
gen (Fig. 10.25). In other words, the carbonyl oxygen is more basic than the
nitrogen.
‘0 SN ‘0@_H
”
Fig. 10.25 In an amide the carbonyl oxygen is more basic than the nitrogen.
To understand what is going on here we first need to identify the HOMO of

the amide. There are two candidates: either the lone pair on the nitrogen, or one
of the lone pairs on the oxygen. As we discussed before, the carbonyl oxygen
can be considered as sp” hybridized and two of these hybrids are occupied by
lone pairs; these hybrids lie in the plane of the molecule and therefore are not
involved in the z system.
From what we know about the energies of AOs (see Fig. 4.34 on p. 57) we
would expect the oxygen lone pair to be lower in energy than that on nitrogen,
so the latter would form the HOMO. However, as we have seen the interaction
of the nitrogen lone pair with the carbonyl z system results in a lowering of the
energy of the lone pair as it moves into the 27 MO (see Fig. 10.19 on p. 161).
In the case of the amide, the nitrogen lone pair is so lowered in energy that it
comes below the oxygen lone pairs thus making the latter the HOMO. So the
best energy match to the LUMO of the proton is the lone pair orbital on the
oxygen.
We can also formulate an explanation of this effect using resonance struc-

tures. In Fig. 10.21 on p. 162 it is illustrated how the lone pair can become
involved in the formation of a C-N z bond. This decreases the ‘availability’ of
the nitrogen lone pair, thus reducing the basicity of the nitrogen.
Amides and acy! chlorides

As we have seen in Section 7.6 on p. 110 acyl chlorides react very rapidly with
nucleophiles, usually leading to subsequent elimination of CI~, whereas, as we
have already noted, amides are rather unreactive. The reaction, or lack of it,
with water (shown in Fig. 10.26) offers a clear demonstration of this difference.
“ 0 H,0
2 O° fo)
Hd VIE RE coal \ / —_—>—s_— | + ce’
R~ No fast R~ Cc No R~ Cc SoH
O
Cw
I H20
——_—s no reaction
R NHs
Fig. 10.26 Acyl chlorides react rapidly with water as a nucleophile, ultimately generating a carboxylic
acid and chloride ions; amides are unreactive to water.
An explanation you will often find given for the high reactivity of the acyl
chloride is that the chlorine is electron-withdrawing and so this increases the
partial positive charge on the carbonyl carbon, making it more reactive towards
nucleophiles. The problem with this argument is that nitrogen is also electron-
withdrawing; in fact chlorine and nitrogen have about the same electronegativ-
ity, so if the presence of an electron-withdrawing chlorine substituent increases
the reactivity of the carbonyl, a nitrogen substituent should do the same — which
plainly it does not.
To sort out what is going on here we have to realize that the presence
of the substituent X in RCOX has two effects. The first is that the electron-
withdrawing nature of X increases the positive charge on the carbonyl carbon
thus increasing its reactivity toward nucleophiles; this is called the inductive
effect. The second is the conjugative effect described on p. 163; in this the
delocalization of a lone pair from X into the carbonyl 2 system decreases the
reactivity of the carbonyl towards nucleophiles by raising the energy of the 2*
LUMO.
As nitrogen and chlorine have about the same electronegativity, and hence
presumably the same inductive effect, we must look to the conjugative effect to
explain the difference in reactivity between the acyl chloride and the amide. It
must be that the conjugative effect (which decreases reactivity) is much greater
for the amide than for the acyl chloride.
The explanation for this is that the orbital involved in forming the delocal-
ized system is a 3p in chlorine and a 2p in nitrogen. The 3p is much larger
than the 2p (see p. 49) and so the overlap with the 2p orbitals which form the
carbonyl m system is poorer (see p. 67) for chlorine than it is for nitrogen.
1@)
\
C—Cl
H
Fig. 10.27 Surface plots of the x MOs of methanoyl chloride; the molecule is viewed from the side with
the C-Cl bond going to the right and the C—O bond going away from us and to the left. Note how there
is little contribution from the chlorine 3p orbital to the 12 MO, indicating that the chlorine lone pairs are
not participating in the 7 bonding.
We have shown the surface plots for the

simplest acyl chloride, methanoyl
The surface plots of the x MOs in methanoyl chloride (HCOCI), shown in
chloride. However, it should be noted that Fig. 10.27, support this explanation. The 12 MO results in bonding between
this molecule is only stable at low the carbon and the oxygen, but this surface plot shows no contribution from
temperatures and is not usually
the chlorine. The 277 MO is mainly just the 3p AO on chlorine. Compare these
encountered as a laboratory reagent.
orbitals with those for methanamide (Fig. 10.18 on p. 161) which show the
nitrogen orbital participating in the 7 system.
Overall, it is clear that the chlorine orbitals are not affecting the bonding
in the 7 system and so the conjugative effect is minimal. This, then, is the
explanation for the high reactivity of the acyl chloride.
In summary, there are two effects, illustrated in Fig. 10.28, of the sub-
stituent X on reactivity:
e the inductive effect involves the electron-withdrawing nature of X in-

creasing the positive charge on the carbonyl] carbon and hence increasing
its reactivity;
e the conjugative effect involves the delocalization of the lone pair from X
into the carbonyl system thereby raising the energy of the LUMO and so
reducing the reactivity.
O O o
il Ci |
R~ SX R~ x Aw Sx
inductive effect conjugative effect
Fig. 10.28 An electronegative substituent X has two effects: firstly, it withdraws electrons from the
carbonyl carbon, thus increasing its reactivity (inductive effect); secondly, its lone pair is delocalized
into the carbonyl 7 system, thus decreasing its reactivity (conjugative effect).
Amides and esters

The reactivity of esters falls between that of amides and acyl chlorides: an ester
is readily hydrolysed by refluxing with dilute alkali, whereas the hydrolysis
of an amide requires much stronger conditions, i.e. higher temperatures and
longer times. Both reactions start with nucleophilic attack by hydroxide on the
carbonyl, as shown in Fig. 10.29.
10.3. Chemical consequences — reactions of carbonyl compounds 167
O HO O° O
HO NP
Cw ——> \/
on > > me
T + R'OH
R OR' R OR' R~ oO
e) HO O°
HO E\I-P Cc a Cc
/ —_> —_>
T
Cc + NH3
R
a“ae™N
NH> R~ ~NHo a7 Ne
Fig. 10.29 Both esters and amides are hydrolysed by aqueous alkali and both reactions are initiated
by nucleophilic attack on the carbonyl by hydroxide. However, whereas the ester is readily hydrolysed
under mild conditions, hydrolysis of the amide needs ‘stronger’ conditions, i.e. prolonged heating.
The lone pair on the oxygen in an ester can interact with the carbonyl 2
orbitals to form a delocalized system, just as in the case of an amide. The
result is that, as was illustrated in Fig. 10.19 on p. 161, the LUMO is raised in
energy thus reducing the reactivity of the carbonyl group.
Figure 10.30 compares the effect of conjugation with a nitrogen versus an
oxygen lone pair; as we have noted before, the lone pair on the oxygen is lower
in energy than that on the nitrogen, as is shown in (b). Clearly, the nitrogen
lone pair has a better energy match with the carbonyl 2* MO than does the
oxygen, so we expect the 32 MO to be raised in energy more in the case of the
amide than of the ester. Since it is this raising in energy of the LUMO which is
responsible for the reduced reactivity of the amide, we conclude that the ester
should be more reactive than the amide.
The inductive effect will also point to the ester being more reactive than
the amide as the oxygen is more electron-withdrawing than the nitrogen. Thus,
both the conjugative and inductive effects work together to make the ester more
reactive than the amide, which is what we observe.
However, esters are still less reactive than acyl chlorides despite the induc-
tive effect of the electronegative oxygen substituent. We therefore conclude
that there must be significant delocalization of the oxygen lone pair in the ester
in order to account for this reduction in reactivity.
(@) 3g (b)
‘ . . 3n
CO n* { 5 CO n* - 5
4 2n 5 4 -
f i nH “——— 2p
COn : ; COn : i
co
Hh N oe)
4U O
amide ester
Fig. 10.30 MO diagrams showing the interaction between the lone pair on nitrogen or oxygen with the
a system of the carbonyl group. Diagram (a) is identical to that of Fig. 10.19 and is appropriate for an
amide; note how the interaction with the lone pair raises the energy of the LUMO (37). Diagram (b)
shows how the situation is modified for an ester; the oxygen lone pair is lower in energy than that on
nitrogen, and so has a poorer energy match with the 7* MO, resulting in a smaller rise in the energy
of the LUMO (37). On this basis, we wouid predict that an ester will be more reactive than an amide.
C-O bond lengths in carbonyl compounds

When discussing the effect of conjugation of the nitrogen lone pair in amides
we noted that good evidence for this was the shortening of the C—N bond length
(Fig. 10.20 on p. 162). Now that we have discussed a range of other carbonyl
compounds it is interesting to compare their C—O bond lengths and see how
they can be rationalized in terms of the bonding models we have developed.
The data are given in Fig. 10.31. ™
1.187% O 1.206. O 1.213. 0 1.220. 0
“i “i ™T “I
C C C om
HsC~ Cl H3C~ OCH3 H3C~ ~CH3 ~—-H3C NH»
conjugative effect becoming dominant
inductive effect becoming dominant
Fig. 10.31 C—O bond lengths, in A, for a series of different carbonyl compounds.
The amide shows the longest C-O bond and we have already described
(p. 162) how this can be attributed to the involvement of the nitrogen lone pair
in the carbonyl z system. The ester has a shorter bond than the amide, and this
is consistent with the view that the oxygen lone pair participates less in the z
system than does the nitrogen lone pair, as was described on p. 167.
The acyl chloride has the shortest bond, and we can rationalize this by
recalling from p. 165 that there is very little participation by the chlorine lone
pair in the z system, and hence none of the weakening of the C-O bond seen
in amides. The z bond is therefore just a 2c-2e bond between the carbon and
the oxygen.
However, the story is a little more complicated than this, as we see that the
C—O bond in the acy] chloride is actually shorter than the bond in the ketone. In
this molecule there are no lone pairs to be involved in the z system, SO we can
take the C—O bond length in the ketone to be unaffected by conjugation. What
is going on here is that the inductive effect is responsible for the shortening of
the C—O bond; we will leave the explanation of why this is to the next chapter
where the story is taken up again in Section 11.2 on p. 176.
In summary, the trend of bond lengths shown in Fig. 10.31 is nicely ex-
plained as resulting from a balance between the conjugative effect, which leads
to lengthening of the bond, and the inductive effect, which leads to shortening
of the bond.
10.4 Stabilization of negative charge by conjugation

In water, a carboxylic acid dissociates to a significant extent, giving H30* and
the carboxylate anion; in contrast, simple alcohols dissociate to such a small
extent that they are not thought of as even weak acids (Fig. 10.32).
In both cases it is an O-H bond which is being broken, but the crucial
difference is that in the carboxylate ion the negative charge is conjugated with
the carbonyl m system and is therefore delocalized on to both oxygens. No
such delocalization is possible for the anion formed from the simple alcohol.
4 Pg o ®
R— q + HsO = _—R-C ‘ + H30
OH \y (-)
R-OH + H0 === R-O + 4H,0°

Fig. 10.32 Carboxylic acids dissociate quite readily in water to give the delocalized carboxylate ion
{see Fig. 10.15 on p. 159). In contrast, simple alcohols hardly dissociate at all; we can attribute the
difference to the extra stability of the delocalized carboxylate anion.
It seems clear, therefore, that the delocalization possible in the carboxylate ion
is responsible for stabilizing this ion. In this section we will explore why this
is SO.
The origin of this stabilization is that the lone pair of electrons on the oxy-
gen (formally the negative charge) becomes involved in the carbonyl 7 system.
If we concentrate on just the p orbitals of the carbon and the two oxygens we
can draw up a simple MO scheme, shown in Fig. 10.33, which explains how
this stabilization comes about.
(a) bb) ag
CO n* CO n* ‘ 7
+ 2p 2n oo 2p
co os CO n i i
Cc=0 O- C=O Oo
Fig. 10.33 Illustration of how the interaction of a negative charge, located in an oxygen 2p orbital, with
an adjacent carbonyl! x system results in a lowering of the energy. Diagram (a) shows the situation
with no interaction and (b) shows what happens when the orbitals interact; there is a decrease in the
total energy of the occupied orbitals. The exact energy of the 27 MO is somewhat uncertain but it is
expected to be close in energy to the oxygen 2p.
In Fig. 10.33 (a) we are supposing that there is no interaction between the
two 7 MOs and the out-of-plane 2p orbital. This orbital has two electrons in it,
as both electrons from the O-H bond are left behind on the oxygen when the
bond breaks to give Ht. In (b) the 2p orbital is now allowed to interact with
the x MOs and this gives rise to three MOs; in fact what we have here is three
2p orbitals in a row, so the three MOs are similar to those given in Fig. 10.16
on p. 159.
Although the pair of electrons originally in the 2p orbital do not change
energy very much as they move into the largely non-bonding 27 MO, the other
pair of electrons are lowered in energy as they move into the I MO. Overall,
therefore, the interaction leads to a lowering of the energy.
It is interesting to compare the carboxylate ion with two ions in which the
oxygens are replaced by CH? groups: replacing one oxygen gives an enolate
anion and replacing two gives an allyl anion which we have already described
on p. 157.
The enolate anion is generated by treating an aldehyde or a ketone with a

strong base; this can remove one of the protons on a carbon adjacent to the
. carbonyl, leaving a negative charge as shown in Fig. 10.34 (a). As in the car-
boxylate ion, the charge is delocalized arid this is illustrated in the diagram by
the use of resonance structures. Removing a proton from a carbon adjacent
to a C=C double bond also gives a delocalized anion (an allyl anion) which iS
shown in Fig. 10.34 (b).
ty
(a) fe) C)
I oO
EX Nowsa t
stron © H Cx! yp H
R base R7 ~e" “~~ R “6
H H
enolate
(b) 5
Ve CHp CHe
ox aH
aa very ry stron
strong Clo H . H
R ’ base R~ ~C~ R~ Sc
H |
H allyl anion
Fig. 10.34 An enolate anion is formed by removing one of the protons on a carbon adjacent to a
carbonyl group, as shown in (a). Removing one of the protons from a carbon adjacent to a C=C
double bond gives an allyl anion, as shown in (b). Both anions are delocalized, but the enolate has
a resonance form in which the negative charge appears on an electronegative element, oxygen. This
confers extra stability and so explains why the enolate anion is much easier to form than the allyl anion.
Comparing the two anions, we can expect that the enolate will be lower
in energy as the negative charge can be delocalized onto the electronegative
oxygen atom. Although the allyl anion is also delocalized, there are no elec-
tronegative atoms onto which the charge can be delocalized. This explains why
we need a much stronger base to form the allyl anion than is needed to form
the enolate.
We have a nice progression here involving these three anions. The car-
boxylate, in which the negative charge can be delocalized onto two oxygens, is
the most stable of the three. Next comes the enolate, in which delocalization
onto one oxygen is possible, and finally the allyl anion in which there are no
electronegative atoms.
Stabilization by larger 7 systems

If the negative charge is conjugated with a more extensive 7 system then the
charge can be delocalized onto more atoms and this is found to confer extra
stability. In general, the more delocalized the charge becomes, the greater the
stabilization.
A good example of this is the anion formed by the ionization of the O-H in
phenol; as shown in Fig. 10.35, the resulting negative charge can be delocalized
round the benzene ring. The extra stability which this delocalization gives the
anion accounts for the observation that simple alcohols (such as ethanol) are
not acidic to a significant extent whereas phenol is a weak acid (indeed, it
used to be known as carbolic acid). The anion formed when a simple alcohol
dissociates is not delocalized, whereas that from phenol is.
OH 0° O O O
(Ss) (S)
ic)
Fig. 10.35 The phenolate anion, formed by the dissociation of phenol, shows extensive delocalization
of the negative charge into the benzene ring. This confers extra stabilization and so accounts for
phenol being a weak acid, in contrast to simple alcohols which are not significantly acidic.
Figure 10.36 shows the HOMO of this phenolate anion. Comparing this to
the resonance structures shown in Fig. 10.35 we see that the carbons on which
the negative charge appears are precisely the ones whose 2p orbitals contribute
to the HOMO.
Fig. 10.36 Surface plot of the HOMO of the phenolate anion, whose resonance structures are shown
in Fig. 10.35; the oxygen is on the left, coming towards us. Note how the atoms whose AQOs contribute
significantly to this MO are the same as those onto which the negative charge can be delocalized.
11 Substitution and elimination
reactions
ny
At the end of Chapter 9 we described that when a nucleophile replaces the

bromine atom in t-butyl bromide, (CH3)3CBr, the rate is found by experi-
ment to be first order and independent of the concentration of the nucleophile
rather than second order (i.e. proportional to both the concentrations of the
bromoalkane and the nucleophile) as is observed with bromomethane:
first order:
(CH3)3CBr + Nu~ —> (CH3)3CNu+ Br rate = k[(CH3)3CBr]
second order:
CH3Br + Nu~ —> CH3Nu + Br- rate = k'[(CH3Br][Nu_].
We are now going to look more closely at these and other reactions using some
of the ideas from Chapter 10 to understand why the rate laws are different.
11.1. Nucleophilic substitution revisited - Sy1 and Sy2

We have seen in Sections 7.5 on p. 107 and 9.1 on p. 131 how bromomethane
and hydroxide react together in a single step to form the products. For the
moment let us simply assert that the reaction between a nucleophile and t-butyl
bromide proceeds via a completely different two-step mechanism. In the first
step, the carbon—bromine bond breaks to form a bromide ion and a reactive
species, called a carbocation or carbenium ion, in which a carbon atom bears a
positive charge. The curly arrow mechanism for this step is shown in Fig. 11.1.
It is essentially the reverse of two ions reacting to form a neutral molecule as
illustrated on p. 97 by the reaction of Ht and H7 ions.
(CHg)3C Lt Br (CHg)sC ® Br
Fig. 11.1 The formation of a carbenium ion from t-butyl bromide.
The carbenium ion is very reactive and in the second step it quickly reacts
with the nucleophile to give the product. The curly arrow mechanism for this
step is shown in Fig. 11.2.
©
(CH3)3C Nu ———» (CH3)3C —Nu
Fig. 11.2 The reaction between the carbenium ion and the nucleophile.
11.1 Nucleophilic substitution revisited — Sy 1 and Sx2 173
transition
state 1
energy
transition
> state 2
intermediate
i"
carbenium ion
Hgcyon Nu®
H3C Br
reactants products
i
Bre HaCyN
H3C
—_
reaction coordinate
Fig. 11.3 An energy profile for the reaction between f-butyl bromide and a nucleophile, Nu~. The
reaction occurs in two steps: the first is the formation of the carbenium ion intermediate; the second
step is the reaction between the carbenium ion and the nucleophile to give the product. The second
transition state is lower in energy than the first.
Figure 11.3 shows an energy profile for the overall reaction. As the reaction
between the positively charged carbenium ion and the nucleophile has such a
small activation energy (F,2), the initial formation of the carbenium ion from
the t-butyl bromide is the rate-determining step (see Section 9.6 on p. 141).
The rate of step (1), the rate at which the intermediate carbenium ion forms,
depends only on the concentration of the t-butyl bromide — the more t-buty]
bromide there is, the faster the carbenium ion will be formed. We can write the
rate of formation of the carbenium ion as:
rate of step (1) = k;[(CH3)3CBr].
As soon as it has been formed, the carbenium ion reacts with the nucle-
ophile to give the products. Since the first step is rate-determining, it follows
that the rate of formation of the product is the same as the rate of step (1):
rate of formation of the product = k[(CH3)3CBr].
This mechanism therefore explains the observed first-order kinetics of the re-
action of t-butyl bromide.
We have two possible mechanisms for these nucleophilic substitution reac-
tions. The reaction between bromomethane and the nucleophile depends on the
concentrations of both reagents and proceeds by both reagents coming together
in a single step. This mechanism is known as an Sn2 mechanism: this stands
for Substitution, Nucleophilic with the ‘2’ indicating that the rate-determining
step involves both reagents and so the rate law is second order.
174 Substitution and elimination reactions
(a) (b)
Nu— Nu ae
Fig. 11.4 Diagram (a) shows that the three hydrogen atoms in bromomethane do not hinder the ap-
proach of a nucleophile in the correct direction to attack the C—Br o*. In contrast, (b) shows that the
methyl groups in t-butyl bromide prevent the nucleophile approaching. The dots in each picture are
used to outline the approximate atomic radii of each of the atoms.
The reaction between t-butyl bromide and the nucleophile proceeds in two
steps — the initial formation of a carbenium ion from the t-butyl] bromide fol-
lowed by the fast reaction of this ion with the nucleophile. This mechanism is
known as an SnI mechanism, the ‘1’ indicating that the rate-determining step
only involves one species i.e. this step is unimolecular or first order.
The question we now need to address is why is there this change in mech-
anism? Why does bromomethane not follow the Sj 1 mechanism and form the
cation CH}? Why does f-butyl bromide not react directly with the nucleophile
following the Sy2 mechanism?
The latter question is relatively easy to answer. We saw in Section 7.5
on p. 109 that in a reaction following the Sn2 mechanism, the attacking
nucleophile must approach from directly behind the bromine leaving group.
Whilst this is possible for a nucleophile attacking bromomethane (as shown in
Fig. 11.4 (a)), for t-butyl bromide, this approach is severely hindered by the
three methyl groups as shown in (b).
The question as to why bromomethane does not undergo an Sn 1 reaction
has a more complex answer. Put simply, the answer is that the carbenium ion
that would be formed from bromomethane, H3C*, is so high in energy that it
cannot readily form. This contrasts with the (CH3)3C* ion which is much more
stable. It turns out that the methyl groups help to stabilize the positive charge
in the (CH3)3Ct ion whereas no stabilization is provided by the hydrogens in
H3C?. Exactly how positive charges in ions may be stabilized in this and other
ways is the subject of the' next section.
11.2 The stabilization of positive charge

We have seen in Section 10.4 on p. 168 how negative charge can be stabilized
by delocalization, i.e. where the electrons are not confined to one atom but are
spread over many in a delocalized MO. Positive charges can be stabilized in
a similar manner but there is one important difference: positive charges them-
selves do not move. To understand this distinction we need to think carefully
about what negative and positive ions are.
A negative ion is simply a molecule (or atom) in which the number of elec-
trons exceeds the total nuclear charge. For example, in BH, the total nuclear
charge is nine (five for boron, plus a total of four from the hydrogens) but there
are ten electrons, giving an overall negative charge of one. Similarly, a posi-
tively charged ion is one in which there are insufficient electrons to compensate
for the nuclear charge. For example, in NH7 the total nuclear charge is 11, but
there are just 10 electrons. In a negative ion we can imagine the electrons mov-
ing from one atom to another as a result of delocalization — for example, in the
enolate anion shown in Fig. 10.34 (a) on p. 170 the negative charge can be on
the carbon or the oxygen.
In a positive ion the charge may also appear on different atoms. However,
this is not due to the positive nuclei moving but is due to the movement of the
negative electrons. The positive charge only appears to move as a result of the
actual movement of electrons.
As with negative charges, the formation of a positively charged species is
not generally favourable unless there are special factors to stabilize it, such as
delocalization; it is this stabilization which we will discuss in this section.
The structure of CHT

On p. 86 we described the bonding in the planar trigonal molecule BH3 using
sp” hybridization of the boron. There are three 2c-2e bonds formed by these
hybrids, leaving an empty 2p orbital pointing out of the plane. The species
CH} is isoelectronic with BH3, meaning that it has the same number of elec-
trons; however, as the nuclear charge is one greater, the molecule has an overall
positive charge.
We can therefore use the same description for the bonding in CHj as we
used for BH3; so the LUMO of CHy is a vacant 2p orbital, shown schemati-
cally in Fig. 11.5. We must not be tempted to say that the ‘positive charge is in
the empty 2p orbital’; electrons occupy orbitals, not positive charges. ho
His
H
We saw in Section 7.2 starting on p. 98 that the presence of the empty 2p
orbital in BH3 (the LUMO) makes it susceptible to attack by a nucleophile
which donates electrons from its HOMO into the BH3 LUMO. The same is Fig. 11.5 The LUMO of CHj is the
true for CHT, but even more so, as the positive charge further increases the out-of-plane 2p orbital.
interaction with the nucleophile, especially if it is also charged. As a result,

CH} is a highly reactive species which, if it could form in solution, would
only have a fleeting existence.
Stabilization by adjacent lone pairs

The interaction of CHF with the HOMO of the nucleophile gives us a clue as
to how the positive charge on carbon can be stabilized. Suppose that we make
a pair of electrons available not from the HOMO of a nucleophile but from an
atom attached to the carbon; to be specific, let us replace one of the hydrogen
atoms with O~. Such a replacement gives us the rather odd looking species
+CH2-O7 depicted in Fig. 11.6 (a). We can imagine that the negative charge
on oxygen is located in a 2p orbital, which is of course adjacent to the empty
2p orbital on carbon. These two orbitals will interact to give a 7 MO — shown
in (b) — which is occupied by the two electrons. In other words, a 7 bond is
formed. The two representations (a) and (b) are simply resonance structures of
the same molecule, methanal, CH2=O as shown in (c).
The MO picture of this interaction is shown in Fig. 11.7; as we have noted
before, the lone pair on oxygen is lower in energy than the carbon orbital.
Interaction with the carbon orbital results in a 2 bonding MO which is lower in
(a) (b) (c)

filled O 2p orbital ‘
empty C 2p orbital
Fig. 11.6 Replacing one of the hydrogen atoms in CH with an O- gives the odd-looking species
shown in (a). The two 2p orbitals interact to form the x MO shown in (b); this is occupied by the two
electrons to give a x bond. The two forms are actually resonance structures of the same molecule,
methanal, H2C=O, as shown in (c).
energy than the original orbital on oxygen. The lone pair of electrons move into
this orbital, thus leading to an overall lowering of the energy. We see, therefore,
how having an adjacent lone pair leads to stabilization of the positively charged
carbon.
Using a lone pair from a negatively charged oxygen is an odd way to stabi-
lize a positive charge on carbon — as we have seen, the net result is the forma-
! .C-O x*
tion of a neutral molecule with no positive charge at all. However, this example
highlights that the stabilization results in the positive charge being no longer
; i Le localized on the carbon, an idea which we will continue to explore.
ot
of 5 i
The stabilizing lone pair does not need to come from a negatively charged
atom; exactly the same effect is observed if an -OH group is used in place of the
tho rt O~. At first glance, it might be tempting to say that the electronegative oxygen
would destabilize an adjacent positive charge; this is not the case. Rather it
Fig. 11.7 MO picture showing how the
energy of the system is lowered by the is because the oxygen has relatively high-energy lone pairs which can interact
interaction of a lone pair on oxygen with an empty orbital in such a way that the charge is stabilized.
(shown on the left) with an empty orbital
As Fig. 11.8 shows, a better way of drawing this species is as a protonated
on carbon (shown on the right).
carbonyl. Whilst the unstabilized carbenium ion CH} is far too reactive to
exist in aqueous solution, protonated carbonyl cations are formed when acid is
added to a solution of the carbonyl compound.
Experimental evidence confirms that the better representation of the cation
vow a“ is as the protonated carbonyl compound with the positive charge on the oxygen
R R R R
Fig. 11.8 An OH group stabilizes an
rather than on carbon. In forming the new C—O z bond, the donation of the
adjacent carbenium ion. A better oxygen electrons lowers the energy of the system just as we saw in Fig. 11.7.
representation of this ionis as a
protonated carbonyl.
The acylium ion
We are now in a position to understand why the C=O bond in an acyl chlo-
ride, such as ethanoyl chloride, is so short, as mentioned on page 168. If
ethanoyl chloride is mixed with silver tetrafluoroborate in an inert solvent such
as nitromethane (CH3NOz) at low temperatures, silver chloride precipitates out
leaving a solution of the salt [CH3CO]*
BF, :
CH3COC] + AgBF4y —> [CH3CO]* BF, + AgCl(s).

The reaction is driven by the precipitation of the insoluble silver chloride. It
might be tempting to think that the structure of the cation is just like ethanoyl
Fig. 11.9 A possible mechanism for the chloride minus the chloride ion as shown in Fig. 11.9. However, the crystal
formation of the acylium cation. structure of the salt has been determined and the cation turns out to be linear
11.2. The stabilization of positive charge 177
with a C—O bond length of 1.108 A, notably shorter than the C-O bond length
in ethanoy! chloride at 1.187 A.
Once the chloride ion has departed, there are only two atoms or groups
around the carbonyl carbon in [CH3CO]*, the methyl group and the oxygen;
it should therefore be no surprise that these are at 180° to each other. We now
have just the situation described above with a positive charge, nominally on the
carbon atom, and an adjacent oxygen atom. The oxygen atom can stabilize the
positive charge by donating a pair of electrons. This is shown in Fig. 11.10.
‘. @ @
O=C—CHs3 =~, :0=C—CH3
Fig. 11.10 The positive charge on the carbon is stabilized by the donation of a pair of electrons from
the adjacent oxygen atom.
The donation of the lone pair from the oxygen formally creates a triple
bond between the oxygen and carbon which explains the shortening of the
bond length. The acylium ion plays an important role in a number of reactions
such as the acylation of aromatic rings.
The way in which the bond length is shortened in the acylium ion allows
us to understand why the CO bond in ethanoy] chloride is shorter than in other
carbonyl compounds. Whilst ethanoy] chloride does not exist as the separate
acylium and chloride ions, the chlorine is electronegative and withdraws some
electron density towards itself, increasing the partial positive charge on the car-
bony] carbon as mentioned on p. 168. This in turn brings down some electron
density from the oxygen lone pairs, strengthening the CO z bond as shown in
Fig. 11.11. Fig. 11.11 The electronegative chlorine
withdraws electrons from the carbon,
Comparisons of the CO bond lengths in propanone, ethanoy! chloride and thereby making it slightly more positively
fluoride and the acylium cation are shown in Fig. 11.12. We have already de- charged and this, in turn, brings down
scribed how the electron-withdrawing chlorine leads to a shortening of the CO more electron density from the oxygen,
thereby strengthening the C=O m bond.
bond. Changing chlorine to the even more electronegative fluorine results in
further shortening of the bond. Finally, in the acylium cation, there is essen-
tially a triple bond between the carbon and oxygen, giving a further reduction
in bond length.
12138. Oo 1187. oO 1160. 0 1.108 0°
C
HsC~ ~CHs HsC~ Cl HsC~ F LL
3
Fig. 11.12 Along the series propanone > ethanoyl chloride — ethanoyl fluoride the partial positive
charge on the carbonyl carbon increases leading to a decrease in the CO bond length (shown in A).
The acylium cation with its full positive charge has one of the shortest CO bond lengths known.
Stabilization by adjacent 7 systems

Other atoms with lone pairs, such as nitrogen, can stabilize positive charges |
in the same way that oxygen does but if no lone pair is available, then other ANE"
filled orbitals may help to stabilize the positive charge. We saw on p. 96 that |
generally after a lone pair, the next highest energy orbital, and hence the next H H
best for stabilizing a positive charge, is a x MO. An example of stabilization Fig. 11.13 The allyl cation. The positive
by such an MO is the allyl cation shown in Fig. 11.13 where the positive charge charge is stabilized by the adjacent 7
bond.
is stabilized by a neighbouring C=C.
3n
oe C-Cnt sO
OO C-Cr me
Fig. 11.14 MO diagram showing how the positive charge on carbon in the allyl cation is stabilized by
interaction with the adjacent x system. The C=C and x* MOs interact with the vacant p orbital
on the adjacent carbon to give three MOs. The lowest, 17, is lower in energy than the combining
orbitals. Since this is the only orbital occupied, the energy of the system has been lowered. MO 27 is
non-bonding and 37 is anti-bonding.
In the allyl cation we have a 2 system formed from three p orbitals: two can
be thought of as originating from the C=C z bond and one is the vacant orbital
on the adjacent carbon. We have already described in Fig. 10.11 on p. 158 the
three MOs which arise from the interaction of these three 2p orbitals; the MOs
are depicted once more in Fig. 11.14. In this figure we imagine that the three
MOs arise from the interaction of the C=C 2 and 2* MOs with the 2p AO on
the adjacent carbon. For example, the 12 MO can be thought of as arising from
an in-phase interaction between the 2p and the C=C m MO. Overall though, the
final result is just the same as for three p orbitals overlapping in a line.
Of these MOs, only the Iz is occupied and this clearly results in a lowering
of energy as the electrons move from the C=C m2 MO. The interaction with the
adjacent w system therefore results in a lowering of the energy — we say that
the positive charge is stabilized. It is interesting to note that it is the electrons
in the 7 system which are lowered in energy by the interaction with an empty
orbital, thus stabilizing the molecule.
In Fig. 10.14 on p. 159 we saw how the allyl anion could be represented by
two different resonance structures or a delocalized structure. The allyl cation
may be represented in an analogous manner as shown in Fig. 11.15.
H H H
Ce
=2a™~
=o
@/C (+) LC.C (+)
Hc“ f CH Hoo™ CH, HoC*~ a CH
Fig. 11.15 On the left, connected by a double-headed arrow, are two resonance structures for the allyl
cation. On the right is an alternative representation in which the dashed line shows the partial x bond
across all three atoms; the atoms onto which the positive charge can be delocalized are indicated by
(+).
Stabilization by adjacent o bonds - o conjugation

If there are no lone pairs or 2 bonds adjacent to the positive charge, the cation
can still be stabilized by the electrons in neighbouring a bonds. This type of
stabilization is called o conjugation.
In order for the o bond to help stabilize the cation, it must be in the correct
orientation. We saw how lone pairs and bonds stabilize the positive charge
when their occupied MOs overlap with the vacant p orbital on the carbon;
effective overlap can only occur if the orbitals involved are in the same plane.
The same is true for 0 conjugation — the o MO must be in the same plane as
the vacantp orbital as shown in Fig. 11.16 (a).
(@) filled C-H o orbital (b)

®\ C) constructive
CQ
y's: Goce,
destructivea)
empty C 2p orbital ! . 2p
Fig. 11.16 In order for an adjacent o bond to help stabilize a cation, the « MO must be able to get into c®
the same plane as the vacant p orbital. This has been achieved in (a) where the C-H o MO involves
the carbon atom next to the one with the positive charge. In the case of CHS, shown in (b), the only «
MOs are at right angles to the vacant p orbital and so no net overlap is possible.
If the orbitals can overlap, they interact as shown in Fig. 11.17. The elec-
trons in the 0 MO are now more delocalized and hence lowered in energy. It is
this lowering in energy of the electrons in the o MO that stabilizes the cation.
Fig. 11.17 The interaction between the o
Figure 11.18 shows the delocalized MO formed essentially from the interac-
bonding MO and the vacant p orbital
tion between the C-H o MOs and the vacant p orbital. It clearly shows that lowers the energy of the electrons in the
some electron density is shared between the two carbon atoms. o MO, thereby stabilizing the cation.
Fig. 11.18 A delocalized MO formed from the interaction between the C-H o MOs and the vacant p
orbital; the electrons can be found in the region between the two carbon atoms. This delocalization
lowers the energy of the cation.
If the a MOs cannot get into the same plane as the vacant p orbital then
no stabilization is possible. This is the case in the CH} cation; the C-H bonds
are at right angles to the vacant p orbital and consequently no net overlap is
possible as illustrated in Fig. 11.16 (b).
The o MO need not be from a C-H bond; a C-C bond, for example, can
also stabilize the cation. The degree of stabilization by o conjugation is not
as significant as conjugation involving a lone pair or 2 bond. The reason for
this is that o MOs are lower in energy than both lone pairs and 7 MOs and so
the degree of overlap between a 0 MO and a vacant p orbital will be less than
between the lone pair or 7 MO and the vacant p orbital. However, the more a
conjugation that is possible, the more stabilized the cation will be.
This finally explains the mystery we started this chapter with: why f-buty]
bromide undergoes a substitution reaction by first forming the carbenium ion
whereas bromomethane follows the Sy2 mechanism instead. In the (CH3)3Ct
ion, C-H o MOs from each of the three methyl groups are able to conjugate
with the vacant p orbital. Figure 1'1.19"shows a delocalized MO where the Pp
orbital on the central carbon is o conjugated to all of the C-H MOs, spreading
electron density over the whole molecule.
The large amount of o conjugation possible in this ion is sufficient for it
to be relatively stable and so be formed readily. In contrast, for CH} no a
conjugation is possible and hence this cation does not form. So on the grounds
Fig. 11.19 A MO in the carbenium ion, of the stability of the carbenium ions, the Syl mechanism is disfavoured for
(CH3)3C*, showing the extensive
delocalization due to the « conjugation. CH3Br and favoured for (CH3)3CBr.
The degree to which o conjugation is possible explains the observation
that the more alkyl groups around the central cation, the more stable the ion, as
shown in Fig. 11.20.
© @ ® (o)
H—cwh Hac — Cw" Hyco — Cw" CH Hac
— cus
H H H CHs
Increasing degree of o conjugation possible
Increasing stability of cation

Fig. 11.20 The more alkyl groups attached to the positive carbon, the more « conjugation is possible
and hence the more stable the carbenium ion.
11.3 Elimination reactions

Mixing a nucleophile, such as cyanide, CN~, with t-butyl bromide gives the ex-
pected substitution product (CH3)3C-CN. However, adding hydroxide (which
we might think of as being a typical nucleophile) to t-butyl bromide does
not yield (CH3)3C—OH but instead an alkene, isobutene (2-methylpropene), as
shown in Fig. 11.21. Rather than the expected substitution reaction occurring,
what we see is an elimination reaction in which the net result is that HBr has
effectively been lost (eliminated) from the t-butyl bromide (no HBr appears in
the products since under basic conditions water and bromide ion are formed).
CHg
| iC}
CHa HC ON + Br
wi : + OH aK H3C OH
H3C'¢ Br
3 ~_N io}
C=CHe + Br + HO
H3C
Fig. 11.21 The reaction between f-butyl bromide and hydroxide does not yield the expected
alcohol
but instead an alkene, isobutene.
11.3 Elimination reactions 181
Let us suppose for the moment that in the elimination reaction the ini-
tial step is as we outlined above, namely the formation of the relatively stable
trimethylcarbenium ion, (CH3)3CT. Why does hydroxide not react with this to
form the substitution product (CH3)3C—OH? We can gain a clue to understand-
ing this by looking again at the delocalized MO of the trimethylcarbenium ion
shown in Fig. 11.19. The orbital shows that there is electron density spread
out over the whole of the ion; this delocalization has a number of important
consequences.
The delocalization removes some electron density from the C-H 0 MOs
and shares it between the two carbons. Removing electron density from the
C—H o MOs means that these o bonds are weakened. However, the C-C bond
is being strengthened since it gains some partial 7 character. This accounts for
the fact that the C—-C bond is shorter in the cation than in the neutral hydrocar-
bon as illustrated in Fig. 11.22.
If we continued this withdrawal of electrons from one of the C-H o bonds |
to the extreme, eventually all of the electron density from the o bond would ~~
Cu 'CH
H3C7h W'S
end up being shared between the two carbon atoms. In other words, eventually 8 CH;
there would be a full C-C m bond and a free proton, Ht. This process is 1.54A
illustrated in Fig. 11.23 where the stabilizing interaction between the C-H o Fig. 11.22 The delocalization in the
bond and the vacant p orbital is shown in (a) and the result of taking this to the trimethylcarbenium ion shortens the C—C
extreme is shown in (b). The corresponding curly arrow mechanism is shown bond length relative to that in the
hydrocarbon.
in the lower part of the diagram.
The mechanism forming the alkene in Fig. 11.23 is not realistic — the car-
benium ion does not spontaneously form the alkene and a proton in solution.
Indeed, as we shall see later, the reverse is true — a proton will attack the alkene
to form the cation. Nonetheless, the 0 conjugation does weaken the C—-H bond
and this means that the hydrogen becomes more acidic. This gives us the clue
as to what happens when a base (such as hydroxide) is added.
In order to form the substitution product, the incoming hydroxide has to
reach the central carbon. Although this is possible, the approach is crowded
by the hydrogens — see Fig. 11.24 (a) — and in fact the hydroxide reacts pref-
erentially with these. Remember that these hydrogens are, as a result of the
NC—Ca" CH 3 Hii) Ao i CH
17®@ @ i
empty C 2p orbital “
Py filled C—C nr orbital
K H®
Hwyo—C CH
CV ech Hinge
yc “CH gun
H
Fig. 11.23 The interaction between the C-H « MO and the vacant p orbital is shown in (a). If carried to
the extreme, this would eventually lead to the complete formation of a C-C m bond and a free proton
shown in (b). The lower part of the diagram shows the curly arrow representation of the same process.
(a) (b)
HO HO -)
Fig. 11.24 Rather than squeeze past the hydrogens to attack the vacant p orbital on the carbon as
shown in (a), the hydroxide just helps to take off the already acidic hydrogen as shown in (b).
@ conjugation, somewhat acidic and so all that happens is that the OH™ ab-
stracts one of these hydrogens (as H*) to give water and the alkene as shown
in Fig. 11.24 (b). The corresponding curly arrow mechanism for this reaction
is shown in Fig. 11.25.
H—o” ) H—O,
f H
(3)
wiC—CX ~ uCH 30 HungmcounCHs
a)
HH CH, H "CHs
Fig. 11.25 The reaction between the trimethylcarbenium ion and the strong base hydroxide to form the
elimination product, isobutene.
We have seen that, instead of forming t-butyl alcohol, hydroxide ion reacts
with t-butyl bromide to give isobutene. It is possible to form t-butyl alcohol
from the bromide but we cannot use hydroxide since it is too strong a base.
Using water instead of hydroxide will form more of the desired substitution
product, although some of the alkene will still form. Water is less basic than
hydroxide and therefore not so good at removing the acidic hydrogens from the
cation. However, it is still perfectly able to react with the reactive trimethylcar-
benium ion. The curly arrow mechanism for the reaction of the trimethylcar-
benium ion with water is shown in Fig. 11.26.
H ‘
|
H
-O Ge H ()]
~) H iV H rox H
Hac
3
—SunCHg oT,
b Stn, a el
| an :
NCH, HaC~ Von’ H3C~ NV Gis
3
Fig. 11.26 The reaction between the trimethylcarbenium ion and the weak base water to form the
substitution product, t-butyl alcohol.
Bases and nucleophiles

As we have seen, in its reaction with a halogenoalkane, hydroxide can act
either as a nucleophile and substitute for the halogen, or it can act as a base
by removing a hydrogen ion. In a way, it is unfortunate that we have two
separate words for this behaviour since it encourages us to think that there is
11.4 Addition reactions — elimination in reverse 183
a fundamental difference in the way the hydroxide reacts; there is not. When
acting as a base, hydroxide (or any other base) is carrying out a nucleophilic
attack but specifically on an H—-X bond rather than on a C—X bond. All that is
needed is for the base/nucleophile to possess a high-energy pair of electrons,
usually a lone pair. The base/nucleophile could be charged, like OH~ or NH;
or neutral, like NH3.
The idea that a base is a hydrogen ion acceptor and that an acid is a hydro-
gen ion donor, is known as the Brgnsted-Lowry theory, named after Johannes
Brgnsted and Thomas Lowry who independently proposed these ideas in the
1920s. A more refined definition of acids and bases was proposed by Gilbert
N. Lewis: a Lewis base is a species which is a good donor of electrons whilst a
Lewis acid is a species which can act as an electron acceptor. The Lewis defi-
nition of a base removes any distinction between something acting as a base or
as a nucleophile!
It is true that some substances are better at removing hydrogen ions than
attacking at other centres; an example being inorganic amides such as sodium
amide, Nat NH; or lithium diisopropylamide, Lit N[CH(CH3)2]3 . Other sub-
stances, such as iodide, I~, are good nucleophiles but poor bases. Most sub-
stances can (and do) act as both nucleophiles and bases.
Exactly why something prefers to attack an H—X bond and remove a hydro-
gen ion, rather than attack any other centre, depends on a number of factors;
these may include the energy of its lone-pair electrons, which other groups or
counter-ions are present, what solvent is used and, of course, exactly what it is
reacting with. These ideas wil] be explored in the future chapters.
11.4 Addition reactions — elimination in reverse

We mentioned above that whilst it is fairly easy for a base such as hydroxide
to remove a hydrogen from the trimethylcarbenium cation, the Ht will not
spontaneously fall off to give an alkene. We are now going to look at the
reverse reaction, adding H* to an alkene to form a carbenium ion, specifically
we shall look at the protonation of isobutene to form the trimethylcarbenium
ion.
On p. 101 we saw that in trying to understand how two reagents might react,
we should first identify the highest energy occupied MO of the whole system
and then the very lowest energy unoccupied MO. Treating our acid as a source
of H+ means that the LUMO must be from this species as it has no electrons;
we will assume that this orbital is a 1s AO. The highest energy electrons must
therefore come from the alkene. There are no lone pairs in the alkene — the
HOMO is the C-C xz bonding MO.
The problem is, for an unsymmetrical alkene like isobutene, there are two
positions in which the incoming Ht could end up being bound, as shown in
Fig. 11.27.
In route (a), the proton adds to the /east substituted carbon atom, meaning
that the positive charge ends up on the carbon atom with the most substituents.
Protonation on the other carbon, as shown in route (b), means that the positive
charge ends up on the least substituted carbon. We have seen on p. 180 that the
(a) ®
ral
Hey
1C=C}h
wiCH
3 ee
‘ ‘\
wic—C
—_—
®@
wiiCHs
‘CH;
H™ NCH H
(b) ®H ou ’ H
Hi JC=Ch wtCH 3 ———_> Hey 1C—Cu,

H” ‘CH HY Yorn’
Fig. 11.27 An unsymmetrical alkene, such as isobutene, may protonate in either of two positions. {n
(a) the proton adds to the carbon atom with the most hydrogens and results in the positive charge
forming on a carbon which has three methyl groups attached. In (b) the proton adds to the carbon
with two methyl groups and results in the positive charge forming on a carbon with just one alkyl group
attached. Route (a) is the preferred choice since this forms the most stable carbenium ion.
most stable cation will be the one where the most o conjugation is possible.
Since hydrogens attached to a positively charged carbon atom cannot stabilize
the cation but adjacent C-H or C-C o bonds can, the most stable cation will be
the one with least hydrogens attached to the positive carbon. Looking at it the
other way round, the most stable carbenium ion is the one in which the positive
carbon has the greatest number of stabilizing alkyl groups attached to it.
Once the carbenium ion has been formed, it will react with any nucleophile
present, as shown in Fig. 11.28. This is exactly the same mechanism we saw
in our discussion of the Sn 1 mechanism.
H Cc—cuncHs H oo,
3 C —_—_ 3 Cc ‘ny CH
CH3
Fig. 11.28 The trimethylcarbenium ion will quickly react with any nucleophile present.
Overall in this reaction H™ is first added to the alkene followed by attack by

a nucleophile. For example when isobutene is mixed with hydrobromic acid,
H3C H
HBr, the first step is for the alkene to be protonated to give the trimethylcarbe-
‘c=c
/ nium ion and then this is attacked by bromide ion giving t-butyl bromide. Such
H3C H
reactions are examples of addition reactions. As Fig. 11.29 shows, addition
‘ reactions are simply the reverse of elimination reactions.
— HBr + HBr
Markovnikov’s Rule
elimination addition
The observation that the proton adds to the carbon with the greater number of
hydrogens attached, is sometimes known as Markovnikov’s rule after the Rus-
sian chemist, Vladimir Markovnikov, who first formulated the rule. However,
Br H
sometimes the rule breaks down. A better version of the rule would be to say
that the proton adds to the carbon atom of the alkene that gives rise to the most
— ~ H
Hsc'4 CN
HC stable cation. The reason for this slight modification is that, as we have seen
Fig. 11.29 Addition of HBr to isobutene is
above, other groups are better at stabilizing positively charged carbon atoms
the reverse of the elimination of HBr from than simple alkyl groups. An example of the application of this modified rule
t-butyl bromide. is shown in Fig. 11.30.
HC6s+. 6
8 LON HC.sb~ 27 26 HaC\Ne LC
(ie protonate i “GHe protonate “ow ~CHe
Cc CHp on carbon 1 C CH> on carbon 2 Cc CH
Hl ~O7 H~ 1 >0O7 H~o~Q7

Fig. 11.30 The structure in the centre can be protonated on either C; or Co. Protonating Cj gives a
carbenium ion which is stabilized by a conjugation alone; protonating Co gives a carbenium ion which
is stabilized by the neighbouring oxygen atom. The oxygen is much better at stabilizing the charge
and so the molecule is protonated preferentially on Co.
The structure in the centre can protonate on either C; or C2. C; has more
hydrogens attached than C2 which has none and so Markovnikov’s rule would
suggest that protonation occurs on C;. Protonating C; does give a carbenium
ion which can be stabilized by o conjugation, but protonating C2 gives a car-
benium ion with a neighbouring oxygen atom. Remember (see p. 176) that
an adjacent oxygen atom stabilizes a positive charge very effectively — much
more so than just o conjugation alone — so this molecule does protonate pref-
erentially on C2. A better way of representing the protonated product is shown
in Fig. 11.31 where the oxygen has donated one of its lone pairs and formed a
mz bond between the C and O. So in this example the molecule does not pro-
tonate on the carbon with the most hydrogens attached but so as to give the Fig. 11.31 After protonating on C» the
most stable cation. Understanding the rule is always better than just applying carbenium ion is stabilized by the oxygen
as shown in this structure.
it blindly!
11.5 E2 elimination
We have seen how t-butyl bromide can eliminate HBr by first losing a bromide
ion to form the stable carbenium ion and then having a proton removed to form
the alkene. HBr can be eliminated from the other bromoalkanes to give alkenes;
for example, 1-bromopropane forms propene when mixed with hydroxide as
shown in Fig. 11.32.
H
Cc ~ S)
H3c~ ~C~ ——* HC Sct 1, + Br
+ H20
Fig. 11.32 Hydroxide reacts with 1-bromopropane to give propene.
However, a carbenium ion could not form from 1-bromopropane since there
would be insufficient o-conjugation in the CH3CH2CH; ion for it to be stable
(see Fig. 11.20 on p. 180). We therefore need to find an alternative mechanism
for the elimination which does not involve the formation of this ion.
We saw on p. 181 that the reason why a base can remove one of the protons
from the trimethylcarbenium ion was that the interaction of the C-H 0 MOs
with the vacant p orbital weakens the C-H bonds making the hydrogens more
acidic. In 1-bromopropane, there is no vacant p orbital for the C-H o MQOs to
interact with. However, the vacant C-Br o*: MO can interact to some degree
with the C-H 0 MOs.
The C-Br o* is not as low in energy as a vacant p orbital so the energy
match between this and the filled C-H o MO is not so good. Nonetheless, if
(a) (b)
e lone pair of base @. _- new o bond between

Be NY base and proton
O es C-H o orbital
Yo, SH
Hs Cv @.. , ny
/%
empty C—Br o* orbital
filled C-C x orbital
lone pair of bromide
al aH
cys SQjn
SH
> too
H3Cii iH
H Br Br 9
Fig. 11.33 Diagram (a) shows the interaction between the filled C-H « MO and the vacant C-Br o*
MO in 1-bromopropane as a base approaches. The products are shown in (b); a new bond has formed
between the base and the proton, a new C-C x bond has formed and the bromide ion has been iost.
The corresponding curly arrow mechanism for the process is shown below the structures.
the filled C-H o MO and the vacant C-Br o* can be aligned in the same plane,
they can have an interaction similar to that between the C-H o MO and the
vacantp orbital (see Fig. 11.23 on p. 181).
The interaction between the C-H o MO and the vacant C-Br o* MO in
1-bromopropane as a base approaches is shown in Fig. 11.33; the curly arrow
mechanism for the reaction is also shown in the lower part of Fig. 11.33.
There are three points to note in this reaction. The first is the formation of
a new bond between the proton and the base. As we saw before, the hydrogen
cannot just fall off by itself as an isolated proton; a base is needed to remove it.
In Fig. 11.33 (b) we see the new o bonding MO that has been formed between
the base and the Ht. The formation of this bond is shown in the curly arrow
mechanism by arrow (1). This arrow tells us that a pair of electrons from the
base is ultimately going to be bonding the base to the hydrogen.
The second point to note is the formation of a new zm bond between the
two central carbon atoms. We can understand this as being a result of the
gradual overlap of the filled C-H o MO and the vacant C-Br o* MO in the
starting material. The bonding interaction between these two orbitals is shown
by the wavy lines in Fig. 11.33 (a). The new z bond that has formed from this
interaction is shown in (b). Curly arrow (ii) tells us two things: that electron
density is moving from the C-H bond as it is breaking and that it builds up in
between the carbons forming the new z bond.
The final point to note is the breaking of the C-Br bond. We could think
of this as arising from the electrons from the C-H o bond moving into the
anti-bonding C—Br o* orbital. As we saw in Chapter 5, filling an anti-bonding
MO cancels out the effects of a filled bonding MO. So the interaction between
(a)
B-H bond partially formed
(-)
By C-H bond
a partially broken
H t H
Howe ©,
H *%
C-C x bond Br?

partially formed
C-Br bond partially broken
Fig. 11.34 The transition state of the reaction between 1-bromopropane and a base is shown in (a).
The new base—proton and C-C m bonds have partially formed and the old C-H and C—Br bonds have
partially broken. Charge is beginning to develop on the bromine atom and is being lost from the base.
Diagram (b) shows a delocalized MO for a transition state analogous to that shown in (a). This shows
that electron density is spread over the whole molecule. Its shape can be understood by comparing
with the orbitals in Fig. 11.33.
the C-H o bonding MO and the C—Br o” simultaneously forms a new C-C x
bond and breaks the C—Br bond. The breaking of the C—Br bond is represented
by curly arrow (iii) which tells us that the electrons from this bond end up on
the bromide ion as a lone pair, as shown in (b).
The important point is that these three steps are all occurring at the same
time. The mechanism is said to be concerted. Fig. 11.34 (a) shows a transition
state for this process. It shows the new bond forming between the base and the
proton, the C-H bond beginning to break, the C-C a bond forming and the
C-Br bond beginning to break. The negative charge that was initially on the
base is being spread over the whole molecule — still partially on the base but
also beginning to form on the bromide.
The full MO picture for this reaction is rather involved but the HOMO for
the transition state is rather revealing and is shown in Fig. 11.34 (b). This is
one, delocalized, MO which can hold just two electrons. It clearly shows that
electrons are being transferred from the base to the bromide leaving group with
the form of the orbital resembling the interactions shown in Fig. 11.33.
Elimination kinetics
Since this elimination mechanism involves the base and the halogenoalkane
coming together in a single encounter, it should not surprise us that this reaction
rate is first order with respect to the concentrations of both the base and the
halogenoalkane, i.e. second order overall:
rate of elimination = k[1-bromopropane]|[base].
This mechanism is known as an ‘E2’ mechanism, i.e. an Elimination reac-

tion in which the rate-determining step involves two species i.e. is bimolecular.
The first elimination mechanism we saw in- Section 11.3, where the bromide
ion first fell off the t-butyl bromide to leave the relatively stable trimethylcar-
benium ion which was then quickly attacked by the base, would be termed an
‘El’ mechanism. Like the Syl mechanism, this is a first-order reaction since
the rate-determining step, the initial loss of the bromide ion, involves only the
t-butyl bromide.
The El and E2 mechanisms are really the extremes of a continuum of pos-
sibilities. In the El, the halide or other leaving group leaves first before the
proton is abstracted. In the E2 mechanism, both events happen together, so in
the transition state the halide ion has partly left and the proton has partly been
abstracted. If the halide is nearer ‘to having left than the proton is, then the
reaction is beginning to look more E1-like.
A pure E!] mechanism is rather less common than the E2. An elimination
reaction will only take place by an El mechanism if a relatively stable cation
can form following the loss of the leaving group and if the base that is present
is not too powerful. If the base is strong, it will get involved and help the
hydrogen come off before the leaving group gets a chance to fall off by itself.
12 The effects of the solvent
Up to now we have neglected a crucial component of reactions — the solvent.

Almost all reactions are carried out in solution, including all the ones going
on in our bodies, but it is all too easy to overlook how much the solvent influ-
ences reactions. A good example is the reaction of t-butyl chloride shown in
Fig. 12.1. As was discussed in Chapter 11, this molecule reacts by first forming
the carbenium ion (CH3)3Ct which then goes on to react with either a nucle-
ophile to form a substitution product, or with a base to form the elimination
product, isobutene.
CH,
|
. . H cue,
substitution SY 7 Nu
aN iC} H3C
CH, u
CHs
| rate determining o/
Hoc we - 0
not CP
3
sep gc“ CH3
cl @
-H H3C _ _
elimination 7 C= CH
H3C
Fig. 12.1 t-butyl chloride reacts first by forming the carbenium ion and then forming either the substi-
tution product or the elimination product. In each case, the rate-determining step for the reaction is the
initial formation of the carbenium ion. '
The rate of formation of either the substitution or elimination product de-

pends only on the rate of the initial step in which the carbenium ion is formed;
the reaction is therefore first-order. The table below gives values of this first-
order rate constant, k;, at 298 K for the reaction of t-buty! chloride in a number
of different solvents. The table also gives the half-life, ty, , of the reaction which
is the time needed for the reactant concentration to halve.
water ethanol acetone benzene pentane

ky/s7! 29x 1072 8.5x 1078 13x 107! 69x 10733 10~!6
ty 24 sec 94 days 169 years 30000 years 220 million years
The table shows that whilst the reaction of f-butyl chloride in water pro-
ceeds quickly with half of the reagent being used up in 24 seconds, the reaction
in pentane is so slow that it essentially does not occur at all. Even changing the
solvent from water to ethanol results in a considerable change in the rate. It is
clear that the solvent has an enormous effect on the rate of this reaction.
The reaction between CH3Br and Cl, shown in Fig. 12.2, proceeds by the
Sn2 mechanism and it is found that the influence of the solvent on the rate of
this reaction is opposite to the case of t-butyl chloride. For example, changing
the solvent from water to propanone (acetone) results in a 10? fold increase in
the rate. Moving to the gas phase, where there is no solvent at all, results in a
rate increase by a factor of 10!5,
a® Vi Cll LH
H"'y CL} OO C (o}
¢ ~Br 42 Br
Fig. 12.2 The substitution reaction between bromomethane and chloride ion proceeds 1 04 times faster
in propanone (acetone) than water, and 1015 times faster when no solvent is present at all.
A further example of the importance of the solvent is the solubility of one

substance in another. Solid NaCl dissolves in water but not in hexane, whereas
the reverse is true for a substance such as naphthalene (Fig. 12.3). We attribute
the ready solubility of ionic compounds such as NaCl in water as being due to
the ability of water to solvate ions — something which hexane is very much less
Fig. 12.3 The structure of the aromatic
hydrocarbon naphthalene. At room
effective at doing.
temperature it is a white solid which The rate-determining step of the reaction of t-butyl chloride involves the
sublimes readily giving off a vapour formation of ions. The fact that the reaction goes so much faster in water than
which repels many insects — hence the
use of naphthalene in moth balls.
in pentane is clearly related to the excellent solvation of ions by water.
In order to appreciate fully how the solvent can affect the reaction we first
need to understand the nature of the different solvents and how they can interact
with ions. It is to this topic that the next few sections are devoted and then,
towards the end of the chapter, we will return to answer the question as to why
the solvent has such a large influence on the rate of a reaction.
12.1 Different types of solvent

Solvents may usefully be classified according to how polar they are, but trying
to quantify this is not particularly straightforward. A useful measure is the
relative permittivity (or dielectric constant) of the solvent. In Section 3.2 on
p. 21 we saw that the attractive force between two oppositely charged ions with
charges c+ and z_ varies with the separation between the two ions, r, according
to:
Z4z_e7
force = ———~
4reqr2
where e is the charge on the electron (the fundamental unit of charge) and éo is
the vacuum permittivity.
In a solvent, the force between the two ions is reduced by a factor called
the relative permittivity, €,, of the medium:
4
Z4z—-e*
force = 5 (12.1)
4megérr
Thus, moving to a solvent with a larger relative permittivity decreases the in-
teraction between the ions, as is illustrated schematically in Fig. 12.4.
Typical values of ¢, are around 2 for hydrocarbon solvents such as hexane
to over 100 for primary and secondary amides. Water — the solvent used in
nature — has a relative permittivity of 80.
Solvents with a relative permittivity greater than about 15 are usually de-
scribed as polar whereas those with smaller values of the relative permittivity
are described as non-polar or apolar. As we shall see, polar solvents are much
better at dissolving ionic salts than are non-polar solvents.
(a) (b) (Cc)

|
—_—_—>— <r
|
2) © S) ©
Fig. 12.4 The force between two ions (represented by the size of the arrows in the diagrams) depends
on the relative permittivity of the solvent. The force is greatest when the ions have nothing in between
them as in a vacuum, represented in (a), and decreases on moving to solvents with progressively
increasing relative permittivities as represented in (b) and (c).
The relative permittivity can easily be determined from simple physical

measurements but rationalizing the particular value for a given solvent is not so
straightforward. The value of e, depends on the dipole moment of the solvent
molecules and also the extent to which they can interact with one another.
Particularly important in this regard is the ability of solvent molecules to form
hydrogen bonds with one another.
Hydrogen bonds
In order for a hydrogen atom in a molecule to form a hydrogen bond, it first
needs to be slightly acidic — usually, this simply means that it is bound to an
electronegative element such as fluorine, oxygen or nitrogen. Anything with
a high energy lone pair can then interact with this acidic hydrogen, resulting
in the lone pair being shared to some extent between the two atoms — in other
words, a partial bond is formed to the hydrogen.
Figure 12.5 illustrates the formation of a hydrogen bond in water, which has
both the required acidic hydrogens (attached to electronegative oxygen) and the
high energy lone pairs on oxygen. Liquid HF is another example of a solvent
in which there is extensive hydrogen bonding between solvent molecules.
(a) 7 (b) 4 (c) 4 (d) 7

/ KF) / / /
Aa :0 0° H® :0 0° H—~O O—H--O
/ \ / \ / \ / \
H H H H H H H H
Fig. 12.5 Illustration of the formation of a hydrogen bond between two water molecules. The elec-
tronegative oxygen withdraws electron density from the O—H bond as shown in (a). In the extreme
case, this would give a hydroxide ion and a proton, H*, shown in (b). Any neighbouring water would
be protonated at the oxygen to give the products shown in (c). In reality the process is not so extreme;
there is still almost a full bond in the water molecule itself and a partial bond — a hydrogen bond -
between neighbouring water molecules. This is usually indicated by a dashed line as shown in (qd).
Some approximate strengths of hydrogen bonds are shown in Fig. 12.6;

for comparison, the strengths of the ‘full’ bonds in these molecules are also
shown. In each case it can be seen that the hydrogen bonds are very much
weaker than the full bonds. The strengths of the hydrogen bonds increase with
the electronegativity of the element to which the hydrogen is attached, so the
hydrogen bonding increases along the series H2S, NH3, H20, HE.
The final point to note from Fig. 12.6 is that the hydrogen bond to an ton
is much stronger than that to a neutral molecule. The strongest hydrogen bond
in the diagram is that between HF and F~ shown in (h). In this species both
382 v __,, 29 _ -373

S—H-“-S—H N—H-~- fag H SH. 22 F—H FH
| | y)] 2 ye) 498
H H Hy qh H~ Ny
(e) HyO 5 (f) rS) (9) fo} (h)

“H. 151 O—H--Cl O—H-~2 -F 169 169 40
of wy [F—H—F]
HO _ H
Fig. 12.6 Approximate bond strengths and hydrogen bond strengths (in kJ mol—1) (a) in hydrogen
sulfide; (b) in ammonia; (c) in water; (d) in hydrogen fluoride; (e) between HzO* and water; (f) between
chloride ion and water; (g) between fluoride ion and water and (h) between fluoride ion and hydrogen
fluoride.
H-F bond lengths and bond strengths are the same. The resulting symmetrical
ion, [F-H-F]-, is so stable that it is readily obtainable as the solid sodium salt,
sodium hydrogenfluoride, NaHF2.
Hydrogen bonds are often responsible for holding molecules in one partic-
ular shape. This is important in proteins since it is the three-dimensional shape
of such molecules which gives them their special properties. Another example
is DNA in which the two strands which are twisted together into the famous
double-helix are held together by hydrogen bonds between the individual base
pairs. As is shown in Fig. 12.7, the base adenine hydrogen bonds with thymine,
and guanine with cytosine.
Thymine i
-O
Hv _H"
- So”
Sc _TA” H
N
oS Guanine |
hain H
Fig. 12.7 DNA exists as a double-helix due to hydrogen bonding. The adenine subunits from one
strand hydrogen bond with the thymine subunits from the other; similarly guanine and cytosine are
hydrogen bonded to one another.
Protic and aprotic solvents

Solvents which have weakly acidic hydrogens capable of forming hydrogen
bonds are called protic solvents. In order for the hydrogens to be acidic they
must be bonded to an electronegative atom, which makes the solvent polar.
Polar solvents which do not contain acidic hydrogens are called polar aprotic.
Figure 12.8 lists several common solvents in general order of polarity and
classifies them as polar protic, polar aprotic and non-polar. Also included in
the table are data on the relative permittivities of the solvent and the dipole
moments of the solvent molecules.
water On,
80.1 (1.85) H H
methanoic acid
i
Cc
51.1 (1.41) H~ OH
Polar protic
solvents
methanol CH30H
33 (1.7)
ethanol CH3CHz0H
25.3 (1.7)
dimethyl sulfoxide |
General order of decreasing polarity
(DMSO) 23w
47.2 (3.96) H3C CHs
dimethyl formamide J Polar aprotic

(DMF) Hu Nyos solvents
38.3 (3.82)
CH,
propanone \|
(acetone) Ue
21.0 (2.88) HsC~ ~CHg
Cl
trichloromethane |
(chloroform) cl we ~
4.8 (1.04) Cl
ethoxy ethane Ho Ho
Cw Cw Non-polar
(diethy! ether)
solvents
4.3 (1.15) HsC O CHg
benzene CoH,
2.3 (0)
Y hexane
1.9 (0)
CeHi4
polar aprotic and

Fig. 12.8 Table showing the classification of some common solvents into polar protic,
number immediately following the name is the relative permittivity at 293 K; the number
non-polar. The
in the gas phase.
italicized in brackets is the dipole moment (in Debye) for the molecule
For polar protic solvents, hydrogen bonding is possible between solvent

but
molecules. This group of solvents tends to have large relative permittivities
hydrogen
small dipole moments. The polar aprotic solvents are not able to form
bonds between solvent molecules. This group tends to have mid-range relative
permittivities but larger dipole moments.
small rela-
Non-polar solvents cannot form any hydrogen bonds and have
tive permittivities and small or zero dipole moments.
12.2 The solvation of ions

We saw on p. 190 that the force of attraction between two ions always decreases
when the ions are transferred from a vacuum to a solvent. For essentially the
same reasons, the Gibbs energy of an isolated ion always decreases on moving
from a vacuum into a solvent.
This change in Gibbs energy on taking an ion from the gas phase to a
solvent (the Gibbs energy of solvatiof’, AG?) can be estimated using the
Born equation:
2,2
ze“N 1
AG? solv = _ Ms
87 €or
(1- Ep=). (
12.2
)
Ee, ————— In this relationship, z is the charge on the ion, é; is the relative permittivity of
929 __20 40 60 80 the solvent, Na is Avogadro’s number and r is the radius of the ion.
Figure 12.9 shows a plot of the Born equation prediction for AG@,,, as a
~ 7200 function of relative permittivity; we can understand its form in the following
way. Since é€, is always greater than 1, the term in the bracket is positive and
Q3 -400 4 so the Born equation predicts that AG@.,,, will always be negative, i.e. there is
always a decrease in Gibbs energy on going from the vacuum to the solvent.
, 3-600 } Also, the term in the bracket is largest (closest.to 1) for large values of ¢,,
a -800 |
i.e. for polar solvents. This means that the greatest decrease in Gibbs energy
occurs when an ion is transferred from a vacuum to a polar solvent; in other
Fig. 12.9 Graph showing the prediction
of the Born equation, Eq. 12.2, for how
words, the Gibbs energy of an ion is lower in a polar solvent than in a non-polar
the Gibbs energy of solvation varies with solvent.
the relative permittivity of a solvent. Note The specific interactions between a charged ion and the solvent depend on
how AG... is always negative and
becomes more so as the relative
the type of solvent. Polar solvents — both protic and aprotic — usually have an
permittivity increases. However, once ¢, oxygen or nitrogen atom with a lone pair which can interact with positive ions
is greater than about 20, there is little as shown for the solvent dimethy] sulfoxide (DMSO) in Fig. 12.10.
further change on AB oy: The radius of
The big difference between polar protic and aprotic solvents arises in the
the ion has been taken as 1 A and z= 1.
solvation of anions. As we saw in Fig. 12.6 (f)-(h), protic solvents are able
to form hydrogen bonds to anions; this is shown for methanol in Fig. 12.11
Ko
(a). In contrast, no hydrogen bonds are possible between anions and aprotic
we / ; polar solvents — these can only solvate anions poorly by aligning their dipole
S
TAO Me moments as shown for DMSO in Fig. 12.11 (b). Non-polar solvents are poor
© Loe ™
at solvating both positive and negative ions.
O, y :0” 5 Soe Solvating different ions

oe
How well an ion is solvated depends not only on the solvent, but on the ion
i |~ itself. The Born equation (Eq. 12.2) tells us that for a given solvent, the Gibbs
73N energy of solvation is proportional to z*/r, i.e. the charge of the ion squared
Fig. 12.10 A polar solvent such as divided by its radius. This means that the Gibbs energy of solvation of an ion
DMSO solvates a cation by using its lone becomes more negative as the ion gets smaller and as the charge on the ion
pairs to interact with the ion.
increases.
It turns out that experimental values of the enthalpy and entropy of solva-
tion also correlate quite well with the value of z7/r. For example, Fig. 12.12
shows a plot of the enthalpies of solvation in water against z7/r for some com-
mon cations.
The enthalpies are clearly grouped depending on the charge on the cation —
the ions with the greatest charge having the most negative enthalpies of solva-
12.2 The solvation of ions 195
(a) CH, (b)

/
4
Oo folm
_ 7 O—CH 3 2 7 ON
Ha ON. ~ ,H Oy oe " N92,
EP v
“HA CH
HCO" \ ~O ° > |
H l. <6
| oreint \
/ “
H3C
Fig. 12.11 Protic solvents, such as methanol, solvate anions by forming hydrogen bonds as in (a).
Aprotic solvents, such as DMSO, only solvate anions weakly by orienting their dipole moments with
the positive end towards the anion as shown in (b).
tion. Within each group, the enthalpy of solvation is inversely proportional to

the radius of the ion: for example, the enthalpy of solvation for the large Ba?+
ion is less negative than that for the smaller Mg?* ion.
Figure 12.13 shows that the entropy of solvation in water correlates quite
well with z*/r, although the correlation is not as good as it is for the enthalpy
of solvation. For the majority of ions the entropy of solvation is negative,
which we can attribute to the ordering of the water molecules as they become
coordinated to the ions. The smaller the ion and the more highly charged it
is the greater restriction it causes and hence the more negative the entropy of
solvation.
zr nm
0 50 100 150 200 250
+ 4 it 4 i it
Nat .
Agr& Xtit Monocations
10004 2 cos
Pa Care
Pb?+ Fe?tMg?+
7 -2000 ig? BX
re) Zn?* Cy2+
€ dications poe
2 -3000+
= x X Lu
5 3+ 3+ Sc3+
7 Bi"* x, x, ye Fe3+
Mn
2 -4000
s Te wome crt
a. 4 trications
3 5000
% -6000 -
a
xTh 4+
oot
5 -7000 xu x24
tetracations
-8000
Fig. 12.12 A plot of the enthalpy of solvation of various cations in water against z*/r shows that there
is a good correlation between these two quantities. The plot also shows that the ions fall into groups
depending on their formal charge. Within each group, the smaller the radius of the ion, the greater the
enthalpy of solvation; thus Mg2*+ has a more negative enthalpy of solvation than Ba@*.
_
Cst*
100 4 XK, monocations 22/r dame!
Nat 50 100 150 200L 250ri
Cut | nl i
0
Lit Bat
entropy of solvation / J K -' mol ~!

-100 -
-200 -
-300 4 Lao
xFe* —Y ABY Cot
-400 4 trications a3
-500 - yth* — tetracations
Zt
-600 -
Fig. 12.13 Plot showing how the entropy of solvation of various cations in water correlates with 2 Ir;
the correlation is not as good as in the case of the enthalpy of solvation (Fig. 12.12). However, the
ions are still in groups depending on their formal charge and, within each group, the smaller the radius
of the ion, the more negative the entropy of solvation. The entropy of solvation for the large singly
charged cations are positive, implying that they disrupt the ordering of the water molecules to some
extent.
What is surprising is that the entropies of solvation for some of the larger
singly charged cations are actually positive; this means that transferring such
an ion from the gas phase to water results in an increase in the disorder in
the solvent. To work out what is going on here we need to recognize that in
pure water there is an extensive network of hydrogen bonding. When an ion is
introduced, some of the water molecules interact with this ion thereby disrupt-
ing the hydrogen bonded network in the solvent: this means that the entropy
within the solvent increases to a small extent. For large singly charged ions,
which only coordinate the water molecules weakly, this increase in the entropy
within the solvent outweighs the decrease in entropy due to the coordination of
the water thus making the entropy of solvation positive.
Another way of putting this is to say that these ions disrupt the hydrogen
bonded network in the solvent, leading to an increase in the entropy. A similar
effect is seen for anions (including all the common singly charged ones), most
of which have positive entropies of solvation in water.
When an ion is introduced into water, it is not only the solvent immediately
surrounding the ion that is affected. A number of water molecules, usually be-
tween four and eight, will be directly coordinated to the ion in the so-called pri-
mary hydration shell as shown in Fig. 12.14. However, the primary hydration
Fig. 12.14 The water molecules directly
shell may also affect the water molecules around it thus forming a secondary
coordinated to the ion form the primary hydration shell. It has been estimated that whilst only four water molecules
hydration shell, shown within the dotted are coordinated to the small Lit ion, up to a total of 22 water molecules may
line. These water molecules affect their
neighbours, binding to them more tightly
be affected in some way by the presence of the ion. The table below gives
than they would in the free solvent estimates for the total number of water molecules associated with various ions,
forming a secondary hydration shell. sometimes called the hydration number.
ion Cst Kt Nat Lit Ca?+ Mg?+ Zp2t+

hydration number 6 7 13 22 29 36 44
The hydration shell around a given ion often results in the ion appearing
much larger in solution than might be expected. For example, while the actual
ionic radii of the gaseous ions of the Group I metals increase as we go down
the Group from Lit to Cs*, the effective radii of the ions and their coordinated
water molecules actually decrease as we go down the Group.
The coordination of water molecules with a given ion may also play an
important part in how soluble a compound is in water, as we shall see in the
next section.
12.3 Solubilities of salts in water

On p. 14 in Chapter 2 we saw that by considering the entropy changes of both
the system and the surroundings we could understand why a salt such as am-
monium nitrate dissolves in water despite the process being endothermic. We
are now going to revisit this whole question of why some salts are soluble and
some are not — not surprisingly the solvent plays a crucial role in this and we
will use the ideas developed so far in this chapter to understand the factors
which affect solubility.
Rather than looking at the entropy change of the system and the surround-
ings, and hence determining the entropy change of the Universe, it is easier
just to consider the Gibbs energy change of the system — we saw in Section 2.7
starting on p. 16 that these two approaches are entirely equivalent. So, when
trying to predict the degree to which a salt, MX, is soluble, we simply need to
look at the Gibbs energy of solution, AG? solution’ for the process
MX(s) —> Mt (aq) + X~ (aq). (12.3)

The crucial thing is the sign of AG€.,..5n: if it is negative then the products are
favoured at equilibrium, i.e. the salt is soluble; if it is positive then the solid salt
is favoured and so it is insoluble (or, more accurately, the salt will be sparingly
soluble).
We can also assess how soluble a salt is by simply looking at the equilib-
rium constant for the reaction of Eq. 12.3. This equilibrium constant is known
as the solubility product, Ksp, and is given by
Ksp = [Mt (aq)leq(X~ (aq) leq;
where you need to recall that solid species do not contribute terms to the ex-
pression for an equilibrium constant, so only the ions from Eq. 12.3 appear in
the expression.
As we saw in Section 8.3 on p. 118, for any equilibrium the standard Gibbs
energy change and the equilibrium constant are related; in this case AGS ojution
and Kp are related by
A G° —_ R T In Ksp .
solution —
In trying to rationalize why a particular salt is either soluble or insoluble,

it is often helpful to break down AG® solution for the salt into the enthalpy and
entropy contributions, AH?) . 4, and ASsoution”
AG olution = A solution — TASsoiution:
It is also helpful to analyse the values of each of these using the cycle shown in
Fig. 12.15. In this cycle the solid salt is, first dissociated to give gas-phase ions;
the enthalpy change for this process is the (standard) lattice enthalpy, AHP rice?
discussed on p. 26. Then the gaseous ions are solvated for which the enthalpy
change is the (standard) enthalpy of solvation of M* plus that for X~
AP Ss! otutionMX)
MX(s) >» Mt(aq) + X-(aq)
AF inttice/ MX)
ae )
AM Sol X )
M*(g) + X(g)
Fig. 12.15 The enthalpy of solution of a salt may usefully be broken down into the lattice enthalpy of
the solid plus the enthalpies of solvation of the gaseous ions. A similar cycle can be constructed for
the entropy of solution.
Soluble salts
We shall first look at some salts which are readily soluble in water; some data
(at 298 K) for three such examples are given in the following table.
salt Ky AG cotution AA otution A SSolution ~ TAS ution

“/kImolt! ski molt! /IK7! molt! / kJ mol!
Ca(NO3)2 410000 —32 —19 +45 —13
MgSO4 16 000 —24 —87 —210 +63
NaNO3 15 —6.7 +20 +90 —27
All three of these salts are soluble in water as shown by the large values of K sp
and the negative values of AG@.,...,,- The contributions to AG solution made
by AAS ution 29d ASSution are Visualized in Fig. 12.16; as it is most directly
relevant, —-TAS® tion Has been plotted, rather than AS? solution’ Let us look at
each salt in turn and work out why it is soluble.
Why is calcium nitrate soluble in water?

The reaction is exothermic, which must mean the enthalpy of solvation is
greater in magnitude than the lattice enthalpy. The entropy also increases,
which we can attribute to the breaking up of the ordered lattice. Both AH°
solution
and —TAS°solution are therefore negative, contributing to a negative AG°
solution’
Why is magnesium sulfate soluble in water?

For this salt the reaction is much more exothermic than for calcium nitrate.
The enthalpy of solvation is now much greater in magnitude than the lattice en-
thalpy, something we can attribute to the fact that there are nwo doubly charged
12.3 Solubilities of salts in water
199
80 [ ~TAS?
60 + 7
© 40
=
3 20 CI
0) —
Lal
—_————_ — ——___1 — es
oe
20 a |
hea ee
| |
AH?
Ca(NO,)., MgSO, NaNO,
Fig. 12.16 Visualization of the separate contributions made by AM solution (light grey bars) and
ASsolution (Aarker grey bars) to the values for AG&..4,,,, (black bars) for three different salts dis-
solving in water. The entropy term has been plotted as — TASS tution 28 this is most directly relevant.
In all three cases AG? solution is negative, indicating that the salt is soluble, but this negative value is
achieved in different ways for each salt.
ions being solvated. However, there is a cost for this strong solvation — the en-
tropy of solvation is negative because many water molecules are being tied up
solvating the ions. However AG2,,,,,,, 18 negative as the large enthalpy term
outweighs the unfavourable positive —TAS®.,,..,, term.
We can say that this salt is soluble because the favourable enthalpy term
outweighs the unfavourable entropy term.
Why is sodium nitrate soluble in water?

For this salt the reaction is endothermic, which must mean that the enthalpy of
solvation is smaller in magnitude than the lattice enthalpy. This should not be
too surprising — both ions are singly charged and as we have seen these are not
strongly solvated in water.
There is a large increase in the entropy when the salt dissolves. We can
attribute this to the loss of the ordered lattice and the weak solvation of these
singly charged ions. Overall, the negative —T AS®.)jo, term outweighs the
positive enthalpy term, resulting in a negative value for AG Qj jtion-
We can say that this salt is soluble because the favourable entropy term
outweighs the unfavourable enthalpy term. These three examples show how a
negative AG° can be achieved using different combinations of AH, and
A ie)
solv’
Sparingly soluble salts

We shall now consider some salts that are only sparingly soluble in water.
These salts will all readily be precipitated out of solution when their ions are
mixed. Since the precipitation reaction does occur whereas the salts do not dis-
solve to a significant degree, rather than looking at the Gibbs energy for the salt
dissolving we shall instead look at the Gibbs energy of precipitation, AGoow
which is A,G° for the reaction
M* (aq) + X” (aq) —> MX(s);
of course, AG°.,ppt — =- AG° solution’

20 =TAS? bo
_ o- _ .— ea LI
2-20; || a |
— | |
2 -4o- | |
60r Hl ac
-80 \H®
AgCl BaSO, CaCO,
Fig. 12.17 Visualization of the separate contributions made by AA ot (light grey bars) and ASeot
(darker grey bars) to the values for AGopt (black bars) for three different salts being precipitated from
solution in water. The entropy term has been plotted as — TASS ot:
As before, the Gibbs energy change may also be broken down into the
enthalpy and entropy of precipitation, AH>,, and AS;
AG? ppt = 4H, ppt — TAS, ot:
Some data (at 298 K) for three sparingly soluble salts are given in the following
table and, as before, the data are visualized in Fig. 12.17.
AGS AHS ASS

7” ppt 7 —TAS® Ppt
salt Ksp Ppt ; ppt
/kJ mol /kJ mol /JK~* mol /kJ mol
AgCl 1.5 x 10710 —56 —66 —34 +10

BaSO4 1.1 x 107? 5] -19 +105 —3)
CaCO, 5.8 x 107? -47 +12 +200 —59
The fact that these three salts are all only sparingly soluble in water is
indicated by the very small values for Ksp. Each AGj,, is negative, indicating
that all the salts will readily precipitate out of solution. We shall look more
closely at each reaction to see why the precipitation occurs.
Why does silver chloride precipitate out of solution?

The formation of the precipitate is an exothermic reaction which means that
the lattice energy is greater than the magnitude of the solvation energy; we can
rationalize this by noting that these singly charged ions are not strongly sol-
vated. There is a decrease in the entropy as the ions go from the solution to
the lattice. This is perhaps to be expected since the free ions are only weakly
solvated and so there is a large loss of disorder when the lattice forms. How-
ever, the resulting positive —TAS>,, term is not large enough to overcome the
negative AH), term, and so AG», is negative.
We can say that precipitation takes place because the favourable enthalpy
term is dominant. Looking at the process the other way round, AgCl is insolu-
ble because of the unfavourable enthalpy term.
Why does barium sulfate precipitate out of solution?

This is an exothermic reaction, although less so than for the precipitation of
silver chloride; we can attribute this to the stronger solvation of the doubly
charged ions. Precipitation is accompanied by an increase in the entropy, lead-

ing to a negative —TAS-,, term. So, both the enthalpy and the entropy terms
favour the reaction. At first it seems odd that the entropy should increase as the
ions form a lattice, however this increase is due to the liberation of the water
molecules that were strongly complexed to the doubly charged ions.
The precipitation of barium sulfate is favoured by both the enthalpy and
entropy change. The opposite reaction — barium sulfate dissolving — is dis-
favoured by both.
Why does calcium carbonate precipitate out of solution?

This is perhaps the most interesting of all the reactions as the formation of the
precipitate is actually endothermic. We can attribute this to the strong solvation
of the doubly charged ions resulting in the solvation energy being larger in
magnitude than the lattice energy. As with barium sulfate, and for the same
reasons, the entropy change on precipitation is positive, resulting in a negative
—T AS,» term which outweighs the unfavourable positive AH, term.
So, calcium carbonate precipitates from solution on account of the large
increase in the entropy due to the liberation of the solvent from the ions; this
favourable entropy change outweighs the unfavourable enthalpy change. To
have a reaction in which the formation of a solid is accompanied by an increase
in entropy seems at first sight rather odd until we recognize the crucial role
which the solvent plays.
Solubilities of salts in other solvents

The last section illustrated how a number of factors contribute to determining
how soluble a given salt is in water. The same ideas can be applied to solvents
other than water but generally these are not as good as water at solvating inor-
ganic salts. The table below shows AH, and AS?,,,, (at 298 K) for Nat ina
number of polar solvents.
water methanamide methanol DMF

(formamide)
AH? / kJ mol! —418 —434 —438 —450
AS°,solv /JK7! mol“! +5.9 —18 —44 —59
It is initially surprising that the enthalpy of solvation becomes more nega-

tive on moving from water to dimethylformamide (DMF). Each of these sol-
vents has a lone pair on oxygen which can solvate the cation in an exothermic
process. However, when the ion is in water it disrupts the hydrogen bonding
between water molecules; the breaking of these hydrogen bonds is an endother-
mic process and this makes the enthalpy of solvation of the ion less negative
than it might otherwise be.
The amount of hydrogen bonding is greatest in water, and reduces as we go
to formamide and then methanol; there is no hydrogen bonding between DMF
molecules. The trend in the enthalpies of solvation thus follows the extent of
hydrogen bonding in the solvent.
The ;trend in AS°,.solv supports this interpretation. Whilst there is actually
; + . .
a small increase in entropy on adding Na* to water due to the disruption of
the hydrogen bonding, for all the other solvents the entropy change is negative
because the solvent molecules become coordinated to the positive ion. Since
there is no hydrogen bonding to disrupt between DMF molecules, placing a
sodium ion in this solvent gives the largest decrease in entropy.
The greater decrease in the entropy on moving to an aprotic solvent to
some extent offsets the fact that the reaction becomes more exothermic. The
net result is that it is often difficult to predict how soluble an ion will be ina
given solvent. “
The reduced solubility of ionic salts in organic polar solvents is often due to
the large unfavourable negative AS<,,, term. For example, whilst the solubility
of NaCl decreases as we go from water to methanoic acid to methanol, the
process is actually endothermic in water, has almost no enthalpy change in
methanoic acid and is exothermic in methanol.
12.4 Acid strengths and the role of the solvent

An acid, which we will denote HA, dissociates in water to give H* and A~
according to the equilibrium:
HA(aq) = H* (ag) + A (aq).

H* exists in water as the solvated hydronium ion, H30*(aq), so the equilib-
rium is more properly written
HA(aq) + H2O(/) = H30* (ag) + AW (aq). (12.4)
This reaction involves breaking a bond (to give ions), which is certainly an
endothermic process, but the solvation of the resulting ions is exothermic and
so can compensate for energy needed to break the bond. The whole process
has many similarities to the case of an ionic solid dissolving in water and so,
on the basis of the previous discussion, we can expect that entropy terms will
play an important role.
In this section we are going to look at two examples of how we can un-
derstand trends in acidity by recognising the important role of the solvent. In
the first example we will see how it is possible to rationalize the observation
that whereas HCI is a very strong acid, HF is rather weak. In the second ex-
ample, we will look at the effect that substitution of H by Cl has on the acidity
of CH3COOH; we will find, perhaps rather surprisingly, that the origin of the
observed increase in acidity is entirely entropic.
The equilibrium constant for the dissociation of an acid (Eq. 12.4) is called
the acid dissociation constant, Ka, which is given by:
x, = HROTIAT
[HA]
As it is the solvent, the fraction of the water present which is actually used to
form H307* is negligible, so the concentration of water is essentially constant.
It is therefore not included in the expression for the equilibrium constant, Ka.
12.4 Acid strengths and the role of the solvent
203
The value of K, can be related to A,G°, and A,G° is computed from A-H°
and A,S° in the usual way:
A,G° = —RT In Ky
A,G° => A,H° —_ TA,S°.
The acidity of HC] and HF

We will discuss the acidity of these two acids in water by analysing their dis-
sociation using the following Hess’ Law cycle:
A,;G°(2)
HA(g) —" Ht(g)+A7(g)
a.6°()] [a,c
A,;G°(4)
HA(aq) — Ht(aq)+A(aq)
The acid is first taken to the gas phase (step 1), where it then dissociates into
gas phase ions (step 2) and then finally these ions are hydrated (step 3); for
simplicity we have written the dissociation as giving Ht rather than H3O7.
Using this cycle we can compute A,G° for step 4 by knowing the values for
steps 1~3: A,G°(4) = A,G°(1) + A;G°(2) + A;G°(3).
In the case of HCI the values of A,G° (in kJ mol~!) for each step are:
HCl(g) 3 Ht(g)+C1-(g)
-4] [-1392
HCl(aq) —S Ht(aq)+Cl-(aq)
The first thing that we note is that the dissociation of HC] in water is very
favourable (A,G°(4) = —42 kJ mol~!), which is as expected as we know that
HC] is a strong acid and so dissociates fully in water. Why this is so can be
rationalized by looking at the enthalpy and entropy terms of the separate steps
in the cycle. These are shown in the table:
A,H°/ kJ mol7! A,S°/ J K7! mol7!
Hel(g) “8 Ht(gy+Ci(g) HCK(g) —> Ht (g) + C1“(g)

Paws
HCl(ag) —8 Ht(aqg)+Cl-(aq) HCl(aq) = Ht(aq)
|-
+ Cl-(aq)
First, consider the enthalpy terms: A, H°(2), the enthalpy of dissociation to

ions in the gas phase, is very large and positive. An explanation for this is that
dissociating a molecule into ions requires a lot more energy than dissociating
into neutral atoms on account of the strong attraction between the ions which
must be overcome in order to separate them.-
However, A,H°(3), the enthalpy of solvation of the ions, is very large and
negative; we can attribute this to the favourable interactions between the ions
and the polar solvent. Overall, this term wins, making A,H°(4), the enthalpy
of dissociation in solution, negative. The relative sizes of the four terms are
visualized in Fig. 12.18; what is striking is how A,H°(2) and A,;H°(3) are
almost equal and opposite, leading to a value for A, H°(4) which is very much
smaller than either A,H°(2) or A,H°(3).
Of the entropy terms, the most important is the large negative value for
A,S°(3), the entropy of solvation of the ions. As we have already noted, the
ordering effect that an ion has on the.solvent is the origin of this reduction
in entropy. The smaller the ion, the greater the effect, and so we should not
be surprised to find that the solvation of H*(g) results in an entropy change
which is significantly negative. As we can see, A,;S°(3) is so negative that it
dominates the other two terms and makes A,5S°(4) negative; this is visualized
in Fig. 12.18.
The dissociation of HC] in water has a (negative) favourable A,H° and a
(negative) unfavourable A,S°. From these values we can compute A,G° (at
298 K) as:
Fig. 12.18 Visualization of the relative
sizes of the ArH° (on the left) and ArS°
(on the right) values of the steps involved
A,G° = A;H° — TA,S°
in the dissociation of HCI in water. The
numbering on the columns refers to the
= —58 — 298 x (—55 x 1077)
steps in the table above. A, H°(1) and = —58 + 16
ArH°(4) are all but invisible on this scale:
this highlights how the small value of = —42 kJ mol7!.
ArH?(4) is determined by the difference
in the two large terms, A, H°(2) and Overall the A,H° term dominates, making A,G° negative, but note that there
ArH?(3).
is rather a fine balance between A,;H° and the —7A,S° term — it would not
take much of a change in either to make A;G° positive.
It is perhaps surprising at first that HF turns out to be rather a weak acid
(only 7% of the HF molecules are dissociated in a 0.1 mol dm~? solution).
However, we can rationalize why this is by using a cycle in the same way as
we did for HCI and comparing the A,H° and A,S° values:
A,H°/kJ mol7! A,S°/J K7! mol7!

HF(g) > Ht(g)+Fo(g) HF(g) > Ht(g)+F7(g)
sf
HF(aq) —>
fs fa
H*+(aq)+F7(aq) HF(aq) —+ Ht (aq) +F-(aq)
Comparing these values to the ones for HC] we see that A,H°(2) is larger
for HF than HC] which just tells us that the HF bond is stronger than the HC]
bond. A,H°(3) is more negative for HF than HCl, and this can be rational-
ized by noting that F~ is smaller than Cl-; as is illustrated in Fig. 12.12 on
p. 195, the smaller an ion, the more negative the enthalpy of hydration. The
bond strength increases by a little more than the amount by which the solvation
energy becomes more negative and as a result A; H°(4), although still negative,
is much less so than for HCl.
When it comes to the entropy values we have a larger negative entropy of
solvation (A,;S°(3)) for Ht + F~ than for Ht + Cl-; again, we can attribute
12.4 Acid strengths and the role of the solvent 205
this to the smaller size of the F~ ion. A,S°(4) is thus more negative for HF
than for HCl.
The value of A,G° at 298 K now turns out to be positive:
A,G° = A,H° — TA,S°

= —9 — 298 x (-71 x 1073)
=-9+2]
= +12kJ mol7!.
In contrast to the case of HCI, the —7 A,S° term now dominates, making A,;G°
positive and so HF is a weak acid.
The comparison with HCI shows how relatively small changes in the values
of A,H° or A,S° for the individual steps can swing us from a strong acid (HCl)
to a weak acid (HF). It also shows the importance of the entropy terms which
in the end are the deciding factors in making HF a weak acid.
Chloroethanoic acids
Ethanoic acid itself is rather a weak acid, but as we substitute chlorine for the
hydrogens in the methyl group the acid strength increases to the point where
trichloroethanoic acid is classified as strong. It is interesting to see how this
trend can be rationalized with the aid of A, H° and A,S° values.
The equilibrium in question is:
CH3_,Cl,COOH(aq) + H20(1) = CH3-nCl,»,COO™ (aq) + H307 (aq)
where n can be 0, ..., 3. We define the acid dissociation constant, Ka, in the
usual way as
[CH3_,Cl,COO7 ][H30T]
a=
{CH3_,Cl, COOH]
When discussing acid strengths it is usual not to give the value of Kg but the
quantity pK, defined as
pKa = — log Ka.
Note that the logarithm is to the base 10, as opposed to the natural logarithm
(In). With this definition a weak acid with a value of Ka which is less than 1
will have a positive value for the pKa. For example, K, for ethanoic acid (at
298 K) is 1.7 x 10~> which is 10~*8, so the pKa is +4.8.
On the other hand, strong acids which have large values of K, have smaller
pK, values, and indeed for the very strongest acids the values are negative.
For example, Ka for HCI (at 298 K) is 2.3 x 10° which is 10+64 so the pKa
is —6.4. The logarithmic scale compresses a wide range of K, values into a
much smaller scale.
The table gives thermodynamic data for the dissociation of chloroethanoic
acids, all at 298 K.
7. 30; , 15,
E2 79 10
10} r 5} / i +
0 0 rn of
AH° -TAS® A,G° AH? -TA,S° 4,G° | 4H° -TA,S° A,G°
-104 | AH® -TAS° AG? | ° [
-20; + “1 OF
-30! L -15t L
CH,COOH CH,CICOOH CHCI,COOH CCI,COOH
Fig. 12.19 Diagrammatic representation of the relative contributions that the ArH°® and —TA,S° terms
make to ArG° for the dissociation of a series of chloroethanoic acids at 298 K. Note that the vertical
scale has been expanded two-fold for the di- and trichloro acids. The diagram clearly indicates that
the value of ArG®° is dominated by the - 7 A,S° term.
CH;COOH CH)CICOOH CHCl,COOH CCI;COOH

A;H° /kJ mol7! —0.08 —4.6 —0.7 +1.2
A,S°/JK7! mol! -91.6 —70.2 —27.0 —5.8
—TA;S°/kJ mol! 4.27.3 +20.9 +8.1 +1.7
ArG° /kJ mol7! +27.2 +16.3 +7.3 42.9
pKa 4.76 2.86 1.28 0.51
In addition to giving the values for A;H° and A,S° we have also given the
value of —TA,S°; the reason for this is that as A-G° = A,H° — TA,S° the
two terms which contribute directly to A;G° are A;H° and -TA,S°.
The first thing we notice is that adding each successive chlorine increases
the strength of the acid, as evidenced by the decreasing pK, values; indeed the
effect is rather dramatic when we recall that the pK, values are on a logarithmic
scale — the K, values in fact change by four orders of magnitude.
A simple explanation for what is happening is to say the electron-
withdrawing chlorine atoms are polarizing the O-H bond in the carboxyl
group, thus making it easier to dissociate. We could also argue that the elec-
tronegative chlorine atoms are helping to stabilize the anion formed on dissoci-
ation. Tempting though both of these arguments are, the data in the table reveal
that they are entirely incorrect.
The crucial observation to make on the data is that for all but the
trichloroethanoic acid the A, H° terms are negligible compared to the —T A, S°
terms; the variation in pK, values is therefore entirely an entropy effect, and
has nothing to do with the strength or otherwise of the O-H bond. Figure 12.19
illustrates in a graphical way the contributions which the A,H° and —TA,S°
terms make for each acid; it is clear that the enthalpy terms play an insignificant
role.
Our explanation for the increase in acidity must therefore centre on the
variation in the entropy terms. The first thing to note is that A,S° is negative;
we have seen this before for the dissociation of HC] and HF. As we noted there,
the origin of this reduction in entropy is the ordering effect which the ions have
on the water.
12.5 How the solvent affects the rate of reaction
207
The A,S° values become less negative as the methyl hydrogens are suc-
cessively replaced by chlorine; we can argue that the electron-withdrawing fe)
chlorine atoms are leading to a spreading out of the negative charge (formally |
residing on the oxygen) across the whole molecule. The charge is therefore less CL. 6
H3C~ ~O
concentrated and so the solvent water molecules are less tightly held, leading
to a smaller reduction in the entropy. Figure 12.20 illustrates the idea.
fe)
o | -
The case of trichloroethanoic acid is perhaps a little different from the oth- NOoos © 3
ers, as for this molecule the A, H° and —T A,S° values are comparable so that oO
we cannot say that the entropy term is dominant. However, it is still clear that
Ho o.
a
the high acidity of this species compared to all of the others is attributable to
the fact that it has the least negative A,S° term. In fact, the A,H° for this acid
5 T z
NM YAO é
is the only one of the series which is positive (i.e. unfavourable), but despite CH ~O ¢
this the trichloro derivative is the most acidic; the less unfavourable entropy | g
n
oO
term is clearly dominating over the enthalpy term. Cl Pi

$- o
This example illustrates two things. The first is that we really need to look O
at the enthalpy and entropy terms to see what is going on in a series of equi- o II
libria — just looking at the equilibrium constants can be rather misleading. The ee CL So
second is that when ions are involved in an equilibrium the contribution that
the entropy of solvation makes is often decisive. oi” Cl
s-
Fig. 12.20 Illustration of how the addition
of successive electron-withdrawing
12.5 How the solvent affects the rate of reaction
chlorine atoms leads to a dispersal of
negative charge for the carboxylate
We are now in a position to understand the effects of changing the solvent on
anion; as a result the solvating water
the rates of the reactions which we discussed at the start of the chapter. molecules are less tightly held and so the
For the reaction of t-buty] chloride, the first step is the formation of the ArS° values for dissociation become less
negative.
trimethylcarbenium and chloride ions. These ions are clearly far more polar
than the starting material and hence are preferentially stabilized by polar sol-
vents. The energy profiles shown in Fig. 12.21 illustrate this point: the solid
line shows a profile for the reaction in a non-polar solvent and the dotted line
for the reaction in a polar solvent.
The intermediate ions are stabilized by the more polar solvent and so are
lower in energy. More importantly, the transition state for the reaction in which
the carbenium ion is formed is also lower in energy since it too is more polar
than the starting material. The starting material, t-butyl chloride, is not charged
and so is little affected by changing the polarity of the solvent.
The net effect is that the activation energy for the reaction is decreased on
moving to the more polar solvent and hence the rate of the reaction increases.
This is exactly what is observed, as was discussed on p. 189.
In contrast, the reaction between CH3Br and Cl~ (Fig. 12.2 on p. 190)
goes faster on moving from water to acetone and is fastest when no solvent
is present. Since this reaction proceeds via the Sn2 mechanism, the rate-
determining step is the reaction between the bromoalkane and the incoming
chloride ion.
The key to understanding the difference in rate is the effect of the solvent
on the chloride anion. In a protic solvent like water or methanol, this anion
is strongly solvated by hydrogen bonding whereas in an aprotic solvent like
acetone, it is less strongly solvated (see Section 12.2 on p. 194). In the Sn2
mechanism, the nucleophile must approach the bromomethane in a specific
geometry (see p. 109) — the hydrogen bonded solvent hinders this approach
and hence the reaction rate increases on moving from the protic solvent water
to the aprotic solvent acetone.
An energy profile for this reaction is: shown in Fig. 12.22. In a protic
solvent, the negatively charged nucleophile is strongly solvated through hy-
drogen bonding. Thus the energy of the starting materials (chloride ion and
bromomethane) are lower in the protic solvent than in the aprotic solvent.
In the transition state, a bond is beginning to form between the chlorine and
carbon and the bond between the bromine and the carbon is beginning to break.
The negative charge is spread over the molecule (mainly on the chlorine and
bromine) rather than being concentrated on just one ion as it is in the starting
materials or products. In other words, the charge is becoming dispersed. As
a result, the solvent is less firmly associated with the transition state than with
the incoming nucleophile before the attack.
The important point is that while the chloride ion and the transition state are
all lowered in energy due to solvation by the protic solvent, the transition state
is lowered less than the chloride ion due to the charge being more dispersed
in the transition state. This means that, as shown in Fig. 12.22, the activation
energy is greater in the protic solvent than in the aprotic solvent. Thus, the
reaction proceeds more quickly in the aprotic solvent, which is exactly what is
found experimentally.
CH;
din
HeC'Yy “SL
rs H3C Cl
® transition
® state
——_ to products
Enon
polar
E polar
CH;
i H3C CH, .
\
H3C''y ae intermediate ions
H3C
reactant ‘sive
products
—————
reaction coordinate
Fig. 12.21 The energy profile for the reaction of t-butyl chloride in a non-polar solvent
is shown by the
solid line; the dotted line shows the profile in a more polar solvent. Whilst the change
in solvent has
little effect on the energy of the uncharged starting material, the charged intermediat
es are lower in
energy in the more polar solvent. As charges are building up in the transition
state for the first reaction
(which leads to the formation of the ions), this too is lowered in energy in
the more polar solvent. The
net effect is that the activation energy for the reaction in the polar solvent
is less than in the non-polar
solvent and hence the reaction rate is greater in the polar solvent.
12.5 How the solvent affects the rate of reaction 209
reactants transition state products
H H ° M
cr + o—er Cl- -b--B; CiI—cC,, +
Hi4 a ‘ ¥‘H
H
energy
E,, aprotic
—_ >
reaction coordinate
Fig. 12.22 The energy profile for the reaction between chloride ion and bromomethane in a polar
aprotic solvent is shown by the solid line and for the same reaction in a protic solvent by the dotted
line. In the protic solvent the free chloride ion is more strongly solvated than the transition state due
to the charge dispersal in the latter. The result is that on moving to a protic solvent the energy of the
transition state is lowered by less than is the energy of the chloride ion. Therefore the activation energy
for the substitution reaction is greater in the protic solvent than in the aprotic solvent.
13 Leaving groups
In earlier chapters we have looked at a number of simple reactions, such as

that between ethanoyl chloride and hydroxide, shown in Fig. 13.1 (a), and the
formation of a carbenium ion from t-butyl bromide, Fig. 13.1 (b).
(a)
i fe) {ou O
Co! OH
_
O
I © Cl
Hsc~ Cl H3C~ [e HsC~ OH
(b) CH3
| @
H3C yok —_—_» H3C WiC —CHg + Br —~, Products
H3C H3C
Fig. 13.1 Two reactions which we have looked at before and which occur readily: (a) the addition-
elimination reaction between ethanoy! chloride and hydroxide; and (b) the formation of the trimethyl-
carbenium ion from t-butyl bromide, which is the first step in the Sy1 and E1 mechanisms.
We now want to understand why some other reactions for which we can
write plausible looking mechanisms do not occur. For example, why does Cl~
not attack ethanoic acid to form ethanoy] chloride, as shown in Fig. 13.2 (a)?
This reaction is simply the reverse of that shown in Fig. 13.1 (a), but it does not
take place. Similarly, the attack on a ketone by Cl” to give an acyl chloride,
as shown in Fig. 13.2 (b), does not take place, neither does the formation of a
carbenium ion by loss of H~, shown in (c). We know that the same carbenium
ion is formed readily by loss of Br~ from t-butyl bromide, so why is it that this
ion cannot be formed by loss of H~ from a hydrocarbon?
(a) O Cl O
er OP I °oH
cl oe C
H3C~ OH H3C~ (9H Hsc~ Cl
(b)
QL
O
y, Eo, 4 Cl
Hac CH nso (OHS > . 7oN
(c) CHs
|
HCY <)wc
DK HeCrrc—cH,
'" @
+ H°
HC H3C
Fig. 13.2 Three reactions which do not occur: (a) the formation of ethanoyl chloride from ethanoic acid
and chioride ion; (b) the formation of ethanoyl chloride from propanone and chloride ion and (c) the
formation of trimethylcarbenium ion from 2-methylpropane.
13.1 Energy profiles and leaving groups 211
We will see that we can answer these questions simply by considering the
relative energies of the species involved. This will lead to the development of
a general set of ideas for predicting which reactions will go readily and which
will not.
13.1. Energy profiles and leaving groups

An energy profile diagram for the reaction between ethanoyl chloride and hy-
droxide is shown in Fig. 13.3. Since we know that the products are favoured
at equilibrium, we have drawn the products (chloride ion and ethanoic acid)
lower in energy than the reactants (ethanoyl chloride and hydroxide ion) by an
amount labelled AE .
To a rough approximation, the only difference between the ethanoy] chlo-
ride and the ethanoic acid is that the C—Cl bond is replaced by a C-OH bond.
At 450 kJ mol~!, the C-OH bond in the carboxylic acid is stronger than the
C-Cl bond in the acy! chloride (350 kJ mol!) and so this change contributes
to AE, the lowering of energy as the reaction proceeds to the products. How-
ever, the relative energies of the OH™ and Cl” ions will also contribute to the
value of AE as these too are reactants and products.
The key species on the energy profile is the tetrahedral intermediate. It is
perhaps not too surprising that it is higher in energy than both the reactants and
the products as the intermediate has four groups crowded around the carbon
whereas the reactants and products have just three. Once formed, the inter-
mediate can either lose Cl~ to form the products or lose OH™ to re-form the
reactants. Since it is harder to break a C—O bond than to break a C-Cl bond,
energy
tetrahedral
intermediate
AE
TO Ci e Lc |
products H3C
—————{
reaction coordinate
Fig. 13.3 An energy profile diagram for the reaction of ethanoy! chloride and hydroxide. The products
are lower in energy than the reactants by an amount labelled AE.
212 Leaving groups
it is easier for the intermediate to lose the Cl~ and form the products. Conse-
quently, we have shown the activation energy for the tetrahedral intermediate
to from the products, (F,2 in Fig. 13.3) as being lower less than the activa-
tion energy required for the tetrahedral intermediate to return to the reactants
(E4,-1).
In these two possible reactions of the tetrahedral intermediate, Cl” and
OH™ are called /eaving groups. The fagt that loss of Cl~ occurs more readily
than loss of OH™ leads us to describe Cl~ as a better leaving group than OH ~.
We can now answer the question as to why Cl~ does not attack ethanoic
acid to form ethanoyl chloride. If the Cl~ were to attack, as in Fig. 13.2 (a),
the tetrahedral intermediate so formed would be the same as that in Fig. 13.3.
As we have seen, for this intermediate the Cl~ is far more likely to leave than
the OH”, and this loss of CI~ will simply regenerate the ethanoic acid. No
ethanoyl chloride is produced as the intermediate preferentially collapses back
to the starting materials.
A similar line of argument can be used to explain why Cl~ does not react
with propanone to form ethanoyl chloride and CH; , as shown in Fig. 13.2
(b). The products from this hypothetical reaction are much higher in energy
than the reactants, as shown in the energy profile diagram in Fig. 13.4. We
have also shown the products being even higher in energy than the tetrahedral
intermediate.
The reason the products are so high in energy is mainly due to the formation
of the CH, ion; it is much harder to break a C-C bond and liberate CH, than
it is to break a C-Cl bond to form Cl~. We describe the Cl~ ion as being
more stable, meaning lower in energy, than the CH; ion. This is because the
energy
Ol 6 “CH,
AA
“Noy
products
AE>| | aE
Cl tetrahedral
“Cc intermediate AE,
H3C~ CH,
reactants
———-
reaction coordinate
Fig. 13.4 Energy profile diagram for the reaction between Cl- and propanone
to form ethanoyl chloride
and CH, | the reaction is shown in Fig. 13.2 (b).
13.2 Leaving group ability 213
electrons in the Cl” ion experience a greater effective nuclear charge than those
on carbon and hence are lower in energy (see Section 4.8 on p. 57).
Even if the chloride ion did attack propanone to form the tetrahedral inter-
mediate (which is entirely possible), this species would eliminate Cl~ and re-
turn to the starting materials in preference to eliminating CH, to form ethanoyl
chloride. We say that CI™ is a much better leaving group than CH3 .
Whilst the forward reaction shown in Fig. 13.4 does not take place, the (+) (-)
reverse reaction can. If care is taken to prevent further reactions, ketones can M—R
be prepared from acy] chlorides by reaction with organometallic compounds, Fig. 13.5 Organometallic compounds are
which essentially act as a source of the alkyl anion — see Fig. 13.5. Once the strongly polarized with the metal positive
and the alky! group (such as a methyl
alkyl! anion has attacked the acy] chloride to form the tetrahedral intermediate,
group) negative. They essentially behave
the loss of chloride ion inevitably follows yielding the ketone. The driving as a source of the R~ anion.
force for this reaction is essentially the high-energy organometallic starting Organocadmium compounds such as
(CH3)oCd are particularly good for
material — both the intermediate and products are lower in energy than this.
forming ketones from acyl chlorides.
Finally we turn to the question as to why the trimethylcarbenium ion
can form from f-butylbromide, as shown in Fig. 13.1 (b), but not from
2-methylpropane, as shown in Fig. 13.2 (c).
Since the same carbenium ion is being produced in each reaction, the dif-
ference must be because of the anion formed. We can therefore say that hydride
ion, H’, is a much worse leaving group than Br~. The reasons are twofold:
firstly, it is harder to break a C-H bond than a C-Br bond and secondly, the
greater effective nuclear charge on Br means that its valence orbitals are lower
in energy than those of hydrogen. This is often expressed by saying that Br7
is ‘more stable’ than the H~ — but we have to remember to be careful about
what we mean by ‘more stable’. As we have seen in the previous chapter, the
stability of an anion is often strongly influenced by its interactions with the
solvent.
13.2 Leaving group ability

In the previous section we saw that Cl~ is a better leaving group than both OH™
and CH; , and that Br~ is a better leaving group than H™. From the study of
many reactions, a general order for leaving group ability has been established;
Fig. 13.6 lists some common leaving groups in order of the ease with which
they can leave a compound.
How good something is as a leaving group depends on two things: (a)
the bond strength between the leaving group and the atom with which it was
initially bonded (usually carbon) and (b) the ‘stability’ of the leaving group
itself once it has been formed.
The very best leaving group is neutral nitrogen, N2. This is a very stable
molecule — once it has been eliminated, there is very little chance that it will
add back again. Further, when bonded in a molecule the N2 group bears a
positive charge which draws electrons towards it; this results in a very weak
C-N bond which may easily be broken, further enhancing the leaving group
ability.
The halide ions appear in the list in the order I~ > Br~ > Cl” > F’: of
them iodide is the best leaving group and fluoride the worst. This order re-
flects the halogen—carbon bond strengths as exemplified by the series CH3l,
214 Leaving groups
leaving compound pK, of conjugate

group acid (R = H)
@
A No R—Np
TsO° R—OTs -6.5

3 1° R—I ~10
5 Br® R—Br -9
D @®
$ H,O R— OH, -1.7
2 cle *R—-Cl -7
2 ®
F ce) _
R—F 3
3
ra R'COO® R'COOR 3-5
2 ®
£ NH3 R—NH3 9
OH® R—OH 15.7

RO R—OR 16-17
H° R—-H 35
very poor ° _
leaving groups} NHe R—NH, 38
CHS R—CH, 48
Fig. 13.6 The left-hand column lists some ions or molecules in order of their ability to act as a leaving
group from an alkyl carbon or a carbonyl group. The middle column shows how each leaving group
occurs in a compound; R represents any alkyl group. TsO is the tosylate group, shown in Fig. 13.7.
The right-hand column gives the approximate pKa value of the conjugate acid of the leaving group,
which is the acid formed when R=H.
CH3Br, CH3Cl and CH3F in which the bond strengths are 237, 293, 352 and
472 kJ mol! respectively.
The order F> > OH” > NH, > CH; reflects the stability of the anions
themselves. The elements formally bearing the negative charge are all in the
O
aN d /_0° same row in the Periodic Table and so the effective nuclear charge experienced
by the valence electrons in these atoms increases in the order F > O> N>C.
Consequently the lowest energy orbitals are in F~ and the highest in CH, ,
which means that F~ is lower in energy (i.e. more stable) than O~ and so on
for the others.
The order TsO™ > RCOO™ > RO™ also reflects the stability of the anion
which is related to the degree of delocalization that is possible in these ions
CH,
(see Section 10.4 beginning on p. 168). The negative charge in the tosylate ion
Fig. 13.7 The structure of the tosylate
can be delocalized over three oxygen atoms, which confers greater stability
group, abbreviated to TsO-.
when compared to the carboxylate anion in which the charge can only be delo-
calized over two oxygen atoms. Both of these ions are lower in energy than an
alkoxide, RO™ (e.g. CH3CH207), in which the charge essentially remains on
the single oxygen atom.
From the list we see that it is always harder for an anion to act as a leaving
group than the protonated form of the same group. For example, OH™ is a
much poorer leaving group than neutral H2O, and NH, is a much poorer leav-
ing group than neutral NH3. Thus one way to transform an -OH group into a
much better leaving group is to protonate it to form -OH; . Now, rather than
the anion OH” leaving, it is the neutral species H2O which leaves.
13.3 Leaving groups and pKa 215
13.3 Leaving groups and pKa

How well an ion (X~) or molecule (Y) can act as a leaving group is to some
extent reflected by the acidity of HX or HYT. HX is called the conjugate acid
of X~, similarly HY* is the conjugate acid of Y; these conjugate acids are
formed simply by adding H* to the relevant leaving group.
The connection between acidity of the conjugate acid and leaving group
ability comes about because similar factors affect both. Acidity is determined
by the H—X bond strength and the stability of X~. As we have seen, the same
factors determine how good X7 is as a leaving group, although we should note
that for a leaving group it is the C—X, rather than H-X, bond strength which is
relevant.
As we saw on p. 205, we usually quantify the strength of an acid using
pK, values. A negative pX, corresponds to an acid which readily dissociates
in water, whereas a positive pK, corresponds to a weak acid which only partly
dissociates, if at all.
The pK, values for the conjugate acids of the leaving groups are given in
Fig. 13.6. It can be seen that the conjugate acids of good leaving groups, such
as TsO”, I~ and H20, have negative pK, values whereas the conjugate acids
of groups which do not ordinarily act as leaving groups, such as H~, NH5 and
CH, , have large positive pK, values.
The same trends that we see for leaving group ability are found in the
strengths of the conjugate acids. For example, the acidity of the hydrogen
halides increases along the series HF, HCl, HBr, HI — which mirrors exactly
the trend in leaving group ability.
The acidity of hydrides increases as we move across the Periodic Table:
CHzg is not at all acidic and hence has an extremely high pK, (around 48), NH3
(pK, = 38) can only be deprotonated by very strong bases, H2O is a very weak
acid (pK, = 15.7) and HF is acidic, but still only weakly so (pK, = 3). As
we saw for leaving groups, this order reflects the stability of the ions formed
on dissociation of the acid, with F~ being the most stable due to the large
effective nuclear charge on fluorine, and CH, being the least stable as a result
of the smaller effective nuclear charge of carbon.
Whilst we see the same general trends in both leaving group ability and
acid strength, the order of leaving group ability from carbon does not match
exactly the order of the acidity of the conjugate acids (see Fig. 13.6). For
example, H2O is a better leaving group than Cl, yet HCl is a stronger acid
than H30*. Such discrepancies are due to the fact that how good something is
as a leaving group depends on its bond strength to carbon, whereas how strong
its conjugate acid is depends on its bond strength to hydrogen, the two are not
necessarily the same.
Leaving group ability is a good guide to working out which reactions will
go and which will not, but often it is not as clear cut as the examples given here.
This brings us to the last topic which is the more complex one of what happens
when there are feasible alternative reactions open to a given set of reactants.
14. Competing reactions
In this final chapter we will look at what happens when there are a number of
possible outcomes for a given set of reactants i.e. when there are alternative
pathways available. To understand and'rationalize which of these pathways is
the dominant one and — more importantly — to understand how to alter the bal-
ance between different pathways, will need us to bring together all the concepts
and ideas we have been developing in the previous chapters.
14.1 Reactions of carbonyls with hydroxide

We will start by thinking about the initial reaction between hydroxide ion and
an aldehyde or ketone. To keep the structures and orbitals as simple as possible,
we shall look at the reaction between hydroxide and ethanal (acetaldehyde). As
we saw in Section 7.3 on p. 102, the hydroxide ion could act as a nucleophile
and attack the C=O m* MO to form a tetrahedral intermediate. The 7* MO
H
is shown in Fig. 14.1. As we have seen before, this MO is essentially just the
H, 7 out-of-phase interaction between the carbon 2p and oxygen 2p AOs; the MO
O C
ao
—C
\
has a greater contribution from carbon than from oxygen.
The hydroxide attacks at the carbonyl carbon for two reasons: (i) due to
H
the effect of the electron-withdrawing oxygen, the carbony] carbon is slightly
Fig. 14.1 The LUMO of ethanal is the 1*
positively charged; and (ii) the 7* LUMO has a greater contribution from the
MO which has a greater contribution from
the carbon 2p AO than from the oxygen carbon AO than from the oxygen. The curly arrow description of the attack by
2p. OH™ is shown in Fig. 14.2 (a).
o> HO ©®
no® VI —@) \s
H~ ~CHg H~ ~CH3
1d) 0 O
WEDS (b) Ho” |
C C
H~ ~CHg H~ ~CHg
Fig. 14.2 In (a) hydroxide attacks the carbonyl carbon, simultaneously breaking the C=O x bond to
form a tetrahedral intermediate. This intermediate breaks down, as shown in (b), by re-forming the 2
bond and eliminating the only realistic leaving group, which is the hydroxide ion which attacked to form
the tetrahedral intermediate in the first place.
Once formed, the only thing the tetrahedral intermediate can do is re-form
the ketone by eliminating the only realistic leaving group ~ the hydroxide ion
that just attacked; CH, and H™ are simply not good enough leaving groups
to compete with the OH™. The curly arrow description of the collapse of the
tetrahedral intermediate is shown in Fig. 14.2 (b).
An equilibrium will be established with both the intermediate and the alde-
hyde present. For an aldehyde such as ethanal, significant amounts of the in-
termediate will be present, but for a ketone, the equilibrium lies very much in
favour of the carbonyl.
Formation of enolates
The other way the hydroxide could react with the carbonyl compound is as a
base; deprotonation leads to an enolate anion of the type shown in Fig. 10.34
on p. 170. The enolate anion formed from ethanal is shown in Fig. 14.3.
enolate anion
Fig. 14.3 A strong base may remove one of the hydrogens from the methyl group in ethanal to form an
enolate anion. This can be drawn with the charge on either carbon or oxygen; the best representation
is probably that with the charge on the more electronegative oxygen.
We shall now look more closely to see why hydroxide can do this. What
we will find is that the C-H bonds on the carbon adjacent to the carbonyl are
weakened as a result of interaction between their orbitals and the m system.
This, combined with the fact that the negative charge on the enolate can be
delocalized on to oxygen, results in these hydrogens being significantly more
acidic than in simple alkanes.
We saw on p. 160 that in an amide there is an interaction between the
nitrogen 2p orbital and the C=O 7 system which leads to the formation of z
MOs which are delocalized over the carbon, nitrogen and oxygen. The ‘lone
pair’ on nitrogen is not, therefore, localized on that atom; rather, it participates
in a delocalized system (see Fig. 10.18 on p. 161).
In an analogous fashion, whilst we might like to think of the C-H and C=O
mz bonds in ethanal as being separate, the computed MOs show there is some
degree of interaction between the C—H o bonding orbitals of the methyl group
and the a system. Figure 14.4 (a)(i) shows a surface plot of a x MO which
contributes to the bonding between the carbon and hydrogens in the CH3 group
in ethanal. The MO may be thought of as arising from an interaction between
the methyl carbon 2p AO and the 1s AOs of the hydrogens above and below
the plane of the o-framework; this interaction is illustrated in Fig. 14.5.
Figure 14.4 (a)(ii) is a surface plot of the same MO but at a lower value
of the wavefunction than in (a)(i). Whilst (a)(i) shows that this MO has most
electron density on the methy! carbon and the attached hydrogens, (a)(i1) shows
that electron density is also, to a lesser extent, delocalized over the 2 system.
Similarly, Fig. 14.4 (b)(i) shows the C=O mw bonding MO resulting from
the in-phase combination of the C and O 2p AOs. As usual, there is a greater
contribution to this MO from the oxygen AO than from the carbon AO. How-
ever, there is also a small contribution from the C-H o bonding MOs on the
methyl carbon, as shown in Fig. 14.4 (b)(ii) which is plotted at a lower value
of the wavefunction.
The x* anti-bonding MO, shown in Fig. 14.4 (c)(i), is made up largely
from the out-of-phase combination of the C and O 2p AOs, with a greater
contribution from the carbon than the oxygen. As can be seen from (c)(ii),
218 Competing reactions
(a)(1)
(a)(Il)
Fig. 14.4 Surface plots of the 1 MOs of ethanal, i.e. the orbitals for which the plane of the molecule is
a nodal plane. The plots in the upper half of the figure are made for a higher value of the wavefunction
than those in the lower half. Thus the plots in the upper half just show where the electron density
is most concentrated, whereas those in the lower half show the atoms on which smaller amounts of
electron density are found. Orbital (a) essentially just contributes to the bonding between the methyl
carbon and its attached hydrogens. Orbital (b) is essentially the C=O x bonding MO and (c) is the 1*
anti-bonding MO. All the orbitals are to some extent delocalized over the oxygen atom, both carbon
atoms and the two out-of-plane hydrogen atoms.
even this MO has a small contribution from the C-H o bonding MOs.
What these three MOs tell us is that there is some weak o conjugation
between the C-H o bonding MOs of the methy] group and the z system; this
C 2p AO is shown schematically in Fig. 14.6. Note that this conjugation is not possible
™ A)
H>_ ae two H 1s AOs
for the C-H bond between the carbonyl carbon and the attached hydrogen as
2 eG
O
mcr » the o orbitals for this bond lie in the plane and so are not able to interact with
aa
H the x system. In contrast, the orbitals from the C-H bonds of the methy] group
lie out of the plane so that interaction with the 2 system is possible.
Flg. 14.5 The p orbital on the methyl
carbon can contribute to the bonding of The effect of this o conjugation is that some electron density is withdrawn
the methyl hydrogens by overlapping with from the methyl C—H bonds towards the oxygen. This weakens the bonds and
the 1s orbitals on the two out-of-plane
makes the hydrogens more acidic than they would be in a simple alkane. We
hydrogens.
saw a similar example on p. 179 where neighbouring C-H o bonds helped
to stabilize a carbenium ion. We also saw how the resulting withdrawal of
electron density electrons from the C-H bonds made the hydrogens sufficiently acidic that they
pulled towards O
could be removed by a base, such as hydroxide, and how this led to elimination
reactions (see Section 11.3 on p. 180).
(2 out-of-plane In ethanal, this o cgnjugation increases the acidity of the hydrogens at-
OOD oe ¥ C-H —H oo orbita
orbital
tached to the carbon adjacent to the carbonyl]. In addition, the enolate anion
which results from the removal by a base of one of these hydrogens is delo-
calized. Taken together, these two effects account for the fact that it is much
p orbitals on C and O
easier to remove H* from either an aldehyde or a ketone than from a simple
Fig. 14.6 The C-H o bonding MOs of the alkane. Typically, these carbonyl compounds have pK, values of around 20, in
methyl group and the 2p orbitals on the
carbonyl carbon and the oxygen can ali
contrast to simple alkanes which have values of between 40 and 50.
interact to form a delocalized set of Figure 14.7 shows two curly arrow descriptions of the formation of an eno-
MOs. Note that for this interaction to be late. Although both mechanisms are acceptable, (b) is preferred since it em-
successful the C—-H bonds must lie out of
the plane of the molecule. Electron
phasizes the role of the C=O bond withdrawing electrons from the C-H bond,
density is pulled towards the oxygen due making the hydrogen acidic in the first place. Also, as we saw in Fig. 10.34 on
to its greater effective nuclear charge. p. 170, when representing the enolate ion it is better to have the negative charge
oF i
aowcal == He Not + HO
,, e
Y*, |
io H
HO
(b) CO "9
7 Cc ~c~ H — ~ «=> Cc So + H50
% |
H H
Fig. 14.7 Two curly-arrow representations of the formation of an enolate. Version (a) shows the depro-
tonation by hydroxide ion to form the enolate; this version does not emphasize the electron-withdrawing
nature of the C=O bond which is what makes the hydrogen acidic in the first place. The effect of the
carbonyl is represented in version (b); in addition, this also suggests the enolate formed has a greater
charge on the oxygen than on the carbon.
on the more electronegative oxygen atom. However, the structure which has
the charge on the carbon does make a contribution, so we should think of the
enolate as having significant charge on both the oxygen and the carbon.
In summary there are two possible reactions that could occur between hy-
droxide and ethanal: the first is the direct attack of the hydroxide into the 7*
MO of the ethanal to form the tetrahedral intermediate; the second is the de-
protonation of one of the methyl hydrogens to form the enolate. The question
is, which reaction actually occurs? The answer to this is simple: they both do.
Both of these reactions are reversible so at any one time the reaction mixture
will contain the unreacted hydroxide and aldehyde, the tetrahedral intermedi-
ate and the enolate. We shall see later that there may also be further reaction
products.
How do we know that both reactions are occurring? Experimental evi-
dence for this comes from studies using isotopically labelled reagents, details
of which are given in the next section.
Isotopic labelling
It is possible to prepare water or hydroxide where the normal 160 atoms or hy-
drogen atoms are replaced with their heavier isotopes !80 and deuterium (D or
2H), When !8Q-labelled hydroxide and water are mixed with '6Q-containing
aldehydes and ketones, it is found that the heavier oxygen isotope is readily in-
corporated into the carbonyl compounds. The mechanism by which this occurs
in shown in Fig. 14.8.
Similarly, if deuterated water (D2O) and deuteroxide (OD~ ) are used, the
methyl hydrogens are rapidly exchanged for deuterium atoms. The mechanism
for this reaction is shown in Fig. 14.9.
h eu" TON
18 fo)
H 18 —H
129% \| _@ ~ Ws oH). NS
a’
H~ ~CHs H~ ~CHg H CHs
4 H H H*
06> oH 109%) 9-H 189 20
: oO. 9. |
Cc NX
H~ ~CHs H~ ~CHs H~ ~CHs
Fig. 14.8 Illustration of how '80 from '8OH- can be incorporated into carbonyl compounds. Step
(a) shows the initial attack of the labelled hydroxide on the ethanal. The best leaving group after this
step is still the labelled hydroxide that just attacked — we need to make the unlabelled oxygen into the
best leaving group. In water, there is extensive hydrogen bonding between the solvent and any -O—
ions; this aids the tetrahedral intermediate in protonating the -O~ and deprotonating the OH group.
Thus step (b) shows the intermediate deprotonating the solvent and (c) shows the labelled hydroxide
deprotonating the other OH group. The net effect of these two steps is that the unlabelled oxygen now
becomes the better leaving group (as OH ) and is then pushed off in step (d) by the O~ to yield the
180-labelled product. All of the steps are reversible, so given the large excess of '80-labelled water
and hydroxide, eventually almost all of the initial 160 in the ethanal will be replaced.
C HH =6__@) Cx. LH
H
ou po
O
D~ D
<0 ? i
u- C re H (b)b HZ Cc Nov H repeating
- steps
ae HZ C Sov D
+ a) an 7
O7D H ep H D D
JU a
Fig. 14.9 A mechanism showing how deuterium from OD~ and D2O can replace the methyl hydrogens
in ethanal. In step (a) one of the out-of-plane hydrogens is removed by the deuteroxide. This step is
reversible so the enolate could piek up this hydrogen again but, given the large excess of the deuterated
water, it is more likely that the enolate will react with a fully deuterated water molecule instead, as
shown in step (b). Rotation of the C-C bond means that any of the methyl hydrogens could eventually
be replaced in this manner by repeating steps (a) and (b). Note that the aldehyde hydrogen could not
be replaced in this manner since, as explained in the text, it is not significantly acidic.
Energy profiles for the alternative reactions

The isotopic labelling experiments show that both reactions are taking place in
solution and that all the reactions are reversible. A schematic energy profile for
both of these reactions is shown in Fig. 14.10.
The attack of the hydroxide into the z* MO to form the tetrahedral inter-
mediate is fast (as shown by the small activation energy for this step, Ea tet) but
we see from the profile that the collapse of the intermediate back to the alde-
14.2
energy Reactions of enolates 221
Es enolate
AE enolate
fo\es
Sow H
H H,O H
tetrahedral ; ;
intermediate starting materials enolate anion
reaction coordinate
Fig. 14.10 A schematic energy level profile for the equilibria between ethanal, the tetrahedral inter-
mediate (formed by the hydroxide attacking the C=O 2*) and the enolate (formed by the hydroxide
removing one of the methyl hydrogens). Overall, the lowest energy species are the starting materials.
As Ea tet is smaller than E2 enolate. the tetrahedral intermediate forms faster than the enolate. Both
intermediates are higher in energy than the reactants, so at equilibrium the concentration of these
intermediates will be small.
hyde and hydroxide is faster still. The intermediate is higher in energy than the
starting materials and so at equilibrium there will be rather little of it present.
The energy profile also shows that the enolate is slower to form than
the tetrahedral intermediate as the activation energy for enolate formation,
Eaenolate, iS higher than for formation of the tetrahedral intermediate, Eo tet.
Formation of the enolate is slow because the methy] hydrogens are only weakly
acidic so there is hardly any hydrogen bonding to these hydrogens. As a result,
a hydroxide is not always present in exactly the right position to take off the
hydrogen.
Since both the enolate and the intermediate are higher in energy than the
starting materials, the equilibrium mixture will contain more aldehyde and hy-
droxide than either of the intermediates.
The enolate anion undergoes many useful reactions and so it is often used
in synthetic chemistry. Of course, at the same time as forming the enolate we
almost certainly will form some of the tetrahedral intermediate, but usually this
is not important as this intermediate does not lead to further reactions.
14.2 Reactions of enolates

The enolate has significant negative charge on both the oxygen and the carbon
next to the carbonyl carbon (see Fig. 14.7 on p. 219). We might therefore expect
that when the anion acts as a nucleophile it can do so either by attacking from
the carbon or the oxygen; as the charge is not equally distributed, we would not
expect the reactivity at these two positions to be the same. These expectations
are borne out by experiment, and in addition it is found that exactly how the
Fig. 14.11 The 7 MOs of the enolate anion. The lqwest energy MO is all in-phase and is shown in (a).
Unlike the allyl anion, the enolate anion is not symfnetrical due to the oxygen on one end; this draws
more electron density towards it in this MO. MO (b) includes an extra node and is higher in energy. It
mainly has electron density on the two end atoms but with a greater contribution from what was the
methyl carbon. MOs (a) and (b) are occupied; MO (c) is the 2* anti-bonding MO and is unoccupied.
enolate reacts depends on what it is reacting with and on the conditions of the
reaction — even changing the solvent or the counter-ion can have a large effect
on the outcome of the reaction.
A species such as the enolate anion which can act as a nucleophile through
two different sites is known as an ambident nucleophile. Calculations suggest
the charges are approximately —0.8 on the oxygen, +0.3 on the carbonyl car-
bon and —0.6 on what was the methyl carbon. The form of the 7 MOs, shown
in Fig. 14.11, is consistent with the idea that most of the charge is on the two
end atoms. This is similar to other cases in which a m system is formed from
the overlap of three p orbitals, e.g. the carboxylate anion (Fig. 10.16 on p. 159)
and amides (Fig. 10.18 on p. 161).
The MOs in Fig. 14.11 are for the free anion. In practice, however, a
counter-ion must also be present, and one of the most commonly used is Lit.
The first question is which end will the counter-ion be most associated with?
Lithium will not form a strong covalent bond with either carbon or oxygen,
simply because the 2s orbital on lithium is not well matched in energy with
either the carbon or oxygen AOs (see the graph of orbital energies on p. 57).
Of the two, the better match is with the carbon orbitals so we might expect the
lithium to be more associated with the carbon end of an enolate as shown in
Fig. 14.12 (a).
(a) ° (b) o Li®

H “
C Sow
Li H a“
Cx“on OH
‘YH H H
Fig. 14.12 In a lithium enolate we might expect the lithium to be covalently bound to the carbon, as in
(a), on the grounds that the carbon AQs are better matched in energy to those on lithium than are the
oxygen AOs. However, the electrostatic interaction, shown in (b), turns out to be much stronger, and
so (b) is the favoured structure.
However, there is an alternative which is for the cation simply to have an

electrostatic interaction with the enolate. Such an interaction will be strongest
between the Li* and the oxygen on the enolate as this is the position with
the highest electron density; the resulting structure can be represented as in
Fig. 14.12 (b). Experimental work confirms that this structure is indeed the
one adopted.
14.2 Reactions of enolates 223
(a) 4
4, ©
al C, I Hruc
A 7 O . : o
‘ H | very little reaction i
aah via oxygen end WON od
| |
H
(b) VI
HC,
O van O H {°
Il 4 H main reaction | |
a Cc Cum
H~ owt via carbon end H~ Sow a
l “ce
H H H
(c) VI
H—C,,
4 H main reaction I |
“Sol via carbon end Hm Sow A
| 4:
H H H
Fig. 14.13 Possible mechanisms for the reaction between the enolate from ethanal and CHgl. In (a)
the reaction is via the oxygen end of the enolate, whereas in (b) the carbon end of the enolate is
involved; experimentally it is found that most of the reaction takes place through the carbon. CHgl is
uncharged so there is very little electrostatic interaction and as a result the reaction is controlled by
orbital interactions. The HOMO of the enolate has the greatest contribution on the end carbon and
therefore the reaction is expected to take place through this position. Mechanism (c) has the same
outcome as (b), i.e. attack through the carbon, but in (c) the enolate has been drawn in a more realistic
manner with the charge shown on the oxygen.
Suppose that we now allow iodomethane, CH3lI, to react with the enolate.
In principle, either end of the enolate could attack the CH3I in an Sn2 mech-
anism to form two alternative products as shown in Fig. 14.13 (a) and (b). As
the reaction ultimately involves adding a methyl group to the enolate, CH3lI is
called a methylating agent or, more generally, an alkylating agent. Experimen-
tally what is observed is that the only product is the one in which the reaction
is via the carbon end of the enolate; our task is to explain why this is.
Iodomethane is not charged so there is going to be very little electrostatic
attraction between it and either end of the enolate anion. As outlined in Sec-
tion 7.5 on p. 107, the poor energy and size matches between the carbon and
the iodine AOs means that the C-I o* orbital will be particularly low in energy
for an unoccupied MO. Thus a good orbital interaction is possible between the
enolate and CH3I; this is in contrast to the case of Lit interacting with the eno-
late where there is a good electrostatic attraction but poor orbital interaction.
As we saw on p. 101 the most significant orbital interaction is between the
HOMO of one species and the LUMO of the other. For iodomethane reacting
with an enolate, this interaction will be between the HOMO of the enolate and
the LUMO of the iodomethane. Figure 14.11 shows that the HOMO of the
enolate (orbital (b)) has a greater contribution on the end carbon than on the
oxygen, so the best orbital interaction is from the carbon rather than the oxygen
end of the enolate. Therefore the reaction takes place through the end carbon
rather than through the oxygen.
The mechanisms for the reaction through both ends of the enolate are
shown in Fig. 14.13 (a) and (b). Although most of the reaction takes place
through the carbon we do not have to draw the enolate with the charge local-
ized on the carbon. We mentioned earlier that probably the best representation
of the enolate ion is to have the negative charge on the oxygen end; using this
representation of the enolate we can still draw a mechanism for the reaction
through carbon, as shown in Fig. 14.13 (c).
Manipulating the reaction

— by changing the alkylating agent
It is possible to increase the amount of the product in which the oxygen be-
comes alkylated, i.e. in which the methyl group is attached to the oxygen. To
do this, we need to make the interaction more charge controlled rather than
being controlled just by orbital energy factors. Replacing CH3I by CH3Br or
CH3Cl accomplishes this. Bromine is more electronegative than iodine, so
CH3Br has a greater dipole moment than CHs3I. Electrostatic interactions are
thus more important, and so there is more tendency for the reaction to take
place through the oxygen of the enolate. The effect is even greater for CH3Cl
which has a greater dipole moment still.
As an example, we will look at how the ratio of carbon- to oxygen-alkylated
product varies in the reaction shown in Fig. 14.14 as the halogen, X, in the
alkylating agent is changed.
When iodoethane is used as the alkylating agent, the ratio of oxygen-
to carbon-alkylated products is 13:87. With bromoethane the ratio becomes
39:61 whilst with chloroethane the ratio increases further to 60:40. Clearly the
amount of the oxygen-alkylated product increases as the halide becomes more
electronegative.
Ko Oo
| ll
H3C~ on C oc.Hs + CH3CHs
athe —X _
CoHs
a ae
~ O
H3C
ww
“om ~OCzHs
re H3c~ ~C~
or OCpHs
H H CoHs
product formed by product formed by
reaction via oxygen reaction via carbon
Fig. 14.14 The reaction between an enolate anda halogenoalkane can result in alkylation
at either Oxy-
gen or carbon. The ratio between these two products depends on, amongst other
things, the halogen
X. It is found that the proportion of the oxygen-alkylated products increases as the
electronegativity
of the halogen increases; we interpret this as the increased dipole moment of the
halogenoalkane
leading to the reaction becoming more controlled by electrostatic interactions.
The amount of oxygen-alkylated product may be increased still further

when ethyl tosylate, CpHs5OTs, is used in place of the halogenoethane. The O
oO
structure of ethyl tosylate and the curly arrow mechanism for the reaction is N Lo) chs
shown in Fig. 14.15. As can be seen in the diagram, it is the CH? carbon of the mo
HH
ethyl group which is attacked by the enolate in an Sy2 mechanism. However,
Enolate
instead of a halogen, this carbon is now attached to an even more electron-
withdrawing oxygen, which increases the partial positive charge on the carbon
still further. This results in the formation of even more of the oxygen-alkylated
product — the oxygen-:carbon-alkylated product ratio is now 88:12. CH;
Notice that although we wanted to have an oxygen attached to the ethyl Fig. 14.15 The curly arrow mechanism
group to increase the charge on the carbon, it would not have been sufficient for the reaction between ethyl! tosylate
simply to use an OH group for this purpose. As we saw in Section 13.2 on and an enolate; in this Sy2 reaction the
enolate can attack through carbon or
p. 213, OH™ is not a good leaving group but the tosylate anion, TsO™, is. oxygen.
Thus, using ethyl tosylate provides an increased positive change on the carbon
as well as a good leaving group.
— by changing the solvent

Changing the solvent can also influence the outcome of the reaction, i.e. the
ratio of the carbon- to oxygen-alkylated products. The data quoted so far for
the reaction shown in Fig. 14.14 was obtained in the solvent hexamethylphos-
phoramide (HMPA) whose structure is shown in Fig. 14.16. This solvent is
classed as polar aprotic (see Section 12.1 on p. 190). It is very good at sol-
vating cations but, as it has no acidic hydrogens, it does not solvate anions so
Fig. 14.16 The structure of the solvent
well, as no hydrogen bonding to them is possible.
hexamethylphosphoramide (HMPA).
Using HMPA as the solvent means that the Kt ion is strongly solvated
whereas the enolate anion is not so strongly solvated and so is free to react. In
contrast, if the reaction is carried out in the polar protic solvent t-buty! alcohol
(2-methyl-2-propanol), even the ethyl tosylate forms exclusively the carbon-
alkylated product.
The reason for this change is because the polar protic solvent forms hydro-
gen bonds to the oxygen end of the anion. These hinder the approach of the
alkylating agent to the oxygen and so reaction through the carbon is favoured.
14.3. Unsymmetrical enolates

There is only one enolate which can be formed from ethanal, but for unsym-
metrical ketones two different enolates can be formed, depending on which
hydrogen is removed. An example is shown in Fig. 14.17; in this ketone there
is an extra methyl group on one of the carbons adjacent to the carbonyl, so two
different enolates can be formed.
By careful control of the reaction conditions we can favour one of the eno-
lates over the other. The first point to note is that we will need to use a strong
base to take off one of the acidic hydrogens; however, we do not want the base
to act as a nucleophile and attack the carbonyl group. One way to avoid this
is to use a bulky base which is simply too large to get in close enough to at-
tack the carbonyl; an example of such a base is triphenylmethyllithium, whose
structure is shown in Fig. 14.18.
(a) <9 cc)

| |
Hi ON AH Hie SG-H
wor 2 does
2
Heer Nd
enolate A 2
(b) O 0°
Base” ) =
H “ Vow H —-. 1H
C cy H3C —C on
Hc™ \ | YH ——_—— rn
HeC — CHe enolate B HeC — CHa
Flg. 14.17 Two different enolates can be formed from an unsymmetrical ketone, depending on which
hydrogen is removed by the base.
A consequence of using such a bulky base is that it is easier for it to remove

the hydrogens from the carbon which does not have the methyl group attached,
simply because these protons are less sterically hindered. The result is that
enolate A is easier to form than enolate B.
(-) (+) However, although enolate A is easier to form than enolate B, it turns out
ww —Li . . .
" that enolate B is actually lower in energy than enolate A. The reason is to do
with different amounts of o conjugation present in the two enolates.
We have seen how any out-of-plane hydrogens can o conjugate with a
neighbouring z system. In enolate B the hydrogens on the methyl group are
able to o conjugate with the 2 system, whereas in enolate A the methyl hy-
Fig. 14.18 The structure of triphenyl- drogens cannot be involved in such conjugation. As we have seen, conjugation
methyliihium, PhgCLi, which is a song
base but, on account of its bulk, a poor &
enerally y
lowers the energy BY of the species,
pecies,
so we therefore expect xpect enolate
enolate B t 0
nucleophile. be lower in energy than enolate A.
This o conjugation means that the electrons in the methyl group are, to a
certain extent, spread over the whole z system. This can be seen in Fig. 14.19
which compares the HOMOs of the two enolates; the extra delocalization in
enolate B is clear.
ae.
io
enolate A enolate B
Fig. 14.19 Surface plots of the HOMOs of enolate A and enolate B shown in Fig. 14.17. In enolate B
the hydrogens on the methyl group are involved in o conjugation with the 7 system, something which
is evident from the way in which the HOMO is spread over all of these atoms. In contrast, no such
conjugation is possible for enolate A, as evidenced by the HOMO. This extra delocalization makes
enolate B lower in energy than A.
energy
wa Cc Now
Hsc™ \ | YH
H2C
— CH,
S) -) (+) cS)
9 (CgHs)30 —Li 9
starting materials
Hi NG -H H oF Newt
Hsc™ \ / i rr al
H2C— CH, HsC— CH»
enolate A enolate B
reaction coordinate
Fig. 14.20 Energy profile for the formation of enolates A and B by the reactions shown in Fig. 14.17.
As explained in the text, enolate A forms more rapidly than enolate B on account of the steric crowding
caused by the methyl group; on the profile this difference in the rates of reaction is shown by the
smaller activation energy for the formation of enolate A. However, enolate B is actually lower in energy
than enolate A, on account of the o conjugation which is possible for B and not A.
Thermodynamic and kinetic products

As we have seen whilst it is easiest, and hence quickest, to form enolate A,
enolate B is lower in energy; this is illustrated in the energy profile shown
in Fig. 14.20. Enolate A is described as the kinetic product as it is formed
fastest, whereas enolate B is described as the thermodynamic product as it
is thermodynamically the most stable enolate and so present in the greatest
amounts at equilibrium. The question is, how can we manipulate the reaction
conditions to favour one enolate or the other?
When an excess of the base is added to the ketone at room temperature,
the ratio of the enolates formed is 72:28 A to B. Once formed, these enolates
cannot return to the ketone since in the presence of excess base there is nothing
acidic for the enolates to pick up a hydrogen from. Now if an alkylating agent
is added which reacts with the enolates the ratio of the two alkylated products
formed will be the same as the ratio of the two enolates.
Returning to the formation of the enolates, if instead of having excess
base we have a slight excess of ketone, then under the same conditions (i.e.
same time and temperature) the ratio of enolates A to B becomes 6:94. Now,
the lower energy (more thermodynamically stable) enolate B is the dominant
species. This comes about because it is now possible for the enolates to return
to the ketone by deprotonating some of the excess ketone present; this is shown
in Fig. 14.21.
<0 i
H,
Hn SoH O ye™ owt
S\ I co7
Hc \ ~ SH
Hoo
° HC — —CHp H/C
Le cunt HoC—CH,
enolate H3C™ \ / YH ketone é 4
HeC — CHo H3C ~ 7% New
ketone hy \ | SH
Het — CHe
enolate
Fig. 14.21 Curly arrow mechanism showing how enolate A can return to the ketone by picking up
H+ from another ketone, forming enolate B in the process; it could just as well have formed another
molecule of enolate A. Similarly, enolate B could react with a ketone to form enolate A or B. In other
words, there is a pathway which allows both enolate and the ketone to come to equilibrium.
Of course, all this reaction does is convert enolates to ketones and vice
versa. However, because the reaction is reversible it allows the enolates and
the ketone to come to equilibrium:
enolate (A or B) + ketone = ketone + enolate (A or B).
Therefore, since enolate B is thermodynamically more stable, more of this

enolate will be present at equilibrium. Subsequent reaction with an alkylating
agent will then lead to more of the final product being formed from enolate B
than from enolate A, provided that the alkylation reaction does not perturb the
equilibrium significantly.
14.4 Substitution versus elimination revisited

The reaction between ethanal and hydroxide is relatively straightforward to
understand since there are only two reasonable possibilities for the initial in-
teraction between the two reagents. For the majority of reactions carried out in
the laboratory, many outcomes are possible and it is the task of the chemist to
try to optimize the conditions to produce the desired product.
We have already seen in Chapter 11 how a base/nucleophile reacts with
halogenoalkanes to form either substitution products or elimination products.
Thus, as shown in Fig. 14.22, 1-bromopropane could react with hydroxide to
form either propan-1-ol’or propene. For less symmetrical halogenoalkanes,
H H H HH
H3C
2"N Sere
yy OH
AZ r
_ H3C~
—
vo"
Lo H
+ H3C~ “or
HH H HH
elimination substitution
product product
Fig. 14.22 The halogenoalkane 1-bromopropane can react with OH~ as a nucleophile to give the
substitution product propan-1-ol, or with OH~ as a base to give the elimination product propene.
H »Br
Cat “Ny, NaOCH
Hy in CH30H
H;CO H '
C3H Cc C3H7 Cy LH
= oh |
H y HoH H
2-hexyl methyl ether 1-hexene
cs
27 % 19%
+ CaH7 c Cw CH3 + C3H

7. ~N C72 Cc _~ H
H
trans-2-hexene
:
cis-2-hexene
42% 12%
Fig. 14.23 The reaction between 2-bromohexane and methoxide, CH307, gives one substitution prod-
uct (the ether) and three different alkenes which arise from elimination reactions. The proportions of
each product are shown.
more than one elimination product is possible; for example, the reaction be-
tween 2-bromohexane and methoxide, CH30-, gives four different products
as shown in Fig. 14.23.
The ether is formed by a substitution reaction between the bromohexane
and methoxide, following the Sy2 mechanism as shown in Fig. 14.24 (a). The
elimination products are formed via the E2 mechanism, which is shown in
Fig. 14.24 (b) for the case of formation of trans-2-hexene.
|
So7 Br ‘ow vor
H Cc Cri,
H C3H7 H C3H7
o)
Br
(b)
CH30 CHay CH30H
H/\ a
co tay witCH
H 4 C3H7 Cp Cc 3ht7
‘ H 5
Br
Fig. 14.24 Curly arrow mechanisms for the reaction of 2-bromohexane with methoxide via (a) an
Sy2 substitution reaction and (b) an E2 elimination reaction. There are three possible products from
elimination reactions (see Fig. 14.23), of which only one is shown here.
Altering the ratio between the substitution and elimination products

How can we make sense of the relative amounts of these products and increase
the yield of one product over the others? In forming the ether, the methoxide
ion is acting as a nucleophile; it is firstly attracted towards the partial positive
charge of the carbon with the attached bromine atom, and then attacks into the
C-Br o* orbital, thereby breaking the C—Br bond. If we can increase the partial
positive charge on the carbon, there will be greater electrostatic attraction to the
CH307 group, thus favouring the substitution reaction.
This is exactly what we see if the bromine atom is replaced by a chlorine or
fluorine atom; whereas the reaction of 2-bromohexane gives 27% of the ether,
2-chlorohexane gives 38% and 2-fluorohexane more still. In contrast, replacing
the bromine by iodine means the partial charge on the carbon is reduced and it
is found that the amount of the ether decreases to 16%.
However, whilst the amount of the substitution product formed increases on
replacing the bromine atom by the more electron-withdrawing fluorine atom,
the rate of the reaction decreases by a factor of over a thousand. This is be-
cause the fluoride is a poorer leaving group than bromide, as was described in
Section 13.2 on p. 213.
It is possible to increase the amount of the substitution product further by
replacing the bromine by a really good leaving group which is still capable of
polarizing the carbon to a significant extent. An example of such a leaving
group is the tosylate (Fig. 13.7 on p. 214); if this is used in place of bromine
the fraction of the substitution product increases to over 70%. Previously, when
looking at the reactions of enolates, we saw how the tosylate group could be
used in a similar way to both increase the partial positive charge of the attached
carbon and also provide an excellent leaving group.
The percentage of the substitution product can be increased still further
by making the elimination reactions more unfavourable. In these reactions,
CH307 is acting as a base, so not surprisingly changing to CH3OH, which is
a much weaker base, reduces the proportion of the elimination products. If use
of this weaker base is combined with replacing the bromine by the excellent
tosylate leaving group, the fraction of the ether rises to 90%.
Energy profiles
Drawing an energy profile in the case where there are several different reaction
pathways is quite difficult since we need to show the relative energies of all
the species and their transition states. Let us consider the case where reactants
X can form four possible products A, B, C, and D, as was the case in the
reaction shown in Fig. 14.23. Figure 14.25 is a schematic energy profile for
these reactions; the reaction coordinate is represented by movement on the xy-
plane and the energy by the height of the surface. The reactants and the four
products are all energy minima, with the reactants being shown in the centre of
the diagram and the four products towards the corners.
On the surface, the dotted lines trace the lowest energy paths between the
reactants and each of the products. The highest energy point along such a path
for each of these paths is the transition state; A? for forming A, and so on.
Fig. 14.25 The lower figure shows a perspective view of a schematic energy profile for the situation
where reactants X can form four possible products A, B, C, and D. The height of the surface indicates
the relative energy of the species and movement in the xy-plane represents the reaction coordinate.
The dotted lines over the surface indicate the paths of minimum energy between reactants and par-
ticular products; the transition state (denoted by t{) is at the maximum energy point along each path.
The upper figure shows a contour plot for the same energy profile; the contours are labelled in order
of increasing energy.
The reactants can form any of the products and given enough energy, any
of the products could return to the reactants. In order to understand which will
be the main products, we need to consider the relative depths of the energy
relative energy
minima, i.e. the relative energies of the reactants and products, and also the
nnn
energy barriers between these minima which must be overcome if the product
is to form. The energy profiles along the paths between the reactants and each
of the products are shown in Fig. 14.26.
wo
As we have seen in Section 9.4 on p. 137, as there is an energy barrier (the

activation energy) which reactions have to overcome, changing the temperature
has a large effect on the rate constant. We also saw (Fig. 9.8 on p. 138) that the
Fig. 14.26 Energy profiles for the paths
greater the activation energy, the larger the change in rate constant for a given between the reactants and each of the
change in temperature. products shown in Fig. 14.25.
If there are several pathways open to a reaction, the one with the lowest
activation energy will always have the largest rate constant. However, the rate
constant for each pathway will respond differently to a change in temperature
on account of the differences in their activation energies. As the temperature is
lowered, all of the rate constants will decrease, but the decrease will be greatest
for those pathways with the highest activation energies. At low enough tem-
peratures, the pathway with the lowest activation energy will therefore come to
dominate over all the others. “eS
Looking at the profiles in Fig. 14.26 we can see that at low temperatures the
main product that will be formed will be A, simply because the activation en-
ergy required to reach this product is the lowest. However, A is not much lower
in energy than the starting materials and so we would not expect the reaction
to go to completion. The next most abundant product at low temperatures will
be D since this has the next lowest activation energy.
As the temperature is increased, we will start to see more of products B and
C. However, it is easier for C to return to the starting materials than it is for B
since the activation energy for C returning to X is smaller than for B returning
to X. Hence we should not expect too much C to form — if there is enough
energy to form it in the first place, there will be enough energy for it to return
to the starting materials and eventually to form the lower energy products.
When the temperature is high enough for all the reactions to be readily
reversible, the ratio of the products will simply be dependent on their relative
energies: since B has the lowest energy, at equilibrium it will be the major
product. D is the species next lowest in energy, so it will be the next most
C3H7 —C=C—CH3 abundant product.
2-hexyne From this energy profile we can see that it is never possible to prepare
much of C since other reactions are always more favourable; if we did want
to prepare C, we would have to use a different reaction which would have a
He different energy profile. Altering the energy of the reactants (for example by
metal catalyst changing a leaving group as we did for the reaction in Fig. 14.23 or even by
changing the solvent) will change the whole energy surface. The energies of
the reactants and products will change, as well as the energy barriers. It might
C3H7 CH3
be that on the new surface C is a more favourable product.
\
For example, in the reaction shown in Fig. 14.23, it is never possible to
C=C
/ \ generate more cis-2-hexene than trans-2-hexene no matter how we alter the
H H
temperature or change the leaving group. This is because the activation en-
cis-2-hexene
ergy for forming the cis isomer is always higher than the activation energy for
Fig. 14.27 Although cis-2-hexene cannot
forming the trans isomer; also the trans isomer is lower in energy than the cis.
be prepared in good yield from the
reaction between a 2-halogenohexane This does not mean that it is impossible to generate the pure cis isomer,
and base shown in Fig. 14.23, other just that it is not possible using this particular reaction. In fact, cis-2-hexene
routes do give a good yield. For example,
a simple way to prepare predominantly
may be prepared in good yield by partial hydrogenation of 2-hexyne using an
this isomer is to partially hydrogenate appropriate metal catalyst as shown in Fig. 14.27. Under these conditions it is
2-hexyne using a metal catalyst. much easier to form the cis isomer than the trans form.
14.5 So why do chemical reactions happen? 233
14.5 So why do chemical reactions happen?

The simple answer to this question 1s that reactions happen because they result
in an increase in the entropy of the Universe — which is also the reason why
anything happens. We should really ask:
Why does this particular chemical reaction increase the entropy of the
Universe?
For an endothermic reaction, it must be because the entropy change of the
reaction is positive. However, for the majority of reactions the answer will be
because the reaction is exothermic and so gives out heat to the surroundings
thus increasing their entropy.
Why is a reaction exothermic?

This is because the particular arrangement of atoms in the products is in some
sense lower in energy than in the reactants. We must remember to consider all
of the reactants and all of the products. It may well be that the deciding factor
is something as subtle as the way in which the species interact with the solvent.
Why is one arrangement of atoms lower in energy than another?

When there is a particularly favourable interaction between two species (atoms
or molecules) we say there is some sort of bond. These bonding interactions
may be electrostatic or covalent in origin. We need to understand the factors
which favour strong ionic interactions and those which favour strong covalent
interactions. In most reactions, both are involved to some degree.
How can we make the product we want?

If the reaction is favourable, all that is needed is to supply enough energy to get
over the barrier. If it is unfavourable (1.e. the products are higher in energy than
the reactants) we will need to couple it to another reaction which ts strongly
favoured, or simply not allow the reaction to come to equilibrium. We also
need to think about reactions which will compete with the one we want and
how these can be avoided or minimized.
...and finally
Chemists are always trying to make new molecules, or devise better ways of
making them ~ they do this partly out of curiosity and partly because new
compounds are needed in every aspect of our life, from pharmaceuticals to
novel materials such as ceramics and semiconductors. To be successful, the
chemist needs to understand why and how reactions occur.
Index
x bond Markovnikov’s Rule for addition to C=C, 184

in carbony] group, 92 of HBr to alkene, 184
in ethene, 90 to C=C, 183
ma molecular orbital AgCl
from p-p overlap, 69 precipitation of, 200
o conjugation, 179 alkylation of enolate anion, 223
effect on stability of enolate anions, 226 effect of changing alkylating agent, 224
in carbonyl compounds, 218 effect of changing solvent, 225
stabilization of carbenium ion, 180 ally] anion
o framework, 90, 152 x molecular orbitals, 157
in carbonyl! compounds, 91 bond lengths in, 158
o molecular orbital formation from propene, 158
from p-p overlap, 68 resonance structures, 159
from s-s overlap, 65 stability compared to enolate anion, 170
trom s-p overlap, 68 allyl cation, 177
a/m symmetry labels, 65 x molecular orbitals, 178
2c-2e bonds, see two-centre two-electron bonds resonance structures, 178
ambident nucleophile, 222
A factor, see pre-exponential factor amide
absolute entropy, 11 z molecular orbitals, 161
acetal basicity, 164
interconversion with ketone, 127 C-N bond length in, 162
acetylene, see ethyne HOMO, 164
acid strength, 202 LUMO, 163
of chloroethanoic acids, 205 molecular orbital diagram for 7 system, 161
of HA in water, analysed using a cycle, 203 planarity of, 160
of HCI in water, 203 reactivity compared to aldehyde, 163
of HF compared with HCI, 204 resonance structures, 162
activation energy, 137 ammonia
apparent. for complex reactions, 148, 149 description of bonding using sp? HAOs, 85
determination of, 138 inversion of, 87
interpretation of, 137 angular node, 47
related to energy profile, 137 general pattern, 52
acyl chloride in 2p, 47
attack on by nucleophiles, 110 in 3p and 3d, 51
energy profile for nucleophilic attack on, 141 angular part of atomic orbital, 47
acylium ion, 176 anti-bonding molecular orbital, 61
shortening of C=O bond in, 177 destructive overlap in, 62
addition energy of, 62
Index 235
indicated by *, 73 bond lengths

nodal plane in, 63 C-C in carbenium ions, 181
Arrhenius equation, 137 in acylium ion, 177
prediction of variation of rate constant with tem- in butadiene and propene, 155
perature and activation energy, 138 in carboxylate anion, 160
atomic energy levels, 35 in propene and the allyl anion, 158
atomic orbital, 36 of C-N in amides, 162
1s, 40, 46 of C=O in carbonyl compounds, 168, 177
2s and 2p, 46 bond order, 74
3s, 3p and 3d, 49 correlation with bond strength, 74
angular nodes in, 47 bonding molecular orbital, 61
degeneracy in hydrogen, 48 constructive overlap in, 62
energies in hydrogen, 45 energy of, 62
names of d orbitals, 52 Born equation, 194
nodal planes in, 47 Born interpretation of wavefunction, 39
pattern of nodes in, 52 Born—Landé equation, 25
quantum numbers for, 44 butadiene
radial and angular parts, 47 xz molecular orbitals for, 153
radial nodes in, 46 bond lengths in, 155
representation as contour plot, 41
representation as density plot, 41 Ca(NO3)2
representation as surface plot, 42 solubility, 198
size of, 50 CaCO3
atomic spectra, 35, 36 precipitation of, 201
ATP carbenium ion
as high energy molecule, 129 competition between attack by OH™ as base or nu-
hydrolysis of, coupled to other reactions, 129 cleophile, 182
effect of solvent polarity on rate of formation, 189
BoH6 formed by protonation of unsymmetrical alkene,
structure of, 83 184
base formed in Sn 1 reaction, 172
Brgnsted-Lowry and Lewis definitions, 183 in E1 elimination, 181
nucleophilic behaviour, 182 shortening of C-C bond lengths in, 181
BaSO4 stabilization by adjacent z systems, 177
precipitation of, 200 stabilization by adjacent lone pairs, 175
benzene stabilization by polar solvent, 207
occupied z molecular orbitals, 152 stabilized by o conjugation, 180
resonance structures, 151 carboxylate anion
BH3 z orbitals for, 159
description of bonding using sp” hybrids, 86 bond lengths in, 160
LUMO, 99 formation of, 159, 169
molecular orbital diagram, 86 resonance structures, 159
reaction with LiH, 98 stabilization of negative charge, 169
BH, centre of inversion, 66
description of bonding using sp” hybrids, 88 CH3Br
reduction of a ketone, 105 LUMO, 108
Bohr model, 40 CH3Cl
Bohr radius, 44 LUMO, 108
236 Index
CHF dissociation energy

description of bonding in, 175 of diatomic, 61
CH4 dissociation reactions
molecular orbital diagram, 84 effect on temperature on, 15
molecular orbitals, 80
two-centre two-electron bonds, 81 E! mechanism
chloroethanoic acids for elimination, 181
acid strengths compared, 205 rate law for,,] 88
CIO E2 mechanism
intermediate in Cl + 03, 139 competition with El mechanism, 188
Co(H]I) and Co(II) halides, stability of, 29 for elimination, 185
complex reaction, 139 geometry for, 186
apparent activation energy in, 148 rate law for, 187
conjugate acid effective nuclear charge, 54
of leaving group, 215 effect on orbital energies, 54
conjugation, 155 effect on orbital size, 55
stabilization of negative charge by, 168 in helium, 54
stabilization of positive charge by, 177 in lithium, 55
conjugative effect, 163 screening and penetration, 56
effect on bond lengths, 168 electron-electron repulsion, 53
reduced in acyl chlorides, 165 electronegativity, 58
Coulomb interaction, 21 related to orbital energies, 59
coupling reactions together, 128 electronic configuration
to form amide bond, 129 of atoms, 36
to form ATP, 129 of homonuclear diatomics of the first row, 73
curly arrows, 98 electrostatic interaction, 21
connection to orbitals, 101 effect of changing relative permittivity, 190
for abstraction of proton from carbenium ion, 182 elementary reaction, 139
for formation of enolate, 219 rate law for, 140
for Ht + H~, 98 elimination, 180
for nucleophilic addition to C=C, 107 competition with substitution, 181, 189, 228
for nucleophilic addition to carbonyl, 104 El mechanism, 181
for reaction between LiH and BHs3, 101 E2 mechanism, 185
for reaction between OH™ and an acyl chloride, geometry for E2 mechanism, 186
112 endothermic reaction
for reaction of methanal with hydride, 104 spontaneous, 14
energies
degeneracy, 48 approximate order for molecular orbitals, 96
lifted in multi-electron atoms, 55 of atomic orbitals for the first row elements, 57
of hydrogen orbitals, 48 of hydrogen orbitals, 45
diatomic molecules of molecular orbitals in homonuclear diatomics,
energy curve for, 60 75
heteronuclear, molecular orbital diagrams for, 76 energy barrier, 132
homonuclear, molecular orbital diagrams for, 71 in inversion of ammonia, 134
dielectric constant, 190 mechanical analogy, 133
dipole moment reactions with no barrier, 135
effect on relative permittivity, 191 why present, 133
origin of, 77 energy curve of diatomic, 60
Index 237
energy minimization, 81 equilibrium

energy profile approach to, 113
effect of change in solvent polarity on, 207 between two species, 114
for attack of Cl~ on a ketone, 212 connection to rate of reaction, 130
for attack of OH™ on acyl chloride, 211 locating the position of, 115
for competing reactions, 230 position of, 5
for formation of enolates from unsymmetrical ke- equilibrium bond length, 61
tones, 227 equilibrium constant, 121
for nucleophilic attack on acyl chloride, 142, 149 effect of temperature on, 123, 144
for OH” + ethanal, 221 related to standard Gibbs energy change, A,;G°,
for oxidation of methanoic acid by Bro, 149 118
for pre-equilibrium, 145 relation to rate constants, 144
for Sn1 reaction, 173, 207 equilibrium geometry, 81
for Sn2 reaction, 134, 207 esters
enolate anion description of bonding using hybrid atomic orbitals,
alkylation of, 223 92
bonding to counter ion, 222 ethane
evidence for, 219 description of bonding using hybrid atomic orbitals,
formation from unsymmetrical ketone, 225 89
formation of, 170, 217 ethene
resonance structures, 170 description of bonding using hybrid atomic orbitals,
enthalpy 90
of precipitation, 200 ethyne
of solution, 198 description of bonding using hybrid atomic orbitals,
enthalpy change 90
relation to heat absorbed, 16 extraction of metals, 124
enthalpy of solvation of ions
correlated with z7/r, 195 first-order rate law, 137
effect on acid strength of HCI, 203 forcing a reaction to the left
entropy by adding an excess of a product, 127
absolute, 11 forcing a reaction to the right
of precipitation, 200 by removing products, 126, 128
of solution, 198 by using excess of the reagents, 126
relation to disorder or randomness, 7 formaldehyde, see methanal
entropy change
as a result of heat flow, 9 g/u symmetry labels, 66
effect of temperature, 10 geometry of attack
of surroundings, 8, 11 in E2 elimination, 186
of surroundings in exo- and endothermic processes, in nucleophilic addition to carbonyl, 105
9 in nucleophilic substitution, 109
of system, 8 Gibbs energy, 16
of Universe, 8 as a function of composition for a reaction, 113
on making ice, 11 molar, 114
entropy of solvation of ions of precipitation, 200
correlated with z2 /r, 196 of solution, 197
effect on acid strength of chloroethanoic acids, 206 of solvation of ions, predicted by Born equation,
effect on acid strength of HC], 203 194
positive value for some ions, 196 Gibbs energy change
238 Index
defined, 16 sp’, 83
in spontaneous process, 17 sp”, 85
of making ice, 17 energies of, 85
related to entropy change of the Universe, 16 formation of, 83
related to equilibrium constant, 117 occupation by lone pairs, 85
standard, A,G° reconciled with delocalized molecular orbitals, 93
defined, 118 used in the description of bonding in
related to equilibrium constant, 118 carbonyl epmpounds, 91
glucose diatomics, 88
production by photosynthesis, 129 ethane, 89
ethene, 90
H> and H> ethyne, 90
molecular orbital diagram, 64 No, 88
H20 nitriles, 93
description of bonding using sp’ hybrids, 85 organic molecules, 89
molecular orbitals for, 94 hydration number, 197
two-centre two-electron bonds in, 81 hydration shells, 196
Haber process, 127 hydrogen bond, 191
HCl between water molecules, 191
acid strength in water, 203 disrupted in water by introduction of ion, 196
He? and He; in DNA, 192
molecular orbital diagram, 65 in protic solvent, 192
HF strength of, 192
acid strength in water, 204 hydrolysis of ATP
molecular orbital diagram, 78 Gibbs energy change of, 129
molecular orbitals, 79
highest occupied molecular orbital, see HOMO ice
HMPA entropy changes on forming, 9, 11
as solvent, 225 Gibbs energy change on forming, 17
HOMO, 95 imine
identifying, 95 interconversion with ketone, 127
LiH, 99 inductive effect, 165
OH, 108 effect on bond lengths, 168
HOMO-LUMO interaction, 100 influencing the position of equilibrium, 122
choice of HOMO and LUMO, 101 by changing concentration, 126
importance in geometry of nucleophilic addition by changing temperature, 122
to carbonyl, 105 intermediate
importance of, 101 carbenium ion in El elimination, 181
LiH and BH3, 100 carbenium ion in Sn reaction, 172
nucleophilic addition to amides and aldehydes, 163 ClO in decomposition of ozone by Cl, 139
nucleophilic addition to C=C, 106 contrasted to transition state, 141
nucleophilic addition to carbonyl, 102 O in decomposition of ozone, 140
nucleophilic substitution, 107 tetrahedral in nucleophilic attack on acyl chloride,
OH™ + CH3Br, 108 110, 141
OH™ reacting with an acyl chloride, 110 inversion at carbon
hybrid atomic orbitals, 83 during nucleophilic substitution, 109
sp. 86 inversion of ammonia
s and p character. 86 energy profile for, 134
Index 239
orbital description, 87 magnetic quantum number, 45

ion pairs, 22 Markovnikov’s rule, 184
compared to lattice, 25 methanal
energy change on forming, 23 description of bonding using hybrid atomic orbitals,
ion—ion repulsion, 22 91
ionic radius, 20, 27 LUMO, 103
table of values, 27 methane, see CH4
trends in, 27 methanoyl chloride
ionic solids, 20 x orbitals for, 166
covalency in, 32 MgSO4
evidence for, 20 solubility, 198
molecular collisions, 131
Ka orientation of, 132
defined, 205
molecular orbital
Kapustinskii equation, 28
allyl anion, z system, 157
lattice allyl cation, 7 system, 178
caution over interpretation of lines, 83 anti-bonding molecular orbital, 61
electron density in, 33 approximate order of energies, 96
benzene, occupied z orbitals, 152
NaCl, 20
bonding molecular orbital, 61
lattice energy, 26
calculated using Born—Landé equation, 25 butadiene, z orbitals for, 153
calculated using Kapustinskii equation, 28 by linear combination of atomic orbitals, 61
lattice enthalpy, 26 carboxylate ion, z orbitals for, 159
delocalized for CH4, 80
Le Chatelier’s Principle, 123
leaving group, 212 energies of bonding and anti-bonding orbitals, 62
protonation to improve ability, 214 energy match and overlap, 67
leaving group ability, 213 ethanal, 7 orbitals for, 218
connection to pK, of conjugate acid, 215 for H20, 94
LiBH4 for heteronuclear diatomics, 76
from LiH + BH3, 98 for homonuclear diatomics, 71
LiH for LiH, BH3 and LiBHg, 99
HOMO, 99 four p orbitals in a row, 153
molecular orbital diagram, 77 from p-p overlap, 68, 69
reaction with BH3, 98 HOMO of phenolate anion, 171
linear combination of atomic orbitals, 61 identifying HOMO and LUMO, 95
lowest unoccupied molecular orbital, see LUMO methanamide, z orbitals in, 161
LUMO, 95 methanoyl chloride, z orbitals for, 166
aldehyde, 163 of H2, 61
amide, 163 propenal, z orbitals for, 156

BH3, 99 reconciled with localized hybrid atomic orbitals,
CH3Br, 108 93
CH3Cl, 108 rules for forming, 66
CHj, 175 size and overlap, 67
ethanoy! chloride, 110 symmetry and overlap, 67
identifying, 95 symmetry labels, 65
methanal, 103 three p orbitals in a row, 157
unsymmetrical due to energy mismatch, 76
Madelung constant, 24 molecular orbital diagram, 62
240 Index
allyl cation 7 system, 178 HOMO, 108

amide, z system, 161 orbital angular momentum quantum number, 45
BH3, 86 representation by letters, 45
CHa, 84 orbital approximation, 53
effect of s-p mixing, 73 effective nuclear charge, 54
for stabilization of positive charge by adjacent lone orbital energies
pair, 176 of atoms in the first row, 57
H2, 64 related to electronegativity, 59
HF, 78 separation of 2s and 2p, 58
homonuclear diatomics, 68, 71 order of reaction, 136
LiH, 78 organometallic compounds
propenal z system, 156 as source of R7, 213
stabilization of positive charge by o conjugation, ozone
179 decomposition of, 140
molecular orbitals destruction by Cl, 139
enolate anion from ethanal, z orbitals for, 222
multi-electron atoms, 53 pKa
orbital energies, 55 defined, 205
for chloroethanoic acids, 206
N2 paramagnetism
description of bonding using hybrid atomic orbitals,
of Bz, 74
88
of O2, 75
molecular orbital description of bonding in, 74
penetration, 56
NaNO;
compared for 3s, 3p and 3d, 57
solubility, 199
greater in 2s than 2p, 56
NH3, see ammonia
peroxides of Group I metals, stability of, 31
NH3 + HCl, 5, 8, 10
phenolate anion
NH{
resonance structures, 171
description of bonding using sp? hybrids, 88
stabilization of negative charge, 171
nitriles
photoelectron spectroscopy, 37
description of bonding using hybrid atomic orbitals,
of atoms, 37
93
position of equilibrium
NO: (dimerization of), 5, 6, 113
influencing, 122
nodal plane, 47
locating, 113
general pattern, 52
in 2p, 47 pre-equilibrium, 145
in 3p and 3d, 51 analysis of oxidation of methanoic acid by Br,
nucleophile 146
base acting as, 182
applicability of, 147
nucleophilic addition energy profile for, 145
to C=C, 106 pre-exponential factor, 137
to carbonyl], 102 determination of, 138
nucleophilic attack on acyl chlorides, 110 interpretation of, 137
nucleophilic substitution, 107 precipitation
attack hindered in f-butyl bromide, 174 of AgCl, 14, 200
choice of mechanism between Sn 1 and Sn2, 174 of BaSOq4, 200
competition with elimination, 228 of CaCO3, 201
principal quantum number, 45
OH dependence of orbital energy on in hydrogen, 45
Index 241
probability density, 40 effect of energy barrier on, 132

propenal effect of temperature on, 137, 231
xz molecular orbitals for, 156 factors which affect, 131
molecular orbital diagram for m7 system, 156 graphical interpretation of, 136
resonance structures, 157 rate law, 135
propene rate-determining step, 141
bond lengths in, 155 in nucleophilic attack on acyl chlorides, 142
reaction
quantization, 35 2-bromohexane with methoxide, 229
evidence for, 36 acyl chloride with water, 165
quantum mechanics, 39 addition of HBr to alkene, 184
quantum numbers, 44 addition to C=C, 183
for atomic orbitals, 44 aldehyde with water to give hydrate, 163
magnetic quantum number, 45 amides with aqueous alkali, 167
orbital angular momentum quantum number, 45 carbenium ion + base, 182
principal quantum number, 45 carbenium ion + H2O, 182
CH3Br + OH, 108, 131, 135
radial distribution function, 42 complex, 139
definition of, 43 dimerization of NO2, 5, 6
interpretation of, 43 elementary, 139
maximum in for ls, 44 elimination to form C=C, 180
of ls, 43 enolate anion with alkylating agent, 223
of 2s and 2p, 48 formation of an acetal, 127
of 3s, 3p and 3d, 50 formation of an amide bond, 128
value at r = 0, 43 formation of an ester, 112, 126
radial node, 46 formation of an imine, 127
general pattern of, 52 H + Bro, 139
in 2s, 46 Ht +H, 97
in 3s and 3p, 50 H> and Bro, rate law for, 140
radial part of atomic orbital, 47 H2 and hy, rate law for, 141
rate constant, 136 LiH + BH; to form LiBHa, 98
effect of temperature on, 137 NH; + HCI, 5, 8, 10
rate law, 135 nucleophilic addition to C=C, 106
first and second order, 137 nucleophilic addition to carbonyl, 102, 216
for El reaction, 188 nucleophilic attack on acyl chloride, 110, 141, 148
for E2 reaction, 187 nucleophilic substitution, 107, 172
for elementary reaction, 140 oxidation of methanoic acid by Br2, 146
for reaction between t-butyl bromide and a nucle- reduction of a ketone with BH, , 105
ophile, 150, 172 synthesis of ammonia using the Haber process, 127
for reaction between CH3Br and a nucleophile, 150, t-butyl bromide + nucleophile, 172
172 reaction coordinate, 134
for reaction between CH3Br and OH~, 137 reactivity of carbonyl compounds
for Sn1 reaction, 173 amides compared with acyl chlorides, 165
for Sn2 reaction, 173 amides compared with aldehydes, 163
order of, 136 amides compared with esters, 166
rate of reaction conjugative effect, 163
connection to equilibrium, 130 inductive effect, 165
effect of concentration on, 131 relative permittivity, 190
242 Index
resonance structures effect on rate of reaction, 207

allyl anion, 159 effect on rate of Sn2 reaction, 189
allyl cation, 178 polar and non-polar (apolar), 190
amides, 162 polarity of, 190
benzene, 151 protic and aprotic, 192
carboxylate anion, 159 table of relative permittivities and dipole moments,
phenolate anion, 171 193
propenal, 157 sp hybrid atomic orbitals, 86
reversible reactions, 143 sp* hybrid atomic orbitals, 85
relation between activation energies of forward and sp® hybrid atomic orbitals, 83
reverse reactions, 143 used in the description of bonding in
rotation about C-C o and z bonds, 90 NH3 and H20, 85
rows of p orbitals NHj and BH;, 88
four, 153 sparingly soluble salts, 199
general form of molecular orbitals, 153 spherical polar coordinates, 40
three, 157 spontaneous process, 8
Rydberg constant, 45 endothermic reaction, 14
exothermic reaction, 13
Sn1, 172 stabilization of high oxidation states, 28
effect of solvent polarity on rate, 207 stabilization of negative charge
rate law for, 173 by conjugation, 168
reaction of t-butyl bromide, 172 stabilization of positive charge
Sn2 by o conjugation, 179
effect of solvent polarity on rate, 189, 207 by adjacent m systems, 177
rate law for, 173 by adjacent lone pairs, 175
S-p mixing standard enthalpy change, A,H°
effect on molecular orbital diagram, 73 defined, 119
screening, 56 determination of, 119
Second Law of Thermodynamics, 8, 15 standard entropy change, A,S°
second-order rate law, 137 defined, 119
solubility, 190 determination of, 120
salts in non-aqueous solvents, 201 standard Gibbs energy change, A,G°
salts in water, 197 defined, 118
soluble salts, 198 determination of, 120
sparingly soluble salts, 199 in terms of A-H° and A,S°, 118
specified by solubility product, 197 related to equilibrium constant, 118
thermodynamic cycle for, 198 significance of sign, 121
solubility product, 197 significance of size, 122
solvation of ions, 194 standard states, 118
anions in protic solvents, 194 steric requirement, 132
enthalpy of, 195 superoxides of Group I metals, stability of, 31
entropy of, 196 symmetry
Gibbs energy change, predicted by Born equation, effect on formation of molecular orbitals, 67
194 symmetry labels, 65
hydration number, 197
hydration shells, 196 temperature
solvent effect on entropy change, 10
effect on rate of formation of carbenium ion, 189 effect on equilibrium constant, 123
Index 243
effect on rate constant, 137

tetrahedral intermediate, 110
collapse of, 111, 212, 216
thermodynamic and kinetic products, 227
tosylate
as leaving group, 214
transition state, 141
for inversion of ammonia, 134
for nucleophilic substitution reactions, effect of sol-
vent polarity on, 207
for reaction between CH3Br and OH, 109, 135
two-centre two-electron bonds
dot and cross picture of, 81
molecular orbital picture of, 82
representation by lines, 82
unhybridized p orbital, 85
valence electrons, 73
water
hydrogen bonding in, 191
solubility of salts in, 197
solvation shells round ions, 196
sparingly soluble salts in, 199
water (other than as a solvent), see H2O
wavefunction, 39
interpretation of, 39
relation to probability density, 40
zeolite clay
for removing water, 127
By tackling the most central ideas in chemistry, Why Chemical Reactions Happen provides
the reader with all the tools and concepts needed to think like a chemist. The text takes
a unified approach to the subject, aiming to help the reader develop a real overview
of chemical processes, by avoiding the traditional divisions of physical, inorganic and
organic chemistry.
Why Chemical Reactions Happen is written for students who are in their first year of a
university chemistry course, or are about to embark on such a course. To understand
how chemical reactions happen we need to know about the bonding in molecules,
how molecules interact, what determines whether an interaction is favourable or not,
and what the outcome will be. Answering these questions requires an understanding
of topics from quantum mechanics, through thermodynamics, to ‘curly arrows’. In this
book all of these topics are presented in a coherent and co-ordinated fashion, showing
how each leads to a deeper understanding of chemical reactions.
Numerous illustrations support the explanations given in the text

Mathematical content is kept to a minimum, making the book suitable for
readers from all backgrounds
James Keeler is a Lecturer and Director of Teaching in

the Department of Chemistry, University of Cambridge,
and Fellow of Selwyn College.
Peter Wothers is a Teaching Fellow in the Department
of Chemistry, University of Cambridge, and Fellow
of St Catharine's College.
Companion web site at:

www.oup.com/uk/best.textbooks/chemistry/keeler/
including illustrations available to download and several appendices
giving more details on selected topics.
Wisi
OXF ORD
UNIVERSITY PRESS
www.oup.com
9
7 80199
99"249 732

Dokumen - Pub Why Chemical Reactions Happen 9780199249732 0199249733

Uploaded by

Copyright:

Available Formats

Dokumen - Pub Why Chemical Reactions Happen 9780199249732 0199249733

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dokumen - Pub Why Chemical Reactions Happen 9780199249732 0199249733

Uploaded by

Copyright:

Available Formats

why chemical

We are grateful to Michael Rodgers, then commissioning editor for science

Cambridge, February 2003.

AND OUR TEACHERS

1 What this book is about and who should read it

What makes a reaction go?

5 Electrons in simple molecules

9 Rates of reaction 131

10 Bonding in extended systems — conjugation 151

11 Substitution and elimination reactions 172

13 Leaving groups 210

14 Competing reactions 216

When we first come across chemical reactions they are presented to us as

e Acyl chlorides react readily with water to give carboxylic acids:

whereas amides are completely unreactive to water;

e In solution 2-bromo-2-methylpropane can form the positively charged

H3C. CH3 CHs

e Areason often given as to why a precipitate of AgCl is formed from Agt

e The result of the reaction of propene with HBr may be predicted by

How this book is laid out

How to read this book

Some practical matters

2c-2e two-centre, two-electron

NH3(g) + HCl(g) —> NHsCl(s).

Mg(s) + 2O2(g) —> MgO(s).

CH3COOH(aq) + H20(J) = CH3COO™ (ag) + H307* (aq).

Another example is the dimerization of NO2 to give N2Oa:

2.1 Energy changes

2NO2(g) —> N2Oa(g).

N204(g) —> 2NO2(g).

2NOz > N20, _: oe N,O, —> f 2NO3

low pressure 1 atmosphere high pressure

2.2 Entropy and the Second Law of Thermodynamics

NH3(g) + HCl(g) —> NH4Cl(s)

2.3 Entropy and heat

In an exothermic process the system gives out heat to the surroundings

In an endothermic process the system takes in heat from the surroundings

IEE entropy of system decreases

NH3(g) + HCl(g) —> NH4Cl(s)

2.4 Role of temperature

smailer increase in entropy {vo

heat absorbed by the surroundings

How to make ice

ASuniv = ASsur + ASsys

ASuniv = ASsur + ASsys

Why all this fuss about making ice?

2.5 Exothermic and endothermic reactions

NH3(g) + HCl(g) —> NH4Cl(s)

in which there is a reduction in entropy of the system. However, as the reaction

A contrast to ammonium nitrate dissolving in water is when solid sodium

As was discussed in Section 2.4 on p. 10, temperature plays an important

N204(g) —> 2NO2(g).

2.6 How do we know all this is true?

2.7 The Second Law and Gibbs energy

ASuniv = A Ssurr + ASsys. (2.2)

ASsurr = dure > (2.3)

Ysurr = —sys:- (2.4)

A Suniv = —4sys + ASsys (2.5)

AGgys = A Hgys _~ TsysA Ssys (2.7)