Molecular Physical Chemistry: José J.C. Teixeira-Dias

José J.C.
Teixeira-Dias
Molecular
Physical
Chemistry
A Computer-based Approach using
Mathematica® and Gaussian
Molecular Physical Chemistry
José J.C. Teixeira-Dias
Molecular Physical
Chemistry
A Computer-based Approach using
Mathematica® and Gaussian
123
José J.C. Teixeira-Dias
University of Aveiro
Aveiro, Baixo Vouga
Portugal
ISBN 978-3-319-41092-0 ISBN 978-3-319-41093-7 (eBook)

DOI 10.1007/978-3-319-41093-7
Library of Congress Control Number: 2016947376
© Springer International Publishing Switzerland 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
This book contains chapters on thermodynamics, chemical kinetics, quantum

chemistry, molecular symmetry, molecular structure, crystals, and water, and is
intended for second-year master’s students in chemistry. It presents the subject
through real examples, discussing the results of molecular orbital calculations
performed by Gaussian on small molecules, exploring and running Mathematica
codes presented at the end of each chapter that enable the student to plot functions,
normalize functions, fit data, solve equations, and test physical models; they are
accompanied by detailed explanations that provide insight and a suitable environ-
ment for active learning. Each chapter contains a glossary of important scientific
and technical terms, and the book includes detailed and complete answers to all
exercises. Since the molecular orbital calculations presented are standard, packages
other than Gaussian can alternatively be used to provide the necessary data. Stu-
dents who are unfamiliar with Mathematica should watch the set of short videos
provided by this software to learn to write and run small programs and follow the
explanations to the selected codes at the end of each chapter. Those who are
familiar with other computational tools can alternatively use them.
Aveiro, Portugal José J.C. Teixeira-Dias
v
Contents
1 Thermodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Ideal Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Kinetic Model of Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Pressure and Temperature . . . . . . . . . . . . . . . . . . . . 4
1.2.2 Distribution of Velocities . . . . . . . . . . . . . . . . . . . . . 6
1.2.3 Mean Free Path . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Van der Waals Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4 Mathematical Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4.1 Exact Differential . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.4.2 Fundamental Theorem of Calculus . . . . . . . . . . . . . . 20
1.4.3 Line Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.5 Thermodynamic Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.6 Heat and Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6.1 Mechanical Work . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.7 First Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.7.1 Heat Capacities. . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.7.2 Calorimeter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.7.3 Standard States . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.8 Reversible Heat Engine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.8.1 Carnot’s Heat Engine . . . . . . . . . . . . . . . . . . . . . . . 33
1.8.2 Absolute Temperature . . . . . . . . . . . . . . . . . . . . . . . 33
1.9 Entropy and the Second Law . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.10 Irreversible Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.10.1 Heat Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.10.2 Gas Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
1.10.3 Diffusion of Matter . . . . . . . . . . . . . . . . . . . . . . . . . 40
1.10.4 Chemical Reaction . . . . . . . . . . . . . . . . . . . . . . . . . 41
1.11 Chemical Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.11.1 Gibbs–Duhem Equation . . . . . . . . . . . . . . . . . . . . . . 43
1.11.2 Ideal Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
1.11.3 Real Gases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
1.11.4 Liquid Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
1.11.5 Pure Liquids and Solids. . . . . . . . . . . . . . . . . . . . . . 47
vii
viii Contents
1.12 Gibbs Energy. . . . . . . . . . . . . . . . . . . . . . . . ............ 47

1.12.1 Chemical Potential and Gibbs Energy
of Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.12.2 Gibbs–Helmholtz Equation. . . . . . . . . . . . . . . . . . . . 50
1.13 Chemical Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
1.14 Gibbs Phase Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1.15 Helmholtz Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
1.16 Surface Tension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
1.16.1 Liquid Droplet in Air . . . . . . . . . . . . . . . . . . . . . . . 55
1.16.2 Capillary Action . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
1.17 Membrane Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
1.18 Electrochemical Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
1.18.1 Nernst Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Mathematica Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2 Chemical Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.1 Rate of a Chemical Reaction . . . . . . . . . . . . . . . . . . . . . . . . . 84
2.2 Experimental Rate Equation. . . . . . . . . . . . . . . . . . . . . . . . . . 86
2.2.1 First-Order Reactions . . . . . . . . . . . . . . . . . . . . . . . 87
2.2.2 Second-Order Reactions. . . . . . . . . . . . . . . . . . . . . . 88
2.2.3 Zeroth-Order Reactions . . . . . . . . . . . . . . . . . . . . . . 89
2.3 Effect of Temperature Change . . . . . . . . . . . . . . . . . . . . . . . . 89
2.4 Elementary Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
2.5 Complex Reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
2.6 Extremely Fast Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.6.1 Neutralization Reaction in Water . . . . . . . . . . . . . . . 96
2.7 Chemical Oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
2.7.1 Brusselator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3 The Schrödinger Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

3.1 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.1.1 Eigenvalues and Eigenfunctions . . . . . . . . . . . . . . . . 115
3.1.2 One-Dimensional Schrödinger Equation . . . . . . . . . . . 115
3.1.3 Hermitian Operators . . . . . . . . . . . . . . . . . . . . . . . . 117
Contents ix
3.1.4 Important Theorems . . . . . . . . . . . . . . . . . . . . . . . . 118

3.1.5 Dirac Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
3.2 Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
3.2.1 Reduced Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.2.2 Classical Treatment . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.2.3 Quantum-Mechanical Treatment . . . . . . . . . . . . . . . . 124
3.2.4 Morse Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.3 Spherical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
3.4 Angular Momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.4.1 Orbital Angular Momentum . . . . . . . . . . . . . . . . . . . 131
3.4.2 Spin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
3.5 Hydrogen Atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
3.6 Antisymmetry Principle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
3.7 Variational Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.8 Born–Oppenheimer Approximation. . . . . . . . . . . . . . . . . . . . . 143
3.9 Hartree–Fock Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.9.1 Slater-Type Orbitals . . . . . . . . . . . . . . . . . . . . . . . . 148
3.9.2 Hartree–Fock Equations. . . . . . . . . . . . . . . . . . . . . . 149
3.9.3 Hartree–Fock–Roothaan Equations . . . . . . . . . . . . . . 150
3.9.4 Correlation Energy . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.10 Density Functional Theory . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.10.1 Electron Probability Density. . . . . . . . . . . . . . . . . . . 152
3.10.2 External Potential . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3.10.3 Functional Derivative . . . . . . . . . . . . . . . . . . . . . . . 153
3.10.4 Hohenberg–Kohn Theorems . . . . . . . . . . . . . . . . . . . 156
3.10.5 Kohn–Sham Method . . . . . . . . . . . . . . . . . . . . . . . . 157
3.10.6 Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
3.11 Perturbation Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
3.11.1 Nondegenerate Energy Level . . . . . . . . . . . . . . . . . . 162
3.11.2 Variational Perturbation Method . . . . . . . . . . . . . . . . 164
3.11.3 Degenerate Energy Level . . . . . . . . . . . . . . . . . . . . . 166
3.12 Time-Dependent Perturbation Theory . . . . . . . . . . . . . . . . . . . 168
3.12.1 Time-Dependent Schrödinger Equation . . . . . . . . . . . 168
3.12.2 Time-Dependent Perturbation . . . . . . . . . . . . . . . . . . 169
3.13 Absorption and Emission of Radiation . . . . . . . . . . . . . . . . . . 172
3.13.1 Spontaneous Emission of Radiation. . . . . . . . . . . . . . 175
3.14 Raman Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
3.14.1 Classical Treatment . . . . . . . . . . . . . . . . . . . . . . . . . 177
3.14.2 Quantum-Mechanical Treatment . . . . . . . . . . . . . . . . 178
3.15 Molecular Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
3.15.1 Computational Methods . . . . . . . . . . . . . . . . . . . . . . 181
3.15.2 Gaussian-Type Functions . . . . . . . . . . . . . . . . . . . . . 182
3.15.3 Standard Basis Sets . . . . . . . . . . . . . . . . . . . . . . . . . 185
x Contents
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
4 Molecular Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

4.1 Symmetry Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
4.2 Point Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
4.3 Matrix Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
4.4 Character Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
4.5 Selection Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
4.6 Molecular Vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
5 Molecular Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

5.1 Electron Probability Density . . . . . . . . . . . . . . . . . . . . . . . . . 244
5.2 Electrostatic Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
5.3 Mulliken Population Analysis . . . . . . . . . . . . . . . . . . . . . . . . 249
5.3.1 Density Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
5.3.2 Minimal Basis Set Calculation for CH4 . . . . . . . . . . . 251
5.4 Natural Bond Orbitals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
5.4.1 Hybrid Atomic Orbitals . . . . . . . . . . . . . . . . . . . . . . 254
5.4.2 Natural Bond Orbitals for CH4 . . . . . . . . . . . . . . . . . 257
5.4.3 Natural Bond Orbitals for H2C=CH2 . . . . . . . . . . . . . 258
5.4.4 Natural Bond Orbitals for HCCH . . . . . . . . . . . . . . 259
5.4.5 CH Hybrids in CH4, H2C=CH2 and HCCH . . . . . . . 260
5.4.6 Molecular Geometries and Electrostatic Potentials . . . . 261
5.5 Potential Energy Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
5.5.1 Intrinsic Reaction Coordinate . . . . . . . . . . . . . . . . . . 264
5.6 Molecular Conformations . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
5.6.1 Ethane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
5.6.2 1,2-Dichloroethane . . . . . . . . . . . . . . . . . . . . . . . . . 269
5.6.3 Boltzmann Distribution . . . . . . . . . . . . . . . . . . . . . . 270
5.7 Chiral Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Contents xi
6 Crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.1 Packing Disks and Spheres . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.1.1 Disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.1.2 Spheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
6.1.3 Hexagonal Close Packing . . . . . . . . . . . . . . . . . . . . 291
6.1.4 Cubic Close Packing . . . . . . . . . . . . . . . . . . . . . . . . 292
6.1.5 Packing Densities . . . . . . . . . . . . . . . . . . . . . . . . . . 292
6.1.6 Occupying Interstices . . . . . . . . . . . . . . . . . . . . . . . 295
6.2 Translation Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
6.2.1 2D Bravais Lattices. . . . . . . . . . . . . . . . . . . . . . . . . 297
6.2.2 3D Bravais Lattices. . . . . . . . . . . . . . . . . . . . . . . . . 301
6.3 Crystal Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
6.3.1 Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
6.3.2 Lattice Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
6.3.3 Cesium Chloride and Sodium Chloride . . . . . . . . . . . 305
6.3.4 Diamond and Zinc Blende . . . . . . . . . . . . . . . . . . . . 306
6.4 X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
6.5 Electrons in Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
6.6 Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
7 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
7.1 Molecular Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
7.2 Enthalpy of Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
7.3 Atomic Charges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
7.4 Dipole Moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
7.4.1 Electric Multipoles . . . . . . . . . . . . . . . . . . . . . . . . . 334
7.4.2 Point Dipole. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
7.4.3 Electric Field Streamlines. . . . . . . . . . . . . . . . . . . . . 337
7.4.4 H2O Dipole and Quadrupole . . . . . . . . . . . . . . . . . . 338
7.5 Molecular Orbitals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
7.5.1 Natural Bond Orbitals . . . . . . . . . . . . . . . . . . . . . . . 341
7.6 Molecular Vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
7.7 Intermolecular Interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . 344
7.7.1 Electrostatic Interaction . . . . . . . . . . . . . . . . . . . . . . 344
7.7.2 Induction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
7.7.3 Dispersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
7.8 Hydrogen Bond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
7.8.1 The Water Dimer . . . . . . . . . . . . . . . . . . . . . . . . . . 351
xii Contents
7.9 Ice Ih . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

7.9.1 Gas Hydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
7.10 Liquid Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
7.11 Phase Diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
7.12 Water as Solvent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
7.12.1 Electric Permittivity . . . . . . . . . . . . . . . . . . . . . . . . 359
7.13 Simple Nonpolar Solutes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
7.13.1 Ostwald Coefficient. . . . . . . . . . . . . . . . . . . . . . . . . 362
7.13.2 Hydrophobic Interaction . . . . . . . . . . . . . . . . . . . . . 365
7.14 Ionic Solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
7.15 Amphipathic Molecules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
7.15.1 Sodium Decanoate Micelles . . . . . . . . . . . . . . . . . . . 373
7.16 Acids and Bases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
7.16.1 Autoionization of Water . . . . . . . . . . . . . . . . . . . . . 376
7.16.2 Acid Ionization Constant . . . . . . . . . . . . . . . . . . . . . 377
7.16.3 Lewis Acids and Bases . . . . . . . . . . . . . . . . . . . . . . 380
7.17 Standard Electrode Potentials . . . . . . . . . . . . . . . . . . . . . . . . . 381
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Further Reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Answers to Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
Credits
Gaussian (www.gaussian.com)
The computational quantum chemistry calculations whose results are presented in
this book were carried out with the Gaussian 09 series of programs.
Authors for Gaussian 09, Revision A.02: M.J. Frisch, G.W. Trucks, H.B.
Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B.
Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F.
Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J.A. Montgomery, Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd,
E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K.
Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega,
J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.
J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz,
J. Cioslowski, and D.J. Fox, Gaussian, Inc., Wallingford CT, 2009.
GaussView (www.gaussian.com)
GaussView is the graphical interface to Gaussian.
Authors for GaussView, Version 5.0.9: Roy Dennington, Todd Keith, and John
Millam, Semichem Inc., Shawnee Mission, KS, 2009.
Natural bond orbitals
A natural bond orbitals analysis of the electron population is included in Gaussian
09.
Reference: NBO Version 3.1, E.D. Glendening, A.E. Reed, J.E. Carpenter,
and F. Weinhold.
Mathematica (www.wolfram.com)
Author of Mathematica: Wolfram Research Inc. Champaign, IL, 2010.
Wolfram Alpha (www.wolframalpha.com)
This is a computational knowledge-based engine that provides answers to questions
raised through the Internet in free-form input of natural language (English).
xiii
xiv Credits
Wolfram MathWorld (mathworld.wolfram.com)

This Internet site, built with Mathematica technology, is considered to be the web’s
most extensive mathematical resource.
CCCBDB (cccbdb.nist.gov)
National Institute of Standards and Technology (NIST) Computational Chemistry
Comparison and Benchmark DataBase
NIST Standard Reference Database Number 101
Release 17b, September 2015, Editor: Russell D. Johson III
This database consists of experimental and computational thermochemical data
for a selected set of more than 1700 gas phase atoms and small molecules.
Character tables (symmetry.jacobs-university.de)
These character tables apply to chemically important point groups.
Handbook of Chemistry and Physics, 2011
On CD-ROM, Version 2011, W.M. Haynes, Editor-in-Chief, CRC Press.
Thermodynamics
1
Abstract
The first three sections of this chapter include the ideal gas equation, the kinetic
model of gases, and the Van der Waals equation. They aim at building a unified
background for thermodynamics. Having in mind those students who might need
some refreshment of the mathematical concepts much used in thermodynamics,
such as exact and inexact differentials, the fundamental theorem of calculus, and
line integrals, we introduce a short section on these mathematical tools, followed
by sections on thermodynamic systems, heat and work, and the first law. In order
to formally arrive at the state function entropy, we show that an arbitrary
reversible cycle can be covered to the required precision by a mesh of
infinitesimal Carnot cycles and arrive at the Clausius inequality for an arbitrary
irreversible cycle. Following Prigogine, a distinction is then established between
entropy transfer across system boundaries and entropy produced inside a system
due to irreversible processes that drive the system to equilibrium. These are
illustrated with heat flow, gas expansion, matter diffusion, and a chemical
reaction in the gas phase. Next, we derive the Gibbs–Duhem equation and obtain
the chemical potential for the ideal gas, real gases, liquid solutions, pure liquids,
and solids, before presenting sections on the Gibbs energy, chemical equilib-
rium, the Gibbs phase rule, and the Helmholtz energy. The last three sections of
this chapter, on surface tension, membrane potential, and the electrochemical
cell, are important applications of the first and second laws of thermodynamics.
At the end of the chapter we present several Mathematica codes (Maxwell
Distribution of Molecular Speeds, Critical Point for Van der Waals Fluid,
3D Plot of the Van der Waals Equation in Reduced Variables, Absolute and
Reduced Temperatures for the Van der Waals Carbon Dioxide, Isothermal and
Adiabatic Transformations on an Ideal Gas Surface, Efficiency of the Carnot
Heat Engine, Gibbs Energy, and Affinity of a Chemical Reaction) that are
interconnected with the main text and contain detailed explanations of important
© Springer International Publishing Switzerland 2017 1

J.J.C. Teixeira-Dias, Molecular Physical Chemistry,
DOI 10.1007/978-3-319-41093-7_1
2 1 Thermodynamics
commands. Some of these explanations present suggestions for the student to

follow. The student will also find at the end of this chapter a glossary of
important scientific and technical terms in thermodynamics and a list of
exercises, whose complete answers are at the end of the book.
1.1 Ideal Gas
In, 1660, Robert Boyle (1627–1691) showed that at a defined temperature T, the
volume V of a gas is inversely proportional to its pressure p, that is,
V ¼ f1 ðTÞ=p ð1:1Þ
where f1(T) is a function of temperature (see Note §1). Edmé Mariotte [1620(?)–
1684] independently confirmed this functional dependence between volume and
pressure at constant temperature. In the last quarter of the eighteenth century,
Jacques Charles (1746–1823) studied the dependence between volume and tem-
perature, having found that the volume is proportional to the temperature at constant
pressure,
V ¼ Tf2 ðpÞ ð1:2Þ
In 1811, Amedeo Avogadro (1776–1856) put forward a hypothesis according to

which equal volumes of any two gases at the same temperature and pressure contain
the same number of molecules. This hypothesis implies that at a defined temper-
ature and pressure, the volume of a gas is proportional to the number of gas
molecules N. Using the Avogadro constant (NA = 6.022 1023 mol−1, see
Appendix), the amount of gas n is given by
n ¼ N=NA ð1:3Þ
Combining (1.1), (1.2) and the Avogadro hypothesis (gas volume proportional to
the amount of gas), one concludes that f1(T) is given by
f1 ðTÞ ¼ constant nT ¼ nRT ð1:4Þ
where R is called the gas constant (R = 8.3144 J mol−1 K−1; see Appendix).
Substitution of (1.4) in (1.1) leads to
pV ¼ nRT ð1:5Þ
1.1 Ideal Gas 3
The gas constant is related to the Boltzmann constant (kB = 1.3806 10−23
J K−1; see Appendix) by
R ¼ kB N A ð1:6Þ
A gas whose pressure p, volume V, temperature T, and amount of substance

n obey (1.5) is called an ideal gas. Since these four variables are linked by (1.5),
three of them determine the fourth. For instance, n, V, and T determine p. A gas
with defined values of pressure, volume, and temperature is said to be in a ther-
modynamic state. For that reason, (1.5) is called the equation of state, and the
variables p, V, T, and n are called state variables. Substitution of (1.6) and (1.3) in
(1.5) leads to
p V ¼ N kB T ð1:7Þ
Dividing both members of (1.7) by V gives
p ¼ qN kB T ð1:8Þ
where qN is the number density, given by
qN ¼ N=V ð1:9Þ
Equations (1.5), (1.7), and (1.8) are equivalent forms of the ideal gas equation of
state.
1.2 Kinetic Model of Gases
A gas can be studied at the macroscopic and molecular levels. The macroscopic
study requires the measurement of variables such as pressure, volume, temperature,
and amount of gas, interrelated by an empirical, relationship involving these vari-
ables (the equation of state), as we have just shown for the ideal gas. In turn, at the
molecular level, the study of gases requires a theoretical model that allows for the
understanding of the dynamical behavior of molecules under the model assump-
tions and the evaluation of macroscopic intensive variables such as pressure and
temperature.
It was Daniel Bernoulli (1700–1782) who first concluded that the pressure of a
gas results from collisions of gas particles with the container walls. At that time, no
one knew about molecules or could predict the order of magnitude of the speed of
gas particles.
The kinetic model of gases explains macroscopic gas properties such as pressure
and temperature from the random motion of molecules and was mostly developed
during the nineteenth century, with main contributions by James Prescott Joule
4 1 Thermodynamics
Fig. 1.1 Molecules of a gas

are in constant random motion
(1818–1889), James Clerk Maxwell (1831–1879) and Ludwig Boltzmann (1844–

1906). This gas model is based on five assumptions. The first states that the average
distance between molecules in a gas is far greater than the dimensions of the
molecules. The second says that the molecules of a gas are in constant random motion
(Fig. 1.1). The third assumption sees each molecule as a sphere making elastic
collisions with other molecules or with the container walls. The fourth assumption
says that the molecules do not interact with each other except during collisions, and
the last assumption states that gas molecules obey Newton’s laws of motion.
At equilibrium, each gas molecule exists in an isotropic space; that is, all
directions emerging from the molecular center of mass are equivalent, and the
choice of directions for a reference system of axes is arbitrary. Therefore, equations
can be deduced for the x velocity component of a molecule and easily generalized
for the y and z components.
In the kinetic model of gases, molecules correspond to rigid spheres, that is,
spheres that cannot be deformed. Therefore, the kinetic energy is not transferred
into potential energy. Kinetic energy is the single form of energy existent in the
kinetic model of gases. Before and after a collision between molecules, the total
kinetic energy (sum of the kinetic energies of both molecules) is invariant. The
same happens with the collision of a molecule with a container wall, but in this
case, the kinetic energy of the container wall is zero, and so the colliding molecule
has the same kinetic energy before and after the collision. In the kinetic model of
gases, all the collisions are elastic.
1.2.1 Pressure and Temperature
Among the molecules that move toward the container wall at the right in Fig. 1.2,
those that collide with the wall within the time interval Dt cannot be at a distance
from the wall greater than vxDt, where vx > 0. The volume occupied by such
molecules is vxDtA, where A is the area of the wall, and the number of such
molecules is given by
vx DtAqN =2 ð1:10Þ
1.2 Kinetic Model of Gases 5
Fig. 1.2 In the time interval

Dt, molecule 1 collides with
the wall at the right
where qN is the number density given by (1.9). The factor of 1/2 in (1.10) accounts
for the fact that one-half of the molecules move toward each of the walls, right and
left. In each collision with the wall at the right, the x component of the velocity
changes sign. Hence, the variation of linear momentum on collision with the wall is
given by mvx − (−mvx) = 2mvx (Fig. 1.3), and the total variation of linear
momentum is 2mvx times the number of molecules (1.10), that is,
mv2x DtAqN ð1:11Þ
According to Newton’s equation, the force is equal to the time derivative of the
linear momentum, which is given approximately by the ratio of (1.11) to Dt,
mv2x AqN ð1:12Þ
since the time interval is very short. Therefore, the pressure (force per unit area) on
the right wall is given by (1.12) divided by A, that is,
Fig. 1.3 Molecule in elastic

collision with a wall
6 1 Thermodynamics
p ¼ mv2x qN ð1:13Þ
The velocity axes are orthogonal, and so we can write
v2 ¼ v2x þ v2y þ v2z ð1:14Þ
For an isotropic space,
v2 ¼ 3v2x ð1:15Þ
Substitution of (1.15) in the second member of (1.13) leads to
p ¼ mv2 qN =3 ð1:16Þ
However, not all molecular velocities have the same absolute value (the same
speed), and consequently the same square speed. In the ensemble average of square
speeds, the average is taken over all N gas molecules in the container,
2 2
v ¼ v1 þ v22 þ þ v2N =N ð1:17Þ
and we can write

p ¼ m v2 qN =3 ð1:18Þ

3kB T=2 ¼ m v2 =2 ¼ hEcin i ð1:19Þ
This equation shows that temperature is proportional to the ensemble average of

the kinetic energy of gas molecules.
1.2.2 Distribution of Velocities
In velocity space, the Cartesian axes correspond to the velocity components vx, vy,
and vz. The probability of finding the velocity vector v = (vx, vy, vz) in the domain
(vx, vx + dvx), (vy, vy + dvy), (vz, vz + dvz) (Fig. 1.4) is given by
PðvÞdvx dvy dvz ð1:20Þ
where the function P(v) is the probability density for the distribution of molecular
velocities, first evaluated by James Clerk Maxwell for a gas in thermodynamic
equilibrium and given by
Fig. 1.4 Velocity space
h i
PðvÞ / exp m v2x þ v2y þ v2z =ð2kB T Þ ð1:21Þ
where the symbol / means that the proportionality constant of the above expression
has been omitted and the absolute value of the exponent of (1.21) is the molecular
kinetic energy m(v2x + v2y + v2z )/2 divided by kBT. The proportionality constant of
(1.21) is given by
Z1 Z1 Z1 h i 3=2
m
1= exp m v2x þ v2y þ v2z =ð2kB T Þ dvx dvy dvz ¼
2pkB T
1 1 1
ð1:22Þ
(E1). Therefore, the probability density is

3=2 h i
m
PðvÞ ¼ exp m v2x þ v2y þ v2z =ð2kB T Þ ð1:23Þ
2p kB T
and is said to be normalized, that is, the integral of the probability density over all
velocity space is equal to 1. The density P(v) is called the Maxwell velocity
distribution. Expression (1.23) can be easily factored as the product of the vx, vy,
and vz components of the velocity vector v, that is,
PðvÞ ¼ Pðvx ÞPðvy ÞPðvz Þ ð1:24Þ
where
1=2
m

Pðvx Þ ¼ exp mv2x =ð2kB T Þ ð1:25Þ
2p kB T
Since P(vx) depends on v2x , both vx and −vx give the same value to the function,
which is symmetric about zero, that is, its mean value is zero. Figure 1.5 shows
P(vx) for molecular oxygen, at 100 and 300 K.
8 1 Thermodynamics
Fig. 1.5 P(vx) at 100 and

300 K, for molecular oxygen.
Figure obtained with
Mathematica
For an isotropic space,
Pðvx Þ ¼ Pðvy Þ ¼ Pðvz Þ ð1:26Þ
and the geometric place for the tips of velocity vectors with the same absolute value
is the volume between two concentric spheres with radii v and v + dv, which is
given by the product of the area of the spherical surface 4pv2 and dv. Thus, using
(1.23) and (1.14), we can write
3=2
m

f ðvÞdv ¼ PðvÞ4pv2 dv ¼ 4p exp mv2 =ð2kB T Þ v2 dv ð1:27Þ
2pkB T
where f(v) is the distribution of molecular speeds [the word speed is used to
express the absolute value of the velocity vector, and the integration of (1.27) is now
between zero and infinity]. Figure 1.6 shows the distributions of molecular speeds
for oxygen, at 100 and 300 K (see Mathematica code M1). It can be seen that
these distributions are no longer symmetric, due to the factor v2, which strongly
weights larger values of molecular speeds, thus widening the higher wing of the
distributions.
Fig. 1.6 Distributions of

molecular speeds for oxygen,
at 100 and 300 K.
Mathematica
At a defined temperature, the speed that corresponds to the maximum of f(v) is

called the most probable speed, vp, and is given by the positive solution of the
equation

@f
¼0 ð1:28Þ
@v T
[f(v) has a single positive stationary point that is the curve maximum]. The positive
solution for (1.28) is
rffiffiffiffiffiffiffiffiffiffiffi
2kB T
vp ¼ ð1:29Þ
m
At 273.15 K, the most probable speed for O2 is 377 m s−1.

As mentioned before, f(v)dv is the probability for finding the speed v. Therefore,
f(v)dv is the weighting factor for the speed v in order to evaluate the mean speed vm
given by the following integral:
Z1 rffiffiffiffiffiffiffiffiffiffiffi
8kB T
vm ¼ f ðvÞvdv ¼ ð1:30Þ
pm
0
At 273.15 K, the mean molecular speed for O2 is 425 m s−1.

The root mean square speed (square root of the mean square speed, rms) is
given by
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uZ1 rffiffiffiffiffiffiffiffiffiffiffi
u
u 3kB T
vrms ¼ t f ðvÞv2 dv ¼ ð1:31Þ
m
0
For O2 at 273.15 K, the rms speed is equal to 461 m s−1 (E2).
1.2.3 Mean Free Path
Consider the mutual diffusion of two gases, oxygen and nitrogen, at a defined
pressure and temperature, after the wall separating them has been removed
(Fig. 1.7a). When the equilibrium is reached, the gas mixture has the same con-
centration everywhere in the system. In a container with a few liters of volume, the
equilibrium state may be reached in approximately 1 h. This result may look sur-
prising, since molecular speeds are on the order of hundreds of meters per second.
While each molecule has a large number of collisions, not all of them drive the
molecule away from the starting point, since some collisions drive the molecule
backward.
10 1 Thermodynamics
(b)
N2
(a)
before some time later ...
O2 N2
O2
Fig. 1.7 a Diffusion of two gases at the same pressure and temperature. b Test molecule of O2
diffusing through N2. Note that the colors used to distinguish the gases are fictitious: both O2 and
N2 do not absorb in the visible spectrum
Let us follow an oxygen molecule on its path through nitrogen gas (Fig. 1.7b).
To simplify matters, let us assume that the nitrogen molecules are fixed. Through
what distance does the oxygen molecule move before it first collides with a nitrogen
molecule? How long does it move before its second/third collision? The average
distance between collisions is called the mean free path.
In this thought experiment, the test molecule O2 and the molecules N2 are
represented by spheres of the same diameter d. The cross section for collision is the
area of a circle of radius d (pd2) (Fig. 1.8a). Between consecutive collisions, this
circle spans a cylinder whose height is equal to the distance traveled by the test
molecule (Fig. 1.8b). The volume of the broken cylinder is called the interaction
volume, and it contains all the nitrogen molecules that collide with the test molecule
within a certain time interval Dt.
(b)
(a)
Fig. 1.8 a Cross section for collision between spheres of equal radius. b All molecules colliding
with the test molecule have their centers of mass within the broken cylinder
A reasonable estimate k for the mean free path can be obtained by dividing the
distance d traveled by the test molecule by the number of collisions Nc, that is, the
number of N2 molecules within the interaction volume (= volume of the broken
cylinder),
k ¼ d=Nc ð1:32Þ
Nc is given by
Nc ¼ Vint qN ð1:33Þ
where qN is the number density of molecular nitrogen and the interaction volume
Vint is
Vint ¼ pd2 d ð1:34Þ
Substitution of (1.34) in (1.33) and then in (1.32) leads to
d 2 1
k¼ pd qN ð1:35Þ
pd2 dqN
This result is based on the less realistic assumption of the nitrogen molecules
remaining fixed while the test molecule travels among them. However, the distance
traveled by the test molecule [d in the numerator of (1.35)] and the length of the
interaction cylinder [d in the denominator of (1.35)] are distinct quantities, which
we name dn and dd, respectively. In fact, the distance traveled by the test molecule
dn is equal to the rms speed times the corresponding time interval,
dn ¼ vrms Dt ð1:36Þ
whereas dd, the real length of the interaction cylinder, is equal to the rms speed of
the test molecule relative to the speed of the colliding nitrogen molecule,
dd ¼ ðvrel Þrms Dt ð1:37Þ
It can be shown (see §2) that

pffiffiffi
ðvrel Þrms ¼ 2vrms ð1:38Þ
Hence, substitution of (1.36) in (1.35) and of (1.38) and (1.37) in (1.35) leads to
the following expression for the mean free path:
dn vrms Dt pffiffiffi 1

k¼ ¼ p ffiffi
ffi ¼ 2 pd 2
qN ð1:39Þ
pd 2 dd qN pd 2 2vrms DtqN
12 1 Thermodynamics
Considering the assumptions of the kinetic model of gases, this expression can
give only an order of magnitude for the mean free path. According to (1.39), the
mean free path is inversely proportional to the cross section for collisions and the gas
number density. In fact, increases of the cross section and of the gas number density
are expected to lead to decreases in the mean free path. Note that the mean free path
is independent of the distance traveled by the test particle [see (1.39)], since this
distance was chosen in an arbitrary way. Considering the approximate nature of
(1.39), one can deal with d as a parameter whose value can be chosen so that a
realistic value is obtained for the collision cross section. In particular, d = 0.3 nm
(= 3 Å) allows one to obtain a reasonable estimate for the mean free path.
The quantity qN can be obtained from the gas pressure and temperature using
(1.8), assuming ideal gas behavior. Alternatively, in order to relate qN to the
molecular diameter d, one can take the ratio of the liquid and gas densities and
equate this ratio to the inverse ratio of the molecular volumes,
ql m=Vl Vg
¼ ¼ ð1:40Þ
qg m=Vg Vl
since a molecule has the same mass m in the liquid and gas phases. For nitrogen, the
liquid density is 800 kg m−3 and the gas density is 1.2 kg m−3. Therefore, the ratio
of densities is approximately equal to 670 ( 800/1.2). In the liquid, the volume of
one molecule is approximately given by d3 (volume of a cube containing one
nitrogen molecule). Hence, applying (1.40) to nitrogen, we can write
Vg
670 ) Vg 670d3 ð1:41Þ
d3
If the definition of number density [see (1.9)] is applied to one molecule in the
gas and (1.41) is used, we obtain
1
qN ¼ Vg ð670d3 Þ1 ð1:42Þ

pffiffiffi 1 pffiffiffi
k¼ 2pd2 qN 670d3 = 2pd2 150d ð1:43Þ
Hence, the mean free path is about 150 times the collision cross-sectional radius
or the particle diameter (see Fig. 1.8a). The average distance between molecules in
the gas is given approximately by
1=3
670d 3 10d ð1:44Þ
that is, is about 15 times smaller than the mean free path.
The Maxwell velocity distribution (1.23) characterizes an equilibrium state. When

a gas is not at equilibrium, the distribution of velocities differs from (1.23), and the
temperature may not be well defined. In a gas, the evolution from an initial velocity
distribution to a Maxwell velocity distribution is quite fast, since it takes the time of a
few molecular collisions (in most cases, this time is shorter than 10−8 s) (E3).
1.3 Van der Waals Equation
An ideal gas cannot liquefy by a decrease of temperature or by compression, since the

ideal gas model ignores intermolecular interactions. Without these, gases would not
liquefy, liquids would not solidify, and liquids and solids could be easily compressed
to negligible volumes. In 1873, Van der Waals [1837–1923] derived an empirical,
equation based on a modification of the ideal gas equation that accounts for the volume
of molecules as impenetrable particles and for attractive interactions that contribute to
reducing the pressure of a gas made of particles of nonzero size. He received the Nobel
Prize in physics in 1910 for “his work on the equation of state for gases and liquids.”
Let us begin by considering the first modification of the ideal gas equation that
accounts for the nonzero volume of molecules. For 1 mole of gas,
Vm b ð1:45Þ
where Vm is the ideal gas molar volume and b accounts for the nonzero volume of
molecules per mole of gas, that is, the excluded molar volume. The second mod-
ification of the ideal gas equation introduced by Van der Waals accounts for the
pressure reduction due to attractive forces between particles. Attraction between
particles reduces their velocities, thus decreasing the number of collisions per time
unit on a unit area of wall. Considering the RT/(Vm − b) pressure being reduced by
dp, we can write
RT
p¼ dp ð1:46Þ
Vm b
Van der Waals took dp inversely proportional to the square of the volume,
a
dp ¼ ð1:47Þ
Vm2
Thus, the Van der Waals equation is given by

a
pþ ðVm bÞ ¼ RT ð1:48Þ
Vm2
where a and b are van der Waals constants (see Table).

14 1 Thermodynamics
Van der Waals constants a and b for selected gases
a (bar L2 mol−2) b (L mol−1)

Helium 0.0346 0.0238
Neon 0.208 0.0167
Argon 1.355 0.0320
Krypton 2.325 0.0396
Xenon 4.192 0.0516
Carbon dioxide 3.658 0.0429
Carbon disulphide 11.25 0.0726
Ammonia 4.225 0.0371
Water 5.537 0.0305
Methane 2.303 0.0431
Ethane 5.580 0.0651
Hexane 24.84 0.1744
Decane 52.74 0.3043
Values taken from the Handbook of Chemistry and Physics, 2011. Alternatively, see Wolfram
Alpha for a particular gas
At a specified temperature, the (p, Vm) curve is an isotherm. Figure 1.9 shows
isotherms for carbon dioxide, from 274 to 344 K, in steps of 10 K. As the tem-
perature of the isotherm increases, the maximum and minimum of the isotherms
gradually become closer, until they merge at the critical temperature. In the
critical isotherm, pressure and molar volume at which maximum and minimum
coincide correspond to an inflection called a critical point. This point is mathe-
matically defined by the following equalities:

@p @2p
¼0 ¼0 ð1:49Þ
@Vm T @Vm2 T
Fig. 1.9 Van der Waals

isotherms for carbon dioxide,
from 274 to 344 K, in steps of
10 K. Figure obtained with
Mathematica
1.3 Van der Waals Equation 15
Solving this system of equations gives the critical volume Vmc and the critical
temperature Tc. The corresponding pressure is the critical pressure pc (see Mathe-
matica code M2). The critical values depend on the van der Waals constants as
follows:
8a a
Vmc ¼ 3b Tc ¼ pc ¼ ð1:50Þ
27bR 27b2
(E4). Note that (1.48) can be written as
pVm3 ðbp þ RTÞVm2 þ aVm ab ¼ 0 ð1:51Þ
showing that the Van der Waals equation is cubic in Vm. When the linear and
quadratic terms cancel, the three roots coincide, meaning that the maximum and
minimum of the isotherm converge to the critical point. For T > Tc, the isotherms
are single-valued, and as the temperature increases, they gradually lose the
inflection point and become similar to the ideal gas isotherms (Fig. 1.9) (E5).
Figure 1.10 presents a Van der Waals isotherm (continuous solid curve) for
carbon dioxide at 270 K. Let us start at E′ in the gas and progress as the volume
decreases. The point at which the gas starts to condense is E. When the liquid is in
equilibrium with the gas, a progressive volume reduction causes more gas to
condense, while the pressure is kept constant during the liquid–gas equilibrium. The
dashed straight line AE represents the real liquid–gas equilibrium, whereas the
oscillating part of the Van der Waals curve that includes a maximum and a mini-
mum is not real, and so the points along the continuous solid curve ABCDE cannot
be experimentally observed. Maxwell determined the horizontal part AE and found
that the areas a (=ABCA) and b (=CDEC) are equal (Maxwell construction).
Progressing along the straight line EA as the volume decreases due to gradual
condensation of the gas, the point A is attained where all the gas has been con-
densed to the liquid. Thus, the steep curve AA′ corresponds to the liquid. The
steepness of this part of the curve is due to the reduced compressibility of the liquid.
Fig. 1.10 Van der Waals isotherm at 270 K, for carbon dioxide. A′A corresponds to the liquid,
the dashed straight line ACE corresponds to the liquid–gas equilibrium, and EE′ corresponds to
the gas. Graph obtained with Mathematica
16 1 Thermodynamics
Fig. 1.11 Van der Waals

isotherms in reduced
variables. The reduced
temperatures correspond to
the temperatures of the Van
der Waals isotherms for
carbon dioxide in Fig. 1.9.
Mathematica
We now consider the critical values (1.50) and use them as units of volume,
temperature, and pressure to obtain the reduced variables
Vmr ¼ Vm =Vmc Tr ¼ T=Tc pr ¼ p=pc ð1:52Þ
If the values of Vm, T, and p taken from (1.52) are substituted in (1.48), one
obtains the Van der Waals equation in reduced variables

3 1 8
pr þ 2
Vmr ¼ Tr ð1:53Þ
Vmr 3 3
(see Mathematica code M3). This is a universal equation in the sense that it does
not include Van der Waals constants specific to any particular fluid. In fact,
comparison of (1.48) with (1.53) shows that a, b, and the gas constant R correspond
to 3, 1/3, and 8/3, respectively, that is, these values of a, b, and R lead to unit values
of the critical variables. Figure 1.11 shows the Van der Waals isotherms in reduced
variables with the reduced temperatures corresponding to the absolute temperatures
of the isotherms for carbon dioxide in Fig. 1.9 (see Mathematica code M4, E6).
1.4 Mathematical Tools
In the following sections of the present chapter, most of the expressions use dif-
ferentials of thermodynamic functions and deal with inexact differentials and line
integrals. So the present section has in mind those students who might need some
refreshment of mathematical concepts such as exact and inexact differentials, the
fundamental theorem of calculus, and line integrals. Those who are aware of these
mathematical concepts might skip this section.
1.4 Mathematical Tools 17
1.4.1 Exact Differential
Recall that the derivative of a function f(x) at x is given by
f ðx þ DxÞ f ðxÞ
f 0 ðxÞ ¼ lim ð1:54Þ
Dx!0 Dx
The slope of the tangent line at x (Fig. 1.12) is the trigonometric tangent of the
angle h
sinðhÞ
tanðhÞ ¼ ð1:55Þ
cosðhÞ
where h is the angle formed by the tangent line and the positive x-axis. Figure 1.12
shows a curve and its tangent at x, and illustrates the difference between
Dy ¼ yðx þ DxÞ yðxÞ ð1:56Þ
and
dy ¼ y0 ðxÞDx ð1:57Þ
Inspection of Fig. 1.12 also shows that the difference between (1.56) and (1.57),
Dy − dy, goes to zero as Dx ! 0. The question is whether Dy − dy approaches
zero faster or slower than Dx. To answer this question, we need to take the limit of
(Dy − dy)/Dx as Dx ! 0, which is given by
Dy dy Dy
lim ¼ lim y0 ðxÞ ¼ y0 ðxÞ y0 ðxÞ ¼ 0 ð1:58Þ
Dx!0 Dx Dx!0 Dx
Since (Dy − dy)/Dx goes to zero as Dx ! 0, it can be concluded that

Dy − dy approaches zero faster than Dx, and so we can write
Fig. 1.12 Illustration of

Dy and dy
18 1 Thermodynamics
Dy dy ¼ eDx lim e ¼ 0 ð1:59Þ

Dx!0
where eDx goes to zero faster than Dx (for example, as Dx2). Substitution of (1.57)
in (1.59) gives
Dy ¼ dy þ eDx ¼ y0 ð xÞDx þ eDx ð1:60Þ
and so
Dy y0 ðxÞ Dx ð1:61Þ
(note the approximation). We now replace Dx and Dy by finite yet arbitrarily small
variations dx and dy and write
dy ¼ y0 ðxÞdx ð1:62Þ
where dy is the differential of y. Equation (1.62) can be rewritten using the fol-
lowing equivalent forms:
dy df
dy ¼ dx df ¼ dx ð1:63Þ
dx dx
Note that dy/dx or the equivalent df/dx represents a limit [see (1.54)], not a
quotient of differentials dy and dx.
For a function f(x, y) of two independent variables, the partial derivative of
f(x, y) with respect to x is given by
@f f ðx þ Dx; yÞ f ðx; yÞ
fx ¼ ¼ lim ð1:64Þ
@x Dx!0 Dx
assuming that this limit exists. Note that in evaluating the rate of change of f(x, y)
with respect to x, the variable y is held constant. For this reason, we sometimes
write (∂f/∂x)y emphasizing that y is kept constant. The partial derivative of f(x, y)
with respect to y is given by
@f f ðx; y þ DyÞ f ðx; yÞ

fy ¼ ¼ lim ð1:65Þ
@y Dy!0 Dy
Consider now the change Df as a result of changes in both independent variables,

Dx and Dy. We can write the following equalities:
Df ¼ f ðx þ Dx; y þ DyÞ f ðx; yÞ

Df ¼ ½f ðx þ Dx; y þ DyÞ f ðx; y þ DyÞ þ ½f ðx; y þ DyÞ f ðx; yÞ
@f @f
Df ¼ Dx þ e1 Dx þ Dy þ e2 Dy
@x @y
@f @f
Df ¼ Dx þ Dy þ e1 Dx þ e2 Dy
@x @y
If e1 ! 0 and e2 ! 0 as Dx ! 0 and Dy ! 0, then Df ! df, and we can write
@f @f
df ¼ dx þ dy ð1:66Þ
@x @y
where df is the total differential of f. For two independent variables x and y, (1.66)
is the equivalent of the second equality of (1.63) for one independent variable.
We now consider the differential
gðx; yÞdx þ hðx; yÞdy ð1:67Þ
and ask whether this is an exact differential, that is, whether there is a function
f(x, y) such that

@f @f
df ¼ dx þ dy ¼ gðx; yÞdx þ hðx; yÞdy ð1:68Þ
@x y @y x
that is,

@f @f
¼ gðx; yÞ ¼ hðx; yÞ ð1:69Þ
@x y @y x
since dx and dy are arbitrary variations. Note that the first-order derivatives of
f(x, y) are related to each other by

@f @f
@ @
@x @2f @y @2f
¼ ¼ ¼ ð1:70Þ
@y @y@x @x @x@y
that is, the mixed partial second derivatives are equal. Substitution of (1.69) in
(1.70) leads to
@gðx; yÞ @hðx; yÞ
¼ ð1:71Þ
@y @x
which is the requirement for (1.67) to be an exact differential (E7).

20 1 Thermodynamics
1.4.2 Fundamental Theorem of Calculus
Recall now that the integral of a continuous function f(x) in the interval [a,b] is
given by
Zb X
f ðxÞdx ¼ lim f ðxi ÞDxi ð1:72Þ
jLj!0
i
b
where [a,b] is divided into n subintervals Dxi and |L| represents the width of the
largest subinterval (Fig. 1.13).
Since f(xi) times Dxi is the area of the rectangle of height f(xi) and width Dxi, the
summation on the right-hand side of (1.72) corresponds to the area of all the
rectangles, and its limit is the area under the curve in the interval [a,b] (Fig. 1.13). If
the upper limit b is replaced by the variable x defined within the interval [a,b], one
obtains the indefinite integral
Zx
F ð xÞ ¼ f ðuÞdu ð1:73Þ
a
where the variable of integration is denoted by u to prevent confusion with the

integral’s upper limit variable x. It can be shown that
F 0 ðxÞ ¼ f ðxÞ ð1:74Þ
i.e., F(x) is the antiderivative of f(x). In order to understand this important result,
known as the fundamental theorem of calculus, we apply the definition of the
derivative of F(x) and use (1.73), obtaining
Fig. 1.13 The integral of the function in the interval [0.5,3.0] is equal to the area under the curve.
Figure obtained with Mathematica
xZþ Dx Zx xZþ Dx
Fðx þ DxÞ FðxÞ ¼ f ðuÞdu f ðuÞdu ¼ f ðuÞdu ¼ f ðnÞDx

a a x
Fðx þ DxÞ FðxÞ ¼ f ðnÞDx
Fðx þ DxÞ FðxÞ
lim ¼ f ðxÞ i:e: F 0 ðxÞ ¼ f ðxÞ
Dx!0 Dx
The last equality of the first line above shows that f(n) is the mean value of f(x) in
the interval Dx. When applied to a general interval [a,b], this equality expresses the
so-called mean value theorem for integrals. In turn, comparison of the second and
third equalities above allows one to conclude that f(n) ! f(x) as Dx ! 0.
1.4.3 Line Integral
Consider now a curve C parametrically described by the position vector of a

particle:
rðtÞ ¼ xðtÞ i þ yðtÞ j þ zðtÞ k ð1:75Þ
Without loss of generality, we can assume that the parameter t is the time
variable. The particle is driven by a force F over the path defined by the position
vector r(t). When the particle undergoes the displacement dr during the time
interval dt, the infinitesimal work dW done by F on the particle is given by the
scalar product F.dr and the integral of dW over the curve C is
Z Z Ztf
dr
W¼ dW ¼ F:dr ¼ F: dt ð1:76Þ
dt
c ti
where ti and tf are the limits of integration (the initial and final values of t), and r(ti)
and r(tf) denote the initial and final positions of the particle on the curve C. The last
equality of (1.76) results from
dr
dr ¼ dt ð1:77Þ
dt
[see (1.63)]. From the last member of (1.76), the curve can be seen as consisting of
a finite number of infinitesimal segments, each of which has a tangent given by dr/
dt defined at each point of the curve. The direction of the tangent varies continu-
ously with t. Since the integral (1.76) depends on the initial and final points of the
path and on the path itself, it is called a line integral or path integral. The
path-dependence of (1.76) is reflected by the scalar product F . dr/dt, since both
F and dr/dt are path-dependent functions. Here dW is an inexact differential.
22 1 Thermodynamics
Exact differentials have an important property: the integral of an exact differ-

ential depends only on the initial and final points of integration (the limits of
integration),
ZB
Df ¼ df ¼ f ðBÞ f ðAÞ ð1:78Þ
A
If the integration of an exact differential is done over a cycle (Fig. 1.14), then the
initial and final points of integration coincide, and
I
df ¼ 0 ð1:79Þ
From Fig. 1.14 and (1.79), we can write

I Z Z
df ¼ df þ df ¼ 0 ð1:80Þ
A!Bð1Þ B!Að2Þ
Where
Z Z Z
df ¼ df ¼ df ð1:81Þ
A!Bð1Þ B!Að2Þ A!Bð2Þ
The equality between the first and the last members of (1.81) allows one to
conclude that the integral of an exact differential is path-independent, because the
paths 1 and 2 that link A and B were arbitrarily chosen. Then it can be concluded
that (1.79) is a necessary and sufficient condition for df to be an exact differential
[necessary condition: if df is an exact differential, then (1.79) is valid; sufficient
condition: if (1.79) is valid, then df is an exact differential].
Fig. 1.14 The integral of an

exact differential over a cycle
is zero
1.5 Thermodynamic Systems 23
1.5 Thermodynamic Systems
Every thermodynamic study begins with a clear distinction between the object of
study, the system, and everything else, the exterior. Thus, the system is absolutely
central to every thermodynamic study. One can say that the system is the “pro-
tagonist” in thermodynamics. It may be a gas, two metal blocks in contact, a liquid
and its gas in equilibrium, a solute in a solution, a chemical reaction. Besides the
system composition, a thermodynamic study requires the knowledge of its
boundaries, since interactions may occur between the system and its surroundings.
On the whole, system and exterior form the universe of the transformation.
Systems can be classified into three types, isolated, closed, and open, according
to the way they interact with the exterior. An isolated system does not exchange
matter or energy with the exterior (Fig. 1.15). The exterior of an isolated system is
irrelevant, and it may be considered nonexistent. Thus, an isolated system coincides
with the universe of the transformation. For this reason, isolated systems have
considerable theoretical importance, in particular in the formulation of the first and
second laws of thermodynamics, since these laws become simpler and more
comprehensive when they are referred to isolated systems. However, the complete
and effective isolation of a system is difficult to realize, since it requires walls that
behave as perfect insulators. A closed system does not exchange matter with its
surroundings, and an open system exchanges matter and energy with its sur-
roundings (Fig. 1.15).
When the variables needed to characterize a system take defined values, the
system is said to exist in a thermodynamic state, and the variables are called state
variables. Pressure p, temperature T, volume V, and amount of substance nk (the
index k distinguishes different substances) are state variables used in describing
equilibrium states. Variables such as volume and amount of substance provide
information about the size of a system and for that reason are called extensive
variables. On the other hand, variables such as pressure and temperature are local
variables. As such, they do not provide information on the size of the system and
are called intensive variables. The ratio of two extensive variables that have the
same size-dependence is an intensive variable. For example, the ratio of mass to
volume is density, an intensive variable.
Fig. 1.15 Systems can be classified according to the way in which they interact with the exterior
24 1 Thermodynamics
An equilibrium state is a time-independent state. An isolated system away from

equilibrium inevitably evolves to an equilibrium state through irreversible pro-
cesses. Examples of these are heat conduction, matter diffusion, and chemical
reactions. When a system is not at equilibrium, additional information is required
on the time-dependence of the irreversible processes that occur in the system before
it reaches the equilibrium state. During an irreversible process, the values of the
state variables are time-dependent, and time is an external variable (a parameter)
measured by a chronometer. In a chemical reaction, the amounts of reactants and
products are related by the reaction stoichiometry, and it is possible to define a
single state variable, called the extent of the reaction, that accounts for the time
evolution of the chemical reaction toward equilibrium.
The first and second laws of thermodynamics involve the energy U and entropy
S, and these thermodynamic state functions can be expressed as functions of the
state variables temperature, volume, and amount of substance:
U ¼ UðV; T; nk Þ S ¼ SðV; T; nk Þ ð1:82Þ
As we will find later on in this chapter, scientific convenience leads to the

definition of new state functions using the energy U, entropy S, and state variables
p, T, V, and nk. Each thermodynamic state is characterized by the corresponding set
of values of the state variables.
1.6 Heat and Work
The intensive variables temperature and pressure are associated with two funda-
mental types of equilibrium, namely thermal equilibrium and mechanical equilib-
rium. Consider an isolated system formed by two blocks of the same metal, at
temperatures T1 and T2, with T1 > T2 (= T1 − dT) (Fig. 1.16). Heat dQ flows
spontaneously from the part of the system at temperature T1 to the part of the
system at temperature T2, until both parts reach the same temperature. One says that
the system has then attained thermal equilibrium. It is assumed that small vari-
ations in volume (part 1 reduces its volume as it cools; part 2 increases its volume
as it warms) can be easily accommodated by the system, since the heat transferred is
Fig. 1.16 Heat flows from higher to lower temperature

1.6 Heat and Work 25
Fig. 1.17 At constant temperature, a frictionless wall moves from higher to lower pressures
small. Note that a system in thermal equilibrium has the same temperature in all
points.
We now consider a closed system at constant temperature formed by a gas in two
compartments with pressures p1 and p2, and p1 > p2 (= p1 − dp). A frictionless wall
separates these compartments and moves spontaneously from higher to lower
pressures (Fig. 1.17). When the pressures become equal, we say that the system has
attained mechanical equilibrium. A system in mechanical equilibrium has the
same pressure at all points.
1.6.1 Mechanical Work
Consider now a gas at constant temperature in a cylindrical container with a fric-

tionless wall (Fig. 1.18a). Assume that the initial gas pressure is greater than the
external pressure, p > pext. Then the gas volume increases, and its pressure
decreases until the pressure becomes the same on both sides of the frictionless wall.
The gas undergoes an isothermal expansion (an increase in volume) and performs
mechanical work on the surroundings. The work is assigned the sign of the energy
(b)
(a)
Fig. 1.18 a The system does work on the exterior during an isothermal expansion. b Apart from
its sign, the mechanical work is equal to the area under the curve p(V)
26 1 Thermodynamics
change produced on the system, that is, the work done on expansion is negative. In
turn, the work done on compression is positive, just like the corresponding energy
change of the system (on compression, the energy of the system increases). For an
infinitesimal quantity of work, one has
F
dW ¼ Fdx ¼ Adx ¼ pdV ð1:83Þ
A
where F is the force exerted on the moving wall or piston and A is the area of the
piston. Integration of (1.83) for the whole expansion leads to
ZVf
W ¼ pdV ð1:84Þ
Vi
where Vi and Vf are the initial and final volumes, respectively. The absolute value of
the work is the area under the curve p(V) (Fig. 1.18b). Work is a path-dependent
quantity, or process-dependent quantity, not a state function (E8). Mathematically,
dW is an inexact differential.
1.7 First Law
Can energy be obtained from “nothing” or simply destroyed to give “nothing”? In

the past, there were many attempts to obtain mechanical work for free. The idea was
based on so-called perpetual motion machines (Angrist 1968). However, none of
those machines were really successful. Indeed, all the experimental knowledge
gathered so far has confirmed that energy changes should always be accounted for.
Even in the field of nuclear reactions, where energy and mass are interchangeable
(E = mc2), energy conservation has been always confirmed overall.
If one considers the universe (= system + exterior) of a transformation, the total
energy is conserved, that is, the energy change for any infinitesimal transformation
is zero:
dUuniv ¼ 0 ð1:85Þ
For an isolated system, the exterior is irrelevant, because an isolated system does
not exchange matter or energy with its surroundings. Therefore,
dUisol ¼ 0 ð1:86Þ
In a closed system, the system can exchange energy with the surroundings in two
ways, through heat flow and through work. Representing by dQ the heat exchanged
between the system and its surroundings during the time interval dt and by dW the
1.7 First Law 27
work exchanged between the system and its surroundings during the same time
interval, then
dU ¼ dQ þ dW ð1:87Þ
Since dW is an inexact differential, dQ has to be an inexact differential, because only

the sum of two inexact differentials can give the exact differential dU. Thus, every
energy change in U depends only on the initial and final states of the transformation,
not on the path followed to convert the initial state into the final state, and so
I
dU ¼ 0 ð1:88Þ
The dQ and dW terms in (1.87) are positive if they correspond to heat and work
transferred from the exterior to the system (in this case, the system gains energy),
and negative if they correspond to heat or work transferred from the system to the
exterior (in this case, the system loses energy). This convention of signs for dQ and
dW is consistent with the system being the “protagonist” in thermodynamics.
dU ¼ dQ pdV ð1:89Þ
At constant volume, the mechanical work is zero, and
dQV ¼ dU ð1:90Þ
where dQV represents the exchanged heat at constant volume (Fig. 1.19). At
constant pressure,
dQp ¼ ðdU þ pdVÞp ¼ ðdU þ pdV þ VdpÞp ¼ ½dðU þ pVÞp ð1:91Þ
where the added term (Vdp)p does not affect the result, because at constant pressure,
dp = 0, and so Vdp = 0. If we define the state variable enthalpy (en- = within;
thalpein = to heat),
H U þ pV ð1:92Þ
Fig. 1.19 Heat at constant

volume (left) and heat at
constant pressure (right)
28 1 Thermodynamics
then substitution of (1.92) in (1.91) leads to
dQp ¼ dH ð1:93Þ
Note that U is a state function, both p and V are state variables, and the product
and sum of two state variables is a state function. This reasoning allows to conclude
that U + pV is a state function.
1.7.1 Heat Capacities
Expressions (1.90) and (1.93) can be used to define isochoric heat capacity (heat
capacity at constant volume, isochoric from Greek isos “equal” + choros
“space”),

dQV @U
CV ¼ ¼ ð1:94Þ
dT @T V
and isobaric heat capacity (heat capacity at constant pressure, isobaric from Greek
isos “equal” + baros “weight”),

dQp @H
Cp ¼ ¼ ð1:95Þ
dT @T p
(E9, E10). Since heat is an extensive variable and temperature is an intensive

variable, heat capacity is an extensive variable (the ratio of one extensive variable to
one intensive variable). Tabulated values found in books of data are values of
intensive quantities (extensive quantities are sample-dependent quantities). For that
reason, it is useful to define
c ¼ C=m ð1:96Þ
where c is the specific heat capacity and m is the mass, usually expressed in grams.
Thus, the specific heat capacity is the heat capacity for 1 g of substance. Whenever
it is possible and appropriate to deal with an amount of substance, then
Cm ¼ C=n ð1:97Þ
where Cm is the molar heat capacity and n is the amount of substance. Figure 1.20
shows the molar heat capacities at constant pressure (1 bar = 100,000 Pa), for two
important gases in the atmosphere, water and carbon dioxide, as functions of
temperature. At ambient temperature and pressure of 1 atm, the heat capacity for
liquid water approximately doubles the heat capacity for gaseous water.
1.7 First Law 29
Fig. 1.20 Molar heat capacities at constant pressure, for water and carbon dioxide in the gaseous
states, as functions of temperature. Values taken from the Handbook of Chemistry and Physics
2011, and graphs obtained with Mathematica
1.7.2 Calorimeter
The heat released or absorbed in a transformation can be experimentally determined

with a calorimeter. Usually, the transformation is made to occur inside a chamber,
at constant volume, and the heat associated with the transformation is determined
from the change in the temperature of the surrounding water. Derived from (1.94),
the basic equation is
dQV ¼ Ccal dT ð1:98Þ
where Ccal is the heat capacity of the calorimeter. In order to determine dQV from
a measured dT, Ccal has to be known beforehand. To this end, one chooses to
(a) (b)
Fig. 1.21 Schematic representation of calorimeters for determining the heat of a saturated
solution (at left) and the heat of combustion of a compound (at right)
30 1 Thermodynamics
measure dT for a known dQV, which can be produced by a current-carrying resistor.

For a current intensity I and a resistance R, the voltage across the resistor is V = IR
and the heat evolved by the resistance in the time interval dt is I2Rdt. Once Ccal is
known (note that Ccal is a calorimeter constant), dQV can be determined from the
change in temperature dT measured by the same calorimeter.
A calorimeter should be suitable to the type of transformation whose heat is to be
determined (Fig. 1.21). For example, the heat of combustion of a compound is
determined in a bomb calorimeter, with the combustion being carried out inside a
chamber pressurized to about 20 atm with pure oxygen to ensure that the com-
bustion is complete.
When the heat to be determined is smaller than 1 J (for example, in the ther-
modynamic study of biological systems), the sample that absorbs or releases heat is
kept at constant temperature by a very sensitive heat exchanger that measures the
exchanged heat. This type of calorimeter is called an isothermal calorimeter.
1.7.3 Standard States
The thermodynamic properties of a system depend on the nature and physical state
(gas, liquid or pure crystalline solid) of its component substances, as well as on the
physical variables such as temperature, pressure, and, in the case of solutions,
concentrations of solutes (partial pressures for gaseous mixtures). Thus, tabulated
data on a system, besides specifying its temperature, rely on a standard pressure p0
and, in the case of solutions, on a standard molality m0 or a standard molarity c0 for
each solute. Without such standard values, tabulated thermodynamic data could be
of little practical use, due to the dispersion of physical conditions. The most
common choice of tabulating data includes a reference temperature equal to
298.15 K, the pressure p0 = 1 bar (= 100,000 Pa), and, in the case of solutions, the
solute molality m0 = 1 mol of solute per kg of solvent, or the solute molarity
c0 = 1 mol of solute per dm3 of solution. The standard state of a pure substance at
a specified temperature is its most stable state (gas, liquid, or pure crystalline solid)
at the standard pressure p0 = 1 bar.
In the gaseous state, the standard state is the hypothetical state in which the gas
has ideal behavior at the standard pressure p0. This means that a real gas at 1 bar of
pressure is not in its standard state. Deviations from ideal behavior result from
intermolecular interactions that depend on the gas, its pressure, and temperature,
and so would not lead to a definable standard state.
The standard state of a pure liquid or a pure solid is its state at the standard
pressure p0. Unlike gases, liquids and solids have reduced compressibility, since
their volumes are little affected by pressure variation.
The standard state of a solute is its hypothetical state in a solution with ideal
behavior, at the standard concentration m0 or c0, and the standard pressure p0. Ideal
behavior corresponds to the absence of solute–solute intermolecular interactions,
implying that the solute vapor pressure is proportional to the mole fraction of the
solute in solution (Henry’s law). At standard concentration m0 or c0 and standard
1.7 First Law 31
pressure p0, the solute in a real solution is not in the standard state, since the
deviations from ideal behavior are solution-specific, that is, do not lead to a
definable standard state.
Absolute energies cannot be determined. Only energy changes can be measured
or evaluated. This applies to the energy U, the enthalpy H, and other thermody-
namic state functions later considered in this chapter.
The enthalpies of formation of all elemental substances in their standard states
are zero, at all temperatures. Hence, at a specified temperature, the standard molar
enthalpy of formation Hfo of a compound is the difference between its molar
enthalpy in its standard state and the enthalpies of its constituent elements in their
standard states. In addition, at a specified temperature, the standard reaction en-
thalpy Hro is the difference between the stoichiometric sum of the enthalpies of
formation of the products and the stoichiometric sum of the enthalpies of formation
of the reactants, with reactants and products being in their standard states (E11).
Tabulated data of standard thermodynamic properties at T = 298.15 K usually
include data for a large number of compounds (see Handbook of Chemistry and
Physics, 2011).
1.8 Reversible Heat Engine
Can heat be converted into work with 100 % efficiency? All existing experimental
knowledge gives a negative answer to this question. Even in idealized conditions of
maximum efficiency, a fraction of the heat is necessarily released to a heat sink
without being converted into work. In contrast, work can be converted into heat
with 100 % theoretical efficiency. In order to understand the consequences of this
asymmetry of nature, we consider the heat engine (a device that converts heat into
work) functioning under idealized conditions of maximum efficiency.
The first thermodynamic studies on the heat engine were strongly impelled by
the great impact of the steam engine on charcoal extraction, transportation, agri-
culture, and industry during the so-called industrial revolution, from the second half
of the eighteenth century to the first half of the nineteenth. In particular, the studies
carried out by Sadi Carnot (1796–1832) led to an understanding of the conditions
for maximum efficiency of the heat engine.
Under idealized conditions of maximum efficiency, the essential workings of the
heat engine are schematically shown in Fig. 1.22. During a complete cycle, the heat
engine absorbs heat Qh from the heat source (hot reservoir) at temperature Th,
converts part of Qh into work W, and discards the remaining heat Qc (= Qh − W) to
the heat sink (cold reservoir) at temperature Tc (note that Qh, W, and Qc represent
absolute values). After performing these actions, the heat engine returns to its initial
state to complete a cycle. Hence, the variations of all state variables in the heat
engine (our system) are zero. The temperatures of the heat source and heat sink are
not affected by the release or capture of the heats Qh and Qc, respectively. Hence,
32 1 Thermodynamics
Fig. 1.22 Essential

workings of an idealized heat
engine
the heat source and the heat sink are assumed to be infinite heat reservoirs, that is to
say, to have infinite heat capacities. The efficiency of the heat engine is given by
W
g¼ ð1:99Þ
Qh
According to the first law,
Qh Qc ¼ W ð1:100Þ
Qc
g¼1 ð1:101Þ
Qh
Thus, the heat engine efficiency is a positive quantity smaller than 1. Maximum
efficiency corresponds to the minimum value of Qc/Qh, that is, to the minimum
value of Qc for the same value Qh. The maximum efficiency of the heat engine
cannot depend on the engine materials or engine gears, since otherwise, the effi-
ciency could in principle be surpassed by an improved choice of these items. This
reasoning leads to the conclusion that the maximum efficiency of a heat engine can
depend only on the temperatures of the heat source and heat sink, sh and sc,
Qc
gmax ¼ 1 ¼ 1 f ðsc ; sh Þ ð1:102Þ
Qh
Carnot’s theorem. Note that under the conditions of maximum efficiency, volume
changes associated with mechanical work cannot produce thermal gradients, for
otherwise, additional heat would be released to the heat sink, and maximum effi-
ciency would not be attained. Hence, the heat engine with maximum efficiency
takes unlimited time to accomplish its operations. Only then can the heat engine
1.8 Reversible Heat Engine 33
(a) (b)
Fig. 1.23 A Carnot engine cycle is formed by an isothermal expansion (AB), an adiabatic expansion
(BC), an isothermal compression (CD), and an adiabatic compression (DA). a p = Rh/V surface
with isothermal and adiabatic curves. b Surface (a) projected in the plane (p, V, 100) seen from the
point (p, V, h) = (0, 10, 200). Ideal gas surface with isotherms and adiabatic curves obtained with
Mathematica
follow a reversible path and attain both thermal and mechanical equilibrium at each
point in the path, thus being a reversible heat engine.
1.8.1 Carnot’s Heat Engine
Carnot’s heat engine consists of an ideal gas that operates over a reversible cycle
formed by steps A ! B (isothermal expansion), B ! C (adiabatic expansion; see
§3), C ! D (isothermal compression), and D ! A (adiabatic compression)
(Fig. 1.23) (see Mathematica code M5). The ideal gas temperatures for the heat
source and the heat sink are represented by hh and hc, respectively.
It can be shown (see §4) that the Carnot engine’s efficiency depends only on the
temperatures of the heat source and heat sink according to the expression
Qc hc
gCarnot ¼ 1 ¼1 ð1:103Þ
Qh hh
where hh and hc are the temperatures given by the ideal gas empirical scale of
temperatures, (E12).
1.8.2 Absolute Temperature
As shown before, the heat engine’s efficiency is a positive quantity less than 1 [see
(1.101)]. For a reversible heat engine, the efficiency depends only on the
34 1 Thermodynamics
temperatures of the hot and cold reservoirs. These two conclusions are combined in
(1.102), which, in turn, leads to
Qc
¼ f ðsc ; sh Þ ð1:104Þ
Qh
This equality applies to a reversible heat engine operating between temperatures

sh and sc (Fig. 1.24a). Consider now two reversible heat engines, one operating
between temperatures sh and s′, and the other operating between temperatures s′
and sc (Fig. 1.24b). The reservoir at the intermediate temperature s′ operates as a
cold reservoir for the first reversible heat engine and absorbs the amount of heat Q′,
and operates as a hot reservoir for the second reversible heat engine, releasing the
same amount of heat Q′. Adding W′ (= Qh − Q′) to W″ (= Q′ − Qc) cancels Q′ and
gives the work W (= W′ + W″ = Qh − Qc). From (1.104), we can write
Qc Q0 0 0 f ðsc ; s0 Þ gðsc Þ
¼ f ðs c ; s Þf ðs ; s h Þ ¼ ¼ ð1:105Þ
Q0 Qh f ðsh ; s0 Þ gsh
where the second equality results from Q′/Qh = 1/(Qh/Q′) = 1/f(sh,s′).

The reversible heat engines of Fig. 1.24 receive the same amount of heat Qh
from the heat source at sh, release the same amount of heat Qc to the heat sink at sc,
produce the same amount of work W = W′ + W″ thus having the same efficiency.
Therefore, the intermediate temperature s′ is irrelevant and can be omitted as an
argument of the function f, thus justifying the last equality of (1.105). In addition,
we can assume that the function g(s) takes the simple form of a constant c times s
and define a temperature scale for which c = 1, i.e.,
Fig. 1.24 Reversible heat (a) (b)

engines: the efficiencies of
(a) and (b) are equal
1.8 Reversible Heat Engine 35
gðsÞ ¼ c s T ð1:106Þ
From (1.104) and (1.105), it can be concluded that the functions f and g can have
only positive values. Thus, the identity of (1.106) defines a scale of absolute
temperatures. Substitution of (1.106) in (1.105) and then in (1.102) leads to
Qc Tc
gmax ¼ 1 ¼1 ð1:107Þ
Qh Th
(see Mathematica code M6). Rearrangement of (1.107) leads to the following

equality for a reversible cycle:
Qh Qc
þ ¼0 ð1:108Þ
Th Tc
Comparison of (1.107) with (1.103) allows on to conclude that the scales of the
ideal gas temperature and absolute temperature coincide.
Consider now a heat engine that operates in a cycle between temperatures Th and
Tc and absorbs Qh from the heat source but produces less work than the corre-
sponding Carnot engine (Wirr < W), since volume changes during production of
mechanical work are associated with thermal gradients. These volume changes lead
c [ Qc Þ. For this irreversible
to irreversible heat discarded to the heat sink ðQirr
cycle,
Qh Qirr
þ c
\0 ð1:109Þ
Th Tc
1.9 Entropy and the Second Law
The equality for a Carnot cycle (1.108) can be generalized to an arbitrary reversible
cycle. To this end, consider a reversible cycle and the enclosed area occupied by a
mesh of Carnot cycles in which the temperatures of the isotherms differ by DT. This
mesh comprises two types of Carnot cycle (Fig. 1.25): those that are completely
inside the arbitrary cycle and share their isothermal and adiabatic transformations
with four neighboring Carnot cycles, and those that are over the cyclic path and
have fewer than four neighboring Carnot cycles. Transformations that are shared by
two contiguous Carnot cycles are carried out in opposite directions in both cycles
and cancel out. The isothermal and adiabatic transformations that are not canceled
describe a zigzag path over the periphery of the arbitrary cycle (Fig. 1.25). In the
limit of infinitesimal Carnot cycles (the difference between temperatures of con-
secutive isotherms tends to zero, DT ! 0, and consecutive adiabatics gradually get
closer), the zigzag path becomes indistinguishable from an arbitrary reversible
cycle. Hence, the generalization of (1.108) leads to
36 1 Thermodynamics
Fig. 1.25 An arbitrary reversible cycle can be covered to the required precision by a mesh of
infinitesimal Carnot cycles. Carnot cycles are carried out in the same direction (in this figure, they
are performed in the clockwise direction). Figure obtained with Mathematica
I
Qh Qc dQrev
þ ¼0 ) ¼0 ð1:110Þ
Th Tc T
Unlike Qh and Qc, which represent absolute values, dQrev includes a positive or
negative sign for absorbed or released heat, respectively. The last equality of
(1.110) shows that dQrev/T is an exact differential [see (1.79)], that is, dQrev/T is the
differential of a state function, which was named by Rudolf Clausius (1822–1888)
entropy (en “inside” + from Greek tropé “transformation”) and is usually rep-
resented by a capital S:
I I
dQrev dQrev
¼0 dS ¼ 0 dS ¼ ð1:111Þ
T T
(E13, E14). In turn, generalization of (1.109) for an arbitrary irreversible cycle

leads to
I
Qh Qirr dQ
þ c
\0 ) \0 ð1:112Þ
Th Tc T
Formulas (1.111) and (1.112) can be combined as follows:

I I
dQ dQ
0 dS ¼ 0 ) dS ð1:113Þ
T T
1.9 Entropy and the Second Law 37
where the inequalities hold for irreversible processes. An irreversible process occurs
in a specific direction, not in its reverse, and evolves to equilibrium. Examples of
irreversible processes are the flow of heat from a higher to a lower temperature, the
motion of a piston from a higher to a lower pressure, the melting of an ice cube at
ambient temperature, the fall of a weight from a higher to a lower altitude.
The last inequality of (1.113), called Clausius’s inequality, shows that dS is
greater than dQ/T when the latter corresponds to an irreversible exchange of heat
between system and surroundings. Hence, for an irreversible process, the difference
between dS and dQ/T is always greater than zero,
dQ
dS [0 ð1:114Þ
T
According to Ilya Prigogine (1917–2003, Nobel Prize in chemistry in 1977), a
distinction should be made between two terms included in the entropy change
dS (Fig. 1.26) (see Kondepudi and Prigogine, Further Reading). The first, deS, is
the transfer of entropy across system boundaries due to energy and matter ex-
changes with the surroundings. In principle, these energy and matter exchanges can
be performed reversibly [deS = dQrev/T; see the last equality of (1.111)], and the
corresponding entropy change deS is either positive or negative (no thermodynamic
law can be formulated on this term). In turn, the second term, diS, is the entropy
produced within the system by irreversible processes that drive the system toward
equilibrium. This term is always positive. The following expressions summarize
these considerations:
dQ
dS ¼ de S þ di S de S ¼ di S [ 0 ð1:115Þ
T
The statement according to which an irreversible process within the system has
always a positive entropy change (diS > 0) is one of the formulations of the second
law of thermodynamics (E20). While the equilibrium is a time-independent state,
an irreversible process has time as an associated external variable, and the arrow of
time points toward equilibrium. Therefore, for an irreversible process,
Fig. 1.26 Irreversible processes that occur inside the system are associated with an increase of
entropy
38 1 Thermodynamics
di S
[0 ð1:116Þ
dt
where diS/dt is called the rate of entropy production.

The first and second laws of thermodynamics enable the determination of energy
and entropy changes, not of energy and entropy absolute values. However, in 1906,
Walther Nernst (1864–1941) found that the entropies of all systems approach zero
as the temperature tends to zero. This statement, usually called the third law of
thermodynamics, has been empirically confirmed and enables the determination of
absolute values for the entropy.
1.10 Irreversible Processes
We now consider examples of irreversible processes such as heat flow, gas

expansion, matter diffusion, and chemical reactions. For each of them, we present
the expressions of diS corresponding to a time interval dt and of the rate of entropy
production diS/dt.
1.10.1 Heat Flow
We begin with an isolated system formed by two blocks of the same metal at two
different temperatures T1 and T2 (Fig. 1.27). An isolated system does not exchange
energy or matter with the surroundings, and so dV = 0, dU = 0, and deS = 0 (E15).
For an infinitesimal heat dQ (dQ is an absolute value) passing from the metal block
at higher temperature to the metal block at lower temperature, we can write

dQ dQ 1 1
di S ¼ þ ¼ dQ [ 0 ð1:117Þ
T1 T2 T2 T1
since the entropy is an additive function, and for the rate of entropy production,
Fig. 1.27 In an isolated system, heat flows irreversibly from a higher to a lower temperature
1.10 Irreversible Processes 39

di S 1 1 dQ
¼ [0 ð1:118Þ
dt T2 T1 dT
with diS being produced in time dt. As we know from experience, heat flows
irreversibly from higher to lower temperature, and this process corresponds to
diS greater than zero.
1.10.2 Gas Expansion
Consider now a closed system at a defined temperature (dT = 0), with fixed boundaries
(dV = 0) and a frictionless wall separating a gas at pressures p1, on one side of the
piston, and p2, on the other side of the piston (p1 > p2) (Fig. 1.28). The changes in
volume of parts 1 and 2 of the system have the same absolute value, and so
dV1 ¼ dV2 dVa ð1:119Þ
The system, consisting of a gas with a frictionless wall separating two parts at
different pressures, exchanges heat with a thermostat in its surroundings, to
maintain a constant temperature. The frictionless wall moves from higher to lower
pressure, subject to the pressure difference p1 − p2 (p1 > p2). Then,
Tde S ¼ dU ð1:120Þ
and, making use of (1.119),
Tdi S ¼ p1 dV1 þ p2 dV2 ¼ ðp1 p2 ÞdVa [ 0 ð1:121Þ
Therefore,
di S p1 p2 dVa
¼ [0 ð1:122Þ
dt T dt
Fig. 1.28 At constant

temperature, a frictionless
wall moves from higher to
lower pressure
40 1 Thermodynamics
1.10.3 Diffusion of Matter
The situation depicted in Fig. 1.29 represents a closed system consisting of two
solutions with the same solute and same solvent, at constant temperature, separated
by a semipermeable membrane that hypothetically allows the passage of the solute
but not the solvent. It is further assumed that the parts of the system have the same
volume, but the amount of solute is greater in part 1 than in part 2. Considering that
the system has fixed boundaries (dV = 0), the changes in the amounts of solute are
associated with irreversible diffusion of matter through the semipermeable mem-
brane. The equation for dU is
dU ¼ TdS þ l1 dn1 þ l2 dn2 ð1:123Þ
where

@U @U
l1 ¼ l2 ¼ ð1:124Þ
@n1 S;n2 @n2 S;n1
are called the chemical potentials for the solute in parts 1 and 2, and
dn1 ¼ dn2 dn ð1:125Þ
where dn1 is negative (the amount of solute in part 1 decreases), dn2 is positive (the
amount of solute in part 2 increases) and dn is the extent of solute spontaneously
diffused through the semipermeable membrane in the time interval dt. From
(1.124), we can conclude that the chemical potential is an intensive variable. The
dn1 and dn2 terms account for changes in energy U due to solute transfer from part
1 to part 2. Expression (1.123) shows that U = U (S, n1, n2), since the volume of the
whole system is kept constant (V is not a variable in this experiment).
Making use of the first equality in (1.115), Eq. (1.123) can be split in two
equations, one referring to the exchange of energy across the system boundaries
(the whole system is kept at constant temperature),
Fig. 1.29 At constant temperature, matter diffuses through the membrane from higher to lower
chemical potential
1.10 Irreversible Processes 41
dU ¼ Tde S ð1:126Þ
the other being due to irreversible changes that occur inside the system, namely, in
the amounts of the solute in parts 1 and 2, dn1 and dn2, and in the corresponding
entropy change diS,
0 ¼ Tdi S þ l1 dn1 þ l2 dn2 ð1:127Þ
Note that addition of (1.126) and (1.127) gives back (1.123). Substitution of
(1.125) in (1.127) leads to
1 l l2
di S ¼ ðl1 dn1 þ l2 dn2 Þ ¼ 1 dn [ 0 ð1:128Þ
T T
and
di S l1 l2 dn
¼ [0 ð1:129Þ
dt T dt
Therefore, diffusion of matter occurs from the higher chemical potential l1 to
the lower chemical potential l2. In the irreversible process of matter flow, l1
decreases, whereas l2 increases until equality between the chemical potentials l1
and l2 is attained at equilibrium.
1.10.4 Chemical Reaction
Consider now a closed system with the following chemical reaction at a defined
temperature T and pressure p,
aAðgÞ þ bBðgÞ ! cCðgÞ þ dDðgÞ ð1:130Þ
where a, b, c, and d are the stoichiometric coefficients. Infinitesimal changes in the

amounts of reactants A and B and products C and D are related to one another by
the reaction stoichiometry and
dnA dnB dnC dnD

¼ ¼ ¼ dn ð1:131Þ
a b c d
where the identity defines the state variable n, called the extent of the chemical
reaction. The stoichiometric coefficients for the reactants are multiplied by −1
because the infinitesimal changes in the corresponding amounts of the reactants, dnA
and dnB, are negative. Equation (1.131) can be rewritten in the more general form
dnk ¼ vk dn ð1:132Þ
42 1 Thermodynamics
where mk takes the values −a, −b, c, d, which are called stoichiometric numbers.
Combining the first law (1.87) and the second law (1.115) and generalizing from
(1.127), we can write
dU ¼ Tde S pdV þ Tdi S þ lA dnA þ lB dnB þ lC dnC þ lD dnD ð1:133Þ
and split this equation into
dU ¼ Tde S pdV 0 ¼ Tdi S þ lA dnA þ lB dnB þ lC dnC þ lD dnD ð1:134Þ
In turn, substitution of (1.131) in the second equation of (1.134) leads to
1
di S ¼ ðlA dnA þ lB dnB þ lC dnC þ lD dnD Þ
T
1
¼ ðalA blB þ clC þ dlD Þdn ð1:135Þ
T
Theophile De Donder (1872–1957) defined the state variable called affinity of
the chemical reaction as the stoichiometric sum of the chemical potentials of the
reactants minus the stoichiometric sum of the chemical potentials of the products,
A ¼ alA þ blB clC dlD ð1:136Þ
(see Kondepudi, Further Reading). Substitution of (1.136) in (1.135) gives
A
di S ¼ dn [ 0 ð1:137Þ
T
Considering now that dn refers to the time interval dt, we can write

di S A dn
¼ [0 ð1:138Þ
dt T dt
where dn/dt is the rate of the chemical reaction. As the chemical reaction proceeds
to equilibrium, where dn/dt = 0, the chemical affinity decreases to zero, a value
attained at equilibrium.
1.11 Chemical Potential
The concept of chemical potential is essential to describing the change of compo-

sition of a mixture, be it a physical change or a chemical reaction. For that reason, it
is important to find a functional dependence between chemical potential and pressure
of gases and between chemical potential and concentration in solutions.
1.11 Chemical Potential 43
1.11.1 Gibbs–Duhem Equation
The functional dependence between chemical potential and pressure uses an

important thermodynamic equation known as the Gibbs–Duhem equation [Josiah
Willard Gibbs (1839–1903), Pierre Duhem (1861–1916)], which shows that the
intensive variables temperature, pressure, and chemical potential are not
independent.
We begin by combining the first and second laws for a closed system with
possible changes in the amounts nk of the system components due to a chemical
reaction and write
X
dU ¼ TdS pdV þ lk dnk ð1:139Þ
k
From this equation, we conclude that

@U @U @U
¼T ¼ p ¼ lk ð1:140Þ
@S V;nk @V S;nk @nk S;V;nj ðj6¼kÞ
From (1.139), we learn that the energy U is a function of entropy S, volume V,

and amount of substance nk for all k values, U = U(S, V, n1, …, nk, …). Here U, S,
V, and nk are extensive variables. Hence, multiplication of S, V, and all the nk by a
factor k is equivalent to multiplying U by the same factor,
U ðkS; kV; kn1 ; . . .; knk ; . . .Þ ¼ kUðS; V; n1 ; . . .; nk ; . . .Þ ð1:141Þ
Differentiating both members of (1.141) with respect to k (differentiation of the

first member requires the chain rule for partial differentiation) leads to
@U @ðkSÞ @U @ðkVÞ X @U @ðknk Þ dðkUÞ

þ þ ¼
@ðkSÞ @k @ðkVÞ @k k
@ðknk Þ @k dk
which simplifies to
@U @U X @U
Sþ Vþ nk ¼ U
@ðkSÞ @ðkVÞ k
@ðknk Þ
Considering now k = 1 (k is an arbitrary scaling factor), we obtain
@U @U X @U
Sþ Vþ nk ¼ U ð1:142Þ
@S @V k
@nk
Substitution of (1.140) leads to

44 1 Thermodynamics
X
U ¼ TS pV þ lk nk ð1:143Þ
k
In turn, differentiation of (1.143) gives

X X
dU ¼ TdS þ SdT pdV Vdp þ lk dnk þ nk dlk ð1:144Þ
k k
If we now subtract (1.139) from (1.144), we obtain

X
0 ¼ SdT Vdp þ nk dlk ð1:145Þ
k
This equality, known as the Gibbs–Duhem equation, shows that the intensive
variables T, p, and lk are not independent variables. For a system with one com-
ponent, the Gibbs–Duhem equation takes the form
dl ¼ Sm dT þ Vm dp ð1:146Þ
where Sm and Vm represent the molar entropy and molar volume, respectively. At a
specified temperature (dT = 0), integration of (1.146) gives
Zl Zp
dl ¼ Vm dp ð1:147Þ
l0 p0
where the lower limits are the chemical potential and pressure of a reference state
(usually, p0 = 1 bar), and the upper limits are the chemical potential and pressure as
variables (this equation is an equality between indefinite integrals). After integration
of the first member, we obtain
Zp
lðp; TÞ ¼ l ðTÞ þ
0
Vm ðp; TÞdp ð1:148Þ
p0
We will now proceed by applying this equation to ideal and real gases, liquid
solutions, pure liquids, and solids.
1.11.2 Ideal Gas
For an ideal gas, Vm = RT/p. Hence,

p
lðp; TÞ ¼ l ðTÞ þ RT ln 0
0
ð1:149Þ
p
Considering an ideal gas mixture, the chemical potential for the k component is
given by

pk
lk ðpk ; TÞ ¼ l0k ðTÞ þ RT ln ð1:150Þ
p0
where the partial pressure for the k component is given by
pk ¼ x k p ð1:151Þ
with xk the molar fraction for the k component and p the total pressure of the
gaseous mixture. Substitution of (1.151) in (1.150) gives

p
lk ðp; T; xk Þ ¼ l0k ðTÞ þ RT ln 0 þ RT ln xk ¼ l0k ðp; TÞ þ RT ln xk ð1:152Þ
p
1.11.3 Real Gases
Let us now consider the chemical potential for a real gas and its deviation from the
corresponding ideal gas that defines the reference state. In order to compare molar
volumes of real and ideal gases, we take the compressibility factor Z defined by
RT
Vm;real ¼ ZVm;ideal ¼ Z ð1:153Þ
p
This definition shows that Z = 1 for the ideal gas. Subtracting Vm,ideal from each
member of the above equation leads to
Z1
Vm;real Vm;ideal ¼ ðZ 1ÞVm;ideal ¼ RT ð1:154Þ
p
Making use of (1.148) for the real and ideal gases and noting that the standard
state is defined by the ideal gas, we can write
Zp
lreal ðp; TÞ ¼ l0ideal ðTÞ þ Vm;real ðp; TÞdp
p0
ð1:155Þ
Zp
lideal ðp; TÞ ¼ l0ideal ðTÞ þ Vm;ideal ðp; TÞdp
p0
46 1 Thermodynamics
Now we take the difference between these equations and use (1.154) to obtain
Zp
Z1
lreal ðp; TÞ ¼ lideal ðp; TÞ þ RT dp ð1:156Þ
p
p0
G.N. Lewis (1875–1946) defined a quantity with pressure units called fugacity
f that gives a pressure-dependence similar to that of the ideal gas,

f
lreal ðp; TÞ ¼ lideal ðp; TÞ þ RT ln ð1:157Þ
p
[compare with (1.149)]. Fugacity is a corrected pressure that takes into account
deviations from the ideal gas’s behavior due to the presence of intermolecular
interactions. When the pressure approaches zero, a real gas acquires ideal gas
behavior. Hence,

f
limp!0 ¼1 ð1:158Þ
p
1.11.4 Liquid Solutions
Diluted nonionic solutions exhibit a chemical potential-dependence on the solute

mole fraction similar to that of ideal solutions [see the last equation of (1.152)],
lk ðT; xk Þ ¼ l0k ðTÞ þ RT ln xk ð1:159Þ
where xk
1 and the dependence on pressure has been ignored, since liquids, and
condensed phases (liquids and solids) in general, have little sensitivity to pressure.
For nonideal solutions, Lewis introduced the concept of activity to keep the same
form of (1.159), and wrote
lk ðT; xk Þ ¼ l0k ðTÞ þ RT ln ak ð1:160Þ
where the activity ak is given by
ak ¼ c k x k ð1:161Þ
and ck is called the activity coefficient (ck = 1 implies ak = xk, that is, ideal
behavior).
1.11.5 Pure Liquids and Solids
Returning to the integration of both members of (1.146), between 0 K and tem-

perature T, and pressure between p0 and p, we obtain
ZT Zp
lðp; TÞ ¼ lðp ; 0Þ
0
Sm ðTÞdT þ Vm dp ð1:162Þ
0 p0
Solids and liquids are little affected by pressure changes. Hence, the molar
volume can be considered approximately a constant. Then,
lðp; TÞ ¼ l0 ðTÞ þ Vm ðp p0 Þ ð1:163Þ
where
ZT
l0 ðTÞ lðp0 ; 0Þ Sm ðTÞdT ð1:164Þ
0
Making use of (1.160) for one component and of (1.163), we can write
l0 ðTÞ þ Vm ðp p0 Þ ¼ l0 ðTÞ þ RT ln a ð1:165Þ
For solids and liquids, l0(T) is several orders of magnitude greater than
VmDp. For instance, for liquid water and Dp = 1 bar (= 100000 Pa), VmDp = 1.8
J mol−1 and l0(T) = −237.1 kJ mol−1 (chemical potential value taken from
Handbook of Chemistry and Physics, 2011). Hence, within a good approximation,
the activity of pure solids and liquids is equal to 1 (ln a 0).
1.12 Gibbs Energy
At constant temperature and pressure, the state function that evolves to a minimum
at equilibrium is the Gibbs energy, defined by the following identity:
G U þ pV TS ð1:166Þ
Differentiating G and using (1.89) and (1.115) for a closed system, we obtain
dG ¼ dU þ pdV þ Vdp Tde S Tdi S SdT ¼ SdT þ Vdp Tdi S ð1:167Þ

48 1 Thermodynamics
Fig. 1.30 At equilibrium, the Gibbs energy attains a minimum, and the reaction affinity is zero.
Graphs obtained with Mathematica
At constant temperature and pressure (dT = 0, dp = 0), we can write
dGT;p ¼ Tdi S 0 ð1:168Þ
This result shows that at constant temperature and pressure, the Gibbs energy
evolves through an irreversible process to a minimum at equilibrium (diS > 0, and
so dGT,p is negative; at the minimum, the entropy is a maximum). When the
irreversible process is a chemical reaction,
X
dGT;p ¼ Tdi S ¼ lk dnk ¼ Adn ð1:169Þ
k
where A is the chemical reaction affinity [see (1.132) and (1.136)]. Making use of
(1.63) in (1.169), we can write

@G @G
¼ lk ¼ A ð1:170Þ
@nk T;p;nj ðj6¼kÞ @n T;p
The second of the above equations shows that the partial derivative of the Gibbs
energy with respect to the extent of the reaction, at constant temperature and
constant pressure, is the negative of the chemical reaction affinity. At equilibrium,
the Gibbs energy attains a minimum, and the chemical reaction affinity is zero
(Fig. 1.30).
1.12.1 Chemical Potential and Gibbs Energy of Formation
The first of the above equations shows that the partial derivative of the Gibbs
energy with respect to the amount of substance of the k component, at constant
temperature, constant pressure, and fixed amounts of other substances in the
reaction mixture, is equal to the chemical potential of the k component of the
1.12 Gibbs Energy 49
chemical reaction mixture. While this is the formal relationship between chemical
potential and Gibbs energy (the chemical potential is the partial molar Gibbs
energy), we can proceed to find a more pragmatic expression between chemical
potential and Gibbs energy as experimentally measurable physical quantities.
Gibbs energy is an extensive state variable. In order to understand the rela-
tionship between standard Gibbs energies of formation and standard chemical
potentials, we substitute (1.143), obtained during the derivation of the Gibbs–
Duhem equation, into (1.166), the definition of Gibbs energy, leading to
X
G¼ lk nk ð1:171Þ
k
For one compound, we can write
G
l ¼ Gm ¼ ð1:172Þ
n
thus concluding that the chemical potential is a molar Gibbs energy, Gm (chemical
potential and molar Gibbs energy have the same units, energy per mole). Therefore,
G is an extensive state function proportional to n, with the proportionality constant
being the chemical potential. However, in order to determine Gibbs energies
experimentally, zeros for the Gibbs energies and chemical potentials need to be
established. As mentioned before, the standard state of a pure substance at a
specified temperature is its most stable state (gas, liquid, or pure crystalline solid) at
the standard pressure p0 = 1 bar. At a specified temperature, the standard molar
Gibbs energy of formation DG0f of a particular compound is the difference
between the molar Gibbs energy of this compound in its standard state and the
Gibbs energies of its constituent elements in their standard states. Since atoms of
elemental substances do not transmute in chemical reactions (only nuclear reactions
outside the thermodynamic realm can lead to transmutation of chemical elements),
all elemental substances in their standard states are assumed to have a standard
chemical potential and a standard molar Gibbs energy of formation equal to zero at
all temperatures,
l0a ¼ DG0f;a ¼ 0 ð1:173Þ
where a generically denotes elemental substances. Having established the zeros for
the standard chemical potentials and the standard molar Gibbs energies of formation
of compounds, we can now conclude that at a specified temperature, the standard
chemical potential of substance k is equal to the standard Gibbs energy of formation
of the same substance,
l0a ¼ DG0f;k ð1:174Þ

50 1 Thermodynamics
where k stands for a pure compound. At a specified temperature, the standard

Gibbs energy of a chemical reaction, DG0r , is the difference between the stoi-
chiometric sum of the standard Gibbs energies of formation of the products and the
stoichiometric sum of the standard Gibbs energies of formation of the reactants,
with reactants and products being in their standard states.
1.12.2 Gibbs–Helmholtz Equation
The Gibbs–Helmholtz equation relates the temperature-dependence of the Gibbs

energy to the corresponding enthalpy value and can be used to determine the
enthalpy variation from data on the temperature variation of the Gibbs energy. To
arrive at the Gibbs–Helmholtz equation, we begin by considering a closed system
at equilibrium (diS = 0) and substitute this equation into (1.167), thus obtaining

@G
¼ S ð1:175Þ
@T p
Combining the definition of Gibbs energy [see (1.166) and (1.92)] with (1.175),
we can write

@G
G ¼ H þT ð1:176Þ
@T p
By dividing both members of this equality by T2 and rearranging, we obtain

1 @G G H
¼ 2 ð1:177Þ
T @T p T 2 T
that is,
1
@ TG H
¼ 2 ð1:178Þ
@T T
If we now apply this equation to the Gibbs energy variation of a chemical

reaction DG and the corresponding enthalpy change DH, we obtain
1
@ DG DH
T
¼ 2 ð1:179Þ
@T T
1.13 Chemical Equilibrium 51
1.13 Chemical Equilibrium
Consider now the chemical reaction
aAðgÞ þ bBðg) ! cCðg) + dDðg) ð1:180Þ
where both reactants and products are gases. Breaking down the Gibbs energy of
the chemical reaction into the Gibbs energies of formation of products and reac-
tants, and using the relationship between Gibbs energy of formation and chemical
potential, we can write
DGr ¼ ½ðcDGf;C þ dDGf;D Þ ðaDGf;A þ bDGf;B Þ ¼ ðclC þ dlD Þ ðalA þ blB Þ

ð1:181Þ
Use of (1.160) allows writing

DGr ¼ cl0C þ dl0D al0A þ bl0B þ RT ln v ð1:182Þ
where the reaction quotient v is given by
acC adD
v¼ ð1:183Þ
aaA abB
An equation equivalent to (1.181) can be written for the standard states of

products and reactants
DG0r ¼ ½ðcDG0f;C þ dDG0f;D Þ ðaDG0f;A þ bDG0f;B Þ ¼ ðcl0C þ dl0D Þ ðal0A þ bl0B Þ

ð1:184Þ
Combining (1.182) and (1.184) leads to
DGr ¼ DG0r þ RT ln v ð1:185Þ
When approaching chemical equilibrium, ΔGr tends to zero, and the reaction
quotient tends to the equilibrium constant,
DGr ! 0 v ! Ke ð1:186Þ
where
acC;eq adD;eq
Ke ¼ ð1:187Þ
aaA;eq abB;eq
52 1 Thermodynamics
Applying (1.186) to (1.185) gives
DG0r ¼ RT ln Ke ð1:188Þ
(E16; see Mathematica code M7). From (1.160) and (1.184), it can be concluded
that the equilibrium constant is a function of temperature only. This important result
is at the origin of the expression “equilibrium constant.”
1.14 Gibbs Phase Rule
In thermodynamics, distinct physical states of matter are called phases. For a

one-component system, the occurrence of different phases and the equilibria
between them can be represented in a graph of pressure as a function of temper-
ature, called a phase diagram. The p−T curves of a typical phase diagram for a
one-component system (Fig. 1.31) indicate two-phase equilibria, namely liquid
gas, solid liquid, and solid gas equilibria. The point t that gives the pressure
and temperature of the three-phase equilibrium solid liquid gas is called the
triple point. The point c is the critical point Tc, the temperature above which a
supercritical fluid cannot be liquefied by increase of pressure.
Let us consider a closed system consisting of two phases, liquid and gas
(Fig. 1.31). At equilibrium, the chemical potentials of the liquid and gas are equal,
lliq ðp; TÞ ¼ lgas ðp; TÞ ð1:189Þ
there are no irreversible processes in the system, and the entropy production is zero
[see (1.169)],

di S lliq ðp; TÞ lgas ðp; TÞ dn
¼ ¼0 ð1:190Þ
dt T dt
Fig. 1.31 Typical phase

diagram for a one-component
system
1.14 Gibbs Phase Rule 53
Hence,

p
l0liq ðTÞ ¼ l0gas ðTÞ þ RT ln 0 ð1:191Þ
p
[see (1.149) and (1.163)], where p is the vapor pressure of the pure liquid. After
rearrangement, we can write

p l0liq ðTÞ l0gas ðTÞ
ln 0 ¼ ð1:192Þ
p RT
(E17). This equation allows us to conclude that the vapor pressure is a function of
temperature, a functional dependence that corresponds to the p–T curve of the
liquid–gas equilibrium in the phase diagram. Equilibria between any two phases of
one-component systems also lead to p-T curves of phase diagrams. Note that these
curves relate two intensive variables, namely p and T.
Besides pressure and temperature, solutions have additional intensive variables
related to composition. Let us consider a system at equilibrium containing C com-
ponents in each of its P phases: how many independent intensive variables or
degrees of freedom does such a system have?
If composition variables are expressed by mole fractions, then the sum of all
mole fractions in each phase equals 1. In particular, for phase a,
X
C
ðaÞ
xk ¼ 1 ð1:193Þ
k¼1
Therefore, we have C − 1 independent mole fractions in each phase and P(C − 1)

independent mole fractions altogether, to which we should add two additional
intensive variables corresponding to pressure and temperature. Then the total number
of intensive variables in the system is equal to P(C − 1) + 2. For a particular
component, say component k, the equations involving P phases are given by
l1k ðp; TÞ ¼ l2k ðp; TÞ ¼ . . . ¼ lPk ðp; TÞ ð1:194Þ
and their number is equal to P − 1. For C system components, we have C(P − 1)

equations. Hence, the number of independent intensive variables in the system, that
is, the number of degrees of freedom f, is equal to the number of intensive variables
P(C − 1) + 2 minus the number of equations involving those variables C(P − 1),
f ¼ PðC 1Þ þ 2 CðP 1Þ ¼ C P þ 2 ð1:195Þ
This result is known as the Gibbs phase rule. For C = 1 and P = 1

(one-component system and one phase), f = 2. For C = 1 and P = 2
(one-component system and two phases in equilibrium), f = 1. Finally, for C = 1
54 1 Thermodynamics
and P = 3 (one-component system and three phases in equilibrium), f = 0 (zero

degrees of freedom corresponds to the triple point).
When the system components are involved in R chemical equilibria, there are
R additional equations expressing the affinities of the chemical reactions,
A 1 ¼ 0 A2 ¼ 0 . . . AR ¼ 0 ð1:196Þ
Each of these equations makes each chemical potential dependent on the

remaining ones, thus reducing by 1 the number of independent chemical potentials
[see the first equation of (1.170)]. For R chemical equilibria, the number of inde-
pendent components is reduced by R. Therefore, the Gibbs phase rule takes the
following more general form:
f ¼ ðC RÞ P þ 2 ¼ Ci P þ 2 ð1:197Þ
where the number of independent components Ci is obtained by subtracting the number

of chemical equilibria R from the number of components C in the system (E18).
1.15 Helmholtz Energy
At constant temperature and volume, the state function that evolves to a minimum
at equilibrium is the Helmholtz energy, defined by the following identity:
F U TS ð1:198Þ
Differentiating (1.198) and using (1.89) and (1.115), we obtain
dF ¼ dU TdS SdT ¼ dU Tde S Tdi S SdT ¼ pdV SdT Tdi S

ð1:199Þ
for a closed system. At constant temperature and volume (dT = 0, dV = 0),
dFT;V ¼ Tdi S 0 ð1:200Þ
where the inequality results from the second law for irreversible processes (diS >
0), and the equality corresponds to equilibrium (diS = 0). Therefore, in a closed
system at constant temperature and volume, the Helmholtz energy evolves through
an irreversible process to a minimum at equilibrium (dFT,V is negative). When the
irreversible process is a chemical reaction,
X
dFT;V ¼ Tdi S ¼ lk dnk ¼ Adn ð1:201Þ
k
1.15 Helmholtz Energy 55
where A is the chemical reaction affinity [see (1.132) and (1.136)]. Making use of
(1.63) in (1.201), we can write

@F @F
¼ lk ¼ A ð1:202Þ
@nk T;V;nj ðj6¼kÞ @n T;V
The second of the above equations shows that the partial derivative of the
Helmholtz energy with respect to the extent of the reaction, at constant temperature
and volume, is the negative of the chemical reaction affinity. At equilibrium, the
Helmholtz energy attains a minimum, and the chemical reaction affinity is zero.
1.16 Surface Tension
1.16.1 Liquid Droplet in Air
Consider now a system formed by a small liquid droplet in the air of a box, at
constant temperature and volume (Fig. 1.32). Since the gravitational attraction on
the droplet is negligible, the droplet takes an approximately spherical shape. It is
found that the diameter of the droplet is smaller than that of a sphere with the same
mass in the liquid bulk. At constant temperature and volume, the liquid droplet
reduces its diameter as the contracting force on its surface due to the surface
tension is balanced by an increase in the inside pressure (Fig. 1.32).
Let p′ and V′ be the pressure and air volume in the box, and p″ and V″ the
pressure and volume of the droplet. Then, using (1.87) and the second equation of
(1.115), we can write
dU ¼ TdS p0 dV 0 p00 dV 00 þ cdR ð1:203Þ
The last term of this expression is the work required to increase the interfacial
area by dR, and the proportionality coefficient c is called surface tension, whose
units are energy per area or force per length. Substitution of (1.203) in the first
equality of (1.199) leads to
Fig. 1.32 At constant temperature and volume, a suspended liquid droplet reduces its diameter in
an irreversible process that ends when the contracting force of the surface is balanced by the
increase in the inside pressure. In this irreversible process, the Helmholtz energy evolves to a
minimum
56 1 Thermodynamics
dF ¼ p0 dV 0 p00 dV 00 SdT þ cdR ð1:204Þ
where

@F
c¼ ð1:205Þ
@R T;V 0 ;V 00
that is, the surface tension is the Helmholtz energy per unit of interfacial area, at
constant T, V′, and V″. At 25 °C, water has a surface tension equal to 71.99 mJ m−2,
the surface tension of mercury is 485.48 mJ m−2, and most organic compounds
have, at temperatures below their boiling points, surface tensions between 15 and
50 mJ m−2 (values taken from Handbook of Chemistry and Physics, 2011).
The contraction of the liquid droplet in air is an irreversible process that makes
the system evolve to equilibrium in a time interval dt. Introducing constant tem-
perature (dT = 0) and constant overall volume (dV′ + dV″ = 0) in (1.204) and
making use of (1.200), we obtain
di S dV 00 dR
T ¼ ðp00 p0 Þ þc ð1:206Þ
dt dt dt
where V″ is the volume of the liquid droplet (V″ = 4pr3/3, dV″ = 4pr2dr) and R is its
surface area (R = 4pr2, dR = 8prdr). Substitution of these values in (1.206) leads to
di S
dr
T ¼ ðp00 p0 Þ4pr 2 þ c8pr ð1:207Þ
dt dt
During the evolution to equilibrium, the expression in square brackets tends to zero
and impels dr/dt to tend to zero, in the same way as a temperature gradient is the cause
for the corresponding irreversible heat flow or a concentration gradient is the cause for
an irreversible flow of matter. Therefore, at equilibrium, both the expression in square
brackets and dr/dt are zero. Hence, the excess pressure in the liquid droplet can be
expressed in terms of the surface tension and the droplet radius,
Dp ¼ p00 p0 ¼ 2c=r ð1:208Þ
1.16.2 Capillary Action
When the lower end of a vertical capillary tube is dipped into water, water rises in the
capillary tube as a manifestation of surface tension. Unlike the liquid droplet in air,
where there was a single interfacial surface, the liquid–air interface, the capillary tube
dipped into water has two interfaces involving water, namely the water–air and
water–glass interfaces. Water wets the glass, and so water rises in the capillary tube and
the meniscus is concave.
1.16 Surface Tension 57
The intersection between the liquid–air interface and the liquid–solid interface is
a circumference of a circle whose radius is equal to the internal radius of the
capillary tube. The angle at which the liquid is in contact with the capillary tube is
called the contact angle and is denoted by h (Fig. 1.33a). If the liquid wets the
glass, then this angle is smaller than p/2 (cosh > 0) and the height of liquid inside
the capillary tube is denoted by h (h > 0) (Fig. 1.33a). Since temperature is kept
constant and the total volume of liquid is also constant, the thermodynamic state
function that evolves to a minimum (equilibrium) under these conditions is the
Helmholtz energy [see the equality in (1.200)]. In the irreversible process leading to
equilibrium, the liquid inside the capillary tube gradually rises, and the pressure qgh
times the volume of the rising column of liquid inside the capillary tube (in units,
pressure volume = energy) is positive and increases. In turn, the surface tension
c cosh times the area of the interior wall of the capillary tube (in units, surface
tension area = energy) is negative and decreases as h increases. Hence, the
minimum of the Helmholtz energy results from a balance between these two
opposing quantities and is given by
Zh
DFðhÞ ¼ ðc cos hÞ2prh þ ðqhgÞpr 2 dh
0 ð1:209Þ
2 2
qgpr h
¼ ðc cos hÞ2prh þ
2
(a) (b)
Fig. 1.33 a For a liquid that wets the glass, one observes capillary rise, and the meniscus is
concave. The contact angle h is less than p/2 and cosh > 0. At equal heights throughout the liquid,
pressure is equal. b For a contact angle h greater than p/2, cosh < 0, and the liquid is depressed in
the capillary tube. This is the case of mercury, for example
58 1 Thermodynamics
Minimization of DF as a function of h leads to the value of h at equilibrium,

that is,
@DFðhÞ 2c cos h
¼ ðc cos hÞ2pr þ qgpr 2 h ¼ 0 ) h ¼ ð1:210Þ
@h qgr
If the liquid does not wet the solid surface and has a high surface tension like that
of mercury, then the contact angle is greater than p/2 (cosh < 0), the liquid is
depressed inside the capillary tube, and the height of liquid inside the capillary tube
is negative (h < 0) (Fig. 1.33b). Note that the terms of (1.209) keep the same signs
as when there is capillary rise.
1.17 Membrane Potential
Consider a closed system at constant temperature and volume, consisting of parts 1

and 2, with potassium chloride aqueous solutions in both of these parts and the
potassium chloride molality greater in part 1. Assume that the membrane that
separates parts 1 and 2 is permeable to potassium ions but prevents chloride ions
from passing through (Fig. 1.34). The transport of potassium ions through a porous
membrane from higher concentration (part 1) to lower concentration (part 2) leaves
an excess of chloride ions in part 1 and an excess of potassium ions in part 2. As a
consequence of this, the solutions of parts 1 and 2 become no longer electrically
neutral, and an electric potential is formed that opposes the continued migration of
potassium ions from part 1 to part 2 (Fig. 1.34). Note that the electrostatic inter-
actions between ions of a solution are very strong. Small variations in the charge
density lead to strong electrostatic forces that tend to restore electric neutrality. This
means that the bulk of solutions 1 and 2 are, to a good approximation, electrically
neutral, and the electrical potential difference due to potassium ion transport
through the membrane is mostly located in the membrane interfacial surfaces liq-
uid–membrane and membrane–liquid (Fig. 1.34). The electric potential due to the
excess of charge on both sides of the membrane eventually prevents the
Fig. 1.34 Irreversible transport of potassium ions through a porous membrane destroys electrical
neutrality in the liquid–membrane and membrane–liquid interfaces
1.17 Membrane Potential 59
continuation of charge transport through the membrane, leading the process of

transport of potassium ions to an end.
In order to interpret the irreversible process of potassium ion transport through
the porous membrane, we start by considering parts 1 and 2 with amounts of
potassium ion n1 and n2, chemical potentials l1 and l2, and electric potentials /1
and /2. Since we have a closed system, the total amount of potassium ions is fixed
and we can write
dn1 ¼ dn2 dn ð1:211Þ
where n represents the extent of potassium ions passing through the membrane. For
each dni, besides the chemical potential associated with mass transfer, there is an
additional contribution of electric energy (= charge electric potential) associated
with the charge transfer. Representing by z the amount of transported ion charge
and by F the Faraday constant, equal to the charge of 1 mole of elementary charges
(F = eNA = 96485.3365 C mol−1, see Appendix), we can write, at constant volume
and temperature (dV = 0, dT = 0),
dU ¼ TdS þ ðl1 þ zF/1 Þdn1 þ ðl2 þ zF/2 Þdn2 ð1:212Þ
Making use of the first equality of (1.115), Eq. (1.212) can be split into two
equations, one for the exchange of energy across system boundaries (the whole
system is kept at constant temperature),
dU ¼ Tde S ð1:213Þ
the other being due to irreversible changes inside the system,
0 ¼ Tdi S þ ðl1 þ zF/1 Þdn1 þ ðl2 þ zF/2 Þdn2 ð1:214Þ
Note that addition of (1.213) and (1.214) gives (1.212). Substitution of (1.211) in
(1.214) and rearrangement of the resulting equation leads to
Tdi S ¼ ðl1 þ zF/1 Þdn ðl2 þ zF/2 Þdn ð1:215Þ
This equation can be written as
Tdi S ¼ ð~
l1 l e
~2 Þdn ¼ Adn ð1:216Þ
~ is the electrochemical potential, defined by

where l
e
l i ¼ li þ zF/i ð1:217Þ
60 1 Thermodynamics
and Ã represents the electrochemical affinity, equal to the difference of electro-

chemical potentials of parts 1 and 2,
e¼e
A l1 e
l 2 ¼ ðl1 l2 Þ þ zFð/1 /2 Þ ð1:218Þ
(see Kondepudi, Further Reading). The transport of potassium ions through the
membrane is an irreversible process. Hence, diS > 0 and from (1.216) we can write
l e
di S ðe l 2 Þ dn Ae dn
¼ 1 ¼ ð1:219Þ
dt T dt T dt
where dt is the time interval that corresponds to the entropy production diS.
At equilibrium, diS/dt = 0, and both the electrochemical affinity Ã and the rate of
the irreversible process dn/dt are zero (dn/dt = 0 is a consequence of Ã being zero).
Then, at equilibrium, the electrochemical potentials are equal,
e
l1 e
l2 ¼ 0 ð1:220Þ
In turn, (1.218) and (1.220) lead to
1
/2 /1 ¼ ðl l2 Þ ð1:221Þ
zF 1
where the difference between electrical potentials is called the membrane potential
(a positive quantity). Substitution of (1.160) in (1.221) leads to

RT a1
/2 /1 ¼ ln ð1:222Þ
zF a2
A porous membrane that allows migration of a specific ion is called an

ion-selective membrane. The mechanism of selective migration can be realized
with a ligand embedded in the membrane, which selectively coordinates the specific
ion, thus promoting its transport through the membrane (Fig. 1.35).
Fig. 1.35 The selective irreversible transport of calcium ions through the membrane is a result of
their coordination by a ligand embedded in the membrane
1.18 Electrochemical Cell 61
1.18 Electrochemical Cell
Just try to submerge a copper wire into an aqueous solution of zinc sulfate (col-
orless). Nothing happens! However, if a zinc plate is immersed in a copper sulfate
solution (blue), the solution gradually becomes colorless, the zinc plate becomes
progressively eroded, and copper is deposited over it. This experiment can be
described by the following chemical equation,
Zn(s) + Cu2 þ ðaqÞ ! Zn2 þ ðaqÞ þ Cu(s) ð1:223Þ
which can be decomposed into the oxidation half-reaction
Zn(s) ! Zn2 þ ðaqÞ þ 2e ð1:224Þ
and the reduction half-reaction
Cu2 þ ðaqÞ þ 2e ! Cu(s) ð1:225Þ
The electrons produced in the oxidation half-reaction are consumed in the

reduction half-reaction. These half-reactions can be carried out in separate beakers
(zinc plate and zinc sulfate solution in one beaker, copper plate and copper sulfate
solution in another) with a salt bridge connecting the solutions and an external
circuit connecting the electrodes (from electric + Greek hodos “way”), as
schematically shown in Fig. 1.36. The electrons produced by the oxidation
half-reaction in the zinc plate go through the external circuit toward the copper
plate, where they are combined with copper ions from the solution and deposited
over the copper electrode (copper ion reduction half-reaction). The whole appara-
tus, an electrochemical cell that converts the Gibbs energy from the oxidation–
reduction reaction (1.223) into electric energy, is called a galvanic cell (from Luigi
Galvani, 1737–1798). The electric energy from a galvanic cell can be used to feed
an electric motor in an external circuit. An electrochemical cell that performs the
Fig. 1.36 Galvanic cell Zn(s) | Zn2+(aq) || Cu2+(aq) | Cu (schematic)

62 1 Thermodynamics
opposite transformation, that is, uses electric energy to drive an oxidation–reduction

reaction, is called an electrolytic cell. Galvanic and electrolytic cells are electro-
chemical cells. From now on, we consider only galvanic cells.
The electrode where the oxidation occurs (anode, from Greek ana + ho-
dos = “up” + “way”) becomes negative due to the electrons produced in the
oxidation half-reaction. In turn, the electrode where the reduction occurs (cathode,
from Greek kata + hodos = “down” + “way”) consumes electrons, thus becoming
positive. In the external circuit, electrons migrate from the anode to the cathode.
Thus, positive current flows to the anode as suggested by the origin of this word.
A salt bridge provides electrical contact between the oxidation and the reduction
half-cells of a galvanic cell. The ideal salt bridge electrolyte is chosen so that the
electric potential differences across both salt bridge–solution junctions cancel each
other, so that they do not affect the electrochemical potential difference. For a 1:1
salt bridge electrolyte, both ions should have approximately the same migratory
speed.
An electrochemical cell can be represented in a convenient and quick way by a
cell diagram, where one vertical bar | represents an interface, and two vertical bars
|| stand for the salt bridge junctions. The cathode half-cell is on the right of the cell
diagram. The following cell diagram,

Zn(s)Zn2 þ ðaqÞCu2 þ ðaqÞCu(s) ð1:226Þ
represents the galvanic cell of Fig. 1.36.
1.18.1 Nernst Equation
The irreversible processes of an ion-selective membrane and a galvanic cell create

an electric potential difference. Just as for an ion-selective membrane, the rate of
entropy production of a galvanic cell is given by
di S Ae dn
¼ ð1:227Þ
dt T dt
where Ã is the electrochemical affinity of the galvanic cell,
dn I
¼ ð1:228Þ
dt nF
I is the electric current intensity, that is, the charge transported per unit of time, and
n is the number of electrons involved in the corresponding oxidation–reduction
reaction. In dealing with a galvanic cell, we use a procedure similar to the one
followed for an ion-selective membrane and begin by evaluating the electro-

chemical affinity (see Kondepudi, Further Reading).
Let us consider a galvanic cell in which the following oxidation and reduction
half-reactions take place:
X ! Xn þ þ ne Yn þ þ ne ! Y ð1:229Þ
For each of these half-reactions, the electrochemical affinity is the difference

between the electrochemical potentials of the “reactants” (chemical species in the
first member) and the electrochemical potentials of the “products” (chemical species
in the second member). The chemical species X, Xn+, Yn+, and Y are assigned to
the corresponding chemical potentials. In turn, in the external circuit of the galvanic
cell, the chemical potential of the electron le is a function of the electronic density
and temperature, physical quantities we assume to be constant. Hence, le is con-
stant, and the electrochemical potential for each electron is given by
e
l e ¼ le F/ ð1:230Þ
where / is the electric potential, and the minus sign is due to the negative charge of
the electron. The electrochemical affinities of the anode and cathode half-reactions
are given by
X ! Xn þ þ ne e a ¼ lX ½lXn þ þ nðle F/a Þ

A
ð1:231Þ
Yn þ þ ne ! Y e c ¼ ½lYn þ þ nðle F/c Þ lY
A
where the superscripts a and c stand for anode and cathode, respectively. Since the
oxidation–reduction reaction is the sum of the oxidation and reduction
half-reactions, the electrochemical affinity of a galvanic cell Ã is also the sum of the
electrochemical affinities of the anode and cathode reactions,
X þ Yn þ ! Xn þ þ Y
ð1:232Þ
e¼A
A e c ¼ ðlX þ lYn þ Þ ðlXn þ þ lY Þ nFð/c /a Þ
ea þ A
This equation should be compared with (1.218), which refers to the electro-
chemical affinity of an ion-selective membrane. In the latter, the irreversible process
is a physical process consisting in the transfer of an ion through the semipermeable
membrane, whereas in the electrochemical cell, the irreversible process is an oxi-
dation–reduction reaction of a galvanic cell.
When Ã = 0, then dn/dt = 0 [see (1.227)], and the current intensity is zero [see
(1.228)]. Hence,
1
De ¼ /c /a ¼ ½ðl þ lYn þ Þ ðlXn þ þ lY Þ ð1:233Þ
nF X
64 1 Thermodynamics
where De = /c − /a, the potential difference between cathode and anode at

zero-current intensity, is called the zero-current cell potential [compare (1.233)
with (1.221) for the membrane potential]. Note that every electrochemical cell has
an internal electric resistance Rint. If the electrical current across the cell is I, then
the potential IRint opposes the electrical potential of the cell. Hence, the zero-current
situation ensures that its internal resistance does not affect the cell potential, since
the product IRint becomes zero when I = 0.
As is well known, chemical potentials depend on the nature and activities of
chemical species according to the following expression:
lk ¼ l0k þ RT ln ak ð1:234Þ
Substitution of this expression in (1.233), with k taking the values X, Yn+, Xn+,
and Y, leads to the Nernst equation

RT a n þ aY
De ¼ De0 ln X ð1:235Þ
nF aX aY n þ
where
1
0
De0 ¼ lX þ l0Yn þ l0Xn þ þ l0Y ð1:236Þ
nF
is the standard cell potential. When chemical equilibrium is reached, the reaction
quotient, that is, the expression within parentheses in (1.235), becomes equal to the
equilibrium constant, and De = 0. Then we can write
RT
De0 ¼ ln Ke ð1:237Þ
nF
At 298.15 K,
log Ke 16:9nDe0 ð1:238Þ
For n = 1 and De0 = 0.5 V, log Ke 8.5, that is, Ke 108.5 3 108.
If the Nernst equation (1.235) is rewritten so that the cathode and anode
potentials are separate terms, one obtains

RT aY RT aX
De ¼ e0;c ln e0;a ln ð1:239Þ
nF aY n þ nF aX n þ
where the expressions in parentheses refer to reduction half-reactions for the

cathode and anode. Note that both arguments of the logarithm functions correspond
to half-reactions written as reductions. Therefore, we can generalize and write
RT aX
Xn þ þ ne ! X e = e0 ln ð1:240Þ
nF aXn þ
where e is the electrode potential and e0 is the standard electrode potential,

which is the electrode potential for unit activities of Xn+ and X (E19, E20).
Notes
§1. Pressure is defined as force per unit area. The SI unit of pressure is the pascal,
1 Pa = 1 N m−2. A column of liquid of uniform density q and height h exerts a
pressure qhg, where g represents the acceleration due to gravity; 1 mmHg is the
pressure exerted by a column of mercury of height 1 mm when g = 9.80665 m s−2 and
q(Hg) = 13.5951 g cm−3. 1 atm = 760 mmHg = 101325 Pa, 1 bar = 100,000 Pa.
Temperature is usually expressed in kelvin (K), degrees Celsius (°C), or degrees
Fahrenheit (°F). The Celsius and Fahrenheit scales are empirical, and use the
melting and boiling temperatures of water at 1 atm of external pressure as refer-
ences. For the Celsius scale, 0 °C corresponds to the normal melting point of water,
and 100 °C corresponds to the normal boiling point of water. In the Fahrenheit
scale, 32 °F corresponds to the normal melting point of water, and 212 °F corre-
sponds to the normal boiling point of water. The Kelvin scale is defined using the
second law of thermodynamics, namely, the efficiency of a reversible heat engine or
Carnot engine. The triple point of water is exactly 273.16 K (0.01 °C) and
611.73 Pa. The Kelvin and Celsius scales are related by T/K = T/°C + 273.15.
§2. Let v1 and v2 be the velocities of the test molecule O2 and of one target molecule
N2. The relative velocity is equal to v1 − v2. The absolute value of this vector is
given by
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðv1 v2 Þ ðv1 v2 Þ ¼ v21 þ v22 2v1 v2
Taking ensemble averages, we can write

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

hðv1 v2 Þ ðv1 v2 Þi ¼ v21 þ v22 2hv1 v2 i
66 1 Thermodynamics
The molecular motions are random, and so the velocities of molecules 1 and 2
are uncorrelated, that is,
h v1 v2 i ¼ 0
In addition, the space of molecular velocities is isotropic,

2 2 2
v1 ¼ v2 ¼ v
Hence,
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffipffiffiffiffiffiffiffiffi pffiffiffi
hðv1 v2 Þ ðv1 v2 Þi ¼ v21 þ v22 2hv1 v2 i ¼ 2 hv2 i ¼ 2vrms
§3. In an adiabatic transformation (from Greek, adiábatos, “impassable”), the

system is thermally isolated from the surroundings, that is, there is no heat
exchanged between the system and its surroundings. The ideal gas equation for an
adiabatic transformation is
TV c1 ¼ constant
Substitution of T by pV/(nR) leads to the equation
pV c ¼ constant
where c is defined by
c ¼ Cmp =CmV
and Cmp and CmV are the isobaric and isochoric molar heat capacities. For a
monatomic ideal gas, c = 5/3.
§4. In the following derivation, we use the definition of mechanical work given by
(1.84) and the expression for the first law that results from integration of both
members of (1.187), DU = Q + W. We assume 1 mole of ideal gas, so we deal with
molar volumes. A Carnot cycle consists of an ideal gas undergoing the following
four steps:
Notes 67
Step A!B (Isothermal Expansion)
In this step, the gas absorbs QAB from the heat source at temperature hh and uses
this heat to do the work WAB on the surroundings. Note that the energy of an ideal
gas depends only on its temperature [see (1.19)]. Hence, for an isothermal trans-
formation, the energy change is zero, and so from the first law, QAB + WAB = 0:
ZB ZB
1
WAB ¼ pdV ¼ Rhh dV ¼ Rhh lnðVB =VA Þ
V
A A
For an expansion, VB > VA, and so WAB is a negative quantity (work done by the
gas on the surroundings).
Step B ! C (Adiabatic Expansion)
Since this step is adiabatic, QBC = 0 and
pB VBc ¼ pC VCc
Using this equality, the work done by the system is given by
ZC ZC
pV c pC VCc VC1c pB VBc VB1c pC V C pB V B
WBC ¼ pdV ¼ dV ¼ þ ¼ þ
Vc 1c 1c 1c 1c
B B
Rðhc hh Þ Rðhh hc Þ
¼ ¼
1c c1
where WBC is a negative quantity (this step corresponds to an expansion,

hB = hh > hc = hC and c > 1).
Step C ! D (Isothermal Compression)
In this step, the work WCD done on the system during the gas compression is
given by
ZD ZD
1
WCD ¼ pdV ¼ Rhc dV ¼ Rhc lnðVD =VC Þ
V
C C
68 1 Thermodynamics
Being a compression, VD < VC, and WCD is a positive quantity. For an isothermal
transformation in an ideal gas, the energy change is zero, and so QCD + WCD = 0
(QCD corresponds to released heat, and is thus a negative quantity).
Step D ! A (Adiabatic Compression)
Since this step is adiabatic, QDA = 0 and
pD VDc ¼ pA VAc
Using this equality, the work done on the system is given by
ZA ZA
pV c pA VAc VA1c pD VDc VD1c
WDA ¼ pdV ¼ dV ¼ þ
Vc 1c 1c
D D
p A VA pD VD Rðhh hc Þ
¼ þ ¼
1c 1c c1
From the results obtained for WBC and WDA, we conclude that WBC + WDA = 0.
The work done on the Carnot reversible cycle is equal to the sum of the works
done in the four steps of the cycle. Using the above results (QAB + WAB = 0,
QBC = 0, QCD + WCD = 0, QDA = 0, and WBC + WDA = 0), we can write
WAB þ WBC þ WCD þ WDA ¼ WAB þ WCD ¼ QAB QCD
Hence, the efficiency for the Carnot cycle is given by
Qc W QAB jQCD j jQCD j WCD

gCarnot ¼ 1 ¼ ¼ ¼1 ¼ 1þ
Qh Qh QAB QAB WAB
Rhc lnðVD =VC Þ Rhc lnðVC =VD Þ hc
¼ 1þ ¼1 ¼1
Rhh lnðVB =VA Þ Rhh lnðVB =VA Þ hh
where W, Qh, and Qc represent absolute values. The equations for the isothermal
transformations (pAVA = pBVB and pCVC = pDVD) and adiabatic transformations
(pBVcB = pCVcC and pDVcD = pAVcA) can be simultaneously solved, provided VB/
VA = VC/VD. This equality justifies the last equation above. Therefore, the Carnot
engine efficiency is given by
Qc hc
gCarnot ¼ 1 ¼1
Qh hh
Mathematica Codes 69
Mathematica Codes
M1. Maxwell Distribution of Molecular Speeds
This Mathematica code evaluates the normalization factor of f(v) in (1.27) and
plots this probability function for a gas of oxygen molecules at 100 and 300 K. The
normalization factor is calculated as the inverse of the definite integral of f(v) in
the range {0,∞} and requires the use of the Mathematica command
Integrate[f(v),{v,0,Infinity},Assumptions!{T>0,m>0,kB>0}]
where v is the integration variable, whose range is between zero and infinity, because v
represents a molecular speed, that is, an absolute value of the velocity vector, and the
option Assumptions indicates that the symbolic quantities T,m, and kB are positive.
They are assigned specific values in lines of code 3, 4, and 5, and line of code 6 outputs
the value of the norm for the oxygen molecule, at an unspecified temperature T.
The Mathematica command
Plot[f[v]/.T !{100,300},{v,0,1200}]
generates plots of functions f(v) in the range {0,1200}, where T is replaced by
100 and 300 (/. stands for the ReplaceAll Mathematica command). This is the
point where the variable T is assigned specific values, namely, 100 and 300. The
physical quantities m, v, and T, and the constants avogadro and kB are in SI
70 1 Thermodynamics
units. If you wish to run the above code for different molecular masses or different
temperatures and do not want to transfer universal constants and mass values from
one run to the next, you should use Clear[avogadro,kB,m] as the first line of
the code. This Mathematica function clears values and definitions of the specified
symbols, making it possible to evaluate the norm on the second line of code using
symbolic constants and variables.
Suggestion: Use the Graphics primitive Text to add the temperatures
“100 K” and “300 K” next to the corresponding curves in the above plot. If nec-
essary, consult Wolfram Documentation in the Help menu of Mathematica with the
words Graphics and Text.
M2. Critical Point for Van der Waals Fluid
This code shows how to use Mathematica to determine the critical values of
volume, temperature, and pressure for a Van der Waals fluid. The code solves the
system of Eqs. (1.49) and confirms the critical values presented in (1.50). The first
line of code defines the p[V,T] function for the Van der Waals equation, and the
second line of code solves the system of Eqs. (1.49) to determine the critical values
of volume and temperature. The expression D[p[V,T],V] gives the partial
derivative of p with respect to V at constant T, and D[p[V,T],{V,2}] gives the
second-order partial derivative of p with respect to V at constant T. The solutions of
the system of equations are assigned to sol, which is a list of a list of length 2 (see
the first line of output; in Mathematica, a list is denoted by {…}). The second line
of output shows that sol[[1]] stands for the first part of sol, that is, a list of the
critical volume and critical temperature. The Print command of the fourth line of
code is used to print the string of characters inside quotation marks (“Critical
Volume”), and the next line of code uses ReplaceAll (/.) to replace V by the
first part of sol, that is, by sol[[1]]. The last line of code solves for the values
of a, b, and R that lead to critical values of volume, temperature, and pressure equal
to 1, thus to the Van der Waals equation in reduced variables.
M3. 3D Plot of the Van der Waals Equation in Reduced

Variables
This Mathematica code plots the Van der Waals equation in reduced variables
[see (1.53)] using the Mathematica command
Plot3D[f[V,T],{V,0.34,3},{T,0.34,3}]
where f[V,T] corresponds to the reduced pressure as a function of the reduced
volume V and reduced temperature T, in the above indicated ranges. Note that a
function of two independent variables is represented by a surface. The Van der
Waals isotherms of Fig. 1.11 are cross sections of this surface by planes that
correspond to defined temperature values. The option Lighting specifies the
simulated lighting used to color the plotted surface. Lighting!“Neutral”
simulates white light.
Suggestion: Run the above Mathematica code line leaving out the options
PlotStyle and Lighting.
72 1 Thermodynamics
M4. Absolute and Reduced Temperatures for the Van der

Waals Carbon Dioxide
This Mathematica code evaluates the critical temperature for carbon dioxide (see the
values of a and b in the Table for Van der Waals constants) and calculates reduced
temperatures corresponding to absolute temperatures from 274 to 344 K in steps of 10.
M5. Isothermal and Adiabatic Transformations on an Ideal Gas

Surface
This Mathematica code plots the ideal gas surface with meshes for two
isothermal and four adiabatic transformations. The option for Plot3D that speci-
fies the number of isothermal and adiabatic transformations is
Mesh!{2,4}
and the option for the corresponding functions is
MeshFunctions!{#1&,#1*#2^(2/3)&}
where #1& is the pure function for an isothermal transformation (#1 is the first
argument of f[T,V]) and #1*#2^(2/3) is the pure function for an adiabatic
transformation (#2 is the second argument of f[T,V]). The adiabatic transfor-
mation is given by TVc −1 = constant, where c is equal to 5/3 for a monatomic ideal
gas (see §3).
Suggestion: Change the above Mathematica code so that it runs for 35

isothermal transformations and 16 adiabatic transformations. Choose a View-
Point from Top. Compare the obtained plot with that of Fig. 1.25.
M6. Efficiency of the Carnot Heat Engine
This Mathematica code calculates the efficiencies of Carnot heat engines [see
(1.107)] when the temperature of the hot reservoir is fixed at 1001 K and the
temperatures of the cold reservoir are 1, 10, 100, and 1000 K. The efficiencies of
Carnot engines are presented as percentages with three significant digits using the
Mathematica command SetPrecision. This code illustrates the use of Table to
obtain two lists of length 4: the list aa of four cold reservoir temperatures and the list
bb composed of the corresponding Carnot efficiencies. These two lists are interleaved
by the use of the Mathematica command Riffle. The command Partition par-
titions the list cc into nonoverlapping lists of length 2 composed of the temperature of
the cold reservoir Tc and the maximum Carnot’s engine efficiency η. Note that the
above code can be compressed into the following single line of code:
Partition[Riffle[Table[10î,{i,0,3}],
SetPrecision[Table[(1.-10î/1001)*100,{i,0,3}],3]],2]
The following code uses ListPlot to plot the list of points (Tc, η), where Tc
varies from 1 to 1000 in steps of 1, and η is indicated as a percentage. This plot
shows a linear dependence of η versus Tc, with η decreasing with the increase of Tc.
74 1 Thermodynamics
M7. Gibbs Energy and Affinity of a Chemical Reaction
Consider a hypothetical chemical reaction R(g) ! P(g), where R and P are ideal
gases and the equilibrium constant is assumed to be 3.0. Taking the Gibbs energy
and the affinity as functions of the extent of the chemical reaction n [see (1.131) and
(1.132)] whose range is [0,1], the partial derivative of the Gibbs energy with respect
to n, at constant temperature and pressure, is equal to the negative of affinity of the
chemical reaction [see second equality of (1.170)].
The following Mathematica code shows how the Gibbs energy of the above
chemical reaction and the corresponding affinity evolve to equilibrium. In order to
obtain the Gibbs energy as a function of n, we write
X
G ¼ G0R þ DG
where G0R is the standard Gibbs energy of the reactant R (in the code, we assume
G0R ¼ 0) and DG is obtained by substituting (1.188) into (1.185), that is, using

v
DG ¼ RT ln
Keq
where v represents the reaction quotient
n
v¼
1n
and Keq is the equilibrium constant. In the code,
i
n¼
1000
where i varies in the range [1,1000]. Therefore, each step of n is equal to 0.001, and
the command ListPlot[data] plots 1000 points. The Mathematica command
Fit finds a least squares fit to data as a linear combination of functions in the list
{1,x,x^2,x^3,x^4}, where x is real and corresponds to i. The value of x at
equilibrium is obtained by the code lines
sol=Solve[bestfit’[x]==0,x][[3]];
eq=x/.sol;
Solving bestfit’[x]==0 leads to three roots, the first two being complex
numbers, the third being real and equal to 750.181. The third root, denoted by sol,
is selected using [[3]] immediately after the Solve command. The value of x at
equilibrium is assigned to eq. Plots of the Gibbs energy G and affinity A are
restricted to the neighborhood of equilibrium.
Suggestion: Write a Mathematica code to obtain the Gibbs energy and the
affinity as functions of the extent of the chemical reaction A + B ! 2C and assume
Keq = 100.
Glossary
Activity Applies to chemical species in nonideal solutions and

corrects mole fractions for departure from ideal
behavior due to intermolecular interactions; it is
obtained as the product of an activity coefficient by the
mole fraction; see (1.161). Ideal behavior corresponds
to the activity coefficient being equal to 1.
76 1 Thermodynamics
Adiabatic transformation A transformation in which the system is thermally

isolated from the surroundings and so no heat is
exchanged between the system and its
surroundings; see §3.
Affinity of a chemical The stoichiometric sum of chemical potentials of
reaction the reactants minus the stoichiometric sum of
chemical potentials of the products; see (1.136).
Affinity and Gibbs energy are related through the
second equality of (1.170). At equilibrium, the
Gibbs energy attains a minimum and the
reaction affinity is zero; see Fig. 1.30 and
Mathematica code M7.
Anode The electrode where oxidation occurs. In a galvanic
cell, the Gibbs energy from an oxidation–reduction
reaction is converted into electric energy. The
oxidation at the anode produces electrons that give
rise to a negative charged electrode and releases
positive ions into solution, thus eroding the anode.
In an electrolytic cell, the positive terminal of the
applied voltage is directly connected to the anode,
and so the anode becomes positive, causing
oxidation to occur. The anode is where oxidation
occurs, whether the cell is galvanic or electrolytic.
Carnot heat engine Consists of an ideal gas that operates over a reversible
cycle formed by isothermal and adiabatic expansions
followed by isothermal and adiabatic compressions;
see Fig. 1.23 and Mathematica code M5. According
to Carnot’s theorem, the efficiency of the Carnot heat
engine depends only on the temperatures of the hot
and cold reservoirs; see (1.102).
Cathode The electrode where reduction occurs. In a galvanic
cell, reduction at the cathode combines electrons
from the electrode with positive ions from solution
giving rise to neutral atoms that are deposited on the
cathode and make it grow. In turn, depletion of
electrons from the cathode makes it positive (the
cathode has a higher electric potential than the
anode). In an electrolytic cell, the negative terminal
of the applied voltage is directly connected to the
cathode, and so the cathode becomes negative, thus
originating reduction. The cathode is where
reduction occurs, whether the cell is galvanic or
electrolytic.
Glossary 77
Chemical potential The partial derivative of energy U with respect to

the amount of substance nk, at constant entropy S,
constant volume V, and constant amounts of other
system components j 6¼ k; see (1.139) and (1.140).
Alternatively, the chemical potential is the partial
derivative of the Gibbs energy G with respect to the
amount of substance nk, at constant temperature T,
constant pressure p, and constant amounts of other
system components j 6¼ k; see (1.169) and the first
equation of (1.170). The chemical potential has
units of energy per mole.
Closed system Can exchange energy, not matter, with its
surroundings; see Fig. 1.15.
Electrochemical affinity For an ion-selective membrane is the difference of
electrochemical potentials across the membrane; see
(1.218) and Fig. 1.34. For an oxidation–reduction
reaction, the electrochemical affinity is the
stoichiometric sum of electrochemical potentials of
the oxidation half-reaction plus the stoichiometric
sum of electrochemical potentials of the reduction
half-reaction; see (1.231) and (1.232). At
equilibrium, the electrochemical affinity is zero: for
an ion-selective membrane, this situation gives the
membrane potential, see (1.221); for a galvanic cell,
one obtains the zero-current cell potential;
see (1.233).
Electrochemical cell A general expression that comprises galvanic and
electrolytic cells.
Electrochemical potential The chemical potential (units of energy per mole)
associated with mass transfer, plus electric energy
per mole (= charge per mole x electric potential)
associated with charge transfer; see (1.217).
Electrolytic cell Uses electric energy to drive an oxidation–reduction
reaction.
Enthalpy A state function defined by H U + pV; see (1.92)
and (1.93).
Entropy A state function usually represented by S that is
central to the second law of thermodynamics. An
arbitrary reversible cycle carried out in a given
direction can be covered to the required precision
by a mesh of infinitesimal Carnot cycles carried out
78 1 Thermodynamics
in the same direction. This enables one to conclude

that the integral of dQrev/T over the arbitrary
reversible cycle is zero, that is, dQrev/T is the
differential of a state function, called entropy; see
(1.110) and (1.111). According to Ilya Prigogine
(1917–2003, Nobel Prize in chemistry in 1977), a
distinction should be made between two terms
included in the entropy change dS. The first, deS, is
the transfer of entropy across system boundaries
due to energy and matter exchanges with the
surroundings. This transfer of entropy can be
performed reversibly, the corresponding entropy
change deS being either positive or negative (no
thermodynamic law can be formulated on this
term). In turn, the second term, diS, is the entropy
produced within the system by irreversible
processes that drive the system toward equilibrium
and is always positive, diS > 0, a statement that
corresponds to the second law of thermodynamics.
Extent of a chemical State variable whose differential is equal to the
reaction differential of the amount of a reactant or a product
of a chemical reaction divided by the corresponding
stoichiometric number; see (1.131) and (1.132).
Stoichiometric numbers coincide with the
stoichiometric coefficients for the products of a
chemical reaction and are equal to the
stoichiometric coefficients multiplied by −1 for the
reactants. The definition of the extent of a chemical
reaction results from the fact that infinitesimal
arbitrary changes in the amounts of reactants and
products are related to one another by the chemical
reaction stoichiometry.
Galvanic cell Converts Gibbs energy from an oxidation–reduction
reaction into electric energy.
Gibbs–Duhem equation Shows that the intensive variables temperature,
pressure, and chemical potential are not
independent variables; see (1.145). For one system
component, see (1.146).
Gibbs energy State function defined by G U + pV − TS; see
(1.166) and (1.168).
Glossary 79
Gibbs phase rule Equates the number of degrees of freedom of a

system at equilibrium with the number of
independent components, minus the number of
phases, plus 2; see (1.197).
Helmholtz energy State function defined by F U − TS; see (1.198)
and (1.200).
Ideal gas Physical model of a gas whose equation of state is
given by (1.5), (1.7), or (1.8). The particles of an
ideal gas are assumed to be point masses (the total
volume of the particles of an ideal gas is zero) and
noninteracting.
Irreversible process Occurs in a specific direction, not in its reverse, and
evolves to equilibrium. In the time interval
dt during an irreversible process, diS > 0 and the
rate of entropy production diS/dt is positive; see
(1.115) and (1.116).
Isobaric heat capacity Partial derivative of enthalpy H with respect to
temperature T at constant pressure; see (1.95). The
isobaric heat capacity has units of energy per
kelvin.
Isochoric heat capacity Partial derivative of energy U with respect to
temperature T at constant volume; see (1.94). The
isochoric heat capacity has units of energy per
kelvin.
Isolated system Does not exchange matter or energy with its
Kinetic model of gases Describes a gas assuming that (i) the average
distance between molecules is far greater than the
molecules’ dimensions, (ii) the molecules are in
constant random motion, (iii) each molecule is seen
as a sphere with elastic collisions with other
molecules and the container walls, (iv) the
molecules do not interact with each other except
during collisions and (v) the gas molecules obey
Newton’s equations of motion. The distribution of
molecular speeds is given by (1.27); see Fig. 1.6 for
an illustration with oxygen molecules at 100 and
300 K.
Mean free path Average distance between gas molecule collisions.
The mean free path is inversely proportional to the
square of the cross section for collisions and to the
gas number density; see (1.39).
80 1 Thermodynamics
Mechanical work In the time interval dt is defined by dW = −pdV; see

(1.83) and Fig. 1.18.
Membrane potential Electric potential difference developed across a
porous membrane, due to the transport of certain
ions, from higher to lower electrochemical
potentials; see (1.219).
Nernst equation Gives the zero-current cell potential as a function of
the reaction quotient of an oxidation–reduction
reaction expressed in terms of activities; see
(1.235).
Number of degrees The number of independent intensive variables of a
of freedom system at equilibrium.
Number of independent For a system at equilibrium is the number of its
components components minus the number of distinct chemical
equilibria involving them.
Open system Can exchange matter and energy with its
Phases Distinct physical states of matter.
Surface tension Helmholtz energy per unit of interfacial area; see
(1.205).
Van der Waals critical Temperature at which the maximum and minimum
temperature of van der Waals isotherms coincide, giving rise to
the inflection called the critical point and to the
critical isotherm; see (1.49), and Mathematica code
M2.
Van der Waals equation Equation of state for a fluid (gas and liquid)
composed of particles with nonzero volumes and
pairwise attractive interactions; see (1.48). This
equation results from modification and
improvement of the equation of state of an ideal
gas. A table with Van der Waals constants of
selected gases appears next to (1.48), and the Van
der Waals isotherms for carbon dioxide can be seen
in Fig. 1.9.
Van der Waals reduced Molar volume, temperature, and pressure of a Van
variables der Waals fluid in units defined by the
corresponding critical values; see (1.52). The Van
der Waals equation in reduced variables is given by
(1.53); see Mathematica codes M3 and M4.
Exercises 81
Exercises
E1. Use Mathematica to confirm the normalization factor for (1.23).

E2. Use Mathematica to confirm (1.29), (1.30), and (1.31) and evaluate the most
probable speed, the mean molecular speed, and the root mean square speed, for O2
at 298.15 K.
E3. Obtain estimates for the mean free path, the average distance between mole-
cules, and the time for one collision of O2 at 273.15 K and 1 atm of pressure.
Assume that the radius of the collision cross section is equal to 0.3 nm.
E4. Use Mathematica to determine the critical values of volume, temperature, and
pressure for the van der Waals equation.
E5. Use Mathematica to plot the isotherms of the van der Waals equation for
carbon dioxide with interactive manipulation of the temperature value.
E6. Use Mathematica to generate a three-dimensional plot of the Van der Waals
equation in reduced variables.
E7. Classify 4x2y3 dx + 3x3y2 dy as an exact or inexact differential.
E8. Show that p dV + V dp, with p and V being the pressure and volume variables,
is an exact differential.
E9. Use the first law to arrive at

@U @U
dQ ¼ dT þ p þ dV
@T V @V T
and derive the following equality:

@U @V
Cp CV ¼ p þ
@V T @T p
E10. The heat capacity of iron is 25.09 J mol−1 K−1. Calculate the entropy change
for 4 mol of iron heated from 273 K to 373 K, assuming that the heat capacity for
iron is constant in this temperature interval.
E11. Evaluate the standard reaction enthalpy for C3H8(g) + 5O2(g) ! 3CO2(g) +
4H2O(l), at 298.15 K ½DHfo ½C3 H8 ðgÞ ¼ 104:7 kJ mol1 ; DHof ½CO2 ðgÞ ¼
393:5 kJ mol1 ; DHof ½H2 OðlÞ ¼ 285:8 kJ mol1 .
E12. Rudolf Clausius expressed the second law in the following way: No process is
possible whose sole result is the transfer of heat from a body of lower temperature
to a body of higher temperature. In turn, Lord Kelvin used the following statement:
No process is possible in which the sole result is the absorption of heat from a
reservoir and its complete conversion into work. Show that Clausius’s and Kelvin’s
statements correspond to equivalent heat engines.
82 1 Thermodynamics
E13. Consider an empirical, ideal gas thermometer defined by the volume as a

function of the gas’s Celsius temperature t at constant pressure, V = V0 + V0 a t.
How is V0 defined? Determine a.
E14. Draw a Carnot cycle in the temperature–entropy space.
E15. Use Mathematica to determine the equilibrium constant for 2NO2(g)
N2O4(g) at 298.15 K {DGof ½NO2 ðgÞ ¼ 51:3 kJ mol1 ; DGof ½N2 O4 ðgÞ ¼ 99:8 kJ mol1 }
and the NO2 and N2O4 mole fractions at equilibrium.
E16. Consider a one-component system with two phases in equilibrium. Apply the
Gibbs–Duhem equation to each of the phases in equilibrium and determine the
derivative of pressure with respect to temperature to obtain Clapeyron’s equation.
Apply this equation to the liquid–vapor transition and assume ideal behavior for
the vapor phase to obtain the Clausius–Clapeyron equation.
E17. Apply the Gibbs phase rule to the equilibrium in a system containing iron, iron
(III) oxide, carbon, oxygen, and carbon dioxide. Determine the number of degrees
of freedom in the system and find the equations that involve the system’s ther-
modynamic variables.
E18. Consider an irreversible process in an isolated system. Show that entropy
tends to a maximum at equilibrium.
E19. Show that all the electrode potentials are defined for one electron involved in
the corresponding reduction half-reaction.
E20. Evaluate the equilibrium constant for 2Ag+ + Cu ! 2Ag + Cu2+ at 298.15 K.
References
Angrist SW (1968) Perpetual motion machines. Sci Am
Further Reading
Cheung C-K, Keough GE, Gross RH, Landraitis C (2009) Getting started with mathematica, third
edn. Wiley, New Jersey
Kondepudi D (2008) Introduction to modern thermodynamics. Wiley, New Jersey
Kondepudi D, Prigogine I (1998) Modern thermodynamics: from heat engines to dissipative
structures. Wiley, New Jersey
Wolfram S (2015) An elementary introduction to the Wolfram Language. Wolfram Media, Inc.
Available online
Chemical Kinetics
2
Abstract
Chemical kinetics is the branch of chemistry that measures rates of chemical
reactions, studies the factors that influence them, designs and prepares new
catalysts, and interprets the results at the molecular level. The independent
variable of chemical kinetics, from the chemical reaction starting moment when
the reactants are mixed to its final moment when equilibrium is reached, is time,
a variable introduced by the second law of thermodynamics for irreversible
processes. The first study of the rate of a chemical reaction is credited to Ludwig
Wilhelmy in 1850 for the decomposition of sucrose (table sugar) into glucose
and fructose, in acid medium. Wilhelmy found that the rate of this chemical
reaction is proportional to the existing amount of sucrose at each instant in the
course of the chemical reaction. This chapter begins with sections on the rate of a
chemical reaction, the experimental rate equation, and the effect of temperature
change. We then consider elementary reactions, complex reactions, and
extremely fast reactions. Most chemical reactions function like one-way streets:
the concentrations of reactants decrease, those of reaction intermediates increase
at first and decrease later, and the concentrations of products increase. However,
for a few reactions far from equilibrium, the concentrations of some intermediate
species oscillate, increasing and decreasing repeatedly. These reactions are
illustrated with the Brusselator, a model chemical oscillator developed in the
Brussels thermodynamic school founded by Prigogine. At the end of this
chapter, the student can find two notes on matrix diagonalization and systems of
first-order linear differential equations, useful for understanding the mathemat-
ical treatment given to the Brusselator, two Mathematica codes (First-Order
Chemical Reaction, Brusselator) with references to expressions in the main text,
detailed explanations for new commands and suggestions for the student to
follow, a glossary that explains important scientific terms, and a list of exercises,
whose complete answers can be found at the end of the book.

DOI 10.1007/978-3-319-41093-7_2
84 2 Chemical Kinetics
2.1 Rate of a Chemical Reaction
Add bleach (typically a solution of sodium hypochlorite) to a food dye in aqueous

solution. This simple experiment can illustrate the change of concentration of the
dye as a function of time. Let us call the dye A and assume that its initial con-
centration is [A]0 = 5.0 10−5 M. The continuous addition of bleach causes
progressive discoloring. The spectral absorbance of such a diluted solution is
proportional to the concentration of the absorbing species (Beer–Lambert law). This
empirical, relationship makes it possible to determine the dye concentration from
absorbance measurements taken at an absorption maximum wavelength (kmax).
Since the initial concentration of the dye is very low, the bleach exists in large
excess as compared with the dye. Hence, the bleach concentration does not
appreciably change during dye discoloring and the chemical reaction can be rep-
resented in the following way
A!B ð2:1Þ
where B represents the colorless form of the dye (the bleach is omitted because its
concentration is approximately constant during the experiment). The following
table gives the concentration of A as a function of time at 1-min intervals.
t/min 0 1 2 3 4 5 6 7 8 9 10
[A]/10−5 M 5.00 3.57 2.49 1.82 1.26 0.89 0.61 0.45 0.30 0.23 0.16
The graph of [A] as a function of t is called a kinetic reaction profile for the
chemical reaction (2.1) and is presented in Fig. 2.1 (see Mathematica code M1).
Once the fitting curve is known (an exponential, as we will soon find out), it is
possible to obtain the derivative at each instant in the considered time range.
Since A is the reactant, [A] is a decreasing function of time, its first derivative is
negative, and the reaction rate is given by −d[A]/dt. The reaction rate defined
using B is given by d[B]/dt, that is, d[B]/dt = −d[A]/dt (E1).
Fig. 2.1 Typical kinetic

reaction profile for
discoloration of a dye.
Mathematica
2.1 Rate of a Chemical Reaction 85
Considering now the chemical reaction
A þ B ! 2C ð2:2Þ
[for example, H2(g) + I2(g) ! 2HI(g)], we can write
dnC ¼ 2ðdnA Þ ¼ 2ðdnB Þ ð2:3Þ
where dnA, dnB, and dnC represent changes in the molar quantities of A, B, and C.
Both dnA and dnB are negative, because A and B are reactants, dnC is positive (C is
the single product) and is twice −dnA or −dnB [see (2.2)]. From (2.3), we can write
dnA dnB dnC

¼ ¼ ð2:4Þ
1 1 2
where the denominators are the stoichiometric numbers (=stoichiometric coeffi-

cients, where the stoichiometric coefficients of reactants are multiplied by −1). For
the generic chemical reaction
aA þ bB ! cC þ dD ð2:5Þ
we can write
dnA dnB dnC dnD

¼ ¼ ¼ dn ð2:6Þ
a b c d
where n is called the extent of the chemical reaction that specifies the variable
composition along the course of the chemical reaction (n is a state variable). If the
chemical reaction (2.5) is carried out at constant volume, then
1 d½A 1 d½B 1 d½C 1 d½D 1 dn

¼ ¼ ¼ ¼ ð2:7Þ
a dt b dt c dt d dt V dt
where the square brackets stand for molarity. Each member of (2.7) defines the rate
of the chemical reaction that can be obtained at any specified instant during the
course of the chemical reaction (E2). Note that (2.7) assumes a time-independent
stoichiometry, that is, the reaction stoichiometry is assumed to be valid at every
instant during the reaction. This assumption is not always true. If the chemical
reaction has intermediate species with lifetimes that enable their detection, then the
stoichiometry is not time-independent, and the rate of the reaction cannot be defined
in a unique way. In these cases, we are really talking about a sequence of chemical
reactions, and n is a state variable for each chemical reaction in the sequence.
2.2 Experimental Rate Equation
Experimental studies show that the rates of many chemical reactions depend on the
concentrations of the reactants according to the following equation:
v ¼ k½Ax ½By ð2:8Þ
where A and B are reactants, the exponents x and y are called partial orders of A
and B, and k is called the rate constant. The overall order of the reaction is equal
to x + y. Since k, x, y, and the concentrations in the second member of (2.8) are
experimentally determined quantities, (2.8) is called the experimental rate equa-
tion. Partial orders x and y usually take the values 1, 2, or 0 (when the rate equation
does not depend on the concentration of a particular reactant, its partial order with
respect to that reactant is zero). Since the reaction rate depends on temperature, the
experimental rate equation is always determined at a specified temperature.
Most chemical reactions have more than one reactant. When the partial order of
one particular reactant is to be experimentally determined, conditions should be set
to guarantee that other reactants do not interfere, and their concentrations will not
vary significantly. For instance, considering a chemical reaction with reactants A
and B, the experimental determination of the partial order of reactant A requires that
[B] be kept approximately constant. Under these conditions, (2.8) takes the form
0
v ¼ k ½Ax ð2:9Þ
where k′ = k[B]y constant, because [B] does not significantly change in the
course of the reaction. Equation (2.9) is called the pseudo rate equation, and k′
represents the pseudo rate constant. The partial order x acts like the overall
pseudo-order for the chemical reaction. The isolation and initial rate methods allow
separation of the concentration variables of a rate equation in order to determine the
partial order of a particular reactant.
In the experimental determination of the partial order of A by the isolation
method, the concentration of B should greatly exceed the initial concentration of A.
If A and B are in the stoichiometric proportion 1:1 and the initial concentrations of
A and B are in the ratio 1:100 (for example, [A]0 = 0.100 M and [B]0 = 10.0 M),
then when 99 % of A has reacted, the change in the concentration of B is only 1 %.
The initial rate method takes the rate equation at the initial instant (t = 0),
v0 ¼ k½Ax0 ½By0 ð2:10Þ
The determination of the partial order of A requires kinetic experiments with

different initial concentrations of A for the same initial concentration of B. For
chemical reactions with products that decompose or interfere during the course of
the reaction, the initial rate method is the sole method available for the kinetic study
of such chemical reactions.
2.2 Experimental Rate Equation 87
2.2.1 First-Order Reactions
The rate equation for a first-order chemical reaction with a single reactant (or a
pseudo first-order chemical reaction) is given by
d½A=dt ¼ k½A ð2:11Þ
Solving this differential equation consists in finding the function f(t) = [A] that
satisfies (2.11) and the initial condition f(0) = [A]0, where [A]0 is the initial con-
centration of A. While it is easy to conclude that the function that satisfies (2.11) is
an exponential [the derivative of expðktÞ is k expðktÞ], we solve (2.11) in a
more general way that can be applied to second-order reactions. We begin by
separating variables [A] and t in each member,
d½A=½A ¼ kdt ð2:12Þ
Now we obtain indefinite integrals of both members. The integration limits for
the first member are [A]0 and [A], and for the second member, are the corre-
sponding values of time, that is, 0 and t. We obtain

ln ½A=½A0 ¼ kt ð2:13Þ
This equation is equivalent to
ln½A ¼ ln½A0 kt ð2:14Þ
that is, the graph of ln[A] as a function of t gives a straight line whose intercept is ln
[A]0 and whose slope is equal to −k. If we now consider the time interval t = t1/2
such that [A] is half its initial value, [A] = [A]0/2, then substitution of these
equalities in (2.13) leads to
t1=2 ¼ ln2=k ð2:15Þ
where the time interval t1/2 is called the half-life. Contrary to what the name might
suggest, the half-life is not half of the time to reaction completion. Equality (2.15),
valid for first-order chemical reactions, shows that the half-life does not depend on
[A]0. This is an important result that applies only to first-order chemical reactions.
Hence, when the half-life of a chemical reaction is independent of the initial con-
centration, we conclude that the chemical reaction has first-order kinetics (Fig. 2.2).
Equality (2.13) can be rewritten in the following equivalent way,
½A ¼ ½A0 ekt ð2:16Þ
which shows that [A] has exponential decay with time. Spontaneous decays of
radioactive atomic nuclei are first-order processes. Adapting the above equalities to
Fig. 2.2 The half-life of a first-order reaction does not depend on [A]0. For the kinetic profile
illustrated above, the half-life is approximately equal to 2 min. Figure obtained with Mathematica
radioactive decay simply requires substitution of [A] by the number of radioactive

nuclei N and of [A]0 by N0. After these substitutions, Eq. (2.11) shows that the rate
of radioactive decay (−dN/dt) is proportional to the number of nuclei that have not
yet decayed (N). In radioactive studies, the half-life t1/2 is preferred to the rate
constant k, in contrast to what happens in chemical kinetics.
2.2.2 Second-Order Reactions
The rate equation for a second-order chemical reaction with a single reactant (or a
pseudo second-order chemical reaction) is given by
d½A=dt ¼ k½A2 ð2:17Þ
Separation of variables leads to
d½A=½A2 ¼ kdt ð2:18Þ
After integration of both members, we obtain
1 1
¼ þ kt ð2:19Þ
½A ½A0
Hence, the graph of 1/[A] as a function of t gives a straight line whose intercept
is 1/[A]0 and whose slope is equal to k. Substitution of [A] by [A]0/2 and of t by t1/2
in (2.19) gives the half-life for a second-order reaction:
2.2 Experimental Rate Equation 89
1
t1=2 ¼ ð2:20Þ
k½A0
Hence, the half-life of a second-order reaction is inversely proportional to the

initial concentration.
2.2.3 Zeroth-Order Reactions
Zeroth order with respect to a particular reactant means that the chemical reaction
rate does not depend on the concentration of that reactant. This can occur when the
reactant is not involved in the slowest step of the reaction (the rate-limiting or
rate-determining step), or when the reaction rate depends on the adsorption of a
particular reactant by a fully covered catalyst surface. Since the surface is saturated,
the reaction rate does not depend on the concentration of the adsorbed reactant.
The rate equation of a zeroth-order chemical reaction for a single reactant is
given by
d½A=dt ¼ k½A0 ¼ k ð2:21Þ
After separation of variables and integration of both members, we obtain
½A ¼ ½A0 kt ð2:22Þ
Thus, the graph of [A] as a function of t gives a straight line whose intercept is
[A]0 and whose slope is equal to −k. Substitution of [A] by [A]0/2 and of t by t1/2 in
(2.22) leads to
t1=2 ¼ ½A0 =ð2kÞ ð2:23Þ
Hence, the half-life of a zeroth-order reaction is proportional to the initial

concentration.
2.3 Effect of Temperature Change
The temperature-dependence of the rate of a chemical reaction lies essentially in the

rate constant [see (2.8)], since the concentration factors expressed in molality are not
affected by temperature (concentrations that involve a volume in their definition, like
molarity, might be slightly affected by temperature variation). Based on observation
of rate constant variations with temperature, Arrhenius (1859–1927; Nobel Prize in
chemistry in 1903) proposed the k(T) empirical, dependence given by

Ea
kðTÞ ¼ A exp ð2:24Þ
RT
where Ea is the Arrhenius activation energy and A is the Arrhenius A-factor

(E3). The logarithm of (2.24) is
Ea
lnkðTÞ ¼ lnA ð2:25Þ
RT
thus showing that ln kðTÞ depends linearly on 1/T. The intercept of the resulting
straight line is ln A and the slope is equal to −Ea/R. The derivative of (2.25) is given
by
dlnkðTÞ Ea
¼ ð2:26Þ
dT RT 2
This equality suggests a way for determining the activation energy from
experimental data on k(T), provided the chemical reaction rate constant follows an
Arrhenius dependence.
Above the troposphere, the stratosphere ranges in altitude from about 11–50 km,
with temperature varying approximately from −60 to −2 °C. The ozone layer is
mainly found in the lower layer of the stratosphere, up to an altitude of approxi-
mately 30 km. As an example of the Arrhenius dependence, consider the
second-order stratospheric chemical reaction
N þ O2 ! NO þ O ð2:27Þ
for which A = 1.5 10−11 cm3 molecule−1 s−1 and Ea/R = 3600 K (values taken
from Handbook of Chemistry and Physics, 2011). The Arrhenius
temperature-dependence of the rate constant of the above chemical reaction is
illustrated in Fig. 2.3.
Fig. 2.3 Rate constant of the chemical reaction N + O2 ! NO + O as a function of temperature

(left), and lnk as a function of 103/T (right), in the temperature range from 200 to 350 K. Note that
the lowest value of 103/T in the horizontal axis of the plot at the right is far from the origin, which
would correspond to an infinite temperature. Thus, the intersection of the straight line with the
horizontal axis is far from the intercept. Graphs obtained with Mathematica
2.3 Effect of Temperature Change 91
Ea/kJ mol−1 exp[–Ea /(RT)]

300 K 600 K
10 2 10−2 1 10−1
50 2 10−9 4 10−5
100 4 10−18 2 10−9
The Arrhenius activation energy has a strong influence on the rate equation. To
illustrate this point, the table shows approximate values of the Arrhenius expo-
nential function for three hypothetical values of the activation energy (10, 50, and
100 kJ mol−1), at two different temperatures (300 and 600 K). At 300 K, the
Arrhenius exponential decreases by a factor of order 10−7 when the activation
energy changes from 10 to 50 kJ mol−1, and by a factor of order 10−9 when the
activation energy changes from 50 to 100 kJ mol−1. In turn, at 600 K, the same
increments in the activation energy lead to exponential decreases by factors of order
10−4. In conclusion, the decrease in the activation energy due to the eventual use of
catalysts and the increase in temperature lead to appreciable increases in the
chemical reaction rates that present the Arrhenius dependence. Note that the catalyst
reduces the activation energy of the chemical reaction, thus leading to an increase in
the reaction rate. The decrease in the activation energy is the result of a change of
mechanism without altering the initial and final states of the reaction (the reactants
and products of the overall chemical reaction).
In the graph of ln k as a function of 103/T [see (2.25)], the extrapolation needed to
obtain the intercept may lead to a large uncertainty in the determination of the A-
factor, since the data points are usually within a short range of 103/T. For instance, the
data points of Fig. 2.3 range from 103/T = 2.85 K−1 (T 351 K) to 103/T = 5.0 K−1
(T = 200 K). At 1000 K, the value of 103/T is 1 K−1, still far from the origin.
2.4 Elementary Reactions
Elementary reactions occur in a single step and so have time-independent stoi-

chiometries and do not have reaction intermediates. An elementary reaction has a
single potential energy maximum in the reaction path as a function of the reaction
coordinate. In contrast, the existence of one reaction intermediate in a chemical
reaction implies an energy minimum in the reaction path between reactants and
products.
Elementary reactions can be classified according to their molecularity, which is
the number of reactant molecules that take part in the reaction. Those reactions that
involve one, two, and, less frequently, three reactant molecules are called uni-
molecular, bimolecular, and trimolecular reactions, respectively. There are few
examples of chemical reactions thought to involve three reactant molecules, and no
reactions are known to involve four reactant molecules, since it is highly
improbable that four molecules collide at the same instant with the energy and
orientation required for a chemical reaction. Note that the term molecularity applies
only to elementary reactions or to individual steps of complex reactions.
Consider a reaction of the following type:
Nu þ RX ! RNu þ X ð2:28Þ
where Nu− stands for a nucleophile and X represents an electronegative atom or an

electronegative group bonded to a tetrahedral carbon atom in the R radical.
A chemical reaction of this type is called a nucleophilic substitution reaction. In
fact, it is a substitution reaction (X is substituted by Nu) resulting from a nucle-
ophilic attack of Nu− on the carbon atom bonded to X:
Nu þ Cd þ Xd ð2:29Þ
If the nucleophilic substitution reaction occurs in a single step, the approach of

Nu− is concerted with the withdrawal of X−. In this case, the transition state
involves both reactant species Nu− and RX, and the nucleophilic substitution
reaction is bimolecular and named SN2. The reaction is said to occur in a bi-
molecular concerted step, and its rate equation is given by
v ¼ k½RX]½Nu ð2:30Þ
These considerations can be illustrated by the reaction of bromomethane with

sodium hydroxide, in methanol, at 25 °C,
OH þ CH3 Br ! CH3 OH þ Br ð2:31Þ
This reaction has second-order kinetics and obeys the following experimental
rate equation:
v ¼ k½CH3 Br][OH] ð2:32Þ
This reaction mechanism has a single step with a single transition state, [HO…
CH3…Br]−. The transition state involves both reactant species, and the nucleophilic
substitution is bimolecular, that is, the mechanism is SN2.
2.5 Complex Reactions
When the mechanism of a chemical reaction consists of more than one step, the reaction
is said to be a complex reaction. Experimental evidence for the existence of one
reaction intermediate leads to the conclusion that the reaction mechanism is formed by
at least two steps. Sometimes, the lifetime of a reaction intermediate enables its isolation
and characterization. However, reaction intermediates usually are reactive species with
2.5 Complex Reactions 93
very low concentrations, thus being difficult to detect. The physical methods generally
used in the detection of short-lived reaction intermediates are spectroscopic methods,
which involve an interaction with electromagnetic radiation.
The rate equation of an elementary reaction can be written once the reaction
stoichiometry is known. Hence, when the rate equation does not reflect the reaction
stoichiometry, we have a complex reaction. For example, the reaction between
hypochlorite and iodide ions in aqueous solution,
ClO þ I ! IO þ Cl ð2:33Þ
has the following experimental rate equation:
v ¼ k½ClO ½I =½OH ð2:34Þ
The presence of the concentration of the hydroxide ion in the rate equation and
its absence from the stoichiometric Eq. (2.33) leads to the conclusion that the
chemical reaction (2.33) has more than one step in its mechanism; that is, it is a
complex reaction.
The reaction of 2-bromo-2-methylpropane, (CH3)3CBr, with sodium hydroxide,
in methanol at 25 °C,
OH þ ðCH3 Þ3 CBr ! ðCH3 Þ3 COH þ Br ð2:35Þ
has the following experimental first-order rate equation:
v ¼ k½ðCH3 Þ3 CBr ð2:36Þ
This rate equation shows that the slowest reaction step involves a single reactant
species, namely, (CH3)3CBr. Having the chemical reaction (2.35) in mind, we can
infer from (2.36) that the first step of (2.35) consists in the dissociation of
(CH3)3CBr in the (CH3)3C+ carbocation and the bromide ion,
ðCH3 Þ3 CBr ðCH3 Þ3 C þ þ Br slow step ð2:37Þ
Considering the reactivity of the (CH3)3C+ carbocation, it is likely that this

chemical species recombines with the bromide ion to form
2-bromo-2-methylpropane, giving rise to an equilibrium. However, when the
(CH3)3C+ carbocation reacts with a hydroxide ion, an alcohol molecule is formed,
ðCH3 Þ3 C þ þ OH ! ðCH3 Þ3 COH fast step ð2:38Þ
This is a fast step, because the hydroxide ion has a stronger nucleophilic char-
acter than the bromide ion. Addition of steps (2.37) and (2.38) brings back the
original overall reaction (2.35). The transition state involves a single reactant
molecule, (CH3)3CBr, and the overall reaction (2.35) represents a unimolecular
nucleophilic substitution, thus being an example of an SN1 mechanism.
The difference in the mechanisms of reactions (2.31) and (2.35) lies in the
different stabilities of the involved carbocations, CH3+ and (CH3)3C+. CH3+ is a
stronger electrophile than (CH3)3C+, because the same positive charge is distributed
over a larger carbocation in (CH3)3C+ as compared with CH3+. Hence, reaction
(2.31) does not have a slow initial step for the dissociation of CH3Br into CH3+ and
Br−, since CH3+ is quite a strong electrophile for that to occur.
2.6 Extremely Fast Reactions
Most reactions of ions in aqueous solution are extremely fast, reaching equilibrium
in times of order 10−10 and 10−12 s. Such reactions cannot be studied using con-
ventional methods that depend on the mixture of reactants, since the diffusion times
(times for migration of reactant molecules until they collide with each other) are
orders of magnitude greater than reaction times. Techniques can be used that apply
an almost instantaneous perturbation, after which the system momentarily leaves
the equilibrium state (concentrations of reactants and products are changed for a
moment) and returns to the equilibrium state, since the duration of the perturbation
is not significant when compared with the reaction half-life. The process of
returning to the equilibrium state is called relaxation. The applied perturbation can
be a shock wave, a pulse of electromagnetic radiation that produces a photo-
chemical reversible reaction (flash photolysis), a sudden temperature jump (T-
jump), or a sudden pressure jump associated with sound absorption in gaseous
systems (Eigen 1954). The study of extremely fast reactions using brief energy
pulses led to the award of the 1967 Nobel Prize in chemistry to Manfred Eigen,
Ronald Norrish, and George Porter.
In order to proceed with the study of an extremely fast chemical reaction, it is
necessary to know the rate equation for the reaction in the forward and reverse
directions. Consider the reaction in aqueous solution
AþBC ð2:39Þ
where A and B are positive and negative ions and C is the chemical species
resulting from the combination of those ions. Assume that the forward reaction is of
first order in A and in B and that the reverse reaction is of first order in C, and
represent by k! and k← the rate constants in the forward and reverse reactions.
When a brief perturbation, for instance a temperature jump, is applied to the
reaction at equilibrium, the A, B, and C concentrations will change with respect to
their equilibrium values according to the following equalities, which preserve the
reaction stoichiometry:
½A ¼ ½Aeq d ½B ¼ ½Beq d ½C ¼ ½Ceq þ d ð2:40Þ

2.6 Extremely Fast Reactions 95
where d represents an infinitesimal change in concentration caused by the pertur-

bation. These equalities can be written in the following equivalent way:
D½A ½A ½Aeq ¼ d D½B ½B ½Beq ¼ d D½C ½C ½Ceq ¼ d

ð2:41Þ
Therefore, the relaxation rate satisfies the following equalities:
dD½A dD½B dD½C dd

¼ ¼ ¼ ð2:42Þ
dt dt dt dt
where, according to (2.39),
dd
¼ k! ½A½B k ½C ð2:43Þ
dt
dd
¼ a bd þ vd2 ð2:44Þ
dt
where

a ¼ k! ½Aeq ½Beq k ½Ceq b ¼ k! ½Aeq þ ½Beq þ k v ¼ k! ð2:45Þ
Since the equilibrium constant of (2.39) is given by
k! ½Ceq
Keq ¼ ¼ ð2:46Þ
k ½Aeq ½Beq
it follows that a = 0. In addition, in (2.44), the second-order term in d is negligible

when compared with the first-order term. Therefore,
dd
bd ð2:47Þ
dt
This result shows that the perturbation d follows first-order kinetics,
d ¼ d0 ebt ¼ d0 et=s ð2:48Þ
where
b ¼ s1 ð2:49Þ
(the units of b are the inverse of time, s−1) (E4). The rate constants k! and k← can
be experimentally determined if two values of b, say b1 and b2, have been previ-
ously determined corresponding to two sets of [A]eq and [B]eq values (set1 and set2).
2.6.1 Neutralization Reaction in Water
Consider water and the neutralization equilibrium
H þ ðaqÞ þ OH ðaqÞ H2 O(aqÞ ð2:50Þ
where H+(aq) stands for the symbolic representation of a protonated water molecule H3O+
or a protonated group of hydrogen-bonded water molecules as H5O2+, H7O3+, H9O4+, ….
The rate constant for the combination of H+ and OH− is equal to k! = 1.4 1011
dm3 mol−1 s−1, that is, the neutralization is an extremely fast reaction, whereas the rate
constant for the reverse reaction (ionization of H2O) is equal to k← = 2.5 10−5 s−1
(Eigen 1967). At equilibrium, making use of (2.46), we can write
k! ½H þ eq ½OH eq k ½H2 O ¼ 0 ð2:51Þ
where [H2O] 55.6 mol dm−3. Hence, the constant for the ionic product of water
at 25 °C is given by
k
½H þ eq ½OH eq ¼ ½H2 O ¼ 1:0 1014 ð2:52Þ
k!
that is, [H+]eq = [OH−]eq = 1.0 10−7 mol dm−3, at 25 °C.
2.7 Chemical Oscillations
Most chemical reactions function like one-way streets: the concentrations of

reactants decrease, those of reaction intermediates increase at first and decrease
later, and the concentrations of products increase. However, for a few reactions far
from equilibrium, the concentrations of some chemical species oscillate, i.e.,
increase and decrease repeatedly. These reactions are called chemical oscillators.
Chemical oscillations differ from pendulum oscillations: the film of a pendulum
cannot be distinguished from the same film run backward (there is no arrow of time
in the pendulum), whereas chemical oscillators are associated with entropy pro-
duction due to irreversibleprocesses that occur in the reacting system.
2.7 Chemical Oscillations 97
The first reported observation of a periodic reaction in homogeneous solution is

due to Bray in 1921 (Bray 1921). At that time, Lotka had reported the mathematical
study of a system of differential equations describing the mechanism of a hypo-
thetical periodic chemical reaction (Lotka 1920). The best-known oscillating
chemical reaction results from experiments carried out by Belousov in 1958 and by
Zhabotinsky in 1964, and is known as the Belousov–Zhabotinsky experiment
(Winfree 1984). For years, the results of the Belousov–Zhabotinsky experiment
were regarded with suspicion, since oscillations are incompatible with the existence
of a Gibbs energy minimum at equilibrium. This apparent incompatibility was
resolved when it was realized that chemical oscillations occur far from equilibrium.
2.7.1 Brusselator
In 1968, Prigogine and Lefever developed a model that shows how a chemical
reaction, far from equilibrium, can pass from a stationary point to an oscillatory
state (Prigogine 1968). This chemical oscillator, often called the Brusselator as a
reminder of the Brussels thermodynamic school founded by Prigogine, consists of
the following mechanism:
k1
A ! X
k2
B þ X ! Y þ D
ð2:53Þ
k3
2X þ Y ! 3X
k4
X ! E
where the inverse reactions are assumed to have negligible rate constants, and the
overall chemical reaction is
AþB ! DþE ð2:54Þ
In order to simplify the notation, A, B, X, and Y stand for [A], [B], [X], and [Y],
respectively. Note that an increase of X, in elementary reactions 1 and 3, is followed
by its decrease, in chemical reactions 2 and 4. A similar oscillatory behavior can be
assigned to Y that decreases in chemical reaction 3, where Y is a reactant, and
increases in reaction 2, where Y is a product.
The rate equations for the reaction intermediates X and Y in (2.53) are given by
the following set of nonlinear coupled differential equations:
dX dY
¼ k1 A k2 BX þ k3 X 2 Y k4 X ¼ k2 BX k3 X 2 Y ð2:55Þ
dt dt
Fig. 2.4 In a flow reactor, the constant supply of A and B keeps the concentrations of these
reactants in the reaction mixture fixed, and the constant removal of D and E maintains the system
far from these reactants equilibrium
where the effects of diffusion of X and Y have been ignored, since we assume that
chemical reactions (2.53) occur in a homogeneous medium, and so there is spatial
uniformity of concentrations of all involved chemical species. In solving this set of
nonlinear differential equations, the initial conditions are such that the corre-
sponding physical system (the reaction mixture) is kept far from equilibrium: A and
B have fixed values, while D and E are constantly removed from the reaction
mixture to prevent reverse chemical reactions from occurring (Fig 2.4). Solving
(2.55) for its stationary-point solutions (dX/dt = 0 and dY/dt = 0) leads to
k1 k2 k4 B
Xs ¼ A Ys ¼ ð2:56Þ
k4 k1 k3 A
where the subscript s stands for stationary point.

In order to assess the linear stability of the stationary point, we consider a
time-dependent perturbation added to the stationary point solutions (2.56) and find
out whether the perturbation increases or decreases in time (see Kondepudi and
Prigogine, Further Reading).
The rate Eq. (2.55) can be written in the following more general way:
dX dY
¼ Z 1 ðX; YÞ ¼ Z2 ðX; YÞ ð2:57Þ
dt dt
The time-dependent perturbations x(t) and y(t) are added to the stationary-point
coordinates Xs and Ys,
X ¼ Xs þ xðtÞ Y ¼ Ys þ yðtÞ ð2:58Þ
Therefore,
dX dx dY dy
¼ ¼ ð2:59Þ
dt dt dt dt
since Xs and Ys are constants of the experiment. The Taylor expansion of Z1 and Z2
in (2.57) about the stationary point leads to

@Z1 @Z1
Z1 ðXs þ x; Ys þ yÞ ¼ Z1 ðXs ; Ys Þ þ xþ yþ
@X s @Y s
ð2:60Þ
@Z2 @Z2
Z2 ðXs þ x; Ys þ yÞ ¼ Z2 ðXs ; Ys Þ þ xþ yþ
@X s @Y s
Since x(t) and y(t) represent small perturbations, only the linear terms in (2.60)
are retained (quadratic and higher-order terms are ignored). Hence, making use of
(2.57) and (2.59), and noting that Z1(Xs, Ys) and Z2(Xs, Ys) are zero by definition of
the stationary point, we conclude that

dx @Z1 @Z1 dy @Z2 @Z2
¼ xþ y ¼ xþ y ð2:61Þ
dt @X s @Y s dt @X s @Y s
or in vector notation,
dx
¼ Kx ð2:62Þ
dt
where
" @Z #
@Z1
@Z
@X s @Z
1
@Y s x
K¼ x¼ ð2:63Þ
2
@X s
2
@Y s
y
with K being the Jacobian matrix. The Jacobian matrix elements quantify the
variation of the Z1 and Z2 rates with respect to changes in X and Y at the stationary
point.
For the Brusselator, the Jacobian matrix is given by

k2 B k4 k3 Xs2
K¼ ð2:64Þ
k2 B k3 Xs2
where Xs is given by the first equality of (2.56). The characteristic equation (see
§1) is
detðK kI) ¼ 0 ð2:65Þ
that is,

k2 B k4 k k3 Xs2
¼ 0 k2 ðk2 B k4 k3 Xs2 Þk þ k3 k4 Xs2 ¼ 0
k2 B k3 Xs2 k
ð2:66Þ
The eigenvalues k± are obtained by solving Eq. (2.66),
p 1
k ¼ ðp2 4qÞ1=2 ð2:67Þ
2 2
where
p ¼ k2 B k4 k3 Xs2 q ¼ k3 k4 Xs2 ð2:68Þ
(E5).
Consider now the following set of parameters k1 = k2 = k3 = k4 = 1.0 and the
initial condition A = 1.0, leading to Xs = 1.0, Ys = B [see (2.56)], p = B − 2, q = 1
[see (2.68)], and p2 − 4q = B2 − 4B. For B = 2, we have p = 0, and the eigen-
values of the Jacobian matrix are pure imaginary. For B = 1.5 or 2.5, we have
p2 − 4q = (±0.5)2 − 4 < 0, and the eigenvalues have a nonzero imaginary
part. The value B = 1.5 leads to p < 0 and to a complex conjugate pair with a
negative real part, whereas B = 2.5 gives p > 0 and a positive real part. We now
return to the initial conditions A = 1.0 and B = 2. From (2.62) and (2.64), we can
write

x_ ðtÞ 1 1 xðtÞ
¼ ð2:69Þ
y_ ðtÞ 2 1 yðtÞ
The eigenvalues of this Jacobian matrix are ±i, and the corresponding eigen-
vectors are (−1 − i, 2)T and (−1 + i, 2)T. Then the general solution to the above
system of differential equations (see §2) is

xðtÞ 1 i it 1 þ i it
¼ c1 e þ c2 e ð2:70Þ
yðtÞ 2 2
After applying e±it = cos t ± isin t, we obtain

xðtÞ a v
¼ cos t þ sin t ð2:71Þ
yðtÞ b d
This equality shows that the x(t) and y(t) functions are linear combinations of
cosine and sine functions. We now define U and V variables by U X − Xs and
V Y − Ys, so that the stationary point is at (Us, Vs) = (0, 0) instead of at (Xs, Ys).
Figure 2.5 considers the same set of parameters (k1 = k2 = k3 = k4 = 1.0) and the
initial condition A = 1.0, with B equal to 1.5, 2.0, and 2.5, and shows U and V as
functions of time in the first plot and V(t) as a function of U(t) in the second plot
(phase trajectory). For B = 1.5, the eigenvalues of the Jacobian matrix are a
Fig. 2.5 Oscillations and phase trajectories for the Brusselator. Figures obtained with
Mathematica
complex conjugate pair with a negative real part, and U and V quickly decay to zero
as the phase trajectories rapidly spiral into the origin. For B = 2.0, the eigenvalues
of Jacobian matrix are pure imaginary, and the trajectories spiral asymptotically to
the origin as t ! ∞ (see Mathematica code M2). In turn, when B = 2.5, the
eigenvalues of the Jacobian matrix are a complex conjugate pair with a positive real
part, and the trajectories quickly approach a stable limit cycle as t ! ∞, with
U(t) and V(t) indefinitely maintaining the same amplitudes and same shapes (the
system is said to have reached permanent oscillations) (E6).
Notes
§1. Matrix Diagonalization
When a linear transformation represented by an n n matrix A is applied to a

vector u represented by an n 1 column vector, another n 1 vector w of the
same vector space is obtained according to the following equation:
Au ¼ w ð2:72Þ
This equality represents a linear transformation within the same vector space,
which is said to preserve the vector length, since u and w have the same length. One
especial and very important case of (2.72) quite frequent in quantum mechanics
occurs when the vector w is equal to a scalar k times the column vector u, that is,
Au ¼ ku ð2:73Þ
where the scalar k and the vector u are to be determined (they are not known
beforehand). Certainly, the trivial solution u = 0 is of no practical interest! Values
of k and nontrivial vectors u that satisfy (2.73) are called eigenvalues and eigen-
vectors, respectively. Equality (2.73) can be rewritten as
ðA kI Þu ¼ 0 ð2:74Þ
where the expression in parentheses is an n n matrix. The equality (2.74) rep-

resents a homogeneous system of linear equations in n unknowns, and so the
condition for a nontrivial solution of (2.74) is
detðA kIÞ ¼ jA kIj ¼ 0 ð2:75Þ
Expansion of this determinant yields an nth-degree polynomial in k. The

polynomial Eq. (2.75) is called the characteristic equation or secular equation.
The n roots of the characteristic equation are the n eigenvalues that satisfy (2.75).
Substitution of each of these eigenvalues in (2.74) enables one to determine the
corresponding eigenvector only to within a multiplicative constant, because if u is a
solution of (2.74), so is any multiple of u.
We can now illustrate the above considerations with the matrix
0 1
1 0 1
A ¼ @1 2 1 A ð2:76Þ
2 2 3
Notes 103
The characteristic equation is

1 k 0 1

detðA kIÞ ¼ 1 2k 1 ¼ ðk 3Þðk 2Þðk 1Þ ¼ 0 ð2:77Þ
2 2 3 k
whose roots are k = 1, 2, 3. We obtain the following eigenvectors:

0 1 0 1
0 0 1 1
k ¼ 1 ) ðA IÞu1 ¼ @ 1 1 1 A u 1 ¼ 0 ) u1 ¼ a @ 1 A ð2:78Þ
2 2 2 0
0 1 0 1
1 0 1 2
k ¼ 2 ) ðA 2IÞu2 ¼ @ 1 0 1 A u2 ¼ 0 ) u 2 ¼ b @ 1 A ð2:79Þ
2 2 1 2
0 1 0 1
2 0 1 1
k ¼ 3 ) ðA 3IÞu3 ¼ @ 1 1 1 Au3 ¼ 0 ) u3 ¼ c@ 1 A ð2:80Þ
2 2 0 2
The multiplicative constants a, b, and c can be determined once the lengths of

u1, u2, u3 are known. The eigenvectors u1, u2, u3 can be grouped as columns of a
3 3 matrix U, and the eigenvalues can be arranged as diagonal entries of the
diagonal matrix D,
0 1 0 1
j j j k1 0 0
U ¼@ u 1 u2 u3 A D¼@ 0 k2 0A ð2:81Þ
j j j 0 0 k3
Using (2.74) for each of the eigenvalues k1, k2, and k3, we can write
0 10 1
j j j k1 0 0
AU ¼ðAu1 ; Au2 ; Au3 Þ ¼ ðk1 u1 ; k2 u2 ; k3 u3 Þ@ u1 u2 u3 A@ 0 k2 0A
j j j 0 0 k3
¼ UD
ð2:82Þ
Note that the product UD gives each eigenvalue corresponding to a column of

the resulting matrix, as we want. On the other hand, the product DU would give
each eigenvalue corresponding to a row of the product matrix. Considering (2.82),
we can write
AU ¼ UD ð2:83Þ
If the eigenvectors are linearly independent, then U is nonsingular, that is, its
determinant is different from zero. Therefore, U−1 exists, and we can multiply
(2.83) from the left by U−1 and obtain
U 1 AU ¼ D ð2:84Þ
Two n n matrices A and B are called similar if B = X−1AX for some non-
singular and therefore invertible matrix X. The transformation A ! X−1AX is a
similarity transformation. Similar matrices represent the same transformation
(operator) in two different bases of the same vector space related by the matrix
X. Equality (2.84) shows a particular kind of a similarity transformation, whose
result is a diagonal matrix. For that reason, (2.84) represents the diagonalization of
the matrix A. The similarity transformation that diagonalizes the matrix A linearly
combines the basis vectors, so that the resulting matrix becomes diagonal in the
new basis.
We now illustrate the concepts of similarity transformation and similar matrices
using a three-dimensional vector space with Cartesian coordinates and the basis
vectors (xyz)T and (rst)T. Consider the 3 3 matrix A that converts (xyz)T into (x′y′
z′)T,
0 0 1 0 1
x x
@ y 0 A ¼ A@ y A ð2:85Þ
0
z z
and the 3 3 matrix B that converts (rst)T into (r′s′t′)T,

0 01 0 1
r r
@ s 0 A ¼ B@ s A ð2:86Þ
0
t t
Matrix X relates these bases: by applying the matrix X to (rst)T and (r′s′t′)T, we
obtain (xyz)T and (x′y′z′)T, respectively,
0 1 0 1 0 0 1 0 01
x r x r
@ y A ¼ X@ s A @ y A ¼ X @ s0 A
0
ð2:87Þ
0 0
z t z t
The change of basis can be carried in the opposite direction, from (xyz)T and (x′y′z
′) to (rst)T and (r′s′t′)T, meaning that X−1 exists, that is, X is nonsingular (its deter-
T
minant is different from zero). If X is a symmetry operation, then as an element of an

algebraic group, it has an inverse. Therefore, substitution of (2.87) in (2.85) gives
0 0 1 0 1 0 01 0 1 0 01 0 1
x x r r r r
@ y0 A ¼ A@ y A ) X @ s0 A ¼ AX @ s A ) @ s0 A ¼ X 1 AX @ s A ð2:88Þ
0 0 0
z z t t t t
Notes 105
Comparison of this result with (2.86) leads to
B ¼ X 1 AX ð2:89Þ
where X−1AX is called a similarity transformation and A and B are said to be

similar matrices, that is, they represent the same linear transformation after a
change of basis implemented by the matrix X. If both coordinate systems are
orthogonal, then the inverse of matrix X is equal to its transpose, that is,
X 1 ¼ X T ð2:90Þ
The matrix X is said to be orthogonal, and XTAX is called an orthogonal

transformation. An important conclusion about changing the vectorial basis within
the same vector space is that a similarity transformation preserves the sum of the
diagonal elements or trace of matrices A and B.
§2. Systems of First-Order Linear Differential Equations
Consider the system of two simultaneous first-order linear equations that corre-
sponds to (2.62),
x_ ¼ Kx ð2:91Þ
Since this system of differential equations has constant coefficients, it is rea-

sonable to expect that a solution might be of the form
xðtÞ ¼ uekt ð2:92Þ
where u and k are to be determined. Substitution of (2.92) into (2.91) and can-
cellation of ekt from both sides gives the eigenvalue equation
Ku ¼ ku ð2:93Þ
Since (2.91) is a system of linear differential equations, if x1(t) and x2(t) are
solutions, then
xð t Þ ¼ c 1 x1 ð t Þ þ c 2 x 2 ð t Þ ð2:94Þ
is also a solution. Therefore, from (2.92) and (2.94) we conclude that the general
solution for (2.91) is given by
xðtÞ ¼ c1 u1 ek1 t þ c2 u2 ek2 t ð2:95Þ
where u1 and u2 are eigenvectors and k1 and k2 are the corresponding eigenvalues
(see McQuarrie, Sect. 11.6, p. 563).
Mathematica Codes
M1. First-Order Chemical Reaction
This Mathematica code does a least squares fit of kinetic data to a model of

exponential decay and plots the resulting exponential function and the kinetic data
points. The first line of code presents a list of data points named data, and the
second line defines the model of exponential decay
model=a Exp[-k t];
In turn, the third line of code,
fit=FindFit[data,model,{a,k},t]
does the fitting, where the command FindFit finds a least squares fit for the
defined model function, and the second line of results presents the function
Function[t,5.01042 e0:344655t ]
where the comma inside square brackets separates the time variable t from the
body of the pure function
5:01042 e0:344655t
The function f[t] is plotted in the last line of code, and the data points are
rendered over the plotted curve by the Mathematica command Epilog, whose
main action is
Map[Point,data]
This Mathematica function Map applies Point to each element of data.
Suggestion: Complete the code to evaluate the half-life to a precision of four
decimal places.
M2. Brusselator
This Mathematica code solves the equations of the Brusselator for k1 = k2 =

k3 = k4 = 1.0, A = 1.0, and B = 2.0. Representing by X(t) and Y(t) the concentra-
tions of the intermediates as functions of time and by Xs and Ys the corresponding
stationary values [for the above set of parameters, Xs = 1.0 and Ys = 2.0, see
(2.56)], the variables U and V are such that X = U + Xs and Y = V + Ys and the rate
equations are solved using the Mathematica command NDSolve, which finds a
numerical solution for ordinary differential equations like the rate equations of the
Brusselator [see (2.55)].
The last line of code plots in a row the U(t) and V(t) functions in the same graph
and V as a function of U (a parametric plot) in a second graph. Removing from the
last code line all the plot style options, we obtain
Row[{Plot[Evaluate[{U[t],V[t]}/.sol],{t,0,mt}],
ParametricPlot[Evaluate[{U[t],V[t]}/.sol],{t,0,mt}]}]
where the Mathematica function Evaluate causes {U[t],V[t]}/.sol
to be evaluated, that is, U[t] and V[t] are replaced by the solutions of the rate
equations sol. Note that the opposite of Evaluate is Hold, which maintains an
expression to which it applies in an unevaluated form. ParametricPlot gen-
erates a plot of V as a function of U, since both of these variables are functions of
time, an external variable, or parameter.
For the above-mentioned initial conditions (k1 = k2 = k3 = k4 = 1.0, A = 1.0,
and B = 2.0), the eigenvalues of the Jacobian matrix are pure imaginary [p = 0 and
q = 1; see (2.67) and (2.68)], and the trajectories spiral asymptotically to the origin
as t ! ∞, as can be concluded by inspection of the phase trajectory shown above.
Suggestion: Using the Mathematica function Table, write a Mathematica code
for solving the equations of the Brusselator for k1 = k2 = k3 = k4 = 1.0, A = 1.0,
and three values of B, namely, B = 1.5, 2.0, and 2.5.
Glossary
Arrhenius equation The empirical exponential dependence of the rate

constant as a function of temperature, proposed by
Arrhenius (1859–1927; Nobel Prize in chemistry in
1903); see (2.24). Contains two empirical
parameters, the Arrhenius A-factor and the Arrhenius
activation energy Ea
Belousov–Zhabotinsky The best-known oscillating chemical reaction,
experiment resulting from experiments carried out by Belousov
in 1958 and Zhabotinsky in 1964. For years, the
results of this experiment were regarded with
suspicion, since oscillations are incompatible with
the existence of a Gibbs energy minimum at
equilibrium. This apparent incompatibility was
solved when it was realized that chemical
oscillations occur far from equilibrium.
Brusselator A model chemical oscillator developed in 1968 by
Prigogine and Lefever, in the Brussels
thermodynamic school founded by Prigogine, that
shows how a chemical reaction, far from
equilibrium, can pass from a stationary point to an
oscillatory state; see (2.53).
Chemical oscillator A complex reaction in which, far from equilibrium,
the concentrations of some chemical species
oscillate, i.e., increase and decrease repeatedly.
Complex reaction A chemical reaction whose mechanism consists of
more than one step, with the slowest step being the
rate-determining step. When existing experimental
evidence points to the occurrence of one reaction
intermediate, then one can conclude that the reaction
mechanism is formed by at least two steps.
Glossary 109
Elementary reaction A chemical reaction that occurs in a single step, has a

time-independent stoichiometry, and does not have
any reaction intermediate. An elementary reaction
has a single potential energy maximum in the
reaction path as a function of the reaction coordinate.
Extremely fast reaction A chemical reaction that reaches equilibrium in times
of order 10−10 and 10−12 s and so cannot be studied
using conventional methods that depend on the
mixture of reactants, since the diffusion times (times
for migration of reactant molecules until they collide
with each other) are orders of magnitude greater than
the above reaction times.
Half-life The time interval required for the concentration of a
reactant or the number of radioactive atoms to
decrease to half its initial value.
Initial rate method Involves measuring the reaction rate of a chemical
reaction at very short times before any significant
changes in the concentrations of reactants occur; see
(2.10).
Isolation method Involves measuring the reaction rate of a chemical
reaction when the concentration of one reactant is
greatly exceeded by the concentrations of all other
reactants so that these do not significantly change
during the reaction. Under these conditions, the rate
equation takes a much simpler form with a pseudo
rate constant; see (2.9) and compare with (2.8).
Kinetic reaction profile The graphical representation of the concentration of
a reactant or product of a chemical reaction as a
function of time.
Overall reaction order The sum of all partial orders in the experimental rate
equation.
Partial order The exponent to which the concentration of a
reactant is raised in the experimental rate equation.
For many chemical reactions, the partial orders are
not equal to the reaction stoichiometric coefficients,
whereas for elementary chemical reactions, partial
orders coincide with the stoichiometric coefficients.
Rate of chemical reaction The time derivative of the concentration of a reactant
or product of a chemical reaction divided by the
corresponding stoichiometric number; see (2.7).
Exercises
E1. Considering the experimental data that led to Fig. 2.1, use Mathematica to
calculate the rate constant for the first-order chemical reaction and the time
elapsed until the concentration of A is reduced to 5 % of its initial value.
E2. Derive (2.7).
E3. For a second-order chemical reaction and the following rate constant values,
use Mathematica for determining the Arrhenius activation energy.
T/K 900 950 1000 1050 1100 1150

k/M−1 s−1 0.01305 0.07686 0.37907 1.60593 5.96669 19.7777
E4. Derive (2.44).

E5. Consider the Brusselator.
(a) Write the set of differential equations for the concentrations of X and Y as
functions of time and determine the stationary point.
(b) Obtain the Jacobian matrix at the stationary point. Assume k1 = k2 = k3 =
k4 = 1.0 and A = 1.0, and determine the Jacobian matrix in terms of B.
(c) Use Mathematica for determining the eigenvalues of the Jacobian matrix for
B = 2.0.
E6. Consider the Lotka–Volterra mechanism
k1
A þ X ! 2X
k2
X þ Y ! 2Y
k3
Y ! B
(a) Write the set of differential equations for the concentrations of X and Y as
functions of time and determine the stationary point.
(b) Determine the Jacobian matrix at the stationary point. Assume k1 = k2 =
k3 = 1 and A = 1.0 and determine the Jacobian matrix for these values.
(c) Use Mathematica for determining the eigenvalues of the Jacobian matrix.
(d) Use Mathematica for solving the kinetic equations, plot U(t), V(t), and the
phase trajectory.
References 111
References
Bray WC (1921) A periodic reaction in homogeneous solution and its relation to catalysis. J Am
Chem Soc 43:1262–1267
Eigen M (1954) Methods for investigation of ionic reactions in aqueous solutions with half-times
as short as 10−9 s. Application to neutralization and hydrolysis reactions Discuss Faraday Soc
17:194–205
Eigen M (1967) Immeasurably fast reactions, Nobel Lecture, December 11
Lotka AJ (1920) Undamped oscillations derived from the law of mass action. J Am Chem Soc
42:1595–1599
Prigogine I, Lefever R (1968) Symmetry breaking instabilities in dissipative systems II. J Chem
Phys 48:1695–1700
Winfree AT (1984) The prehistory of the Belousov-Zhabotinsky oscillator. J Chem Educ 61:661–
663
Further Reading
Kondepudi D, Prigogine I (1998) Modern thermodynamics: from heat engines to dissipative

structures. Wiley
McQuarrie D (2003) Mathematical methods for scientists and engineers. University Science Books
Mortimer M, Taylor P (eds) (2002) Chemical kinetics and mechanism. Series The molecular
world. The Open University
The Schrödinger Equation
3
Abstract
The equation of motion of macroscopic particles was proposed by Isaac Newton
toward the end of the seventeenth century. At the beginning of the twentieth
century, several scientific advances led to the conclusion that classical mechanics
was unable to correctly describe the mechanical behavior of subatomic particles.
In particular, the electron, with its small mass of about 1 over 2 103 the proton
mass, exhibits both particlelike and wavelike properties. This chapter aims at
providing the student with the important background theory on quantum
chemistry. To this end, it contains sections on operators, the harmonic oscillator,
spherical coordinates, angular momentum, the hydrogen atom, the antisymmetry
principle, the variational method, Born–Oppenheimer approximation, the
Hartree–Fock method, and density functional theory, followed by sections on
perturbation theory, time-dependent perturbation theory, absorption and emis-
sion of radiation and Raman scattering, and a final section on molecular
calculations that comprises computational methods and standard basis sets. At
the end of this chapter, the student finds two notes, one on the particle in a
one-dimensional box and the other on the two-particle rigid rotor, that
complement the main text without diverting the student’s attention from the
sequence of the chapter sections. Next, several Mathematica codes (Wave
Equation, Helmholtz Equation, Harmonic Oscillator, Spherical Harmonics,
Determinants, Systems of Homogeneous Linear Equations, Normalization
Constants for Slater-Type Orbitals, Functional Derivative, STO versus a
Gaussian Function at the Origin, Fitting Gaussian Functions to a 1s Hydrogen
Orbital, Product of Gaussian Functions) are interconnected with the main text,
along with detailed explanations of new Mathematica commands. Some of these
have suggestions for the student to follow. Finally, the student will find a
glossary of important scientific and technical terms and a list of exercises, whose
complete answers are presented in a section at the end of the book.

DOI 10.1007/978-3-319-41093-7_3
114 3 The Schrödinger Equation
3.1 Operators
An operator is a rule that transforms a given function into another function. For
example, the derivative of f(x) results from applying the operator d/dx to f(x),

0 d
f ðxÞ ¼ f ðxÞ ð3:1Þ
dx
^ and B
Two operators A ^ are said to be equal if
^ ¼ Bf
Af ^ ð3:2Þ
for all functions f. The composition of two operators is defined by

^ Bf
A ^ Bf
^ A ^ ð3:3Þ
where the operator on the right is first applied to f, and the operator on the left is
applied to the function in parentheses. The composition of operators is associative:

^ B
A ^ ¼ A
^C ^B ^
^ C ð3:4Þ
However, the composition of operators is not necessarily commutative, unlike

constants and functions, whose product is always commutative. The commutator
of two operators is defined by

^ B
A; ^ A^B
^B
Â^ ð3:5Þ
For two commuting operators, their commutator is equal to the null operator,
and the order of application of both operators is irrelevant. The operator d/dx and a
constant c commute,

d d d d d
cf ðxÞ ¼ c f ðxÞ ) ;c ¼ cc ¼0 ð3:6Þ
dx dx dx dx dx
but the operators d/dx and x do not commute,

d d d d d
½x f ðxÞ ¼ f ðxÞ þ x f ðxÞ ) ;x ¼ xx ¼1 ð3:7Þ
dx dx dx dx dx
A linear operator satisfies the following equalities:
^ ½f ðxÞ þ gðxÞ ¼ Af
A ^ ðxÞ þ AgðxÞ
^ ^ ½c f ðxÞ ¼ c A
A ^ f ðxÞ ð3:8Þ
3.1 Operators 115
where f and g are arbitrary functions and c is an arbitrary constant. Examples of

linear operators are x2, d/dx, d2/dx2. Operators like cos and log are not linear.
3.1.1 Eigenvalues and Eigenfunctions
When the result of applying a linear operator to a function f is a constant times the
same function,
^ ¼ af
Af ð3:9Þ
we say that f is an eigenfunction, a is an eigenvalue, and (3.9) is the eigenvalue

equation (eigen, from German adjective meaning characteristic). For example, e3x
is an eigenfunction of the operator d/dx with eigenvalue equal to 3,

d 3x
e ¼ 3e3x ð3:10Þ
dx
(E1, E2).
3.1.2 One-Dimensional Schrödinger Equation
The classical form of the kinetic energy in one dimension x is
p2x
ð3:11Þ
2m
where px is the linear momentum along the x-axis, whose corresponding quantum
mechanical operator is given by
h @
^px ¼ ð3:12Þ
i @x
(E3, E4).
The time-independent Schrödinger equation is given by the eigenvalue equation
^ k ¼ Ek wk
Hw ð3:13Þ
where the operator is called the Hamiltonian (the energy operator), wk is the
eigenfunction, and Ek is the corresponding energy eigenvalue. For a system con-
sisting of one particle in one dimension, the Hamiltonian is given by
^ ¼ h d þ VðxÞ
2 2
H ð3:14Þ
2m dx2
where the first term is the kinetic energy operator (E5), the second term is the
potential energy operator, and
h
h ¼ ð3:15Þ
2p
Multiplication of (3.13) by the complex conjugate of wk followed by integration
over all coordinate space leads to
Zþ 1 Zþ 1
^ k dx
wk Hw ¼ Ek wk wk dx ¼ Ek ð3:16Þ
1 1
where
Zþ 1 Zþ 1
wk wk dx ¼ jwk j2 dx ¼ 1 ð3:17Þ
1 1
The square of the absolute value of the wave function, |wk(x)|2, is the probability
density for finding the particle at state wk and point x, and the integral of |wk(x)|2
over the entire domain of x is 1, since wk is normalized. The physical meaning of
|wk|2 as a probability density implies that wk is single-valued and continuous, has
continuous first derivatives, and is quadratically integrable, that is, the integral of
|wk|2 is finite and so can be normalized. A function that meets these requirements is
called a well-behaved function.
The Schrödinger equation cannot be derived any more than the Newton equation of
classical mechanics can be derived. In fact, these equations cannot be derived from
first principles, since they express first principles for the mechanical behavior of
macroscopic bodies and subatomic particles. In spite of that, we can imagine the line
of thought followed by Schrödinger to arrive at his famous and less intuitive equation.
To this end, the classical wave equation is combined with the De Broglie expression
that connects linear momentum, a particle characteristic, with wavelength, a wave
feature. We begin by writing the following one-dimensional wave equation,
@ 2 uðx; tÞ 1 @ 2 uðx; tÞ
¼ ð3:18Þ
@x2 v2 @t2
where u is any twice differentiable function, and v represents the speed of the
propagating wave. The Mathematica code M1 solves this differential equation,
showing that its general solution is a linear combination of any twice differentiable
functions of arguments x − vt and x + vt. Functions with these arguments are called
waves and propagate to the right or forward direction (argument x − vt) and to the
left or backward direction (argument x + vt) (E6).
A propagating wave conveys two kinds of periodicity: space periodicity in the
x direction with wavelength k as its characteristic parameter, and time periodicity with
3.1 Operators 117
period s, the inverse of frequency m, as its characteristic parameter. For a forward

wave, we can take x/k − t/s, since both x/k and t/s are dimensionless quotients and the
minus sign indicates a forward-propagating wave. A sine or cosine wave is called a
harmonic wave. The units of the argument of a sine function are radians, so x/k − t/s
should be multiplied by 2p, giving 2p (x/k − t/s), where 2p /k is the angular
wavenumber k and 2p/s = 2pm is the angular frequency x. Therefore, a sine for-
ward wave is given by Asin(kx − xt), where A is the wave amplitude (E7).
Assuming separation of variables x and t and expressing the time-dependence of
u(x, t) by a cosine function, we can write
uðx; tÞ ¼ wðxÞ cos xt ð3:19Þ
Substitution of (3.19) in (3.18) gives
d2 wðxÞ
þ k2 wðxÞ ¼ 0 ð3:20Þ
dx2
where v = mk and k is the angular wavenumber. The Mathematica code M2 solves

this differential equation, known as the Helmholtz equation (see also §1). We now
consider the particle energy as a sum of the kinetic and potential energies,
p2x
E¼ þ VðxÞ ð3:21Þ
2m
and use the De Broglie equation
kpx ¼ h ð3:22Þ
Substitution of (3.21) and (3.22) in (3.20) gives the one-dimensional

Schrödinger equation
h2 d2 wð xÞ
þ V ð xÞwð xÞ ¼ Ewð xÞ ð3:23Þ
2m dx2
(E8, E9, E10).
3.1.3 Hermitian Operators
Consider now an arbitrary well-behaved function f and a set g1, g2, …, gi,… of
well-behaved functions that obey the same boundary conditions of f and satisfy the
equality
X
f ¼ c i gi ð3:24Þ
i
where the ci are coefficients. The set of functions g1, g2, …, gi, … is said to be a
complete set of functions and can be used to expand any arbitrary function that
obeys the same boundary conditions of the functions gi.
Now consider a linear operator Â that represents the physical quantity A. When
the system is described by the state function w, the average value of Â is given by
Z
h Ai ¼ ^
w Awds ð3:25Þ
where the integral is over the full range of coordinates of the wave function. The
average value of a physical quantity is a real number, so
h Ai ¼ h A i ð3:26Þ
and
Z Z

^
w Awds ¼ ^
w Aw ds ð3:27Þ
A linear operator that satisfies this equality for all well-behaved functions is
called a Hermitian operator. Replacing w by one eigenfunction of the Hermitian
operator, one easily concludes that the eigenvalues of a Hermitian operator are real
numbers. The above definition of Hermitian operator can be shown to be equivalent
to the following,
Z Z

^
f Agds ¼ ^ ds
g Af ð3:28Þ
for arbitrary well-behaved functions f and g (E11).
3.1.4 Important Theorems
We are now going to prove a few important theorems in quantum mechanics.

Theorem 1 The eigenvalues of a Hermitian operator are real numbers.
^ and its general eigenvalue equation

Consider the Hermitian operator A
^ k ¼ ak gk
Ag ð3:29Þ
Using the definition of Hermitian operator, we write

Z Z

^ k ds
gk Ag ¼ ^ k ds
gk Ag ð3:30Þ
3.1 Operators 119

Z

ak ak gk gk ds ¼ 0 ð3:31Þ
Since gk is a quadratically integrable function, we have
ak ¼ ak ð3:32Þ
that is, the eigenvalue ak is real. Therefore, the eigenvalues of a Hermitian operator
are real numbers.
Theorem 2 Two linear operators with a common complete set of eigenfunctions
commute.
If two linear operators have a common complete set of eigenfunctions g1, g2, …,
gi, …,
^ i ¼ ai gi
Ag ^ i ¼ bi gi
Bg ð3:33Þ
where ai and bi are eigenvalues, then they commute, i.e., their commutator is the
null operator, and consequently,

A^B
^B
Â^ f ¼0 ð3:34Þ
for every arbitrary well-behaved function f. To arrive at (3.34), express f as a linear

combination of the complete set of functions gi [see (3.24)] and use these as
eigenfunctions common to both operators [see (3.33)].
Theorem 3 If two Hermitian operators commute, then there exists a common
complete set of eigenfunctions for them.
^ and
In order to demonstrate this theorem, we consider two Hermitian operators A
^ and write the eigenvalue equation for A,
B ^
^ i ¼ ai gi
Ag ð3:35Þ
^ to both members, we obtain

If we now apply B

B ^ i ¼B
^ Ag ^ ðai gi Þ ¼ ai Bg
^ i ð3:36Þ
Being Hermitian, B ^ and B

^ is linear, and this justifies the last equality. Since A ^
commute, we can write

^ Bg
A ^ i ¼ ai Bg
^ i ð3:37Þ
^ with eigenvalue ai. Assuming

that is, the function in parentheses is an eigenfunction of A
^
that the eigenvalues of A are nondegenerate (degenerate eigenvalues are equal), then
each eigenvalue corresponds to one and only one linearly independent eigenfunction.
Therefore, gi and Bg ^ i are linearly dependent functions, i.e., they are proportional,
^ i ¼ bi gi
Bg ð3:38Þ
and consequently gi is an eigenfunction of B: ^ We have just proved that the

^
eigenfunctions of A are also eigenfunctions of B ^ when the eigenvalues of A ^ are
nondegenerate.
Now suppose that the eigenfunctions of A ^ are r-fold degenerate, that is, there are
^
r eigenfunctions of A with the same eigenvalue. While any linear combination of
these eigenfunctions is an eigenfunction of A^ with the same eigenvalue (this can be
easily proved; try the twofold degenerate case, for example), it is not certain that the
chosen eigenfunctions of A ^ are also eigenfunctions of B:
^ Linear combinations of the
^ ^
eigenfunctions of A that are eigenfunctions of B diagonalize the matrix of B, ^ since
this matrix diagonalization finds linear combinations of the basis functions that
correspond to eigenfunctions of B. ^ A complete derivation of Theorem 3 can be
found in Merzbacher, Further Reading, Sect. 10.4.
Theorems 2 and 3 allow us to conclude that the commutativity of two linear
operators is a necessary and sufficient condition for having a common complete set
of eigenfunctions. Hermitian operators represent observables. Hence, when two
Hermitian operators A ^ and B^ commute, they share a common complete set of
eigenfunctions. Hence, the corresponding eigenvalues ai and bi are definite and
sharp values that can be simultaneously assigned to observables A and B.
Theorem 4 Two eigenfunctions of a Hermitian operator that correspond to dif-
ferent eigenvalues are orthogonal.
^ and the eigenvalue equations corresponding

Consider the Hermitian operator A
to different eigenvalues,
^ ¼ af
Af ^ ¼ bg
Ag ð3:39Þ

Z
ðb aÞ f gds ¼ 0 ð3:40Þ
We conclude that f and g are orthogonal, since b 6¼ a.

Theorem 5 Two independent eigenfunctions of a Hermitian operator that corre-
spond to the same eigenvalue (degenerate eigenvalue) may not be orthogonal, but
they can always be replaced by linear combinations that become orthogonal.
3.1 Operators 121
^ that
Consider two independent eigenfunctions of the Hermitian operator A
correspond to the same eigenvalue (degenerate eigenvalue),
^ ¼ sf
Af ^ ¼ sg
Ag ð3:41Þ
If we now assume that f and g are not orthogonal, we can take f and g + cf (note
that g + cf is also an eigenfunction of Â with the same eigenvalue s) and choose
c such that these two independent eigenfunctions become orthogonal, i.e.,
Z R
f gds
f ðg þ cf Þds ¼ 0 ) c ¼ R ð3:42Þ
f f ds
In conclusion, a Hermitian operator has an orthonormal set of eigenfunctions

whose eigenvalues are real numbers. This set of functions can be used to expand
any quadratically integrable and well-behaved function with the same boundary
conditions of the eigenfunctions; i.e., it is a complete set of functions.
3.1.5 Dirac Notation
A notation introduced by Dirac, very useful in quantum mechanics, uses the fol-
lowing symbolism:
Z Z

f Agds ¼ f
A
^ ^
g f gds ¼ hf jgi ð3:43Þ
where the integrals are over the full range of coordinates of the wave functions. We
have
Z Z

f gds ¼ fg ds ð3:44Þ
so
hf jgi ¼ hgjf i ð3:45Þ
3.2 Harmonic Oscillator
The atoms of a molecule oscillate about their equilibrium positions. For an isolated
molecule (a molecule in the gas phase at low pressure), suitable spectroscopic
methods show that these oscillations have well-defined frequencies. In order to
understand the vibrations of molecules, we begin by separating the translations and
rotations. For a diatomic molecule AB, the motions of atoms A and B involve six
Fig. 3.1 In the diatomic

molecule AB, the motions of
atoms A and B can be
described using two reference
systems, one fixed in the
laboratory (xlab, ylab, and zlab
axes), the other locked to the
molecule with one of its axes
coincident with the
internuclear axis (C is the
center of mass)
Cartesian coordinates, three for each atom. Let us consider two systems of refer-
ence, one fixed in the laboratory, the other locked to the molecule, with one of its
axes coincident with the internuclear axis (Fig. 3.1). The motions of the atoms can
now be separated into the center of mass motion described by three translations,
corresponding to the components of the vector R(t) along the xlab, ylab, and zlab axes,
two rotations, around two orthogonal axes perpendicular to the internuclear axis
(for a linear molecule, rotation around the internuclear axis does not lead to
physically distinguishable configurations and so is not considered), and one
vibration, described by the variation in the internuclear distance between atoms A
and B (Fig. 3.1).
A molecule with N atoms needs 3N Cartesian coordinates as functions of time
(three Cartesian coordinates per atom) to fully describe the motions of its atoms.
Then, 3N coordinates include three coordinates for the center of mass motion
(molecular translations), three coordinates for the rotations of a nonlinear molecule
around three orthogonal axes, or two coordinates for the rotations of a linear
molecule, and 3N − 6 or 3N − 5 coordinates for the vibrations of a nonlinear or a
linear molecule, respectively.
3.2.1 Reduced Mass
Consider now a diatomic molecule AB as a system of two point masses defined by

the vectors rA and rB. The total kinetic energy is the sum of the kinetic energies of
atoms A and B, that is,
3.2 Harmonic Oscillator 123
2 2
1 drA 1 drB
T ¼ mA þ mB ð3:46Þ
2 dt 2 dt
If we now consider the molecule AB with its center of mass defined by the
vector R, then we can write
RM ¼ rA mA þ rB mB ð3:47Þ
where
M ¼ mA þ mB ð3:48Þ
The stretching vibration is described by the following coordinate:
r ¼ rB rA ð3:49Þ
Substitution of (3.47), (3.48), and (3.49) in (3.46) leads to

2
1 dR 1 dr 2
T¼ M þ l ð3:50Þ
2 dt 2 dt
where l is called the reduced mass, which is defined by the equality
1 1 1
þ ð3:51Þ
l mA mB
(E12, E13).
3.2.2 Classical Treatment
In the classical one-dimensional harmonic oscillator model, a single point mass

m is connected to a rigid wall by a perfectly elastic spring (Fig. 3.2, left). The rigid
wall is equivalent to an infinite mass. When the perfectly elastic spring is distended
or contracted, the force that restores equilibrium is proportional to the shift from the
equilibrium position,
F ¼ kx ð3:52Þ
where x(t) = s(t) − se, with se representing the equilibrium position. The propor-
tionality constant k is called a force constant. Use of the Newton equation in (3.52)
gives
d2 x
m ¼ kx ð3:53Þ
dt2
Fig. 3.2 (Left) A single mass linked to a rigid wall by a spring. (Right) Two masses linked by a
spring (R is the center of mass of the system)
This is the differential equation for the classical one-dimensional harmonic

oscillator. According to (3.53), the second derivative of x(t) is proportional to x(t),
with the proportionality constant being equal to −k/m. Therefore, a sine or a cosine
function or a linear combination of them satisfies (3.53). Note that the sine and
cosine functions differ by a phase shift of p/2, and the following sine function,
x ¼ A sinðxt þ bÞ ð3:54Þ
is equivalent to a linear combination of sine and cosine functions, where x is the

angular frequency of the oscillator (x = 2pm; units radian s−1), while A and b are
constants that result from integration of (3.53) (E14). The maximum and minimum
values of x (A and −A) correspond to the maximum and minimum of the sine
function in (3.54) (1 and −1). Substitution of (3.54) in (3.53) leads to
rffiffiffiffi
k
x¼ ð3:55Þ
m
3.2.3 Quantum-Mechanical Treatment
We now write
dV
F¼ ð3:56Þ
dx
where V is the potential energy. Substitution of (3.56) in (3.52) gives
dV
¼ kx ð3:57Þ
dx
whose integration leads to

1
V ¼ kx2 ð3:58Þ
2
on the assumption that the integration constant is zero (the potential energy always
has an arbitrary additive constant, so we can take it as zero). The Taylor expansion
of the potential energy around the potential energy minimum, Vmin, is given by

dV 1 d2 V
V ¼ Vmin þ þ x2 þ ð3:59Þ
dx min 2 dx2 min
The first term can be chosen to be zero, and the second term is zero because the
minimum is a stationary point, and the slope at a stationary point is zero. Therefore,
we can write, after taking (3.58) and (3.59) into account,

d2 V
¼k ð3:60Þ
dx2 min
that is, the curvature of the potential energy curve in the minimum is equal to the
harmonic oscillator force constant. Terms of order higher than x2 account for an-
harmonicity and are not included in the harmonic oscillator model.
The one-dimensional Schrödinger equation for the harmonic oscillator [(3.23)
with the potential energy given by (3.58)] is
h2 d2 wv ð xÞ 1 2
þ kx wv ð xÞ ¼ Ev wv ð xÞ ð3:61Þ
2m dx2 2
where v is the vibrational quantum number (v = 0, 1, 2, …). After x is substituted

by z using x = z/√a with a mx/h, the eigenfunctions for (3.61) become expressed
as
wv ¼ Nv ez =2 Hv ðzÞ
2
ð3:62Þ
where the normalization constants are given by

1=4
1 1
Nv ¼ pffiffiffiffiffiffiffiffi ð3:63Þ
p 2v v!
(see Mathematica code for obtaining these normalization factors in M3), and Hv are
the polynomial factors known as Hermite polynomials (E15, E16), defined by
2 dv z2
Hv ðzÞ ð1Þv ez e ð3:64Þ
dzv
These polynomials satisfy the recurrence formula

Hv þ 1 ðzÞ ¼ 2zHv ðzÞ 2vHv1 ðzÞ ð3:65Þ
(E17), and the first five are given by
H0 ¼ 1
H1 ¼ 2z
H2 ¼ 4z2 2
H3 ¼ 8z3 12z
H4 ¼ 16z4 48z2 þ 12 ð3:66Þ
Figure 3.3 shows the squares of the harmonic oscillator eigenfunctions (3.62)
(probability density curves), for v from 0 to 5 (see Mathematica code in M3). The
number of nodes in each eigenfunction equals the value of the vibrational quantum
number.
If we now substitute back z by x using z = √ax, we obtain the first four harmonic
oscillator normalized wave functions expressed in the original x variable:
a1=4
eax =2
2
w0 ¼
p
a1=4 1 pffiffiffi
pffiffiffi eax =2 2 ax
2
w1 ¼
p 2
a1=4 1
pffiffiffi eax =2 4ax2 2
2
w2 ¼
p 8
a1=4 1 h pffiffiffi pffiffiffi i
pffiffiffiffiffi eax =2 8 a x3 12 ax
2 3
w3 ¼ ð3:67Þ
p 48
Fig. 3.3 Probability density curves for the harmonic oscillator as functions of z. This figure was
obtained with Mathematica
pffiffiffiffiffiffiffiffiffiffiffiffi
where the normalized factors are (a/p)1/4 ½1= ð2v v!Þ (see Mathematica code in M3
for verifying these normalizing factors and confirming that these wave functions are
orthogonal).
The existence of boundary conditions for the wave functions (these must go to
zero as x goes to infinity) implies that Ev = (v + 1/2)hx, where v = 0, 1, 2, …, that
is, the energies of the vibrational levels depend linearly on the vibrational quantum
number v, and the separation between consecutive energy levels is hx. At 0 K, the
harmonic oscillator quantum number is zero (v = 0); the harmonic oscillator energy
is called the zero-point energy and is equal to hx=2. For a molecule as a collection
of harmonic oscillators, the total zero-point energy is the sum of all zero-point
energies.
3.2.4 Morse Potential
The single vibrational mode of a diatomic molecule is the stretching of its inter-
nuclear distance. An empirical representation of the potential energy function for a
diatomic molecule, more realistic than the harmonic oscillator, is the so-called
Morse potential energy function
h i2
UðRÞ ¼ D 1 ebðRRe Þ ð3:68Þ
The Taylor expansion of this function about the equilibrium point Re is
ðR Re Þ2 00 ðR Re Þ3 000
UðRÞ ¼ UðRe Þ þ ðR Re ÞU 0 ðRe Þ þ U ðRe Þ þ U ðRe Þ þ
2! 3!
ð3:69Þ
where
2
dU dU
UðRe Þ ¼ 0 ¼0 ¼k
dR Re dR2 Re
rffiffiffiffiffiffi
d3 U k d4 U 7k2
¼ 3k ¼ ð3:70Þ
dR3 Re 2D 4
dR Re 2D
where k = 2Db2 (E18). Substitution of (3.70) in (3.69) leads to

rffiffiffiffiffiffi
k k k 7k2
UðRÞ ¼ ðR Re Þ2 ðR Re Þ3 þ ðR Re Þ4 þ ð3:71Þ
2 2 2D 48D
The first nonzero term is the harmonic term. Cubic, quartic, and following terms
account for the Morse potential anharmonicity.
Fig. 3.4 Morse function fitted to the points obtained by scanning the bond distance of F2. The
parabola represents the harmonic oscillator potential energy curve. This figure was obtained with
Mathematica
Figure 3.4 shows the potential energy curves for the harmonic oscillator
(parabola) and the Morse function fitted to the points in black obtained by deter-
mining the potential energy of the system of two fluorine atoms at predefined
internuclear distances (scan of the bond distance for F2, Gaussian 09 B3LYP/
cc-pVTZ calculation; for now, the student should simply consider this a reliable and
accurate calculation for its purpose, leaving the details of the level of calculation
and the basis of functions to be understood later on, in Sect. 3.15). The minimum of
the parabola was set to be coincident with the minimum of the potential energy
curve for F2. Both curves (Morse function and parabola) have the same curvature at
their minima. It can be seen that the harmonic oscillator deviates increasingly from
the calculated potential energy curve for F2 as the amplitude of vibration increases.
Large-amplitude vibrations such as the stretching vibrations involving light atoms
as in H2, CH, NH, OH, FH, F2, have important anharmonic corrections. Unlike the
harmonic oscillator curve, the Morse potential function can be well fitted to the
points of the potential energy curve of F2.
3.3 Spherical Coordinates
The potential energy function for the electron of a hydrogen atom is a function of
the distance from the electron at point (x, y, z) to the origin at the nucleus, and thus a
spherically symmetric function. One says that the hydrogen atom is an example of a
central force problem. The set of coordinates adequate to a central force problem
are spherical coordinates, (r, h, /), where r is the distance from the origin to the
point (x, y, z) (0 r < ∞), h is the angle between the positive z-axis and the
r vector to the point (0 h p), and / is the angle between the positive x-axis
3.3 Spherical Coordinates 129
Fig. 3.5 Cartesian and

spherical coordinates
with the projection of the r vector in the xy-plane (0 / < 2p) (Fig. 3.5). Right
triangle trigonometry can easily lead to the following equalities:
x ¼ r sin h cos / y ¼ r sin h sin / z ¼ r cos h ð3:72Þ
From these expressions, one can obtain

z y
r ¼ ðx2 þ y2 þ z2 Þ1=2 cos h ¼ tan / ¼ ð3:73Þ
ðx2 þ y2 þ z2 Þ1=2 x
In Cartesian coordinates, the infinitesimal volume corresponds to a paral-

lelepiped with edges equal to dx, dy, and dz, that is, dv = dxdydz. In spherical
coordinates, two of the edges of the infinitesimal volume are curved, since they are
associated with angular coordinates.
We begin by considering h and / constant. The intersection of the horizontal
circle marked in Fig. 3.6 (r, h constants) with the vertical semicircle that contains
the z-axis and the projection of the r vector on the xy-plane (r, / constants) is a line
segment of length r (Fig. 3.6). The length of the infinitesimal edge between r and
r + dr (not shown in Fig. 3.6) is given by dr,
h; / ¼ constant r ! r þ dr edge1 ¼ dr ð3:74Þ
We now consider r and / constant. The intersection of the spherical surface

(r constant) with the vertical plane that contains the z-axis and the projection of the
r vector on the xy-plane (/ constant) is a vertical semicircle marked by a continuous
line in Fig. 3.6. An infinitesimal arc between h and h + dh (not shown in Fig. 3.6)
is given by rdh,
r; / ¼ constant h ! h þ dh edge2 ¼ rdh ð3:75Þ

Fig. 3.6 The vertical

semicircle corresponds to
constant r and / variables,
whereas the horizontal circle
corresponds to constant r and
h
The last edge of the infinitesimal volume in spherical coordinates is obtained by

considering r and h constant. The intersection of the spherical surface (r constant)
with the conical surface (h constant) is a horizontal circle marked in Fig. 3.6. An arc
of this circle between / and / + d/ (not shown in Fig. 3.6) is given by rsinhd/
(r sinh is the projection of the r vector on the xy-plane),
r; h ¼ constant / ! / þ d/ edge3 ¼ r sin hd/ ð3:76Þ
Thus, the infinitesimal volume in spherical coordinates is given by the product

edge1 edge2 edge3, which, after substitution of (3.74), (3.75), and (3.76),
leads to
dvsph ¼ r 2 sin hdrdhd/ ð3:77Þ
3.4 Angular Momentum
In classical mechanics, the motion of one particle around a center of force can be
characterized by two kinds of momenta, linear and angular momenta. Let r be the
vector from the center of force to the particle position,
r ¼ ix þ jy þ kz ð3:78Þ
and v the corresponding velocity vector,

3.4 Angular Momentum 131
dr dx dy dz
v ¼ i þj þk ð3:79Þ
dt dt dt dt
The classical definitions of the linear and angular momenta are given by
p mv ð3:80Þ
and
Lrp ð3:81Þ
i.e.,

i j k
L ¼
x y z
ð3:82Þ

px py pz
where the components of L along the x, y, and z axes are
Lx ¼ ypz zpy Ly ¼ zpx xpz Lz ¼ xpy ypx ð3:83Þ
(E19).
3.4.1 Orbital Angular Momentum
The quantum-mechanical operators of the linear momentum [see (3.12)] are given
by
@ @ @
^px ¼ ih ^py ¼ ih ^pz ¼ ih ð3:84Þ
@x @y @z
In turn, the angular momentum associated with one-electron motion around a

center of force is called orbital angular momentum. Substitution of (3.84) in
(3.83) leads to the following expressions for the orbital angular momentum
operators:

^ @ @ ^ @ @ ^ @ @
h y z
Lx ¼ i Ly ¼ ih z x Lz ¼ ih x y
@z @y @x @z @y @x
ð3:85Þ
From these operators we can arrive at

L^x ; L ^z
^y ¼ ihL L^y ; L ^x
^z ¼ ihL ^z ; L
L ^y
^x ¼ ihL ð3:86Þ
In order to derive these equalities, the above commutators should be applied to

an arbitrary wave function f, and the equality between mixed partial second
derivatives, ∂2f/∂x∂y = ∂2f/∂y∂x, should be used (E20). Note that (3.83), (3.85),
and (3.86) obey the circular symmetry x ! y ! z ! x.
The operator for the square of angular momentum magnitude is given by
L ^L
^2 ¼ L ^¼L
^2x þ L
^2y þ L
^2z ð3:87Þ
This operator commutes with each angular momentum component,

2 2 2
^ ;L
L ^x ¼ 0 L ^y ¼ 0
^ ;L ^ ;L
L ^z ¼ 0 ð3:88Þ
(E21). In conclusion, the square of the angular momentum operator of an isolated

atom commutes with each of its components. However, no two components
commute with each other. Therefore, the square of the angular momentum operator
and only one of its angular momentum components, for example the z-component,
have a common complete set of eigenfunctions. In spherical coordinates, the
angular momentum operators involve only the variables h and /. In particular,

^2 ¼ h2 @2 @ 1 @2
L þ cot h þ ð3:89Þ
@h2 @h sin2 h @/2
^z ¼ ih @
L ð3:90Þ
@/
where the last expression suggests a separation of the variables h and / in the
eigenfunctions. The one-electron eigenvalue equations and quantum numbers for
the square and the z-component of the angular momentum operators are given by
^2 Y‘m ðh; /Þ ¼ ‘ð‘ þ 1Þh2 Y‘m ðh; /Þ ‘ ¼ 0; 1; 2; . . .

L
ð3:91Þ
^z Y‘m ðh; /Þ ¼ mh Y‘m ðh; /Þ m ¼ ‘; ‘ þ 1; . . .; 0; . . .; ‘ 1; ‘
L
where the eigenfunctions

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2‘ þ 1ð‘ mÞ! m
Y‘m ðh; /Þ ¼ P ðcos hÞeim/ ð3:92Þ
4pð‘ þ mÞ! ‘
are called spherical harmonics and Pm ‘ are the associated Legendre polynomials
(note that in Y‘m and Pm
‘ , m is a superscript, not an exponent) (see §2). Concise and
useful definitions of associated Legendre polynomials, implemented in Mathe-
matica as LegendreP½‘; m; cos h, and spherical harmonics, implemented in
Mathematica as SphericalHarmonicY[‘,m,h,/], can be obtained by
querying Wolfram Alpha at www.wolframalpha.com. The interested reader can find
3.4 Angular Momentum 133
Fig. 3.7 Polar plots of associated Legendre polynomials P00 , P01 , and P02 , and three-dimensional
plots of their corresponding spherical harmonics, obtained with Mathematica
in Wolfram MathWorld, at mathworld.wolfram.com, more detailed mathematical

explanations on the associated Legendre polynomials and spherical harmonics. The
Mathematica code M4 presents the spherical harmonics expressions for ‘ 2 and
the spherical harmonics 3D-plots for ‘ 2 and m = 0.
The obvious way of plotting a function f(h) depicts h as an angle and |f(h)| as the
distance from the origin to the point [h, |f(h)|]. Such a plot is called a polar plot.
Since the distance from the point to the origin is the absolute value |f(h)|, the sign of
f(h) has to be separately marked over each region of space. Note that a polar plot
emphasizes the symmetry of the plotted function, not its shape. For example, the
polar plot for a constant is a circle, not a horizontal line. Figure 3.7 shows polar
plots of the associated Legendre polynomials
1
P00 ðcos hÞ ¼ 1 P01 ðcos hÞ ¼ cos h P02 ðcos hÞ ¼ ð3 cos2 h 1Þ
2
and three-dimensional plots of the corresponding spherical harmonics (E22, E23).
3.4.2 Spin
The yellow color imparted to a flame by sodium atoms results from emission from
the electronic configuration 1s22s22p63p1 to the ground state configuration
1s22s22p63s1. While this spectral feature of atomic sodium atoms is usually called
the sodium D line, it is in fact associated to two closely spaced spectral lines. In
1925, Uhlenbeck and Goudsmit explained this fine structure by proposing that the
electron has an intrinsic angular momentum called the spin angular momentum,
or more simply, spin, with two possible values of its projection along a preferred
axis such as the axis of an applied magnetic field (Fig. 3.8). Later, in 1928, Dirac
developed the relativistic quantum mechanics of the electron, where the electron
spin arises naturally. In nonrelativistic quantum mechanics, the spin must be con-
sidered an additional hypothesis. While the orbital angular momentum we have
Fig. 3.8 Possible

orientations of the spin vector
lie on the surface of cones
whose axes coincide with the
z-axis
presented above is classically associated to one-electron motion around a center of

force, the spin angular momentum has no classical mechanics equivalent.
The spin angular momentum operators are postulated to be Hermitian operators
and obey commuting relationships analogous to the orbital angular momentum
equivalents. For one electron, the spin eigenvalue equations and spin quantum
numbers are
^S2 w ¼ sðs þ 1Þh2 w s ¼ 1=2

ð3:93Þ
^Sz w ¼ ms hw ms ¼ 1=2; 1=2
where the +1/2 and −1/2 values of ms correspond to a and b spin eigenfunctions
that form a complete set of orthonormal functions, i.e., normalized and mutually
orthogonal functions,
Z Z Z Z
a adr ¼ b bdr ¼ 1 a bdr ¼ b adr ¼ 0 ð3:94Þ
Since the spin is a quantum concept without classical equivalent, (3.94) should
be seen as merely expressing formal equations.
3.5 Hydrogen Atom
The Hamiltonian for the hydrogen atom is given by
^ ¼ h r2 þ VðrÞ
2
H ð3:95Þ
2l
3.5 Hydrogen Atom 135
where
@2 @2 @2
r2 þ þ ð3:96Þ
@x2 @y2 @z2
is the Laplacian operator and l is the reduced mass of the electron–nucleus system
that satisfies the following equality:
1 1 1
þ ð3:97Þ
l me mN
For the lightest atom protium, the mass of the nucleus is about 2 103 times the
mass of the electron. Therefore, the second term on the right-hand side of (3.97) is
negligible, and l is approximately equal to the electron mass,
l me ð3:98Þ
The potential energy operator V(r) is given by
e2
VðrÞ ¼ ð3:99Þ
4pe0 r
where r is the electron–nucleus distance and e0 is the vacuum permittivity (see

Appendix).
In spherical coordinates, the Laplacian operator takes the form
@2 2@ 1 ^2
r2 ¼ þ 2 2L ð3:100Þ
@r 2 r @r r h
Therefore, the Hamiltonian for the hydrogen atom commutes with the square of
the angular momentum operator [see (3.89)] and with the z-component of the
angular momentum operator [see (3.90)], and we can write
^ L
½H; ^2 ¼ 0 ^ L
½H; ^z ¼ 0 ð3:101Þ
(E24). These equalities allow us to conclude that there is a complete set of

eigenfunctions common to the Hamiltonian for the hydrogen atom, the square of the
angular momentum operator, and its z-component. The square of the angular
momentum operator and its z-component do not involve the coordinate r [see (3.89)
and (3.90)], so the radial factors act like constants with respect to the angular
momentum operators. Hence, the eigenfunctions for the hydrogen atom are
obtained by multiplying radial functions by spherical harmonics,
wn‘m ¼ Rn‘ ðrÞY‘m ðh; /Þ ð3:102Þ

with the radial factors and spherical harmonics being labeled by the n, ‘, and
m quantum numbers. Alternatively, ‘ can be substituted by a letter, where the first
letters of the words of spectroscopic origin sharp, principal, diffuse, and funda-
mental stand for ‘ = 0, 1, 2, and 3, respectively. After the letter f, the alphabetical
order is followed, but j is omitted, and s and p are skipped because they have
already been used for ‘ = 0 and 1,
‘ 0 1 2 3 4 ...
s p d f g ...
Boundary conditions for the radial factors imply the introduction of the quantum
number n. For n 3 and the hydrogen atom (Z = 1) wave functions in atomic
units (unit of length is a0 = 0.52917721092 10−10 m; see Appendix), the radial
factors are given by
R1s ¼ 2er
1 1
R2s ¼ pffiffiffi ð2 rÞer=2 R2p ¼ pffiffiffi rer=2
2 2 2 6
2 2 r=3 4 4
R3s ¼ pffiffiffi ð27 18r þ 2r Þe R3p ¼ pffiffiffi ð6 rÞrer=3 R3d ¼ pffiffiffiffiffi r2 er=3
81 3 81 6 81 30
ð3:103Þ
(E25, E26).
The probability of finding the electron with spherical coordinates lying in the
ranges r to r + dr, h to h + dh, and / to / + d/ is given by

2
jwn‘m ðr; h; /Þj2 r 2 sin hdrdhd/ ¼ R2n‘ ðrÞr 2 dr
Y‘m ðh; /Þ
sin hdhd/ ð3:104Þ
The probability of finding the electron in the range r to r + dr, independently of

h and /, is given by
R2n‘ ðrÞr 2 dr ð3:105Þ
because the spherical harmonics are normalized functions, that is,
Z2p Zp

m

Y ðh; /Þ
2 sin hdhd/ ¼ 1 ð3:106Þ
‘
0 0
The jRn‘ ðrÞj2 r 2 are called radial distribution functions (Fig. 3.9).
3.6 Antisymmetry Principle 137
Fig. 3.9 Radial distribution functions for the 1s, 2s, and 2p hydrogen orbitals. Compare 2s and
2p and note the 2s lobes separated by a node. This graph was obtained with Mathematica
3.6 Antisymmetry Principle
In classical mechanics, each point of a particle trajectory is described by its position

and linear momentum coordinates (six coordinates). Identical particles can be
distinguished by distinct trajectories. However, electrons cannot be assigned indi-
vidual trajectories due to the uncertainty principle that states that position and
momentum cannot be simultaneously specified. Each electron of an atom or a
molecule is assigned four variables: three spatial coordinates that specify the
position vector (for example, x, y, z) and the spin variable that takes the value a for
ms = +1/2 or b for ms = −1/2. If we represent the set of these four variables by s,
the wave function for two electrons can be written as W(s1, s2). Consider now the
exchange or permutation operator that interchanges all four coordinates of electrons
1 and 2,
^ 12 Wðs1 ; s2 Þ ¼ Wðs2 ; s1 Þ
P ð3:107Þ
Applying this operator twice has no net effect, since the initial configuration is
restored,

P ^ 12 P
^ 212 Wðs1 ; s2 Þ ¼ P ^ 12 Wðs2 ; s1 Þ ¼ Wðs1 ; s2 Þ
^ 12 Wðs1 ; s2 Þ ¼ P ð3:108Þ
Therefore, the eigenvalue of the square of the permutation operator is 1, and

consequently, the eigenvalue of the permutation operator is either +1 or −1.
However, only −1 is consistent with experimental data for electrons, that is,
^ 12 Wðs1 ; s2 Þ ¼ Wðs1 ; s2 Þ
P ð3:109Þ
Combining (3.107) and (3.109) leads to
Wðs2 ; s1 Þ ¼ Wðs1 ; s2 Þ ð3:110Þ
and the wave function is said to be antisymmetric with respect to the interchange
of coordinates of two electrons including the spin coordinates (antisymmetry
principle). If s2 = s1, then
Wðs1 ; s1 Þ ¼ Wðs1 ; s1 Þ ð3:111Þ
and consequently,
Wðs1 ; s1 Þ ¼ 0 ) jWðs1 ; s1 Þj2 ¼ 0 ð3:112Þ
Therefore, two electrons with the same spin have zero probability of being found
at the same point. Since the wave function is a well-behaved function and so is
continuous, the probability of finding two electrons with the same spin in the same
region of space should be small. This consequence of the antisymmetry principle is
called the Pauli exclusion principle.
How to build a many-electron wave function W from one-electron functions, so
that W obeys the antisymmetry requirement? First of all, we take each one-electron
function as being obtained by the product of a spatial factor and one of the spin
functions a or b. For electron 1, we can write
wa ðs1 Þ ¼ /a ðr1 Þrð1Þ ð3:113Þ
where r can be either a or b. Each one-electron spatial factor is called an orbital,

and the product of a spatial factor and the corresponding spin factor is called a
spin-orbital. In calculations, the same spatial factor is usually combined with
different spin factors a and b.
Consider now a three-electron system and the product of three spin-orbitals,
wa ð1Þwb ð2Þwc ð3Þ ð3:114Þ
In order to make electrons indistinguishable, all permutations of the electron

labels must be considered. In addition, for each permutation of the coordinates of
two electrons, the sign must be changed, and so we obtain
wa ð1Þwb ð2Þwc ð3Þ wa ð2Þwb ð1Þwc ð3Þ þ wa ð2Þwb ð3Þwc ð1Þ

ð3:115Þ
wa ð3Þwb ð2Þwc ð1Þ þ wa ð3Þwb ð1Þwc ð2Þ wa ð1Þwb ð3Þwc ð2Þ
The one-electron functions wi are orthonormal, so integration of the square of the

absolute value of (3.115) gives 1 + 1 + 1 + 1 + 1 + 1 = 6, and the normalizing
pffiffiffi pffiffiffiffi
constant is 1= 6 ¼ 1= 3!. Changing the simple product wave function (3.114)
into a normalized antisymmetric wave function W can be formally represented by

the operator A with the normalization constant included, where
W ¼ A½wa ð1Þwb ð2Þwc ð3Þ ð3:116Þ
Hence, applying A to a simple product of n one-electron orthonormal functions

gives an antisymmetric linear combination of n! permutations with the normal-
pffiffiffiffi
ization constant 1= n!. The resulting normalized function W is antisymmetric
with respect to the exchange of a pair of electron coordinates, thus satisfying the
antisymmetry requirement.
An efficient way of obtaining all permutations of the electron coordinates con-
sists in taking the normalized “determinant” of one-electron functions called a
Slater determinant. A 3 3 Slater determinant is given by

w ð1Þ wb ð1Þ wc ð1Þ
a
W¼ pffiffiffiffi
wa ð2Þ wb ð2Þ wc ð2Þ
ð3:117Þ
3!
w ð3Þ w ð3Þ wc ð3Þ
a b
Permuting a pair of electron coordinates is equivalent to exchanging lines in the

Slater determinant (each line of the Slater determinant corresponds to a particular
electron coordinate). For instance, permuting rows 1 and 2 leads to



w ð1Þ w ð1Þ w ð1Þ
¼
ð3:118Þ

a b c

a b c


a b c a b c
[note that normalizing constants have been omitted in (3.118)]. In addition, two
equal columns make the Slater determinant equal to zero, since

wa ð1Þ wa ð1Þ wc ð1Þ

w ð2Þ wa ð2Þ wc ð2Þ
¼ 0 ð3:119Þ

a

w ð3Þ wa ð3Þ wc ð3Þ
and the corresponding probability density becomes zero. Therefore, no two elec-
trons of an atom or a molecule can have the same spin-orbital. This conclusion
shows that the Pauli exclusion principle is a corollary of the antisymmetric behavior
of the wave function.
If columns 1 and 3 of the Slater determinant in (3.117) are substituted by
normalized orthogonal combinations of these columns, the resulting determinant
remains the same or simply changes its sign (the corresponding probability density
is not altered),

w ð1Þ þ w ð1Þ

a pffiffi c w ð1Þ wa ð1Þw pffiffi c
ð1Þ

b


w ð2Þ þ w ð2Þ
2 2

a pffiffi c wb ð2Þ wa ð2Þw pffiffi c
ð2Þ
¼
w ð2Þ w ð2Þ wc ð2Þ
ð3:120Þ

2 2

a b

wa ð3Þ þ wc ð3Þ wa ð3Þwc ð3Þ


pffiffi w ð3Þ pffiffi
a b c
2 b 2
The Mathematica code M5 illustrates the above properties of determinantal

functions and confirms (3.118), (3.119), and (3.120) for 3 3 matrices and their
determinants.
Consider now the following Slater determinant for a two-electron atom or
molecule:

wa ðs1 Þ wb ðs1 Þ
1
Wðs1 ; s2 Þ ¼ pffiffiffiffi
¼ pffiffiffi ½wa ðs1 Þwb ðs2 Þ wa ðs2 Þwb ðs1 Þ ð3:121Þ
2! wa ðs2 Þ wb ðs2 Þ
2
Taking each spin-orbital as the product of a common spatial factor /a by a and

b, one obtains a closed shell configuration and the following wave function:

ðS¼0;MS ¼0Þ 1
/a ðr1 Það1Þ /a ðr1 Þbð1Þ
Wðs1 ; s2 Þ ¼ pffiffiffiffi
2! /a ðr2 Það2Þ /a ðr2 Þbð2Þ
ð3:122Þ
1
¼ /a ðr1 Þ/a ðr2 Þ pffiffiffi ½að1Þbð2Þ að2Þbð1Þ
2
Electrons 1 and 2 have antiparallel spins, so the sum Ms = ms1 + ms2 is zero.
Therefore, the total spin is zero (S = 0), and the number of distinct values of MS, or
electron-spin multiplicity (2S + 1), is equal to 1 (the state is called a singlet).
Consider now an open shell configuration with electrons 1 and 2 in different
orbitals, both with spins a. The corresponding Slater determinant is the product of
an antisymmetric product of spatial factors and the simple product of spin
functions:

1
/a ðr1 Það1Þ /b ðr1 Það1Þ
Wðs1 ; s2 ÞðS¼1;MS ¼1Þ ¼ pffiffiffiffi
2! /a ðr2 Það2Þ /b ðr2 Það2Þ
ð3:123Þ
1
¼ pffiffiffi ½/a ðr1 Þ/b ðr2 Þ /a ðr2 Þ/b ðr1 Það1Það2Þ
2
This electronic state corresponds to Ms = ms1 + ms2 = 1/2 + 1/2 = 1. If both

spin factors are given by b, then Ms = ms1 + ms2 = (−1/2) + (−1/2) = −1, and the
total wave function is

1
/a ðr1 Þbð1Þ /b ðr1 Þbð1Þ
Wðs1 ; s2 ÞðS¼1;MS ¼1Þ ¼ pffiffiffiffi
2! /a ðr2 Þbð2Þ /b ðr2 Þbð2Þ

ð3:124Þ
1
¼ pffiffiffi ½/a ðr1 Þ/b ðr2 Þ /a ðr2 Þ/b ðr1 Þbð1Þbð2Þ
2
The wave function that corresponds to Ms = ms1 + ms2 = 0 multiplies an

antisymmetric product of spatial factors by a linear combination of symmetric
products of spin functions and can be expressed by the following linear combi-
nation of Slater determinants:

1
Wðs1 ; s2 ÞðS¼1;MS ¼0Þ ¼ þ

2

ð3:125Þ
1 1
¼ pffiffiffi ½/a ðr1 Þ/b ðr2 Þ /a ðr2 Þ/b ðr1 Þ pffiffiffi ½að1Þbð2Þ þ bð1Það2Þ
2 2
Note that each individual determinant of (3.125) is not an eigenfunction of the

square of the spin angular momentum operator and so does not correspond to a pure
spin state. However, the linear combination of determinants (3.125) is an eigen-
function of the square of the spin angular momentum operator and is called a
spin-adapted wave function. Wave functions (3.123), (3.124), and (3.125) belong
to the same triplet (S = 1, 2S + 1 = 3). For more than two electrons, the wave
function cannot be factorized into spatial and spin coordinates.
3.7 Variational Method
Solving the Schrödinger equation for atoms or molecules with more than one
electron requires the use of approximate methods, the most powerful of these being
based on the variational theorem. Consider a system with a Hamiltonian whose
lowest-energy eigenvalue is E0, and a well-behaved function / that satisfies the
boundary conditions of the system. According to the variational theorem,
R
^
/ H/ds
R
E0 ð3:126Þ
/ /ds
The eigenfunctions wk of the Hamiltonian satisfy the eigenvalue equations
^ k ¼ Ek wk
Hw ð3:127Þ
and form a complete orthonormal set of functions that can be used to expand /,
X
/¼ ak wk ð3:128Þ
k
To prove (3.126), we substitute (3.128) into (3.126) and use (3.127) to obtain
R P P P
^
/ H/ds k aj ak Ek djk k a k a k Ek
W R ¼ P P ¼ P
j

E0 ð3:129Þ
/ /ds j k aj ak djk k ak ak
where djk is the Kronecker delta (djk = 0, for j 6¼ k, and djk = 1, for j = k) and the
inequality is justified because E0 is the lowest-energy eigenvalue.
If / is a linear combination of n linearly independent functions that satisfy the
boundary conditions of the system,
X
n
/¼ cj fj ð3:130Þ
j¼1
the set of functions fj is said to be a set of basis functions with the coefficients
being the variation parameters. Assuming real basis functions and real coefficients,
then
R P P
^
/ H/ds k cj ck Hjk
W¼ R ¼P P
j
ð3:131Þ
/ /ds j k cj ck Sjk
where
Z Z
Sjk ¼ fj fk ds Hjk ¼ ^ k ds
fj Hf ð3:132Þ
Note that W is a function of another function, in this case /. Thus, W is called a

functional of /, and the functional dependence is represented by a square-brackets
notation, W[/]. While a function is a rule that associates a number with another
number [for example, f(x) = 2x2 + 3 associates the number x = 3 with the number
2x2 + 3 = 21], a functional is a rule that associates a number (the functional value)
with a function.
Minimization of W[/] with respect to the coefficients can be achieved by solving
the system of equations
@W
¼0 ð3:133Þ
@cj
which is equivalent to
@K @D
W ¼0 ð3:134Þ
@cj @cj
where
XX XX
K¼ cj ck Hjk D¼ cj ck Sjk ð3:135Þ
j k j k
(E27). Illustration of the above equations with a set of two basis functions leads to
3.7 Variational Method 143
@K @D
@c1 ¼ 2c1 H11 þ 2c2 H12 @c1 ¼ 2c1 S11 þ 2c2 S12
@K @D ð3:136Þ
@c2 ¼ 2c1 H12 þ 2c2 H22 @c2 ¼ 2c1 S12 þ 2c2 S22
The Hamiltonian is Hermitian, and we have assumed real basis functions, so

H12 = H21 and S12 = S21. Substitution of (3.136) in (3.134) gives

ðH11 WS11 Þc1 þ ðH12 WS12 Þc2 ¼ 0
ð3:137Þ
ðH12 WS12 Þc1 þ ðH22 WS22 Þc2 ¼ 0
(E28, E29, E30). A nontrivial solution for this system of equations (the trivial
solution c1 = c2 = 0 corresponds to a null / everywhere) requires a vanishing
determinant of the coefficients, i.e.,

H11 WS11 H12 WS12

¼0 ð3:138Þ

H12 WS12 H22 WS22
Expansion of (3.138) gives an equation of second degree in W whose roots are

W(1) and W(2). The lowest root W(1) is an approximation to the lowest energy
eigenvalue. Substitution of W(1) in (3.137) gives a system of linear homogeneous
equations with the coefficients c(1) (1)
1 and c2 as unknowns. The determinant of the
coefficients is zero, so the equations are equivalent, that is, they differ by a constant
factor. Therefore, we have two unknowns and only one independent equation. This
system of equations is called underdetermined, and its solution gives the ratio of the
coefficients, c(1) (1)
2 /c1 . Normalization of /
(1)
enables one to determine c(1) (1)
1 and c2 .
A nontrivial solution for a system of n linear homogeneous equations in n
unknowns requires a zero determinant corresponding to (3.138) in the 2 2 case.
Being a system of homogeneous equations, the unknowns can be multiplied by an
arbitrary constant. This means that we are left with n equations and n − 1
unknowns, which are the ratios c2/c1, c3/c1, …, cn/c1 of the coefficients. In order to
solve for the values of all coefficients, we need to add a new equation that nor-
malizes /. The Mathematica code M6 illustrates this process of solving systems of
homogeneous linear equations with 2 2 and 3 3 coefficient matrices.
3.8 Born–Oppenheimer Approximation
Consider the nonrelativistic Hamiltonian for a molecule

0 1
B C
^ h2 X r2a
h2 X 2
e2 BBX X Za Zb X X Za XX 1C C
H¼ ri þ B þ C ð3:139Þ
2 a ma 2me i 4pe0 B a b [ a rab rai r C
|fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflffl} @ i a i j [ i ij A
|fflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflffl{zfflfflfflfflfflffl}
T^N Tê ^NN
V ^Ne
V êe
V
where a and b refer to nuclei, i and j to electrons. The first and second terms on the
right side of (3.139) are operators for the kinetic energies of the nuclei and elec-
trons, and the last three terms are potential energy operators for the nucleus–nucleus
repulsions, electron–nucleus attractions, and electron–electron repulsions. Benzene,
a typical medium-size molecule, has 12 nuclei and 42 electrons corresponding to 36
nuclear coordinates and 126 electron coordinates. Solving the Schrödinger equation
for this molecule is a formidable task. However, nuclei are much heavier than
electrons: one proton mass is about 1.8 103 times the electron mass, and one 12C
nucleus mass is about 2.2 104 times the electron mass. This suggests an
approximation whereby electron and nuclear coordinates are dealt with separately,
with the electrons carrying out their motions while the nuclei are fixed. The Born–
Oppenheimer approximation deals with the separation between electron and
nuclear coordinates in two steps. In the first step, the kinetic energy of the nuclei is
assumed to be zero and the Schrödinger equation
ê þ V
ðH ^NN Þwe ðsi ; Ra Þ ¼ UðRa Þwe ðsi ; Ra Þ ð3:140Þ
is solved, where
^ e ¼ Tê þ V
H ^Ne þ V
êe ð3:141Þ
is the electronic Hamiltonian, si represents the electron coordinates (spins inclu-

ded), and Ra stands for the nuclear position coordinates. Multiplying both members
of (3.140) by the complex conjugate of the wave function and integrating with
respect to the electron coordinates leads to the potential energy U(Ra), which
depends parametrically on the nuclear coordinates, U(Ra) = Ee + VNN(Ra), thus
becoming the potential energy for the nuclear motion. The second step of the Born–
Oppenheimer approximation considers the Schrödinger equation for the nuclear
motion
^ N wN ðRa Þ ¼ EN wN ðRa Þ
H ð3:142Þ
where
^ N ¼ T^N þ UðRa Þ
H ð3:143Þ
When (3.142) is solved using a molecule-fixed reference frame, the molecular

rotations and translations are not accounted for, and the energy eigenvalues EN
correspond to internal degrees of freedom, which are the molecular vibrations.
For a diatomic molecule,
0 2
^NN ¼ ZZ e
V ð3:144Þ
4pe0 R
3.8 Born–Oppenheimer Approximation 145
Fig. 3.10 A U(R) function for the ground electronic state of F2 obtained by scanning the
internuclear distance (Gaussian 09 B3LYP/cc-pVTZ calculation). The inset shows the zero-point
energy level marked with 0 and the first, second, and third stretching energy levels marked with 1,
2, and 3, respectively. This figure was obtained with Mathematica
where Z and Z′ are atomic numbers and R is the internuclear distance. The potential
energy curve U(R) is a function of R, and the difference between the energy at
infinite internuclear distance and the energy at the minimum Re is called the
equilibrium dissociation energy,
De ¼ Uð1Þ UðRe Þ ð3:145Þ
Figure 3.10 shows points obtained by solving the electronic Schrödinger

equation (3.140) for two fluorine atoms at defined values of the internuclear dis-
tance R (scan of R). These points were obtained by scanning the internuclear
distance R, and a curve was fitted to the points using a Morse potential to which
R−6, R−8, and R−10 terms were added. The inset presents vibrational levels obtained
by solving the nuclear Schrödinger equation (3.142), that is, by performing a fre-
quency calculation. The energy level 0 corresponds to the zero-point energy,
0.002392 Eh = 6.28 kJ mol−1, marked from the minimum of the curve (see the
appendix for conversion factors). The 1–0 transition corresponds to the harmonic
frequency 1049.88 cm−1 = 12.56 kJ mol−1. Levels 2 and 3 were drawn assuming
harmonic behavior, that is, the 2–1 and 3–2 transitions were made equal to the
harmonic frequency (anharmonicity was ignored). All these calculations were
carried out using the system of programs Gaussian 09, with the B3LYP method of
calculation and the cc-pVTZ basis set (the student can leave the technical details of
this calculation to Sect. 3.15).
3.9 Hartree–Fock Method
The electron–electron repulsion operator of an atomic or molecular Hamiltonian

includes the distance rij between any pair of electrons [see (3.139)]. This variable
makes the electron coordinates interdependent, since each electron position depends
on the positions of all other electrons due to instantaneous repulsions. The inter-
dependence of the electron coordinates implies that the coordinates of the electrons
cannot be separated, and consequently, one-electron wave functions (orbitals) are
approximations, not exact functions. Hartree (1897–1958) developed a method to
optimize one-electron functions as factors of many-electron simple product wave
functions. Later, Slater (1900–1976) and Fock (1898–1974) extended the method to
antisymmetric linear combinations of simple products (Slater determinants). In the
Hartree and Hartree–Fock methods, instantaneous electron–electron repulsions are
replaced by interactions between each electron and the average potential due to all
other electrons. This approximation enables the use of one-electron functions (or-
bitals) in the formation of many-electron wave functions for atoms and molecules.
Consider an atom with two electrons and the following simple product of spatial
orbitals:
wð0Þ ð1; 2Þ ¼ f ð0Þ ð1Þgð0Þ ð2Þ ð3:146Þ
where f and g are the products of radial factors by spherical harmonics, the former
being the variation functions. Zero superscripts indicate that the one-electron wave
functions are initial guesses, a concept that will soon be understood, and 1 and 2
inside parentheses stand for spherical coordinates (r1, h1, /1) and (r2, h2, /2) of
electrons 1 and 2. The average repulsion potential “felt” by electron 1 is obtained by
integrating the density of electron 2 over all its positions (integral over
dv2 ¼ r22 sin h2 dr2 dh2 d/2 ). In atomic units (see Appendix), this average potential is
given by
ð0Þ
2
ð0Þ g ð2Þ
V1 ð1Þ ¼ dv2 ð3:147Þ
r12
(E31). The above potential is a function of the coordinates of electron 1 only, since
the integration over dv2 removes the coordinates of electron 2, thus replacing
3.9 Hartree–Fock Method 147
instantaneous repulsions between electrons 1 and 2 by a fictitious repulsion between

the charge of electron 1 and the charge of electron 2 averaged over the entire
volume of the coordinates of electron 2. Further integration of (3.147) with respect
to the angular coordinates of electron 1, (h1, /1),
R 2p R p ð0Þ
ð0Þ 0 V1 ð1Þ sin h1 dh1 d/1
V1 ðr1 Þ ¼ 0
R 2p R p ð3:148Þ
0 0 sin h1 dh1 d/1
leads to a function of r1. Likewise, for electron 2, we can write
ð0Þ
2
ð0Þ f ð1Þ
V2 ð2Þ ¼ dv1 ð3:149Þ
r12
and
R 2p R p ð0Þ
ð0Þ 0 V2 ð2Þ sin h2 dh2 d/2
V2 ðr2 Þ ¼ 0
R 2p R p ð3:150Þ
0 0 sin h2 dh2 d/2
Having written the one-electron potentials, we can now proceed to write

one-electron Schrödinger equations that lead to improved approximations over
f(0)(1) and g(0)(2). In atomic units,

1 2 Z ð0Þ ð1Þ
r1 þ V1 ðr1 Þ f ð1Þ ð1Þ ¼ e1 f ð1Þ ð1Þ ð3:151Þ
2 r1

1 Z ð0Þ ð1Þ
r22 þ V2 ðr2 Þ gð1Þ ð2Þ ¼ e2 f ð1Þ ð2Þ ð3:152Þ
2 r2
where the second terms inside square brackets are the electron–nucleus attractions
(Z is the atomic number), and r1 and r2 are the distances from electrons 1 and 2 to
the atomic nucleus.
Note that the Hamiltonians for the above equations depend on one-electron wave
functions. In particular, V(0) (0) (0) (0)
1 (r1) depends on g (2), and V2 (r2) depends on f (1).
(1) (1)
Orbitals f (1) and g (2), obtained by solving these equations, are used to evaluate
V(1) (1)
1 and V2 , and these potential operators are introduced back into the one-electron
Schrödinger equations to obtain f(2)(1) and g(2)(2). This iterative procedure con-
tinues until no significant change is found in the one-electron wave functions and
corresponding energies, from one iteration to the next. Then, one says that
self-consistency has been attained. Energy eigenvalues e1 and e2 include the 1–2
and 2–1 repulsions, respectively. These repulsions are the same, so the sum e1 + e2
contains the electron–electron repulsion twice. To obtain the total energy, the
repulsion between electron densities |f(1)|2 and |g(2)|2 has to be subtracted from the
sum e1 + e2,
ZZ
jf ð1Þj2 jgð2Þj2
E ¼ e1 þ e2 dv1 dv2 ¼ e1 þ e2 J12 ð3:153Þ
r12
where J12 is called the Coulomb integral.

The above self-consistent field (SCF) procedure can be easily generalized for a
product of spatial orbitals. However, a simple product wave function is not
antisymmetric with respect to the exchange of spatial and spin coordinates of all
electrons. Considering an antisymmetrized product of spin-orbitals (a Slater
determinant), the total energy includes exchange integrals in addition to Coulomb
integrals. The Hartree–Fock total energy for a closed shell configuration of a
molecule with n electrons, i.e., n/2 doubly occupied molecular orbitals, is given by
X
n=2 n=2 X
X n=2
EHF ¼ 2 ei ð2Jij Kij Þ þ VNN ð3:154Þ
i¼1 i¼1 j¼1
(E32, E33). In atomic units, Coulomb integrals are given by

ZZ
/i ð1Þ/i ð1Þ/j ð2Þ/j ð2Þ
Jij ¼ dv1 dv2 ð3:155Þ
r12
and exchange integrals by

ZZ
/i ð1Þ/j ð1Þ/j ð2Þ/i ð2Þ
Kij ¼ dv1 dv2 ð3:156Þ
r12
(E34, E35). While each Coulomb integral represents the repulsion between two
orbital densities, each exchange integral can be physically associated with the
interaction between two overlap densities and results from antisymmetry in the
wave function. Integrations over electron coordinates in (3.155) and (3.156) have
the effect of dealing with Coulombic and exchange electron–electron interactions,
not as real instantaneous repulsions, but as averages over the spatial coordinates of
the electrons. This is the price the Hartree–Fock method pays for dealing with
one-electron functions (orbitals) and, at the same time, having to calculate integrals
involving the electron–electron distance rij.
3.9.1 Slater-Type Orbitals
Hartree–Fock calculations on atoms frequently use Slater-type orbitals. In atomic

units, a Slater-type orbital (STO) is given by
Nnf r n1 efr Y‘m ðh; /Þ ð3:157Þ
where n, ‘, and m are quantum numbers, f is the orbital exponent, and the nor-
malization constant is given by
ð2fÞn þ 1=2
Nnf ¼ ð3:158Þ
½ð2nÞ!1=2
The Mathematica code M7 computes the normalization constants of radial

factors of Slater-type orbitals, for n 3 (E36, E37).
3.9.2 Hartree–Fock Equations
When the Hartree–Fock SCF procedure is carried out for a closed shell electronic
configuration, that is, for a nondegenerate ground state (electron-spin multiplicity
2S + 1 = 1), the wave function is given by a single Slater determinant of
spin-orbitals. The orthonormal orbitals that minimize the Hartree–Fock energy
(3.154) satisfy the Hartree–Fock equations,
^
Fð1Þ/i ð1Þ ¼ ei /i ð1Þ ð3:159Þ
where ei is an eigenvalue (an orbital energy), /i is the corresponding eigenfunction

(a spatial orbital), and the Fock operator in atomic units is given by
1 X Za X
n=2

^
Fð1Þ ¼ r21 þ 2^Jj ð1Þ K
^ j ð1Þ ð3:160Þ
2 a
r1a j¼1
where the subscript a runs over the nuclei, and the Coulomb and exchange op-
erators are defined by the following equalities:
Z
/j ð2Þ/j ð2Þ
^Jj ð1Þ/i ð1Þ ¼ /i ð1Þ dv2 ð3:161Þ
r12
and
Z
/j ð2Þ/i ð2Þ
^ j ð1Þ/i ð1Þ ¼ /j ð1Þ
K dv2 ð3:162Þ
r12
Equations (3.159) have to be solved iteratively until self-consistency is attained,

since the Fock operator depends on /j for j 6¼ i [see (3.160)].
3.9.3 Hartree–Fock–Roothaan Equations
In 1951, Roothaan expanded /i in one-electron basis functions vt centered on the

nuclei of the molecule,
X
b
/i ¼ vt cti ð3:163Þ
t¼1
where b is the total number of basis functions. Substitution of (3.163) in the Fock
equations (3.159) and multiplication by the complex conjugate of vs followed by
integration leads to
X
b
ðFst ei Sst Þcti ¼ 0 s ¼ 1; 2; . . .; b ð3:164Þ
t¼1
where
Z Z
Fst ¼ ^ t dv
vs Fv Sst ¼ vs vt dv ð3:165Þ
Equations (3.164) are called Hartree–Fock–Roothaan equations and form a

system of linear homogeneous equations in which each unknown ei corresponds to
a set of b coefficients cti (t = 1, 2, …, b) that define a normalized /i. The nontrivial
solution of (3.164) implies
detðFst ei Sst Þ ¼ 0 ð3:166Þ
This equation leads to a polynomial whose unknowns are the orbital energies ei.
Solving the equations that result from the substitution of each determined ei in
(3.164) and normalizing /i gives the coefficients cti, where t = 1, 2, …, b. The
F matrix elements are obtained from the Fock operator, which in turn depends on
the /i orbitals. Therefore, the Hartree–Fock–Roothaan equations are solved by an
iterative process that ends when self-consistency is reached.
Rearranging (3.164) leads to
X
b X
b
Fst cti ¼ Sst cti ei s ¼ 1; 2; . . .; b ð3:167Þ
t¼1 t¼1
which suggests the use of matrix formalism. In fact, the first member shows that
each row of the F matrix (a fixed s value, with t varying from 1 to b) multiplies a
column of the matrix of coefficients (with a defined i, with t varying from 1 to b).
Repeating this procedure for all values of s and subsequently all values of i gives
the product of the matrices F and C. In the second member of (3.167), the matrix
S replaces the matrix F of the first member and each column of the matrix of
coefficients C is multiplied by the corresponding orbital energy value. In matrix
formalism, Eq. (3.167) are expressed as
FC ¼ SCe ð3:168Þ
3.9.4 Correlation Energy
A Hartree–Fock SCF wave function takes the interactions between electrons only in
an average way, whereas interelectronic repulsions are in reality instantaneous.
Thus, the difference between the exact nonrelativistic energy and the Hartree–Fock
energy is defined as the correlation energy
Ecorr ¼ Eexact EHF ð3:169Þ
According to the variational theorem (3.126), the exact nonrelativistic energy is

lower than the Hartree–Fock energy. Therefore, the correlation energy is always
negative.
One apparently obvious method of dealing with electron correlation involves the
explicit inclusion of electron–electron distance in the wave function. However, due
to the occurrence of many difficult integrals involving the electron–electron dis-
tance, this method can be applied only to very small systems, one “historical”
example being the helium atom (Coulson and Neilson 1961). Another method of
dealing with correlation is configuration interaction, whereby the wave function is
taken as a linear combination of ground and excited configurations,
X
W¼ ci U i ð3:170Þ
i
and each Ui corresponds to a single Slater determinant or to a linear combination of

a few Slater determinants in open shell configurations. Significant contributions in
(3.170) require excited configurations with the same symmetry as the ground state
configuration.
For a specified method of calculation and set of basis functions, the correlation
correction can be obtained as a percentage using the formula (Ecalc − EHF)/Ecalc
100. For small molecules, energy values can be readily found in CCCBDB (see
Credits). Considering the B3LYP method and the aug-cc-pVTZ basis set (the
student can find the technical details of this method of calculation and basis set
explained in Sect. 3.15), the correlation correction is about 0.5 % for H2O and
approximately 0.8 % for CH4 and H3CCH3. In chemical terms, these errors cor-
respond to important energies, roughly 3 to 4 times the energy of one CH bond in
CH4 (435 kJ mol−1).
3.10 Density Functional Theory
3.10.1 Electron Probability Density
Consider the nondegenerate ground state wave function of one molecule,

W0(s1, …, sn), where sk represents the set of three spatial coordinates (xk, yk, zk) and
one spin coordinate of electron k (k = 1, …, n). This wave function contains
3n spatial coordinates of the n electrons. The probability of finding electron 1 in
volume dv1 irrespective of its spin and positions and spins of other electrons is
given by
Z Z
2
dpðr1 Þ ¼ ... jW0 ðs1 ; . . .; sn Þj dv2 . . .dvn dr1 . . .drn dv1 ð3:171Þ
where the integrals extend to the spatial coordinates of electrons 2 to n and to the
spin coordinates of electrons 1 to n. The electrons are indistinguishable, so the
probability for finding any electron, not just electron 1, in the volume dv, is n times
(3.171). Therefore, the ground state electron probability density is
Z Z
2
q0 ðrÞ ¼ n ... jW0 ðs1 ; . . .; sn Þj dv2 . . .dvn dr1 . . .drn ð3:172Þ
q0(r) is a one-electron function central to the density functional theory, as we are

going to find out. Its units are volume−1 (in atomic units, 1/bohr3; see Appendix).
The electron charge density is obtained by multiplying q0 by the electron charge
−e. Since the electron charge in atomic units is −1, the electron charge density is
given in atomic units by −q0.
Note that the integral of q0(r) over dv [see (3.172); integration of the right-hand
side of (3.172) is over dv1] gives n, the total number of electrons of the molecule,
since the ground state wave function is normalized,
Z
q0 ðrÞdv ¼ n ð3:173Þ
3.10.2 External Potential
We now consider the pure electronic Hamiltonian (3.141) and the electronic
Schrödinger equation (3.140), within the Born–Oppenheimer approximation
(Sect. 3.8). The electronic Schrödinger equation is solved for fixed positions of the
nuclei, since nucleus–nucleus repulsions simply add a constant to the electronic
Hamiltonian in the electronic Schrödinger equation (3.140). The set of nuclei in
their fixed positions in the molecule, that is, the nuclear framework of the molecule,
originates on each electron i the potential given by
3.10 Density Functional Theory 153
X Za
vðri Þ ¼ ð3:174Þ
a
rai
in atomic units. In density functional theory, (3.174) is called external potential,

because it reflects the nuclear framework of the molecule, which is external to the
system of electrons. Considering that we know the number of electrons of the
molecule, it follows that the nuclear framework of the molecule, and consequently,
the external potential, determine the molecular Hamiltonian, since the kinetic
energy operator and the electron–electron repulsion operator are the molecule’s
independent operators. The sum of v(ri) over all electrons gives
X
n
^Ne ¼
V vðri Þ ð3:175Þ
i¼1
Since the electrons are indistinguishable and v(r) commutes with W0 [v(r) is
simultaneously an operator and a function, and as such, commutes with another
function], we can write
Z Z !
X
n
. . . W0 vðri Þ W0 dv1 . . .dvn dr1 . . .drn
Z Z
i¼1 Z
¼ vðrÞn 2
. . . jW0 j dv2 . . .dvn dr1 . . .drn dv ð3:176Þ
Using (3.172), we obtain the average value of the nuclei–electron attractions

Z
VNe ½q0 ¼ vðrÞq0 ðrÞdv ð3:177Þ
3.10.3 Functional Derivative
We now derive an expression that enables us to evaluate functional derivatives, and

in particular, obtain the functional derivative of VNe with respect to q. We begin by
considering the set of all continuous plane curves connecting two given points
a and b, and the length associated with each curve. This is a simple example of a
functional, a rule that associates a number, in this case the length of each curve or
functional value, with a function, in this example each curve connecting points
a and b. The length of each plane curve is given by the following integral:
Zb Zb Zb

2 1=2
1=2
J¼ ds ¼ dx þ dy
2
¼ 1 þ y02 dx ð3:178Þ
a a a
where dy = y′dx. This integral sets a correspondence between its value J, a real
number, and the derivative of each plane curve connecting the given points a and b,
the function y′. The integral J is a functional of y′, and this is represented by J[y′].
We are interested in finding the curve with the shortest length. Of course, this is a
trivial question, since the shortest path between two points is a straight line.
However, solving this problem in a systematic way requires evaluation of functional
derivatives, an important question in density functional theory. To this end, let us
consider the more general functional
Zb
J½y ¼ Fðx; y; y0 Þdx ð3:179Þ
a
and an increment h(x) to the function y(x),
yðxÞ ) yðxÞ þ hðxÞ ð3:180Þ
so that h(a) = h(b) = 0 (a and b are fixed endpoints, so the variations at these points
are zero). The corresponding increment for the functional (3.179) is
Zb
DJ ¼ J½y þ h J½y ¼ ½Fðx; y þ h; y0 þ h0 Þ Fðx; y; y0 Þdx ð3:181Þ
a
and the Taylor expansion of the integrand gives
Zb

DJ ¼ Fy h þ Fy0 h0 dx þ ð3:182Þ
a
where Fy and Fy′ represent partial derivatives of F with respect to y and y′, and the dots
represent terms of order higher than 1, relative to h and h′. The differential of J[y] is
Zb

dJ ¼ Fy h þ Fy0 h0 dx ð3:183Þ
a
(note the use of the symbol d to represent a functional differential and compare (3.183)
with (3.179)). Integration by parts of the second term inside square brackets leads to
Zb Zb Zb

b d
dJ ¼ 0
Fy h þ Fy0 h dx ¼ Fy hdx þ Fy0 h
a
Fy0 hdx
dx
a a a
Zb
d
¼ Fy Fy0 hdx ð3:184Þ
dx
a
since the second term is zero because h(a) = h(b) = 0. At a stationary point,
Zb
d
dJ ¼ Fy Fy0 hðxÞdx ¼ 0 ð3:185Þ
dx
a
Since h(x) is arbitrary, we can conclude that
d
Fy F y0 ¼ 0 ð3:186Þ
dx
This equation, called the Euler–Lagrange differential equation, is central to
the calculus of variations. The necessary condition for J[y] to have an extremum
(maximum or minimum) is that y(x) satisfy the Euler–Lagrange equation (see
Gelfand and Fomin, Further Reading, Chap. 1, Sect. 4).
Returning now to the example of the curve in a plane with the shortest length
between two points a and b and applying the Euler–Lagrange equation, we obtain
Fy = 0, Fy’ = y′/√(1 + y′2), dFy′/dx = 0, and y′ = constant, that is, y(x) = Ax + B, as
anticipated.
Since h(x) is an arbitrarily small variation in y(x) [see (3.180)], we can represent
h(x) by dy and rewrite the last equality of (3.184) as
Zb
d
dJ ¼ Fy Fy0 dydx ð3:187Þ
dx
a
The functional or variational derivative of J is given by
dJ d
¼ F y F y0 ð3:188Þ
dy dx
(a proof of this statement can be found in Gelfand and Fomin, Further Reading,
Chap. 1, Sect. 7). Hence, the Euler–Lagrange equation (3.186) can be formulated
as expressing that the functional derivative at a stationary point is zero, a result that
is also applicable to an ordinary derivative of a function at a stationary point.
We now consider (3.177) for a trial electron density q and, using (3.179), write
F(r, q) = v(r)q(r) and apply (3.188). Since F does not explicitly depend on q′, its
derivative with respect to q′ is zero. In turn, the derivative of F with respect to q is
v(r). Hence, using (3.188), we conclude that the functional derivative of VNe with
respect to q is equal to the external potential v(r),
dVNe ½qðrÞ
¼ vðrÞ ð3:189Þ
dqðrÞ
This result connects the external potential v(r) with the average nuclei–electrons
attraction potential VNe. The Mathematica code M8 calculates several functional
derivatives and confirms (3.189).
3.10.4 Hohenberg–Kohn Theorems
The ground state wave function of a molecule is a function of 4n electron coor-

dinates (3n spatial coordinates and n spin coordinates), whereas the electron
probability density is a function of only three spatial coordinates. Unlike the wave
function, the electron probability density of a molecule is an experimentally
determined observable that can be obtained from X-ray diffraction intensities of
molecular crystals.
For a nondegenerate ground state, Pierre Hohenberg and Walter Kohn demon-
strated in 1964 that the ground state electron probability density q0 uniquely
determines the Hamiltonian operator, the ground state molecular energy, and all
other ground state molecular properties (Hohenberg and Kohn 1964). This theorem
was proved by reductio ad absurdum, that is, by showing that its contradiction
would lead to a false conclusion. In particular, it was assumed that the same ground
state electron density q0 can give rise to two external potentials differing by more
than a constant, and consequently, to two different n-electron Hamiltonians. This
premise leads to a logical impossibility, thus proving that the ground state electron
probability density q0 uniquely determines the Hamiltonian operator, the ground
state molecular energy, and all other ground state molecular properties. The ground
state electron density q0 determines v(r), which is a system-dependent operator,
since it involves the nuclear framework of the molecule. In turn, the external
potential determines the electronic Hamiltonian, because the kinetic energy operator
and the electron–electron repulsion operator are universal operators, dependent only
on the number of electrons of the molecule.
The electronic Hamiltonian is the sum of the operators for the kinetic energy of
electrons and the potential energy for the electron–nucleus attractions and electron–
electron repulsions [see (3.141)]. Consequently, we can write
E0 ¼ Te þ VNe þ Vee ð3:190Þ
where E0, Te, VNe, and Vee represent ground state averages of the Hamiltonian,
electron kinetic energy operator, electron–nucleus attractions, and electron–electron
repulsions, respectively. Since E0 is a functional of the electron density q0 and VNe
is a known functional of q0 [see (3.177)], we conclude that Te and Vee are also
functionals of q0 and write
Z
E0 ½q0 ¼ Te ½q0 þ vðrÞq0 ðrÞdv þ Vee ½q0 ð3:191Þ
Note that the Hohenberg–Kohn theorem does not tell us how to calculate E0
from q0, since Te and Vee are unknown functionals of q0.
For a trial electron density q(r) that satisfies
Z
qðrÞdv ¼ n ð3:192Þ
and q(r)
0 for all r, Hohenberg and Kohn also proved that
Z
E½q ¼ Te ½q þ vðrÞqðrÞdv þ Vee ½q
E0 ½q0 ð3:193Þ
that is, the true ground state electron density minimizes the energy functional. This
variational theorem shows that no trial electron density can give a lower ground state
energy than the exact ground state electron density. Note that the validity of this theorem
requires the use of the electronic Hamiltonian, as mentioned above [see (3.190)].
3.10.5 Kohn–Sham Method
Let us now recall the first Hohenberg–Kohn theorem, which tells us that the ground
state electron probability density q0 uniquely determines the ground state molecular
energy and all other ground state molecular properties. However, this theorem does
not tell us how to calculate E0 from q0, since Te and Vee in (3.191) are unknown
functionals of q0, nor does it provide us with a method for evaluating q0 without
having first to obtain the ground state wave function.
In 1965, Kohn and Sham considered a fictitious reference system (we label this
system by the subscript s) formed by n noninteracting electrons, with each electron
being subject to the effective potential vs(ri), which was chosen so that the ground
state electron density of the reference system qs is equal to the exact ground state
electron density of the real molecule, qs = q0 (Kohn and Sham 1965). Note that the
electron probability density is a function of 3 variables and is obtained from a wave
function of 4n variables by integration over 4n-3 variables [see (3.172)]. The
connection between this noninteracting system and the target molecule is provided
by the first Hohenberg–Kohn theorem, since the ground state electron density of the
reference system qs, being equal to q0, uniquely determines the ground state
molecular energy and all other ground state molecular properties of our real target
system of interacting electrons. The noninteracting electrons of the fictitious system
enable us to define one-electron wave functions (spin-orbitals) and form a Slater
determinant of the lowest-energy Kohn–Sham spin-orbitals,

ws;0 ¼
vKS KS KS
1 v2 . . .vn k ¼ hi a
vKS KS
k þ 1 ¼ hi b
vKS KS
ð3:194Þ
where k varies from 1 to n, i varies from 1 to n/2 (for n electrons, the closed shell
configuration has n/2 occupied orbitals), and each spin-orbital vk is the product of a
Kohn–Sham orbital hi and the spin function a or b. Note that the Slater
pffiffiffiffi
determinant includes the normalizing factor 1= n!. Just as the orthonormal orbitals
that minimize the Hartree–Fock energy satisfy the Fock equations, the Kohn–Sham
orthonormal orbitals that minimize the ground state energy satisfy the Kohn–Sham
eigenvalue equations
^hKS ð1ÞhKS ð1Þ ¼ eKS hKS ð1Þ ð3:195Þ

i i i
where the one-electron Kohn–Sham Hamiltonian in atomic units is given by
^hKS ð1Þ ¼ Tê ðr1 Þ þ vs ðr1 Þ ð3:196Þ
and the electron density of the Slater determinant of spin-orbitals is
X
n=2

KS
2
q0 ¼ qs ¼
h
ð3:197Þ
i
i¼1
(E38).
We now recall Eq. (3.191), which represents the first Hohenberg–Kohn theorem,
Z
E0 ½q0 ¼ Te ½q0 þ vðrÞq0 ðrÞdv þ Vee ½q0 ð3:198Þ
with Te and Vee unknown functionals of q0. The kinetic energy operator is a sum of
one-electron kinetic energy operators. Hence, for the noninteracting system of Kohn
and Sham, it can be shown that the kinetic energy is given by
Z Z " #
X
n
Te;s ½q0 ¼ ... ws;0 Tê ðiÞ ws;0 dv1 . . .dvn dr1 . . .drn
n Z
i¼1
X
¼ ðhKS ^
i Þ ð1ÞTe ð1Þ hi ð1Þdv1
KS
ð3:199Þ
i¼1
(E39). Note that this kinetic energy is not equal to the kinetic energy of the real
interacting system, even if the noninteracting and interacting systems share the
same electron density. Therefore, we can write
Te ½q0 ¼ Te;s ½q0 þ DTe ½q0 ð3:200Þ
where DTe is the difference between the average kinetic energy of the molecule and
the average kinetic energy of the reference system. We now consider the third term
on the right-hand side of (3.198) and write
ZZ
1 q0 ðr1 Þq0 ðr2 Þ
Vee ½q0 ¼ dv1 dv2 þ DVee ½q0 ð3:201Þ
2 r12
where the second term of the right-hand side of this equation is the difference
between the average electron–electron repulsion energy of the molecule and the
electrostatic electron–electron repulsion obtained from the ground state electron
density q0. In order to understand the 1/2 factor of the first term of (3.201), we need
to go back to (3.172) and find that each q0 factor involves a multiplication by n, that
is, it runs over n electron labels that are different in both q0 factors, as required by
the operator 1/rij (i 6¼ j). For example, for electrons with labels 1, 2, and 3, the
six-dimensional integration of (3.201) applies to a sum over electron pairs 1–2, 1–3,
2–1, 2–3, 3–1, and 3–2. Hence, each electron–electron pair is repeated, and the
factor 1/2 corrects for that double counting. Considering now the physical meaning
of the electron charge density in atomic units −q0, the products −q0(r1)dv1 and
−q0(r2)dv2 in the six-dimensional integral of (3.201) represent electronic charges in
the volume elements dv1 and dv2. Therefore, on the right hand-side of (3.201), the
first term represents the electrostatic electron–electron repulsion energy.
Substitution of the result of (3.199) in (3.200) and subsequent substitution of the
resulting Eq. (3.200) and of (3.201) in (3.198) leads to
n Z
X
E0 ½q0 ¼ ðhKS ^
i Þ ð1ÞTe ð1Þhi ð1Þdv1
KS
i¼1
Z ZZ ð3:202Þ
1 q0 ðr1 Þq0 ðr2 Þ
þ vðrÞq0 ðr Þdv þ dv1 dv2 þ Exc ½q0
2 r12
where the delta terms DTe and DVee are grouped in the so-called
exchange-correlation energy functional defined by
Exc ½q0 DTe ½q0 þ DVee ½q0 ð3:203Þ
The energy components of Exc[q0] are the kinetic correlation energy, which is
the DTe difference between the kinetic energy for the real molecule and the kinetic
energy of the reference system, the exchange energy, which results from the
antisymmetric characteristic of the wave function [see (3.194)], the Coulombic
correlation energy, which accounts for the influence of electron correlation on the
electrostatic electron–electron repulsions described by the first term of the
right-hand side of (3.201), and the self-interaction correction, which results from
the fact that in the electrostatic electron–electron repulsion energy, the electron
density q0(r1) is different from zero in dv1, and so electron 1 erroneously interacts
with itself, the same thing happening with q0(r2) and dv2.
Since the exact form of the exchange-correlation energy functional is unknown,
one needs to resort to approximate functionals to describe it. Having arrived at the
expression for the ground state energy functional [see (3.202)], we can now be a bit
more explicit about the external potential operator of (3.196) for the one-electron
Kohn–Sham Hamiltonian
X Za Z
^ 1 qðr2 Þ
hKS ð1Þ ¼ Tê ðr1 Þ þ vs ðr1 Þ ¼ r21 þ dv2 þ vxc ð1Þ ð3:204Þ
2 a
r1a r12
where vxc(r) is the one-electron exchange-correlation potential, defined by the

derivative of the exchange-correlation energy functional,
dExc ½q0 ðrÞ

vxc ðrÞ ð3:205Þ
dq0 ðrÞ
In analogy with (3.189), this definition implies that

Z
Exc ½q0 ¼ vxc ðrÞq0 ðrÞdv ð3:206Þ
[see (3.177)]. This equation results from (3.205), since we defined vxc(r) from
Exc[q0], not the other way round.
3.10.6 Overview
Let us now summarize the main steps in the derivation of the density functional
theory equations. The most important statement lies in the first Hohenberg–Kohn
theorem, which says that the ground state electron probability density q0 uniquely
determines the Hamiltonian operator, the ground state molecular energy, and all
other ground state molecular properties.
Within the Born–Oppenheimer approximation, the electronic Hamiltonian is the
sum of operators for the kinetic energy of electrons, potential energy of the elec-
tron–nucleus attractions, and potential energy of electron–electron repulsions.
Based on the first Hohenberg–Kohn theorem, we conclude that the electronic
energy and its component terms are functionals of q0. However, this theorem does
not tell us how to calculate E0 from q0, since Te and Vee are unknown functionals of
q0.
The Kohn–Sham method overcomes this difficulty by considering a fictitious
reference system formed by n noninteracting electrons, with each electron subject to
an unknown effective potential, which in turn is supposed to originate the exact
ground state electron density for the real molecule.
The fictitious reference system of noninteracting electrons leads to one-electron
functions (i.e., spin-orbitals), thus enabling the calculation of the average value of
the kinetic operator for the corresponding Slater determinant. Since this kinetic
energy is not equal to the kinetic energy of the real interacting system, one needs to
determine the difference DTe between the average kinetic energy of the real
molecule and the average kinetic energy of the reference system.
In turn, the average electron–electron repulsion energy is another unknown
functional. However, since the electrostatic electron–electron repulsion obtained
from the ground state electron density q0 can in principle be calculated, the
difference DVee between the average electron–electron repulsion energy of the real
molecule and the electrostatic electron–electron repulsion of the fictitious nonin-
teracting system remains to be calculated.
Now the delta terms DTe and DVee are grouped in the so-called
exchange-correlation energy functional, whose exact form is unknown. For this
reason, in the various DFT methods of calculation, approximate functionals are
used to describe the exchange-correlation energy functional.
One practical question remains to be answered: how to obtain the electron
density of the target molecule? The answer to this question is provided by the
variational Hohenberg–Kohn theorem, which shows that the true ground state
electron density minimizes the energy functional obtained using the exact electronic
Hamiltonian. This theorem suggests a way of improving the trial electron density,
obtained as the sum of the squares of the absolute values of the spatial Kohn–Sham
orbitals [see (3.197)], by varying the electron density, so as to minimize the energy
functional. Like the one-electron Hartree–Fock equations (3.159), the one-electron
Kohn–Sham equations (3.195) also have to be solved iteratively, since the
one-electron Kohn–Sham Hamiltonian depends on all one-electron hi functions [see
(3.197) and (3.204)].
Note the parallelism between the Hartree–Fock and Kohn–Sham methods. Both
assume fictitious systems in order to define one-electron functions (orbitals).
However, in the Hartree–Fock method, each electron is subject to the average
Coulomb repulsion of the remaining electrons, to which is added the exchange
potential that results from the antisymmetry requirement. In turn, in the Kohn–
Sham method, a fictitious reference system is formed by n noninteracting electrons,
with each electron being subject to the effective potential vs(ri) so chosen that the
ground state electron density of the reference system qs is equal to the exact ground
state electron density of the real molecule, qs = q0. The Kohn–Sham equa-
tions (3.195) and (3.196) correspond to the Hartree–Fock equations (3.159) and
(3.160). Both Hartree–Fock and Kohn–Sham methods can be expressed by matrix
formalisms, whereby Hartree–Fock orbitals are replaced by Kohn–Sham orbitals,
and the Fock matrix is replaced by a Kohn–Sham matrix.
3.11 Perturbation Theory
After the variational method, the major method for approximately solving the
Schrödinger equation is perturbation theory, applicable to a system B whose
Schrödinger equation cannot be solved but can be obtained from the Schrödinger
equation for system A by the introduction of a small perturbation. A hydrogen atom
in the presence of a small uniform electric field illustrates the use of perturbation
theory. In this section, both the unperturbed Hamiltonian (system A) and the per-
turbation are time-independent operators.
3.11.1 Nondegenerate Energy Level
Consider the time-independent Schrödinger equation
^ k ¼ Ek wk
Hw ð3:207Þ
where wk represents the eigenfunctions of the perturbed system (system B) and

assume that we can solve the Schrödinger equation for the unperturbed system
(system A)
^ 0 wð0Þ ¼ Eð0Þ wð0Þ

H ð3:208Þ
k k k
ð0Þ
where wk represents the wave function for the unperturbed nondegenerate level
ð0Þ ^ differs slightly from H^ 0 by the perturbation
with energy Ek . Suppose that H
^,
operator H 0
^ ¼H
H ^0
^0 þH ð3:209Þ
We introduce the scaling parameter k so that the perturbation is applied grad-

ually and there is a continuous change from the unperturbed to the perturbed
system. Therefore,
^ ¼H
H ^ 0 þ kH
^0 ð3:210Þ
where k varies between 0, for the unperturbed system, and 1, for the perturbed
system when the perturbation is fully “switched on.” We now consider a nonde-
ð0Þ ð0Þ
generate eigenfunction w0 and corresponding eigenvalue E0 of the unperturbed
Hamiltonian H^ 0,
^ 0 w0ð0Þ ¼ E0ð0Þ w0ð0Þ

H ð3:211Þ
and the Schrödinger equation for the system with the perturbation fully “switched
on,”
^ 0 ¼ E0 w0
Hw ð3:212Þ
Substitution of (3.210) in (3.212) shows that both the eigenfunction and the
eigenvalue depend on k,
w0 ¼ w0 ðs; kÞ E0 ¼ E0 ðkÞ ð3:213Þ
where s represents the space and spin coordinates. We can now expand in Taylor
series the eigenfunction and the eigenvalue in powers of k, obtaining
3.11 Perturbation Theory 163

@w0 k2 @ 2 w 0 ð0Þ ð1Þ ð2Þ
w0 ¼ ðw0 Þk¼0 þ k þ þ ¼ w0 þ kw0 þ k2 w0 þ
@k k¼0 2! @k k¼0
2
ð3:214Þ

@E0 k2 @ 2 E 0 ð0Þ ð1Þ ð2Þ
E0 ¼ ðE0 Þk¼0 þ k þ þ . . . ¼ E0 þ kE0 þ k2 E0 þ
@k k¼0 2! @k2 k¼0
ð3:215Þ
where

ðkÞ 1 @ k w0 ðkÞ 1 @ k E0
w0 E0 ð3:216Þ
k! @kk k¼0 k! @kk k¼0
We assume that
D
E
ð0Þ
w0
w0 ¼ 1 ð3:217Þ
and expand w0 using (3.214) to obtain

D
E D
E D
E D
E
ð0Þ
ð0Þ
ð0Þ ð0Þ
ð1Þ ð0Þ
ð1Þ
w0
w0 ¼ 1 ¼ w0
w0 þ k w0
w0 þ k2 w0
w0 þ ð3:218Þ
|fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl}
1 0 0
Substitution of (3.210), (3.214), and (3.215) in (3.212) gives

^ 0 Þ wð0Þ þ kwð1Þ þ k2 wð2Þ þ
^ 0 þ kH
ðH 0 0 0

ð0Þ ð1Þ ð2Þ ð0Þ ð1Þ ð2Þ
¼ E0 þ kE0 þ k2 E0 þ w0 þ kw0 þ k2 w0 þ ð3:219Þ
We now equate the coefficients of equal powers of k and obtain

k0 : ^ 0 Eð0Þ wð0Þ ¼ 0
H ð3:220Þ
0 0

k1 : H^ 0 E0ð0Þ w0ð1Þ þ H^ 0 E0ð1Þ w0ð0Þ ¼ 0 ð3:221Þ

k2 : ^ 0 E0ð0Þ w0ð2Þ þ H
H ^ 0 E0ð1Þ w0ð1Þ E0ð2Þ wð00Þ ¼ 0 ð3:222Þ
Equation (3.220), the unperturbed Schrödinger equation, is automatically satis-

fied [see (3.211)]. We now multiply the other two equations on the left by the
ð0Þ ^ 0 is a Hermitian
complex conjugate of w0 and integrate over all space. Since H
operator, we obtain the first- and second-order corrections to the energy
D E D E
ð0Þ
^ 0
ð0Þ ð0Þ
^ 0
ð1Þ
E0 ¼ w0
H E0 ¼ w0
H
ð1Þ ð2Þ
w0 w0 ð3:223Þ
ð1Þ
We now expand the unknown function w0 in the complete set of orthonormal
eigenfunctions of the unperturbed Hamiltonian,
X XD
E
ð1Þ ð0Þ ð0Þ
ð1Þ ð0Þ
w0 ¼ ak w k ¼ wk
w0 wk ð3:224Þ
k6¼0 k6¼0
since the unperturbed and perturbed systems are assumed to have the same
ð1Þ ð0Þ
boundary conditions. The expansion coefficient a0 is zero because w0 and w0 are
orthogonal [see (3.218)]. Multiplying (3.221) on the left by the complex conjugate
ð0Þ
of wk and integrating over all space gives
D
E D E
ð0Þ
ð1Þ ð0Þ
^ 0
ð0Þ
¼ wk
H
ð0Þ ð0Þ
Ek E0 wk
w0 w0 )
|fflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflffl}
ak D E ð3:225Þ
ð0Þ
^ 0
ð0Þ
D
E w H w
ð0Þ
ð1Þ k 0
ak ¼ wk
w0 ¼ ð0Þ ð0Þ
Ek E0
Hence,
D E
ð0Þ
^ 0
ð0Þ
ð1Þ
X wk
H w0
ð0Þ
w0 ¼ ð0Þ ð0Þ
wk ð3:226Þ
k6¼0 Ek E0

D ED E
ð0Þ
^ 0
ð0Þ ð0Þ ^ 0
ð0Þ
ð2Þ
X w0
H wk wk H w0
E0 ¼ ð0Þ ð0Þ
ð3:227Þ
k6¼0 Ek E0
thus enabling us to conclude that the second-order energy can be derived from the
first-order correction to the wave function. Also, it can be shown that the third-order
energy could also be obtained from the first-order wave function, and in general, the
(2n + 1)th energy correction can be derived from the nth-order perturbation wave
function.
3.11.2 Variational Perturbation Method
Perturbation expressions like (3.226) and (3.227) that involve infinite sums are not
suitable for quantitative evaluations. However, the variational principle can be used
to provide an upper bound on the second-order perturbation energy. To this end,

ð1Þ
multiplying (3.221) on the left by the complex conjugate of w0 and integrating
over all space leads to
D

E D

E
ð1Þ
^ 0 ð0Þ
ð1Þ ð1Þ
^ 0 ð1Þ
ð0Þ
w 0
H E0
w0 ¼ w0
H E0
w0 ð3:228Þ
ð0Þ
We now multiply (3.222) on the left by the complex conjugate of w0 , integrate
over all space, and apply the orthogonality relationships of (3.218). Using the
Hermitian characteristics of the Hamiltonian and the perturbation operator gives
D

E
ð1Þ
^ 0 ð1Þ
ð0Þ ð2Þ
w0
H E0
w0 ¼ E0 ð3:229Þ
Combining (3.228) and (3.229) allows us to write

D

E
ð1Þ
^ 0 ð0Þ
ð1Þ ð2Þ
w0
H E0
w0 ¼ E0 ð3:230Þ
We consider now a guess function / and express the variational principle

ð1Þ
Eq. (3.126) for / w0 as
D

E
ð1Þ
^ 0 ð0Þ
ð1Þ
/ w0
H E0
/ w0
0 ð3:231Þ
This expression gives

D

E D

E

^0 ð0Þ

^0 ð0Þ
ð1Þ
/
H E0
/ /
H E0
w0
D

E D

E
ð1Þ
^ 0 ð0Þ
ð1Þ
^ 0 ð0Þ
ð1Þ
w0
H E0
/ þ w0
H E0
w0
0 ð3:232Þ

D

E D

E D

E

^0 ð0Þ

^0 ð0Þ
ð1Þ ð1Þ
^ 0 ð0Þ
ð2Þ
/
H E0
/ /
H E0
w0 w0
H E0
/
E0 ð3:233Þ
We now return again to Eq. (3.221), multiply it on the left by the complex
^ 0 , and substitute in
conjugate of /, integrate, use the Hermitian characteristics of H
(3.233), obtaining
D

E D

E D

E

^0 ð0Þ

^0 ð1Þ
ð0Þ ð0Þ
^ 0 ð1Þ
ð2Þ
/
H E0
/ þ /
H E0
w0 þ w0
H E0
/
E0 ð3:234Þ
The first member of this equation is a functional of /, J[/], which is an upper

bound to the second-order perturbation energy (E40).
3.11.3 Degenerate Energy Level
An energy level with a degeneracy degree d has d linearly independent wave

functions,
ð0Þ ð0Þ ð0Þ

w0;1 ; w0;2 ; . . .; w0;d ð3:235Þ
which are eigenfunctions of the unperturbed Hamiltonian,

^ 0 E0ð0Þ wð0Þ ¼ 0
H ð3:236Þ
0;k
Note that labels 1, 2, …, d do not necessarily imply that the corresponding states
are the lowest-lying states. The wave functions (3.235) are assumed to be
orthonormal. Every linear combination of them is still an eigenfunction of the
unperturbed Hamiltonian with the same eigenvalue,
X
d
^ 0 E0ð0Þ
H
ð0Þ
ak w0;k ¼ 0 ð3:237Þ
k¼1
Assuming that the perturbation splits the degeneracy of the energy level, there is
a unique normalized eigenfunction w0 with eigenvalue E0 for the total Hamiltonian,

H^ E0 w 0 ¼ 0 ð3:238Þ
However, there is a problem in defining limk!0 w0 , since there is an infinite

number of linear combinations of the wave functions (3.235) that satisfy the
Schrödinger equation for the unperturbed Hamiltonian and correspond to the same
energy. Let us assume that
ð0Þ
X
d
ð0Þ
lim w0 ¼ /0 ¼ ak w0;k ð3:239Þ
k!0
k¼1
How can we determine the coefficients ak? As previously done for the nonde-
generate energy level, the eigenfunction w0 and the eigenvalue E0 can be expanded
as Taylor series in powers of k, yielding
ð0Þ ð1Þ ð2Þ

w0 ¼ /0 þ kw0 þ k2 w0 þ . . . ð3:240Þ
ð0Þ ð1Þ ð2Þ

E0 ¼ E0 þ kE0 þ k2 E0 þ ð3:241Þ
Substitution of (3.240) and (3.241) in the Schrödinger equation for the perturbed
system (3.238) leads to
h i
^ 0 E0ð0Þ þ kE0ð1Þ þ k2 E0ð2Þ . . .
^ 0 þ kH
H
ð0Þ ð1Þ ð2Þ
/0 þ kw0 þ k2 w0 þ ¼ 0
ð3:242Þ
Equating the coefficients of the k1 terms, one obtains

k1 : H^ 0 E0ð0Þ w0ð1Þ þ H^ 0 E0ð1Þ /0ð0Þ ¼ 0 ð3:243Þ
that is,
X
d
^ 0 E0ð0Þ w0ð1Þ þ H
H ^ 0 E0ð1Þ ð0Þ
ak w0;k ¼ 0 ð3:244Þ
k¼1
ð0Þ
We now multiply both members by the complex conjugate of w0;m , where m is
one of the states of the d-fold degenerate unperturbed level, and integrate over all
space, obtaining
X
d
^ mk
0 ð1Þ
ak H E0 dmk ¼ 0 ð3:245Þ
k¼1
where
D E
ð0Þ
^ 0
ð0Þ
^ mk
H 0
¼ w0;m
H w0;k ð3:246Þ
Equation (3.245) is a set of d linear homogeneous equations in the unknowns ak,

with nontrivial solutions, provided that

^ mk
0 ð1Þ
det H E0 dmk ¼ 0 ð3:247Þ
ð1Þ
This is an equation of degree d in E0 , with d roots that are first-order correc-
ð0Þ
tions to E0 ,
ð0Þ ð1Þ ð0Þ ð1Þ ð0Þ ð1Þ

E0 þ E0;1 ; E0 þ E0;2 ; . . . E0 þ E0;d ð3:248Þ
If all roots of (3.247) are distinct, the degeneracy is completely removed to first
order. Assuming distinct roots for (3.247), we introduce each of them in (3.245),
solve this set of equations for a2/a1, a3/a1, …, ad/a1, and find a1 by normalization of
ð1Þ
(3.239). For each root of (3.247), E0;k , the set of coefficients determines the cor-
responding zeroth-order function (3.239).
3.12 Time-Dependent Perturbation Theory
3.12.1 Time-Dependent Schrödinger Equation
Before presenting the time-dependent perturbation theory, we consider an unper-

^ 0 . The time
turbed system whose time-independent Hamiltonian is represented by H
evolution of the state function W of this undisturbed quantum-mechanical system
0
is given by the time-dependent Schrödinger equation,
h @W0 ^ 0 W0
¼H ð3:249Þ
i @t
where W0 is a function of the coordinates s and time t, W0(s, t). If our system is in a
stationary state of energy E0, then we can use the time-independent Schrödinger
equation
^ 0 W0 ¼ E0 W0
H ð3:250Þ
and substitute this equation in (3.249), obtaining
h @W0
¼ E 0 W0 ð3:251Þ
i @t
Solving this differential equation is easy, since E0 is a constant, and we obtain
W0 ðs; tÞ ¼ AðsÞ expðiE0 t=h) ð3:252Þ
where A(s) = W0(s, 0). Hence, A(s) can be identified with the eigenfunction of the
time-independent Schrödinger equation for a stationary state w0(s),
^ 0 w0 ðsÞ ¼ E0 w0 ðsÞ
H ð3:253Þ
that is, A(s) = w0(s). Therefore,
W0 ðs; tÞ ¼ expðiE0 t=hÞw0 ðsÞ ð3:254Þ
Note that
W0 W0 ¼ w0 w0 ð3:255Þ
since the exponential factor of W0 in (3.254) is a phase factor eia, and as such does
not affect the probability density given by (3.255).
3.12 Time-Dependent Perturbation Theory 169
Let us now consider w0k ðsÞ to be eigenfunctions of the time-independent

Schrödinger equation
^ 0 w0k ¼ Ek0 w0k

H ð3:256Þ
Equation (3.249) can be rewritten as

^ 0 W0 ¼ 0
ih@=@t H ð3:257Þ
Since the operator in parentheses is linear, every linear combination of solutions

of the time-dependent Schrödinger equation (3.249) such as
X
W0 ðs; tÞ ¼ ck expðiEk0 t=hÞw0k ðsÞ ð3:258Þ
k
is a solution of (3.249). The time-dependence of (3.258) is given by the expo-

nentials of imaginary exponents, which act as phase factors.
3.12.2 Time-Dependent Perturbation
^ 0 ðtÞ. The time-dependent

We now consider a time-dependent perturbation H
Schrödinger equation is then given by
h @W ^ 0 ^ 0
¼ H þ H ðtÞ W ð3:259Þ
i @t
where W(s, t) is the new state function of the perturbed system. If we specify a value
for time, for example t = t0, then
X
Wðs; tÞ ¼ bk ðt0 Þ expðiEk0 t=hÞw0k ðsÞ ð3:260Þ
k
since (3.259) and (3.249) have time-independent Hamiltonians, and so the

time-dependence of the state functions is expressed by phase factors [see (3.258)].
Hence, the solution of (3.259) is given by
X
Wðs; tÞ ¼ bk ðtÞ expðiEk0 t=hÞw0k ðsÞ ð3:261Þ
k
where the time-dependence of the coefficients bk is due to the perturbation. Sub-

stitution of (3.261) in (3.259) gives
P " #
@
h bk ðtÞ expðiEk0 t=hÞw0k X
k ^ 0 ^ 0
¼ ðH þ H Þ bk ðtÞ expðiEk t=hÞwk
0 0
i @t k
ð3:262Þ
We now differentiate the first member with respect to time, use the linear
^ 0 and H
behavior of H ^ 0 and apply (3.256)
X dbk ðtÞ
h X
exp iEk0 t h w0k þ bk ðtÞ exp iEk0 t h Ek0 w0k
i k dt k
X 0 0 X 0 0
¼ bk ðtÞ exp iEk t h Ek wk þ
0 ^ wk
bk ðtÞ exp iEk0 t h H
k k
Canceling equal terms on both sides of this equation gives
X dbk ðtÞ
h X 0 0
^ wk
exp iEk0 t h w0k ¼ bk ðtÞ exp iEk0 t h H ð3:263Þ
i k dt k
In order to select the time derivative of bn(t) on the left-hand side of this
equation, we multiply both members by w0 n , integrate over the spatial and spin
coordinates, and use

w0n jw0k ¼ dnk ð3:264Þ
obtaining
dbn ðtÞ iX
^ 0 jw0k

¼ bk ðtÞ expðixnk tÞ w0n jH
dt h k
where
En0 Ek0
xnk ¼
h
As with time-independent perturbation theory [see (3.210)], we introduce now
the scaling parameter k, rewrite the above equation as
dbn ðtÞ iX
^ 0 jw0k

¼ bk ðtÞ expðixnk tÞ w0n jkH ð3:265Þ
dt h k
and expand the coefficients bk(t) in powers of k (see Ballentine, Further Reading,
Sect. 12.5),
3.12 Time-Dependent Perturbation Theory 171
ð0Þ ð1Þ ð2Þ

bk ðtÞ ¼ bk þ k bk ðtÞ þ k2 bk ðtÞ þ ð3:266Þ
We now substitute (3.266) in (3.265), collect equal powers of k, and obtain, to

first order,
i X ð0Þ
ð1Þ
dbn ðtÞ
¼ ^ 0 jw0k
bk expðixnk tÞ w0n jH ð3:267Þ
dt h k
and to second order,
i X ð1Þ
ð2Þ
dbn ðtÞ
¼ ^ 0 jw0k
bk ðtÞ expðixnk tÞ w0n jH
dt h k
In (3.267), the time-independent zeroth-order coefficients correspond to the

absence of perturbation and to stationary states. In addition, integration of (3.267)
ð1Þ
enables us to obtain the first-order coefficient bn ðtÞ.
Before the time-dependent perturbation H ^ 0 ðtÞ is applied at t = 0, let us assume
ð0Þ
that the system is in the nondegenerate stationary energy level Ei0 , and so bi =1
ð0Þ
and bk = 0 for k 6¼ i, that is,
W0i ¼ expðiEi0 t=hÞw0i ð3:268Þ
The perturbation is switched on in the time interval 0 < t < s, during which the
coefficients of (3.261) become variable, and so there are no stationary states. For
times t
s, the perturbation vanishes, and the coefficients retain the values they
had at time s, that is, bk(t > s) = bk(s). The probability of having a nondegenerate
state with energy Ef0 is given by |bf(s)|2 [see (3.261)]. In order to obtain this
probability, we begin by using (3.267) and write
ð1Þ
dbf ðtÞ i X ð0Þ D

E
¼ bk expðixfi tÞ w0f
H ^ 0
w0k
dt h k
i D

E
¼ expðixfi tÞ w0f
H
^ 0
w0i ð0Þ
ðbk ¼ dki Þ ð3:269Þ
h
We now integrate both members between t = 0 and t = s, use (3.266) for k = 1,
ð0Þ
and note that bf = 0, obtaining
Zs D
i
0
0E
w0f
H
^
wi expðixfi tÞdt
ð1Þ
bf ðsÞ bf ðsÞ ¼ ð3:270Þ
h
0
Finally, we take the square of the absolute value of both members to calculate
the probability of having the state with energy Ef0 :

s
2

Z

2 1
D 0
0
0 E

bf ðsÞ

wf H^ wi expðixfi tÞdt

ð3:271Þ
2

h

3.13 Absorption and Emission of Radiation
We now consider the interaction of a molecule with electromagnetic radiation,

where the molecule is dealt with quantum-mechanically but the electromagnetic
radiation is treated classically. A plane polarized electromagnetic radiation consists
of one electric field and one magnetic field whose amplitudes are sinusoidal waves
that travel in phase in mutually perpendicular planes. Electrons in atoms or
molecules interact mostly with the electric field. The mathematical expression for a
plane polarized electric field wave is given by
E x ðz; tÞ ¼ E 0 sinðxt kzÞ ð3:272Þ
where the electric field E 0 is the amplitude of E x , is polarized in the x direction,

propagates along z, and
2p 2p
x¼ ¼ 2pv k¼ ¼ 2p~v ð3:273Þ
s k
where x is the angular frequency (s is the period of the radiation and m its fre-
quency), and k is the angular wavenumber (k is the wavelength, its inverse being
the wavenumber). The electric field (the electric force per unit charge) interacts with
charges in the molecule, electrons, and nuclei. The force on charge q is
F ¼ E x q ¼ dV=dx, where V is the potential energy. Integration of the latter
equation gives V ¼ E x qx. Therefore, for a distribution of charges and an x-
polarized electric field, the perturbation is given by
X X
^ 0 ðtÞ ¼ E x
H qa xa ¼ E 0 sinðxt kzÞ qa x a ð3:274Þ
a a
where qa represent the charges and xa their positions along x. Substitution of (3.274)
in (3.270) gives
Zs *

+

X
iE 0

bf ðsÞ w0f
qa xa
w0i sinðxt kzÞ expðixfi Þdt ð3:275Þ
h
a

0
3.13 Absorption and Emission of Radiation 173
Note that the wavelength of ultraviolet radiation is of order 102 nm, and visible
and infrared radiations have larger wavelengths. Hence, the term kz in the argument
of the sine function is small, since k is proportional to the inverse of the wavelength
[see (3.273)]. In addition, the order of magnitude of the largest dimension of a
typical molecule is 1 nm, and so w0i and w0f are vanishingly small for distances
greater than 1 nm. Hence, using the so-called long wavelength approximation, we
can replace (3.275) by the simpler expression
Zs *

+

X
iE 0

bf ðsÞ w0f
qa xa
w0i sinðxtÞ expðixfi tÞdt ð3:276Þ
h
a

which, in turn, can be replaced by

*

+ Zs

X
iE 0

bf ðsÞ w0f
qa xa
w0i sinðxtÞ expðixfi tÞdt ð3:277Þ
h
a

since w0i and w0f are time-independent wave functions and the dipole moment
operator does not explicitly depend on time. Substitution of
expðixtÞ expðixtÞ
sinðxtÞ ¼ ð3:278Þ
2i
in (3.277) gives
*

+ Zs
E0
X

0

bf ðsÞ wf
qa xa
w0i exp½iðxfi þ xÞt exp½iðxfi xÞt dt
2
h
a
0
ð3:279Þ
In turn, use of
Zs
expðasÞ 1
expðatÞdt ¼ ð3:280Þ
a
0
leads to
*

+
E0
X
exp½iðxfi þ xÞs 1 exp½iðxfi xÞs 1

0
bf ðsÞ 0
wf
q x
w
2
hi
a a a
i xfi þ x xfi x
ð3:281Þ
The first and second terms of the expression in braces lead to 0/0 indeterminate
forms for x = −xfi and x = xfi, respectively, which can be solved by taking the
following limits:
exp½iðxfi xÞs 1
lim ¼ is ð3:282Þ
xfi x!0 xfi x
(E41). These limits can be easily obtained by the quotient of derivatives of the
numerator and denominator, according to l’Hôpital’s rule.
Considering the square of the absolute value of the expression in braces in
(3.281), Fig. 3.11 plots the resulting band shape as a function of x in a neigh-
borhood of xfi. The maximum at x = xfi corresponds to an ascending transition
from i to f that leads to absorption of energy from the electromagnetic radiation. In
turn, the maximum at x = −xfi corresponds to a downward transition from f to
i that leads to stimulated emission of electromagnetic radiation, a radiation process
found in lasers.
The probability of the spectroscopic transition from i to f is given by the square
of the absolute value of bf [see (3.281)]. If the radiation has additional electric field
components along y and z, then we have to add the squares of the absolute values of
(3.281) for the corresponding y and z terms and separate the square of the band
shape factor in (3.281) (see Fig. 3.11), being left with

D E
2
D

E
2
D

E
2

0

^x jw0i
þ
w0f
l

wf jl ^y
w0i
þ
w0f
l
^z
w0i
ð3:283Þ
For the i to f transition probability to be significant, at least one of the above

integrals should be different from zero. We come back to this point in Sect. 4.5 to
discuss the implications of molecular symmetry.
Fig. 3.11 Band shapes for emission (x = −xfi) and absorption (x = xfi) as functions of x. The
units are arbitrary. Figure obtained with Mathematica
3.13 Absorption and Emission of Radiation 175
3.13.1 Spontaneous Emission of Radiation
Both absorption and stimulated emission of radiation are induced by radiation and
have transition probabilities proportional to the spatial density of radiation, d(x).
The probability of absorption per unit of radiation density and unit of population of
the initial state is called for historical reasons the Einstein B coefficient for
absorption and is equal to the Einstein coefficient for stimulated emission, thus
being represented by the same letter. The coefficient for absorption and stimulated
emission is proportional to the square of the transition dipole moment (3.283).
Let us consider two nondegenerate quantum states a and b, where b is the
excited state. The equality between Einstein coefficients for absorption and stim-
ulated emission implies the occurrence of transitions that are independent of the
radiation density, thus being associated with spontaneous emission. If spontaneous
emission did not exist, state b would acquire the same population as state a, and
then it would not be possible to observe the transition between these states spec-
troscopically. The probability of spontaneous emission per unit of population of
the initial state is called for historical reasons the Einstein A coefficient.
At equilibrium, the increase in the population of state b is balanced with the
decrease of this population, and we can write
Bdqa Bdqb Aqb ¼ 0 ð3:284Þ
where qa and qb are the populations of the a and b states, the first term is the
probability for absorption, the second term stands for the probability of stimulated
emission, and the last term accounts for the probability of spontaneous emission. At
equilibrium and a specified temperature, qa and qb are related by the Boltzmann
distribution, which is given by
qb ¼ qa expðhxba =kTÞ ð3:285Þ
where xba = xb − xa and k is the Boltzmann constant (k = 1.3806488 10−23

J K−1; see Appendix). Substitution of (3.285) in (3.284) leads to
A ¼ Bd ½expðhxab =kT Þ 1 ð3:286Þ
We now need to know the radiation density d. Using the spatial density of
radiation in equilibrium with the blackbody at a defined temperature, we write,
according to Planck’s law,
d / x3 ½expðhx=kT Þ 11 ð3:287Þ
where the symbol / stands for “proportional to.” Substitution of (3.287) in (3.286)
gives
expðhxab =kTÞ 1
A / Bx3 ð3:288Þ
expðhx=kTÞ 1
As mentioned before [see (3.281) and Fig. 3.11], B takes nonnegligible values in
the resonance condition, that is, when x = xab. Hence, (3.288) can be replaced by
A / Bx3 ð3:289Þ
This result shows that spontaneous emission is increasingly important for high
frequencies and justifies the difficulties in obtaining laser radiation in the ultraviolet
region.
3.14 Raman Scattering
In 1928, the Indian physicist C.V. Raman (1888–1970) experimentally showed that
a small fraction of a monochromatic beam of visible light is scattered by a sample
with change of frequency (inelastic light scattering) and that the observed shifts in
frequency depend on the chemical structure of the molecules responsible for the
scattering. For this discovery, Raman was awarded the 1930 Nobel Prize in physics.
Light scattering involves incident and scattered photons. Most of the latter are
scattered with the same frequency as the incident photons (elastic or Rayleigh
scattering, E42), and a small fraction of them are scattered at a different frequency
(inelastic or Raman scattering). In Rayleigh scattering, the energies of the photon
and molecule are separately conserved, whereas in Raman scattering, energy con-
servation is observed for the whole radiation–molecule system, that is, frequency
changes in the radiation and molecule are of equal magnitude but opposite signs. In
Raman scattering, Stokes transitions are those in which the scattered photons lose
energy to the molecule, whereas anti-Stokes transitions are those in which the
scattered photons gain energy from the molecule. Figure 3.12 presents the Rayleigh
and Raman spectra of carbon tetrachloride (liquid) according to the IUPAC
(International Union of Pure and Applied Chemistry) recommendations, where the
abscissa is a linear scale occupied by the Raman shifts presented in cm−1 and
decreasing from left to right (the frequency of the scattered light increases from left
to right, as it should), and the ordinate is linear and proportional to the intensity.
The intensity of Raman scattering is several orders of magnitude lower than the
intensity of Rayleigh scattering. The distribution of Stokes and corresponding
anti-Stokes lines on each side of the Rayleigh line is symmetric in frequency, but
Stokes lines are more intense than the corresponding anti-Stokes lines (Fig. 3.12).
For this reason, a Raman spectrum usually records only the Stokes side of the
spectrum.
3.14 Raman Scattering 177
Fig. 3.12 Rayleigh and Raman spectra of carbon tetrachloride (liquid) presented according to
IUPAC recommendations. Details of band shapes and band intensities should be considered
schematic. This figure was obtained using Mathematica with each band simulated by a linear
combination of Lorentzian and Gaussian functions
3.14.1 Classical Treatment
The electric field of electromagnetic radiation induces an electric dipole moment in

a molecule that is given by
lind ¼ aE ð3:290Þ
where the induced dipole moment and the electric field are represented by column
vectors in the (xyz)T basis, and the polarizability tensor is represented by a 3 3
symmetric matrix
0 1
axx axy axz
a ¼ @ ayx ayy ayz A ð3:291Þ
azx azy azz
(E43). In the following equalities of this subsection, we assume that the polariz-
ability is spherical (axx = ayy = azz = a and off-diagonal elements equal to zero) to
make the equations simpler without unnecessary subscripts. The electric field of a
plane-polarized electromagnetic radiation varies in time with the frequency of the
radiation m0 and is given by
E ¼ E 0 cosð2pv0 tÞ ð3:292Þ
In turn, the polarizability changes with the vibrational coordinate involved in the
Raman transition, and its expansion in a Taylor series gives

@a
a ¼ a0 þ Qþ ð3:293Þ
@Q 0
where the vibrational coordinate Q has a specific frequency and is given by
Q ¼ A cosð2pvtÞ ð3:294Þ
Substitution of (3.292), (3.293), and (3.294) in (3.290) gives, to first order,

@a
lind ¼ a0 E 0 cosð2pv0 tÞ þ AE 0 cosð2pv0 tÞ cosð2pvtÞ ð3:295Þ
@Q 0
Using simple trigonometry [cos (a + b) = cosa cosb − sina sinb], one obtains

1 @a
lind ¼ a0 E 0 cosð2pv0 tÞ þ AE 0 fcos½2pðv0 þ vÞt þ cos½2pðv0 vÞtg
2 @Q 0
ð3:296Þ
This expression shows that the frequency m0 corresponds to the Rayleigh scat-
tering and involves the static polarizability a0, whereas the Stokes and anti-Stokes
Raman frequencies m0 − m and m0 + m involve the polarizability derivative with
respect to the vibrational coordinate (∂a/∂Q)0 (E44). A Raman active vibration
requires

@a
6¼ 0 ð3:297Þ
@Q 0
with the Raman intensity being proportional to the square of the polarizability
derivative with respect to the vibrational coordinate.
3.14.2 Quantum-Mechanical Treatment
While the quantum-mechanical treatment of absorption and stimulated emission of

radiation led to the electric dipole transition integrals in (3.283), the
quantum-mechanical treatment of light scattering considers the first-order induced
transition electric dipole moment, whose q component (q can be x, y, or z) is given by
X n o
lðq1Þ ¼ aqr fi E r0 exp i x0 xfi
fi
r¼x;y;z
where er0 is the electric field amplitude along r (r can be x, y, or z) (see Long,
Further Reading, Sect. 4.2). Separation of the electric field [see (3.290)] in this
expression leads to the transition polarizability, whose qr component between the
initial and final states i and f is given by
0D

1
1 X @ wf
^
0
lq
w0r w0r
l
^r
w0i lr
w0r w0r
l
w0f
^ ^q
w0i
aqr ¼ þ A ð3:298Þ
fi h r6¼i;f xri x0 iCr xrf þ x0 þ iCr
where w0i , w0f , and w0r are time-independent unperturbed wave functions for the
initial state i, the final state f, and the intermediate state r (these wave functions are
assumed to be real), l ^q is the dipole moment operator, x0 is the angular frequency
of the incident radiation,
x0 ¼ E0 =h ð3:299Þ
and
xri ¼ xr xi ð3:300Þ
The relation between the time-dependent and time-independent wave functions

for state r is given by
Wr ¼ wr exp½iðxr iCr Þt ð3:301Þ
where Cr is the half-width of state r, which is related to the lifetime sr through the
energy–time uncertainty principle, as shown below:
1
DEDt h ) DxDt 1 ) 2Cr sr 1 ) sr ð3:302Þ
2Cr
Each state r is not stationary, that is to say, is not a solution of the time-independent
Schrödinger equation, and so does not have a defined energy, and is therefore called a
virtual state. In turn, the initial and final states i and f are stationary, have defined
energies, i.e., DEi = DEf = 0 (Ci = Cf = 0), and so their lifetimes are infinite (the
physical attributes of states i and f never change). As a result,
Wi ¼ wi expðixi tÞ Wf ¼ wf expðixf tÞ ð3:303Þ

Note that (3.298) corresponds to a second-order energy correction of perturba-

tion theory and to a first-order correction in the wave function [see (3.225), (3.226),
and (3.227) and compare (3.298) with (3.227)].
We now consider the transition polarizability (3.298) and the cases that result
from increasing the angular frequency x0 of the exciting radiation (see Long,
Further Reading, Sect. 4.3). The first case is that in which x0 is much smaller than
xri and xrf for all states r and the half-widths Cr are small relative to xri and xrf [Cr
can be neglected in both terms of (3.298)]. Since the denominators of (3.298) then
become approximately equal to xri and xrf and the summation over r is infinite, no
direct information can be obtained on the virtual states r, and this type of Raman
scattering, called normal Raman scattering (Fig. 3.13a), is essentially a
ground-state property. The states r that significantly contribute are those for which
the numerators are different from zero; that is, the transition dipole moments from
the state i to a virtual state r and from a state r to the final state f are both different
from zero, and the weighting of these contributions is given by the inverse of the
corresponding denominators, which are approximately equal to xri and xrf, as we
have mentioned above.
When x0 approaches one xri transition frequency, the denominator
xri − x0 − iCr becomes appreciably smaller, due to partial cancellation in the
difference xri − x0, and this type of light scattering is called preresonance Raman
scattering (Fig. 3.13b).
When x0 xri for a particular state r or a limited number of states r, then
xri − x0 − iCr becomes approximately equal to iCr, the corresponding states
r dominate the sum in (3.298), and the Raman intensity can drastically increase; this
(a) (b) (c) (d)
Fig. 3.13 Raman scattering processes: a normal Raman scattering, x0 « xri; b preresonance
Raman scattering, x0 ! xri; c discrete resonance Raman scattering, x0 xri; d continuum
resonance Raman scattering, x0 in the range of dissociative continuum levels. The potential energy
curve was obtained with Mathematica using a Morse function to which terms in x−6, x−8, and x−10
were added
type of scattering is called discrete resonance Raman scattering (Fig. 3.13c).

Detailed information on the selected states r and their lifetimes can then be
obtained. Since the denominator xrf + x0 + iCr of (3.298) involves the sum of xrf
and x0, cancellation does not occur, and so the corresponding terms become rel-
atively unimportant in the sum over r.
Finally, when x0 falls in the range of dissociative continuum levels, we have
what is called continuum resonance Raman scattering (Fig. 3.13d).
3.15 Molecular Calculations
Molecular calculations are usually performed within the Born–Oppenheimer

approximation and comprise three major methods for solving the Schrödinger
equation: ab initio, density functional theory (DFT), and semiempirical methods.
3.15.1 Computational Methods
The Schrödinger equation of ab initio or first principles calculations does not

include adjustable parameters, and all integrals are evaluated exactly. The electronic
Hamiltonian of these calculations includes fundamental physical constants such as
the Planck constant, electron mass, and electron charge. Therefore, ab initio cal-
culations do not lead to compensation errors, thus having a good reputation,
especially in comparison with semiempirical calculations. However, in molecules
with tens of atoms, they may be prohibitively time-consuming.
A method due to Moller and Plesset (1934) and applied to many-electron sys-
tems uses perturbation theory to obtain an approximation to the electron correlation
energy. The unperturbed Hamiltonian is the sum of Fock operators, and the per-
turbation is the difference between the true interelectronic repulsions and the
Hartree–Fock interelectronic potential energy. When the energy is truncated at the
second-order perturbation correction, the method is called MP2 for Moller–
Plesset-2.
DFT calculations are a good alternative to ab initio calculations because they are
less time-consuming and present results of comparable quality. A popular molec-
ular DFT calculation much used in this book is known by the acronym B3LYP,
which stands for Becke’s 3-parameter exchange functional (Becke 1993) and Lee,
Yang, and Parr correlation functional (Lee et al. 1988; Miehlich et al. 1989). DFT
calculations with the B3LYP exchange-correlation functional generally give
geometries and frequencies in close agreement with experimental results. When the
number of atoms in a molecule is sufficiently large that the DFT calculation
becomes too time-consuming, semiempirical calculations are a possible alternative.
In semiempirical calculations, molecular integrals with negligible values are
taken as zero, and integrals that can be assigned some physical meaning are replaced
by parameters obtained by fitting experimental data (empirical parameters). These
approximations make the calculations faster, thus becoming amenable to molecules

with hundreds of atoms. Semiempirical methods are frequently used with success in
organic molecules. These have a reduced number of different atoms, the most
common being C and H, with O, N, P, and S occurring less frequently, and a limited
number of local geometries, the more frequently encountered being the tetrahedral
and planar triangular geometries. The use of empirical parameters inevitably leads to
compensation of errors and to results whose agreement with experimental values is
more apparent than real.
3.15.2 Gaussian-Type Functions
Orbitals of a quantum-mechanical molecular calculation are expanded in a set of

basis functions, usually Gaussian-type functions or linear combinations of these,
centered on the molecule’s nuclei. A Cartesian Gaussian is given by the general
expression
Nijk;a xi yi zk ear
2
ð3:304Þ
where Nijk,a is the normalizing constant, x, y, and z represent Cartesian coordinates

with the origin at a particular nucleus, i, j, and k are nonnegative integers, r is the
distance to the nucleus, and a is a positive orbital exponent. Values of the sum
i + j + k equal to 0, 1, and 2 correspond to s-, p- and d-type Gaussians, respec-
tively. Each Slater-type orbital (STO) centered on a particular atom of a molecule is
approximately given by a linear combination of Gaussian-type functions:
X
/ di gi ð3:305Þ
i
Linear combinations of Gaussian functions are called contracted Gaussians,

and the individual Gaussians of each contracted Gaussian are called primitive
Gaussians. The number of Gaussians of a particular contracted Gaussian is called
the degree of contraction, and the coefficients are called contraction coefficients.
The potential energy of attraction between an electron and an atomic nucleus
becomes infinite at the nucleus, since then the electron–nucleus distance is zero. As
a result of this singularity, the wave function has a discontinuous first derivative at
the nucleus, and the atomic orbital is said to have a cusp at the nucleus. This
behavior contrasts with that of a Gaussian function that is continuous and has a
continuous first derivative. Mathematica code M9 illustrates this point and plots a
1s STO and an s-type Gaussian function for comparison (Fig. 3.14).
Consider now the normalized 1s radial function of a hydrogen atom,
R10 ðrÞ ¼ 2er ð3:306Þ

3.15 Molecular Calculations 183
Fig. 3.14 While a 1s STO (black) has a cusp at the nucleus, the s-type Gaussian function (gray)
is continuous in both the function and its derivative. These graphs were obtained with
Mathematica
where r represents the electron–nucleus distance in bohrs (see Appendix). The

angular factor for the 1s orbital, the spherical harmonic Y00 , is given by the constant
1/(2√p). Figure 3.15 illustrates the fitting of (3.306) with a linear combination of
three normalized s-type Gaussian functions, whose exponents were taken from the
STO-3G basis set for the hydrogen atom, and the coefficients were adjustable
parameters (see Mathematica code M10; E45).
Consider two-electron integrals of a molecular calculation. When the orbitals are
expanded in a basis set of functions, the two-electron integrals may involve basis
functions centered on one, two, three, or four nuclei (Fig. 3.16). In particular,
integrals involving basis functions centered on three and four nuclei are very
time-consuming. Gaussian functions make the calculation much easier, because the
product of Gaussian functions located at two different positions with different
Fig. 3.15 At the left, the figure shows the radial factors of three s-type Gaussian functions and the
1s orbital of the hydrogen atom. At the right, a linear combination of the Gaussian functions is
fitted to the 1s orbital of the hydrogen atom. The inset shows the cusp of the 1s orbital at the
nucleus, contrasting with the absence of a cusp of the Gaussian function. These graphs were
Fig. 3.16 Gaussian functions centered at four points are reduced to two Gaussian functions
centered at two points
Fig. 3.17 The product of two Gaussian functions centered at different points is a third Gaussian
function centered on an intermediate point. These graphs were obtained with Mathematica
exponents is a third Gaussian function located at an intermediate point (Fig. 3.17).

To prove this statement, consider the following one-dimensional Cartesian Gaus-
sians centered at A and B,
2 2
gA ðxÞ ¼ Na eaðxAÞ gB ðxÞ ¼ Nb ebðxBÞ ð3:307Þ
The product of these Gaussians gives the following Gaussian function:

2
gC ðxÞ ¼ gA ðxÞgB ðxÞ ¼ Na Nb ed ecðxCÞ ð3:308Þ
3.15 Molecular Calculations 185
where
ab aA þ bB
d¼ ðA BÞ2 c ¼ aþb C¼ ð3:309Þ
aþb aþb
Equation (3.308) can be easily proved by showing that the sum of the exponents
of gA and gB is equal to the exponent of gC, that is,
h i
aðx AÞ2 þ bðx BÞ2 cðx CÞ2 þ d ¼ 0 ð3:310Þ
or, after substitution of (3.309),

2 2 aA þ bB ab 2
aðx AÞ þ bðx BÞ ða þ bÞ x þ ðA BÞ ¼ 0
aþb aþb
ð3:311Þ
This result is easily proved using Mathematica (see Mathematica code M11).
3.15.3 Standard Basis Sets
We now consider some of the more popular standard basis sets. We begin with a
minimal basis set, i.e., a basis set with one Slater-type orbital (STO) per each
inner-shell or valence-shell atomic orbital. The STO-3G basis set is a minimal basis
set, where each STO is represented by a linear combination of three Gaussian
functions. For the carbon atom, this basis set consists of five STOs (one 1s + one
2s + three 2p STOs) and three Gaussian functions for each STO, that is, a total of
15 Gaussian functions.
A double-zeta basis set consists of two STOs with distinct f orbital exponents
for each atomic orbital, irrespective of this being an inner-shell or a valence-shell
orbital. In turn, a split-valence basis set consists of two or more STOs per
valence-shell atomic orbital, and only one STO per inner-shell atomic orbital. The
3-21G basis set is a valence double-zeta basis set consisting of three primitive
Gaussian functions per inner-shell atomic orbital and a split-valence shell with one
contracted Gaussian function as a linear combination of two primitive Gaussian
functions and an additional single primitive Gaussian function. Another example of
a valence double-zeta basis set is given by the 6-31G basis set. The basis set
6-311G is an example of a valence triple-zeta basis set.
Atomic orbitals become distorted and have their centers shifted upon molecule
formation. They are said to be polarized. Polarization is accounted for by adding
STOs with ‘ quantum number greater than the maximum ‘ of the valence shell of
the ground state atom. In the CH4 molecule, the 6-31G** basis set adds a set of five
d-type Cartesian Gaussian functions to the carbon atom (the first star) and a set of
three p-type Cartesian Gaussian functions to each hydrogen atom (the second star).
In turn, the 6-31++G** basis set includes a set of four Gaussian functions, one s-
type and three p-type functions, with very low exponents in the carbon atom (the
first plus sign) and one s-type Gaussian function with very low exponent in each
hydrogen atom (the second plus sign). These functions fall off very slowly as
r increases and are called diffuse functions (E46).
Considering an sp valence shell, the inclusion of orbitals of higher angular
momentum in the basis set should take into account the effect of each additional
orbital in the lowering of the total electronic energy through its contribution to the
correlation energy. For example, the lowering of energy associated with the addi-
tion of a set of d orbitals is approximately of the order of magnitude of one f orbital.
In turn, the addition of a second set of f orbitals is equivalent to the lowering of
energy caused by one g orbital. These facts were considered by Dunning, who
proposed the so-called correlation-consistent (cc) basis sets (Dunning 1989). For
the carbon atom, the cc-pVTZ (correlation-consistent polarized valence triple-zeta)
basis set consists of 4s, 3p, 2d, 1f sets of contracted Gaussian functions
(4 1 + 3 3 + 2 5 + 1 7 = 30 functions). For the hydrogen atom, we
have 3s, 2p, 1d sets of contracted Gaussian functions (3 1 + 2 3 + 1 5 =
14 functions). These basis sets may be augmented with diffuse functions by adding
the prefix aug- (for example, aug-cc-pVTZ).
Computing a thermodynamic energy, such as the atomization energy or the
enthalpy of formation, with chemical accuracy means attaining an accuracy of
1 kcal/mol. However, the largest errors in ab initio thermodynamic calculations
result from basis set truncation. Therefore, to achieve chemical accuracy, the total
energy should be computed from the results of a series of calculations that
extrapolate to the complete basis set limit.
Notes
§1. Particle in a One-Dimensional Box
A particle of mass m subjected to a potential energy that is infinite everywhere

along the x-axis, except for a line segment of length ‘ where it is zero, is a particle
in a one-dimensional box. The time-independent Schrödinger equation is given by
2 d2 w
h
þ VðxÞwðxÞ ¼ EwðxÞ 0\x\‘: VðxÞ ¼ 0 0
x
‘: VðxÞ ¼ 1
2m dx2
ð3:312Þ
Outside the box, that is, for V(x) = ∞, it is easy to conclude that the wave
function is zero, i.e., the probability of finding the particle outside the box is zero,
since substitution of V(x) = ∞ in the Schrödinger equation leads to
Notes 187
d2 w 1 d2 w
2
¼ 1wðxÞ ) ¼ 1 ) wðxÞ ¼ 0 ð3:313Þ
dx wðxÞ dx2
Inside the box, V(x) = 0, and the Schrödinger equation is
d2 w
þ k2 wðxÞ ¼ 0 ð3:314Þ
dx2
where
ð2mEÞ1=2
k¼ ð3:315Þ
h
The solution for a differential equation like (3.314) is a linear combination of
sine and cosine functions,
wðxÞ ¼ A cosðkxÞ þ B sinðkxÞ ð3:316Þ
(see M2), where A and B are determined by the boundary conditions. At x = 0,

w(0) = 0, the sine function is zero, cos(0) = 1, and so A = 0. Therefore,
wðxÞ ¼ B sinðkxÞ ð3:317Þ
The second boundary condition allows one to determine the energy, which
becomes dependent on the quantum number n. In fact, at x = l, w(l) = 0, B 6¼ 0,
and
" #
ð2mEÞ1=2 ð2mEÞ1=2
sin ‘ ¼0) ‘ ¼ np ð3:318Þ
h h
For n = 0, one has E = 0 and k = 0 [see (3.315)]. From (3.314), we conclude

that w(x) is zero everywhere, and so n = 0 is not allowed. Therefore, from the last
equation we conclude that
n2 h2
E¼ n ¼ 1; 2; 3. . . ð3:319Þ
8ml2
Using (3.315) and (3.318), we obtain
np
k¼ ð3:320Þ
‘
B is the normalization constant of the wave functions, which can be easily deter-
mined using Mathematica (see Answer to E8) and gives
1=2
2
B¼ : ð3:321Þ
‘
Therefore,
1=2
2 npx
wn ðxÞ ¼ sin ð3:322Þ
‘ ‘
§2. Two-Particle Rigid Rotor
Consider a two-particle system with the particles held at a fixed distance d from
each other. The energy of the rotor is entirely kinetic energy (rotational energy).
Since the potential energy is zero, the Hamiltonian is
^ ¼ h r2
2
H ð3:323Þ
2l
where
1 1 1
¼ þ ð3:324Þ
l m1 m2
In spherical coordinates, the Laplacian operator takes the expression
@2 2@ 1 ^2
r2 ¼ þ L ð3:325Þ
@r 2 r @r r 2 h2
[see (3.100)]. Since the rotor is rigid, r is constant (r = d), and therefore we can
write
1 ^2
r2 ¼ L ð3:326Þ
d 2 h2
^ ¼ 1 ^2
H L ð3:327Þ
2ld2
Therefore, the Schrödinger equation is given by
^ 1 ^2 m
Hwðh; /Þ ¼ Ewðh; /Þ ) L YJ ðh; /Þ ¼ EJ YJm ðh; /Þ ð3:328Þ
2ld2
Notes 189
[see first line of (3.91)], and the energy levels are
1 JðJ þ 1Þh2
JðJ þ 1Þ
h2 YJm ðh; /Þ ¼ EJ YJm ðh; /Þ ) EJ ¼ J ¼ 0; 1; 2; . . .
2ld2 2ld2
ð3:329Þ
Note that the energy levels depend on J only, but the eigenfunctions depend on
J and m [see (3.328) and (3.329)], where mℏ is the z-component of the rigid rotor
angular momentum. This means that each specified J level has a (2J + 1)-degen-
eracy, since m takes values from −J to J corresponding to different orientations of
the rigid rotor angular momentum projection on a preferential direction that is
quantized.
The moment of inertia I of a system of n particles about a specified axis is
defined as
X
n
I mi ri2 ð3:330Þ
i¼1
where mi is the mass of the ith particle and ri is the distance from the particle to the
axis along the perpendicular to the axis. For a two-particle rigid rotor, we have
I ¼ m1 r12 þ m2 r22 ð3:331Þ
where
d ¼ r1 þ r2 ð3:332Þ
and r1 and r2 are measured from the center of mass, and so
m1 r1 ¼ m2 r2 ð3:333Þ
From Eqs. (3.324), (3.331), (3.332), and (3.333), it can be concluded that
I ¼ ld 2 ð3:334Þ
This result can be easily confirmed using Mathematica (see E26).

Mathematica Codes
M1. Wave Equation
This Mathematica code considers the one-dimensional wave equation (3.18) and
uses the Mathematica function DSolve for solving the differential equation for the
function u with independent variables t and x. The symbols @x;x u[t,x] and @t;t u
[t,x] stand for second-order partial derivatives of u[t,x] with respect to x and
t. In general, C[1] and C[2] are default parameters generated by Mathematica,
but in this case. they represent arbitrary functions. In the second line of code, the
first level of sol (a list of one sublist; see the first line of results) is replaced in u
[t,x], thus giving the general solution of the differential equation presented in a
more familiar way (see second line of results).
Suggestion: Insert the option GeneratedParameters!f in the first line of
code. What is the effect of this change?
M2. Helmholtz Equation
This Mathematica code solves the so-called Helmholtz equation [see (3.20)].
The notation @x;x D[f[x],{x,2}] represents the second derivative of f[x] with
respect to x; C[1] and C[2] are default generated parameters, which in this case
represent arbitrary coefficients of a linear combination of Cos and Sin functions
(see second line of results).
M3. Harmonic Oscillator
Consider the one-dimensional Schrödinger equation for the harmonic oscillator and
its eigenfunctions [see (3.62)], which are the product of a Gaussian function and
polynomial factors known as Hermite polynomials. The built-in Mathematica
function HermiteH[n,z] generates Hermite polynomials for nonnegative values
of n. The following line of code gives the first six Hermite polynomials.
The normalization of an eigenfunction requires the integral of the square of the

eigenfunction [see (3.17)],
wn ¼ Nn /n
Z1 Z1
wn dz ¼ Nn
2 2
/2n dz ¼ 1
1 1
Therefore, the normalization of a harmonic oscillator eigenfunction implies the

inverse of the square root of the integral of the square of the eigenfunction, as
shown below:
1
Nn ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
R1 2
/n dz
1
The normalization factors are given by
1
pffiffiffiffiffiffiffiffiffi
p1=4 2n n!
a result that can be easily confirmed by the following Mathematica code:
The following Mathematica code plots the first six probability densities of the
one-dimensional Schrödinger equation for the harmonic oscillator. The first line of
code defines the function f[n,z], and the symbol := (SetDelayed) considers the
right-hand side of the expression to be the delayed value of the left-hand side. The
right-hand side is evaluated each time the left-hand side appears. The result of this
code is presented as a Table of graphs.
Suggestion: Plot the first six Hermite polynomials in the range {z,-5,5}.
The following code confirms that the wave functions (3.67) are normalized and
mutually orthogonal:
M4. Spherical Harmonics
Mathematica has the built-in function SphericalHarmonicY for spherical

harmonics. The following code lines give the mathematical expressions for
spherical harmonics with ‘ = 0, 1 and 2:
The following Mathematica code generates 3D plots in which the absolute value
of the spherical harmonic is plotted as the distance from the origin to the point, for
the angles h and /. The option Lighting!“Neutral” specifies that the sim-
ulated lighting is a source of white light.
M5. Determinants
The following Mathematica code illustrates the properties of determinantal func-

tions by confirming (3.118), (3.119), and (3.120) using 3 3 matrices and their
determinants:
M6. Systems of Homogeneous Linear Equations
The following Mathematica code illustrates the process of solving systems of

homogeneous linear equations using 2 2 and 3 3 matrices [see (3.134),
(3.136), (3.137), and (3.138)]. In the 2 2 matrix, a11 corresponds to H11-W(1)S11,
a12 corresponds to H12-W(1)S12, a21 corresponds to H12-W(1)S12, a22 corresponds
to H22-W(1)S22, and the determinant of the 2 2 matrix is given by (3.138):
M7. Normalization Constants for Slater-Type Orbitals
The following Mathematica code computes the normalization constants of the

radial factors of Slater-type orbitals for n 3 and confirms the general expression
shown in (3.158) for n 3.
M8. Functional Derivative
Mathematica provides a built-in function for evaluating functional derivatives. It is

named VariationalD and requires Needs[“VariationalMethods`”].
This instruction loads an otherwise not included file in the available packages of a
Mathematica session that is necessary for evaluating functional derivatives. The
following Mathematica code illustrates the evaluation of functional derivatives [see
(3.188) (the equation for obtaining the functional derivative) and (3.189)]:
M9. STO Versus a Gaussian Function at the Origin
The following Mathematica code compares the derivatives at the origin of a

1s Slater-type orbital (STO) and an s-type Gaussian function. Since the derivative
of the STO is not a continuous function at the origin, one needs to evaluate the left
and right derivatives. The Mathematica function Limit finds the limiting value of
a function, in this case, the ratio of differences in the dependent and independent
variables, as the difference in the independent variable tends to zero. The Limit
function has the option Direction, which takes the values 1 and −1 corre-
sponding to increasing and decreasing approaches to the limit, respectively:
M10. Fitting Gaussian Functions to a 1s Hydrogen Orbital
The following Mathematica code begins by obtaining normalization factors for the
1s function of the hydrogen atom and three s-type Gaussian functions. Being s func-
tions, they depend only on the r coordinate, not on the angular spherical coordinates h
and /, and consequently, integration is over R2(r)r2dr [see (3.105)]. After normal-
ization, the code confirms the normalizing constants. In order to find numerical values
for the Gaussians coefficients that give the best fit to data, the code prepares a list of
lists called data with regularly spaced points from the function f[r] and defines the
variable model. FindFit finds numerical values for the coefficients c1, c2, and
c3 that lead to the best fit of model to data. In the line of code
c1 ¼ c1=:fit½½1
fit[[1]] stands for the first part of fit, which is c1!0.122444, c1/.fit[[1]]
replaces c1 by 0.122444, and c1=c1/.fit[[1]] assigns 0.122444 to c1.
Similar lines of code are repeated for c2 and c3. The resulting values of c1, c2,
and c3 are presented as output. The smallest coefficient (c1) corresponds to the
Gaussian function with the highest exponent, that is, the function that does more
effective fitting near the nucleus:
M11. Product of Gaussian Functions
The following line of Mathematica code shows that the product of Gaussian
functions located at two different positions with different exponents gA and gB is a
third Gaussian function gC located at an intermediate point [see (3.308)], by
demonstrating that the sum of the exponents of gA and gB is equal to the exponent of
gC [see (3.311)]. This code illustrates the use of FullSimplify, which simplifies
its argument (the expression to which it applies), and of TrueQ, which yields the
result True or False for the expression to which it applies:
Another way of arriving at the same conclusion consists in substituting (3.309)

in the left-hand side of (3.310), FullSimplify the result, and obtaining zero.
The following Mathematica code illustrates the product of two Gaussians

located at different positions and with different exponents by plotting the Gaussian
factors and the Gaussian result:
Glossary
Ab initio calculations Do not include adjustable parameters, and all

molecular integrals are evaluated exactly. The
electronic Hamiltonian of these calculations
includes fundamental physical constants such as the
Planck constant, electron mass, and electron charge.
Therefore, these calculations do not lead to
compensation errors, as semiempirical
calculations do.
Anharmonic Applied to an oscillating system that is not
undergoing simple harmonic motion. The Taylor
expansion of the Morse potential energy function
about the equilibrium point Re (the minimum)
gives a term involving the square of (R − Re) that
represents the harmonic term and cubic, quartic,
and following powers of (R − Re) that account for
the Morse potential anharmonicity; see (3.71).
Antisymmetry principle States that the wave function changes sign when
the coordinates of two electrons including the spin
are interchanged; see (3.110) for a wave function
of two electrons.
Born–Oppenheimer Deals with the separation between electron and
approximation nuclear coordinates in the Schrödinger equation in
two steps. In the first step, the kinetic energy of the
nuclei is assumed to be zero, and the Schrödinger
equation is solved for the electronic Hamiltonian;
see (3.140). The second step solves the Schrödinger
equation for the nuclear motion; see (3.142).
Commutator Is defined by ½A;^ B
^ A ^B^B ^ The order of
^ A.
application of two commuting operators is
irrelevant (their commutator is equal to the null
operator).
Correlation energy Given by the difference between the exact
nonrelativistic energy and the Hartree–Fock
energy of a molecule. A Hartree–Fock SCF wave
function takes the interactions between electrons
in an average way (each electron “feels” the
average repulsion of all the other electrons),
whereas interelectronic repulsions are
instantaneous in reality, and the motions of
electrons in a molecule are correlated
(interdependent).
Glossary 201
Density functional theory Based on a theorem of Pierre Hohenberg and

Walter Kohn published in 1964 that says that the
electron probability density of a nondegenerate
ground state uniquely determines the Hamiltonian
operator, the ground state molecular energy, and
all other ground state molecular properties.
Dirac notation A notation introduced by Dirac to simplify the
representation of molecular integrals; see (3.43).
Functional A rule that associates a number with a function;
see (3.178) for an example of a functional.
Functional derivative For a functional J depending on the function
y(x) and an increment h(x) on the function y(x), the
functional derivative of J[y] with respect to y,
dJ/dy, is the limit of the quotient of J[y + h]
− J[y] and the area Dr lying between the curve
h(x) and the x-axis for an increment Dx, (J[y + h] -
J[y])/Dr, as Dr ! 0. The functional derivative is
given by (3.188).
Harmonic oscillator A single point mass m connected to a rigid wall by
a perfectly elastic spring. When the perfectly
elastic spring is distended or contracted, the force
that restores equilibrium is proportional to the shift
from the equilibrium position, F = −kx; see (3.52).
Hartree–Fock A method for calculating atomic and molecular
energies whereby for a closed-shell electronic
configuration, the wave function is a single Slater
determinant of spin-orbitals that minimizes the
total energy. Each electron is subject to the
average Coulomb repulsion of all the other
electrons, to which is added the exchange
potential, which results from the antisymmetry
requirement of the wave function.
Hermitian operator A linear operator that satisfies (3.27) or (3.28) for
all well-behaved functions. The eigenvalues of a
Hermitian operator are real numbers.
Morse potential The empirical representation of the potential
energy function for a diatomic molecule given by
U(R) = D {1 − exp[−b(R − Re)]}2; see (3.68).
Reduced mass For a diatomic oscillator with masses mA and mB,
its inverse is equal to the sum of the inverses of the
masses mA and mB; see (3.51).
Semiempirical Have molecular integrals with negligible values

calculations taken as zero and integrals that can be assigned
some physical meaning replaced by parameters
obtained by fitting experimental data (empirical
parameters). These approximations make
calculations faster and amenable to molecules with
hundreds of atoms. The use of empirical
parameters inevitably leads to compensation of
errors and to results whose agreement with
experimental values is more apparent than real.
Spherical harmonics Are eigenfunctions of the square and z-component
of the orbital angular momentum operators; see
(3.91) and (3.92). The square of the angular
momentum operator is a function of h and /, and
the z-component of the angular momentum is a
function of /; see (3.89) and (3.90).
Spontaneous emission Independent of radiation density. The probability
of radiation of spontaneous emission per unit of population of
the initial state is usually called the Einstein
coefficient A. Spontaneous emission is
proportional to the cube of frequency [see
(3.289)], thus becoming increasingly important for
high frequencies.
Stimulated emission Induced by radiation. Both absorption and
of radiation stimulated emission of radiation are induced by
radiation and have transition probabilities
proportional to the spatial density of radiation. The
probability of absorption per unit of radiation
density and unit of population of the initial state is
usually called the Einstein coefficient B for
absorption and is equal to the Einstein coefficient
for stimulated emission, thus being represented by
the same letter.
Exercises 203
Exercises

1 1
E1. Find the eigenvalues and eigenvectors of the matrix A ¼ .
4 2
E2. Use
0 Mathematica 1 to find the eigenvalues and eigenvectors of the matrix
1 0 1
A ¼ @ 1 2 1 A.
2 2 3
E3. Show that ∂/∂x is a linear operator.
E4. Calculate the commutator between ∂/∂x and x.
E5. Relate the kinetic energy operator of the Hamiltonian to the linear
momentum operator.
E6. Let u be any twice differentiable function. Show that u(x − vt) is a solution
of the wave equation and a forward propagating wave with speed v.
E7. Confirm that Acos(kx − xt) is a solution of the wave equation.
E8. Consider a particle in a one-dimensional box. Use Mathematica to deter-
mine the constant B for the wave functions wn(x) = Bsin(npx/‘).
E9. Use Mathematica to plot the wave functions for the particle in a
one-dimensional box for n = 1, 2, 3.
E10. Use Mathematica to solve the equation y″(x) + y′(x) − 6y(x) = 0.
E11. Beginning with (3.27) for the definition of a Hermitian operator, arrive at
(3.28). [Hint: Consider w as the linear combination f + cg, where c is an arbi-
trary complex number.]
E12. Derive (3.50) for a diatomic molecule AB.
E13. Write the equation to obtain the reduced mass of a system of four masses,
mA, mB, mC, and mD.
E14. Show that (3.54) is a solution of the differential Eq. (3.53) of the classical
harmonic oscillator.
E15. Use Mathematica to list the first six Hermite polynomials.
E16. Use Mathematica to find the normalization constants of the first six har-
monic oscillator eigenfunctions assuming that a is equal to 1 and confirm that
the normalization constants satisfy the formula [p1/4√(2nn!)]−1.
E17. Confirm the recurrence equality (3.64) for H4(z), H3(z), and H2(z).
E18. Confirm (3.70).
E19. Derive (3.83).
E20. Derive the first of the equalities (3.86).
E21. Derive the first of the equalities (3.88).
E22. Use Mathematica to obtain polar plots of |sin(nh)|, where n takes integer
values from 1 to 4 and h ranges from 0 to p.
E23. Use Mathematica to write an expression for Y20 (h, 0) and make a polar plot
of this function.
E24. Derive (3.101).
E25. Write R2s in SI units.

E26. Consider the two-particle rigid rotor and the definition of moment of
inertia given by I ¼ m1 r12 þ m2 r22 . Using Mathematica, verify I = ld2, where
d = r1 + r2, l is the reduced mass, and m1r1 = m2r2.
E27. Derive (3.134).
E28. Use Mathematica to solve the system of inhomogeneous linear equations
x + y − z = 2, 2x − y + 3z = 5, 3x + 2y − 2z = 5.
E29. Use Mathematica to multiply two 3 3 matrices.
E30. Show that the equations of the system of homogeneous linear equa-
tions (3.137) give the same ratio of coefficients c2/c1.
E31. Write the equivalent of (3.147) for an atom with n electrons.
E32. Why is the term VNN included in (3.154)?
E33. Write the Hartree–Fock energy expression for the ground-state 1s22s2
electron configuration of the beryllium atom.
E34. Show that exchange integrals occur only between spin-orbitals of the same
spin.
E35. Use Dirac notation to express (3.155) and (3.156).
E36. Use Mathematica to obtain the normalization constant for the radial factor
of a Slater-type orbital with n = 2 and verify the result.
E37. Write (3.157) in SI units.
E38. Confirm (3.197), using the wave function (3.122).
E39. Confirm (3.199), using the wave function (3.122).
E40. Consider a hydrogen atom in its ground state perturbed by a uniform
electric field applied in the z direction and represent the excited state by a 2pz
Slater-type orbital. Use Mathematica to minimize and plot the variational per-
turbation upper bound to the second-order energy correction as a function of the
f2pz exponent.
E41. Confirm (3.282).
E42. The intensity of Rayleigh scattering is proportional to the inverse of the
fourth power of wavelength. Explain why the sky is blue.
E43. Use Mathematica to show that the trace of the polarizability tensor and
its anisotropy, whose square is given by c2 = [(axx − ayy)2 + (ayy − azz)2 +
(azz − axx)2 + 6 ða2xy þ a2yz þ a2zx Þ]/2, are invariants under an orthogonal trans-
formation due to a rotation of x and y around z, by an angle h.
E44. Find the ratio of intensities of a Stokes line over the corresponding
anti-Stokes line.
E45. Use Mathematica to plot the function e−c|x| with interactive manipulation
of the parameter c between 0.7 and 2.0, in steps of 0.1. Denote c by f and
present the function for f = 1.5.
E46. Using the basis set 6-311G*, evaluate the energy difference EMP2 − EHF in
kcal mol−1 for the C2H6 molecule and the chemical reaction C2H4(g) +
H2(g) ! C2H6(g), at 298 K (Suggestion: go to CCCBDB). Comment on the
obtained energy differences.
References 205
References
Becke AD (1993) Density-functional thermochemistry. III. The role of exact exchange. J Chem
Phys 98:5648–5652
Coulson CA, Neilson AH (1961) Proc. Phys. Soc. London 78:831–837
Dunning TH Jr (1989) Gaussian basis sets for use in correlated molecular calculations. I. The
atoms boron through neon and hydrogen. J Chem Phys 90:1007–1023
Hohenberg P, Kohn W (1964) Inhomogeneous electron gas. Phys Rev 136:B864–B871
Kohn W, Sham LJ (1965) Self-consistent equations including exchange and correlation effects.
Phys Rev 140:A1133–A1138
Lee C, Yang W, Parr RG (1988) Development of the Colle-Salveti correlation-energy formula into
a functional of the electron density. Phys Rev B 37:785–789
Miehlich B, Savin A, Stoll H, Preuss H (1989) Results obtained with the correlation-energy
density functionals of Becke, and Lee, Yang and Parr. Chem Phys Lett 157:200–206
Moller C, Plesset MS (1934) Note on an approximation treatment for many-electron systems. Phys
Rev 46:618–622
Further Reading
Ballentine LE (1998) Quantum mechanics: a modern development. World Scientific Publishing

Co. Pte. Ltd.
Gelfand IM, Fomin SV (2000) Calculus of variations. Dover Publications, Inc.
Hameka HF (2004) Quantum mechanics: a conceptual approach. Wiley
Koch W, Holthausen MC (2001) A chemist’s guide to density functional theory, 2nd edn.
Wiley-VCH
Levine IN (2013) Quantum chemistry, 7th ed. Pearson Education Inc.
Long DA (2002) The Raman effect: a unified treatment of the theory of Raman scattering by
molecules. Wiley
Merzbacher E (1998) Quantum mechanics. 3rd ed. Wiley
Parr RG, Yang W (1989) Density-functional theory of atoms and molecules. Oxford University
Press
Schatz GC, Ratner MA (1993) Quantum mechanics in chemistry. Prentice-Hall International
Molecular Symmetry
4
Abstract
Symmetry manifests itself in the shapes of things, from macroscopic objects to
molecules and atoms. Within the Born–Oppenheimer approximation, the nuclear
framework of a molecule at equilibrium defines its symmetry, which reflects on
molecular orbitals and molecular vibrations. This chapter begins by defining
symmetry operations and presenting the symmetry elements of several
molecules. We show that the set of symmetry operations of a molecule form a
group and present a dichotomy diagram for finding the point group in a
systematic way. We then represent reflection in a plane, rotation about an axis,
and inversion in a point by 3 3 matrices in the (x, y, z)-basis, using these to
illustrate the concepts of conjugate elements and class. The implications of
symmetry on transition dipole moments and transition polarizabilities lead to
selection rules, which are explained and illustrated. The last section of this
chapter consists of the analysis of symmetry properties of molecular vibrations
in order to identify point groups. At the end of this chapter, the student can find
several Mathematica codes (Classes of Symmetry Operations, Multiplication
Table for the C2v Group, Selection Rule for a Particle in a One-Dimensional
Box, Selection Rule for the Two-Particle Rigid Rotor, Selection Rule for the
Harmonic Oscillator) with references to the main text and detailed explanations
of new commands, a glossary of important scientific terms, and a list of
exercises, whose complete answers can be found after the Appendix.
4.1 Symmetry Operations
A symmetry operation converts the initial configuration of an object into an

indistinguishable final configuration. If the symmetry operation is not observed
during its application, it is impossible to decide whether it was applied, because the
DOI 10.1007/978-3-319-41093-7_4
208 4 Molecular Symmetry
initial and final configurations are indistinguishable. In this chapter, the objects
acted on by symmetry operations are molecules with their nuclei at equilibrium
positions. Each symmetry operation is carried out by a symmetry element. For a
finite object such as a molecule, a symmetry element can be a plane, an axis, a
point, or a combination of an axis and a plane. Symmetry operations and the
corresponding symmetry elements are represented by the same letters, with sym-
metry operations shown in italics.
The symmetry of the H2O molecule results from the equivalence of its OH
bonds. There are three operations that express this symmetry (Fig. 4.1). The first is
the anticlockwise rotation by 2p/2 radians (180º) around the C2 axis of rotation
(this axis coincides with the z axis; anticlockwise is the direction of an angle in the
plane). The second is the reflection through the xz plane of reflection, rxz (r =
sigma is the Greek letter for s; Spiegel is the German word for mirror). The third
symmetry operation is the reflection in the yz plane of reflection ryz.
Consider now the benzene molecule at equilibrium, with two regular concentric
and parallel hexagons, one formed by the carbon nuclei, the other formed by the
hydrogen nuclei (Fig. 4.2). The most evident symmetry element of the benzene
molecule is the sixfold rotation axis C6, perpendicular to the reflection plane rh
defined by the molecule’s nuclei. The axis C6 coincides with the rotation axes C3
and C2. As the highest-order axis, C6 is the principal axis of the molecule.
Reflection planes that contain C6 are of two types, namely vertical reflection planes
rv, which contain diametrically opposed CH bonds, and diagonal reflection planes
rd, which bisect the angles formed by rv planes. If a Cartesian coordinate system is
set up with its origin at the center of the regular hexagon formed by the carbon
nuclei, the exchange of Cartesian coordinates (x, y, z) by (−x, −y, −z) for any point
in space leads to an indistinguishable molecular configuration (Fig. 4.2). This
symmetry operation is called inversion and is represented by i, with the corre-
sponding symmetry element being a point at the origin of the Cartesian reference
system, called the center of symmetry i. Note that every inversion symmetry
operation can be seen as resulting from a 2p/2 rotation around one C2 axis, followed
by a reflection in rh. Hence, the center of symmetry i is said to be equivalent to a
twofold rotation–reflection axis S2. A symmetry operation around C6 followed by
C2
σxz
z
O O O
σyz
z y
z x
Fig. 4.1 Nontrivial symmetry elements of the H2O molecule
4.1 Symmetry Operations 209
Fig. 4.2 The center of

symmetry of a benzene
molecule is at the center of the
regular hexagon formed by
the nuclei of the carbon atoms
1 3 1 1
C3 C2
B B B B
2 3 1 2 2 3 3 2
Fig. 4.3 C3 and C2 rotations in BF3. Fluorine atoms are numbered in order to show the effect of
the symmetry operation
a reflection in rh gives a sixfold rotation–reflection around the axis of symmetry S6.

In turn, combining C3 with rh yields a rotation–reflection axis of symmetry S3.
Let us now consider the BF3 molecule. The structural equivalence of the B–F
bonds, all in the same horizontal plane of reflection rh, points to the existence of a
principal threefold rotation axis C3 perpendicular to the molecular plane. In addi-
tion, each B–F bond axis is a twofold symmetry axis C2 (Fig. 4.3). In turn, each
perpendicular plane that contains one B–F bond axis is a vertical plane of sym-
metry, rv. A 2p/3 rotation around the C3 axis followed by a reflection in rh is a
threefold rotation–reflection symmetry operation around S3.
One of the symmetry elements of the CH4 molecule is a fourfold rotation–
reflection symmetry axis S4 (Fig. 4.4). However, in this molecule, both the fourfold
rotation axis C4 and the perpendicular plane of reflection r are not individual
symmetry elements of CH4.
A symmetry operation around a rotation–reflection axis S1 involves a 2p rotation
followed by a reflection in the perpendicular plane. These operations make S1
equivalent to a reflection in a horizontal plane, S1 = rh. In turn, an S2 rotation–
reflection axis involves a p rotation around the axis followed by a reflection in the
perpendicular plane, the whole operation being equivalent to the inversion, S2 = i,
as we have mentioned before.
Fig. 4.4 The fourfold rotation–reflection symmetry operation on the CH4 molecule is a 2p/4
rotation (90º) followed by reflection in the perpendicular plane that contains the carbon nucleus
F F F F
H H Br Br
C C C C
H H H H
Cl Cl Cl Cl
Fig. 4.5 Object-to-mirror image relationship for CH2ClF and CHBrClF
Consider now the chlorofluoromethane (CH2ClF) and bromochlorofluoromethane

(CHBrClF) molecules. Figure 4.5 shows pairs of these molecules in the object–image
relationship given by a plane mirror. The distinction between them is clear: while the
two CH2ClF molecules correspond to the same molecule, since a p rotation of one of
them followed by a translation makes the object and image superimposable, the
CHBrClF molecules are not superimposable. In fact, in order to convert one CHBrClF
molecule into the other molecule of the pair, two bonds need to be broken and
exchanged. The CHBrClF compound is said to be chiral (the word chiral has its origin
in the Greek word for hand), since the molecules of the object–image pair are not
superimposable, like the left and right hands of a person. From the symmetry point of
view, chiral molecules do not have any symmetry rotation–reflection axis Sn.
4.2 Point Groups
If A and B represent two symmetry operations for the same molecule, the sequential
application of these symmetry operations, A followed by B, defines the product of
symmetry operations, denoted by BA, with A being applied first. Considering the
definition of symmetry operation, we conclude that the product of two symmetry
4.2 Point Groups 211
operations is also a symmetry operation for the same molecule (the product is said
to be an internal mathematical operation). In addition, the action corresponding to
a particular symmetry operation can always be reversed. This means that every
symmetry operation of a particular molecule has an inverse. Note that the product
of a symmetry operation by its inverse is equivalent to “doing nothing,” that is, to
the identity E (from the German word Einheit for unity).
The set of symmetry operations for a particular molecule obeys the following
requirements:
(1) the product of any two elements, including the product of an element and itself,
is a member of the same set of symmetry operations (this is called the closure
requirement);
(2) each symmetry operation has an inverse;
(3) there is an identity operation, which we represent by E;
(4) the product is associative, that is, (AB)C = A(BC) (this property results from
the concept of symmetry operation).
Having these properties, one says that the set of symmetry operations of a
particular molecule belongs to a point group (all symmetry elements have a point
in common which is kept unaltered by any symmetry operation of the group). The
number of elements of a group is called the order of the group. When all products
of symmetry operations are commutative, BA = AB, the group is said to be com-
mutative or abelian. Group theory is a branch of mathematics.
We now consider a systematic way of finding the point group of a molecule. We
begin by identifying groups of molecules with high symmetry, including linear
molecules. These are generally called special groups (Fig. 4.6). The internuclear
axis of linear molecules is a principal axis of infinite order C∞, since no matter
how small the rotation around the principal axis is, a new indistinguishable
molecular configuration is always obtained. There are two groups for linear
molecules, depending on whether they have a center of symmetry (Fig. 4.7). Linear
molecules with no center of symmetry like HCl, HCN, and HCCCl (Fig. 4.7),
belong to the C∞v group, whereas linear molecules with a center of symmetry, like
CO2, HCCH (Fig. 4.7), and H2, belong to the D∞h group. Molecules with the
shape of regular polyhedrons exhibit high symmetry and belong to the groups we
classify as special groups. Regular tetrahedral molecules like CH4, CCl4 (Fig. 4.7),
and P4 belong to the Td group (tetrahedral group). A regular octahedral molecule
like SF6 (Fig. 4.7) belongs to the Oh group (octahedral group).
If a molecule whose group of symmetry is to be identified does not belong to any
of these groups of high symmetry, one should look for the presence of a Cn
symmetry axis. A molecule that does not have a Cn axis of symmetry may belong to
one of the three following groups: Cs, if the molecule, like HOCl (Fig. 4.8), has
only one plane of symmetry; Ci, if the molecule, like CHClF-CHClF in the trans
conformation (Fig. 4.8), has only one center of symmetry; C1, if the molecule, like
CHBrClF (Fig. 4.8), does not have any element of symmetry.
Fig. 4.6 Dichotomy for finding the point group of a molecule
Fig. 4.7 Molecules with high symmetry
If the molecule does not belong to any special group or to any of the three last
groups, it should have at least one Cn symmetry axis (possibly n = 2). Then, one
should ask whether an S2n symmetry element (n 2) collinear with the Cn axis of
symmetry is present. If the answer to this question is no, then one should ask
whether there are nC2 axes of symmetry perpendicular to the Cn axis of symmetry.
Answering no to this question leads us to three groups of symmetry, namely, Cn
(HOOH as an example of the C2 group; Fig. 4.9), Cnh (trans-FHC = CHF as an
4.2 Point Groups 213
Fig. 4.8 Molecules that do not have a Cn axis of symmetry
Fig. 4.9 Molecules with a principal axis of symmetry Cn but no C2 perpendicular axes
example of the C2h group; Fig. 4.9); and Cnv (NH3 as an example of the C3v group;
Fig. 4.9).
Molecules with nC2 axes of symmetry perpendicular to the Cn axis of symmetry
belong to groups Dn [tris(ethylenediamine)cobalt (III) cation, [Co(en)3]3+ with
en = H2NCH2CH2NH2, as an example of the D3 group; Fig. 4.10], Dnh (ethylene as
an example of the D2h group; Fig. 4.10) and Dnd (allene as an example of the D2d
group; Fig. 4.10).
Fig. 4.10 Molecules with nC2 axes of symmetry perpendicular to the principal axis of symmetry
Fig. 4.11 The

1,3,5,7-tetrafluorooctatetraene
molecule
If the answer to the question in the presence of an S2n symmetry element (n 2)

is yes, then the molecule belongs to an S2n group (the 1,3,5,7-tetrafluorooctatetraene
molecule is shown in Fig. 4.11 as an example of the S4 group) (E1).
4.3 Matrix Representations
A transformation that converts the vector (x, y, z)T to (x′, y′, z′)T (the superscript
T stands for “transpose,” thus giving the corresponding column vector) is mathe-
matically represented by a matrix whose elements are the coefficients of the linear
combinations of the original coordinates. Such a transformation is linear and applies
within the same vector space. Let us find the representation matrices for symmetry
operations, namely the identity, reflection, inversion, rotation, and rotation–reflec-
tion. Starting with the identity E and the column vector (x, y, z)T, we can write
0 1 0 1 0 10 1
x0 x 1 0 0 x
@ y0 A ¼ E @ y A ¼ @ 0 1 0 A@ y A ð4:1Þ
z0 z 0 0 1 z
since this operation “does nothing,” that is, x′ = x, y′ = y, and z′ = z. One says that
the 3 3 identity matrix represents the identity operation and the column vector
(x, y, z)T is the basis for this representation.
For the reflection through the xz plane, rxz, we write
4.3 Matrix Representations 215
0 1 0 1 0 10 1
x0 x 1 0 0 x
@ y0 A ¼ rxz @ y A ¼ @ 0 1 0 A@ y A ð4:2Þ
z0 z 0 0 1 z
because x′ = x, y′ = −y, and z′ = z. Inversion is given by the equality

0 1 0 1 0 10 1
x0 x 1 0 0 x
@ y0 A ¼ i@ y A ¼ @ 0 1 0 A@ y A ð4:3Þ
z0 z 0 0 1 z
since x′ = −x, y′ = −y, and z′ = −z.

Figure 4.12 represents the counterclockwise rotation of a vector in the xy plane
by an angle h around the z-axis starting from an angle a with the positive x-axis.
Simple trigonometry enables us to write
x ¼ r cos a y ¼ r sin a ð4:4Þ
and
x0 ¼ r cosða þ hÞ ¼ r ðcos a cos h sin a sin hÞ ¼ x cos h y sin h

ð4:5Þ
y0 ¼ r sinða þ hÞ ¼ r ðcos a sin h þ sin a cos hÞ ¼ x sin h þ y cos h
Hence, in matrix notation, we can write

x0 cos h sin h x
¼ ð4:6Þ
y0 sin h cos h y
Considering now the basis formed by the column vector (x, y, z)T, we write
0 1 0 1 0 10 1
x0 x cos h sin h 0 x
@ y0 A ¼ Cn @ y A ¼ @ sin h cos h 0 A@ y A ð4:7Þ
z0 z 0 0 1 z
Fig. 4.12 Counterclockwise

rotation around the z axis of a
vector by an angle h in the xy
plane
θ
α
where Cn represents the counterclockwise rotation of a vector in the xy plane by an

angle h around the z-axis that coincides with the Cn axis (E2).
The rotation–reflection operation Sn is given by the product rxyCn (the plane of
reflection is perpendicular to the Cn axis, which is coincident with the z-axis), and
so we can write
0 1 0 1 0 10 10 1
x0 x 1 0 0 cos h sin h 0 x
@ y0 A ¼ Sn @ y A ¼ @ 0 1 0 A@ sin h cos h 0 A@ y A
z0 0 z 0 0 1 10 10 0 1 z
cos h sin h 0 x
¼ @ sin h cos h 0 A@ y A ð4:8Þ
0 0 1 z
(E3, E4, E5).

We now consider the C2v point group. The matrix that represents C2 can be
easily obtained from (4.7) using h = 2p/2 radians, and the matrices that represent
the reflections are diagonal, because the planes of reflection coincide with xz and yz
planes. Hence,
0 E 1 0 C2 1 0 rxz 1 0 ryz 1
1 0 0 1 0 0 1 0 0 1 0 0
@0 ð4:9Þ
1 0A @ 0 1 0A @0 1 0A @ 0 1 0A
0 0 1 0 0 1 0 0 1 0 0 1
We now pass to the C3v point group and consider the group of nontrivial symmetry
operations C3, C23 (= C3C3), rv1, rv2, and rv3. Assuming that C3 coincides with the
z-axis and rv1 uses yz as the plane of reflection (Fig. 4.13), then C3 and C23
are represented by the following 3 3 matrices [h(C3) = 2p/3 radians and
h (C23) = 4p/3 radians, see (4.7)],
0 C3 1 0 C32 1
pffiffi pffiffi
1 23 0 p12ffiffi 23 0
B pffiffi32 C B C ð4:10Þ
@ 12 0A @ 3 1 0A
2 2 2
0 0 1 0 0 1
and rv1 is represented by

0 1
1 0 0
@ 0 1 0A ð4:11Þ
0 0 1
We now consider a new basis (x′ y′ z′)T obtained from (x y z)T by a counter-
clockwise rotation of 2p/3 radians around the z-axis that is coincident with both z′
and C3 (Fig. 4.13):
4.3 Matrix Representations 217
Fig. 4.13 Horizontal

projections of vertical planes
of the C3v point group
0 1 0 1
x0 x
@ y0 A ¼ C3 @ y A ð4:12Þ
z0 z
Inspection of Fig. 4.13 shows that the change of basis (4.12) converts rv1 into rv2.
Therefore, rv2 is obtained from rv1 by a similarity transformation B = X−1A X,
where X is C3 [rv1 and rv2 are represented by similar matrices; see (2.89):
rv2 ¼ C31 rv1 C3 ð4:13Þ
Following the same reasoning, rv3 is obtained from rv1 by another similarity
transformation in which X is C23 (the change of basis is due to C23):
1
rv3 ¼ C32 rv1 C32 ð4:14Þ
1 T
Note that C31 ¼ C3T and C32 ¼ C32 , since both (x y z)T and (x′ y′ z′)T are
orthogonal systems of axes [see (4.7)]. Substitution of (4.10) and (4.11) in (4.13)
and (4.14) leads to
0 rv2 pffiffi 1 0 rpv3ffiffi 1

1
23 0 1 3
0
B 2pffiffi3 C B p2ffiffi3 2 C ð4:15Þ
@ 12 0 A @ 12 0A
2 2
0 0 1 0 0 1
Hence, the matrices that represent the symmetry operations of the C3v group in the
basis (x y z)T are [see (4.10), (4.11), and (4.15)]
E 0 C3 1 0 C32 1 rv1 0 rv2 1 0 rv3 1

0 1 pffiffi pffiffi 0 1 pffiffi pffiffi
1 0 0 1 23 0 1 3
0 1 0 0 1
23 0 1 3
0
@0 1 0A B pffiffi32 C B p2ffiffi3 2 C @ 0 1 B 2pffiffi3 C B p2ffiffi3 2 C
@ 12 0A @ 12 0A 0A @ 12 0A @ 12 0A
2 2 2 2
0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1
ð4:16Þ
In addition, C3 and C23 are also related by similarity transformations B = X−1A X,

where X can be rv1, rv2, or rv3. A similarity transformation preserves the sum of the
diagonal matrix elements, or trace; in group theory, character (E6). The character
of the identity symmetry operation E is always equal to the dimension of the basis.
The characters of the symmetry operations of the C3v point group in the basis
(x y z)T are given by
E C3 C32 rv1 rv2 rv3

ð4:17Þ
3 0 0 1 1 1
The symmetry operations of the C3v point group (E, C3, C23, rv1, rv2, rv3) can be
grouped in three sets, {E}, {C3, C23}, and {rv1, rv2, rv3}, formed by symmetry
operations having the same characters and related by similarity transformations.
Each of these sets of symmetry operations is called a class, and the symmetry
operations within one specified class are said to be mutually conjugate (E7).
Character tables group symmetry operations in classes. Using matrix representa-
tions of the symmetry operations for the C3v point group in the basis (x y z)T [see
(4.10), (4.11), (4.15)], the first Mathematica code of M1 confirms that {C3, C23},
and {rv1, rv2, rv3} are two classes of the C3v point group, and the second code of
M1 enables one to observe rTv1.C3(h).rv1 “behind the scenes.”
4.4 Character Tables
The product of two symmetry operations of a point group is a member of the same
point group. The multiplication table of a point group presents the results of the
products of symmetry operations BA, where A and B are symmetry operations of the
group and A is applied first (A and B are presented as columns and rows of the
multiplication table). Using matrix representations of the symmetry operations of
the C2v point group in the basis (x y z)T [see (4.9)], the Mathematica code M2 builds
the multiplication table of the C2v point group. Inspection of this multiplication
table shows that each symmetry operation of the C2v group coincides with its own
inverse, since the diagonal is occupied by the identity. In addition, the upper and
lower matrix triangles of the multiplication table are equal (the product matrix is
symmetric), meaning that the order of multiplication in a product is irrelevant; that
is, symmetry operations commute, and C2v is an abelian point group. Among the
nontrivial symmetry operations C2, rxz, and ryz, the product of two of them is equal
4.4 Character Tables 219
to the one left out. A set of matrices or numbers (a number is a 1 1 matrix)

consistent with the multiplication table is called a group representation.
Multiplication table of the C2v group
E C2 rxz ryz
E E C2 rxz ryz
C2 C2 E ryz rxz
rxz rxz ryz E C2
ryz ryz rxz C2 E
Consider now a molecular orbital for H2O. The multiplication table for the C2v
symmetry group can be used to find how a particular molecular orbital transforms
when a symmetry operation is applied to it. Applying the identity operator to a
function does not modify it in any conceivable way, the result being the same as if
the function had been multiplied by 1, E /i = /i. Consider now the diagonal of the
C2v multiplication table: the identity operation can be replaced by the square of any
other symmetry operation, that is,
E ¼ C2 C2 ¼ rxz rxz ¼ ryz ryz ð4:18Þ
Applying each member of (4.18) to a function, one obtains
E/i ¼ C2 C2 /i ¼ rxz rxz /i ¼ ryz ryz /i ¼ /i ð4:19Þ
Therefore, the eigenvalues for the nontrivial symmetry operations of H2O (C2, rxz,
ryz) should be either +1 or −1. Since the product of two of the nontrivial symmetry
operations (C2, rxz, ryz) is equal to the one left out [this can be confirmed by
inspection of the elements (2,3) (second row, third column), (2,4), (3,2), (3,4), (4,2),
and (4,3) of the multiplication table], either all of the corresponding eigenvalues of
the C2v point group have to be equal to +1, or two of them are equal to –1. Each set
of these eigenvalues consistent with the multiplication table of the group is called
an irreducible representation of the C2v group (the word “irreducible” means that
these representations cannot be further decomposed into simpler ones). The table
that contains the characters of the irreducible representations of the group is called
the character table.
The irreducible representations with 1 in the axis of highest order (the so-called
principal axis; for the C2v group, the C2 axis) are identified by A. In turn, those
representations with −1 in the principal axis are identified by B. The subscript 1 is
used when the eigenvalue of the symmetry operation rxz is equal to +1. The
representation with 1 as eigenvalues in all symmetry operations of the group reflects
the maximum symmetry within the group and is called a totally symmetric species.
Character table of the C2v point group
E C2 rxz ryz
A1 1 1 1 1 z x2, y2, z2
A2 1 1 −1 −1 Rz xy
B1 1 −1 1 −1 x, Ry xz
B2 1 −1 −1 1 y, Rx yz
(http://symmetry.jacobs-university.de)
The character table also presents the functions x, y, z, x2, y2, z2, xy, xz, yz, and the
rotations Rx, Ry, and Rz around the x-, y-, and z-axes. The z-coordinate transforms
like the totally symmetric species (A1), since the z-axis coincides with C2
(Fig. 4.14). Inspection of Fig. 4.14 for the orientation of the x- and y-axes shows
that x transforms like B1 and y transforms like B2. The x, y and z functions represent
translations along these axes and so transform like the dipole moment components.
The function xy transforms like A2 (= B1 B2), xz and yz transform like B1
(= B1 A1) and B2 (= B2 A1), respectively, and x2 (B1 B1 = A1),
y2 (B2 B2 = A1), and z2 (A1 A1 = A1) are totally symmetric. Under symmetry
operations, the polarizability matrix elements transform according to the corre-
sponding subscripts (axx transforms as x2, axy transforms as xy, …).
The rotations around the x-, y-, and z-axes are carried out by the corresponding
torques (Fig. 4.15). In particular, by applying the symmetry operations of C2v to the
torque for Rx, we conclude that Rx transforms like the symmetry species B2 (see
Fig. 4.15 and the character table). Similar reasoning applied to the torques for Ry
and Rz leads to their assignments to B1 and A2, respectively.
Inspection of the C2v character table reveals that the irreducible representations
behave like “orthogonal vectors.” In fact, the dot product of two irreducible rep-
resentations is given by
CTi Cj ¼ h dij ð4:20Þ
C2
σxz
z
O O O
σ yz
z
z x y
Fig. 4.14 Nontrivial symmetry elements of the H2O molecule

4.4 Character Tables 221
Fig. 4.15 Rotations Rx, Ry,

and Rz for the H2O molecule
z
(C2v group) and the
corresponding torques
Rz
Ry
y
Rx
where the character of each irreducible representation is represented by a column

vector, h is the order of the group (= number of symmetry elements; for C2v, h = 4),
and dij is the Kronecker delta (dij = 0 for i 6¼ j, dij = 1 for i = j).
The character of matrix representation (4.9) (= set of characters of the symmetry
operations) is given by
C2v E C2 rxz ryz

ð4:21Þ
C 3 1 1 1
A reducible representation C can be decomposed as a linear combination of irre-

ducible representations of the group
X
C¼ ai Ci ð4:22Þ
where
CT Ci
ai ¼ ð4:23Þ
h
since substitution of (4.22) into (4.23) using (4.20) gives an identity. Applying
(4.23) to C in (4.21) leads to
C ¼ A1 þ B1 þ B2 ð4:24Þ
a result that is not at all surprising, since A1, B1, and B2 are the irreducible
representations for the functions x, y, and z that form the basis for C (see the
character table of the C2v point group).
We now consider the character table for the C3v group (http://symmetry.jacobs-
university.de) and note that the irreducible representation of the point group labeled
E is two-dimensional, since it is shown by the character of the identity operation
E and the basis functions grouped in pairs, on the right-hand side of the irreducible
representation E (E8).
The characters of the symmetry operations of the C3v point group in the (x y z)T
basis [see (4.16)] are given by
C3v E 2C3 3rv

ð4:25Þ
C 3 0 1
Use of (4.20), (4.22), and (4.23) leads to
C ¼ A1 þ E ð4:26Þ
a result that can be easily confirmed by inspection of the character table for the C3v
point group.
E 2C3 3rv
A1 1 1 1 z x2 + y2, z2
A2 1 1 −1 Rz
E 2 −1 0 (x, y) (Rx, Ry) (x2–y2, xy) (xz, yz)
4.5 Selection Rules
A function f(x) is said to be even if f(x) = f(−x). Functions like x2, cos x, exp(−x2),
|x| are even functions. In turn, g(x) is said to be an odd function if g(x) = −g(−x).
Functions like x, x3, sin x are odd functions. Let us now consider the integral of a
function h(x) defined for −a x a. It is easy to show that h(x) can be
decomposed into the sum of the even function f(x) = [ h(x) + h(−x)]/2 and the odd
function g(x) = [ h(x) − h(−x)]/2. If the odd function is integrable, then its integral
over the symmetric interval [−a, a] is identically zero. Thus, we can conclude that
only the even function contributes to the integral of f(x) over the symmetric interval
[−a, a], with this integral being precisely the same as twice the integral over the
interval [0, a]. Generalization of this result enables one to conclude that the integral
of f(s) over the entire configuration space is significant when f(s) is a basis for a
4.5 Selection Rules 223
representation that coincides with or includes the totally symmetric irreducible

representation of the group.
We now recall a result obtained in the previous chapter [see (3.283)] showing
that the probability of a spectroscopic transition in absorption from stationary state
i to stationary state f is proportional to
D E2 D E2 D E2 D E2
0
^x jw0i þ w0f l
w f jl ^y w0i þ w0f l
^z w0i ¼ w0f jl
^jw0i ð4:27Þ
The spectroscopic transition from i to f is active if at least one of the transition

dipole moment integrals is different from zero. Assuming that this applies to the
first term of the sum (4.27), then the product of functions

w0f x w0i ð4:28Þ
belongs to a representation that coincides with or contains the totally symmetric

representation of the group. In other words, the product of functions

w0f w0i ð4:29Þ
is the basis for a representation that coincides with or contains the irreducible
representation for the function x.
When
D E
^jw0i ¼ 0
w0f jl ð4:30Þ
the transition from i to f is forbidden. Allowed transitions have

D E
^jw0i 6¼ 0
w0f jl ð4:31Þ
Consider a charged particle in a one-dimensional box (see 3.§1). When electro-

magnetic radiation is absorbed or emitted, the corresponding selection rule indi-
cates that the quantum number must change by an odd integer, as illustrated by the
Mathematica code M3. In turn, the Mathematica code M4 verifies that the selection
rule for a two-particle rigid rotor (see 3.§2), with an associated dipole moment for
the electromagnetic radiation to interact with, is DJ = ±1, where J is the rotational
quantum number. In addition, the selection rule for a harmonic oscillator (see
Sect. 3.2) indicates that for absorption and stimulated emission of radiation, one has
Dn = ±1, where n is the vibrational quantum number (see Mathematica code M5).
Considering now a Raman scattering transition from i to f, we recall that the xy
polarizability component is given by
0D ED E D E
1
0 0
1 X @ wf jl
0
^x jw0r w0r l
^y wi ^y wr w0r jl
w0f l ^x jw0i
axy ¼ þ A ð4:32Þ
fi h r6¼i;f xri x0 iCr xrf þ x0 þ iCr
This expression suggests that the Raman selection rule depends on the product of
transition dipole moments in the numerators of (4.32),
D E

w0f j xjw0r w0r j yjw0i ð4:33Þ
and involve a summation over all virtual states r. However, since no direct infor-
mation can be obtained about the virtual states r, except perhaps in discrete reso-
nance Raman, where a limited number of them dominate the sum over r, the general
condition for
D 0 0E
axy fi ¼ wf axy wi ð4:34Þ
to be nonzero is that
D E
w0f jxyjw0i ð4:35Þ
belongs to a representation that contains the totally symmetric species, a condition

that can be easily extended to other components of the polarizability tensor.
4.6 Molecular Vibrations
We now determine the symmetry species of the vibrations of a molecule and find
the numbers of distinct fundamental frequencies in infrared and Raman spectra. In
small molecules, these numbers may eventually be used to distinguish isomers or
identify geometries corresponding to a known molecular stoichiometry.
We begin by considering the cis-1,2-dichloroethene molecule (Fig. 4.16). This
molecule belongs to the C2v point group, whose character table is
E C2 rxz ryz
A1 1 1 1 1 z x2, y2, z2
A2 1 1 −1 −1 Rz xy
B1 1 −1 1 −1 x, Ry xz
B2 1 −1 −1 1 y, Rx yz
4.6 Molecular Vibrations 225
y
Cl x Cl
C C
H H
Fig. 4.16 cis-1,2-Dichloroethene (C2v group) with system of Cartesian axes. The latter should be
centered on each nucleus of the molecule, thus yielding a basis of 18 coordinates
The set of unit vectors directed along the Cartesian axes and centered on each
nucleus of the molecule forms a basis of 18 vectors (= 3 N, where N = 6 is the
number of atoms in the molecule) that can describe all molecular motions, namely
translations, rotations, and vibrations. Therefore, these vectors form the basis for a
reducible representation of the C2v group. The diagonal elements of the matrices
that represent the symmetry operations can be easily obtained without the need for
determining the whole representation matrices. Simple and intuitive rules can be
applied. In particular, when a vector moves its origin under the action of a sym-
metry operation, one should count zero for the diagonal element of the corre-
sponding transformation matrix. In turn, if a vector remains unaltered under the
action of a symmetry operation, one should count 1. When a vector rotates by h
radians, the trace of the corresponding matrix is given by 2cosh + 1 [see (4.7)].
Thus, applying the symmetry operations to the set of the basis vectors (see
Fig. 4.16) gives the following reducible representation:
C2v E C2 rxz ryz

ð4:36Þ
C 18 0 0 6
Using (4.22) and (4.23), we decompose C into the irreducible representations of

the group, obtaining
C ¼ 6A1 þ 3A2 þ 3B1 þ 6B2 ð4:37Þ
Inspection of the character table of the C2v group shows that the irreducible rep-
resentations of the group that correspond to rotations Rx, Ry, and Rz are B2, B1, and
A2, and those that correspond to translations x, y, and z are B1, B2, and A1. Thus, we
can write
Crot ¼ A2 þ B1 þ B2
ð4:38Þ
Ctrans ¼ A1 þ B1 þ B2
Subtracting (4.38) from (4.37) leads to the reducible representation for the
molecular vibrations
Cvib ¼ 5A1 þ 2A2 þ B1 þ 4B2 ð4:39Þ
Irreducible representations that transform like x, y, and z (A1, B1, B2) correspond to
infrared active vibrations, and those that transform like x2, y2, z2, xy, xz, and yz (A1,
A2, B1, B2) correspond to Raman active vibrations. Hence,
CIR ¼ 5A1 þ B1 þ 4B2

ð4:40Þ
CR ¼ 5A1 þ 2A2 þ B1 þ 4B2
Therefore, the cis-1,2-dichloroethene molecule has 10 infrared active fundamental

vibrations and 12 Raman active fundamental vibrations.
We now consider the trans-1,2-dichloroethene molecule (Fig. 4.17) that belongs
to the C2h group, whose character table is
Character table of the C2h group
E C2 i rh
Ag 1 1 1 1 Rz x2, y2, z2, xy
Bg 1 −1 1 −1 Rx, Ry xz, yz
Au 1 1 −1 −1 z
Bu 1 −1 −1 1 x, y
Cl y
x H
C C
H
Cl
Fig. 4.17 trans-1,2-Dichloroethene (C2h group) with system of Cartesian axes. The latter should
be centered on each nucleus of the molecule thus yielding a basis of 18 coordinates
In the labeling of the irreducible representations, the subscripts g (from German

gerade, even) and u (from German ungerade, odd) are used when the character for
the inversion is positive and negative, respectively.
nucleus of the molecule forms a basis of 18 vectors and gives the following
reducible representation:
C2h E C2 i rh
ð4:41Þ
C 18 0 0 6
Using (4.22) and (4.23), we decompose C into the irreducible representations of the
group, obtaining
C ¼ 6Ag þ 3Bg þ 3Au þ 6Bu ð4:42Þ
Inspection of the character table reveals that
Crot ¼ Ag þ 2Bg
ð4:43Þ
Ctrans ¼ Au þ 2Bu
We now subtract (4.43) from (4.42) to obtain the reducible representation for the
molecular vibrations
Cvib ¼ 5Ag þ Bg þ 2Au þ 4Bu ð4:44Þ
From the character table, we can conclude that the symmetry species of the infrared
active and Raman active vibrations are
CIR ¼ 2Au þ 4Bu

ð4:45Þ
CR ¼ 5Ag þ Bg
The symmetry species for the infrared active vibrations are of the u-type, whereas
the symmetry species for the Raman active vibrations are of the g-type (E9).
Inspection of the character tables leads one to conclude that those groups that
include inversion as a symmetry operation have mutual exclusion of the infrared
and Raman active vibrations, i.e., the infrared active vibrations are Raman inactive;
and conversely, the Raman active vibrations are infrared inactive. Comparison of
(4.40) with (4.45) leads one to conclude that the cis and trans isomers can be
distinguished by the number of infrared active and Raman active fundamental
vibrations.
Consider now a molecule of general formula XY3, with the X atom bonded to
each of the Y atoms. For this molecular connectivity, the molecule can be pyra-
midal (C3v group) or planar (D3h group). As we will find, this question can be
answered by symmetry analysis of vibrations, in particular, by the number of
distinct fundamental frequencies.
Fig. 4.18 A pyramidal XY3

molecule includes, as
C3
symmetry elements, a
threefold rotation axis C3 and
three vertical planes of
reflection
N
H
H H
A pyramidal XY3 molecule (Fig. 4.18) belongs to the C3v point group, whose
character table is given by
Character table of the C3v group
C3v E 2C3 3rv

A1 1 1 1 z x2 + y2, z2
A2 1 1 −1 Rz
E 2 −1 0 (x, y) (Rx, Ry) (x2–y2, xy) (xz, yz)
We now consider the set of unit vectors directed along the Cartesian axes and
centered on each nucleus of the XY3 molecule. These form a basis of 12 vectors
that correspond to the following reducible representation:
C3v E 2C3 3rv

ð4:46Þ
C 12 0 2
Decomposing C into the irreducible representations of the group leads to
C ¼ 3A1 þ A2 þ 4E ð4:47Þ
and inspection of the character table reveals that
Crot ¼ A2 þ E
ð4:48Þ
Ctrans ¼ A1 þ E
Hence we have
Cvib ¼ 2A1 þ 2E ð4:49Þ
and the following infrared active and Raman active vibrations:
CIR ¼ 2A1 þ 2E
ð4:50Þ
CR ¼ 2A1 þ 2E
Thus, a pyramidal XY3 molecule (C3v group) has four distinct fundamental fre-
quencies of vibration, all infrared and Raman active. Note that each E species
corresponds to two degenerate vibrational modes and a single distinct frequency.
The planar XY3 molecule (Fig. 4.19) belongs to the D3h group, whose character
table is
D3h E 2C3 3C2 rh 2S3 3rv

A′1 1 1 1 1 1 1 x2 + y2, z2
A′2 1 1 −1 1 1 −1 Rz
E′ 2 −1 0 2 −1 0 (x, y) (x2 − y2, xy)
A″1 1 1 1 −1 −1 −1
A″2 1 1 −1 −1 −1 1 z
E″ 2 −1 0 −2 1 0 (Rx, Ry) (xz, yz)
nucleus of the XY3 molecule form a basis of 12 vectors that correspond to the
following reducible representation:
C 3, S 3
Y
X Y
Y
Fig. 4.19 A planar XY3 molecule includes, as symmetry elements, a threefold rotation axis C3
that coincides with a threefold rotation-reflection axis S3, three twofold rotation axes C2 that are
aligned with each of the X–Y bonds, and four planes of symmetry: one horizontal plane and three
vertical planes
D3h E 2C3 3C2 rh 2S3 3rv

ð4:51Þ
C 12 0 2 4 2 2
Expressing this reducible representation as a linear combination of irreducible

representations of the group leads to
C ¼ A01 þ A02 þ 3E0 þ 2A002 þ E00 ð4:52Þ
The symmetry species for rotations and translations are
Crot ¼ A02 þ E00

ð4:53Þ
Ctrans ¼ E0 þ A002
Hence,
Cvib ¼ A01 þ 2E0 þ A002 ð4:54Þ
and
CIR ¼ 2E0 þ A002

ð4:55Þ
CR ¼ A01 þ 2E0
In conclusion, a planar XY3 molecule (D3h group) has three distinct fundamental
frequencies of vibration, both in the infrared and Raman spectra. Thus, it is possible
to distinguish the geometries of an XY3 molecule by the number of distinct fun-
damental frequencies of vibration, since this result differs from that obtained for a
pyramidal XY3 molecule [see (4.50)].
In Raman spectroscopy, the incident radiation is a plane-polarized electromag-
netic wave. Analyzing the polarization of the scattered light may provide useful
information on the symmetry of vibrations. In order to carry on this polarization
analysis, we need to interpose a polarization analyzer in the way of the scattered
light, just before this enters the spectrometer. This polarization analyzer acts like a
filter with two perpendicular orientations that allow parallel and perpendicular
polarized scattered light go through. Then, two Raman spectra are recorded: one
records the intensity IVV (the subscripts stand for Vertical, Vertical) the other
records IVH (the subscripts stand for Vertical, Horizontal). The depolarization
ratio of a Raman band is given by
R
IVH ð~mÞd~v
qp ¼ R ð4:56Þ
IVV ð~mÞd~v
I0
molecule
IVV
IVH
Fig. 4.20 Geometry for observing the depolarization ratio in a Raman spectrum
where the integrals extend over each band of the Raman spectrum. (Fig. 4.20).
Theoretical considerations enable one to conclude that the depolarization ratio of
a Raman band lies in the range
0 qp 3=4 ð4:57Þ
A Raman band is said to be depolarized when qp = 3/4, polarized when

0 < qp < ¾, and completely polarized when qp = 0. When the symmetry of a
molecule is not preserved during a particular vibration (non-totally symmetric
vibration), the depolarization ratio takes the value 3/4. For totally symmetric
vibrations, qp is smaller than 3/4.
Consider now the tetrahedral molecules shown in Fig. 4.21 belonging to the
point groups C3v, C2v, Cs, and C1. The following table shows the symmetry species
for the vibrations of these molecules. Note that the numbers of polarized bands, that
is, of totally symmetric vibrations, are different for these point groups, namely, 3 for
C3v, 4 for C2v, 6 for Cs, and 9 for C1. Thus, recording of the Raman spectra and
measuring depolarization ratios enables one to assign each of these molecules to its
point group (E10).
X
X X Y Y
H H
C H C C C
H
H H
H X X
X Z
C3v C2v Cs C1
Fig. 4.21 Tetrahedral molecules belonging to different point groups; X, Y, and Z represent
different halogen atoms
Symmetry species for vibrations of different point groups
Point group Symmetry species

C3v 3A1 + 3E
C2v 4A1 + A2 +2B1 + 2B2
Cs 6A′ + 3A″
C1 9A
Mathematica Codes
M1. Classes of Symmetry Operations
In the following code, the symmetry operations of the C3v group are represented by
3 3 matrices in the orthogonal basis (x y z)T, with the matrices for rv2 and rv3
being obtained by similarity transformations on rv1, with X equal to C3 and C23, and
the concept of class is illustrated. Since (x y z)T is an orthogonal system of axes, the
inverses of matrices are obtained simply by transposing them. The Mathematica
function Transpose transposes a matrix, and If[condition,t,f] gives t if
condition evaluates to True, and f if it evaluates to False.
Consider the similarity transformation r.T

v1.C3.rv1 with C3 being a function of h.
We begin by representing rv1 and C3(h) by the following matrices,

1 0 cos h sin h
rv1 ¼ C 3 ð hÞ ¼ ð4:58Þ
0 1 sin h cos h
calculate C3(h).rv1,

cos h sin h 1 0 cos h sin h
C3 ðhÞ rv1 ¼ ¼ ð4:59Þ
sin h cos h 0 1 sin h cos h
and rTv1.C3(h).rv1

1 0 cos h sin h cos h sin h
ðrv1 ÞT C3 ðhÞ rv1 ¼ ¼
0 1 sin h cos h sin h cos h
ð4:60Þ
We now assume that a = 0 and r = 1 in (4.61) (see Fig. 4.12),

x 1
x ¼ r cos a ¼ 1 y ¼ r sin a ¼ 0 ) ¼ ð4:61Þ
y 0
and obtain

x0 1 cos h sin h 1 cos h
0 ¼ C 3 ð hÞ ¼ ¼
y0 0 sinh cos h 0 sin h
x 1 cos h sin h 1 cos h
0 ¼ C3 ðhÞ rv1 ¼ ¼
y0 0 sinh cos h 0 sin h
x T 1 cos h sin h 1 cos h
¼ ðrv1 Þ C3 ðhÞ rv1 ¼ ¼
y0 0 sin h cos h 0 sin h
ð4:62Þ
The following code plots (x’,y’)T for C3(h), C3(h).rv1 and rTv1.C3(h).rv1 using
ParametricPlot. This Mathematica function plots the (x’, y’)-coordinates as
functions of a parameter (= external variable).
The first plot shows the arc of a circle of radius 1 (r = 1) drawn in the coun-
terclockwise direction for h from 0 to 2p/3. The second plot shows the same
rotation in the same direction (counterclockwise) after the initial point of the
rotation was reflected on rv1 (see Fig. 4.13). In turn, the third plot shows that the
effect of multiplying on the left by rTv1 causes the curve to reflect by rv1 (note that
rTv1 = rv1, since the matrix representing rv1 is diagonal), and this reflection amounts
to a change of direction from counterclockwise to clockwise. On the whole, these
plots enable one to follow the rTv1.C3(h).rv1 orthogonal transformation “behind the
scenes.”
M2. Multiplication Table for the C2v Group
The following Mathematica code builds the multiplication table for the group C2v.
To this end, the symmetry operations are represented by matrices in the basis
(x y z)T [see (4.9)]. Note the way the command If[condition,t,f] is used,
with the field f being used to include another If. For presenting the results, the
function Array builds a 4 4 array called a and uses the command Grid to
present a two-dimensional grid with items of size 2 by 2, where the elements of the
array are included:
M3. Selection Rule for a Particle in a One-Dimensional Box
The following code calculates transition dipole moment integrals for a charged
particle in a one-dimensional box. The first line defines the length of the box as 1,
the second line presents the general expression for the one-dimensional box wave
functions, the third line shows the values of the quantum numbers for the con-
sidered transitions (the list associated with i is outermost, that is, i changes more
slowly than n), and the fourth line of code calculates the transition dipole moments
for these transitions. The results indicate that for nonzero transition moments, the
quantum number must change by an odd integer, ±1, ±3, …. The last line of code
plots the integrand for the 1!2 transition, so that we can understand why the
nonzero calculated transition dipole moments are negative. The option
Filling!Axis specifies that the area under the plotted curve should be filled to
the axis. Note that the factor x of the integrand for the transition dipole moment
weights more high values of x, thus giving rise to a negative area greater than the
positive area and so to a negative integral.
M4. Selection Rule for the Two-Particle Rigid Rotor
The following code calculates transition dipole moment integrals for the
two-particle rigid rotor, whose eigenfunctions are spherical harmonics (see 3.§2).
The first line shows the quantum numbers for the considered transitions, and the
second line calculates the transition dipole moments and shows that allowed
transitions have DJ = ±1 (J is the rotational quantum number). Since the spherical
harmonics are functions of the angular coordinates, x should be substituted by cosh,
since x = r cosh, and r is a constant. In addition, the integrand should include an
additional sinh factor coming from the volume element, which is given by sinh dh
d/. The last line of code plots the integrand for the 0!1 J-transition. Note that the
command for the plot is SphericalPlot3D, since the integrand is a function of
angular coordinates h and / (Plot3D plots f as a function of x and y):
M5. Selection Rule for the Harmonic Oscillator
The following code calculates transition dipole moment integrals for the harmonic
oscillator. The first line of code represents the eigenfunctions as functions of z (= √a
x) [see (3.62) and (3.63)], the second line shows the quantum numbers for the
considered transitions, and the third line calculates the corresponding transition
dipole moments. The results indicate that nonzero transition moments correspond to
Dn = ±1, where n is the vibrational quantum number. The following lines of code
plot the eigenfunctions and the integrands of the transition dipole moment integrals
involved in one allowed transition (the 0!1 transition, where the integrand is an
even function) and one forbidden transition (the 0!2 transition, where the inte-
grand is an odd function):
Glossary
Abelian group A point group in which all products of

symmetry operations are commutative
Allowed transition A nonzero transition dipole moment or nonzero
transition polarizability
Character The sum of the diagonal elements of a matrix
representing a symmetry operation
Character table A two-dimensional table whose rows are
irreducible representations of the point group and
whose columns are classes of symmetry
operations in the group. A character table also
presents the functions x, y, z, x2, y2, z2, xy, xz, yz,
Glossary 239
and the rotations Rx, Ry and Rz around the x-, y-,

and z-axes distributed by the irreducible
representations of the group they belong to
Class A complete set of mutually conjugate
symmetry operations, which are related to each
other by similarity transformations B = X−1 A X
with respect to any other symmetry operation
X of the group. Symmetry operations within the
same class have the same character, since the
similarity transformation preserves the trace
Conjugate elements A and B are related by a similarity
transformation B = X−1 A X with respect to a
third symmetry operation X of the group and
have the following properties: (1) Every group
element is conjugate to itself. (2) If B is
conjugate to A, then A is conjugate to B. (3) If
A is conjugate to B and C, then B and C are
conjugate to each other
Even function A function f(x) such that f(x) = f(−x). The
integral of an even function from −a to a is
different from zero, because the integrals from
−a to 0 and from 0 to a are equal
Forbidden transition A zero transition dipole moment or zero
transition polarizability
Group A set of elements (which may be symmetry
operations of a particular molecule) that obey
the following mathematical requirements:
(1) the product of any two elements, including
the product of an element and itself, is a member
of the group; (2) the group contains an identity
element; (3) each element has an inverse; (4) the
product is associative, i.e., (AB)C = A(BC)
Irreducible representation A representation that cannot be decomposed
into representations of smaller dimension
Odd function A function f(x) such that f(x) = −f(−x). The
integral of an odd function from −a to a is zero,
because the integrals from −a to 0 and from 0
to a differ only by their sign
Point group A group of symmetry operations of a finite object
such as a molecule that has at least one point
unchanged under every symmetry operation of
the group
Selection rule Identifies the quantum states of a molecule that
give rise to allowed transitions
Symmetry operation Converts the initial configuration of an object

into an indistinguishable final configuration. If
the symmetry operation is not observed during its
application, it is impossible to decide whether it
was applied, because the initial and final
configurations are indistinguishable. Each
symmetry operation is carried out by a symmetry
element that can be an axis of rotation, a plane of
reflection, a point of inversion, or an axis of
rotation combined with a perpendicular plane of
reflection
Exercises
E1. What is the principal axis of symmetry in each of the following molecules:
(a) CHCl3; (b) PF5; (c) SF6; (d) C5H5−. Give the point group of each of the
following molecules: (e) CH2Cl2; (f) BF3; (g) C6H6; (h) CO2; (i) CH4.
E2. Use Mathematica to show that the transpose of the matrix representation of
Cn in the basis (x, y, z)T is equal to its inverse.
E3. Use Mathematica to find the matrix representation for S4 in the basis
(x, y, z)T.
E4. Use Mathematica to show that the transpose of a product of 3 3 matrices
is equal to the product of transposes of matrix factors in reverse order,
(A.B)T = BT.AT.
E5. Use Mathematica to show that the inverse of a product of 3 3 matrices is
equal to the product of the inverses of the matrix factors in reverse order,
(A.B)−1 = B−1.A−1.
E6. Prove that square matrices B and A related by a similarity transformation
(B = X−1AX) have the same sum of their diagonal elements.
E7. Show that
(a) every element of a group A is conjugate to itself;

(b) if A is conjugate to B, then B is conjugate to A;
(c) if A is conjugate to B and to C, then B is conjugate to C.
E8. Use Mathematica to find the multiplication table for the C3v group. Is this an
abelian group?
E9. Considering a molecule with a center of symmetry, then the symmetry
species for infrared active vibrations are of u-type, whereas the symmetry
species for Raman active vibrations are of g-type. Explain.
E10. Find the reducible representations for vibrations of tetrahedral molecules
belonging to the point groups C3v, C2v, Cs, and C1.
Exercises 241
Further Reading
Cotton, F.A.: Chemical applications of group theory, 3rd edn. Wiley, New York (1990)
Kettle, S.F.A.: Symmetry and structure (readable group theory for chemists). Wiley, New York
(1995)
Vincent, A.: Molecular symmetry and group theory, 2nd edn. Wiley, New York (2001)
Molecular Structure
5
Abstract
The first two sections of this chapter define electron probability density and
electrostatic potential and illustrate these topics with the water molecule. The
ways of apportioning electrons of a molecule to different regions of space are
performed by population analysis methods. We illustrate the Mulliken popu-
lation analysis method for the methane molecule and the natural bond orbitals
method for methane, ethene, and ethyne molecules. Next, we present a typical
potential energy surface with one first-order saddle point obtained using a
combination of Morse functions and calculate intrinsic reaction coordinates for
the isomerization reaction HCN ! CNH and the symmetric reaction Cl− +
H3CCl ! ClCH3 + Cl−. Potential energy profiles for the rotations around the
C–C bonds of ethane and 1,2-dichloroethane are presented and discussed. In
particular, the potential energy profiles for the staggered conformation of ethane
and the synclinal and antiperiplanar conformations of 1,2-dichloroethane,
combined with vibrational calculations, enable one to estimate the amplitudes of
the corresponding torsional motions. The last section of this chapter comprises
chiral molecules, the Cahn–Ingold–Prelog rules for distinguishing R and
S enantiomers of carvone, the polarimeter, and the way optically active
molecules interact with plane-polarized electromagnetic radiation. At the end of
this chapter, the student can find several Mathematica codes (Natural Bond
Orbitals for CH4, Potential Energy Surface, Right and Left Helices, Optical
Rotation) with detailed explanations of new commands, a glossary of important
scientific terms, and a list of exercises, whose complete answers are given after
the appendix.

DOI 10.1007/978-3-319-41093-7_5
244 5 Molecular Structure
5.1 Electron Probability Density
Consider a molecule with n electrons. The wave function for the ground state can be
written as
Wðs1 ; . . .; sn Þ ð5:1Þ
where each si represents the set of three spatial and one spin coordinates of electron
i. The probability of finding electron 1 in volume dv1 for any position of the
remaining electrons and any spin of all electrons is given by
Z Z
dpðr1 Þ ¼ ... W0 ðs1 ; . . .; sn ÞW0 ðs1 ; . . .; sn Þdv2 . . .dvn dr1 . . .drn dv1 ð5:2Þ
where r1 = (x1, y1, z1) is the position vector in some reference system of Cartesian
axes, the integral over the spatial coordinates runs from electrons 2 to n, and the
integral over the spin coordinates runs from 1 to n. The electrons are indistin-
guishable (the labeling of electrons is arbitrary), and so the probability of finding
any electron in a volume dv is n times (5.2),
dpðrÞ ¼ q0 ðrÞdv ð5:3Þ
where
Z Z
q0 ðrÞ ¼ n ... W0 ðs1 ; . . .; sn ÞW0 ðs1 ; . . .; sn Þdv2 . . .dvn dr1 . . .drn ð5:4Þ
is the electron probability density for the molecule’s ground state.

The ground state electron probability density is a function of three independent
variables that are the coordinates of each point in the three-dimensional space
around the molecule’s nuclei,
q0 ¼ q0 ðx; y; zÞ ð5:5Þ
Having four variables (x, y, z, and the function value q0), this function cannot be
plotted as a whole, since the world we live in is a three-dimensional world. In order
to reduce the number of variables by one, we take the set of all points with
coordinates (x, y, z) that satisfy the equation
q0 ðx; y; zÞ ¼ c ð5:6Þ
where c is a constant called an isovalue (from Greek isos “equal”). The above
equation can be solved to obtain the z value as a function of x and y, that is,
q0 ðx; y; zÞ ¼ c ) z ¼ f ðx; yÞ ð5:7Þ

5.1 Electron Probability Density 245
Fig. 5.1 Total electron probability density of the H2O molecule represented by density surfaces
with isovalues 0.04000, 0.00400, and 0.00040 (Gaussian 09 B3LYP/cc-pVTZ calculation with
figure presented by GaussView)
Being a function of two variables, f(x, y) represents a surface [recall that a function
of one variable y = g(x) represents a curve], and the surface derived from (5.6) is
called an isodensity surface, because it is the set of points for which q0 is equal to a
specified isovalue.
Figure 5.1 presents three closed surfaces of the electron probability density for
H2O. Note that the surface with density isovalue 0.04000 is inside the surface with
isovalue 0.00400, which, in turn, is totally contained within the surface whose
isovalue is 0.00040. This relationship between surfaces with different isovalues
resembles that of a set of Russian dolls: a smaller doll fits inside a larger one. The
probability of finding any of the 10 H2O electrons within the 0.00040 isodensity
surface should be close to 100 %, and so this isodensity surface can be used to
represent the overall shape of the molecule.
5.2 Electrostatic Potential
The force exerted by point charges q1 and q2 on each other is given by

q1 q2
F¼ 2
ð5:8Þ
4pe0 r12
where r12 is the distance between the charges, and e0 is the permittivity of vacuum
(e0 = 8.854187817 10−12 C2 N−1 m−2, see Appendix), and the potential energy
of interaction is
q1 q2
Ep ¼ ð5:9Þ
4pe0 r12
Consider now a simple charge distribution formed by two charges equal in

magnitude and of opposite signs, in fixed positions (Fig. 5.2). Is the dominant
electrostatic influence on point P1 due to the positive or the negative charge? And
on point P2? Intuitively, the electrostatic influence of the negative charge on point
P1 should be greater than that of the positive charge on the same point, since the
negative charge is closer to P1 than the positive charge. In turn, at point P2, the
closer positive charge results in a dominant electrostatic effect. This suggests that
the effect of each charge should be weighted by the inverse of the distance to the
considered point. In addition, when a point is equally distant from two charges of
the same sign and different magnitudes, the greater charge exerts the stronger
electrostatic effect. Hence, apart from the factor 1/(4pe0), at P1 the sum is negative,
meaning that the dominant influence comes from the negative charge,
q qþ q q
P1 ðr1 \r1 þ Þ þ ¼ þ \0
r1 r1 þ r1 r1 þ
where q is the absolute value of each charge. In turn, at P2, the sum is positive,
meaning that the dominant influence comes from the positive charge,
q qþ q q
P2 ðr2 þ \r2 Þ þ ¼ þ [0
r2 r2 þ r2 r2 þ
The electrostatic influence of a distribution of charges on a particular point

implies allocating a charge to that point, since charges interact with each other [see
(5.9)]. However, the physical quantity to be determined cannot depend on a charge
that does not belong to the original distribution of charges. In other words, the
additional charge has to be so small that it does not significantly change the electric
field created by the charge distribution. This infinitesimal positive charge at point
P is called a test charge qt. Let us now represent by w(∞ ! P) the reversible work
done by the electric field in providing the electrostatic force that transports qt from
infinity to the considered point P (Fig. 5.3). The electric field is defined in each
point of space. Hence, the work done by the electric field to transport the test charge
Fig. 5.2 The point P1 “feels” a dominant electrostatic influence from the nearer negative charge,
whereas P2 “feels” a dominant electrostatic influence from the nearer positive charge
5.2 Electrostatic Potential 247
Fig. 5.3 The electrostatic potential at point P is equal to the reversible work done by the electric
field in the transport of an infinitesimal test charge from infinity to point P, divided by the test
charge
between two points in space depends only on these points, not on the path followed
between them (the electrostatic force is a conservative force). In order to obtain a
function defined at each point in space that does not depend on the test charge qt,
one divides w(∞ ! P) by qt, thus obtaining the electrostatic potential defined by
wð1 ! PÞ
/ðPÞ ð5:10Þ
qt
The electrostatic potential energy of qt at infinity is null, because at infinity, qt does

not interact significantly with the charges of the distribution of charges.
The electrostatic potential difference between points A and B is given by the
integral of the electric field over a path between these points,
ZB
/B /A ¼ EðrÞ dr ð5:11Þ
A
where E is the electric field vector, whose magnitude is the ratio of the electrostatic
force exerted on the test charge qt and this charge, and dr is the infinitesimal
displacement vector along the path (A and B are the lower and upper limits of the
integral). The electric field belongs to the system. When the electric field vector and
the infinitesimal displacement vector have the same direction, E(r) dr > 0, the
system does work on the exterior, thus releasing energy, and so the work is neg-
ative. In turn, when E(r) dr < 0, the exterior does work on the system, and the
work is positive. When A is infinity (the test charge at infinity does not interact with
the charge distribution, and so the electric potential at infinity is zero, /∞ = 0), and
B is the point P, the electrostatic potential /P due to a point charge is given by
ZP Zr
q 1 q
/P /1 ¼ EðrÞ dr¼ dr ¼ ð5:12Þ
4pe0 r2 4pe0 r
1 1
The generalization of (5.12) for a charge distribution is straightforward, since the

electrostatic potential is an additive function, that is, the electrostatic potential of a
charge distribution is the sum of the electrostatic potentials derived from all char-
ges. In atomic units, where the unit of charge is e (the elementary charge), the unit
of length is a0 (the Bohr radius), and 4pe0 = 1 (see Appendix), we can write
X qi
/ðPÞ ¼ ð5:13Þ
i
riP
where the charge qi labeled by the subscript i includes not only its magnitude, but
also its sign, and riP stands for the distance between the charge i and the point
P where the electrostatic potential /(P) is evaluated.
Consider now a point on the 0.00040 isodensity surface of H2O (Fig. 5.1). The
electrostatic potential of a molecule is the sum of two main contributions, one
positive, given by the electrostatic potential produced by the nuclei with atomic
numbers Zn, the other negative, given by the electrostatic potential resulting from
the electron charge density defined at each point in space. The electron charge
density is obtained by multiplying the electron probability density q0 by the elec-
tron charge −e. Since the electron charge is −1 in atomic units, the electron charge
density is given in these units by −q0, and the electron charge in the volume
element dv is −q0dv. Hence,
X Zn Z
q0 ðrÞ
/ðPÞ ¼ dv ð5:14Þ
n
rnP reP
where the second term is a triple integral over dx dy dz (= dv) (E1).

Figure 5.4 shows the electrostatic potential for the H2O molecule mapped over
the total density surface with isovalue equal to 0.00040 using rainbow colors. Dark
red indicates the negative extreme value of the electrostatic potential. Dark blue is
used when the electrostatic potential reaches the extreme positive value. Positive
values of the electrostatic potential in a particular region of space indicate a
dominant influence of the nuclei in that region, which is said to be electrophilic
(from Greek philos “loving”). In turn, negative values of the electrostatic potential
in a particular region of space point to a dominant influence of the electrons, and
that region is said to be nucleophilic. An electrophilic region of a molecule has a
tendency to attract and interact with a nucleophilic region of another molecule.
Such an interaction may result in a chemical reaction. Electrostatic potential maps
for the active sites of enzymes are of great value in designing new drugs.
5.3 Mulliken Population Analysis 249
Fig. 5.4 Electrostatic potential for the H2O molecule mapped over the total density surface with
isovalue equal to 0.00040 (Gaussian 09 B3LYP/cc-pVTZ calculation with figure presented by
GaussView)
5.3 Mulliken Population Analysis
The integral of the ground state electron density of a molecule over all space around
the nuclei [see (5.4)] is equal to the total number of the molecule’s electrons,
Z
q0 ðrÞdv ¼ n ð5:15Þ
There are multiple ways of apportioning n to different regions of space in the

molecule, such as atoms and overlap regions, or atoms and bonding regions. The
methods of so doing are generally called population analysis and were developed
initially by Mulliken, who assigned the total electron charge of a molecule to its
atoms and overlap regions (Mulliken 1955).
5.3.1 Density Matrix
The ground state total electronic wave function for a closed-shell configuration is
given by the following Slater determinant of molecular spin-orbitals,
h i
/ /
W0 ¼ ð1; . . .; nÞ ¼ /1 /
1 2 2 . . ./n=2 /n=2 ð5:16Þ
where each column corresponds to a molecular spin-orbital, and the bar over /i
indicates beta spin-orbitals (E2). The spatial factors of the molecular spin-orbitals
are given by linear combinations of Slater-type atomic orbitals
X
b
/i ¼ csi vs ð5:17Þ
s¼1
Substitution of (5.16) and (5.17) in (5.4) gives
X
n=2 b X
X b X
n=2
q0 ðrÞ ¼ 2 /j /j ¼ 2 crj csj vr vs ð5:18Þ
j¼1 r¼1 s¼1 j¼1
and integration over all space leads to
Z b X
X b X
n=2 b X
X b
q0 ðrÞdv ¼ 2 crj csj Srs ¼ Prs Srs ¼ n ð5:19Þ
r¼1 s¼1 j¼1 r¼1 s¼1
where Prs represents the rs density matrix element and Srs is the corresponding
overlap integral,
X
n=2 Z
Prs ¼ 2 crj csj Srs ¼ vr vs dv ð5:20Þ
j¼1
Both the density and overlap matrices are symmetric matrices. For real coefficients,
X
n=2 X
n=2
Prs ¼ 2 crj csj ¼ 2 crj cTjs ¼ 2Cr CsT ð5:21Þ
j¼1 j¼1
where Cr represents the r-row vector [1 (n/2) vector] that multiplies the transpose
of the s-row vector [(n/2) 1 vector]. The result of this vector product is a number,
the Prs element of the density matrix. In matrix notation, we can write
P ¼ 2CCT ð5:22Þ
where the letters in boldface represent matrices.

The double summation over the atomic basis functions in the last equation of
(5.19) includes diagonal terms for which r = s and Srr = 1 (the atomic basis
functions are normalized) and off-diagonal terms for which PrsSrs = PsrSsr (the
lower and upper matrix triangles of the matrices P and S are equal, that is, these
matrices are symmetric). Hence,
b X
X b X
b b X
X b
n¼ Prs Srs ¼ Prr þ 2 Prs Srs ð5:23Þ
r¼1 s¼1 r¼1 r [ s s¼1
where the off-diagonal elements correspond to the matrix lower triangle (the row
number is greater than the column number).
5.3.2 Minimal Basis Set Calculation for CH4
In order to illustrate the Mulliken population analysis method without having to

deal with extremely large matrices, we turn to a minimal basis set calculation for
CH4. This calculation was performed by the system of programs Gaussian 09, at
the Hartree–Fock level, with the basis set STO-3G (HF/STO-3G calculation). The
five doubly occupied lowest-energy molecular orbitals (CH4 has 10 electrons) are
mathematically expressed by the coefficients (eigenvectors) and corresponding
energies (eigenvalues), as shown in the molecular orbital coefficient matrix. On
application of the molecule’s symmetry operations, these eigenvectors transform
like two of the five irreducible representations of the Td group, namely A1 and T2
(use the character table and see Fig. 5.5).
Molecular orbital coefficients

A1 A1 T2 T2 T2
Eigenvalues (eV) −300.1 −24.8 −14.2 −14.2 −14.2
C 1s 0.99193 −0.22143 0.00000 0.00000 0.00000
C 2s 0.03836 0.62759 0.00000 0.00000 0.00000
C 2px 0.00000 0.00000 0.00000 0.00000 0.57197
C 2py 0.00000 0.00000 0.57197 0.00000 0.00000
C 2pz 0.00000 0.00000 0.00000 0.57197 0.00000
H 1s1 −0.00701 0.18057 0.30082 0.30082 0.30082
H 1s2 −0.00701 0.18057 −0.30082 0.30082 −0.30082
H 1s3 −0.00701 0.18057 0.30082 −0.30082 −0.30082
H 1s4 −0.00701 0.18057 −0.30082 −0.30082 0.30082
Character table for the Td point group
E 8C3 3C2 6S4 6rd
A1 1 1 1 1 1 x2 + y2 + z2
A2 1 1 1 −1 −1
E 2 −1 2 0 0 (2z2 − x2 − y2, x2 − y2)
T1 3 0 −1 1 −1 (Rx, Ry, Rz)
T2 3 0 −1 −1 1 (x, y, z) (xy, xz, yz)
http://symmetry.jacobs-university.de
Fig. 5.5 Molecular orbitals for CH4 represented by surfaces with isovalues equal to 0.06 and the
corresponding symmetry species of the Td point group. Lobes with different grades of gray
correspond to different signs. HF/STO-3G Calculation performed by Gaussian 09, and molecular
orbitals presented by GaussView
The first relevant feature of these orbitals is their delocalization over the whole
molecule. In particular, all the doubly occupied orbitals have nonzero coefficients in
each of the four hydrogen atoms’ 1s orbitals. Note that the molecular orbital with the
lowest energy is essentially an atomic orbital of the carbon atom core, since the largest
coefficient refers to the carbon 1s orbital (the 2p orbitals have zero coefficients), and
the coefficients for the hydrogen atoms’ 1s orbitals are close to negligible. The
molecular orbital with the second-lowest energy has its largest coefficient on the
carbon 2s orbital. These two molecular orbitals transform like the A1 irreducible
representation of the Td group, i.e., they are totally symmetric orbitals. The three
molecular orbitals of highest energy are degenerate (they have the same energy
eigenvalue). Each of these molecular orbitals has a C2 axis that is coincident with each
of the x-, y-, and z-axes and transforms like each of these coordinates. In addition, each
of the C2 symmetry axes coincides with an S4 axis. The S4 rotation–reflection sym-
metry operation includes a 2p/4 rotation around the S4 axis followed by a reflection in
the plane perpendicular to the axis. Therefore, application of S4 twice is equivalent to
one C2 symmetry operation. Note that the coefficients of degenerate molecular orbitals
are not unique, since these can always be replaced by linear combinations that are
orthogonal and correspond to the same energy eigenvalue.
The five molecular orbitals of methane are represented in Fig. 5.5 by surfaces
whose isovalues are equal to 0.06. Considering the orientation and the different
signs of the two lobes of each of the triple degenerate molecular orbitals, it can be
concluded that the character of the S4 rotation–reflection symmetry operation is
equal to −1. Hence, these three molecular orbitals transform as the symmetry
species T2.
Density matrix
C 1s C 2s C 2px C 2py C 2pz H 1s1 H 1s2 H 1s3 H 1s4
C 1s 2.06590
C 2s −0.20184 0.79067
C 2px 0.00000 0.00000 0.65429
C 2py 0.00000 0.00000 0.00000 0.65429
C 2pz 0.00000 0.00000 0.00000 0.00000 0.65429
H 1s1 −0.09387 0.22611 0.34412 0.34412 0.34412 0.60826
H 1s2 −0.09387 0.22611 −0.34412 −0.34412 0.34412 −0.11568 0.60826
H 1s3 −0.09387 0.22611 −0.34412 0.34412 −0.34412 −0.11568 −0.11568 0.60826
H 1s4 −0.09387 0.22611 0.34412 −0.34412 −0.34412 −0.11568 −0.11568 −0.11568 0.60826
The density matrix [matrix whose general elements are Prs; see (5.21)] is a
symmetric matrix, and so we present only the lower matrix triangle (E3). In
addition, the elements of the full Mulliken population analysis matrix (see below)
are given by PrsSrs [see (5.19)], since they correspond to the product of the density
matrix elements Prs and the corresponding overlap integrals Srs. These can be
obtained by dividing the Mulliken population matrix elements PrsSrs by the cor-
responding density matrix elements Prs (E4). Note that the diagonal elements for
the density matrix and the full Mulliken population analysis are the same, since the
diagonal overlap integrals are 1, Srr = 1.
Full Mulliken population analysis matrix
C 1s 2.06590
C 2s −0.05013 0.79067
C 2px 0.00000 0.00000 0.65429
C 2py 0.00000 0.00000 0.00000 0.65429
C 2pz 0.00000 0.00000 0.00000 0.00000 0.65429
H 1s1 −0.00594 0.11187 0.09331 0.09331 0.09331 0.60826
H 1s2 −0.00594 0.11187 0.09331 0.09331 0.09331 −0.01993 0.60826
H 1s3 −0.00594 0.11187 0.09331 0.09331 0.09331 −0.01993 −0.01993 0.60826
H 1s4 −0.00594 0.11187 0.09331 0.09331 0.09331 −0.01993 −0.01993 −0.01993 0.60826
Gross orbital populations
1.99201 1.18802 1.02753 1.02753 1.02753 0.93433 0.93433 0.93433 0.93433
The sum of the elements of the Mulliken population matrix over one particular
orbital, that is, over a row or a column, leads to the gross orbital population (see
above), and the sum of the gross orbital populations over all atomic orbitals gives
the total number of electrons [see (5.23)]. In turn, subtracting the sum of gross
orbital populations belonging to a particular atom in the molecule from the total
number of electrons in the isolated atom gives the Mulliken atomic charges (C
−0.262620, H1 0.065670, H2 0.065670, H3 0.065670, H4 0.065670), whose sum
(the total charge of the molecule) is zero. If the matrix elements of the Mulliken
population analysis matrix corresponding to a specified atom are summed, one
obtains the population condensed to that atom (see below) (E5)
Populations condensed to atoms (all electrons)
1C 2H 3H 4H 5H
1 C 4.71918
2 H 0.38586 0.608263
3 H 0.38586 −0.019929 0.608263
4 H 0.38586 −0.019929 −0.019929 0.608263
5 H 0.38586 −0.019929 −0.019929 −0.019929 0.608263
The strong dependence between the Mulliken population analysis results and the
basis set used in the calculation enable us to conclude that this population analysis
method is of reduced practical utility.
5.4 Natural Bond Orbitals
An alternative to the Mulliken population analysis consists in condensing the

electron density of a molecule into the lowest few orbitals that, in localized 1-center
and 2-center regions of the molecule, have the maximum electron occupancy and
correspond to Lewis-like electron pairs localized in atoms and chemical bonds. In
order to illustrate the fundamental concepts of the natural bond orbital method,
we consider the methane molecule and take full advantage of the equivalence of the
four C–H bonds.
A regular tetrahedral molecule like CH4 can be included in a cube, as Fig. 5.6
shows: the carbon atom occupies the center of the cube, two of the hydrogen atoms
occupy the vertices along a diagonal of one of the faces, and the other two hydrogen
atoms are located on the perpendicular diagonal of the opposed cube face
(Fig. 5.6a). The angle a to the center of a regular tetrahedron is called the tetra-
hedral angle (Fig. 5.6b). Simple trigonometry can show that the cosine of this
angle is equal to −1/3 (a 109.47°), and Cartesian coordinates can be easily
assigned to the vertices of a regular tetrahedron, as shown in Fig. 5.6c (E6).
5.4.1 Hybrid Atomic Orbitals
The electronic configuration of an isolated carbon atom is 1s2 2s2 2p2. Since all
directions that emerge from the nucleus of an isolated atom are equivalent (the
space around the nucleus is isotropic), the 2p electrons of an isolated carbon atom
are uniformly distributed over the 2px, 2py, and 2pz orbitals, which share the same
electron density. Now consider the carbon atom of a CH4 molecule and four
equivalent independent linear combinations of the 2s, 2px, 2py, and 2pz orbitals
directed to the hydrogen atoms on the vertices of a regular tetrahedron. These
Fig. 5.6 a CH4 inscribed in a cube. b Tetrahedral angle. c Cartesian coordinates for the vertices
of a regular tetrahedron
5.4 Natural Bond Orbitals 255
orbitals can be made orthonormal and are called hybrid atomic orbitals. Having
spherical symmetry, the 2s orbital participates in them with the same coefficient (we
can begin by assigning 1, apart the normalization constant for the whole hybrid
orbital). The subscript in each of the 2px, 2py, 2pz orbitals indicates how these
atomic orbitals transform under a symmetry operation. For example, 2px transforms
like the x coordinate, that is, 2px = x f(r), where r is the distance from the point to
the origin. Therefore, the 2px, 2py, and 2pz orbitals act like unit vectors along the
x-, y-, and z-axes, and their coefficients in the hybrid orbitals are given by the
corresponding Cartesian coordinates of each vertex to which the hybrid orbital is
directed (Fig. 5.6c), apart from the normalization constant.
The hybrid orbital directed to the (1, −1, −1) vertex is given by
Nð2s þ 2px 2py 2pz Þ ð5:24Þ
where N is the normalization constant. Since the atomic orbitals are orthogonal and
normalized, the normalization constant of the hybrid atomic orbital is given by the
inverse of the square root of the sum of the squared orbital coefficients. In partic-
ular, for (5.24) we can write
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
N ¼ 1= ð þ 1Þ2 þ ð þ 1Þ2 þ ð1Þ2 þ ð1Þ2 ¼ ð5:25Þ
2
We now consider an spk hybrid orbital
1 pffiffiffi
hA ¼ pffiffiffiffiffiffiffiffiffiffiffi s þ kpA ð5:26Þ
1þk
where pA is a contracted p orbital (a linear combination of primitive p orbitals)

directed to atom A. Since the s and p orbitals centered on the same atom are
orthogonal, the squares of the coefficients of s and pA,
1 k
fs ¼ fp ¼ ð5:27Þ
1þk 1þk
are the s and p fractions, and k is the hybridization parameter, given by the ratio
fp/fs (E7, E8, E9). Note that the sum of fs and fp is equal to 1.
Let us now consider the spkA and spkB hybrids directed to atoms A and B,
1 pffiffiffiffiffi 1 pffiffiffiffiffi
hA ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi s þ kA pA hB ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi s þ kB pB ð5:28Þ
1 þ kA 1 þ kB
The orthogonality of hA and hB leads to

Z
1
Sp A p B ¼ pA pB dv ¼ pffiffiffiffiffiffiffiffiffiffi ð5:29Þ
kA kB
(E10). The scalar product of two unit vectors directed to atoms A and B is given by
dÂ :d^B ¼ cos x ð5:30Þ
These unit vectors are aligned with pA and pB orbitals, which in turn behave like
unit vectors whose scalar product is given by the overlap integral of pA and pB.
Hence, we can equate the second members of (5.29) and (5.30) and write
1
pffiffiffiffiffiffiffiffiffiffi ¼ cos x ð5:31Þ
kA kB
an equality known as Coulson’s theorem. Applying (5.31) to equivalent spk hybrids

leads to
1
k¼ ð5:32Þ
cos x
For k = 3 (sp3 hybrids), x is equal to the tetrahedral angle (109.4712°), for k = 2

(sp2 hybrids), x = 120°, and for k = 1 (sp hybrids), x = 180º.
If we now express the spk hybrid in terms of the px, py, and pz orbitals,
h ¼ cs s þ cx px þ cy py þ cz pz ð5:33Þ
then k is given by
p fraction c2x þ c2y þ c2z

k¼ ¼ ð5:34Þ
s fraction c2s
This expression shows that the hybridization parameter k depends on the coeffi-
cients of s and p orbitals and so on the level and basis set of the calculation. In
addition, (5.34) allows one to conclude that the k value of an spk atomic hybrid
orbital cannot be restricted to integer values 3, 2, and 1. The integer values of k are
imposed by symmetry and correspond to four equivalent hybrid orbitals (Td point
group, example CH4), three equivalent hybrid orbitals (D3h point group, example
BF3) and two equivalent hybrid orbitals (D∞h point group, example CO2),
respectively.
5.4.2 Natural Bond Orbitals for CH4
Consider now the HF/STO-3G scan of the CH bond length RCH in CH4. Keeping
the regular tetrahedron geometry, all HCH angles are equal to the tetrahedral angle,
and the CH bonds have the same length. With these geometric constraints, there is
only one independent variable, the CH bond length, RCH. For large values of RCH,
the interactions between the carbon atomic orbitals and the hydrogen atomic
orbitals are negligible, and so all orbitals reduce to those of isolated atoms. For RCH
smaller than 2.0 Å (this value may vary slightly according to the calculation level
and the basis set of functions), the p fraction becomes equal to ¾, i.e., sp3 hybrids
are formed. Figure 5.7 shows the occupation of bond orbitals and the total
molecular energy as functions of RCH. At RCH = 1.08 Å, the occupation of the
orbitals attains a maximum, and the total molecular energy reaches a minimum.
Molecular orbitals localized in the bonds that attain the maximum occupation
allowed by the Pauli exclusion principle (n = 2) are called natural bond orbitals
(NBOs). The carbon 1s orbital remains essentially a doubly occupied orbital of the
carbon atom core and is thus called a natural atomic orbital (NAO). Each of the
natural bond orbitals is a linear combination of one sp3 hybrid on the carbon atom
and the 1s orbital on the hydrogen atom to which the sp3 hybrid is directed. In
matrix form, we have
Orbital coefficients
C 1s 1.0000 0.0000 0.0000 0.0000 0.0000
C (sp3)1 0.0000 0.7253 0.0000 0.0000 0.0000
C (sp3)2 0.0000 0.0000 0.7253 0.0000 0.0000
C (sp3)3 0.0000 0.0000 0.0000 0.7253 0.0000
C (sp3)4 0.0000 0.0000 0.0000 0.0000 0.7253
H 1s1 0.0000 0.6885 0.0000 0.0000 0.0000
H 1s2 0.0000 0.0000 0.6885 0.0000 0.0000
H 1s3 0.0000 0.0000 0.0000 0.6885 0.0000
H 1s4 0.0000 0.0000 0.0000 0.0000 0.6885
The density matrix that corresponds to this coefficient matrix can be obtained by
applying (5.22); it is given by
Density matrix
C 1s C (sp3)1 C (sp3)2 C (sp3)3 C (sp3)4 H 1s1 H 1s2 H 1s3 H 1s4
C 1s 2.0000
C (sp3)1 0.0000 1.0521
C (sp3)2 0.0000 0.0000 1.0521
C (sp3)3 0.0000 0.0000 0.0000 1.0521
C (sp3)4 0.0000 0.0000 0.0000 0.0000 1.0521
H 1s1 0.0000 0.9987 0.0000 0.0000 0.0000 0.9481
H 1s2 0.0000 0.0000 0.9987 0.0000 0.0000 0.0000 0.9481
H 1s3 0.0000 0.0000 0.0000 0.9987 0.0000 0.0000 0.0000 0.9481
H 1s4 0.0000 0.0000 0.0000 0.0000 0.9987 0.0000 0.0000 0.0000 0.9481
Fig. 5.7 Occupation of bond orbitals and total molecular energy of CH4, as functions of RCH.
Natural bond orbitals (NBOs) occupation and total molecular energy as functions of RCH, for
regular tetrahedral arrangements of CH4. Results from an HF/STO-3G calculation performed by
Gaussian 09 and graphs obtained with Mathematica
The Mathematica code M1 shows that the eigenvalues obtained by diagonal-

izing the above density matrix correspond to the maximum occupation number of
the natural bond orbitals (n = 2), and the sum of the eigenvalues is equal to the total
number of electrons of the molecule.
Figure 5.8 presents the natural bond orbitals for CH4 represented by surfaces
with isovalues equal to 0.06 (lobes with different shades of gray correspond to
different signs). Unlike the canonical molecular orbitals that are delocalized over
the CH4 molecule and transform like the symmetry species of the Td point group,
the natural bond orbitals are equivalent orbitals localized in the C–H bonds, have
maximum occupation (n = 2) and are in direct correspondence with Lewis dot
structures.
5.4.3 Natural Bond Orbitals for H2C=CH2
We now turn to the ethene molecule (ethene is the systematic name for ethylene),
H2C=CH2, where each carbon atom is bonded to three neighboring atoms: two
hydrogen atoms and one carbon atom. The total number of valence electrons in the
molecule is 12, four from each carbon atom, one from each hydrogen atom. The
corresponding six electron pairs include four in the C–H bonds and two in the
carbon–carbon double bond.
The natural bond orbitals for the valence electrons of the ethene molecule are
represented by surfaces with isovalue equal to 0.06 and are shown in Fig. 5.9 by
increasing order of energy. The lower row shows a natural bond orbital that has a
nonzero electron density along the internuclear bond axis and results from a
head-on overlap of atomic orbitals on both carbon atoms. It is called a r bond. The
middle row shows the C–H bonds that are equivalent by symmetry and have the
same energy (are degenerate), and the natural bond orbital in the upper row is
formed by two lobes, one above the plane of the nuclei, the other below, thus
exhibiting zero electron density in the internuclear axis and in the plane of all the
Fig. 5.8 Natural bond orbitals for CH4 represented by surfaces with isovalues equal to 0.06, one
per CH electron bond pair. HF/STO-3G Calculation by Gaussian 09, representation of molecule
and natural bond orbitals presented by GaussView
Fig. 5.9 Natural bond orbitals for H2C=CH2 illustrated by surfaces with isovalue equal to 0.06,
one for each electron bond pair. They are ordered by increasing energy: one CC r bond, four
degenerate CH r bonds, and one CC p bond (Gaussian 09 HF/STO-3G calculation; natural bond
orbitals are drawn by GaussView)
nuclei (nodal plane). This bond results mainly from overlap of p orbitals with
parallel axes that are perpendicular to the molecular plane and is called a p bond.
5.4.4 Natural Bond Orbitals for HCCH
In the ethyne molecule (ethyne is the systematic name for acetylene), HCCH,
each carbon atom is bonded to two neighboring atoms: one carbon atom and one
hydrogen atom. The ethyne molecule is linear. The total number of valence elec-
trons in the ethyne molecule is 10, four from each carbon atom, one from each
hydrogen atom. The corresponding five electron pairs include two electrons pairs
for the carbon–hydrogen bonds and three electron pairs for the carbon–carbon triple
Fig. 5.10 Doubly occupied natural bond orbitals for the ethyne molecule are represented by
surfaces with isovalue 0.06 and are shown by order of increasing energy: one CC r bond, two
degenerate CH r bonds, two CC p bonds. HF/STO-3G Calculation performed by Gaussian 09,
and natural bond orbitals visualized by GaussView
bond (one r and two p bonds). The natural bond orbitals are presented in Fig. 5.10
by order of increasing energy, that is, E(r C–C bond) < E(degenerate CH
bonds) < E(degenerate p CC bonds).
5.4.5 CH Hybrids in CH4, H2C=CH2 and HCCH
Considering now the CH hybrid orbitals in the methane, ethene, and ethyne
molecules (find details of the calculations in Sect. 5.4.6), a clear trend can be
observed, namely, the increasing s orbital fraction in the spk hybrid orbitals, as we
go from CH4, to H2CCH2 and HCCH (see table below). The greater s fraction of the
hybrid orbitals in the ethyne molecule allows negative charge to be held closer to
the nucleus in the carbon atom, thus resulting in a greater acidic character of the
hydrogen atoms in ethyne.
Carbon atom spk hybrids and the corresponding s orbital fractions
Molecule CH hybrids s fraction CC hybrids s fraction
CH4 sp3 0.25 – –
H2C=CH2 sp2.37 0.30 sp1.45 0.41
HCCH sp1.09 0.48 sp0.91 0.52
5.4.6 Molecular Geometries and Electrostatic Potentials
The calculation for determining a molecular geometry consists in systematically

varying the geometric variables until an energy minimum is reached. Nucleus–
nucleus and electron–electron interactions involve particles of the same electric
charge, thus being repulsive and corresponding to positive energy variations, that is,
to energy variations that increase the total molecular energy. In turn, electron–
nucleus interactions are attractive and involve negative energy variations, thus
contributing to lower the molecular energy. When a balance between attractive and
repulsive interactions is achieved for certain values of the geometric variables, the
molecular energy reaches a stationary point, that is, a point with zero slope, and the
calculation is called geometry optimization. In order to decide what type of sta-
tionary point (minimum or maximum) is obtained, a frequency calculation should
be carried out using the stationary point geometry. If first derivatives increase in the
vicinity of a stationary point (positive second-order derivatives), the stationary point
corresponds to an energy minimum, and a frequency calculation will not encounter
any imaginary vibrational frequency.
In order to calculate molecular geometries in close agreement with experimental
values, we used a B3LYP calculation with a valence triple-zeta basis set (cc-pVTZ).
The optimized geometric parameters for CH4, C2H4, and C2H2 and the corre-
sponding experimental values obtained from CCCBDB are shown in Fig. 5.11. The
geometry optimization of CH4 leads to a regular tetrahedral arrangement of the
hydrogen atoms around the central carbon atom, that is, a Td point group symmetry,
and a CH bond length equal to 1.088 Å, in close agreement with the experimental
value 1.087 Å (CCCBDB). The calculated geometric parameters of the ethene
molecule compare very well with the corresponding experimental CCCBDB values,
and those obtained for the ethyne molecule differ from the experimental CCCBDB
values by orders of magnitude of 10−3 Å.
We used the optimized geometries to obtain the electrostatic potentials mapped
over isodensity surfaces with isovalues equal to 0.000400 (Fig. 5.12). In agreement
with the variation of the s orbital fraction values, the hydrogen atoms of ethyne
Fig. 5.11 Ball and bond representation of the ethane, ethene, and ethyne molecules with
calculated versus experimental (CCCBDB) geometric parameters (bond lengths are in angstroms,
bond angles in degrees, experimental values in parentheses). B3LYP/cc-pVTZ calculations
performed by Gaussian 09 and molecular representations obtained with GaussView
Fig. 5.12 Total density surfaces (isovalue = 0.000400) mapped with the corresponding electro-
static potentials, from left to right, for methane, ethene, and ethyne molecules. B3LYP/cc-pVTZ
calculation performed by Gaussian 09, and representations of molecules and electrostatic
potentials visualized by GaussView
exhibit a strong positive electrostatic potential, in contrast with the hydrogen atoms
of ethene and methane.
5.5 Potential Energy Surfaces
Consider now the following hypothetical second-order reaction in gaseous phase:
A þ BC ! AB þ C ð5:35Þ
The system of three atoms, A, B, and C, can be described by three internal coor-
dinates, namely, the internuclear AB distance rAB, the internuclear BC distance rBC,
and the bond angle \ABC. For a defined angle, for example, \ABC ¼ p radians,
the potential energy difference DU between the potential energy of the interacting
three-atom system (A…B…C) and the sum of the energies of the three isolated
atoms, A, B, and C, is a function of variables rAB and rBC, DU = f(rAB, rBC). This
function describes a surface called a potential energy surface.
From the mathematical point of view, a general potential energy surface can be
characterized by its stationary points, i.e., by the points whose derivative is zero,
5.5 Potential Energy Surfaces 263
Fig. 5.13 Maximum, minimum, and saddle point at (0,0), for the functions −x2 − y2, x2 + y2 and
x2 − y2. Figure obtained with Mathematica
@U
¼ 0 k ¼ 1; 2; . . .; 3N 6 ð5:36Þ
@rk
where rk is each of the 3N − 6 coordinates for a system with N atoms, and each
stationary point can be classified as a maximum, a minimum, or a saddle point
(Fig. 5.13). In a first-order saddle point, the function takes a maximum in one of
the 3N − 6 internal coordinates and minima in the remaining 3N − 7 coordinates.
The potential energy surface of Fig. 5.14 corresponds to a linear triatomic
system XHH and was obtained using Morse functions (see Baggott et al. 1988). It
shows two valleys approximately parallel to the corresponding coordinate axes and
an upper pass between them through one first-order saddle point. The reactant
Fig. 5.14 Typical potential energy surface obtained using a combination of Morse functions.
Right view shows the minimum energy path passing by the first-order saddle point. Figure obtained
with Mathematica
channel and the product channel correspond to the chemical reaction X + H2 !

XH + H (see Mathematica code M2). The minimum-energy path converts reactants
into products, passes through the first-order saddle point, and defines the reaction
coordinate. The molecular state that corresponds to the maximum energy is the
transition state.
5.5.1 Intrinsic Reaction Coordinate
When the reaction coordinate begins at the saddle point and goes down the hill
toward the reactant channel (negative values of the reaction coordinate) and toward
the product channel (positive values of the reaction coordinate), the obtained
coordinate is called intrinsic reaction coordinate. Suggestively, the mathematical
method used in the calculation for obtaining the intrinsic reaction coordinate is the
method of steepest descent.
The linear triatomic molecules HCN (hydrogen cyanide) and HNC (hydrogen
isocyanide) are isomers. Conversion of HCN into HNC implies hydrogen migra-
tion, where the transition state is given by a cyclic triangular structure (Fig. 5.15).
The results presented in this figure show that the HCN ! HNC reaction is
endothermic.
Consider now the general reaction
Nu þ RX ! RNu + X ð5:37Þ
where Nu− stands for a nucleophile and X represents an electronegative atom

bonded to a tetrahedral carbon atom in the R radical. A chemical reaction of this
Fig. 5.15 Potential energy variation as a function of the intrinsic reaction coordinate for the
isomerization reaction HCN ! CNH (Gaussian 09 B3LYP/cc-pVTZ calculation). Figure obtained
with Mathematica
5.5 Potential Energy Surfaces 265
Fig. 5.16 Potential energy variation as a function of the intrinsic reaction coordinate for the
reaction Cl− + H3CCl ! ClCH3 + Cl− (Gaussian 09 B3LYP/aug-cc-pVTZ calculation). Fig-
ure drawn by Mathematica
type is called a nucleophilic substitution reaction. In fact, it is a substitution

reaction (X is substituted by Nu) resulting from a nucleophilic attack of Nu− on the
carbon atom bonded to X:
Nu þ Cd þ Xd ð5:38Þ
If the nucleophilic substitution reaction occurs in a single step, the approach of Nu−
is concerted with the withdrawal of X−. In this case, the transition state involves
both reactant species Nu− and RX, the reaction is bimolecular and named SN2, and
is said to occur in a bimolecular concerted step, with its rate equation given by
v ¼ k½RX½Nu ð5:39Þ
Consider now the bimolecular nucleophilic substitution reaction
CI þ H3 CCI ! CICH3 þ CI ð5:40Þ
whose potential energy variation as a function of the intrinsic reaction coordinate is

shown in Fig. 5.16. In this academic example of a chemical reaction, products and
reactants can be distinguished only by isotopic substitution and the conversion of
reactants into products implies the inversion of the methyl group, a movement that
recalls that of an umbrella suddenly struck by a blast of wind. In the transition state,
the CH3+ carbocation presents a planar symmetrical configuration with the carbon
atom equidistant from the chloride ions.
5.6 Molecular Conformations
5.6.1 Ethane
We now consider the ethane molecule, H3C–CH3, and the rotation of one of its
methyl groups around the CC axis. The distances between the hydrogen atoms of
the distinct methyl groups vary in the course of this methyl rotation (Fig. 5.17a) and
so does the total molecular energy. In order to define the relative position of the
methyl groups, four atoms bonded in sequence H(1)C(2)C(3)H(4) are considered,
and the angle measured in the clockwise direction between the H(1)C(2)C(3) and C
(2)C(3)H(4) planes is called a dihedral angle (Fig. 5.17b). Note that an angle
between two planes is measured in the plane perpendicular to the common edge.
Therefore, the dihedral angle for the ethane molecule is determined in the plane
perpendicular to the CC bond axis (the plane of the paper in Fig. 5.17c). We
consider the dihedral angle to vary in the range between 0° and 360°. A 60°
dihedral angle for the atomic sequence H(1)C(2)C(3)H(4) corresponds to a 300°
dihedral angle for the atomic sequence H(4)C(3)C(2)H(1).
The ethane molecule has one CC bond and six CH bonds. Around each tetra-
hedral center (each carbon atom), there are six bond angles (three CCH angles and
three HCH angles), whose sum covers the whole space, thus being equal to 4p
steradian (a constant). Hence, around each carbon atom there is one bond angle
redundancy, meaning that there are five independent bond angle variables, not six.
One additional coordinate is needed to define the relative position of the methyl
groups, that being one HCCH dihedral angle. Altogether, we have seven bond
lengths, plus 5 + 5 bond angles, plus one dihedral angle, making a total of 18
(= 3N − 6, where N = 8) geometric variables. In the potential energy plot of
Fig. 5.18, the dihedral angle is the independent variable. Changing the dihedral
angle will make the remaining 3N − 7 = 17 variables adjust their values to each
new value of the dihedral angle. One says that the dihedral angle scan is accom-
panied by relaxation of the molecular geometry (it is a relaxed scan).
The plot of the potential energy of the ethane molecule H3C–CH3 as a function of
the dihedral angle HC–CH is called a potential energy profile for the rotation
around the C–C bond (Fig. 5.18). As shown in this figure, for every initial value of
the dihedral angle, the potential energy retakes the same value after a dihedral angle
(a) (b) (c)
Fig. 5.17 a In the ethane molecule, methyl group rotation around the CC axis changes the
distance between hydrogen atoms, thus causing the total molecular energy to vary. b Numbering of
atoms involved in the dihedral angle. c The angle between 1-2-3 and 2-3-4 planes is measured in
the plane perpendicular to the CC axis. Representation of molecules visualized by GaussView
5.6 Molecular Conformations 267
Fig. 5.18 Potential energy profile for the rotation of a methyl group in the ethane molecule.
Staggered and eclipsed conformations correspond to stationary points, minimum, and maximum,
respectively (Gaussian 09 B3LYP/cc-pVTZ calculation). Graph drawn with Mathematica
variation of 120° (= 360°/3). When the dihedral angle HC–CH is equal to 0°

(= 360°), 120°, or 240°, the molecular energy reaches a maximum with respect to the
dihedral angle. Looking at the molecule along the CC axis, the hydrogen atoms in
the methyl group more distant from the observer are eclipsed by the hydrogen atoms
of the methyl group closer to the observer, leading to a greater repulsion between
hydrogen atoms of different methyl groups and, consequently, to a maximum in the
potential energy profile. This is the eclipsed conformation. In turn, when the dihedral
angle HC–CH is equal to 60°, 180°, or 300°, the molecule is at an energy minimum
in the potential energy profile. This is the staggered conformation.
Fig. 5.19 CC bond length of the ethane molecule as a function of the dihedral angle (Gaussian
09 B3LYP/cc-pVTZ calculation; graph obtained with Mathematica)
The relaxation of geometric variables during the scan of the dihedral angle can be
illustrated by the variation of the CC bond length. The observed variation of the CC
bond length with the dihedral angle is in phase with the potential energy profile for
the molecule, since the CC bond is along the threefold rotation axis of the molecule.
The largest increase of the CC bond length attenuates the effect of CH interactions of
distinct methyl groups in the eclipsed conformation (Fig. 5.19). In turn, the smallest
values of the CC bond length correspond to the staggered conformation.
The ethane molecule has 18 normal modes of vibration (N = 8, 3N − 6 = 18).
Among these, the normal mode with the lowest frequency is the torsional mode,
which is the out of phase oscillation of the methyl groups, whose atomic dis-
placement vectors are shown in Fig. 5.20 for the staggered conformation, with the
vectors applied to the hydrogen atoms of the same methyl group pointing in the
same direction, either clockwise or counterclockwise, whereas the vectors applied
to hydrogen atoms of different methyl groups point in opposite directions.
The calculated harmonic frequency for the torsional mode in the staggered con-
formation is 306 cm−1 (Gaussian 09 B3LYP/cc-pVTZ calculation). Figure 5.20
shows the potential energy curve with the minimum at the dihedral angle equal to 60°
(staggered conformation), the zero-point energy level (level 0), and the first torsion
level (level 1), which is 306 cm−1 above the zero-point level. The 0–1 transition
corresponds to the fundamental transition, whereas the 0–2 transition is the first
overtone (not shown). The amplitude of torsion in level 1 can be roughly estimated
from Fig. 5.20 as ±30°. This means that each methyl group spans an arc of about 60°,
and the total amplitude of motion relative to each other methyl group spans an arc of
about 120° (the methyl groups rotate in opposite directions, one clockwise, the other
counterclockwise). Hence, each methyl group probes practically all dihedral angles
corresponding to different conformations around the staggered conformation.
Fig. 5.20 Potential energy curve for the staggered conformation of ethane with harmonic energy
levels for the zero-point energy and the v = 1 energy level of the torsional mode (Gaussian 09
B3LYP/cc-pVTZ calculation; molecule representation obtained with GaussView; graph obtained
with Mathematica)
5.6.2 1,2-Dichloroethane
1,2-Dichloroethane is a toxic carcinogenic colorless liquid, mainly used in making

chemicals involved in plastics, rubber, and synthetic textile fibers. Its production is
primarily achieved by a catalyzed addition of chlorine to ethene. The relaxed scan
of the dihedral angle ClC–CCl leads to the potential energy profile of Fig. 5.21.
This is a symmetric profile with respect to the dihedral angle 180°, where the
energy reaches a global minimum. In fact, from 180° to 360°, the curve is the
reflected image of the curve from 0° to 180° (this symmetric behavior is also
observed with respect to the dihedral angle 0°, where the energy reaches a global
maximum). At the dihedral angle 0°, the conformation is synperiplanar, whereas
at the dihedral angle 180°, is antiperiplanar. These conformations have reflection
planes of symmetry that contain the nuclei of atoms in the ClCCCl dihedral
sequence. The global energy minimum is obtained for the dihedral angle equal to
180°. From 0° to 180°, there are a local minimum (synclinal conformation) and a
local maximum (anticlinal conformation). The energy difference between the
global maximum and the global minimum in 1,2-dichloroethane is roughly four
times the corresponding energy difference for the ethane molecule, an effect that can
be ascribed to the ClCl and ClH repulsive interactions involving the lone
electron pairs of the bulky chlorine atoms.
Fig. 5.21 Potential energy profile for the 1,2-dichloroethane molecule. Chlorine atoms are
represented in dark gray. Synclinal and antiperiplanar conformations correspond to minima,
synperiplanar and anticlinal conformations to maxima (Gaussian 09 B3LYP/cc-pVTZ calcula-
tion; molecule representation obtained with GaussView; graph obtained with Mathematica)
Fig. 5.22 Potential energy curves for the synclinal (left) and antiperiplanar conformations with
the corresponding zero-point energy and v = 1 torsional levels and the displacement vectors for the
corresponding vibrations (Gaussian 09 B3LYP/cc-pVTZ calculation; molecule representation
obtained with GaussView; graph obtained with Mathematica)
Figure 5.22 shows the potential energy curves for the synclinal and antiperi-
planar conformations, the displacement vectors for the corresponding torsional
vibrations at the calculated values of 110 cm−1 for the synclinal conformation and
116 cm−1 for the antiperiplanar conformation, and the zero-point energy and v = 1
levels. The large volumes of the chlorine and carbon atoms hide their small dis-
placement vectors. On each ClH2 set of atoms, the displacement vectors point in the
same direction, either clockwise or counterclockwise, whereas the vectors on dif-
ferent ClH2 sets of atoms point in opposing directions. Note that the carbon atoms
move slightly, so that the molecule’s center of mass is kept still. Looking now at the
amplitudes of the synclinal and antiperiplanar torsional motions in the v = 1 levels,
we find that these can be roughly estimated from Fig. 5.22 as ±12°. The substi-
tution of hydrogen in ethane by chlorine in 1,2-dichloroethane increases the
reduced mass of the torsional oscillator, thus reducing the torsional frequency [see
(3.55)] and, consequently, the amplitude of the torsional vibration. Unlike the
methyl groups in ethane, the ClH2C- groups of 1,2-dichloroethane have small
amplitudes of torsional oscillations around their corresponding equilibrium posi-
tions. Thus, the synclinal and antiperiplanar conformations of 1,2-dichloroethane
correspond to spectroscopically distinguishable conformational isomers, or
conformers.
5.6.3 Boltzmann Distribution
Consider now a sample containing 1,2-dichloroethane in equilibrium at a specified

temperature. What are the molecular populations of the antiperiplanar and synclinal
conformers? For a system in thermodynamic equilibrium, the population of the

energy level ei (= the number of molecules with this particular energy) is given by
the Boltzmann distribution (Ludwig Boltzmann, 1844–1906),
gi exp½ei =ðkB TÞ

fi ¼ ð5:41Þ
U
where
X
U¼ gk exp½ek =ðkB TÞ ð5:42Þ
k
T is the equilibrium temperature, kB is the Boltzmann constant

(= 1.38065 10−23 J K−1; see Appendix) and gi is the number of quantum states
with energy ei, that is, the degeneracy degree (levels with a single quantum state
have gi = 1). Being a summation over k, U is independent of k and represents the
total molecular population of all levels. Note that the range of each fi is between 0
and 1, and the sum over all fi equals 1. Hence, each fi represents a population
fraction. If the exponent of the exponential factor in (5.41) and (5.42) is multiplied
and divided by Avogadro’s constant, then the energy per molecule ei is substituted
by the energy per mole, Ei = NAei (NA is Avogadro’s constant), and the Boltzmann
constant is replaced by the gas constant R = NAkB. Then we obtain
gi exp½Ei =ðRTÞ
fi ¼ ð5:43Þ
U
where
X
U¼ gk exp½Ek =ðRTÞ ð5:44Þ
k
For a system of two levels with degeneracy degrees g1 and g2, we can write
f2 =g2
DE21 ¼ E2 E1 ¼ RT ln ð5:45Þ
f1 =g1
From (5.43) we can obtain the quotient fi/gi that appears in (5.45), for i = 1 and
i = 2 (E11).
Vibrational spectroscopic studies in solution with a nonpolar solvent have to be
carried out in order to obtain reliable experimental values concerning individual
conformers. The first step consists in assigning vibrational bands to distinct con-
formers. This is not a simple task! Usually, it is accomplished with the help of
relatively accurate frequency calculations for the considered conformers and of
additional spectroscopic information that might be gathered. In the vibrational
spectrum, one selects a pair of bands of the same vibrational mode in distinct
conformers that allows a reliable determination of their intensities. The ratio of
intensities of those bands equals the ratio of populations for the corresponding
conformers.
5.7 Chiral Molecules
Molecules that are mirror images of one another have the same atoms and the same
bonds, i.e., the same connectivity. Can we necessarily say that they are the same
molecule with identical physical and chemical properties? A test that enables us to
answer this question determines whether molecules that are mirror images of one
other can be superimposed. Superposition is a virtual test, since molecules are
impenetrable objects. Two molecules that are mirror images of each other and
cannot be superimposed have a relationship analogous to a person’s left and right
hands.
Two chlorofluoromethane molecules, CH2ClF, that are mirror images of each
other can be superimposed (Fig. 5.23). Superposition implies a C2 rotation of one
of the molecules followed by its translation to match the positions of the atoms of
the other molecule. The CH2ClF molecules of this pair represent the same com-
pound, with specific physical and chemical properties.
Replacement of the hydrogen atom of each CH2ClF molecule away from the
observer by one bromine atom results in two bromochlorofluoromethane molecules,
CHBrClF (Fig. 5.23). Unlike CH2ClF, the CHBrClF molecules exhibit a new
feature: they are mirror images of each other but cannot be superimposed. No
combined rotation and translation of one CHBrClF molecule of the pair can lead to
superposition of both molecules. In fact, these molecules are to each other as the left
and right hands of a person, and are thus called chiral objects (from Greek kheir,
hand). Being isomers, they are also called enantiomers (from Greek
enantios = opposite).
How are chirality and symmetry related? For a compound like CH2ClF having a
tetrahedral carbon atom as the central atom and not being chiral, we concluded that
a C2 rotation of the molecule followed by a translation leads to superposition with
the other molecule of the object-to-image pair of molecules. In turn, a chiral
molecule does not admit any Sn symmetry operation (Sn = rCn), including a
reflection (r = S1) or an inversion (i = S2).
Fig. 5.23 Left Ball and bond representations of two superposable chlorofluoromethane molecules
with an object-to-image plane mirror relationship to each other. Right Ball and bond
representations of two nonsuperposable bromochlorofluoromethane molecules with an
object-to-image relationship in a plane mirror to each other. Molecular representations obtained
with GaussView
5.7 Chiral Molecules 273
Fig. 5.24 R and

S enantiomers of
bromochlorofluoromethane
Molecules that are mirror images of each other and cannot be superimposed have
a chiral center. The Cahn–Ingold–Prelog rules [Sidney Cahn (1899–1981, Bri-
tish chemist), Christopher Kelk Ingold (1893–1970, British chemist), and Vladimir
Prelog (1906–1998, Croatian chemist), Nobel prize in chemistry in 1975 for
advances in the stereochemistry of organic molecules and chemical reactions]
assign different priorities to the atoms or groups bonded to a particular chiral center.
These priorities follow the decreasing order of the atomic numbers. The atom or
group with the lowest priority points away from the observer. If there is one
hydrogen atom bonded to the chiral center, then the hydrogen atom is pointing
away from the observer. If the decreasing priority of the atoms or groups bonded to
the chiral center is in the clockwise direction, the chiral center is assigned an
R (from Latin rectus = right) descriptor, whereas for the counterclockwise direc-
tion, the chiral center is assigned an S (from Latin sinister = left) descriptor
(Fig. 5.24).
Fig. 5.25 R and

S enantiomers of carvone
Carvone is an interesting chiral molecule: R-carvone smells like spearmint, S-

carvone smells like caraway, a plant whose seeds are used as seasoning. The fact
that enantiomers usually lead to widely different smells suggests that the olfactive
receptors are chiral.
The R and S enantiomers of carvone are shown in Fig. 5.25. The chiral center is
the tetrahedral carbon atom surrounded by curved arrows. Apart from the hydrogen
atom pointing away from the observer, the atoms in the first layer of atoms bonded
to the chiral center, CCC, score the same priority, since all of them are carbon
atoms. The priority rules consider one double bond equivalent to two single bonds.
Thus, applying priorities to the second layer of atoms, we obtain CCC >
CHH=CHH. The tie between the last two members of the second layer is solved in
the third layer, where the priority rules lead to OOC > CCH (see the following
table). On the whole, applying the Cahn–Ingold–Prelog priority rules to carvone
leads to the following sequence in order of decreasing priority: 1 = –C(=CH2)CH3,
2 = –C(H2)C=O, 3 = –C(H2)CH=C, 4 = H, as shown by the curved arrows around
the chiral center in Fig. 5.25 (E12).
Layers of atoms around the chiral center of carvone and Cahn–Ingold–Prelog priorities
Layer Atoms Priorities
1 CCC C=C=C
2 CCC, CHH, CHH CCC > CHH=CHH
3 OOC, CCH OOC > CCH
A right-handed helix follows a clockwise spiral path, whereas a left-handed

helix follows a counterclockwise spiral path. Right- and left-handed helices that
have the object-to-image relationship in a plane mirror cannot be superimposed
(Fig. 5.26), and therefore represent chiral structures. Helicoid structures frequently
appear in biological macromolecules such as nucleic acids and proteins. They also
exist in quartz, where SiO4 tetrahedra that share vertices form helicoid structures of
a single kind (right- or left-handed helices). From the mathematical point of view, a
right-handed helix can be obtained by combining a circle drawn in the clockwise
Fig. 5.26 Left- and right-handed helices having the object-to-image relationship in a plane mirror
are not superimposable. Figure obtained with Mathematica
direction (a right-handed circle) with a perpendicular line segment. In a similar way,

the left-handed helix results from a circle drawn in the counterclockwise direction
(a left-handed circle) combined with a perpendicular line segment. The Mathe-
matica code M3 illustrates this point.
Before we find out how the enantiomers of a chiral molecule interact with
electromagnetic radiation, we briefly consider a plane-polarized electromagnetic
radiation. This consists of one electric field and one magnetic field with sinusoidal
amplitudes that travel in phase in mutually perpendicular planes (Fig. 5.27). Usu-
ally, electrons in molecules interact mostly with the electric field. For that reason,
the plane of polarization of the radiation is referred to the electric field plane.
The electric field component of plane-polarized electromagnetic radiation has
two kinds of periodicity: space periodicity shows how the radiation progresses in a
defined direction, and its characteristic parameter is the wavelength k, whereas time
periodicity shows how the electric field amplitude changes with time at a defined
point along the propagation axis and is represented by the period s. The inverse of
period is the (linear) frequency m. The mathematical expression for a
plane-polarized electric field polarized in the z direction and propagating along the
x-axis is given by
E z ðx; tÞ ¼ E z0 cosðkx xt þ hÞ ð5:46Þ
where h is the phase, k is the angular wavenumber, x is the angular frequency,
2p 2p
k¼ ¼ 2p~v x ¼ ¼ 2pv ð5:47Þ
k s
and E z0 is the electric field maximum amplitude, whose square of the absolute value
is the energy carried by the wave.
Fig. 5.27 Plane-polarized electromagnetic radiation propagating along the x-axis. The electric
field oscillates in the vertical plane. Figure obtained with Mathematica
Fig. 5.28 Schematic setup for the main functions of a polarimeter

The enantiomers of a chiral molecule interact with electromagnetic radiation in

symmetric but distinct ways: one of the enantiomers rotates the plane of polar-
ization of light by a defined angle in the clockwise direction, whereas the other
enantiomer rotates the plane of polarization by the same angle but in the counter-
clockwise direction.
When unpolarized radiation passes through a polarizer, plane-polarized elec-
tromagnetic radiation is obtained. If this radiation travels across a solution whose
solute is a pure enantiomer, the polarization plane rotates by a specific angle and
direction. For an observer to see light in the polarimeter (Fig. 5.28), the analyzer
has to be rotated by the same angle of rotation, in the same direction, provided
polarizer and analyzer were initially aligned. The solute molecule is said to be
optically active or to exhibit optical activity. The magnitude and direction of the
angle of rotation a depend on the chemical nature of the involved enantiomer, its
concentration in solution c, and the interior length d of the sample tube, given by
a ¼ ½aTk cd ð5:48Þ
where [a] is called the specific rotation, i.e., the rotation for unit values of c and
d. The temperature and wavelength dependencies of a are included in the specific
rotation [a], measured for specified values of T and k. The yellow sodium D line
near the wavelength 589 nm is commonly used in measurements.
A plane-polarized electric field wave can be decomposed into left- and
right-circularly polarized electric field waves (Fig. 5.29) that, in an optically active
medium, have different refractive indices, thus propagating at different speeds.
Fig. 5.29 A left-circularly polarized wave added to a right-circularly polarized wave gives a
plane-polarized wave. Figure obtained with Mathematica
Fig. 5.30 On traversing an optically active medium, left- and right-circularly polarized waves
change phase difference progressively, and the resulting wave gradually rotates. In a polarimeter,
the views along the sample tube show the rotation angle for increasing lengths of the sample tube.
Hence, the phase relationship between left- and right-circularly polarized waves
changes progressively, and the resulting wave gradually rotates (Fig. 5.30). The
Mathematica code M4 illustrates this point.
Mathematica Codes
M1. Natural Bond Orbitals for CH4
Using the results of a Gaussian 09 HF/STO-3G calculation on CH4, in particular

the coefficient matrix that presents the five natural orbitals of CH4 in terms of the
1s carbon orbital, four sp3 hybrid atomic orbitals on carbon, and four 1s hydrogen
atom orbitals, this Mathematica code shows that the eigenvalues obtained by
diagonalizing the density matrix are 2 for the occupied orbitals (the maximum
occupation number in accord with the Pauli exclusion principle) and 0 for the
unoccupied orbitals, and the sum of the eigenvalues is equal to the total number of
electrons of the molecule. The density matrix is obtained from the coefficient matrix
using (5.22).
The Mathematica command Chop replaces real numbers whose absolute values
are smaller than 10−4 by 0. The command Transpose transposes a matrix, and
the commands Eigenvalues and Eigenvectors give the eigenvalues and
eigenvectors of a square matrix.
M2. Potential Energy Surface
This potential energy surface is modeled for a linear triatomic system XHH as the
sum of three Morse functions [see (3.68)], with one of them representing an
interaction term (see Baggott et al. 1988). The channels linked by a first-order
saddle point correspond to the chemical reaction X + H2 ! XH + H. The disso-
ciation energies for the reactant and product channels are equal to 20 and 50
arbitrary units of energy, and the force constants for the H2 and XH diatomics are
equal to 40 and 80 arbitrary units of energy per square length. The equilibrium point
for each Morse function is defined at 1.2.
M3. Right- and Left-Handed Helices
The clockwise screwing motion of a right-handed helix moves the helix away from
the observer. From the mathematical point of view, a right-handed helix can be
obtained by combining a circle drawn in the clockwise direction with a perpen-
dicular line segment. In a similar way, a left-handed helix can be seen as the result
of combining a circle drawn in the counterclockwise direction with a perpendicular
line segment. The next two Mathematica codes illustrate the mathematical ways of
forming right- and left-handed helices. They use the ParametricPlot3D
command, which produces a three-dimensional curve in which the x, y, and
z component functions depend on a single parameter (an external variable) that can
be interpreted as the time variable:
The following code uses the Mathematica command Manipulate to contin-

uously vary the parameter range, thus visualizing the formation of the helix in the
upward direction:
Suggestion: Write a Mathematica code to form right- and left-handed helices that
are mirror images of each other.
M4. Optical Rotation
This Mathematica code simulates a wave approaching the observer and gradually
rotating its plane of polarization as it passes through an optically active medium.
The wave is built by adding left- and right-circularly polarized waves with slightly
different phases:
Suggestion: Change the code so that the incoming wave rotates to the coun-
terclockwise direction.
Glossary
Cahn–Ingold–Prelog rules Set of rules named after organic chemists

S. Cahn (1899–1981), C.K. Ingold
(1893–1970), and V. Prelog (1906–1998) that
assign different priorities to the atoms or groups
bonded to a chiral center. These priorities
follow the decreasing order of the atomic
numbers, and the atom or group with lowest
priority points away from the observer. If the
decreasing priority of the atoms or groups

bonded to the chiral center and closer to the
observer is in the clockwise direction, the chiral
center is assigned an R descriptor, whereas for
the counterclockwise direction, the chiral
center is assigned an S descriptor
Conformers Isomers that can be interconverted by internal
rotation around a single bond
Electron probability density Probability of finding any of the n electrons of a
molecule in volume element dv, for any
position of the remaining electrons and any spin
of all electrons
Electrostatic potential Work done by an electric field to transport a test
charge (an infinitesimal positive charge) from
infinity to a specified point in a charge
distribution divided by the test charge. The
electrostatic potential for a molecule is usually
mapped over an isodensity surface with a
specified isovalue (for example, 0.00040) using
rainbow colors, where dark blue is used when
the electrostatic potential reaches the extreme
positive value (electrophilic region), and dark
red indicates the extreme negative value of the
electrostatic potential (nucleophilic region)
Enantiomers Chiral molecules that are mirror images of each
other
Intrinsic reaction coordinate Elementary reaction coordinate that starts at the
saddle point and progresses down the hill
toward the reactant channel (negative values of
the reaction coordinate) and toward the product
channel (positive values of the reaction
coordinate)
Molecular geometry optimization Calculation for determining the geometry of a
molecule that consists in systematically
varying the geometric variables (bond lengths,
bond angles, dihedral angles) until an energy
minimum is reached. This type of molecular
calculation includes two steps, the first leading
to a stationary point (a minimum or a
maximum), the second step being a frequency
calculation to verify whether the stationary
point corresponds to a minimum. If first
derivatives increase in the vicinity of the
stationary point (positive second-order
Glossary 283
derivatives), the stationary point corresponds to

an energy minimum, and then the frequency
calculation will not produce any imaginary
vibrational frequency (in Gaussian, imaginary
frequencies are presented as negative values)
Mulliken population Method developed by Mulliken for
analysis method apportioning the total electron charge of a
molecule to its atoms and overlap regions
Natural bond orbital method Population analysis method that apportions the
electron charge of a molecule to Lewis-like
electron pairs localized in chemical bonds and
atoms
Natural orbital Molecular orbital localized in a bond or an atom
with maximum electron occupation allowed by
the Pauli exclusion principle (number of
electrons = 2)
Optical activity Ability of a chiral molecule to rotate the plane
of polarization of light
Polarimeter Instrument for determining the effect of a
substance in rotating the plane of polarization
of light. A polarimeter essentially consists of a
light source, a polarizer, a sample tube, and a
polarization analyzer
Population analysis Multiple ways of apportioning the total electron
charge of a molecule to different regions of
space in the molecule, such as atoms and
overlap regions (Mulliken population analysis
method) or atoms and bonding regions (natural
bond orbitals method)
Potential energy surface A surface that gives the potential energy of a
system of three atoms A, B, and C as a function
of the internuclear distances A–B and B–C, with
the bond angle \ABC being kept constant.
A typical potential energy surface shows two
valleys (the reactant channel and the product
channel) approximately parallel to the
corresponding Cartesian coordinate axes and
an upper pass between them through one
first-order saddle point (see Fig. 5.14). The
minimum energy path converts reactants into
products, passes through the first-order saddle
point, and defines the reaction coordinate
Exercises
E1. Explain the expression for the electrostatic potential of a molecule.

E2. Consider a calculation for the H2 molecule with a minimal basis set and find
the coefficients of the 1s orbitals in terms of the overlap integral between these
orbitals.
E3. Consider the HF/STO-3G calculation on the CH4 molecule. Starting with
the coefficient matrix, use Mathematica to obtain the corresponding density
matrix.
E4. Consider the density matrix and the full Mulliken population matrix of the
Gaussian HF/STO-3G calculation on the CH4 molecule. Find the overlap
integral between the 1s2 and 1s3 hydrogen orbitals.
E5. Consider the full Mulliken population matrix of the Gaussian HF/STO-3G
calculation on the CH4 molecule. Find the populations condensed to the carbon
atom and to the overlap region between the carbon atom and hydrogen atom 1
(1s1 orbital).
E6. Derive the value for the tetrahedral angle.
E7. Write expressions for the sp3 hybrid orbitals centered on atom C of CH4 and
show that these hybrids are orthonormal.
E8. Use symmetry and Mathematica to write the expressions for the sp2 hybrid
orbitals centered on atom B of BF3 and show that these hybrids are orthonormal.
E9. Write expressions for the sp hybrid orbitals centered on atom C of CO2 and
show that these hybrids are orthonormal.
E10. Derive (5.29).
E11. The analysis of the vibrational spectra of 1,2-dichloroethane shows that the
population of the most abundant antiperiplanar conformer is about 80 % at
ambient temperature (El Youssoufi et al. 1998). Assuming that only conformers
1 and 2 exist, calculate the energy difference between these conformers at
ambient temperature.
E12. The molecular formula for tartaric acid is HOOC–CH(OH)–CH(OH)–
COOH. Identify and classify the stereoisomers for this molecule.
References
Mulliken RS (1955) Electronic population analysis on LCAO-MO molecular save functions I.
J Chem Phys 23:1833–1840
Baggott JE, Caldow GL, Mills IM (1988) J Chem Soc Faraday Trans 2(84):1407–1422
El Youssoufi Y, Herman M, Liévin J (1998) The ground electronic state of 1,2-dichloroethane I.
Ab initio investigation of the geometrical, vibrational and torsional structure. Molec Phys
94:461–472
References 285
Further Reading
Levine IN (2013) Quantum chemistry, 7th edn. Pearson Education Inc., Upper Saddle River
Weinhold F, Landis CR (2012) Discovering chemistry with natural bond orbitals. Wiley, Hoboken
Crystals
6
Abstract
We begin this chapter by describing modes for packing disks and spheres, with
particular attention to three-dimensional close-packed modes. Next, we consider
the concept of translation symmetry, central to crystallography, and present two-
and three-dimensional Bravais lattices. We show the distribution of the most
representative structures of metals in the periodic table of the elements and build
the cesium chloride, sodium chloride, diamond, and zinc blende structures, using
their corresponding translation vectors. X-ray diffraction is the most widely used
experimental method for solving the structure of crystals. For this reason, we
give a brief introduction to the subject and explain the optical transform of an
X-ray diffraction experiment. Next, we consider and illustrate the formation of
bands in solids, introduce semiconductors and the Fermi–Dirac distribution of
electrons for metallic conductors. At the end of this chapter, the student can find
several Mathematica codes (Packing of Disks, Hexagon of Disks, Disk Layers,
The Third Dimension, HCP Structure, 2D Square Point Lattice, 2D Hexagonal
Point Lattice, 3D Bravais Lattices, CsCl Structure, NaCl Structure, From Energy
Levels to Bands, Fermi–Dirac Distribution) that enable one to visualize the
crystalline structures considered in this chapter and include detailed explanations
of new commands, a glossary of important scientific terms, and a list of
exercises, whose answers can be found after the Appendix.
6.1 Packing Disks and Spheres
6.1.1 Disks
Consider the packing of identical disks in a plane. Assuming disks of radius 1/2, the
distance between centers of tangent disks is 1. In square packing, the (x,y) coordinates
DOI 10.1007/978-3-319-41093-7_6
288 6 Crystals
of the disks’ centers describe a square (Fig. 6.1). For hexagonal close packing, the
evaluation of coordinates of the disks’ centers requires simple trigonometry of right
angle triangles, as shown in Fig. 6.1.
Figure 6.2 shows square and hexagonal packing modes of identical disks. The
Mathematica code M1 builds the disk lattices of this figure. In square packing, each
disk is tangent to four neighboring disks, whereas in hexagonal packing, each disk
is tangent to six neighboring disks (Fig. 6.2) whose centers are the vertices of a
regular hexagon. This arrangement reveals the existence of a sixfold rotational
symmetry axis. The Mathematica code M2 illustrates this point and generates the
picture shown in Fig. 6.3.
As shown in Fig. 6.4, the centers of three close-packed disks are the vertices of
an equilateral triangle, whose center is the center of the interstice that can be
determined making use of simple trigonometry for a right triangle.
The Mathematica code M3 considers two close-packed layers of identical disks,
where the second layer is displaced so that its disks’ centers coincide with the
Fig. 6.1 Coordinates of centers of disks of radius 1/2, in square and hexagonal packing modes.
Fig. 6.2 Square and hexagonal packing modes of identical disks. Figure obtained with
Mathematica
6.1 Packing Disks and Spheres 289
Fig. 6.3 Disks exhibiting

sixfold rotational axis for
close-packed identical disks
on the plane. Figure obtained
with Mathematica
Fig. 6.4 (x, y) coordinates

for the center of the interstice
of three close-packed
identical disks in a plane.
Disks drawn with
Mathematica
interstices of the first layer (Fig. 6.5). The translation vector that shifts the second
layer of disks with respect to the first layer is shown in Fig. 6.4. The coordinates of
its origin are (1, √3/2), and the coordinates of its tip at the center of the interstice are
(3/2, 2/√3). Hence, the translation vector is given by
pffiffiffi
1 3
t ¼ ^x þ ^y ð6:1Þ
2 6
6.1.2 Spheres
We begin by considering four equivalent and mutually tangent spheres of radius 1/2
whose centers form a regular tetrahedron that can be drawn inside a cube (Fig. 6.6).
The distance between center 4 and the 1-2-3 plane is given by √2/√3, as can be
290 6 Crystals
Fig. 6.5 Close-packed disk layer and a second close-packed layer of disks in gray displaced so
that its disks’ centers coincide with the interstices of the first layer. Figure obtained with
Mathematica
Fig. 6.6 Right triangle trigonometry determines the distance between center 4 and the 1-2-3 plane
shown by simple trigonometry of right triangles (Fig. 6.6). Sphere 4 is nested on the
interstice of spheres 1, 2, and 3. Therefore, √2/√3 is the vertical distance between the
centers of two horizontal close-packed layers of spheres of radius 1/2. The vector
that goes from center 1 to center 4 [= vector (4-c) + vector (c-1), where vector (4-c)
is obtained from Fig. 6.6 and vector (c-1) is given by (6.1)] is
pffiffiffi rffiffiffi
1 3 2
t ¼ ^x þ ^y þ ^z ð6:2Þ
2 6 3
(E1).
Figure 6.7 shows a regular tetrahedron formed by the centers of four mutually
tangent spheres. The Mathematica code M4 uses the threefold rotation symmetry
axis to draw Fig. 6.7 (E2).
Fig. 6.7 Regular tetrahedron formed by the centers of four mutually tangent spheres.
6.1.3 Hexagonal Close Packing
The close packing of identical spheres on one layer poses no great difficulty: each
sphere is tangent to six neighboring spheres, whose centers form a regular hexagon.
If projected on the plane of the paper, the spheres are seen as tangent disks with
interstices appearing as triangles of curved sides, half of them with an upside
vertex, the other half with a downside vertex (Fig. 6.8). The second layer of spheres
nests on interstices with downside vertices, since these correspond to depressions
on the upper surface of the layer of spheres. Unlike the first layer, where the
downside and upside curved triangles are equivalent sites (rotation of the whole
structure by 180° transforms downside curved triangles into upside curved triangles
and vice versa), the second layer has two kinds of interstices: those that pass
through both layers and those that encounter spheres of the first layer (Fig. 6.8).
Given the option to lay down the third layer, we nest the spheres of this layer on
the second layer interstices that encounter first layer spheres, keeping open the
interstices that pass through all layers. Then the third and first layers coincide on
projection (Fig. 6.8). If successive layers are labeled by capital letters with the first
layer being labeled A, then the second layer is B and the third layer is A again (layer
sequence ABAB… with spheres of alternating layers overlaying one another). This
structure is called hexagonal close-packed (hcp), with the representative polyhe-
dron being a hexagonal prism (Fig. 6.8). The Mathematica code M5 builds
Fig. 6.8.
Fig. 6.8 Hexagonal close packing of rigid spheres with equal radius. Figure obtained with
Mathematica
292 6 Crystals
The number of spheres tangent to any one sphere (number of nearest neighbors)
is called coordination number. The coordination number of the hcp structure is 12.
To obtain this number, we consider layer A and two additional B layers, one above,
the other below, to form the BAB sequence. The central sphere of layer A is tangent
to six spheres on the same equatorial layer and to three spheres on each of the B
layers, above and below the central sphere of layer A.
6.1.4 Cubic Close Packing
We now go back to the first two layers of close-packed spheres and nest the sphere
of the third layer on the interstice that passes through B and A (Fig. 6.9). The
resulting sequence of layers ABCABC… leads to the cubic close-packed (ccp)
structure, which is also a face-centered cubic (fcc) structure, since the represen-
tative polyhedron is a face-centered cube. Layers A and B correspond to planes that
pass through face diagonals of three cube faces. The central sphere of the cube top
face is tangent to four spheres on this face, four spheres on the layer below, and four
spheres on the layer above (the latter is not shown in Fig. 6.9) thus having the
coordination number 12.
6.1.5 Packing Densities
Spheres cannot occupy 100 % of the volume of a crystal repeating unit. The effi-
ciency of occupation of the three-dimensional space by rigid spheres can be
measured by the packing density, which is a ratio of two volumes, the volume of
all spheres contained in the repeating unit (a parallelepiped; note that the cube is a
particular kind of parallelepiped) over the volume of the repeating unit. In order to
evaluate the packing density, one should take into account the contribution of each
sphere to the repeating unit of the crystal, which depends on the position of the
sphere in the parallelepiped of the repeating unit. Each sphere located on a paral-
lelepiped vertex is shared by eight repeating units (Fig. 6.10). Thus, each repeating
unit contains 1/8 of one sphere. If the sphere occupies the midpoint of one edge,
then each parallelepiped contains 1/4 of the sphere. In turn, if the sphere occupies
the center of a face, it is shared by two parallelepipeds, so 1/2 belongs to each
parallelepiped.
Fig. 6.9 Cubic close packing of spheres with equal radius. Figure obtained with Mathematica
Fig. 6.10 Spheres occupying a vertex, the midpoint of an edge, and the center of a face contribute
to each crystal repeating unit with 1/8, 1/4, and 1/2 of a sphere, respectively
If one layer of the hcp structure (say layer A) is displaced to nest over yet
uncovered interstices of the preceding layer, then it transforms itself into layer C of
the ccp structure, and the ABA sequence of the original hcp structure is converted
into the ABC sequence of the ccp structure. This reasoning suggests that the hcp
and ccp structures have the same coordination number (6.12) and the same packing
density, since they are both close-packed structures differing by displaced layers of
spheres.
In 1611, the astronomer and mathematician Johannes Kepler (1571–1630)
conjectured that it would be impossible to exceed the packing density of the cubic
close-packed structure. In the nineteenth century, Karl Friedrich Gauss (1777–
1855) showed that the hexagonal close-packed structure reached the maximum
packing density for a lattice of identical spheres. Finally, in 1998, Thomas Hales
proved Kepler’s conjecture, thus concluding that both the hexagonal and the cubic
close-packed structures reach the maximum packing density for a regular structure
of identical spheres.
In the cubic close-packed (ccp) structure, the central sphere of a cube face is
tangent to the spheres on the vertices (Fig. 6.11). Therefore, we can write
Fig. 6.11 In the cubic close-packed structure, spheres are tangent along the face diagonals.
Arrangement of spheres obtained with Mathematica
294 6 Crystals
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi
4r = a2 þ a2 ) a ¼ 2 2r ð6:3Þ
where a is the edge and r is the sphere’s radius. The packing density is given by
4
8 18 þ 6 12 3 pr
3
4 4 pr 3 p
¼ p3ffiffiffi 3 ¼ pffiffiffi 0:74 ð6:4Þ
a3 2 2r 3 2
(E3).
The body-centered cubic (bcc) structure has a coordination number equal to 8,
since the central sphere is tangent to eight spheres on the vertices of the cube
(Fig. 6.12). The right triangle formed by one cube main diagonal, one cube face
diagonal, and one cube edge allows one to determine the cube edge a as a function
of the sphere’s radius r:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi
4
4r ¼ a2 þ dface 2 ¼ a2 þ 2 a2 ¼ 3 a ) a ¼ pffiffiffi r ð6:5Þ
3
Thus, the packing density is given by

4 pffiffiffi
8 18 þ 1 3 pr
3
2 4 pr 3 3p
¼ 3 3 ¼ 0:68 ð6:6Þ
a3 p4ffiffi r
8
3
This packing density is smaller than those of the hcp and ccp structures: the bcc
structure is not a close-packed structure.
The simple cubic (sc) structure presents spheres on the cube vertices (Fig. 6.13),
and each sphere is tangent to six spheres on the vertices of a regular octahedron.
The packing density is given by

8 18 43 pr 3 43 pr 3 p
¼ ¼ 0:52 ð6:7Þ
a3 ð2rÞ3 6
Fig. 6.12 In the body-centered cubic structure, spheres are tangent along the cube’s main
diagonals. Arrangement of spheres obtained with Mathematica
Fig. 6.13 Spheres are tangent along the cube edges of the simple cubic structure. Figure obtained
with Mathematica
The coordination numbers and packing densities of the above-mentioned

structures of identical spheres are summarized in the following table:
Structure Coordination number Packing density

Hexagonal close-packed (hcp) 12 0.74
Cubic close-packed (ccp) 12 0.74
Body-centered cubic (bcc) 8 0.68
Simple cubic (sc) 6 0.52
6.1.6 Occupying Interstices
The interstices of close-packed structures of identical spheres are of two types:

octahedral, occupying the center of a regular octahedron whose vertices are the
centers of six spheres, and tetrahedral, occupying the center of a regular tetrahedron
whose vertices are the centers of four spheres (Fig. 6.14).
Imagine the octahedral interstice occupied by a small sphere tangent to the
surrounding larger spheres. The ratio of the smaller radius and the larger radius can
be easily obtained from Fig. 6.15 and is given by
Fig. 6.14 Close-packed layers of identical spheres have octahedral and tetrahedral interstices.
296 6 Crystals
Fig. 6.15 Octahedral interstice occupied by one sphere. The diameter of each larger sphere is
arbitrarily equal to 1. Figure obtained with Mathematica
pffiffiffi
rs 2rs 21
¼ ¼ 0:414 ð6:8Þ
r 2r 1
where rs is the radius of the smaller sphere that occupies the octahedral interstice.
Consider now the tetrahedral interstice occupied by a sphere tangent to its
neighboring spheres. From the right triangle of Fig. 6.16, we can write
1
1 ¼ ð1 þ rs Þ sin ða = 2Þ ) rs ¼ 1 ð6:9Þ
sin ða = 2Þ
and consequently,
rs sin ða = 2Þ 1
1
1
¼ ¼ 1 0:225 ð6:10Þ
r 1 sin ½arccos ð1 = 3Þ = 2
where a is the tetrahedral angle, and the radii of the larger spheres are set arbitrarily
equal to 1.
Let us now consider the simple cubic structure and the occupation of the cubic
interstice by a sphere tangent to its eight neighboring spheres. The ratio of the
radius of the smaller sphere and the radius of the larger sphere can be easily
obtained from Fig. 6.17 and is given by
pffiffiffi
rs 2rs 31
¼ ¼ 0:732 ð6:11Þ
r 2r 1
Fig. 6.16 Tetrahedral interstice occupied by one sphere. The marked angle is half of the
tetrahedral angle. The radii of the larger spheres are set arbitrarily equal to 1. Arrangement of
spheres obtained with Mathematica
Fig. 6.17 Simple cubic interstice occupied by one sphere. The diameter of each larger sphere is
arbitrarily equal to 1. Arrangement of spheres obtained with Mathematica
The values of the ratios rs/r and the corresponding coordination numbers are
collected in the following table:
rs/r Environment Coordination number

0.225 − (0.414) Tetrahedral 4
0.414 − (0.732) Octahedral 6
0.732 − 1.000 Cubic 8
The lower value of each rs/r range has been determined assuming that the
spheres of larger radii preserve their initial structure and the sphere of smaller radius
is tangent to its surrounding spheres. In each rs/r range, values greater than the
lower value force the larger spheres to drift away from each other, no longer being
tangent. At each upper limit in parentheses, the structure abruptly changes, and its
coordination number increases.
6.2 Translation Symmetries
In this section, we consider the concept of translation symmetry, central to crys-

tallography. We begin with a discrete two-dimensional array of equivalent points
that describes the periodicity of the original crystal structure, regardless of the
physical content (spheres, atoms, molecules) ascribed to its points, since these are
devoid of any physical content and simply emphasize the translation symmetry.
6.2.1 2D Bravais Lattices
Figure 6.18 shows a finite section of an infinite discrete array of points with a
square arrangement. For each point P of this array, it is possible to find a pair of
integers (n1, n2) (positive or negative integers and zero included) such that the
positional vector of a point P is given by
298 6 Crystals
Fig. 6.18 Square point lattice and its p/4 counterclockwise rotation. Figure obtained with
Mathematica
p ¼ n1 v1 þ n2 v2 ð6:12Þ
where v1 and v2 are linearly independent vectors (noncollinear vectors). Each point
on the lattice corresponds to a defined pair of integers (n1, n2). For instance,
P1 = v1 + 2v2 and P2 = −2v1 − v2 correspond to (n1, n2) equal to (1, 2) and (−2,
−1), respectively. The vector Eq. (6.12) generates every point of the 2D square
array. Therefore, all points are equivalent, since every point has the same
arrangement of neighbors. In the case of Fig. 6.18, v1 and v2 are orthogonal and are
given by
v1 ¼ a ^x v2 ¼ a ^
y ð6:13Þ
where the scalar a is the lattice constant, vectors v1 and v2 are called primitive
translation vectors, and the array of points is a two-dimensional Bravais square
lattice. All points that satisfy the following equality cover the gray area:
r ¼ x 1 v1 þ x 2 v2 0 xi 1 ð6:14Þ
When this area is translated by all vectors p, the two-dimensional space is filled
without overlapping itself or leaving voids, and such an area is called a
two-dimensional unit cell. The Mathematica code M6 uses (6.12) and (6.13) to
build a square point lattice like the one of Fig. 6.18 and applies a p/4 counter-
clockwise rotation to the resulting square point lattice.
Both arrays of points of Fig. 6.18 represent the same lattice, since the translation
vectors w1 and w2 result from v1 and v2 by a p/4 counterclockwise rotation,
w1 ¼ Rp=4 v1 w2 ¼ Rp=4 v2 ð6:15Þ

6.2 Translation Symmetries 299
where Rp/4 is the rotation matrix

cos h sin h
Rh ¼ ð6:16Þ
sin h cos h
for h = p/4,

1 0
v1 = v2 = ð6:17Þ
0 1
and

1 1 1 1
w1 = pffiffiffi w2 = pffiffiffi ð6:18Þ
2 1 2 1
From (6.17) and (6.18), we obtain
1 1
w1 = pffiffiffi ðv1 þ v2 Þ w2 = pffiffiffi ðv2 v1 Þ ð6:19Þ
2 2
where 1/√2 is the normalizing factor (v1 and v2 are normalized vectors).
The choice of primitive translation vectors is not unique. For instance, we can
take v1 and v1 + v2 (Fig. 6.19), since these are linearly independent vectors.
However, v1 and v1 + v2 are not orthogonal, do not have the same length, and so do
not reflect the symmetry of the square lattice. Therefore, they are not a convenient
choice of primitive vectors for a square array.
Consider now a two-dimensional lattice where each point is surrounded by six
points on the vertices of a regular hexagon (Fig. 6.20). The translation vectors of
this hexagonal lattice are given by
Fig. 6.19 The translation

vectors v1, v1 + v2 do not
reflect the symmetry of a
simple square lattice.
Mathematica
300 6 Crystals
Fig. 6.20 Two-dimensional

hexagonal Bravais lattice.
Mathematica
pffiffiffi
1 3
v1 ¼ a ^x v2 ¼ a ^x þ y^ ð6:20Þ
2 2
where a is the lattice constant and these vectors correspond to the sides of a unit-cell
parallelogram. The Mathematica code M7 uses (6.12) and (6.20) to build a
hexagonal point lattice like the one of Fig. 6.20 (E4).
In two-dimensional space, there are five Bravais lattices, namely square, rect-
angular, oblique, hexagonal, and centered rectangular lattices (Fig. 6.21).
The black and gray points of Fig. 6.22 correspond to motifs of a distinct nature,
where each black point has four gray points as closest neighbors, and conversely,
each gray point has four black points as closest neighbors. This structure cannot be
described by a single set of primitive translation vectors, since the points marked
black and gray cannot simultaneously correspond to integer values of n1 and n2 in
(6.12). The dark gray vector is given by
v1 þ v 2
t¼ ð6:21Þ
2
When this vector is applied to every point of the black lattice, it generates the
gray lattice and vice versa. One says that the periodic structure of Fig. 6.22 results
from merging two square point lattices. The resulting structure can be considered a
two-dimensional equivalent of the cesium chloride structure. The tri-dimensional
Fig. 6.21 Bravais lattices in two-dimensional space. Lattices and translation vectors obtained
with Mathematica
Fig. 6.22 Two square lattices merge with one another. When the orange translation vector is
applied to every point on the black point lattice, the gray lattice is obtained. Figure obtained with
Mathematica
CsCl structure results from merging two simple cubic lattices, one for cesium ions,
the other for chloride ions (cesium and chloride ions have approximately the same
radius).
6.2.2 3D Bravais Lattices
A three-dimensional Bravais lattice is an infinite array of discrete points such that

for every point P of the array defined by the positional vector p, it is possible to find
a set of positive or negative integers (zero included) (n1, n2, n3) that satisfy the
following vector equation:
p ¼ n1 v1 + n2 v2 þ n3 v3 ð6:22Þ
where v1, v2, and v3 are linearly independent translation vectors (noncollinear
vectors) that span the lattice. The lengths of cell edges (a, b, c) and the angles
between them (a, b, c) characterize the lattice systems.
In three-dimensional space, there are 7 lattice systems and 14 Bravais lattices.
The lattice systems are cubic (a, a, a; a = b = c = 90°), tetragonal (a, a, c; a =
b = c = 90°), orthorhombic (a, b, c; a = b = c = 90°), monoclinic (a, b, c;
a = c = 90°, b 6¼ 90°), triclinic (a, b, c; a, b, c 6¼ 90°), trigonal (a, a, a; a, b,
c 6¼ 90°), and hexagonal (a, a, c; a = b = 90°, c = 120°). The Bravais lattices are
simple cubic, body-centered cubic, face-centered cubic, simple tetragonal,
body-centered tetragonal, simple orthorhombic, base-centered orthorhombic,
body-centered orthorhombic, face-centered orthorhombic, simple monoclinic,
base-centered monoclinic, simple triclinic, simple trigonal, and simple hexagonal
(see Mathematica code M8). To extend group theory considerations from the
molecule (point groups; see Chap. 4) to the crystalline state (space groups),
translations need to be considered.
302 6 Crystals
We now illustrate the translation vectors v1, v2, and v3 for a few crystal struc-
tures. Starting with the simple cubic structure, its translation vectors are given by
v1 ¼ a ^x v2 ¼ a ^y v3 ¼ a ^z ð6:23Þ
as suggested by the symmetry of the structure (Fig. 6.23, E5). Polonium provides a
rare example of the simple cubic structure.
We now consider the body centered cubic structure, whose translation vectors
are given by
a
v1 ¼ a ^x v2 ¼ a ^y v3 ¼ ð^x þ ^y þ ^zÞ ð6:24Þ
2
where v3 is along the cube main diagonal (Fig. 6.24, E6), and the projection of v3
along, ^
x; ^
y, and ^z has length a/2. The alkali metal cesium is an example of the
body-centered cubic structure.
The translation vectors for the face-centered cubic structure are given by
a a a
v1 = (^y þ ^z) v2 ¼ (^x þ ^z) v3 ¼ (^x þ ^y) ð6:25Þ
2 2 2
Fig. 6.23 Translation vectors for the simple cubic structure. Figure obtained with Mathematica
Fig. 6.24 Translation vectors for the body-centered cubic structure. Figure obtained with
Mathematica
(Figure 6.25, E7). Aluminum gives an example of the face-centered cubic

structure.
The hexagonal close-packed structure is formed by two Bravais lattices, one
spanned by the translation vectors
pffiffiffi
1 3
v1 ¼ a ^x v2 ¼ a ^x þ y^ v3 ¼ c ^z ð6:26Þ
2 2
the other obtained by translation by the vector

v1 v2 v3
t¼ + þ ð6:27Þ
3 3 2
(Figure 6.26, E8). Zinc is an example of the hexagonal close-packed structure.
Fig. 6.25 Translation vectors for the face-centered cubic structure. Figure obtained with
Mathematica
Fig. 6.26 Translation

vectors for the hexagonal
close-packed structure.
Mathematica
304 6 Crystals
6.3 Crystal Structures
6.3.1 Metals
The hexagonal close-packed (hcp), cubic close-packed (ccp) and body-centered

cubic (bcc) structures are the most representative structures for the metals of periods
2–6 and groups 1–14 of the periodic table (Fig. 6.27). Mercury, gallium, indium,
and tin are not among the metals with these structures. On the whole, there is a
tendency for metals of the same group to share the same crystalline structure, as
illustrated, for the bcc structure by groups 1 (IA), 5 (VB), and 6 (VIB), for the hcp
structure by groups 3 (IIIB), 4 (IVB), and 12 (IIB), and for the ccp structure by
groups 10 (VIIIB) and 11 (IB).
6.3.2 Lattice Energy
Consider now an ionic crystal of general formula XmYn. The energy of formation of
one mole of XmYn from infinitely separated ions,
mXn þ ðgÞ þ nYm ðgÞ ! Xm Yn ðsÞ ð6:28Þ
is called the lattice energy and is invariably negative (energy released by the
system). The lattice energy (electrostatic energy of attraction and repulsion between
ions in one mole of XmYn) plus the energy term associated with the volume change
at constant pressure (the ions Xn+ and Ym− are in the gaseous state) is called the
lattice enthalpy. This energy is usually determined by the Born–Haber cycle, a
multistep thermochemical cycle in which the enthalpy of one of the steps is the
lattice enthalpy due to German scientists, Max Born (1882–1950; Nobel Prize in
physics in 1954 for his research achievements on the statistical interpretation of
quantum mechanics) and Fritz Haber (1868–1934; Nobel prize in chemistry in
1918, for his research on the synthesis of ammonia). When the lattice energy is not
bcc ccp hcp

Li Be
Na Mg Al
K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga
Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn
Cs Ba Lu Hf Ta W Re Os Ir Pt Au Hg Tl Pb
Fig. 6.27 Distribution of body-centered cubic, cubic close-packed, and hexagonal close-packed
structures on the periodic table of elements
6.3 Crystal Structures 305
tabulated, the Kapustinskii equation can be used to obtain a calculated value

(Kapustinskii 1956).
6.3.3 Cesium Chloride and Sodium Chloride
Cesium chloride is formed by equal amounts of cesium and chloride ions. Its crystal
structure results from merging two simple cubic lattices, one for the Cs+ ions, the
other for the Cl− ions. Its lattice energy is about −670 kJ mol−1 (Handbook of
Chemistry and Physics, 2011). If one Cl− is at the origin, the translation vectors for
the chloride ion lattice are given by (6.23), and the vector that translates this lattice
to produce the cesium ion lattice is given by
a
t ¼ (^x þ ^y þ ^zÞ ð6:29Þ
2
whose magnitude is equal to half of the main diagonal of a cube with chloride ions
at the vertices. The Mathematica code M9 builds the cesium chloride structure with
the Cs+ and Cl− ions being represented by spheres (Fig. 6.28).
Sodium chloride is formed by equal amounts of sodium and chloride ions. Its
crystal structure results from merging two face-centered cubic lattices, one for Na+
ions, the other for Cl− ions (Fig. 6.29). Its lattice energy is about −767 kJ mol−1
(Handbook of Chemistry and Physics, 2011). If one Cl− is at the origin, the
translation vectors for the chloride ion lattice are given by (6.25), and the vector that
translates this lattice to generate the sodium lattice is given by
a
t ¼ ^z ð6:30Þ
2
The Mathematica code M10 builds the sodium chloride structure, with the Na+
and Cl− ions represented by spheres (Fig. 6.29).
Fig. 6.28 The cesium chloride structure with the Cs+ and Cl− ions represented by spheres (Cs+
light gray spheres, Cl− gray spheres). Figure obtained with Mathematica
306 6 Crystals
Fig. 6.29 The sodium chloride structure with the Na+ and Cl− ions represented by spheres (Na+
light gray spheres, Cl− gray spheres). Figure obtained with Mathematica
6.3.4 Diamond and Zinc Blende
In diamond, each carbon atom is covalently bonded to four carbon atoms on the
vertices of regular tetrahedra. The diamond lattice consists of two face-centered
cubic lattices that interpenetrate each other. The translation vectors for one of these
lattices is given by
n1 v1 þ n2 v2 þ n3 v3 ð6:31Þ
where v1, v2, and v3 are given by (6.25), and (n1, n2, n3) are integers (which can be
positive, negative, or zero). The translation vectors for the other lattice of diamond
is given by
n1 v1 þ n2 v2 þ n3 v3 þ t ð6:32Þ
where t is given by
a
t ¼ (^x þ ^y þ ^z) ð6:33Þ
4
The magnitude of t is one-fourth that of the cube’s main diagonal. For the sake
of clarity, the spheres that schematically represent carbon atoms in Fig. 6.30 have a
radius much smaller than the covalent radius of the carbon atom, and darker and
lighter spheres represent carbon atoms belonging to distinct lattices.
Zinc blende and sphalerite are names for a mineral whose composition is mostly
zinc sulfide with a variable percentage of zinc being replaced by iron (II). The iron
presence gives a dark and glossy appearance to zinc blende, similar to that of
galena, an iron (II) sulfide. The likely confusion of zinc blende with galena is at the
origin of the word sphalerite, from the Greek word sphaleron, meaning misleading.
The word blende gives the same idea, but the origin of blende is German (blen-
den = to blind, also to deceive). In this text, zinc blende is used to refer to one of the
6.3 Crystal Structures 307
Fig. 6.30 Diamond structure

with an illustration of the
vector t. Figure obtained with
Mathematica
crystalline forms of zinc sulfide, ZnS, consisting of two merged face-centered cubic
lattices, one for the sulfur atoms, the other for the zinc atoms.
Figure 6.31 shows two distinct main cubes that schematically represent the same
zinc blende crystal. The cube on the right results from the one on the left by shifting
the latter by the vector t. This vector is directed along the diagonal of the cube, and
its length is one-fourth that of the main diagonal [see (6.33) and Fig. 6.31]. The
translation of the atoms of the main cube on the left by the vector t leads to the zinc
blende structure on the right of Fig. 6.31. One of these main cubes (the one on the
left) shows sulfur atoms represented by light gray spheres on the vertices and face
centers and four zinc atoms (gray spheres) totally contained in the repeating unit. In
turn, the cube on the right shows zinc atoms on the vertices and face centers and
four sulfur atoms totally contained in the repeating unit.
Fig. 6.31 Zinc blende structure with an illustration of the vector t in the left-hand drawing:
light-gray spheres represent sulfur atoms, and gray spheres represent zinc atoms. Left- and
right-hand drawings represent two main cubes of the same crystal. Figure obtained with
Mathematica
308 6 Crystals
If we divide each main cube into eight smaller and equal cubes, we conclude that
only four of these are occupied by ZnS4 tetrahedra on the left of Fig. 6.31, and by
SZn4 tetrahedra on the right of this figure. On the whole, each of the main cubes
corresponds to the formula Zn4S4, since there are 8 (1/8) + 6 (1/2) = 4 atoms
of one of the elements (sulfur or zinc) on the vertices and centers of the faces and
four atoms of the other element totally (zinc or sulfur) included in the repeating
unit. If carbon atoms substitute zinc and sulfur atoms and the lattice is correctly
adjusted to make all bond lengths equal, one obtains the diamond structure.
6.4 X-Ray Diffraction
Atoms and monatomic ions have dimensions of order 1 Å (= 10−10 m), and elec-
tromagnetic radiation with wavelength of this order of magnitude consists of
X-rays. In 1912, Max von Laue (1879–1960, Nobel Prize in physics 1914) showed
that crystals diffract X-rays and at the same time, William Henry Bragg and William
Lawrence Bragg, father and son, solved several crystalline structures using X-ray
diffraction and shared the Nobel Prize in physics in 1915. One of the first structures
to be determined was that of sodium chloride. Nowadays, X-ray diffraction is the
most widely used method for solving the structure of crystals, in particular that of
macromolecules of biological importance such as proteins.
After being irradiated with X-ray waves, each atom acts like a secondary source
of X-rays. When an X-ray beam with defined wavelength impinges on a crystal,
multiple X-ray beams with the same wavelength emerge that can leave their imprint
(diffraction pattern) on an X-ray sensitive support (Fig. 6.32). As found by W.H.
and W.L. Bragg, diffracted X-rays beams result from reflection of X-rays by parallel
planes of the same atoms or ions in a crystal (Fig. 6.33). The condition for con-
structive interference requires that the path difference between X-ray waves
reflected by successive planes of atoms be a multiple of the wavelength,
2d sin h = nk ð6:34Þ
where d is the spacing between adjacent planes of atoms, h the angle of incidence,
n the diffraction order, and k the X-ray wavelength.
Fig. 6.32 X-ray diffraction

experiment (schematic)
6.4 X-Ray Diffraction 309
Fig. 6.33 Bragg diffraction
Each set of parallel planes of the same atoms or ions in the crystal acts like one
three-dimensional diffraction grating. The diffracted X-ray intensity is mathemati-
cally related to the electronic density of the atoms by Fourier transforms. Solving a
diffraction pattern consists in determining the crystalline structure that leads to the
obtained diffraction pattern. This is a complex and time-consuming task that
requires the expertise of an X-ray crystallographer, sophisticated software, and
adequate computational resources.
In order to clarify the relationship between the diffraction pattern and the
crystalline structure on purely qualitative and empirical grounds, a change of scale
can be applied to an X-ray diffraction experiment, with the X-ray beam being
replaced by a laser beam in the visible, and the 3D crystal being substituted by
ordered patterns printed on 2D projection slides. Experiments like this are called
optical transforms of the X-ray diffraction experiment (Lisensky et al. 1991).
Optical transforms correspond to a change of scale by a factor of order 104 on
the X-rays’ wavelength (kX-ray 10−10 m, klaser 10−6 m) and a factor of order
106 on the motif dimensions (an atom diameter 10−10 m, a motif on the slide
10−4 m). The diffraction pattern of an optical transform is obtained on a screen
about 20 m from the slide, which is about 20 cm from the laser. For the slide
perpendicular to the incident laser beam, the condition for constructive interference
(Fraunhofer diffraction, Fig. 6.34) is given by
d sin h = nk ð6:35Þ
Fig. 6.34 Fraunhofer

diffraction
310 6 Crystals
At such a distance between slide and screen, the diffraction pattern is about 1°
wide (1° = 0.0175 radians) and sinh h. Considering first-order diffraction
(n = 1), (6.35) can be replaced by
dh k ð6:36Þ
in the optical transform. This expression allows one to correlate changes in the
ordered pattern on the slide (changes in d) with changes produced in the optical
transform (changes in h). For the same k, d and h are inversely related.
The method of optical transforms has been used to show that B-DNA structural
parameters such as the spacing between base pairs and the radius of the vertical
section of the backbone structure can be obtained from the optical transform of the
B-DNA X-ray diffraction pattern (Lucas et al. 1999).
6.5 Electrons in Solids
Electrons in atoms and molecules have discrete energies and are said to occupy
defined energy levels, since such can be confirmed by spectroscopic methods.
However, in a solid, the proximity between an atom and its neighbors leads to
significant overlap between electronic wave functions and to electron delocaliza-
tion, resulting in large numbers of energy levels with closely spaced energies,
which in practical terms, form a continuum of energy levels, or band.
Consider a planar hydrocarbon with CC bonds of uniform bond distance and
uniform bond order (1.5) and use the Hückel empirical method to calculate the
energy levels of the system of p electrons as the number of carbon atoms increases
(McWeeny 1979). In this model system, each carbon atom has one electron in one
p orbital perpendicular to the backbone of carbon atoms. The overlap between
adjacent p orbitals leads to p electron delocalization over the entire carbon back-
bone (Fig. 6.35).
In its original form, Hückel’s method was conceived to deal with delocalized p
electron systems and uses two parameters, a and b, which represent the ionization
energy of a carbon atom and the interaction energy between pp orbitals of adjacent
Fig. 6.35 The p electron system of a linear polyene hydrocarbon can inspire a model for a
unidimensional crystal. Figure obtained with Mathematica
6.5 Electrons in Solids 311
carbon atoms, respectively. The parameters a and b are negative quantities. For a
linear polyene, the k molecular orbital energy calculated by Hückel’s method is
given by
p
ek = a + 2b cos k ð6:37Þ
N þ1
where N is the number of carbon atoms. For k = 1 (the lowest energy level) and
N ! ∞,
lim e1 ¼ a þ 2b ð6:38Þ
N!1
In turn, for k = N (the highest energy level) and N ! ∞,
lim eN ¼ a 2b ð6:39Þ
N!1
Making use of (6.37), we see that the energy difference between consecutive
levels is given by

pðk þ 1Þ pk
De = ek þ 1 ek ¼ 2b cos cos
N þ1 N þ1
p ð2k þ 1Þp
¼ 4b sin sin ð6:40Þ
2ðN þ 1Þ 2ðN þ 1Þ
(E9).
With a linear polyene with N ! ∞ taken as a model for a unidimensional
crystal, the energy levels are limited by the lowest and highest energies [(6.38) and
(6.39)] that do not depend on k. In addition, from (6.40), it can be concluded that
the energy difference between consecutive levels decreases as N increases. There-
fore, for N ! ∞, the energy difference (6.40) tends to zero, and the energy levels
are replaced by a continuum of levels called a band (Fig. 6.36). The Mathematica
code M11 draws this figure.
Fig. 6.36 Energy levels for the p electron delocalized system of a linear polyene, calculated by
Hückel’s method, as the number of carbon atoms increases from 10 to 100, in steps of 10.
312 6 Crystals
6.6 Semiconductors
Experimental information on the electronic structure of solids can be obtained from

values of electrical and thermal properties. In particular, the electrical resistivity and
its thermal behavior can lead to a general classification of materials and details of
their electronic structure. Electrical resistance is defined by R V/I, where V and
I represent an electric potential difference and an electric current intensity,
respectively. The variable R is an extensive variable. For a cylindrical material of
length L and uniform sectional area S, one has R = q L/S, where q is the electrical
resistivity (q is an intensive variable and its inverse is the electrical conductivity
r = 1/q) and the thermal coefficient of electrical resistivity is given by
1 dq
a¼ ð6:41Þ
q dT
with units K−1 or °C−1. Typical values for the resistivity of metals are of order of
magnitude 10−8 X m, and they increase with temperature. In particular, the resis-
tivities of iron, beryllium, gold, and copper increase almost linearly with temper-
ature, except at very low temperatures, where the increase is exponential
(Fig. 6.37). Solids with resistivity of this order of magnitude that increase with
temperature are called metallic conductors. With high values of electrical resis-
tivity, we find materials called electrical insulators, such as glasses with electrical
resistivities 1010–1014 X m and polytetrafluoroethylene (PTFE, trademark Teflon)
with resistivities 1022–1024 X m. Materials with intermediate values of electrical
resistivities (10−5–10 X m) are called semiconductors. Semiconductors have the
Fig. 6.37 Resistivity as a function of temperature for iron, beryllium, gold, and copper. Data
values taken from Handbook of Chemistry and Physics, 2011 and figure obtained with
Mathematica
6.6 Semiconductors 313
electric current transported by electrons, as occurs with metals, but unlike metals,
they exhibit an electrical resistivity that decreases exponentially with temperature
when they are extremely pure, and the resistivity may even decrease by several
orders of magnitude when they are doped with trace quantities of an electrical active
material.
The valence band of insulators is fully occupied with valence electrons, and the
band gap between the valence band and the lowest energy unoccupied band is far
greater than the energy of the medium (energy associated with the thermal motion
of atoms, sometimes referred to as thermal energy as its name suggests, it is zero at
T = 0 K). Thus, electrons of the highest energy in the valence band, that is, at the
band top, do not have unoccupied neighboring levels available to be excited.
Materials with this band structure are insulators.
Unlike the valence band of electrical insulators, the valence band of metallic
conductors is not fully occupied, and a temperature increase causes more collisions
between excited electrons and atoms, thus leading to an increased resistivity. The
mean number of electrons in each energy level is given by the Fermi–Dirac
distribution, a variant of the Boltzmann distribution that takes into account the
Pauli exclusion principle. At a defined temperature T, i.e., at thermal equilibrium,
the mean number of electrons in energy level e is given by
1
/ðeÞ ¼ ð6:42Þ
1 þ exp ½ðe lÞ = ðkB TÞ
where kB is the Boltzmann constant and l is such that when e = l, then / = 1/2,
i.e., 50 % of the energy level with energy l is occupied. The inverse function of
(6.42) that gives the energy e as a function of the population of electrons / is

1
e ¼ l þ kB T ln 1 ð6:43Þ
/
This function is plotted in Fig. 6.38 for sodium (l = 3.24 eV), at 100, 1200,
4000, and 10,000 K (see Mathematica code M12, E10).
Semiconductors have a band structure qualitatively similar to that of electrical
insulators, with their valence band being fully occupied. However, unlike insula-
tors, the band gap for semiconductors is small and can be overcome by thermal
excitation of electrons at temperatures greater than a minimum value defined by the
band gap [Fig. 6.39a]. The excitation of electrons from the valence band leaves the
electron holes behind. In the presence of an electric field, excited electrons and
electron holes move in opposite directions. In fact, an electron moves in the
opposite direction to the applied electric field, whereas an electron hole moves in
the direction of the applied electric field, thus behaving like a positive charge. In a
semiconductor, both the electrons of the conduction band and the holes in the
valence band contribute to the electric current.
314 6 Crystals
Fig. 6.38 The Fermi–Dirac distribution for sodium (l = 3.24 eV) at T = 100, 1200, 4000, and
10,000 K shown as energy as a function of electrons population. Graph obtained with
Mathematica
Consider now a silicon crystal and the substitution of a few of its atoms by
arsenic in a mixing proportion of about one arsenic atom per one million atoms of
silicon (doping of silicon with arsenic). Arsenic has one more valence electron than
silicon. Thus, one arsenic atom is equivalent to a positive ion, isoelectronic with a
silicon atom, plus one electron added to the silicon lattice. At such a small mixing
proportion, it is unlikely that the crystalline network of silicon will be significantly
modified or disrupted, an undesirable effect that would cause an abrupt increase in
electrical resistivity. The addition of one electron per arsenic atom to the silicon
lattice results in a small donor band, just below the unoccupied band [Fig. 6.39b],
with an energy gap smaller than the medium thermal energy Therefore, thermal
excitation of electrons to the unoccupied band becomes possible, and the electrical
resistivity decreases with temperature (n-type semiconductor).
If the silicon crystal is doped with boron, whose atoms have one valence electron
fewer than silicon, then each boron atom is equivalent to a negative ion, isoelectronic
with silicon, plus one electron hole added to the silicon lattice that acts like a positive
charge. Formation of electron holes leads to charge separation (electrons and holes
move in opposite directions). Since energy is required to separate charges, the
acceptor band forms just above the fully occupied valence band [Fig. 6.39b].
6.6 Semiconductors 315
Fig. 6.39 Valence bands for semiconductors (schematic). a Valence band for pure semiconductor
at T = 0 K and T > 0 K. b Valence bands for n-type and p-type semiconductors
Therefore, electrons from the valence band are excited to the acceptor band using the
medium thermal energy and combine with holes in the acceptor band, causing a
decrease of electrical resistivity with temperature (p-type semiconductor).
Mathematica Codes
M1. Packing of Disks
The following Mathematica code builds disk lattices for square packing, whereby
each disk is tangent to four neighboring disks, and hexagonal packing, whereby
each disk is tangent to six neighboring disks. The Mathematica graphics primitive
for a disk is
Disk[{x,y},r]
where x and y are the Cartesian coordinates of the center of a disk with radius r. For a
square lattice of disks with radius 1/2, each of the Cartesian coordinates of the disk
centers varies in one-unit increments, and so the array of coordinates are given by
Table[Disk[{i,j},1/2],{i,3},{j,3}]
where i and j correspond to x and y, respectively. Hence, a fixed x and variable y
correspond to a row of disks, whereas a fixed y and variable x give a column of
316 6 Crystals
disks. Note that the outermost j of the Table function changes more rapidly than
the innermost i. This means that the disks are organized by rows, and the coor-
dinates of the disk centers in the square packing are given by
Table[{i,j},{i,3},{j,3}]
EdgeForm[Black] is a graphics directive that specifies that the edge of the
disks should be drawn in Black. In turn, the graphics directive LightGray
indicates that the disks should be filled in this color.
For a hexagonal close packing columns with odd j have x=i, and columns with
even j have x=i+1/2. In the Mathematica code, we use Boole[EvenQ[j]],
which gives the result 0 if EvenQ[j] is False (j is odd) and produces 1 if
EvenQ[j] is True (j is even). In addition, by inspection of the triangle of disks
in Fig. 6.1, it can be found that y=(√3/2)j. Hence, the hexagonal close packing
of disks can be obtained with
Table[Disk[{i+Boole[EvenQ[j]]/2,Sqrt[3]/2 j},1/2],{i,3},{j,3}]
and the coordinates of disk centers are given by
Table[{i+Boole[EvenQ[j]]/2,Sqrt[3]/2 j},{i,3},{j,3}]
M2. Hexagon of Disks
The close packing of disks in a plane shows that each disk is tangent to six disks,
whose centers are the vertices of a regular hexagon. The following Mathematica
code presents a hexagonal arrangement of disks around a central disk, by taking
advantage of its sixfold rotational symmetry axis. The Mathematica code that leads
to this arrangement is
Graphics[{EdgeForm[Black],Lightgray,Disk[{0,0},1/2],
Table[{EdgeForm[Black],Lightgray,Disk[{Cos[2pi/6],
Sin[2pi/6],1/2]},{i,0,5}]
The coordinates of the disks centers are given by
{{0,0},Table[{Cos[2pi/6],Sin[2pi/6]},{i,0,5}]
318 6 Crystals
M3. Disk Layers
In the picture on the right, this Mathematica code considers two close-packed
layers of identical disks, where the second layer is displaced so that its disk centers
coincide with the interstices of the first layer. The translation vector that shifts the
second layer of disks with respect to the first layer is shown in Fig. 6.4, where the
coordinates of its origin and its tip at the center of the interstice are (1, √3/2) and
(3/2, 2/√3), respectively. Therefore, the translation vector is given by
pffiffiffi
1 3
t ¼ ^x þ ^y
2 6
[see (6.1)]. Opacity varies from 0 to 1, where 0 corresponds to perfect trans-

parency, and Show presents several graphics combined.
Suggestion: Modify the above code so that it prints the coordinates of the disk
centers shown in gray.
M4. The Third Dimension
The Mathematica Graphics3D primitive for a sphere is

Sphere[{x,y,z},r]
where x, y, and z are the Cartesian coordinates of the center of a sphere with
radius r. In order to determine the z coordinate of a second horizontal layer of
close-packed spheres, we begin by considering four equivalent and mutually tan-
gent spheres of radius 1/2 whose centers form a regular tetrahedron inside a cube
(see Fig. 6.6). The distance between center 4 and the plane 1-2-3 is given by √2/√3,
as can be shown by simple trigonometry. Sphere 4 is nested on the interstice of
spheres 1, 2, and 3. Therefore, √2/√3 is the vertical distance between two horizontal
close-packed layers of spheres of radius 1/2. The vector that goes from center 1 to
center 4 is equal to the sum of vectors (4-c) and (c-1) and is given by
pffiffiffi rffiffiffi
1 3 2
t ¼ ^x þ ^y þ ^z
2 6 3
The figure below on the right shows a regular tetrahedron formed by four
mutually tangent spheres. The Mathematica line of code that leads to the
arrangement of three spheres on the left takes advantage of the threefold rotational
symmetry axis and is given by
Graphics3D[Table[{Gray,Sphere[{f*Sin[2pi/3],
f*Cos[2pi/3],0.0,1/2]},{i,0,2}]
where
f=1/Sqrt[3]
and
f*Sin[2pi/3]=x and f*Cos[2pi/3]=y
(see Fig. 6). Viewpoint!{0,0,Infinity} specifies a plan view from above,
supposedly from infinite distance.
320 6 Crystals
M5. HCP Structure
The following Mathematica code draws a figure for the hexagonal close packing of
disks and rigid spheres with equal radius:
M6. 2D Square Point Lattice
The following Mathematica code uses (6.12) and (6.13) to build a square point
lattice to which is applied a p/4 anticlockwise rotation (RotationMatrix rotates
2D vectors anticlockwise). Each point on the lattice corresponds to a defined pair of
integers {n1,n2}. Both arrays of points represent the same lattice, since the
translation vectors w1 and w2 result from v1 and v2 by a p/4 counterclockwise
rotation [see (6.15)]. The Mathematica command
M7. 2D Hexagonal Point Lattice
The following Mathematica code builds a hexagonal point lattice.

322 6 Crystals
M8. 3D Bravais Lattices
The following code uses the built-in Mathematica command LatticeData to

present images for the 14 Bravais lattices corresponding to 7 crystallographic
systems. Note that this command may require Internet connectivity:
324 6 Crystals
M9. CsCl Structure
The following Mathematica code builds the cesium chloride structure with the Cs+
and Cl− ions represented by spheres. If one Cl− is at the origin, the translation
vectors for the chloride ion lattice are given by (6.23) (Fig. 6.23) (simple cubic
lattice), and the vector that translates this lattice to produce the cesium ion lattice is
given by (6.29). The code forms two lattices, one for Cs+, the other for Cl−. One of
the most important Mathematica commands in this code is Cases, which gives a
list of elements that match a pattern. The output of Cases is a list of True and
False values. Another important command is Pick, which selects elements of a
list of True and False values. To illustrate the use of these two Mathematica
commands, we consider the following line,
Cases[{{-4.12},{-6.18},{2.06},{6.18}},{x_}!x<5&&x>-5]
whose output is
{True,False,True,False}
and
Pick[{aaa,bbb,ccc,ddd},{True,False,True,False}]
whose output is
{aaa,ccc}
Note that the above two lines of code can be condensed into the following single
line:
Pick[{aaa,bbb,ccc,ddd},Cases[{{-4.12},{-6.18},{2.06},
{6.18}},{x_}!x<5&&x>-5]]
The following Mathematica code uses Pick and Cases combined in the same
lines of code to define latticeCs and latticeCl:
M10. NaCl Structure
This Mathematica code builds the sodium chloride structure with the Na+ and Cl−
ions represented by spheres. This structure results from the interpenetration of two
face-centered cubic structures, one for the chloride ions, the other for the sodium
ions. The translation vectors for the face-centered cubic structure are given by
(6.25). If one Cl− is at the origin, the vector that translates the lattice of Cl− ions to
generate the sodium ion lattice is given by (6.30):
326 6 Crystals
M11. from Energy Levels to Bands
This Mathematica code draws energy levels for the p electron delocalized system of
a linear polyene, calculated by Hückel’s method [see (6.37)], as the number of
carbon atoms increases from 10 to 100, in steps of 10:
M12. Fermi–Dirac Distribution
The following Mathematica code plots the Fermi–Dirac distribution (6.42) for
sodium (l = 3.24 eV) at T = 100, 4000, and 10,000 K, showing the electron
population as a function of energy:
Glossary
Bravais lattice An infinite array of discrete points in

three-dimensional space such that for every point P
of the array defined by the positional vector p, it is
possible to find a set of integers (may be positive,
negative, or zero) (n1, n2, n3) that obey the following
vector equation p = n1v1 + n2v2 + n3v3, where v1,
v2, and v3 are linearly independent translation
vectors (noncollinear vectors) that span the lattice
Coordination number For a system of rigid spheres of the same radius is
the number of closest neighbors of each sphere
Electronic band The range of energies an electron may have in a
solid. The electronic band structure of a solid is
formed by electronic bands and band gaps that are
the ranges of energies the electron may not have
328 6 Crystals
Fermi–Dirac distribution Gives the probability of occupancy of energy

levels by electrons. It is a variant of the Boltzmann
distribution that takes into account the Pauli
exclusion principle
Optical transform A two-dimensional array of dots or other motifs on
a slide that with a beam of visible laser light
produces a diffraction pattern. With respect to an
X-ray diffraction experiment, an optical transform
corresponds to a reduction from 3D, on the real
crystal, to 2D, on the slide, a change of scale on
the X-ray’s wavelength and on the motif’s
dimensions. The diffraction pattern of an optical
transform is obtained on a screen about 20 m from
the slide, which is about 20 cm from the laser
Packing density Of rigid spheres is the ratio of the volume of all
spheres contained in a repeating unit and the
volume of the repeating unit. The packing density
measures the efficiency of occupation of the
three-dimensional space by rigid spheres
Semiconductor Material with electrical resistivity intermediate
between that of a metallic conductor and an
insulator in which electric current is transported by
electrons, as happens with metals, but unlike
metals, exhibits an electrical resistivity that
decreases exponentially with temperature when it
is extremely pure and decreases by several orders
of magnitude when it is doped with trace quantities
of an electrically active material
Unit cell The smallest parallelepiped that contains all
structural and symmetric information to build a
crystal lattice by translation, without overlapping
itself or leaving voids
Exercises
E1. Change the numerical values of Fig. 6.6 so that the new values refer to
identical spheres of radius r.
Exercises 329
E2. Use Mathematica to draw three tangent disks of radius 1/2 whose centers form
an equilateral triangle and print coordinates of the disks’ centers. Discuss the
results.
E3. Calculate the packing density for the hexagonal close-packed structure.
E4. Consider a two-dimensional structure formed by the merging of two Bravais
point lattices of black and gray points, where each black point is the center of an
equilateral triangle of gray points with the vertex pointing downward and each
gray point is the center of an equilateral triangle of black points with the vertex
pointing upward. Find the translation vector for converting one Bravais lattice
into the other, and use Mathematica to represent the whole two-dimensional
structure (black and gray points).
E5. Use Mathematica to represent a cube of spheres in the sc structure. Assume
r = 1/2 and take the lattice constant a so that the spheres become tangent along
the cube edge.
E6. Use Mathematica to represent a cube of spheres in the bcc structure.
Assume r = 1/2 and take the lattice constant a so that the spheres become
tangent along the cube main diagonal.
E7. Use Mathematica to represent a cube of spheres in the fcc structure. Assume
r = 1/2 and take the lattice constant a so that the obtained structure is cubic
close-packed.
E8. Use Mathematica to represent a hexagonal prism of spheres in the hcp
structure. Assume r = 1/2 and take the lattice constants a and c so that the
obtained structure is close-packed.
E9. Use Mathematica to arrive at the last equality of (6.40).
E10. Consider the Fermi–Dirac distribution function for metallic sodium at
7000 K. Calculate the energy in eV of the level whose occupation fraction is equal
to 0.25.
References
Kapustinskii AF (1956) Lattice energy of ionic crystals. Q Rev Chem Soc 10:283–294
Lisensky GC, Kelly TF, Neu DR, Ellis AB (1991) Simulating diffraction experiments in
introductory courses. J Chem Educ 68:91–96
Lucas AA, Lambin Ph, Mairesse R, Mathot M (1999) Revealing the backbone structure of B-DNA
from laser optical simulations of its X-ray diffraction diagram. J Chem Educ 76:378–383
McWeeny R (ed) (1979) Coulson’s Valence, 3rd edn. Oxford University Press, Oxford
Further Reading
Ashcroft NW, Mermin ND (1976) Solid state physics. Brooks/Cole Publishers Co.
Atkins PW (1998) Physical chemistry, 6th edn. W. H. Freeman and Company
Smart, L. and Gagan, M., editors, The Third Dimension, The Open University, Series The
Molecular World, 2002
Water
7
Abstract
Water determines life on earth to an extent greater than any other substance. We
know it in three physical states: as a liquid, in the oceans, seas, lakes, and rivers;
as a solid, in lakes covered with ice, in glaciers, and in the North and South
Poles; as a vapor, in the Earth’s atmosphere. Despite having a simple molecular
formula, water has unique structural and thermodynamic properties: it is denser
in the liquid than in the solid states (ice floats in liquid water), and it melts when
ice is subject to high pressure. In addition, the specific heat capacity of liquid
water is about four times that of dry air, and its surface tension is about three
times that of ethanol. The computer simulation of water by molecular dynamics
requires large computational resources and is still is a topic of current research.
In this chapter, we consider several properties of the water molecule including its
geometry, enthalpy of formation, atomic charges, dipole moment, molecular
orbitals, molecular vibrations, and hydrogen bond. The remaining sections of
this chapter take the standpoint of thermodynamics and consider normal ice,
liquid water, the phase diagram of water, water as a solvent, simple nonpolar
solutes, ionic solutions, amphipathic molecules, acids and bases, and standard
electrode potentials. The Mathematica codes presented at the end of the chapter
(Representation of the Water Molecule, Electrostatic Potential Contours for a
Dipole, Interactive Manipulation of Charge, Vector Field Streamlines, Luzar’s
Model, Micelle-Monomers Equilibrium, Critical Micelle Concentration, Weak
Acid HA) are accompanied with detailed explanations and provide insight into
the corresponding subjects. Following the Mathematica codes, we present a
glossary of important scientific and technical terms on water and aqueous
solutions and a list of exercises, whose complete answers can be found after the
Appendix.

DOI 10.1007/978-3-319-41093-7_7
332 7 Water
7.1 Molecular Geometry
The experimental values of the OH bond length and the HOH bond angle of a water
molecule are ROH = 0.958 Å and aHOH = 104.4776° (CCCBDB), showing that H2O
is a symmetric nonlinear triatomic molecule (Fig. 7.1; see Mathematica code
M1). The geometry optimization of a water molecule carried out by the system of
programs Gaussian 09 at the B3LYP level of calculation with the basis set cc-pVTZ
gives ROH = 0.9611 Å and aHOH = 104.5296° (Fig. 7.1).
Figure 7.2 shows the energy profiles for the scans of the bond length ROH when
aHOH is 104.53°, and of the bond angle aHOH when ROH is 0.961 Å. Note that the
HOH bond angle differs from the tetrahedral angle (109.47°) by approximately 5°, a
value much smaller than the HOH bond angle difference with respect to the right
angle, which is about 19°.
0.961 Å
(0.958 Å)
O
104.53o
(104.48o)
Fig. 7.1 Ball and bond representation of the H2O molecule with calculated and experimental
(CCCBDB) values of the geometric variables ROH and aHOH (experimental values are shown in
parentheses). Molecular representation obtained with GaussView
Fig. 7.2 At left, scan of the bond length ROH when aHOH is equal to 104.53°. At right, scan of the
bond angle aHOH with ROH equal to 0.961 Å. Note that the energy scales of the above plots are
quite different. Graphs obtained with Mathematica
7.2 Enthalpy of Formation 333
7.2 Enthalpy of Formation
For a gas phase water molecule, the experimental value of the enthalpy of formation
at 298.15 K is −241.81 kJ mol−1 (CCCBDB). Computing a molecular energy with
chemical accuracy means attaining an accuracy of 1 kcal mol−1 (4.2 kJ mol−1).
Since the largest errors in ab initio calculations result from basis set truncation,
achieving chemical accuracy requires a series of calculations that extrapolate to the
complete basis set limit. This type of calculation was carried out by the system of
programs Gaussian 09 using the composite method W1BD (Martin and de Oliveira
1999; Parthiban and Martin 2001), where W stands for Weizmann Institute and BD
for Brueckner Doubles (Brueckner orbitals). Within W1BD, the geometry opti-
mization is carried out by a B3LYP/cc-pVTZ calculation.
The Schrödinger equation is obtained within the Born–Oppenheimer approxi-
mation, and the electron wave function and its corresponding energy Uelec depend
parametrically on the nuclear coordinates. The energy Uelec is evaluated at 0 K. In
order to convert this energy from 0 to 298.15 K, the thermal energy given by the
sum of the translation, rotation, and vibration contributions (Utrans + Urot + Uvib) at
298.15 K needs to be added to Uelec. Gaussian evaluates the thermal energy and the
enthalpy at 298.15 K (the enthalpy H is defined as U + pV, with the product pV
being equal to RT for a mole of an ideal gas). In order to obtain the enthalpy of
formation of H2O(g), the same composite calculation method has to be applied to
H2O, H2, and O2, since the chemical reaction for the formation of one mole of water
in the gas phase is given by H2(g) + (1/2) O2(g) ! H2O(g). The calculated value of
DHof at 298.15 K is −245.23 kJ mol−1, a value that differs from the experimental
value by less than 1 kcal mol−1.
7.3 Atomic Charges
The electrostatic potential for the H2O molecule mapped with rainbow colors over a
total density surface with isovalue equal to 0.00040 (Fig. 5.4) shows that the hy-
drogen atoms of the water molecule are strongly electrophilic and the oxygen atom
is strongly nucleophilic.
The electrostatic potential of a molecule can be used to determine its distribution
of atomic charges. To this end, the molecular electrostatic potential / and the
electrostatic potential η that results from a set of point charges qa located on the
molecule’s nuclei are evaluated on the nodes i of a grid. For a neutral molecule like
H2O the sum of all the atomic charges is zero, and so the charge centered on the
oxygen nucleus is −2 times the charge on each of the hydrogen atoms, qO = −2 qH.
Hence, in atomic units, η is given by
X qa 2qH qH qH
gi ¼ ¼ þ þ ð7:1Þ
a
rai rOi rH1 i rH2 i
334 7 Water
where i represents a grid node. By the least squares optimization procedure, the
atomic charge qH can be determined by minimizing
X
HðqH ; rOi ; rH1 i ; rH2 i Þ ¼ ðgi /i Þ2 ð7:2Þ
i
with respect to qH, that is, by solving
@H
¼0 ð7:3Þ
@qH
The following table shows the molecular electrostatic potential (MEP) atomic
charges and the Mulliken atomic charges for H2O obtained from B3LYP calcula-
tions with different basis sets. It is shown that the introduction of polarized func-
tions on oxygen and hydrogen atoms significantly improves the results, as can be
expected. In addition, molecular electrostatic atomic charges obtained from large
basis sets with diffuse functions give the same atomic charges. Note that Mulliken
charges vary with the basis sets somewhat erratically.
MEP Mulliken
Method/basis set qO qH qO qH
B3LYP/6-31G −0.897 0.448 −0.718 0.359
B3LYP/6-31G(d,p) −0.737 0.368 −0.610 0.305
B3LYP/6-311G(d,p) −0.749 0.375 −0.474 0.237
B3LYP/cc-pVTZ −0.703 0.352 −0.435 0.218
B3LYP/cc-pVQZ −0.698 0.349 −0.490 0.245
B3LYP/aug-cc-pVTZ −0.680 0.340 −0.354 0.177
B3LYP/aug-cc-pVQZ −0.680 0.340 −0.587 0.294
7.4 Dipole Moment
7.4.1 Electric Multipoles
In this section, we deal with simple charge distributions and their electrostatic
potentials. We begin by considering charge distributions that correspond to pure
electric multipoles (charge, dipole, quadrupole, and octupole) and their corre-
sponding numbers of poles expressed as powers of 2 (Fig. 7.3). Since a charge can
be either positive or negative, the number of poles is given by a power of 2.
The contours for the electrostatic potential of a charge q, in the xy-plane and
atomic units, are given by
7.4 Dipole Moment 335
Fig. 7.3 Electric multipoles (schematic) with the corresponding number of poles expressed as a
power of 2
q
/ðx; yÞ ¼ ð7:4Þ
r
where r represents the distance between the point P with coordinates (x, y, 0) and
the charge located at the origin (0, 0, 0). These contours are concentric circles
whose center is the point charge q (Fig. 7.4), since the electrostatic potential pro-
duced by a point charge depends only on the distance to the charge. All the radial
directions emerging from the charge display the same variation of the electrostatic
potential, that is, they are equivalent directions, and the space around the central
charge is said to be isotropic.
The electrostatic potential of a dipole is

1 1
/ðx; yÞ ¼ q ð7:5Þ
r1;0 r1;0
where q represents the absolute value of the charges (the minus sign in the second
member of this expression accounts for the sign of the corresponding negative
charge), and r−1,0 and r1,0 represent the distances between the point P in the xy-
plane and the negative and positive charges with coordinates (−1, 0, 0) and (1, 0, 0),
respectively (Fig. 7.4; see Mathematica codes M2 and M3; E1, E2, E3).
Fig. 7.4 Contours in the xy-plane for electrostatic potentials produced by one charge, one dipole,
and one quadrupole. Figures obtained with Mathematica
336 7 Water
The electrostatic potential of a quadrupole is given by (Fig. 7.4).

1 1 1 1
/ðx; yÞ ¼ q þ ð7:6Þ
r0;1 r0;1 r1;0 r1;0
7.4.2 Point Dipole
As mentioned earlier, two electric charges +q and −q separated by a small distance

d form an electric dipole. We now represent the distances between the positive and
the negative charges, on the one hand, and the point P in the xy-plane, on the other,
by r+ and r− (Fig. 7.5), and write the electrostatic potential as

1 1 r r þ
/ðx; yÞ ¼ q ¼q ð7:7Þ
r þ r r þ r
Substitution of
" 2 #1=2 " 2 #1=2
d d
rþ ¼ x þ y2 r ¼ xþ þ y2 ð7:8Þ
2 2
in (7.7) followed by Taylor expansion of the resulting function of d about the point
d = 0 gives
d2 00 d3
/ðdÞ ¼ /ð0Þ þ d/0 ð0Þ þ / ð0Þ þ /000 ð0Þ þ ð7:9Þ
2! 3!
where
qx qxð15x2 9r 2 Þ
/ð0Þ ¼ 0 /0 ð0Þ ¼ /00 ð0Þ ¼ 0 /000 ð0Þ ¼ ð7:10Þ
r3 4r 7
Fig. 7.5 Geometric variables

for a dipole
In turn, substitution of (7.10) in (7.9) leads to
q d x q d3 xð5x2 3r 2 Þ
/ðdÞ ¼ þ þ Oðd5 Þ ð7:11Þ
r3 8r 7
Since the exchange of positive and negative charges changes the sign of the
electrostatic potential, this is an odd function of d [/(d) = −/(−d)], and conse-
quently, the even powers of d in (7.9) are zero. If d represents an arbitrarily small
separation between the charges of the dipole, then we can approximate /(d) to the
first term of (7.11) (the terms in d3 and O(d5) become negligible compared with the
d term) and write
l ¼ qd ð7:12Þ
where l is the magnitude of the point dipole vector. Note that x = r cosh (E4).
7.4.3 Electric Field Streamlines
From the definition of electrostatic potential, (see 5.10), one can write
V ¼ qt / ð7:13Þ
where V is the potential energy, qt is the test charge, and / is the electrostatic
potential. The x and y components of the electric force are given by
@V @/ @V @/
Fx ¼ ¼ qt Fy ¼ ¼ qt ð7:14Þ
@x @x @y @y
and the x and y components of the electric field are
Fx Fy
Ex ¼ Ey ¼ ð7:15Þ
qt qt
@/ @/
Ex ¼ Ey ¼ ð7:16Þ
@x @y
Note that
@/ ^ @/ ^
iþ j ¼ r/ ð7:17Þ
@x @y
is the gradient vector of the electrostatic potential in the xy-plane, and ∇ represents
the del, or nabla, operator indicative of the vector differential operator.
338 7 Water
Consider now a contour in the xy-plane described by the x and y variables as

functions of a parameter t (an external variable). If t represents the variable time, as
it usually does in physical problems, then

@x ^ @y ^
v¼ iþ j ð7:18Þ
@t t0 @t t0
is the velocity vector tangent to the contour at (x0, y0) (this point corresponds to
time t0). The dot product of the gradient of the electrostatic potential and the
velocity vector is given by

@/ @x @/ @y d/
r/:v ¼ þ ¼ ¼0 ð7:19Þ
@x x0 ;y0 @t t0 @y x0 ;y0 @t t0 dt t0
This dot product is zero, because the electrostatic potential does not vary on a
contour curve by the definition of contour. Hence, at (x0, y0), the gradient of the
electrostatic potential is perpendicular to the velocity vector, which in turn is tan-
gent to the contour at that point. The gradient vector points in the direction of the
greatest rate of increase of the electrostatic potential. From (7.16) and (7.17), one
concludes that the electric field vector in the xy-plane is
E xî þ E y^j ¼ r/ ð7:20Þ
and so, at a defined point of the xy-plane, the electric field vector points in the
direction of the greatest decrease of the electrostatic potential. Figure 7.6 shows a
plot of the electrostatic potential contours and the electric field streamlines for a
dipole (Fig. 7.6, left) and a point dipole (Fig. 7.6, right; see Mathematica code
M4). It can be seen that the streamlines move away from the positive charge (the
source) and converge to the negative charge (the sink), since these are the direc-
tions of the greatest decrease of the electrostatic potential. In the case of the point
dipole, source and sink coincide (the streamlines move away and converge to the
point dipole).
7.4.4 H2O Dipole and Quadrupole
We now consider the atomic charges for the water molecule derived from the
electrostatic potential obtained from a Gaussian 09 B3LYP/aug-cc-pVTZ calcula-
tion (see Sect. 7.3). Since the electrostatic potential is an additive function, it can be
equated to a sum of electrostatic potentials associated with nonnull electric multi-
poles and represented by the contours in the molecular plane. Figure 7.7 shows the
contours of the total electrostatic potential of H2O as an approximate sum of the
contours for the electrostatic potential of the dipole term associated with the vector
Fig. 7.6 Contours and streamlines show the direction of the electric field, for a dipole (left) and a
point dipole (right). Figures obtained with Mathematica
sum of the O–H bond dipoles and the electrostatic potential for a quadrupole term
(see E5). These results show that besides the dipole, the H2O molecule has also a
significant contribution from the quadrupole.
7.5 Molecular Orbitals
In Hartree–Fock and DFT calculations, the wave function for a closed shell
molecule is a Slater determinant of one-electron wave functions (spin-orbitals) that
are eigenfunctions of effective one-electron Hamiltonians called Fock and Kohm–
Fig. 7.7 Contours for the electrostatic potential of the H2O molecule derived from the distribution
of atomic charges (Gaussian09 B3LYP/aug-cc-pVTZ calculation) as an approximate sum of the
contours for a dipole and a quadrupole. Figures obtained with Mathematica
340 7 Water
Sham operators (E6). For a molecule, these one-electron wave functions are called
molecular orbitals and are expressed as linear combinations of basis functions
centered at each atom of the molecule.
C2v character table
E C2 rxz ryz
A1 1 1 1 1 z x2, y2, z2
A2 1 1 −1 −1 Rz xy
B1 1 −1 1 −1 x, Ry xz
B2 1 −1 −1 1 y, Rx yz
The H2O molecule belongs to the C2v symmetry group, whose symmetry ele-
ments are shown in Fig. 7.8. The symmetry species of this group are presented in
the corresponding character table (see above). Upon application of the C2v sym-
metry operations, the H2O molecular orbitals transform as the group symmetry
species. For this reason, the molecular orbitals are labeled by the corresponding
symmetry species. Being one-electron functions, they are labeled by lowercase
letters.
The water molecule has ten electrons that are allocated to the five molecular
orbitals with lower energy (each molecular orbital is occupied by two electrons with
antiparallel spins). The five doubly occupied molecular orbitals of H2O are shown
in Fig. 7.9 by surfaces of defined isovalues, with the corresponding energies in eV,
C2
σ xz
z
O O O
σ yz
z y
z x
Fig. 7.8 Symmetry elements for the H2O molecule. Molecular representations obtained with
GaussView
1a1 2a1 1b2 3a1 1b1

(-520.3 eV) (-27.4 eV) (-14.4 eV) (-10.5 eV) (-8.4 eV)
Fig. 7.9 Molecular orbitals for the H2O molecule represented by surfaces whose isovalue is equal
to 0.02000 (Gaussian 09 B3LYP/cc-pVTZ calculation)
7.5 Molecular Orbitals 341
in parentheses, and a1, b1, b2 name the molecular orbitals by their symmetry types.
The orbitals that have a common symmetry label are numbered by increasing
energy (for instance, 1a1, 2a1, 3a1). The occupied molecular orbital of lowest energy
(1a1) is essentially restricted to the 1s oxygen atom core electrons, which are not
significantly involved in the O–H bonds.
7.5.1 Natural Bond Orbitals
Figure 7.10 shows the natural bond orbitals BD1 and BD2 and the natural atomic
orbitals LP1 and LP2 for the H2O molecule. These natural orbitals have the fol-
lowing approximate expressions:
LP2 pO
LP1 sp0:86
O
ð7:21Þ
BD2 0:85 sp3:33
O!H2 þ 0:52 sH2
BD1 0:85 sp3:33
O!H1 þ 0:52 sH1
The natural bond orbitals BD1 and BD2 are approximate localized in the OH
bonds and are equivalent orbitals by molecular symmetry, since the OH bonds of
H2O are equivalent bonds. In turn, LP1 and LP2 are approximate localized in the
oxygen atom (lone pairs or natural atomic orbitals).
LP 2
LP 1
BD 1 BD 2
Fig. 7.10 Surfaces of isovalue equal to 0.0200 represent the natural bond orbitals BD1 and BD2
and the oxygen atom lone pairs LP1 and LP2 (Gaussian 09 B3LYP/cc-pVTZ calculation)
342 7 Water
7.6 Molecular Vibrations
The H2O molecule has N = 3 atoms and consequently 3N – 6 = 3 vibrational

coordinates. Two of these are O–H stretchings, the third being the H–O–H angle
bending. According to the H2O symmetry (C2v symmetry group), the equivalent
O–H oscillators cannot oscillate separately: they oscillate either in phase
(symmetric stretching) or out of phase (antisymmetric stretching). In the sym-
metric stretching, both H atoms move apart from the oxygen atom or get closer to it
synchronously, that is, in phase. Apart from a normalization factor, the symmetric
stretching coordinate is the sum of the OH stretching coordinates and transforms
like the A1 symmetry species. In turn, in the antisymmetric stretching, one of the H
atoms moves apart from the oxygen atom and the other H atom gets closer to it, that
is, they move out of phase. Both pass by their extreme positions at the same time.
Apart from a normalization factor, the antisymmetric stretching coordinate is the
difference of the OH stretching coordinates and transforms like the B2 symmetry
species. The H–O–H bending is a totally symmetric vibration, that is, it transforms
as the symmetry species A1.
The vibrational modes of a molecule that transform like the species of its
symmetry group are called normal vibrational modes. The antisymmetric
stretching, the symmetric stretching, and the bending are the normal vibrational
modes of the H2O molecule (Fig. 7.11) whose scaled wavenumbers (the scaling
factor for the calculation B3LYP/cc-pVTZ is equal to 0.965; see CCCBDB) are
3767, 3670 and 1581 cm−1, respectively. These frequencies are less than 1% away
from the experimental values (3756, 3657, 1595 cm−1; see CCCBDB).
Displacement vectors and scaled wavenumbers for the calculated normal vibrational modes of H2O
m/cm−1 3767 3670 1581
x y z x y z x y z
O 0.00 0.07 0.00 0.00 0.00 0.05 0.00 0.00 0.07
H 0.00 −0.56 0.43 0.00 0.59 −0.39 0.00 −0.43 −0.56
H 0.00 −0.56 −0.43 0.00 −0.59 −0.39 0.00 0.43 −0.56
3767 3670 1581
O O O
H H H H H H
Fig. 7.11 Normal vibrational modes of H2O with the wavenumbers in cm−1 after scaling by the
factor 0.965 (CCCBDB) and the corresponding displacement vectors (Gaussian 09 B3LYP/
cc-pVTZ calculation). In the normal mode at 3767 cm−1, the displacement vector for the OH
oscillator on the right has its origin on the H atom and is directed toward the O atom, thus being
hidden by the segment that represents the OH bond
The calculated displacement vectors are shown in the above table. Note that the
x-axis is perpendicular to the molecular plane (the yz-plane) (see Fig. 7.8). Since the
displacement vectors lie in the molecular plane, all the x-components of the dis-
placement vectors are zero. The numbers highlighted in bold correspond to the
largest displacements of the hydrogen atoms, and the small displacements of the
oxygen atom move this atom so that the center of mass of the molecule stays
unshifted (a center of mass motion is a translation, not a vibration).
Displacement vectors and scaled wavenumbers for the calculated normal modes of vibration of
DOH
m/cm−1 3720 2701 1386
x y z x y z x y z
O 0.00 −0.05 0.03 0.00 0.09 −0.07 0.00 0.00 0.09
H 0.00 0.81 −0.59 0.00 0.11 −0.01 0.00 −0.55 −0.70
D 0.00 0.01 0.03 0.00 −0.80 −0.58 0.00 0.28 −0.35
Protium (1H) and deuterium (D = 2H) have the same atomic number Z = 1, and
so both are hydrogen atoms. Within the Born–Oppenheimer approximation, the
H2O and DOH molecules have the same molecular orbitals. However, the vibra-
tional frequencies depend on the reduced masses of the oscillators, and a protium by
deuterium substitution breaks the equivalence of the oxygen–hydrogen oscillators.
The reduced masses of the OH and OD oscillators in unified mass units are
approximately equal to 16 1/(16+1) = 16/17 and 16 2/(16+2) = 32/18,
respectively. Hence,
rffiffiffiffiffiffiffiffi pffiffiffi
xOH lOD
2 ð7:22Þ
xOD lOH
(E7).
The calculated scaled wavenumbers and displacement vectors for the normal
modes of DHO are shown in Fig. 7.12. The normal modes at 3720 and 2701 cm−1
have predominant contributions of the OH and OD stretching vibrations, respec-
tively (see table of atomic displacements). The large variation in the reduced mass
3720 2701 1386
O O O
H D H D H D
Fig. 7.12 Normal vibrational modes of DHO with the wavenumbers in cm−1 after scaling with
the factor 0.965 (CCCBDB), and the corresponding displacement vectors (Gaussian 09 B3LYP/
cc-pVTZ calculation)
344 7 Water
of the oxygen–hydrogen oscillator that resulted from the protium by deuterium

substitution has the effect of localizing the OH and OD stretching vibrations.
Hence, the vibrational mode at 3720 cm−1 is essentially the OH stretching, whereas
the mode at 2701 cm−1 has a predominant contribution from the OD stretching
(note the wavenumbers’ shift of more than 100 cm−1 with respect to the OH
stretching) (E8, E9).
7.7 Intermolecular Interactions
The existence of intermolecular interactions can be easily inferred from simple

experimental observation: gases would not liquefy and liquids would not solidify if
there were no attractive long-range intermolecular interactions. Repulsive interac-
tions occur at short range and prevent solids and liquids from being compressed
into much smaller volumes. Hence, long-range and short-range interactions can be
inferred from the simple existence of liquids and solids and their reduced
compressibility.
Intermolecular interactions can be classified as electrostatic, induction, and
dispersion, according to the physical phenomena that produce their long-range
behavior, which leads to energies that vary with powers of inverse distance. In turn,
short-range interactions are mostly due to repulsion and exchange between elec-
trons of different molecules, and they decrease exponentially with distance.
Long-range and short-range interactions can be illustrated by the potential energy
curves for the interaction between two closed-shell atoms or monatomic ions, since
the corresponding potential energy curves provide useful insight into the funda-
mental features of the interactions.
7.7.1 Electrostatic Interaction
Sodium and fluoride ions have the same number of electrons (10) and the same
closed-shell electron configuration, 1s22s22p6. The potential energy curve for
Na+…F– shows a deep minimum at approximately 2.0 Å that results from a balance
between repulsive and attractive interactions (Fig. 7.13). For interatomic distances
smaller than that of the minimum, the dominant interaction has its origin in the
repulsion and exchange between electrons of different atoms (exchange of electrons
of different atoms requires overlap between the electron densities of both atoms)
and corresponds to a curve that decreases with increasing internuclear distance. In
turn, the leading long-range interaction between the positive and negative charges
of Na+ and F– is electrostatic, mostly of the charge-to-charge type. This interaction
is approximately described by a −1/R curve that represents the attractive wing of the
potential well (Fig. 7.13) and tends asymptotically to zero, thus meaning that the
additive constant in the −1/R fitting is set to zero energy. At R = 6.0 Å, the
potential energy curve is well below zero, and the charge-to-charge electrostatic
7.7 Intermolecular Interactions 345
Fig. 7.13 Potential energy

curve for Na+ … F–
(Gaussian 09 B3LYP/
aug-cc-pVTZ calculation).
Curves are separately fitted to
short-range repulsive and
long-range attractive wings of
the potential energy curve.
The long-range fitting curve
corresponds to the
1/R distance dependence. The
zero energy was chosen so
that the fitting curve tends to
zero as R ! ∞.
Mathematica
interaction is still strong, since electrostatic interactions are relatively strong

interactions.
Electrostatic interactions involve permanent electric moments of the interacting
molecules, and generally speaking can be of the charge-to-charge, charge-to-dipole,
dipole-to-dipole, dipole-to-quadrupole, … types. Usually, the leading electrostatic
interaction comes from the first significant electric moment in each of the inter-
acting molecules. The interaction between permanent 2n-pole and 2 m-pole electric
moments varies as 1/R(n+m+1). For a charge–charge interaction, n = 0 and m = 0,
and the distance-dependence is given by 1/R. A charge–dipole interaction corre-
sponds to n=0 and m=1 with the resulting electrostatic interaction varying with
distance as 1/R2. In turn, a dipole–dipole interaction corresponds to n = 1 and
m = 1, that is, the interaction varies with distance as 1/R3. Electrostatic interactions
can be attractive or repulsive and involve permanent electric moments of the
interacting molecules, thus being pairwise additive. For an assembly of molecules
A, B, and C, pairwise addition means that the total electrostatic energy is given by
the sum of the electrostatic interactions associated with the permanent electric
moments of the AB, BC, and CA pairs.
Since the electron density of atoms or molecules is polarizable, a permanent
electric moment in one molecule gives rise to an induced electric moment into the
other molecule. Contributions like these are included in the induction interaction, as
we now consider.
346 7 Water
7.7.2 Induction
The sodium ion is isoelectronic with a neon atom, and both have the same number
of electrons (10) and the same electronic configuration, 1s22s22p6. Gaussian
calculations for the Na+…Ne system were carried out using the Moller–Plesset
perturbation method to fourth order of perturbation (MP4) with the aug-cc-pVTZ
basis set [MP4(sdq)/aug-cc-pVTZ calculation]. The potential energy curve for
Na+…Ne (Fig. 7.14) shows an energy minimum at 2.55 Å whose depth is more
than one order of magnitude smaller than the minimum observed for Na+…F–. The
charge of Na+ induces a dipole in the neon atom with the negative end closer to the
sodium ion. The resulting charge-induced dipole attractive interaction is called
induction.
In general, the interaction between a permanent 2n-pole and an induced 2 m-pole
varies as 1/R2(n+m+1). For a charge-induced dipole interaction, n=0 and m=1, the
distance dependence is given by 1/R4. Induction interactions are clearly nonaddi-
tive, since the multipole moments induced by a third molecule interact, being
subject to mutual change. The weak solubility of oxygen in water can be approx-
imately described by the interaction between the dipole moment of the water
molecule and the induced dipole moment in the diatomic oxygen molecule.
Fig. 7.14 Potential energy

curve for Na+…Ne [Gaussian
09 MP4(sdq)/aug-ccpVTZ
calculation]. Curves are
separately fitted to the
short-range repulsive and
long-range attractive wings of
the potential energy curve.
The long-range fitting curve
corresponds to the 1/R4
dependence. The zero energy
was chosen so that the 1/R4
fitting curve tends to zero as
R ! ∞. Figure obtained with
Mathematica
7.7.3 Dispersion
Helium and neon boil at 4.22 K and 27.07 K, with these boiling temperatures
corresponding to RT values approximately equal to 0.04 and 0.2 kJ mol−1. Since
these energies are well below 1 kcal mol−1 (4.2 kJ mol−1), the calculation of the
interaction energy between two rare gas atoms cannot be obtained from the dif-
ference between the total energy of the system of the interacting atoms (EA…A) and
the sum of the individual energies of each atom (2EA), since the errors of these large
numbers are far greater than the resulting energy difference. The interaction
between two rare gas atoms corresponds to a tiny perturbation, and so should be
evaluated as such by a perturbation method. To this end, Gaussian 09 calculations
for the He…He and Ne…Ne systems were carried out using the Moller–Plesset
perturbation method to fourth order of perturbation (MP4) with the aug-cc-pVTZ
basis set [MP4(sdq)/aug-cc-pVTZ calculation] (Aziz and Slaman 1991; Woon
1991). The potential energy curves for He…He and Ne…Ne systems present very
shallow minima (approximately 0.07 and 0.2 kJ mol−1) at interatomic distances
around 3.05 and 3.15 Å, respectively, with the minimum for Ne…Ne being about
nine times deeper than for He…He (see insets in the potential energy curves of
Fig. 7.15).
Isolated atoms are isotropic systems with respect to their nuclei, and all per-
manent electric moments of neutral atoms are null. Therefore, the observed minima
Fig. 7.15 Potential energy curves for He…He and Ne…Ne as functions of the internuclear
distance [Gaussian 09 MP4(sdq)/aug-ccpVTZ calculation]. Curves are fitted to the short-range
repulsive wings of the potential energy curve. Details of the minima and the curves fitted to the
attractive wings of the very shallow potential wells are shown in the insets. Energy zeros were
chosen so that the long-range fitting curves tend to zero as R ! ∞. Figures obtained with
Mathematica
348 7 Water
in the potential energy curves of He…He and Ne…Ne systems cannot derive from
interactions involving permanent electric moments. What is the origin of the
observed minima? Why is the minimum of Ne…Ne deeper than that of He…He?
Electrons try to avoid each other due to instantaneous repulsions. While the
instantaneous asymmetry in the electron distribution of a single atom averages to
zero over time, the asymmetry in the electron distribution of one atom produced by
correlation between electrons of different atoms can be described in each atom by
successive induced electric moments (induced dipole, induced quadrupole, …) and
depends on the internuclear distance. Figure 7.16 schematically illustrates how an
instantaneous dipole in the electron distribution of one atom induces a dipole in the
electron distribution of the other atom. The resulting induced dipole-induced dipole
interaction is attractive, since induced dipoles always reorient to produce attraction.
In general, the interaction between induced 2n- and 2m-poles varies as 1/R2(n+m+1).
An induced dipole-induced dipole interaction (n = 1 and m = 1) corresponds to
1/R6 dependence. The correlation between electrons of different atoms can also
result from the contributions of induced electric moments of order greater than 1,
for instance, induced dipole-induced quadrupole (1/R8) and induced quadrupole-
induced quadrupole (1/R10) interactions. The overall resulting attraction is called
dispersion. Due to correlation between electrons of different atoms, neon atoms
lead to a larger well depth than helium atoms, since the dispersion energy increases
with the number of electrons of the interacting atoms. An example of this is the fact
that chlorine is a gas, bromine a liquid, and iodine a solid.
Since the electron densities of interacting atoms or molecules are polarizable,
dispersion always occurs, that is, it is found in every molecular interaction. Its 1/R6
distance-dependence makes it approximately additive, despite being due to an
interaction between induced electric moments. For small or medium-size mole-
cules, dispersion is a weak interaction, but it becomes significant and sometimes
even decisive when the remaining interactions are null or negligible.
Dispersion interactions between alkane molecules contribute appreciably to
determining the physical state of a hydrocarbon. In fact, the magnitude of the
dispersion interactions increases with the number of electrons in the atoms located
on the molecular surface (for alkane molecules, these are hydrogen atoms). This
means that the magnitude of the dispersion interaction increases with the area of the
molecular surface and the number of hydrogen atoms on it. For example,
2,2-dimethyloctane (a branched decane) has a normal boiling point at 155 °C,
Fig. 7.16 Correlation between electrons of different atoms leads to attraction involving an
induced dipole-induced dipole interaction (schematic)
whereas the boiling point of n-decane is 174 °C (the linear alkane has a larger
surface area than the branched alkane).
The boiling points of linear alkanes are plotted in Fig. 7.17 as a function of the
number of hydrogen atoms. Since the general formula for an alkane is CnH2n+2, an
increase of one carbon atom corresponds to an increase of two hydrogen atoms, and
linear alkanes that have four or fewer carbon atoms are gases at ambient temper-
ature. Long hydrocarbon chain molecules with a large number of hydrogen atoms in
the interfacial area between molecules have appreciable dispersion energies. This
can be inferred from their high boiling points: at ambient temperature and normal
pressure, linear alkanes with more than seven carbon atoms are liquids, and linear
alkanes with more than 17 carbon atoms are solids.
Electrostatic, induction, dispersion, and exchange-repulsion interactions are
fundamental interactions. We now consider a specific interaction in which the
hydrogen atom exerts a major role: the hydrogen bond.
7.8 Hydrogen Bond
Consider one hydrogen fluoride molecule interacting with one water molecule. The
potential energy curve as a function of the H…O distance (Fig. 7.18) presents an
energy minimum in which the hydrogen atom covalently bonded to the fluorine
atom interacts with the oxygen atom of the water molecule. The relaxed scan of the
H…O distance leads to an energy minimum whose depth is approximately
37 kJ mol−1 (Gaussian 09 B3LYP/aug-cc-pVTZ calculation). Acting like a bridge
between two electronegative atoms, fluorine and oxygen, the hydrogen atom
establishes a weak directional interaction with one of the lone pairs in the oxygen
atom. At the minimum, the F–H bond length (0.945 Å) is slightly longer than the
bond length of an isolated hydrogen fluoride molecule (0.924 Å), and the H…O
Fig. 7.17 Normal boiling

points of linear alkanes versus
the number of hydrogen
atoms. Figure obtained with
Mathematica
350 7 Water
internuclear distance (1.692 Å) almost doubles the F–H bond length. The F–H…O
interaction is a hydrogen bond and the FH…OH2 complex can be called a
hydrogen bond heterodimer, with the HF and H2O molecules having distinct roles:
HF provides the hydrogen atom (it is called the donor), whereas H2O provides the
lone pair in the oxygen atom (it is called the acceptor).
The hydrogen chloride–water system provides another example of a hydrogen
bond heterodimer, ClH…OH2. This time, the relaxed scan of the H…O distance
leads to a minimum, whose calculated depth is approximately 20 kJ mol−1
(Gaussian 09 B3LYP/aug-cc-pVTZ calculation).
Isodensity surfaces mapped with electrostatic potentials can be used to illustrate
the hydrogen bond formation in the FH…OH2 heterodimer (Fig. 7.19). The elec-
trostatic attractive interaction between the electrophilic region of the hydrogen atom
in the hydrogen fluoride molecule and the nucleophilic region of the oxygen atom
in the water molecule leads to the formation of the hydrogen bond heterodimer. The
complementary functions of the donor and acceptor molecules are well illustrated
by this electrostatic potential color map.
Note that each hydrogen bond heterodimer corresponds to a single closed iso-
density surface, that is, the hydrogen fluoride and the water moieties do not have
distinct isodensity surfaces. This means that the H…O interaction has some partial
covalent character, which leads to the directional properties of the hydrogen bond.
The fact that the expression “hydrogen bond” proposed by M.L. Huggins in 1919
prevailed over “hydrogen bridge” is probably due to the recognition of the covalent
character of the hydrogen bond. In addition, the heterodimer exhibits stronger and
larger electrophilic regions in the hydrogen atoms of the water molecule than in the
isolated water molecule and a stronger and larger nucleophilic region in the fluoride
atom of the heterodimer than in the corresponding isolated hydrogen fluoride
molecule. These features point to the polarization of the FH…OH2 heterodimer
with respect to the individual molecules.
Fig. 7.18 Relaxed scan of

H
the H…O distance in FH…
OH2 (Gaussian 09
B3LYP/aug-cc-pVTZ F H O
calculation). Figure obtained E/kJmol -1
with Mathematica. Molecular 50 1.692 Å H

representations obtained with
GaussView 40
30
20
10
0 RHO /
2 3 4 5 6 7
7.8 Hydrogen Bond 351
-5.000e-2 5.000e-2
H
… O
F H O F
H H
Fig. 7.19 Isodensity surfaces mapped with electrostatic potentials for HF and H2O molecules,
and the hydrogen bond heterodimer FH…OH2 (Gaussian 09 B3LYP/aug-cc-pVTZ calculation)
0.329 0.345
-0.408 0.408 + -0.658 -0.460 0.377 -0.607
0.329 0.345
Fig. 7.20 Atomic charges fitted to electrostatic potentials of HF, H2O, and FH…OH2 (Gaussian
09 B3LYP/aug-cc-pVTZ calculation)
Atomic charges fitted to electrostatic potentials, before and after hydrogen bond
formation, are shown in Fig. 7.20. It can be seen that the hydrogen atoms of the
water molecule are more positively charged in the hydrogen bond heterodimer than
in the isolated water molecule, and the fluoride atom is more negatively charged in
the hydrogen bond heterodimer than in the isolated hydrogen fluoride molecule.
Thus, the distribution of atomic charges reveals polarization of the donor and
acceptor molecules in the hydrogen bond heterodimer, as we already found by
inspection of the electrostatic potential map. As a result of this polarization, the
hydrogen fluoride and the water moieties are no longer neutral, due to the occur-
rence of charge transfer within the heterodimer, a fact that stresses the partial
covalent character of the hydrogen bond. Only the sum of all the atomic charges in
the whole hydrogen bond heterodimer is zero.
7.8.1 The Water Dimer
Consider the hydrogen bond water dimer, HOH…OH2. Figure 7.21 shows the
potential energy curve as a function of the H…O distance for the Cs configuration,
where the plane of reflection contains the HOH…O nuclei. The relaxed scan of the
H…O distance leads to an energy minimum, whose depth is approximately
352 7 Water
19 kJ mol−1 (Gaussian 09 B3LYP/aug-cc-pVTZ calculation). In the O–H…O

fragment, the O–H bond length (0.970 Å) is slightly longer than the O–H bond
length of an isolated water molecule (0.960 Å), the H…O distance (1.953 Å) is
slightly longer than double the O–H bond length, and the OO distance (2.917 Å) is
slightly shorter than the sum of the O–H and H…O distances (2.923 Å), meaning
that the O–H…O interaction is not exactly linear (the calculated value of the
O–H…O angle is equal to 172.1°). These structural features follow trends similar to
those found for the FH…OH2 heterodimer.
The isodensity surfaces mapped with electrostatic potentials suggest that the
electrostatic attraction between the electrophilic region on the HOH donor molecule
and the nucleophilic region of the oxygen atom on the OH2 acceptor molecule
precedes the formation of the water dimer (Fig. 7.22). The electrophilic and
nucleophilic regions are wider in the hydrogen bond dimer than in the isolated
water molecule, pointing to the polarization of the water molecules in the water
dimer. In addition, the individual water molecules that form the water dimer do not
have distinct isodensity surfaces (Fig. 7.22), and the electrostatic potential in the
internuclear H…O region is close to zero (note the green color), thus pointing to the
partial covalent contribution of the hydrogen bond. In order to confirm this feature
of the hydrogen bond, we now consider the intermolecular vibrations of the water
dimer (Fig. 7.23).
The water dimer has N = 6 atoms, that is, 3N − 6 = 12 vibrational modes. Six of
these are intramolecular vibrations mostly localized in each water molecule, cor-
responding to the antisymmetric and symmetric stretchings and to the bending
mode in each H2O molecule of the dimer. The remaining six vibrational modes are
intermolecular vibrations that comprise the out of plane and in plane bends, the
hydrogen bond stretch, the twist, the wag, and the torsion vibrations (Fig. 7.23). By
their own intermolecular nature, these vibrations extend to both water molecules, a
Fig. 7.21 Relaxed scan of

the H…O distance in the
hydrogen bond water dimer
(Gaussian 09 B3LYP/
aug-cc-pVTZ calculation). E/kJmol - 1
1.954 Å
Mathematica. Molecular 25
representations obtained with

GaussView 20
15
10
0 RHO /
1.5 2.0 2.5 3.0 3.5 4.0
7.8 Hydrogen Bond 353
-5.000e-2 5.000e-2
•••
Fig. 7.22 Isodensity surfaces mapped with electrostatic potentials before and after hydrogen
bond formation for the HOH…OH2 dimer (Gaussian 09 B3LYP/aug-cc-pVTZ calculation)
Fig. 7.23 Intermolecular

vibrational modes of the water
dimer with calculated
(Gaussian 09 B3LYP/
aug-cc-pVTZ calculation) and
experimental (CCCBDB)
frequencies. Calculated
frequencies involve scaling
with the factor 0.968
(CCCBDB)
feature that confirms the covalent character of the hydrogen bond. In particular, the
hydrogen bond stretching at 143 cm−1 (calculated frequency 185 cm−1) is a con-
sequence of the partial covalent character in the O…H interaction.
7.9 Ice Ih
In ice each water molecule has both donor and acceptor hydrogen bond function-
alities, and it is involved in four hydrogen bonds directed to the vertices of a regular
tetrahedron occupied by the nuclei of oxygen atoms of the neighboring water
molecules. In fact, the water molecule at the center of this tetrahedral arrangement
acts as a donor in two hydrogen bonds and as an acceptor in the other two hydrogen
354 7 Water
O"
O"
O"
O"
O"
Fig. 7.24 Pentamer of water molecules in ice Ih. This structure is stable only as part of ice Ih
bonds (Fig. 7.24). Each water molecule at one vertex of the tetrahedral arrangement
also takes the role of a central water molecule in a new tetrahedron not shown in
Fig. 7.24. This tetrahedral pattern leads to the formation of hexagonal channels in
ice Ih (I stands for 1 in roman numerals, and the subscript h stands for hexagonal)
that result from the stacking of hexagonal cycles of hydrogen-bonded water
molecules in the chair configuration (Fig. 7.25). At atmospheric pressure (1 atm =
101.325 kPa), the melting point of ice Ih is 273.15 K.
More than ten different forms of ice are known in distinct p-T regions of the
phase diagram of water. Ice polymorphism (variety of hydrogen bonding archi-
tectures of water molecules in solid phase) results from the flexibility that structures
based on O–H…O hydrogen bond networks have when subject to high pressures
and very low temperatures.
7.9.1 Gas Hydrates
The diversity of solid structures of water is not restricted to ice forms. Under
appropriate conditions of pressure and temperature, small gas molecules like CH4,
NH3, and H2S become trapped inside convex polyhedra formed by
hydrogen-bonded water molecules. A convex polyhedron has no indentations or
holes, that is, every linear segment connecting two points inside a convex poly-
hedron is entirely contained within it. The Swiss mathematician Leonhard Euler
(1707–1783) discovered in 1751 that a convex polyhedron with V vertices, E edges,
and F faces satisfies the equation V – E + F = 2. Oxygen atoms of
hydrogen-bonded water molecules are at the vertices of the polyhedra whose edges
correspond to O–H…O hydrogen bonds The sharing of faces by the convex
polyhedra of hydrogen-bonded water molecules gives rise to solid forms of water
that include small gas molecules and are called gas hydrates. Without the trapped
molecules, the network of hydrogen-bonded water molecules would collapse into
liquid water or a conventional ice form.
Large deposits of methane hydrates are found on the deep ocean floor of con-
tinental edges, at depths of hundreds of meters and pressures of hundreds of bars,
near volcanic fissures from which gases like methane can burst. Methane from
7.9 Ice Ih 355
Fig. 7.25 Left Regular tetrahedron of hydrogen-bonded water molecules in ice seen as a pyramid.
Right Ice fragment showing one hexagonal cycle of hydrogen-bonded water molecules in the chair
configuration. For clarity, some hydrogen atoms covalently bonded to oxygen atoms are omitted
methane hydrate reservoirs doubles the amount of fossil fuel in the world (see Seuss
et al. 1999). However, gas hydrates become unstable at temperatures above 0 °C
and pressures below those found in the submarine deposits where they were
formed.
7.10 Liquid Water
In the liquid state, the hydrogen-bonded network is dynamic, less ordered than in
ice, and the hydrogen bonds around each water molecule are distorted with little
directionality. In addition, the distribution of hydrogen bonds is strongly dependent
on temperature. According to a simple model due to Luzar (see Luzar et al. 1983;
Teixeira and Luzar 1999), it is assumed that in liquid water, no water molecule can
form more than four hydrogen bonds with adjacent water molecules, that is, there
are five groups of water molecules according to the number of hydrogen bonds each
water molecule can form (4, 3, 2, 1, 0). In order to evaluate the population of each
of these groups of water molecules at a defined temperature, we need to express the
fraction of formed hydrogen bonds as a function of temperature. Neglecting
hydrogen bond cooperativeness, each pair of oxygen atoms of adjacent water
molecules can be in one of two states, one for a formed hydrogen bond, the other
for a broken hydrogen bond. According to this model, the ratio of populations of
these two states (broken over formed hydrogen bonds) is given by
ð1 phb Þ e
¼ a exp ð7:23Þ
phb RT
where phb is the probability of one hydrogen bond 1 − phb is the probability of a
broken hydrogen bond, and a and e are empirical parameters whose values a = 178
356 7 Water
and e = 13.8 kJ mol−1 better describe the experimental results (E10). Rearranging
(7.23), we obtain
1 a exp½e=ðRTÞ
phb ¼ 1 phb ¼ ð7:24Þ
1 þ a exp½e=ðRTÞ 1 þ a exp½e=ðRTÞ
The binomial distribution to obtain the fraction of water molecules in each of the
five groups (k takes the values 4, 3, 2, 1, 0) is given by

4 4!
¼ 0k4 ð7:25Þ
k k!ð4 kÞ!
For k = 4, 3, 2, 1, and 0, (7.25) takes the values 1, 4, 6, 4, 1, respectively. For

example, there are six distinct ways [4!/ (2! 2!) = 6] for one water molecule to form
two hydrogen bonds (Fig. 7.26).
Combining (7.24) and (7.25), we can write
4k
4 k 4 fa exp½e=ðRTÞg
phb ð1 phb Þ4k ¼ 0k4 ð7:26Þ
k k f1 þ a exp½e=ðRTÞg4
Figure 7.27 shows the hydrogen bond distributions given by the Luzar model, at
-30 °C (supercooled water), 10 °C, 50 °C and 90 °C, with maxima at k = 4, 3, 2, and
1 hydrogen bonds, respectively. The Mathematica code M5 shows how the dis-
tribution of hydrogen bonds varies with temperature between 273 and 373 K, in
steps of 1 K.
Fig. 7.26 The number of distinct ways for one water molecule to form two hydrogen bonds (solid
lines) is given by 4!/(2! 2!) = 6
- 30 o C 10 o C 50 o C 90 o C
0.5 0.5 0.5 0.5
0.4 0.4 0.4 0.4
0.3 0.3 0.3 0.3
0.2 0.2 0.2 0.2
0.1 0.1 0.1 0.1
0.0 0.0 0.0 0.0
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
Fig. 7.27 Distribution of hydrogen bonds in water at defined temperatures, according to the
Luzar model. This figure was obtained with Mathematica
7.10 Liquid Water 357
The main limitations of the Luzar model are due to the fact that the binomial
distribution is restricted to five groups of water molecules (k takes the discrete
values 4, 3, 2, 1, and 0) and to the neglect of cooperativeness in hydrogen bond
formation (phb does not depend on k). These limitations prevent us from extracting
quantitative conclusions from the above results. Nevertheless, the plots show two
important general trends, namely, the hydrogen bond distribution in liquid water is
strongly dependent on temperature, and as temperature increases, the distribution
maximum moves to lower numbers of hydrogen bonds.
7.11 Phase Diagram
The phase diagram of water shows ice–liquid, liquid–vapor and ice–vapor curves
for biphasic equilibria and the triple point at the intersection of these three curves
(Fig. 7.28). The distinguishing feature is the negative slope of the curve for the ice–
liquid equilibrium, which appears to be infinite. Considering an isothermal trans-
formation that crosses this curve and is carried out in the direction of increasing
pressures, from ice to liquid water, it is shown that ice melts when sufficient
pressure is applied to it. Since increase of pressure always leads to increased
compactness, one concludes that liquid water is denser than ice, that is, ice fluc-
tuates on liquid water, and this is because the hexagonal channels of ice Ih produce
a less dense structure than that of liquid water.
In the liquid–vapor equilibrium, the chemical potentials of the liquid and vapor
are equal,
lliq ðTÞ ¼ lgas ðp; TÞ ð7:27Þ
Fig. 7.28 Phase diagram of water for ordinary pressures (Haynes 2011). The triple point is at
T = 273.16 K and p = 0.611657 kPa. The critical point (not shown) occurs at T = 647.10 K and
p = 22064 kPa. This figure was obtained with Mathematica
358 7 Water
where T is the boiling temperature at pressure p. From (7.27) and (1.149) (this
expression gives the chemical potential of an ideal gas, an assumption that will soon
be justified), we can write

p
l0liq ðTÞ ¼ l0gas ðTÞ þ RT ln 0 ð7:28Þ
p
where p0 is the reference pressure (p0 = 1 bar = 100 kPa). Then,
l0gas ðTÞ l0liq ðTÞ RT gðTÞ ð7:29Þ
where
gðTÞ ¼ ln½p0 =pðTÞ ð7:30Þ
At 298.15 K, p = 3.1699 kPa (Haynes 2011) and

100
gð298:15Þ ¼ ln 3:5 ð7:31Þ
3:1699
At 298.15 K, one has RT 2.5 kJ mol−1, and the difference between the
standard chemical potentials of vapor and liquid water in equilibrium is approxi-
mately 9 kJ mol−1 (2.5 3.5). Considering that the chemical potential is a
partial molar Gibbs energy and the water vapor at such a low pressure
(p = 3.1699 kPa) can be considered an ideal gas, (7.29) provides an estimate of the
Gibbs energy necessary for vaporizing one mole of liquid water at 298.15 K.
The liquid–gas equilibrium curve has a terminal point called the critical point
(not shown in Fig. 7.29) at T = 647.10 K and p = 22,064 kPa. Beyond the tem-
perature of the critical point, the vapor cannot be liquefied by pressure increase. The
critical point can be approached through a transformation that follows the liquid–
Fig. 7.29 Starting with the liquid–vapor equilibrium, an increase of temperature at constant
volume leads the system to the critical point and beyond
7.11 Phase Diagram 359
vapor curve and is carried out at constant volume on a closed system consisting of
liquid water in equilibrium with its vapor. As temperature increases, the pressure
and the density of the vapor increase, while the density of the liquid decreases (the
average density of liquid and vapor inside the sample is constant, since both the
total mass of water and the volume are fixed). When approaching the critical point,
the interfacial surface that distinguishes liquid from vapor gradually disappears, and
an extremely dense vapor is formed called supercritical fluid (Fig. 7.29).
7.12 Water as Solvent
For a solution of one solute, it is possible to group the general types of inter-
molecular interactions as solute–solvent, solvent–solvent, and solute–solute. In a
diluted solution of a nonionic solute, solute–solute interactions can in principle be
neglected, since there is a small probability of two solute molecules being at an
interaction range distance. However, for solutions of ionic compounds, the elec-
trostatic interactions between positively and negatively charged ions may be
important even for diluted solutions, since the potential energy of interaction
between two point charges varies as the long-range function 1/R, where R is the
distance between the charges. For aqueous diluted solutions, the solvent concen-
tration is approximately equal to 55.5 M (1000 g dm−3/18 g mol−1). Considering
a 0.10 M nonionic solution, for each solute molecule there are about 5.5 102
H2O molecules. Solute–water interactions may appreciably affect the hydrogen
bond network and the structure of water in a neighborhood of each solute molecule.
Solubility is the concentration of a saturated solution where the solute in solution
is in equilibrium with the pure solute. The sodium chloride solubility in water at 25
°C is about 6.15 mol of sodium chloride per kilogram of water; that is, a saturated
sodium chloride solution contains nine water molecules for each pair of sodium–
chloride ions. At the same temperature, the solubility of carbon tetrachloride in
water is about 0.0042 mol per kilogram of water, meaning that for each carbon
tetrachloride molecule, there are more than 13,000 water molecules. Mixing carbon
tetrachloride with water leads to distinct phases, and thus one usually says that
carbon tetrachloride and water are immiscible liquids, that is, they do not (signif-
icantly) mix. On a molar basis, sodium chloride is approximately 1500 times more
soluble in water than in carbon tetrachloride.
7.12.1 Electric Permittivity
We are now interested in finding a solvent property that may quantify the solvent
polarity. Being a molecular substance, the solvent is a nonconducting material or
dielectric. When we apply an electric field, charges are slightly displaced from their
equilibrium positions, and, in the liquid phase, the total molecular dipoles (per-
manent plus induced dipoles) tend to become aligned with the applied electric field.
360 7 Water
If the polarization is uniform, the displaced charges cancel in neighboring mole-

cules and so appear only on the surfaces (Fig. 7.30). The induced electric field
opposes the external electric field. Therefore, within the dielectric, the applied
electric field is reduced, due to the internal electric field. The ability of the dielectric
to store charge is called capacitance, denoted by C. For a parallel-plate capacitor,
if the charges on the plates are +q and −q and the electric potential between the
plates is V, the capacitance is given by
C ¼ q=V ð7:32Þ
[the SI unit of capacitance is called the farad (F) after Michael Faraday (1791–
1867); 1 F = 1 C V−1]. The capacitance of a parallel-plate capacitor can be related to
its geometry and the electric nature of the material between plates as
C ¼ e A=d ð7:33Þ
where A is the plate area, d is the plate separation, and e is the permittivity of the
material between the plates. The permittivity of vacuum is a universal constant (eo =
8.854187816 10−12 F m−1). Usually, the permittivity of a dielectric is equated to the
product of the permittivity of vacuum by a dimensionless factor called the relative
permittivity er, where
e ¼ e0 er ð7:34Þ
The potential energy of interaction between two point charges is given by

q1 q 2
E¼ ð7:35Þ
4per
where e is the permittivity of the material in which the charges q1 and q2 are
immersed, and r is the distance between them. Both in (7.33) and (7.35), the
Fig. 7.30 An external electric field induces an electric field inside the dielectric. Inset shows
details of induced dipoles in the molecules of the dielectric
7.12 Water as Solvent 361
Fig. 7.31 Argon and methane solubilities in water given by mole fractions (note the 105 scaling
factor) and the solubility of sodium chloride given as molality, as functions of temperature
(Haynes 2011). This figure was obtained with Mathematica
permittivity can be seen as a scaling factor that measures the polarity of the
medium.
At ambient temperature, the relative permittivity of water is 78.3553. This means
that the potential energy of attraction between sodium and chloride ions in water is
78 times smaller than the potential energy of attraction for the same ions, at the
same distance, in free space. Hence, water is quite efficient for promoting the
dissociation of sodium chloride. In an aqueous sodium chloride solution, the local
electric field created by the Na+ and Cl– ions reorients the neighboring water
molecules around the ions by aligning their dipole moments with the electric fields
generated by the ions. In contrast with water, the relative permittivity of carbon
tetrachloride at ambient temperature is 2.2280. Since electrostatic interactions are
much less important in a solvent of high permittivity than in a solvent of low
permittivity, sodium chloride is very soluble in water but is insoluble in carbon
tetrachloride.
7.13 Simple Nonpolar Solutes
Simple nonpolar solutes like argon and methane have extremely small solubilities
in water that decrease with increasing temperature, in contrast with the large sol-
ubilities of electrolytes like sodium chloride that increase with temperature
(Fig. 7.31). In order to understand these trends, we consider, on the one hand, the
potassium ion–water molecule interactions, which are dominated by electrostatic
interactions between the potassium ion charge and the dipole moment of neigh-
boring water molecules (the electric field produced by the potassium ion aligns the
dipole moments of surrounding water molecules, thus producing a solvated
potassium ion, [K(OH2)x]+), and on the other hand, the absence of charge and any
electric moment in the argon atom, implying the nonexistence of electrostatic
interactions between the argon atom and the water molecules, with these becoming
available to form hydrogen bonds between them (Fig. 7.32). An increase in
362 7 Water
Fig. 7.32 Schematic representation of the solvation in water of a potassium ion and an argon
atom
temperature leads to the disruption of the argon atom cavities by progressively

breaking hydrogen bonds between water molecules on the surfaces of these
cavities.
The argon and methane cavities are reminiscent of the convex polyhedra formed
by hydrogen-bonded water molecules that include gas molecules and are the
building blocks of solid structures known as gas hydrates. Being in the liquid phase,
the mobility of the water molecules on the surfaces of the nonpolar solute cavities is
expected to be large when compared with the reduced mobility of the hydrogen-
bonded water molecules that form the convex polyhedra of gas hydrates.
The extremely low solubilities in water of simple nonpolar solutes like argon and
methane make it possible to look at their aqueous solutions mainly as water subject
to small perturbations on the hydrogen bond network of water (solvent–solvent
interactions) around each solute molecule.
7.13.1 Ostwald Coefficient
Consider now a gas like argon or methane in contact with water. A decrease in the
volume of the gas as a result of its dissolution in water, at a defined temperature and
constant pressure of 1 atm (=101,325 Pa), corresponds to an irreversible process
leading to equilibrium (Fig. 7.33). The ratio of the gas volume change DVg to the
volume of the aqueous solution Vsln, at equilibrium,

DVg
cs ¼ ð7:36Þ
Vsln eq
is called the Ostwald coefficient and can be used to measure the gas solubility in
water.
The Ostwald coefficient can also be defined as the ratio of the molar densities of
the nonpolar solute in the liquid to gas phases at equilibrium,
7.13 Simple Nonpolar Solutes 363
Fig. 7.33 Schematic representation of an irreversible process of dissolution in water of a

nonpolar solute like argon or methane, at a constant pressure of 1 atm (=101,325 Pa) and defined
temperature, leading to a decrease in the volume of the gas phase
!
DVg ns =Vsln qs;sln
cs ¼ ¼ ¼ ð7:37Þ
Vsln eq ns =DVg eq qg
eq
where ns is the amount of the gas solute dissolved in water, and qs,sln is the molar
density of the solute in the liquid phase. At equilibrium, the whole system has two
phases, two independent components, and two degrees of freedom, which are
temperature and pressure. The chemical potentials of the solute in the liquid and in
the gas phases are equal,
ls;sln ¼ lg ð7:38Þ
since the aqueous solution is saturated in the gas. The chemical potential of the gas,
assumed to be ideal (the solute is nonionic and extremely diluted), is

p
lg ¼ l0g ðTÞ þ RTln 0 ð7:39Þ
p
and the gas pressure is
p ¼ qg RT ð7:40Þ
Substitution of (7.40) in (7.39) gives

lg ¼ l0g ðTÞ þ RTln RT=p0 þ RTlnqg ¼ l0g p0 ; T þ RTlnqg ð7:41Þ
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
l0g ðp0 ;T Þ
We now consider the chemical potential of the solute in the liquid phase. The
extremely low solubility of a nonpolar solute like argon or methane in water and, as
a result, the random distribution of its molecules in water lead to an ideal solution.
Hence, we can assume unitary activity coefficients and write
364 7 Water
ls;sln ¼ l0s;sln ðTÞ þ RT ln xs;sln ð7:42Þ
where xs,sln is the solute mole fraction in the liquid phase (the solution). Considering
the extremely low solubility of the nonpolar solute in water, then
ns;sln ns;sln =Vsln qs;sln

xs;sln ¼ ¼ ð7:43Þ
nS nS =Vsln qS
where Vsln is the volume of the liquid phase, and nS and qS are the amount of
solvent and the molar density of the solvent, respectively. Substitution of (7.43) in
(7.42) leads to
ls;sln ¼ l0s;sln ðTÞ RTlnqS þ RTlnqs;sln ¼ l0s;sln ðqS ; T Þ þ RTlnqs;sln ð7:44Þ

|fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
l0s;sln ðqS ;T Þ
At equilibrium, the chemical potentials of the nonpolar solute in the liquid and
gas phases are equal, that is,

l0s;sln ðqS ; TÞ þ RTln qs;sln eq ¼ l0g p0 ; T þ RTln qg eq ð7:45Þ
Rearranging this equation and using the last equality of (7.37), we obtain

l0s;sln ðqS ; T Þ l0g p0 ; T ¼ RTlncs ð7:46Þ
This equation shows that the Ostwald coefficient is a function of temperature.

Considering the physical meaning of the chemical potential as a partial molar Gibbs
energy, then the first member of (7.46) is the Gibbs energy for the transfer of one
mole of the nonpolar gas from the gas to the liquid phase,
DGg!sln ðTÞ ¼ RTlncs ðTÞ ð7:47Þ
where

DGg!sln ðTÞ l0s;sln ðqS ; TÞ l0g p0 ; T ð7:48Þ
Using the Gibbs–Helmholtz equation [see (1.179)], we can write

1
@ DGg!sln DHg!sln
T
¼ ð7:49Þ
@T T2
Therefore, the temperature-dependence of the Ostwald coefficients allows one to

determine the DGg!sln temperature-dependence and to find DHg!sln at each defined
temperature. The corresponding entropy DSg!sln is then obtained from
7.13 Simple Nonpolar Solutes 365
DGg!sln ¼ DHg!sln TDSg!sln ð7:50Þ
At ambient temperature (298.15 K), the approximate values of the thermody-

namic state functions for argon and methane are as given in the following table.
Solute DGg!sln/ kJ mol−1 DHg!sln/ kJ mol−1 DSg!sln/ J mol−1 K−1

Argon 8 –10 –60
Methane 8 –11 –65
Negative entropy values point to the formation of hydrogen bonds involving

water molecules on the surfaces of the solute cavities. As heat is released due to the
formation of these hydrogen bonds, enthalpy values become negative. Note that
DHg!sln and TDSg!sln present approximately parallel temperature dependencies
(Fig. 7.34). Since the negative values of TDSg!sln slightly exceed the negative
values of DHg!sln, the values of DGg!sln become positive and small.
7.13.2 Hydrophobic Interaction
We have been dealing with nonpolar solutes of small molecules that are gases when
pure. However, when the pure nonpolar solute is a liquid and its solubility in water
is exceeded, liquid phase separation occurs between water and the liquid nonpolar
solute. Considering a gradual increase of the nonpolar solute, it is inevitable to
conclude that the liquid nonpolar solute phase results from coalescence of solute
cavities (Fig. 7.35). This is a gradual and progressive irreversible process that
begins well before phase separation with the coalescence of solute cavities, with
Fig. 7.34 Thermodynamic functions for the transfer of one mole of the nonpolar solute from the
gas phase to the ideal solution, in the liquid phase, as functions of temperature [Ostwald
coefficients for argon in water were taken from Ben-Naim and Baer (1963), and those for methane
in water were taken from Yaacobi and Ben-Naim (1973)]. This figure was obtained with
Mathematica
366 7 Water
one solute molecule forming larger cavities with two solute molecules, followed by
the formation of cavities with three solute molecules, four solute molecules, and so
on, until large clusters of solute molecules can eventually form a distinct liquid
phase.
A nonpolar solute in water that leads to liquid phase separation is called an oil.
A typical example of an oil is n-decane, a linear alkane with ten carbon atoms per
molecule. The process of coalescence of solute cavities cannot ignore the relevant
role of water in the formation of hydrogen bonds between water molecules on the
surfaces of the solute cavities (hydrogen bonds mainly oriented as tangents to the
cavity surface), and in the important contribution that the high surface tension of
water gives to the reduction of the surface area during coalescence of solute cav-
ities. In fact, the ratio of surface area to volume of a cavity decreases as solute
cavities coalesce (Fig. 7.35). In the absence of the guest molecule, solute cavities
could not exist. They would become unstable and would collapse. The dispersion
interaction between nonpolar solute molecules in the solute cavities is often called
hydrophobic interaction.
7.14 Ionic Solutions
Due to its high relative permittivity water can easily dissociate strong electrolytes,
yielding solutions that conduct electricity (ionic solutions). Consider the dissocia-
tion of a strong electrolyte
Cn Am ! nCm þ þ mAn ð7:51Þ
whose equation for electric neutrality is given by
nðm þ Þ þ mðnÞ ¼ 0 ð7:52Þ
Fig. 7.35 Coalescence of oil-in-water cavities leads to a decrease in the surface area per oil
molecule and to an increase in dispersion interactions between oil molecules
7.14 Ionic Solutions 367
The individual activity coefficients of ions cannot be experimentally determined,

since the solutions are electrically neutral, i.e., they contain positively charged ions
and negatively charged ions that cannot be separated into different systems.
However, in order to arrive at the chemical potential of the electrolyte (a measurable
thermodynamic quantity), we begin with the chemical potentials of the cation and
anion
lCm þ ¼ l0Cm þ þ RTlnðcCm þ xCm þ Þ lAn ¼ l0An þ RTlnðcAn xAn Þ ð7:53Þ
For each mole of the electrolyte, the mole fractions of the cation and anion [see
(7.51)] are given by
n m
cation: anion: ð7:54Þ
nþm nþm
Applying these mole fractions to (7.53) and adding the resulting equations, we
obtain the chemical potential for the electrolyte, given by
lsalt ¼ l0salt þ RTlnxsalt þ RTlncsalt ð7:55Þ
where
nlCm þ þ mlAn nl0 m þ þ ml0An

lsalt ¼ l0salt ¼ C
n þ m n þm ð7:56Þ
1=ðn þ mÞ 1=ðn þ mÞ
xsalt ¼ xnCm þ xm
An csalt ¼ cnCm þ cm
An
Note that the arithmetic mean of the chemical potentials of the cation and anion
corresponds to the geometric mean of mole fractions and activity coefficients, since
these are arguments of a logarithm function. For sodium chloride, m = 1 and n = 1,
and the mean activity coefficient is given by
csalt ¼ ðcC þ cA Þ1=2 ð7:57Þ
Most thermodynamic data express concentration variables in the molality scale.

For this reason, we now convert mole fractions to molalities. Molality is the amount
of solute per kilogram of solvent. Unlike molarities, molalities are not defined in
terms of a volume of solution and so do not depend on temperature and are not
affected by addition of other solutes to the solution. Starting with the definition of
mole fraction of a solute, we can write, for a diluted solution,
ns ns ns =1 kg solvent ms
xs ¼ ¼ ¼ ð7:58Þ
ns þ nS nS nS =1 kg solvent mS
where ns and nS are the amounts of solute and solvent and ms and mS are the solute
and solvent molalities. For water as solvent of a diluted solution, mS 55.55 mol
368 7 Water
(kg water)−1. Note that 1/mS 0.018 kg mol−1, the molar mass of water expressed
in kilograms (E11).
We have previously considered the chemical potential for a nonideal solution in
terms of the activity of a solute
lðTÞ ¼ l0 ðTÞ þ RT ln as ð7:59Þ
where
as ¼ c s x s ð7:60Þ
Substitution of (7.58) in (7.60) and then in (7.59) and rearrangement of the

resulting equation leads to
lðTÞ ¼ l0 ðT; mS Þ þ RT lnðcs ms Þ ð7:61Þ
where
l0 ðT; mS Þ ¼ l0 ðTÞ RT ln mS ð7:62Þ
In an ionic solution, competition between thermal motion, which tends to ran-

domly distribute ions in solution, and coulombic interaction, which attracts ions of
opposite charge and repels ions with a charge of the same sign, leads to a slight
imbalance that favors the coulombic attraction between ions of opposite charge.
This is because the latter tend to stay around a particular ion longer than ions of the
same charge, thus leading, over time integration, to a spherical charge distribution
called ionic atmosphere (Fig. 7.36). Around a particular anion, the preponderance
of cations over anions (anions are repelled) leads to a positively charged distribu-
tion that is the ionic atmosphere of the anion. In turn, around a particular cation, the
preponderance of anions over cations (cations are repelled) leads to a negatively
charged distribution that is the ionic atmosphere of the cation. Since the net charge
within each ionic atmosphere is nearly zero, every macroscopic volume, no matter
how small it is, is electrically neutral. The distance over which a charge q is
shielded by the charge distribution of ions of opposite charge is called the Debye
length.
The Debye–Hückel theory, formulated in 1923 by Peter Debye (1884–1966) and
Eric Hückel (1896–1980), can evaluate departures from ideal behavior in ionic
solutions. This theory works well for diluted solutions. The Debye–Hückel
limiting law calculates the activity coefficient of an electrolyte at very low con-
centrations in water,
log csalt ¼ Ajzc za jI 1=2 ð7:63Þ

7.14 Ionic Solutions 369
+
+
Fig. 7.36 Schematic representation of motion of ions around a particular cation. Note that anions
are attracted by the cation, whereas cations are repelled. Therefore, anions remain longer than
cations within the ionic atmosphere
where zc and za are the charge numbers of the cation and anion, A = 0.509 for water
(note that this value uses a base-10 logarithm), and I represents the ionic strength
which is given by
1 X 2
I¼ zk mk =m0 ð7:64Þ
2 k
where the summation is over all ions in solution, zk is the charge number of ion k,
mk is its molality, and m0 = 1 mol (kg water)−1 (E12). The following figure presents
the mean activity coefficients of sodium chloride, potassium chloride, and hydrogen
chloride as functions of the solute molality, and shows that only extremely diluted
ionic solutions have activity coefficients equal to 1 (Fig. 7.37).
Fig. 7.37 Mean activity coefficients of sodium chloride, potassium chloride, and hydrogen
chloride as functions of solute molality in aqueous solutions. Data taken from Haynes (2011). This
figure was obtained with Mathematica
370 7 Water
7.15 Amphipathic Molecules
Sodium decanoate, CH3(CH2)8COO– Na+, has a carboxylate group, –COO–, bon-

ded to a linear alkyl group with nine carbon atoms CH3(CH2)8. These molecular
fragments have distinct polarity features that can be illustrated by the electrostatic
potential of the decanoate ion mapped over an isodensity surface (iso-
value = 0.00040; Fig. 7.38). The carboxylate group, a hydrophilic fragment, has a
strongly negative electrostatic potential due to its negative charge equally dis-
tributed over the oxygen atoms. In turn, the alkyl chain, a hydrophobic molecular
fragment, has an electrostatic potential mostly represented by the green color that
corresponds to low or null electrostatic potentials, a feature that is clearly observed
after the third carbon atom, since the first three carbon atoms are under the close
influence of the carboxylate ion. Molecules that exhibit both hydrophilic and
hydrophobic properties are called amphipathic molecules (amphi = both + Greek
pathikos, from pathos = experience).
How do amphipathic molecules organize in water? When a solid formed by
amphipathic molecules is gradually added to water in small amounts, the surface
tension is one of the properties that directly reflects the course of molecular events
in the medium, with the amphipathic molecule having a clear preference for being
adsorbed on the liquid–air and liquid–solid interfaces. In this way, the amphipathic
molecule prevents its hydrophobic fragment from being exposed to water, to which
it directs its hydrophilic group (Fig. 7.39). Therefore, as the concentration of the
amphipathic molecule gradually increases, the surface tension of the mixture
decreases with respect to the surface tension of pure water (71.99 mJ m−2 at 25°C),
and for this reason, the amphipathic solute is generally called a surfactant (see
Vargaftik et al. 1983). In the strict technical sense, a surfactant is a compound that
for a concentration not greater than 0.01 M, lowers the water surface tension by
30 mJ m−2 or more (see Berg 2010). As soon as the liquid–air and liquid–solid
interfaces become totally covered by the surfactant the progressive increase in its
concentration ceases to cause appreciable change on the surface tension of the
medium, and the graph of the surface tension versus the initial concentration of the
surfactant exhibits an approximately horizontal pattern (Fig. 7.39). Since the
Fig. 7.38 Electrostatic potential mapped on isodensity surface with isovalue 0.00040 for the
decanoate ion [Gaussian 09 B3LYP/6-31G(d,p) calculation]. On the left, the isodensity surface
was intentionally made transparent to visualize the molecular geometry
7.15 Amphipathic Molecules 371
surfactant adsorption to the interfaces has reached a maximum, the progressive

increase in the surfactant leads to organized structures inside the medium, as the
amphipathic molecules find a new way of preventing their hydrophobic fragments
from being exposed to water. This time, they form organized structures where the
hydrophilic groups are exposed to water and the hydrophobic chains take shelter
inside organized structures called micelles (Fig. 7.39). The surfactant initial con-
centration at which micelles are formed is called critical micelle concentration,
CMC. The latter corresponds to an energy minimum that results from a balance
between attractive dispersion interactions between the tails of the amphipathic
molecules inside the micelle and the repulsive electrostatic interaction involving the
heads of the amphipathic molecules. A deep energy minimum leads to a defined
critical micelle concentration and a specific aggregation number. Experimental
kinetic studies reveal two kinds of relaxation processes involving micelles: one
involves fast exchange of monomers between micelles and the aqueous phase and
corresponds to a relaxation time of order 10−6 s; the other is the dissociation of
micelles into monomers, which has a relaxation time of order 10−3 s (Oh and Shah
1993).
Let us assume the following equilibrium:
nM1 Mn ð7:65Þ
where M1 represents a monomer and Mn stands for a micelle with aggregation

number equal to n. If C0 is the total concentration of monomers, C1 is the con-
centration of monomers in aqueous phase, Cn is the concentration of monomers in
the micelles and no equilibrium other than (7.65) exists, then
C0 ¼ C1 þ Cn ð7:66Þ
Dividing both members by C0 leads to
Fig. 7.39 As an amphipathic solid is gradually added to water, the amphipathic molecules begin
by covering the interfacial liquid–air and liquid–solid surfaces of the beaker, and the surface
tension of the medium decreases. Once the critical micellar concentration is reached, micelles are
formed, and the surface tension shows an approximately horizontal variation (note that the graph is
merely schematic)
372 7 Water
1 ¼ x1 þ xn ð7:67Þ
where x1 is the mole fraction of the monomers in the bulk solvent (=C1/C0) and xn is
the mole fraction of the monomer in the micelles (=Cn/C0). Under ideal conditions
(negligible monomer–micelle interactions), the equilibrium constant for (7.65) is
given by
ðCn =nÞ
Kn ¼ ð7:68Þ
C1n
where the numerator is the concentration of micelles. Substitution of
Cn ¼ xn C0 C1 ¼ ð1 xn ÞC0 ð7:69Þ
in (7.68) gives
xn
Kn ¼ C 1n ð7:70Þ
nð 1 x n Þ n 0
After taking the 1/n root of both members and rearranging, we obtain
x1=n h i
ð1=n1Þ
xn ¼ 1 C 0 n
¼ f C0 ; x1=n
n ð7:71Þ
ðnKn Þ1=n
This equation cannot be algebraically solved with respect to xn, since this
variable appears in both members. Therefore, it needs to be solved numerically by
iteration, where an initial guess of xn is introduced in the second member of (7.71),
and this member of the equation is solved to give an improved value of xn in the
first member of (7.71). This value is again introduced in the second member,
leading to a new and presumably better value of xn. Then, we can write
h i
1=n
xnði þ 1Þ ¼ f C0 ; xnðiÞ ð7:72Þ
For n = 20 and Kn = 102n, the graph of xn as a function of C0/mM is presented in

Fig. 7.40 (see Mathematica code M6 for details of the iteration process) and shows
that the mole fraction of monomers in the micelles, xn, exhibits an initial abrupt
increase for small values of C0, and for C0 greater than 200 mM, xn enters a
platform with small variations. The Mathematica code M7 shows how to obtain an
estimate of the critical micelle concentration (CMC) by determining C0 through the
interception of tangents to the curve (Fig. 7.40).
Fig. 7.40 Single equilibrium

between n monomers and one
micelle for n = 20 and
Kn = 1040, with the CMC
value approximately equal to
21 mM. This figure was
(see Mathematica code M3)
7.15.1 Sodium Decanoate Micelles
For C0 > CMC, sodium decanoate in water forms spherical micelles with a more or
less defined aggregation number. The methyl protons of the decanoate ion in the
bulk solvent and in the micelles “feel” distinct environments that can be probed
using 1H nuclear magnetic resonance. Nuclear magnetic resonance (NMR) spec-
troscopy measures frequencies and intensities of spin transitions of nuclei with
nonnull spin quantum number. In order to observe a nuclear spin transition, a
magnetic field has to be applied to split the otherwise degenerate spin states. The
chemical shift of a nucleus is given by
ðm mref Þ
d¼ 106 ð7:73Þ
mref
where m is the frequency of the considered nucleus and mref is the frequency of a
reference, that for 1H NMR is usually tetramethylsilane, Si(CH3)4. Since the order
of magnitude of the frequency difference m − mref is 10−6 mref, one usually includes
in (7.73) a factor 106, thus expressing the chemical shift in parts per million (ppm).
The chemical shift of a nucleus is independent of the applied magnetic field, since
both m − mref and mref are proportional to the applied magnetic field.
Consider now the single equilibrium (7.65) between the monomers of the
decanoate ion in the bulk (D2O) and the decanoate ion micelles. Our model system
(see Andrade-Dias et al. 2007) consists of two proton states, one for the methyl
group protons of the decanoate ion in the bulk, the other for the methyl group
protons of the decanoate ion in the micelles. Nuclear magnetic resonance cannot
distinguish the chemical shifts of the methyl group protons in both of these
374 7 Water
environments, due to fast exchange in the NMR time scale between the monomer in
the bulk and in the micelles. Hence, the observed chemical shift is the average of
the chemical shifts of the methyl group protons in both of these environments,
weighted by the corresponding mole fractions (fast exchange regime),
d ¼ x1 d01 þ xn d0n ð7:74Þ
where d 01 is the chemical shift for the methyl protons when x1 = 1, i.e., when all the
monomers are in the bulk solvent, and d0n is the chemical shift for the methyl
protons when xn = 1, i.e., when all the monomers are in the micelles Substitution of
(7.67) in (7.74) and rearrangement of the obtained equation leads to
Dd ¼ xn Dd0 ð7:75Þ
where
Dd ¼ d d01 Dd0 ¼ d0n d01 ð7:76Þ
Substituting the first equality of (7.69) in (7.75) leads to
C0 Dd ¼ Cn Dd0 ð7:77Þ
For C0 < CMC, micelles do not occur and Cn = 0, that is, the first member of
(7.77) is equal to zero. For C0 > CMC, the concentration of the monomers in the
micelles (Cn) increases linearly with C0, since Dd0 is a constant in the experiment.
Figure 7.41 shows C0 Dd versus C0 for sodium decanoate in D2O (D = 2H). It
can be seen that the general pattern predicted by Eq. (7.77) is approximately fol-
lowed. The addition of an ionic compound like sodium fluoride increases the ions
Fig. 7.41 Dd.C0 versus C0/mM for sodium decanoate in deuterated water (left), and with sodium
fluoride ([NaF] = 300 mM) added (right). The intercepts of straight lines with the C0 axis give
approximate values for the critical micelle concentrations (Left, CMC110 mM; Right,
CMC 73 mM). This figure was obtained with Mathematica
of opposite charge in the medium, thus reducing the repulsion between the polar
heads of the amphipathic molecules in the micelles and lowering the critical micelle
concentration.
7.16 Acids and Bases
When ammonia is in contact with hydrogen chloride, a white cloud of ammonium

chloride is obtained that can be deposited as a white solid over a smooth cold
surface. Ammonia is a gas, whose normal boiling point is Tb = −33 °C; hydrogen
chloride is another gas, whose normal boiling point is Tb = −85 °C, and ammo-
nium chloride is an ionic solid formed by NH4+ and Cl− ions, with lattice energy
−712 kJ mol−1 (see Kapustinskii 1956). The chemical reaction
HCl(g) þ NH3 ðgÞ ! NH4þ Cl ðsÞ ð7:78Þ
corresponds to the transfer of a hydrogen ion from one HCl molecule that converts
itself into a Cl- ion to one NH3 molecule that produces an ammonium ion NH4+.
This reaction illustrates the concepts of acid and base that were developed by J.N.
Bronsted (1879–1947, Danish) and T.M. Lowry (1874–1936, British) in 1923: the
acid is the chemical species that donates one hydrogen ion; the base is the chemical
species that accepts one hydrogen ion.
Let us substitute ammonia by water. In contrast to (7.78), the reactants now have
different physical states, since pure hydrogen chloride is a gas and water is a liquid.
The greater density of water “tilts” all molecular events on its side. In a 0.10 M
aqueous solution, for each mole of solute there are about 5.5 102 H2O molecules!
Water has the important role of being the solvent, intervening in all chemical
processes that occur in solution. The chemical equation is
HCl(g) þ H2 O(aq) ! H3 O þ ðaqÞ þ Cl ðaqÞ ð7:79Þ
where HCl acts as a Bronsted–Lowry acid by donating a hydrogen ion to one H2O
molecule and leaving behind a chloride ion, and H2O acts as a Bronsted–Lowry
base by receiving one hydrogen ion to form one hydronium ion, H3O+ (this ion has
a triangular pyramidal geometry). The symbol aq in curved brackets indicates that
the chemical species to which it applies is solvated by water. Since each hydrogen
chloride molecule is surrounded by water molecules, each hydrogen chloride
molecule does not need to migrate to form an acid–base pair.
For aqueous solutions with concentration not exceeding 1 M, scientific studies
have concluded that an acid–base reaction includes H2O molecules for solvation of
the acid and/or the base, that is, H2O molecules mediate the hydrogen ion transfer
between acid and base. When mediated by one H2O molecule, the hydrogen ion
transfer becomes one order of magnitude slower (1 ps) than the unmediated
transfer. In turn, concentrated aqueous solutions present a bimodal mechanism,
376 7 Water
since acid–base pairs also occur in solution, leading to the direct transfer of a
hydrogen ion between acid and base (see Rini et al. 2003; Mohammed et al. 2005).
7.16.1 Autoionization of Water
The water molecule has both acid and base functionalities. The following chemical
equation describes the autoionization equilibrium of water:
H2 O(aq) þ H2 O(aq) H3 O þ ðaqÞ þ OH ðaqÞ ð7:80Þ
From the thermodynamic point of view, every chemical species requires the
definition of a standard state. However, there are no substances formed by ions with
the same charge, since the extremely high electrostatic repulsion between the ions
would prevent the formation of a stable substance. In aqueous solution, it is
established by definition that at all temperatures,
DHf0 ½H þ ðaqÞ 0 S0 ½H þ ðaqÞ 0 DG0f ½H þ ðaqÞ 0 ð7:81Þ
and as a consequence,
Cp ½H þ ðaqÞ ¼ 0 ð7:82Þ
where Cp is the isobaric heat capacity. Since (7.81) and (7.82) are thermodynam-
ically defined for the hydrogen ion H+, thermodynamics represents the autoion-
ization of water by the chemical equation
H2 O(aq) H þ ðaqÞ þ OH ðaqÞ ð7:83Þ
The following table presents the thermodynamic quantities of the chemical

species present in (7.83), at 298.15 K (see Haynes 2011).
T = 298.15 K DH0f /kJ mol−1 DG0f /kJ mol−1 S0/J mol−1 K−1
+
H (aq) 0 0 0
OH–(aq) −230.0 −157.2 −10.9
H2O(L) −285.8 −237.1 70.0
Considering (7.81) and (7.83), we can write

DG0w ðTÞ ¼ l0H þ ðTÞ þ l0OH ðTÞ l0H2 O ðTÞ ¼ RT ln Kw ðTÞ ð7:84Þ
where Kw represents the ionization constant of water,

7.16 Acids and Bases 377
Fig. 7.42 pKw as a function

of temperature, in the range
0–350 °C [for the data used,
see Bandura and Lvov
(2006)]. This figure was
aH þ aOH
Kw ¼ aH þ aOH ð7:85Þ
aH 2 O
where ak is the activity of chemical species k (note that aH2 O 1). At 25 °C, the
constant of autoionization of water is approximately equal to 1.0 10−14. Fig-
ure 7.42 shows the temperature dependence of -logKw, from 0 to 350 °C. The
minimum of this curve occurs at 250 °C, with −logKw = 11.2.
Since H3O+(aq) is equivalent to H+(aq) + H2O(aq) from the stoichiometric point
of view, the activity of the hydronium ion is defined by
aH 3 O þ ¼ aH þ aH 2 O ð7:86Þ
This definition implies that the constant of the equilibrium
H3 O þ H2 O þ H þ ð7:87Þ
is 1, which, in turn, means that H3O+(aq) and H+(aq) + H2O(aq) are also equivalent
from the thermodynamic point of view. Note that the hydronium species H3O+
exists in water as such and as chemical species resulting from solvation with
different numbers of water molecules, such as H5O2+, H7O3+, and H9O4+. On the
other hand, H+ is a proton that really exists only in the transition state of an acid–
base reaction like (7.78). From (7.86) and aH2 O 1, it can be concluded that the
equilibrium constant for the autoionization of water as expressed by (7.80) is
numerically equal to the equilibrium constant for the autoionization of water
obtained from (7.83).
7.16.2 Acid Ionization Constant
Consider now a weak acid HA and the ionization equilibrium in water

378 7 Water
HA(aq) þ H2 OðaqÞ H3 O þ ðaqÞ þ A ðaqÞ ð7:88Þ
This chemical equation is equivalent to
HA(aq) H þ ðaqÞ þ A ðaqÞ ð7:89Þ
whose thermodynamic equilibrium constant is given by

aA aH þ
KHA ¼ ð7:90Þ
aHA
Applying logarithms to both members and multiplying by −1, we obtain

aHA
log KHA ¼ log aH þ þ log ð7:91Þ
aA
The hydrogen ion activity is usually expressed by the solution’s pH, defined by
pH log aH þ ð7:92Þ
In addition, the acid pKHA is defined by
pKHA log KHA ð7:93Þ
For very diluted aqueous solutions, we can assume that the activity coefficients are
approximately equal to 1, and the substitution of (7.92) and (7.93) in (7.91) leads to
xHA
pH pKHA log ð7:94Þ
xA
where the mole fractions satisfy the following equation:
xHA þ xA ¼ 1 ð7:95Þ
ð1 xA Þ
pH pKHA log ð7:96Þ
xA
Rearranging (7.96) and using (7.95), we obtain
1 1
xA xHA ð7:97Þ
1 þ 10ðpKHA pHÞ 1 þ 10ðpHpKHA Þ
Figure 7.43 shows, at left, xHA and xA as functions of pH for pKHA = 4, and xHA
as a function of pH for pKHA = 2, 3, 4, 5, and 6; xHA and xA are S-shaped functions
of pH that sum to 1 and cross at the function value 1/2 when pH = pKHA.
Fig. 7.43 xHA and xA as functions of pH for pKHA = 4, at left, and xHA as a function of pH, for
pKHA = 2, 3, 4, 5, and 6, at right. This figure was obtained with Mathematica
The way a solution resists a change in pH can be evaluated by dxA /dpH.

Figure 7.44 shows, at left, the pH as a function of xA for pKHA = 4 (this illustrates a
universal titration curve since xA is not affected by volume change, as occurs
when the independent variable in a regular titration is the added volume of the base)
and dxA /dpH as a function of xA , at right [note that dxA /dpH is the inverse of the
derivative of pH = f(xA ); see Mathematica code M8]. It can be concluded that the
solution attains maximum buffer capacity at xA = 0.5, when HA and A– have the
same mole fractions.
We now consider (7.90) and convert mole fractions in molalities [see (7.58)],
obtaining
aA aH þ xA xH þ cA cH þ m A m H þ cA cH þ K a cA cH þ

KHA ¼ ¼ ¼ ¼ ð7:98Þ
aHA xHA cHA mHA mS cHA mS cHA
where
mA mH þ
Ka ¼ ð7:99Þ
mHA
Fig. 7.44 pH as a function of xA for pKHA = 4, at left, and dxA /dpH as a function of xA with
maximum at xA ¼ 0:5, at right. This figure was obtained with Mathematica
380 7 Water
is called the acidity constant or acid ionization constant. Chemical species

involved in thermodynamic expressions and in table values usually have their con-
centration expressed in molality. The thermodynamic values for the formation of ions
in water at 298.15 K are referred to a standard state whose concentration is 1 molal.
7.16.3 Lewis Acids and Bases
We now consider chemical reactions that retain the essential characteristics of an

acid-base reaction, yet involve acids that may not contain ionizable hydrogen
atoms. The chemical reaction between ammonia and trimethylborane is one of such
reactions and results in the formation of a compound of formula (CH3)3B–NH3,
ðCH3 Þ3 B þ NH3 ! ðCH3 Þ3 BNH3 ð7:100Þ
Trimethylborane is a gas consisting of B(CH3)3 molecules, with normal boiling

point Tb = −20.2 °C. The boron atom is electron-deficient, i.e., it is an electrophile.
By contrast, the nitrogen atom of ammonia is a nucleophile. The complementary
features of boron and nitrogen atoms cancel out when the compound (CH3)3B-NH3
is formed. In the trimethylborane molecule, the boron atom lies in the plane defined
by the carbon atoms, whereas in the compound (CH3)3B-NH3, the boron atom and
the carbon atoms form a triangular pyramid.
The elementary process of addition of a hydrogen ion to the nitrogen atom of
ammonia,
H þ þ NH3 ! NH4 þ ð7:101Þ
shares with (7.100) the essential features of an acid–base chemical reaction: both
the boron atom of trimethylborane and the hydrogen ion are electrophiles (the
hydrogen ion is an extremely strong electrophile) that bond to ammonia through the
lone electron pair of nitrogen. Hence, (7.100) is an acid–base reaction and
trimethylborane an acid, despite lacking any ionizable hydrogen atom.
Gilbert Lewis (1875–1946), known for his association with the valence electron
pair structures of molecules, captured the similarity between reactions (7.100) and
(7.101) and proposed an acid–base concept that comprises both of these reactions.
According to Lewis, an acid is a chemical species that contains an electrophilic
center and attracts a nonbonding electron pair from the base (the nucleophile), thus
forming a covalent bond between the electrophilic and the nucleophilic atoms or
ions. Hence, (7.100) is a Lewis acid –base reaction, where B(CH3)3 is the Lewis
acid, NH3 is the Lewis base, and (CH3)3B-NH3 is the Lewis complex.
The elementary process of addition of one hydrogen ion (a Lewis acid) to the
oxygen atom of a water molecule (a Lewis base),
H þ þ H2 O ! H3 O þ ð7:102Þ
leads to formation of one hydronium ion (a Lewis complex). This elementary acid–
base reaction is similar to (7.101).
A different example of a Lewis acid–base reaction is given by a carbon dioxide
molecule reacting with the hydroxide ion to form the hydrogenocarbonate ion,
CO2 ðaqÞ þ OH ðaqÞ ! HCO

3 ðaqÞ ð7:103Þ
In this example, carbon dioxide is the Lewis acid, the hydroxide ion is the Lewis
base, and the hydrogenocarbonate ion is the Lewis complex.
The solvation of a metal cation can also illustrate a Lewis acid–base reaction in
which the Lewis acid is the metal cation, the coordinating water molecules are the
Lewis bases, and the obtained aquo-complex is the Lewis complex. For instance,

2 þ
Cu2 þ þ 6H2 O ! CuðH2 OÞ6 ð7:104Þ
Lewis and Bronsted–Lowry acid–base reactions are of different types: while the
first is an addition reaction (the Lewis acid reacts with the Lewis base to form the
Lewis complex), the second is a displacement reaction (the hydrogen ion is
transferred from the Bronsted–Lowry acid to the Bronsted-Lowry base). According
to Lewis nomenclature, the Bronsted–Lowry acid is the Lewis complex of a
hydrogen ion (a Lewis acid) with a Lewis base, and a Bronsted–Lowry acid–base
reaction is a chemical reaction of the following general kind:
complexL þ baseL ! base0L þ complex0L ð7:105Þ
where the subscript L stands for “Lewis.”

The Lewis acid concept is more general than the Bronsted–Lowry acid concept.
Considering the base, there are no differences between the Lewis and the Bronsted–
Lowry concepts: both the Lewis and the Bronsted–Lowry bases are nucleophiles
that have at least one pair of nonbonding electrons in one of their electronegative
atoms (typically, N, O, F). However, while the Bronsted–Lowry acid and the Lewis
acid are both electrophiles, the Lewis acid may not have an ionizable hydrogen
atom.
7.17 Standard Electrode Potentials
Thestandardelectrode potential defined to be zero is the standard hydrogen

electrode potential, which corresponds to the half-reaction
1
H þ ðaqÞ þ e ! H2 ðgÞ ð7:106Þ
2
382 7 Water
where the molality of the hydrogen ion is given by m(H+) = 1 mol (kg of solvent)−1,
and the pressure of molecular hydrogen by p(H2) = 101.325 kPa (Fig. 7.45). The
thermic coefficient for the standard hydrogen electrode potential, deo/dT, is also
zero by definition, meaning that the above definition of the standard hydrogen
electrode potential applies to all temperatures.
The International Union of Pure and Applied Chemistry (IUPAC) proposed a
method for determination of hydrogen ion activity that uses the Harned cell,
1
H2 ðgÞ þ AgCl(s) ! Ag(s) þ H þ ðaqÞ þ Cl ðaqÞ ð7:107Þ
2
(see Covington 2011; Haynes 2011). This cell involves the silver chloride–silver
electrode
AgClðsÞ þ e ! AgðsÞ þ Cl ðaqÞ ð7:108Þ
and the hydrogen electrode (7.106) and does not use any liquid junction. The
method measures the potential difference between two Harned cells designed in
such way that the obtained value does not depend on the standard hydrogen
electrode potential, which is zero by definition, since this electrode potential is
canceled in the potential difference. The measured potential difference depends only
on the hydrogen ion activity of the hydrogen chloride aqueous solution in one of the
cells in which HCl is present at a fixed molality [e.g., m = 0.01 mol (kg of
solvent)−1], and the mean activity coefficient is known (e.g., at 298.15 K, c± =
0.904). The method enables one to determine the hydrogen ion activity by the
solution pH [see (7.92)] at different temperatures.
Half-equations (7.106) and
1
H2 OðaqÞ þ e ! H2 ðgÞ þ OH ðaqÞ ð7:109Þ
2
Fig. 7.45 Gaseous hydrogen

adsorbed on the surface of the
platinum plate of a hydrogen
electrode. Platinum acts as a
catalyst
7.17 Standard Electrode Potentials 383
involve the same oxidation states of hydrogen (+1, 0), in acid and alkaline solu-
tions, respectively. Since the activities of H+ and OH− are interdependent by the
ionization constant of water, the above half-equations are equivalent from the
physical point of view, that is, they correspond to the same electrode potential,
whether it is obtained by half-equation (7.106),
1=2
RT aH2
e¼ ln ð7:110Þ
F aH þ
or by half-equation (7.109),
1=2
RT aOH aH2
e ¼ e0H2 O=H2 ln ð7:111Þ
F aH 2 O
The equality between the second members of these equations leads to
RT
e0H2 O=H2 ¼ ln Kw ð7:112Þ
F
where Kw is the ionization constant of water. At 298.15 K, the standard hydrogen

electrode potential in alkaline solution is e0H2 O=H ¼ 0:828 V (E13).
2
The following table presents several standard electrode potentials at 298.15 K
(25 °C) and a pressure of 101.325 kPa (1 atm).
Standard electrode potentials (Values taken from Haynes 2011)
Cl2/Cl– Ag+/Ag Fe3+/Fe2+ Cu+/Cu Cu2+/Cu AgCl/Ag+Cl– H+/H2 Fe2+/Fe Zn2+/Zn Li+/Li
0
e /V 1.35827 0.7996 0.771 0.521 0.3419 0.22233 0.00000 −0.447 −0.7618 −3.0401
Using (1.188) and (1.237), we can write
DG0 ¼ nFDe0 ð7:113Þ
Taking into consideration the zero of standard electrode potentials (the hydrogen
standard electrode potential), (7.113) can be replaced by
G0 ¼ nFe0 ð7:114Þ
Let us now consider a metal M with oxidation states Mm+ and Mn+ (n > m).
Being an extensive thermodynamic function, the Gibbs energy is additive, and so
we can write
G0n0 ¼ G0nm þ G0m0 ð7:115Þ

384 7 Water
where subscripts n0, nm, and m0 express the reductions n ! 0, n ! m, m ! 0,

respectively. Dividing both members of (7.115) by −F and using (7.114), we obtain
ne0n0 ¼ ðn mÞe0nm þ me0m0 ð7:116Þ
This result allows us to determine any one of these standard potentials, provided
we know the other two. For instance, if (7.116) is applied to the standard potentials
of iron in acidic solution, we can write
3e030 ¼ e032 þ 2e020 ð7:117Þ
(E14, E15).
Let us consider now manganese and its various oxidation states in acidic solu-
tion: Mn(VII) as MnO4–, Mn(VI) as MnO42–, Mn(IV) as MnO2, Mn3+, Mn2+, and
Mn. The Latimer diagram for this element shows the standard electrode potentials
connecting the various oxidation states of the element (Fig. 7.46).
A graph of ne0 as a function of n is called Frost diagram, where each point
corresponds to the pair of values (n, ne0) and the slope between two consecutive
points m and n is given by
ðn mÞe0nm =ðn mÞ ð7:118Þ
Figure 7.47 shows a Frost diagram for manganese, in acidic aqueous solution.
This diagram can be obtained using the standard electrode potentials of the Latimer
diagram of Fig. 7.46, with the values to be plotted on the y-axis being obtained by
multiplying the number of electrons involved in the reduction half-reaction by the
corresponding standard electrode potential.
By the Latimer diagram, it can be seen that the reduction pair Mn3+/Mn2+ has a
standard reduction potential, e0 = 1.5415 V, greater than that of the pair Mn2+/Mn,
e0 = −1.185 V. Therefore, the oxidation–reduction reaction involving Mn3+, Mn2+,
and Mn in acidic solution is
2Mn3 þ þ Mn ! 3Mn2 þ ð7:119Þ
An oxidation-reduction reaction like (7.119) in which two oxidation states of the

same element transform into the intermediate oxidation state is called a compro-
portionation reaction, and the Frost diagram is concave at the intermediate oxidation
Fig. 7.46 Latimer diagram for manganese with its various oxidation states in acidic aqueous
solution. The standard electrode potentials are expressed in volts (V) and were extracted from
Haynes (2011), with those inside parentheses having been calculated. This figure was obtained
with Mathematica
7.17 Standard Electrode Potentials 385
Fig. 7.47 Frost diagram for

manganese in acidic aqueous
solution. This figure was
state (Fig. 7.47). In turn, an oxidation–reduction reaction in which one intermediate

oxidation state of a chemical element transforms into two oxidation states of the same
element is called a disproportionation reaction, and the corresponding Frost diagram
is convex at the intermediate oxidation state (E16, E17, E18).
Mathematica Codes
M1. Representation of the Water Molecule
The following Mathematica code represents the H2O molecule by an intersection of

spheres using the command Sphere[p,r], which represents a sphere of radius r
centered at the point p:
386 7 Water
Suggestion: Modify the code to give specularity to the H2O representation. To

this end, consult Wolfram Documentation in the Help Section of Mathematica. Try
different specular exponents.
M2. Electrostatic Potential Contours for a Dipole
The following Mathematica code generates contour plots in the xy-plane for the
electrostatic potential of a distribution of two charges (−1 and 1), located on the y-
axis, at y = 0.5 and y = −0.5. The code begins by defining the function f[q,p,
r], which gives the electrostatic potential of a static distribution of electric charges
[see (5.13)]. The Mathematica command Lighter[Gray,i] gives the fraction i
of Gray (i=0 corresponds to Gray and i=1 to White), and the Mathematica
command ContourPlot generates the contours of a function in the xy-plane,
which is the electrostatic potential of a dipole:
-
- - -
Suggestion: Try Table[Lighter[Gray,i],{i,0,1,1/6}] with other

colors.
388 7 Water
M3. Interactive Manipulation of Charge
This Mathematica code uses Manipulate to control the absolute value of the
charge in the electrostatic potential of a dipole:
3
- 0.1
2
- 0.4
- 0.2
- 0.3
0 0
-1 0.3
0.4 0.2
-2
0.1
-3
-3 -2 -1 0 1 2 3
Suggestion: Run the code and observe the contour changes as the absolute value
of the charge is increased.
M4. Vector Field Streamlines
We begin by considering a dipole with a charge −1 at (x, y) = (−1, 0) and charge 1

at (x, y) = (1, 0). For this dipole, the electrostatic potential in the xy-plane is given
by
1 1
/ðx; yÞ ¼
r þ r
where
h i1=2 h i1=2
r þ ¼ ðx 1Þ2 þ y2 r ¼ ðx þ 1Þ2 þ y2
Substitution of these equalities in the electrostatic potential gives
1 1 1 1
/ðx; yÞ ¼ ¼h i1=2 h i1=2
r þ r
ðx 1Þ2 þ y2 ðx þ 1Þ2 þ y2
The x- and y-components of the electric field are given by
@/ x1 xþ1
Ex ¼ ¼h i3=2 h i3=2
@x
@/ y y
Ey ¼ ¼h i3=2 h i3=2
@y
For a point dipole with charges −1 and 1 separated by an infinitesimal distance

d, we have
dx dx
/ðdÞ ¼ ¼
r 3 ðx2 þ y2 Þ3=2
and
@/ 3dx2 d @/ 3dxy
Ex ¼ ¼ Ey ¼ ¼
@x ðx2 þ y2 Þ5=2 ðx2 þ y2 Þ3=2 @y ðx2 þ y2 Þ5=2
The above equations are used in the following Mathematica code, where we
have set d = 0.01. The command StreamPlot plots the vector field as a function
of x and y. Observe the difference between the contours for the dipole and the point
dipole:
390 7 Water
4 4
2 2
0 0
-2 -2
-4 -4
-4 -2 0 2 4 -4 -2 0 2 4
Suggestion: Run the above code with different values for d, for example 0.1 and
0.5, and observe the differences.
M5. Luzar’s Model
Assume that in liquid water, there are five types of H2O molecules according to the
number of hydrogen bonds each molecule can form, namely, 0, 1, 2, 3, or 4. This
Mathematica code uses (7.26) to calculate the distribution of hydrogen bonds in
water according to Luzar’s model as temperature varies between 273 and 373 K, in
steps of 1
K. The Mathematica function Binomial[n,k] gives the binomial
n 4
coefficient , and the following line of code calculates for k = 0, 1, 2, 3, 4:
k k
35
30
25
20
15
10
0 1 2 3 4
Suggestion: Run the above Mathematica code and find how the hydrogen bond
network changes with temperature.
M6. Micelle–Monomers Equilibrium
The following Mathematica code shows how to solve (7.71) by the iteration process
given by (7.72). The independent variable represented in the code by x is the total
concentration of monomers C0, and # is the mole fraction of monomers in the
micelles xn that appears in both members of (7.72).
In the code below, the Mathematica function Nest applies three times (three
iterations) to 1-(x*10^-3)^(1./n-1)*(#/(n*10^(2 n)))^(1/n)&. In
this expression [see (7.71)], x in millimolar units is multiplied by 10−3 to convert
into molar units, and the equilibrium constant Kn is made equal to 102n. The result
of the third iteration is given by iteration[[3]]:
392 7 Water
Suggestion: use a pocket calculator to verify 6.09465 in the result for

iteration[[3]].
M7. Critical Micelle Concentration
The following Mathematica code shows how to obtain an estimate for the critical
micelle concentration (CMC 21 mM) through the intersection of tangents to the
xn-versus-C0 curve, at C0 = 10 mM and C0 = 800 mM. The intersection of these
tangents is the point at which h[x] and g[x] (the derivatives of f[x]) are equal.
In order to solve the equation that expresses this equality, the code uses FindRoot
to search a numerical root starting from x=50:
M8. Weak Acid HA
The way in which the solution of a weak acid HA resists a change of pH can be
determined by evaluating dxA =dpH. The following Mathematica code considers
the pH as a function of xA [see (7.96)] for pKHA = 4, and evaluates dxA =dpH
by taking the inverse of the derivative of pH = f(xA ), as a function of xA , at
right. Note that the range of xA values is [0.001, 0.999], since 0 and 1 would give
rise to singularities. It is shown that the solution attains maximum buffer capacity at
xA = 0.5, when HA and A– have equal mole fractions. The code makes use of
Line for representing the dashed line segments joining a sequence of points:
394 7 Water
The following Mathematica code plots xA and dxA =dpH as functions of pH
[see (7.97)] for pKHA = 4:
Glossary
Bronsted–Lowry acid A proton (H+) donor. A Bronsted–Lowry base is a

proton acceptor.
Buffer solution A solution that resists pH changes and, for an
aqueous solution of weak acid HA, attains
maximum buffer capacity at xA ¼ 0:5, when HA
and A– have the same mole fractions.
Comproportionation An oxidation–reduction reaction whereby two
chemical species with different oxidation numbers
in one element form a product in which that same
element has an intermediate oxidation number.
Critical micelle concentration The concentration of a surfactant above which
micelles form.
Debye–Hückel limiting law Enables one to determine the activity coefficient
of an electrolyte at very low concentrations in
water as a function of the ionic strength [see
(7.63)], a concept that was introduced in 1921 by
Lewis (1875–1946) and Randall (1888–1950)
and is a measure of the concentration of ions in a
solution [see (7.64)].
Debye–Hückel theory Was formulated in 1923 by Peter Debye (1884–
1966) and Eric Hückel (1896–1980). It evaluates
departures from ideal behavior in dilute ionic
solutions
Debye length The distance over which a charge in an ionic

solution is shielded by the charge distribution of
ions of opposite charge.
Disproportionation An oxidation–reduction reaction in which a
reactant with one element in a particular
oxidation state is simultaneously oxidized and
reduced, giving two different products.
Lewis acid An electrophile that accepts an electron pair from
a Lewis base to form a Lewis adduct.
Micelle An aggregate of amphipathic molecules with
colloidal dimensions (from 1 to 1000 nm) that in
aqueous solution has the hydrophilic head of the
amphipathic molecules exposed to the solvent
and the hydrophobic tails sequestered in the
micelle center.
Ostwald coefficient The ratio of the gas volume change DVg due to the
irreversible dissolution of the gas in water to the
volume of the aqueous solution Vsln, at
equilibrium [see (7.36) and Fig. 7.33].
Quadrupole Charge distribution with null total charge and null
total dipole that consists of four charges of equal
magnitude, two positive and two negative.
Surfactant Amphipathic molecule that decreases the surface
tension of an aqueous mixture with respect to the
surface tension of pure water (71.99 mJ m−2 at
25 °C). In the strict technical sense, a surfactant is
a compound that for a concentration not greater
than 0.01 M, lowers the water surface tension by
30 mJ m−2 or more.
Exercises
E1. Considering the dipole moment definition, for a static distribution of point
charges, show that the dipole moment vector points to the positive charge.
E2. Show that the dipole moment of a charge distribution whose total charge is
zero is independent of the origin.
E3. Considering the electrostatic potential atomic charges of H2O, qO = −0.680
and qH = 0.340, and the geometric values ROH = 0.9611 Å and aHOH =
104.53°, obtain the molecular dipole moment.
396 7 Water
E4. Use Mathematica to plot the angular functions involved in the d and d3
terms of (7.13). Show that those functions are proportional to the Y01(h,0) and
Y03(h,0) spherical harmonics in the plane.
E5. Use Mathematica to plot the contours of the electrostatic potential of a
quadrupole with the charges (−0.225, −0.225, 0.225, 0.225) at the x,y-coordi-
nates (0., 0.1176), (0., −1.05922), (−1.0, −0.48), and (1., −0.48), respectively.
Compare these contours with those of the quadrupole term for H2O (Fig. 7.7).
E6. At CCCBDB (cccbdb.nist.gov), find the energy of H2O at equilibrium (no
zero-point correction) given by calculations obtained with the B3LYP method
and the STO-3G, 3-21G, 6-31G, 6-311G*, 6-311G**, 6-311+G(3df,2p), 6-311
+G(3df,2pd), cc-pVQZ, aug-cc-pVQZ basis sets. Use Mathematica to plot these
data with the energy in eV. Comment on the obtained graph.
E7. Derive (7.22).
E8. Specify the symmetry species of the normal modes of H2O and DHO.
E9. Chlorine has two natural isotopes, 35Cl and 37Cl, whose natural abundances
are in the approximate ratio of 3:1. The breathing mode (symmetric stretching of
the carbon tetrachloride molecule can be observed in the Raman spectrum of this
liquid. Give an estimate of the relative Raman intensities of the distinct isotopic
lines of the breathing mode of the carbon tetrachloride molecule and the fre-
quency difference in cm−1 between the lines corresponding to the isotopic
combinations (35, 35, 35, 35) and (35, 35, 35, 37). The breathing vibration of
the (35, 35, 35, 35) isotopic combination occurs at approximately 459 cm−1
(CCCBDB).
E10. Consider phb as a function of e, for T = 298.15 K, and assume a = 178
kJ mol−1. Use Mathematica to plot the function phb(e) and its first derivative
with respect to e, and determine the inflection point of phb(e).
E11. Consider a dilute solution of one particular nonpolar solute and convert
mole fraction into molarity. What is the value of the conversion factor when the
solvent is water?
E12. Calculate the mean activity coefficient for a 0.001 molal aqueous solution
of CaCl2.
E13. An oxidizing agent that converts H2O into O2 and a reducing agent that
reduces H2O to H2 are not thermodynamically stable in aqueous solution.
Derive the potentials of the O2/H2O and H2O/H2 reductions as functions of pH
at constant temperature (T = 298.15 K). Assume that pO2 = pH2 = p0 = 1 atm.
Making use of Mathematica, plot the obtained functions and fill with gray the
region of thermodynamic stability for water.
E14. Consider MnO2, Mn3+, and Mn2+ in acidic aqueous solution. Evaluate
e0(MnO2/Mn3+), knowing that e0(MnO2/Mn2+) = 1.224 V and e0(Mn3+/Mn2+)
= 1.5415 V.
E15. Consider MnO4–, MnO42–, and MnO2 in acidic aqueous solution. Evaluate
e0(MnO42–/MnO2), knowing that e0(MnO4–/MnO42–) = 0.558 V and e0(MnO4–/
MnO2) = 1.679 V.
E16. Consider e0(Fe3+/Fe2+) = 0.771 V and e0(Fe2+/Fe) = -0.447 V and an
acidic aqueous solution consisting of Fe(III), Fe(II), and Fe(0). Use
Exercises 397
Mathematica to obtain the Frost diagram for iron in acidic aqueous solution.
How are Fe3+, Fe2+, and Fe(s) related, by disproportionation or compropor-
tionation? Explain.
E17. Consider e0(Cu2+/Cu+) = 0.153 V and e0(Cu+/Cu) = 0.521 V and an acidic
aqueous solution consisting of Cu(II), Cu(I), and Cu(0). Use Mathematica to
obtain the Frost diagram for copper. How are Cu2+, Cu+, and Cu(s) related, by
disproportionation or comproportionation? Explain.
E18. Consider the electrochemical concentration cell Ag|Ag+||Ag+|Ag, and
represent by aa and ac the silver ion activities at the anode and cathode,
respectively. In which electrode, anode or cathode, is the activity of the silver
ion greater? Why? Write the Nernst equation for this cell.
References
Martin JML, de Oliveira G (1999) J Chem Phys 111:1843
Parthiban S, Martin JML (2001) J Chem Phys 114:6014
Aziz RA, Slaman MJ (1991) An examination of ab initio results for the helium potential energy
curve. J Chem Phys 94:8047–8053
Woon DE (1991) Benchmark calculations with correlated molecular wave functions. V. The
determination of accurate ab initio intermolecular potentials for He2, Ne2, and Ar2. J Chem
Phys 100:2838–2850
Seuss E, Bohrmann G, Greinert J, Lausch E (1999) Flammable ice. Sci Am
Luzar A, Svetina S, Zeks B (1983) The contribution of hydrogen bonds to the surface tension of
water. Chem Phys Lett 96:485–490
Teixeira J, Luzar A (1999) Physics of liquid water. Structure and dynamics. In: Bellissent-Funel
MC (ed) Hydration processes in biology, NATO ASI series, Series A: Life sciences, vol 305.
Kluwer, Dordrecht, pp 35–65
Haynes WM (Editor-in-Chief) (2011) Handbook of chemistry and physics on CD-ROM, version
2011. CRC Press, Taylor and Francis Group
Ben-Naim A, Baer S (1963) Method for measuring solubilities of slightly soluble gases in liquids.
Trans. Faraday Soc. 59:2735–2738
Yaacobi M, Ben-Naim A (1973) Hydrophobic interaction in water-ethanol mixtures. J Solution
Chem 2:425–443
Vargaftik NB, Volkov BN, Voljak LD (1983) International tables of the surface tension of water.
J Phys Chem Ref Data 12(3):817–820
Berg JC (2010) An introduction to interfaces and colloids—the bridge to nanoscience. World
Scientific Publishing Co. Pte. Ltd
Oh SG, Shah DO (1993) The effect of micellar lifetime on the rate of solubilization and detergency
in sodium dodecyl sulfate solutions. J Amer Oil Chemists Soc 70:673–678
Andrade-Dias C, Lima S, Teixeira-Dias JJC (2007) From simple amphiphilic to surfactant
behavior: analysis of 1H NMR chemical shift variations. J Colloid Interface Sci 316:31–36
Kapustinskii AF (1956) Lattice energy of ionic crystals. Q Rev Chem Soc 10:283–294
Rini M, Magnes B-Z, Pines E, Nibbering ETJ (2003) Real-time observation of bimodal proton
transfer in acid-base pairs in water. Science 301:349–352
Mohammed OF, Pines D, Dreyer J, Pines E, Nibbering ETJ (2005) Sequential proton transfer
through water bridges in acid-base reactions. Science 310:83–86
Bandura AV, Lvov SN (2006) The ionization constant of water over wide ranges of temperature
and density. J Phys Chem Ref Data 35(1):15–30
398 7 Water
Covington AK (2011) pH scale for aqueous solutions. In: Haynes WM (Editor-in-Chief)

Handbook of chemistry and physics on CD-ROM, version 2011. CRC Press, Taylor and
Francis Group
Further Reading
Levine IN (2013) Quantum chemistry, 7th edn. Pearson Education Inc.

Lynden-Bell RM, Morris SC, Barrow JD, Finney JL, Harper CL Jr (2010) Water and life: the
unique properties of H2O. CRC Press, Taylor & Francis Group
Atkins P, Paula J (2006) Physical chemistry for the life sciences. Oxford University Press and
W.H. Freeman and Company
Kondepudi D (2008) Introduction to modern thermodynamics. Wiley, New York
Appendix
Physical Constants
Constant Symbol SI value

Speed of light in vacuum c 299,792,458 m s−1
Elementary charge e 1.602176565 10−19 C
Permittivity of vacuum e0 8.854187817 10−12 C V−1 m−1
Planck constant h 6.62606957 10−34 J s
Electron mass me 9.10938291 10−31 kg
Proton mass mp 1.672621777 10−27 kg
Neutron mass mn 1.674927351 10−27 kg
Deuteron mass md 3.34358348 10−27 kg
Avogadro constant NA 6.02214129 1023 mol−1
Faraday constant F =eNA 96,485.3365 C mol−1
Boltzmann constant kB 1.3806488 10−23 J K−1
Gas constant R =kBNA 8.3144621 J mol−1 K−1
Atomic mass constant mu 1.660538921 10−27 kg
Hartree energy Eh 4.35974434 10−18 J
Bohr radius a0 =e0h2/(pmee2) 0.52917721092 10−10 m
Electron volt eV 1.602176565 10−19 J
P.J. Mohr, B.N. Taylor, and D.B. Newell (2011), “The 2010 CODATA
Recommended Values of the Fundamental Physical Constants” (Web Version
6.0). This database was developed by J. Baker, M. Douma, and S. Kotochigova.
Available: http://physics.nist.gov/constants. National Institute of Standards and
Technology, Gaithersburg, MD 20899. Alternatively, each of these physical con-
stants can be obtained from the computational knowledge engine Wolfram Alpha at
www.woframalpha.com.

DOI 10.1007/978-3-319-41093-7
400 Appendix
Conversion Factors
1 hartree = 27.211385 eV = 2625.4996 kJ mol−1

1 eV = 96.4853365 kJ mol−1
1 cm−1 = 0.011962657 kJ mol−1
1 cal = 4.184 J
1 eÅ = 4.8032 D
Atomic Units
Action h/(2p) 1.054571726 10−34 J s

Charge e 1.602176565 10−19 C
Length a0 0.52917721092 10−10 m
Mass me 9.10938291 10−31 kg
Electric constant 4pe0 1.112650056 10−10 C V−1 m−1
Frequently Used Prefixes
Factor Name Symbol Factor Name Symbol

18 −3
10 exa E 10 milli m
1015 peta P 10−6 micro l
1012 tera T 10−9 nano n
109 giga G 10−12 pico p
106 mega M 10−15 femto f
103 kilo k 10−18 atto a
Greek Alphabet
Alpha A a Iota I i Rho P q

Beta B b Kappa K j Sigma R r
Gamma C c Lambda K k Tau T s
Delta D d Mu M l Upsilon ! t
Epsilon E e Nu N m Phi U /
Zeta Z f Xi N n Chi X v
Eta H η Omicron O o Psi W w
Theta H h Pi P p Omega X x
Appendix 401
Integrals
Mathematica instruction Output
Integrate[1/x,x] Log[x]
Integrate[x^n,x] x^(1+n)/(1+n)
Integrate[Sin[x],x] -Cos[x]
Integrate[Cos[x],x] Sin[x]
Integrate[Tan[x],x] -Log[Cos[x]]
Integrate[Exp[-x^2],x] 1/2 Sqrt[p] Erf[x]
Integrate[Exp[-b x^2],{x,0,Infinity}, Sqrt[p]/(2 Sqrt[b])
Assumptions!b>0]
Extensive mathematical data can be obtained from Wolfram Alpha at

www.woframalpha.com.
Character Tables of Point Groups

Available at http://symmetry.jacobs-university.de
Answers to Exercises
Chapter 1—Thermodynamics
E1. Mathematica Code and Results

DOI 10.1007/978-3-319-41093-7
404 Answers to Exercises

Answers to Exercises 405
E7.
4x2 y3 dx þ 3x3 y2 dy ¼ gðx; yÞdx þ hðx; yÞdy
¼ 12x2 y2 ¼ 9x2 y2
@y @x
6¼
@y @x
Therefore, 4x2y3dx + 3x3y2dy is an inexact differential.
E8.
@p @V
pdV þ Vdp ¼ dðpVÞ ¼1 ¼1
@p @V
Therefore, pdV +Vdp is an exact differential.
E9.
dU ¼ dQ þ dW

@U @U
dU ¼ dT þ dV dW ¼ pdV
@T V @V T

@U @U @U @U
dT þ dV ¼ dQ pdV dQ ¼ dT þ p þ dV
@T V @V T @T V @V T

@U @V
dQp ¼ Cp dTp CV ¼ dVp ¼ dTp
@T V @T p

@U @U
dQp ¼ dTp þ p þ dVp
@T V @V
T
@U @V
Cp CV ¼ p þ
@V T @T p
E10.
dQ CdT
dS ¼ dS ¼
T T
Z373 Z373
CdT dT 373
DS ¼ ¼C ¼ 4 25:09 ln ¼ 31:3 J K1
T T 273
273 273
E11.
C3 H8ðgÞ þ 5O2ðgÞ ! 3CO2ðgÞ þ 4H2 Oð‘Þ

DH0r ¼ 4DHf0;H2 O‘ þ 3DHf0;CO2ðgÞ DHf0;C3 H8ðgÞ
¼ 4ð285:8Þ þ 3ð393:5Þ ð104:7Þ ¼ 2219 kJ mol1
E12.
E13.
V ¼ V0 þ V0 at
t ¼ 0 C ) V ¼ V0

@V 1 @V
¼ V0 a ) a ¼
@t p V0 @t p
E14.
In the temperature–entropy plot of the Carnot cycle, horizontal lines are

isotherms (T = constant), and vertical lines are adiabatic transformations (Q = 0, that
is, S = constant). See §4 to find out that DSAB ¼ QAB = hh ¼ WAB = hh ¼
R ln ðVB = VA Þ, where DSAB > 0, that is, SB >SA, because VB >VA.
E15.
2NO2ðgÞ N2 O4ðgÞ DG0r ¼ DG0f ;N2 O4ðgÞ 2DG0f ;NO2ðgÞ

¼ 99:8 2 51:3 ¼ 2:8kJmol1
DG0r ¼ RT ln Ke
2:8 103 ¼ 8:314462 298:15 ln Ke Ke ¼ 3:09
x x 2 2x
xN 2 O 4 ¼ ¼ xNO2 ¼
2 2x þ x 2 x 2x
xN2 O4 xð2 xÞ
Ke ¼ 2 ¼ ¼ 3:09 ) x ¼ 0:7266
xNO2 ð2 2xÞ2
xN2 O4 ¼ 0:5706 xNO2 ¼ 0:4294
E16.
dla ¼ Sm;a dT þ Vm;a dp dlb ¼ Sm;b dT þ Vm;b dp
at equilibrium: la ¼ lb dla ¼ dlb
dp DSm DHm
Sm;a dT þ Vm;a dp ¼ Sm;b dT þ Vm;b dp ) ¼ ¼
dT DVm TDVm
dp DHm DHm DHm p dp DHm
¼ ) dT
dT TðVm;g Vm;‘ Þ TVm;g RT 2 p RT 2
E17.
P ¼ 4ð3 solid phases þ 1 gas phaseÞ

CðsÞ þ O2ðgÞ CO2ðgÞ 4FeðsÞ þ 3O2ðgÞ 2Fe2 O3ðsÞ
Ci ¼ 5ðcomponentsÞ 2ðchemical equilibria) = 3
P þ f ¼ Ci þ 2 4þf ¼3þ2 f ¼1
variables: T; pO2 ðgÞ ; pCO2 ðgÞ ; pt
equations: pt ¼ pO2 ðgÞ þ pCO2 ðgÞ pCO2 ðgÞ =pO2 ðgÞ ¼ K1 ðTÞ pO2 ðgÞ ¼ K2 ðTÞ
Therefore, temperature is the degree of freedom for this system.

E18.
Applying the first and second laws to an isolated system leads to
dU ¼ Tde S pdV dU ¼ 0; dV ¼ 0; de S ¼ 0 di S 0
Therefore, S attains a maximum at the end of the irreversible process, when
equilibrium is reached.
E19.
RT aX
X nþ þ ne ! X e ¼ e0 ln
nF aXnþ

1 nþ 1 RT aX 1=n 0 RT aX
X þ e ! X e ¼ e0 ln ¼e ln
n n F aX nþ nF aX nþ
E20.
RT
2Agþ þ Cu ! 2Ag þ Cu2þ De0 ¼ lnKe logKe 16:9 nDe0
nF
Agþ þ e ! Ag e0Agþ = Ag ¼ 0:7996 V
Cu2þ þ 2e ! Cu e0Cu2þ =Cu ¼ 0:3419 V
0
De0 ¼ 0:7996 0:3419 ¼ 0:4577 V Ke 1016:9 nDe
¼ 1016:920:4577 ¼ 1015:5 ¼ 3:2 1015
Standard electrode potential values taken from Handbook of Chemistry and
Physics, 2011.
Chapter 2—Chemical Kinetics

ln0:05 ln0:05
½A ¼ ½A0 ekt 0:05½A0 ¼ ½A0 ekt t¼ ¼ ¼ 8:7 min
k 0:345
E2.
aA þ bB ! cC þ dD
dnA dnB dnC dnD
¼ ¼ ¼ ¼ dn
a b c d
½A ¼ nA =V ½B ¼ nB =V ½C ¼ nC =V ½D ¼ nD =V
d½A 1 dnA d½B 1 dnB d½C 1 dnC d½D 1 dnD
¼ ¼ ¼ ¼
dt V dt dt V dt dt V dt dt V dt
1 d½A 1 d½B 1 d½C 1 d½D 1 dn
¼ ¼ ¼ ¼
a dt b dt c dt d dt V dt

E4.
A þ BC
[A] ¼ [A]eq d [B] ¼ [B]eq d [C] ¼ [C]eq þ d
dd
¼ k! [A][B] k [C] ¼ k! [A]eq d [B]eq d k [C]eq þ d
dt
dd
¼ k! [A]eq [B]eq k [C]eq k! [A]eq + [B]eq þ k d þ k ! d2

dt |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |{z}
a¼0 b v
E5.
(a)
dX dY
¼ k1 A k2 BX þ k3 X 2 Y k4 X ¼ Z1 ðX; YÞ ¼ k2 BX k3 X 2 Y ¼ Z2 ðX; YÞ
dt dt
(
0 ¼ k1 A k2 BXs þ k3 Xs2 Ys k4 Xs Xs ¼ kk14A
0 ¼ k2 BXs k3 Xs2 Ys Ys ¼ kk21 kk43 BA
(b)
@Z @Z !
@Z
1
@X s @Z
1
@Y s k2 B k4 þ 2k3 Xs Ys k3 Xs2
¼
2
@X s
2
@Y s
k2 B 2k3 Xs Ys k3 Xs2

k2 B k4 k3 Xs2 B1 1
¼ ¼
k2 B k3 Xs2 B 1
(c) Mathematica code and results

E6.
(a)
dX dY
¼ k1 AX k2 XY ¼ Z1 (X; Y) ¼ k2 XY k3 Y ¼ Z2 (X; Y)
dt ( dt

0 ¼ k1 AXs k2 Xs Ys Xs ¼ kk32
0 ¼ k2 Xs Ys k3 Ys Ys ¼ kk1 A
2
(b)
@Z @Z !
@Z
1
@X s @Z
1
@Y s k1 A k2 Y s k2 Xs 0 k3 0 1
¼ ¼ ¼
2
@X s @Y
2 k2 Ys k2 Xs k3 k1 A 0 1 0
s
(c) Mathematica code and results
(d) Mathematica code and results

Chapter 3—The Schrödinger Equation
E1.
Ax ¼ kx (A kI)x ¼ 0 det(A kI) = jA kI j ¼ 0
where I is the identity matrix (all its elements are equal to zero except those along
the main diagonal, which equal one), k represents the eigenvalues, and x the
corresponding eigenvectors.

1 1 1 k 1
A¼ ¼ 0 k2 k 6 ¼ 0 k1 ¼ 3 k2 ¼ 2
4 2 4 2 k

1 1 x1 3x1 a
k1 ¼ 3 ¼ x2 ¼ 4x1 eigenvector
4 2 x2 3x2 4a

1 1 x1 2x1 b
k2 ¼ 2 ¼ x2 ¼ x1 eigenvector
4 2 x2 2x2 b
where a and b are constants.
E3.
@ @f @g @ @f
ðf þ gÞ ¼ þ ðcf Þ ¼ c
@x @x @x @x @x
E4.

@ @ @f @f @f @
; x f ¼ ðxf Þ x ¼ f þ x x ¼ f ) ;x ¼ 1
@x @x @x @x @x @x
E5.

h @ p2x 1 h @ 2 h2 @ 2
^px ¼ ¼ ¼
i @x 2m 2m i @x 2m @x2
E6.
@s @s
uðx vtÞ ¼ uðsÞ s ¼ x vt ¼1 ¼ v
@x @t
@u @u @s @u @ 2 u @ 2 u
¼ ¼ ¼
@x @s @x @s @x2 @s2
@u @u @s @u @ 2 u @ v @u
@s 2@ u
2
@2u 1 @2u
¼ ¼ v ¼ v ¼ v ¼
@t @s @t @s @t2 @s @s2 @x2 v2 @t2
If the argument x of the function u(x) is changed by x −a, where a is a positive
number, then the function u(x − a) is displaced to the right of u(x) by a. If we let
a =vt, where v is a positive number and t is time, then the function displacement is
continuous and increases with time.
E7.
@s @s
uðsÞ ¼ Acos(kx xtÞ s = kx xt ¼k ¼ x
@x @t
@u @u @2u 2@ u
2
¼ k ¼ kAsin(kx xtÞ ¼ k ¼ k2 Acos(kx xtÞ
@x @s @x2 @s2
@u @u @2u 2@ u
2
¼ ðxÞ ¼ xAsin(kx xtÞ ¼ x ¼ x2 Acos(kx xtÞ
@t @s @t2 @s2
1 @2u 1 @2u 1 2 @2u 1 @2u
¼ 2 2 k ¼ 2 x2 2
k @x
2 2 x @t k 2 @s 2 x @s

E11.
Z Z
b
w Awds ¼ b
w Aw ds w ¼ f þ cg
Z Z h i
b b þ cgÞ ds
ðf þ cgÞ Aðf þ cgÞds ¼ ðf þ cgÞ Aðf
Z Z Z Z
b ds þ c c g Agds
f Af b b ds þ c f Agds
þ c g Af b
Z Z Z Z
b
¼ f Af ds þ c c g Ag ds þ cb b b
f Ag ds þ c g Af ds
Since Â is Hermitian, the first terms on the left- and right-hand sides of this equation
are equal, and so are the second terms. Therefore,
Z Z Z Z

c b ds
g Af b
f Ag ds ¼ c b
g Af b
ds f Agds
Since c is an arbitrary complex number, the factors in square brackets should be

zero. Therefore,
Z Z Z Z
b b b b
g Af ds ¼ f Ag ds f Agds ¼ g Af ds
E12.
RM ¼ rA mA þ rB mB RM ¼ rA mA þ rB mB
rmA ¼ rA mA þ rB mA rmB ¼ rA mB rB mB
mA mB
rB ¼ R þ r rA ¼ R r
M M
2 2 2
1 drA 1 drB 1 dR 1 dr 2
mA þ mB ¼ M þ l
2 dt 2 dt 2 dt 2 dt
E13.
1 1 1 1 1
¼ þ þ þ
l mA mB mC mD
E14.
dx d2 x
x ¼ A sinðxt + bÞ ¼ xA cosðxt + bÞ 2
¼ x2 A sinðxt + bÞ
rffiffiffiffi dt dt
k d2 x
x= k ¼ mx2 m 2 ¼ mx2 A sinðxt + bÞ = kx
m dt

E17.
Hnþ1 ðzÞ ¼ 2zHn ðzÞ 2nHn1 ðzÞ n¼3 H4 ðzÞ ¼ 2zH3 ðzÞ 2 3H2 ðzÞ
H2 ðzÞ ¼4z 2
2
H3 ðzÞ ¼ 8z 12z H4 ðzÞ ¼ 16z4 48z2 þ 12
3

16z4 48z2 + 12 ¼ 2z 8z3 12z 6 4z2 2 ¼ 16z4 48z2 þ 12
E18.

2
UðRÞ ¼ D 1 ebðRRe Þ e Þ ¼ 0
UðR
dU ¼ 2Db ebðRRe Þ e2bðRRe Þ
dU ¼0
dR dR qffiffiffiffiffi
2 R¼Re
d2 U ¼ 2Db2 ebðRRe Þ þ 2e2bðRRe Þ
dU ¼ 2Db2
¼ k b ¼ k
dR 2 dR 2
qffiffiffiffiffi
2D
3 R¼Re
d3 U ¼ 2Db3 ebðRRe Þ 4e2bðRRe Þ
dU ¼ 6Db3 ¼ 3k 2D k
dR 3 d4R R¼Re
3
d4 U ¼ 2Db4 ebðRRe Þ þ 8e2bðRRe Þ

dU ¼ 14Db4 ¼ 7k
2
dR 4 dR 4R¼Re 2D
E19.

i j k

L¼ x y z

px py pz

y z z x x y
L¼Lx i + Ly j + Lz k ¼ i þ j + k
p y p p
z z px px p
y
Lx ¼ y pz z py Ly ¼ z px x pz Lz ¼ x py y px
E20.
Let f be an arbitrary wave function.

Lx b
b L y f ¼ y^pz z^py ðz^px x^pz Þf ¼ y^pz ðz^px Þ yx^p2z z2 ^py ^px þ zx^py ^pz f

h
¼ y^px þ yz^pz ^px yx^p2z z2 ^py ^px þ zx^py ^pz f
i

b
Ly b
L x f ¼ ðz^ pz Þ y^pz z^py f ¼ zy^px ^pz z2 ^
px x^ px ^py xy^p2z þ x^pz z^py f

h
^ ^ ^
¼ zy^px pz z px py xy^pz þ x^py þ xz^pz py f
2 2
^
i
Mixed partial second derivatives are equal, for example, @ 2 f =@x@y ¼

@ f =@y@x. Thus,
2
h h h i
b
Lx b
Ly b L x f ¼ y^px x^py f ¼ b
Ly b L z f ¼ ih b
Lzf b
Lx; b
L y ¼ ih b
Lz
i i
E21.
h i h i h i h i
b
L2; b
Lx ¼ bL 2x þ b
L 2y þ b
L 2z ; b
Lx ¼ bL 2y ; b
Lx þ bL 2z ; b
Lx
by b
¼L Ly Lbx Lbx L by þ b
by L Lz b
Lz b
Lx Lbx b
Lz b
Lz

¼ bLy b
Ly bLx b
Ly bLx b
Ly þ b Ly b
Lx b
Ly bLx b
Ly b
Ly

þ b Lz b
Lz b
Lx b
Lz b
Lx bLz þ b Lz bLx b Lx b
Lz b Lz b
Lz
h i h i h i h i
¼b
Ly b Ly; b
Lx þ b Ly; bLx b Lz b
Ly þ b Lz; b
Lx þ b Lx b
Lz; b Lz
Ly b
¼ ih b Lz b
L z ih b Lz b
L y þ ih b Ly b
L y þ ih b Lz
¼0
When the parameter n increases, more points are sampled, and the function
changes more rapidly.

E24.

b @2 @ 1 @2 b @
L 2 ¼
h2 þ coth þ L z ¼ ih
@h2 @h sin2 h @/2 @/
b ¼
H
h2 2 @2 2 @
r þ VðrÞ r2 ¼ 2 þ
1 b2
L
2l @r r @r r 2 h2
h i h2
h2
h i
b; b
H L 2 ¼ r2 þ VðrÞ; b
L 2 ¼ r2 ; b L 2 þ VðrÞ; b
L2 ¼ 0 þ 0 ¼ 0
2l 2l
h i h2
h2 2 b
h i
b b
H ; L z ¼ r þ VðrÞ; L z ¼ r ; L z þ VðrÞ; b
2 b Lz ¼ 0 þ 0 ¼ 0
2l 2l
E25.
The square of the radial factor is a probability density, thus having units of inverse
of volume, that is, a−3
0 , where a0 = 0.52917721092 10
−10
m is the atomic unit
of length. Therefore, all radial factors come with the factor a−3/2 . In SI units,
3=2 0
we have R2s ¼ 2pffiffi2 2 a0 e

a0 r = ð2a Þ
r 0
. In atomic units, a0 = 1 and R2s ¼
1 ffiffi
p
2 2
ð2 r Þer = 2 .
E27.
R
/ Hb / ds K Xn
W½/ ¼ R ¼ /¼ cj fj
/ / ds D j¼1

K @K @D
@W @ D @cj K @cj
@K @D
@cj W @cj @K @D
D
¼0 ¼ ¼ ¼0 W ¼0
@cj @cj D2 D @cj @cj
0 1 0 1 0 1
a b c r s t arþbuþcy asþbvþcz atþcwþbx
@d e fA @u v x A ¼ @ drþeuþfy dsþevþfz dtþfwþex A
g h i y z w grþhuþiy gsþhvþiz gtþiwþhx
E30.

ðH11 WS11 Þ ðH12 WS12 Þ ðH11 WS11 Þ ðH12 WS12 Þ

ðH WS Þ ðH WS Þ ¼ 0 ðH WS Þ ¼ ðH WS Þ
12 12 22 22 12 12 22 22

ðH11 WS11 Þc1 þ ðH12 WS12 Þc2 ¼ 0
ðH12 WS12 Þc1 þ ðH22 WS22 Þc2 ¼ 0
8 8
> ðH11 WS11 Þ c1 þ c2 ¼ 0 >
> > c2 ¼ ðH11 WS11 Þ
< ðH WS Þ <c ðH12 WS12 Þ
12 12 1
>
> ðH WS12 Þ > ðH WS12 Þ
c1 þ c2 ¼ 0 >
c
: 12 : 2 ¼ 12
ðH22 WS22 Þ c1 ðH22 WS22 Þ
E31.

n Z 2
X
n X gj
V1 ¼ V1j ¼ dvj
j¼2 j¼2
r1j
E32.
Because equation (3.154) applies to molecules. A molecule has more than one
nucleus, and VNN represents the nucleus–nucleus repulsion energy. This energy
term does not exist for atoms, because they have one nucleus.
E33.
When applied to atoms, (3.154) does not include the energy term VNN. Note that the
sums in (3.154) are over doubly occupied orbitals. From the definitions of Coulomb
and exchange integrals, (3.155) and (3.156), and the fact that electron labels 1 and 2
are dummy variables, it can be inferred that Jij =Jji andKij =Kji. In addition, Kii =Jii.
Therefore, applying (3.154) to the 1s22s2 electron configuration leads to
EHF = 2(e1s + e2s ) J1s1s J2s2s 2ð2J1s2s K1s2s Þ
E34.
ZZ
wi ð1Þwj ð1Þwj ð2Þwi ð2Þ
Kij ¼ ds1 ds2
r12
If wi ¼ /i a and wj ¼ /j b
ZZ
wi ð1Þwj ð1Þwj ð2Þwi ð2Þ
Kij ¼ ds1 ds2
r12
ZZ Z Z
/i ð1Þ/j ð1Þ/j ð2Þ/i ð2Þ
¼ dv1 dv2 a ð1Þbð1Þdr1 b ð2Það2Þdr2 ¼ 0

r12
due to integration over the spin variables.
E35.

1 1
Jij ¼ /i ð1Þ/j ð2Þ /i ð1Þ/j ð2Þ Kij ¼ /i ð1Þ/j ð2Þ /j ð1Þ/i ð2Þ
r12 r12
E37.
The square of the absolute value of a Slater-type orbital is a probability density,

thus having units of the inverse of volume, that is, a−3
0 . Therefore, all radial factors
come with the factor a−3/2
0 . In SI units,
n1
r ð2fÞnþ1=2 1
Nn;f efr=a0 where Nn;f ¼
a0 ½ð2nÞ!1=2 a03=2
In atomic units, a0 = 1, and the radial factor of a Slater-type orbital is given by
ð2fÞnþ1=2
Nn;f r n1 efr where Nn;f ¼
½ð2nÞ!1=2
E38.
pffiffiffi
W ¼ /(1)/ð2Þ½a(1)b(2) a(2)bð1Þ/ 2
Z
q0 ð1Þ ¼ W W dv2 dr1 dr2
Z Z
1
¼ / ð1Þ/ ð2Þ/ð1Þ/ð2Þdv2 ½a ð1Þb ð2Þ
2
a ð2Þb ð1Þ½að1Þbð2Þ að2Þbð1Þdr1 dr2
1
¼ / ð1Þ/ð1Þ ð1 0 0 þ 1Þ ¼ j/j2
2
E39.
pffiffiffi
W ¼/(1)/(2)[a(1)b(2) a(2)b(1)]/ 2
Z h i Z
b b b ð1Þ/ð1Þdv1
W T ð1Þ þ T ð2Þ Wdv1 dv2 dr1 dr2 ¼ 2 / ð1Þ T
E40.
^0 ¼ r 1 @2 2@
2
1 ^2 ^2 ^ 0 ¼ z
H r2 ¼ þ 2L L Y‘m ¼ ‘ð‘ þ 1ÞY‘m H
2 r @r 2 r @r r
1 1
R1s ¼ 2er Y0;0 ¼ pffiffiffi 1s ¼ pffiffiffi er 2pz ðSTOÞ ¼ R2f Y1;0
2 p p
! rffiffiffi !
5=2
2f 1 3 1
¼ pffiffiffi refr cos h ¼ pffiffiffi f5=2 refr cos h
3 2 p p
Z1 Zp Z2p
f5=2 ðf þ 1Þr 4
z ¼ r cos h h1sjzj2pz i ¼ e r dr 2
cos h sin hdh d/
p
0 0 0
Z1 pffiffiffiffiffi
4 32 f5
¼ f5=2 eðf þ 1Þr r4 dr ¼
3 ð1 þ fÞ5
0

r2 1 r2 1

1s 1s ¼ 0:5
R2f R2f
2 r 2 r
Z1
2f5 1
¼ e2fr ðf2 r2 4fr þ 2rÞdr ¼ ð2 3fÞf3
3 6
0
0 0
¼ h1sjzj1si ¼ 0 J ¼ R2f H
^ R2f 1sH
^ 1s
ð1Þ
E0
ð2Þ
þ 2h1sjzj2pz i E0 minJðf ¼ 0:844Þ
Mathematica code and results:
E41.
eiss 1 dðeiss 1Þ=ds
s ¼ xnm
x lim ¼ ¼ is
s!0 s ds=ds
E42.
The lowest and highest wavelengths of the visible region are approximately equal to
400 and 800 nm. The 400 nm wavelength is 24 = 16 times more intensely scattered
than the wavelength at 800 nm. This explains the color of the sky.
E44.
The intensity of the Stokes line is proportional to the population of level a in the
vibrational transition a !b, whereas the intensity of the corresponding anti-Stokes
line is proportional to the population of the excited level b in the vibrational
transition b !a. Hence, using the Boltzmann distribution and m4-dependence of the
Stokes and anti-Stokes lines, we can write
qb ¼ qa exp( hvba =kTÞ vba ¼ vb va

IStokes ðv0 vba Þ4 qa ðv0 vba Þ4
¼ 4
= exp(hvba =kTÞ
IantiStokes ðv0 þ vba Þ qb ðv0 þ vba Þ4
It can be concluded that the ratio of the Stokes and anti-Stokes intensities
increases rapidly as mba increases.
E46.
For C2H6 at 298 K, EMP2/6-311G*− EHF/6-311G* −177.7 kcal mol−1, whereas for the
chemical reaction C2H4(g) + H2(g) ! C2H6(g) at 298 K, EMP2/6-311G*− EHF/6-311G*
1.7 kcal mol−1. The MP2 correlation correction for the ethane molecule
corresponds to 0.36 % of the Hartree–Fock energy and is about the energy of two
C–H bonds. For the chemical reaction, the MP2 correlation correction is less than
1 % of the absolute value of the corresponding value for the molecule. This finding
suggests that the correlation energy is mainly due to the electrons of each electron
pair. Now, since the number of electron bond pairs is conserved for reactants and
products, the correlation correction is appreciably lower than for a molecule. In
addition, the correlation correction for a reaction is not necessarily negative, since it
is the correlation correction for the products minus the correlation correction for the
reactants.
Chapter 4—Molecular Symmetry
E1.
(a) C3 (b) C3 (c) C4 (d) C5 (e) C2v (f) D3h (g) D6h (h) D∞h (i) Td

E6.
X XX XX X
Bkk ¼ X 1 ki Aij Xjk ¼ Aij Xjk X 1 ki ¼ Aij X X 1 ji
k X
k i;j X k i;j i;j
¼ Aij dji ¼ Aii
i;j i
ðQEDÞ
E7.
(a)
A ¼ X 1 AX ) XAX 1 ¼ A
(b)
B ¼ X 1 AX ) XBX 1 ¼ A
(c)
A ¼ X 1 BX A ¼ Y 1 CY
Y 1 CY ¼ X 1 BX C ¼ YX 1 BXY 1
1
C ¼ XY 1 B XY 1 C ¼ Z 1 BZ

E C3 C23 rv1 rv2 rv3

C3 C23 E rv2 rv3 rv1
C23 E C3 rv3 rv1 rv2
rv1 rv3 rv2 E C23 C3
rv2 rv1 rv3 C3 E C23
rv3 rv2 rv1 C23 C3 E
This point group is not abelian (not commutative).

E9.
Infrared active vibrations transform like x, y, and z, and for a point group with a center
of inversion, these functions are of plus–minus- or u-type. In turn, Raman active
vibrations transform like the quadratic functions x2, y2, z2, xy, xz, and yz, and for a
point group with a center of inversion, these functions are of plus–plus- or g-type.
E10.
C3v E 2C3 3rv

C 15 0 3
C ¼ 4A1 + A2 þ 5E
Crot ¼ A2 + E Ctrans ¼ A1 + E
Cvib ¼ 3A1 + 3E
C2v E C2 rxz ryz
C 15 1 3 3
C ¼ 5A1 + 2A2 þ 4B1 + 4B2
Crot ¼ A2 + B1 + B2 Ctrans ¼ A1 + B1 + B2
Cvib ¼ 4A1 + A2 + 2B1 + 2B2
Cs E rh
C 15 3
C ¼ 9A + 6A00
0
Crot ¼ A0 + 2A00 Ctrans ¼ 2 A0 + A00

Cvib ¼ 6 A0 + 3 A00
C1 E
C 15
C ¼ 15 A
Crot ¼ 3 A Ctrans ¼ 3 A
Cvib ¼ 9 A
Chapter 5—Molecular Structure
E1.
X Zn Z
q0 ðrÞ
/ðPÞ ¼ dv
n
rnP reP
The first term of this expression is the electrostatic potential, which results from
the discrete distribution of nuclei of the molecule. The second term results from the
continuous distribution of electronic charge, where q0 ðrÞdv is the electronic charge
on the volume element dv centered at P.
E2.
1 1
W0 ð1; 2Þ ¼ pffiffiffi j/1 að1Þ/1 bð2Þj ¼ pffiffiffi /1 ð1Þ/1 ð2Þ½að1Þbð2Þ að2Þbð1Þ
2 2
/1 ¼ MO /1 ¼ c11 v1 þ c21 v2 v1 ; v2 ¼ AOs v1 ¼ 1s1 v2 ¼ 1s2
Z
q0 ðrÞ ¼ 2 W0 ð1; 2ÞW0 ð1; 2Þdv2 dr1 dr2 q0 ðrÞ ¼ 2/1 /1 ¼ 2c211 v21 þ 2c221 v22 þ 4c11 c21 v1 v2
Z Z
q0 ðrÞdv ¼ 2c211 þ 2c221 þ 4c11 c21 S12 ¼ 2 S12 ¼ v1 v2 dv
1
c211 þ c221 þ 2c11 c21 S12 ¼ 1 c211 ¼ c221 jc11 j ¼ jc21 j ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2ð1 þ S12 Þ
E4.
Full Mulliken population matrix element between 1s2 and 1s3: −0.01993.
Density matrix element between 1s2 and 1s3: −0.11568.
Overlap matrix element between 1s2 and 1s3: −0.01993/−0.11568 = 0.172286.
E5.
Population condensed to C atom: 2.06590 + 0.79067 + 0.65429 3 +

(−0.05013 2) = 4.71918.
Population condensed to the overlap region between C and H1 atoms:
−0.00594 + 0.11187 + 0.09331 3 = 0.38586.
E6.
pffiffiffi
df2 ¼ a2 + a2 df ¼ 2 a
2 2
dc df a 2 pffiffiffi
¼ þ dc ¼ 3 a
2 2 2
rffiffiffi rffiffiffi
df dc a a 2 2
¼ sin sin ¼ a ¼ 2 arcsin a ¼ 109:4712
2 2 2 2 3 3
E7.
hA ¼ 12 ðs þ px þ py þ pz Þ hB ¼ 12 ðs px þ py pz Þ
hC ¼ 12 ðs þ px py pz Þ hD ¼ 12 ðs px py þ pz Þ
SAA ¼ 14 ð1 þ 1 þ 1 þ 1Þ ¼ 1 SAA ¼ SBB ¼ SCC ¼ SDD ¼ 1
SAB ¼ 14 ð1 1 þ 1 1Þ ¼ 0 SAB ¼ SAC ¼ SAD ¼ SBC ¼ SBD ¼ SCD ¼ 0
pffiffiffi
1 pffiffiffi 1 pffiffiffi px 3py
hA ¼ pffiffiffi ðs þ 2px Þ hB ¼ pffiffiffi s þ 2
3 3 2 2
pffiffiffi
1 pffiffiffi px 3py
hC ¼ pffiffiffi s þ 2 þ
3 2 2
Z
1 1
SAB ¼ hA hB dv ¼ 1 þ 2 0 ¼ 0 SAB ¼ SAC ¼ SBC ¼ 0
3 2
SAA ¼ SBB ¼ SCC ¼ 1
E9.
1 1
hA ¼ pffiffiffi ðs þ px Þ hB ¼ pffiffiffi ðs px Þ
2 2
Z
1
SAB ¼ hA hB dv ¼ ð1 1Þ ¼ 0 SAA ¼ SBB ¼ 1
2
E10.
1 pffiffiffiffiffi 1 pffiffiffiffiffi
hA ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi s þ kA pA hB ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi s þ kB pB
1 þ kA 1 þ kB
1 pffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi
ShA hB ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Sss þ kA kB SpA pB þ kA SpA s þ kB SspB
ð 1 þ kA Þ ð 1 þ kB Þ
1 pffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi
¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ kA kB SpA pB þ kA 0 þ kB 0 ¼ 0
ð 1 þ kA Þ ð 1 þ kB Þ
pffiffiffiffiffiffiffiffiffiffi 1
1 þ kA kB SpA pB ¼ 0 SpA pB ¼ pffiffiffiffiffiffiffiffiffiffi
kA kB
E11.
1 = antiperiplanar conformer (g1 = 1) 2 = synclinal conformer (g2 = 2)

f2
f2 / g2 eE2 =ðRTÞ ¼ 2eE2 =ðRTÞ f1 / eE1 =ðRTÞ ¼ 2eðE2 E1 Þ=ðRTÞ
f1

f2 E 2 E1 2f1
ln ¼ DE21 ¼ E2 E1 ¼ RTln
2f1 RT f2

2 0:80
¼ 2:5ln ¼ 5:2 kJ mol1
0:20
E12.
The naturally occurring form is the R,R stereoisomer. The R,S stereoisomer is
optically inactive.
Chapter 6—Crystals
E1.
Comparing with Fig. 6.6, all distances now have a scaling factor 2r.
E2.
Considering Fig. 6.6, the distance from point c (the intersection of the threefold axis
with the plane of the disks) to the disk centers is equal to d = 1/√3. Hence, x =
d cosh and y = d sinh. Note that when r = 1/2, d = 1.
Inspection of the above (x,y) coordinates shows that disks are drawn in the
counterclockwise direction in the first two graphs and in the clockwise direction in
the last two graphs.
E3.
Consider a hexagonal prism as repeating unit (see Fig. 6.8). Each sphere in a prism
vertex is shared by six repeating units, thus being considered with a factor 1/6 in the
packing density. The centers of the bottom and top of the hexagonal prism are
occupied by one sphere that is shared by two repeating units, thus having the factor
1/2. Note that there are three spheres inside the hexagonal prism counting 1 each of
them. The volume of the spheres in one repeating unit is given by

1 1 1 4
6 + 6 þ 2 þ 3 pr 3 ¼ 8pr 3
6 6 2 3
In turn, the volume of the hexagonal prism can be obtained by adding the volume of
three parallelepipeds whose height is given by 2 2(√2/√3)r (see answer to E1).
The edge of the base parallelogram of each parallelepiped is given by 2r, which
multiplies by r√3. Therefore, the volume of the hexagonal prism is given by
pffiffiffi
pffiffiffi 2 2 pffiffiffi
3 2r r 3 2 pffiffiffi r ¼ 3 8 2r 3
3
The packing density of the hcp structure of identical spheres is obtained by
dividing the volume of the spheres contained within one hexagonal prism as
repeating unit of the hcp structure by the total volume of the hexagonal prism,
8pr 3 p
pffiffiffi ¼ pffiffiffi 0:74
3 8 2r 3 3 2


E10.

1 1
/ðeÞ ¼ ) e ¼ l þ kB T ln 1
1 þ exp½ðe lÞ=ðkB TÞ /

1
l ¼ 3:24 eV T ¼ 7000 K ) e ¼ 3:24 þ kB 7000 ln 1 ¼ 3:90 eV:
0:25
Chapter 7—Water
E1.
P
The definition of the dipole moment vector is given by l ¼ k qk rk :
For a pure electric dipole, that is, two electric charges +q and −q separated by a
distance d, the above expression is consistent with the dipole moment vector being
directed from the negative to the positive charge. In fact, if we take the origin of the
dipole moment at the positive charge, the positional vector r+ is zero (qr+ = 0), and
the positional vector that points to the negative charge r− is multiplied by the
negative charge −q, thus giving a dipole moment vector l = qr+ + (−q)r− = −qr−
that points to the positive charge.
E2.
X X X X X
qi r0i ¼ qi (ri + R) ¼ qi r i þ R qi ¼ qi ri
|fflffl{zfflffl}
0
where R is the vector that changes the origin.
E3.
lðH2 OÞ ¼ 2 0:3400:9611 cosð104:53=2p=180Þ4:8032 ¼ 1:92 D.

-
- - -
This graph shows that for the same molecule (H2O) and method of calculation
(B3LYP), the increase of the basis set lowers the ground state energy, as stated by
the variational principle (see Sect.3.7).
E7.
sffiffiffi rffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi
k 16 2 16 1 xOH lOD 32 17
x¼ lOD ¼ lOH ¼ 2
l 18 17 xOD lOH 18 16
E8.
H2O belongs to the C2v point group. The antisymmetric stretching transforms as B2;
the symmetric stretching and the bending transform as A1.
DHO belongs to the Cs point group. The normal modes are totally symmetric
(transform as A′) because all displacement vectors are in the molecular plane.
E9.
ð35; 35; 35; 35Þ 34 ¼ 81

ð35; 35; 35; 37Þ 4 33 4 ¼ 108
ð35; 35; 37; 37Þ 6 32 6 ¼ 54
ð35; 37; 37; 37Þ 4 3 4 ¼ 12
ð37; 37; 37; 37Þ 1¼1
Relative intensities of the distinct isotopic lines: 81, 108, 54, 12, 1.
ð35; 35; 35; 35Þ l ¼ 35 þ 35 þ 35 þ 35

1 1 1 1 1
l ¼ 8:75
ð35; 35; 35; 37Þ l0 ¼ 35 þ 35 þ 35 þ 37
1 1 1 1 1
l0 8:87
rffiffiffiffi r ffiffiffiffiffiffiffiffiffi
x0
l 0 8:75
¼ jx xj ¼ 459 459 3 cm1
x l0 8:87
12.8
E11.
For a diluted solution,
ns ns ns = 1 mol dm3 cs
xs ¼ ¼ ¼ ¼ cs VS
ns þ nS nS nS = 1 mol dm3 cS
where cs is the solute molarity, cS is the solvent molarity (for water, cS = 55.55 mol
dm−3). Note that the inverse of cS is the molar volume of the solvent in dm3 (for
water, Vm = 0.018 dm3 mol−1).
E12.
X
I ¼ (1=2) z2k mk m0 ¼ (1=2) 22 0:001 þ ( 1)2 0:002 ¼ 0:003

k
logcsalt ¼ Ajzc za jI 1=2 ¼ 0:509 2 0:0031=2 ¼ 0:05576 csalt ¼ 0:880
E13.
T ¼ 298:15 K pO2 ¼ pH2 ¼ p0 ¼ 1 atm a H2 O 1

RT a2H2 O
O2 þ 4H þ þ 4e ! 2H2 O eO2 = H2 O ¼ e0O2 = H2 O ln ¼ 1:229 0:05916 pH
4F ðpO2 = p0 Þ a4H þ
RT pH2 = p0
2H þ þ 2e ! H2 eH þ = H2 ¼ ln 2 ¼ 0:05916 pH
2F aH þ
An oxidizing agent that converts H2O into O2 and a reducing agent that converts
H2O to H2 are not thermodynamically stable in aqueous solution. Redox couples
that are thermodynamically stable in water lie between the above sloping straight
lines e = e(pH) (region marked gray).
E14.
2e042 ¼ e043 þ e032 e043 ¼ 2e042 e032 ¼ 2 1:224 1:5415 ¼ 0:907
E15.
3e074 ¼ e076 þ 2e064 2e064 ¼ 3e074 e076 ¼ 3 1:679 0:558 ¼ 2:240

E18.
Temperature is assumed to be 298.15 K, and subscripts c and a stand for cathode

and anode. Galvani cells have positive D e.
RT RT
ec = lnaAg þ ðcÞ = 0.05916 log aAg þ ðcÞ ea ¼ lnaAg þ ðaÞ = 0.05916 log aAg þ ðaÞ
F F
aAg þ ðcÞ
ec [ ea De = ec ea = 0.05916 log [ 0 ) aAg þ ðcÞ [ aAg þ ðaÞ
aAg þ ðaÞ
Index
A 6-31++G**, 188
Ab initio calculations, 183, 204, 339 6-311G, 187, 210, 340, 402
Absorption of radiation. See Einstein A and B aug-cc-pVTZ, 153, 188, 271, 340, 345,
coefficients 351–353, 356–359
Acidity, relative, 386 cc-pVTZ, 130, 148, 188, 251, 255,
Acids 266–268, 270, 273, 274, 276, 338, 340,
Bronsted–Lowry, 381, 387, 391 346–349
Lewis, 386, 387, 392 STO-3G, 185, 187, 257, 402
Activation energy. See Arrhenius activation Basis set
energy double-zeta, 187
Activity coefficient, 46, 369, 373, 374, 384, polarization, 187, 236, 281, 287, 288
388, 392, 403 split-valence, 187
Addition reaction, 387 triple-zeta, 187
Adiabatic valence double-zeta, 187
compression, 33, 68, 76, 79 valence triple-zeta, 188
expansion, 33, 67, 76, 79 Bernoulli, 3
transformation, 1, 35, 66, 68, 72, 76, 79 Boiling points of linear alkanes, 355
Affinity of chemical reaction, 42, 54, 74, 76, 79 Boltzmann distribution
Angular momentum Boltzmann constant, 3, 177, 277, 317
orbital, 133, 136, 188, 207 degeneracy degree, 168, 277
spin, 135, 143 population fraction, 277
Anharmonicity. See Morse function Bond orbital
Antiderivative, 20 p, 265
Antisymmetric function, 141 r, 265
Antisymmetry principle, 139, 140, 205 Bond orbitals, natural, 264–266, 283, 288, 347
Arrhenius activation energy, 93, 113 Born–Oppenheimer approximation, 145, 154,
Arrhenius A-factor, 92, 111 162, 205, 339, 349
Associated Legendre polynomials, 134, 135 Boyle. See Ideal gas
Avogadro Bracket notation, 144
constant, 2 Bravais lattices
hypothesis, 2 three-dimensional, 305, 333
Axis of rotation, 246 two-dimensional, 302–304
Buffer capacity, 385, 401
B
B3LYP method, 148, 153, 402 C
Basis functions Cahn–Ingold–Prelog rules applied to
3-21G, 187, 402 bromochlorofluoromethane, 216, 278, 279
6-31G, 187, 340, 376, 402 carvone, 279
6-31G**, 187 Calculus, fundamental theorem of, 20, 157

DOI 10.1007/978-3-319-41093-7
450 Index
Calorimeter antiperiplanar, 275, 276, 289

bomb, 30 dihedral angle, 272–275
isothermal, 30 eclipsed, 273
Carnot cycle, 35, 66, 68, 84 for 1,2-dichloroethane, 276
Carnot heat engine, 33, 35, 65, 73, 76, 79 for ethane, 274
Carnot’s theorem, 32, 76, 79 potential energy profile, 272, 273, 275
Carvone, 280 relaxed scan, 272, 275, 355–358
Center of symmetry, 214, 217, 247 staggered conformation, 273, 274
Central force, 130 synclinal, 275–277
Cesium chloride, 330 synperiplanar, 275, 276
Characteristic equation, 104 torsional mode, 274
Character tables Conformational isomer or conformer, 276, 287
C2h character table, 219, 232 Coordination number, 296, 298, 301, 334
C2v character table, 224–226, 227, 228, Correlation energy
230, 231, 237, 241, 346 coulombic, 161
C3v character table, 219, 222, 223, 228, kinetic, 161
233, 235, 237, 238, 247 Coulomb integral, 150
D3h character table, 233, 235, 236, 262 Coulson theorem, 262
Charles' law. See Ideal gas Critical point, 14, 52, 70, 79, 82, 363, 364
Chemical equilibrium, constant, 51 Crystal structure
Chemical oscillator cesium chloride, 309
Belousov–Zhabotinsky reaction, 99, 111 diamond, 311, 312
Brusselator, 99, 101, 103, 110, 111, 113 metals, 308, 316, 334
characteristic equation, 101, 104 sodium chloride, 310, 331
stationary point, 9, 99–102, 111, 127, 157, unidimensional model, 314
267, 269, 273, 288 zinc blende, 311
Chemical potential Cubic close packing (ccp), 296, 297, 308
ideal gas, 45
liquid solutions, 44, 46 D
pure liquids and solids, 30, 44, 47 Density, number, 3, 5, 11, 12, 78, 82
real gas, 30, 45, 46 Density functional theory (DFT), 154, 156,
Chemical reaction 162, 183, 206
addition, 387 Density matrix, 255, 258, 263, 264, 283, 289
affinity of, 48, 55 Derivative, mixed partial second, 19, 134
displacement reaction, 387 Diagonalization, matrix, 106, 122
extent of, 24, 55, 75, 77, 81, 87 Diamond, 310
first-order, 19, 89, 90, 95, 108 Dichloroethane, 1,2
quotient, 51, 64, 74, 78, 82 anticlinal conformation, 276
rate of, 42, 86, 87, 91, 112 antiperiplanar conformation, 276
second-order, 89, 90, 113 conformers (conformational isomers), 276
stoichiometric coefficients, 77, 81, 112 potential energy profile, 276
zeroth-order, 91 synclinal conformation, 275
Chiral center, 278–280, 287 synperiplanar conformation, 275
Chiral molecule, 216, 278, 279, 281, 288 torsional modes, 274
Clausius’s inequality, 37 Differential
Collision, 3–5, 9, 10, 12, 13, 78, 82, 83, 317 equation, 89, 99, 110, 118, 119, 126, 170,
Complete set of functions, 120 189, 192
Complex reaction, 94, 111 exact, 19, 22, 36
Comproportionation, 390, 391, 403 inexact, 16, 21, 26, 27
Computational methods, 183 total, 19
Condensed population, 259, 289 Dihedral angle, 272, 273, 275, 288
Configuration interaction, 153 Dirac notation, 123, 206
Conformation Displacement reaction, 387
anticlinal, 275, 276 Disproportionation, 391, 392, 403
Index 451
Dissociation energy, equilibrium, 147 Elementary reactions

Double-zeta basis set, 187 bimolecular concerted step, 94, 271
molecularity, 93
E nucleophilic substitution reaction SN2, 94,
Efficiency of heat engine, 31, 32, 73 271
Eigenfunctions, 117, 120, 122, 123, 128, 134, potential energy surface, 269, 285, 288
137, 143, 164, 168, 170, 191, 243, 345 Emission of radiation
Eigenvalues, 102, 122, 123, 264, 283 spontaneous, 177, 178, 208
Eigenvectors, 102, 257, 284 stimulated, 176, 177, 180, 208, 229
Einstein Enantiomers
A coefficient, 177 R and S, 280
B coefficient, 177 Enthalpy
Electric multipoles, 340, 344 of formation, 31, 339
Electrochemical affinity, 60, 62, 63, 76, 80 standard reaction enthalpy, 31, 83
Electrochemical cell Entropy
anode, 62, 63 etymology, 76, 79, 80
cathode, 63 rate of entropy production, 38, 62, 78, 81
cell diagram, 62 Equation of state, 3, 78, 79, 81, 82
electrode potential, 65, 84, 387, 388, 390 Equilibrium constant, 52, 74, 84, 97, 378, 383,
electrolytic cell, 62, 76, 77, 79, 80 399
galvanic cell, 61–63, 76, 77, 79–81 Equilibrium dissociation energy, 285
Nernst equation, 62, 64, 78, 82 Equilibrium
oxidation half-reaction, 61, 62, 77, 80 mechanical, 24, 25, 33
reduction half-reaction, 61, 63, 64, 77, 80, thermal, 24, 33, 317
84, 390 Equilibrium state
salt bridge, 61, 62 state functions, 24, 171, 371
standard cell potential, 64 Ethane
standard electrode potential, 65, 387, 389, conformations, 272–275
390 eclipsed conformation, 272, 273
zero-current cell potential, 64, 77, 78, 80, potential energy profile, 273
82 saddle point, 269, 270, 288, 289
Electrochemical potential, 60, 63, 76, 77, 80 staggered conformation, 273, 274
Electromagnetic radiation torsional mode, 274
plane polarized, 174 Ethene (ethylene), 264, 268
Electron charge density, 154, 161, 254 acidity, relative, 386
Electron probability density natural bond orbitals, 249
isodensity surface, 251, 254, 288, 356–358 Ethyne (acetylene), 266, 267
Electrons in solids acidity, relative, 386
bands, 237, 332, 334 natural bond orbitals, 249
model, 314 Euler–Lagrange differential equation, 157
Electron-spin multiplicity, 142, 151 Exact differential, 17, 19, 22, 27, 83
Electrostatic potential Exchange
definition, 343, 388, 402 energy, 23, 27, 37, 38, 40, 59, 76–78,
electrophilic region, 254, 288, 356, 358 80–82, 161
for a charge distribution, 254, 288 integral, 150, 210
for a molecule, 129, 145, 288, 339, 340, operator, 151
346, 357 Exchange-correlation
for CH4, H2C=CH2, HCCH, 266 energy functional, 161, 163
for electric multipoles, 340, 341 potential, 162
mapped over isodensity surface, 268, 358, Extent of chemical reaction, 48, 74, 77, 81
359, 376 Exterior, 23, 25, 27, 253
nucleophilic region, 254, 288, 356, 358 External potential, 154, 157, 158, 161
452 Index
Extremely fast reactions phase diagram, 52, 53

relaxation, 96, 97 Glossary, 75, 79, 111, 204, 245, 287, 333, 391
Gravity acceleration, 65
F Gross orbital population, 259
Face-centered cubic (fcc), 296, 297, 307,
309–311, 331 H
Fermi–Dirac distribution Half-life, 89–91, 96, 109
for metallic conductors, 317 Hamiltonian
First law, 26, 32, 42, 66, 83 pure electronic, 146, 154, 158, 162, 183,
Fock equations, 152, 160 204
Fock operator, 151, 183 Harmonic oscillator
Force constant, 125, 285 classical treatment, 125
Frequency, angular,, 119, 126, 174, 181, 182, force constant, 127
281 Hermite polynomials, 127, 192
Frost diagram, 390, 391, 403 quantum-mechanical treatment, 126, 180
Full Mulliken population analysis matrix, 258 reduced mass, 125, 209, 349
Function(s) zero-point energy, 129, 147, 274–276
complete set of, 121, 123 Hartree–Fock method, 148, 150, 151, 153, 160,
cusp at the nucleus, 184, 185 163, 205, 210, 257, 345
even, 228, 244, 246 Hartree–Fock–Roothaan equations, 152
odd, 228, 244, 246, 343 Heat
set of basis, 144, 153, 184 calorimeter, 29, 30
well-behaved, 118–121, 123, 140, 143, 207 capacity, 28, 83, 382
Functional, derivative, 17, 155–158, 200, 206 Carnot heat engine, 33, 73
Fundamental theorem of calculus, 16 Carnot’s theorem, 32
sink, 31, 32, 34, 35
G source, 31, 33, 34, 67
Gas(es) Heat capacity
gas constant, 2, 16, 277 isobaric, 28, 66, 78, 81
ideal, 3, 12, 13, 15, 30, 33, 35, 44–46, 66, isochoric, 28, 66, 78, 81
68, 72, 74, 76, 78, 79, 81, 84, 339, 364 molar, 28, 29, 66
kinetic model of, 3, 4, 12, 78, 82 specific, 28
Gaussian-type functions Heat engine
Cartesian, 184, 186, 188 efficiency, 31–33
contracted, 184, 187, 188 irreversible, 35, 37, 40, 41, 48, 52, 54–58,
contraction coefficients, 184 60, 63, 81, 98, 368, 369, 371, 392
degree of contraction, 184 reversible, 31, 33
orbital exponent, 184 Heisenberg uncertainty principle, 139, 181
primitive, 184, 187 Helix
product of, 186, 192, 203 as a chiral object, 278
s-, p-, d-type, 184, 185, 188, 200, 201 left-handed, 280, 286
Geometry optimization right-handed, 280, 286
for CH4, 267 Helmholtz
for H2C=CH2, 266, 267 energy, 54–57, 78, 81, 82
for HCCH, 266, 267 equation, 119
Gibbs–Duhem equation, 43, 44, 77, 81, 84 Hermite polynomials, 192, 194, 209
Gibbs energy Hermitian operator, 119–123, 207, 209
of formation, standard, 48–51 Hexagonal close packing (hcp), 295, 320, 324
Gibbs–Helmholtz equation, 50, 370 Hohenberg–Kohn theorems, 158–160, 162
Gibbs phase rule Hohenberg–Kohn variational theorem, 159,
components, 53, 54 163
degrees of freedom, 53, 78, 81, 84 Hückel method, 314, 315, 332
independent components, 54, 78, 81 Hybrid atomic orbitals
phase, 52–54, 78, 81, 84 CH hybrids, 267
Index 453
for CH4, 266 Isochoric heat capacity, 28, 66

hybridization parameter, 261 Isothermal
s and p fractions, 261 compression, 26, 33, 67, 76, 79
Hybridization parameter, 262 expansion, 25, 33, 76, 79
Hydrogen atom, 130, 136, 137, 163, 185, 188, Isotropic space, 4, 6, 8
201, 210, 260, 263, 264, 266–268,
272–274, 278, 279, 283, 289, 339, 340, K
349, 354–356, 361, 386, 387 Kinetic experimental method
Hydrogen bond initial rate, 88
acceptor, 356–359 isolation, 88
atomic charges, 339, 340, 345, 402 Kinetic model of gases
ClH...OH2, 356 Bernoulli, 3
donor, 318, 356–359, 391 Boltzmann, 4
FH...OH2, 356–358 cross section for collision, 10, 12
partial covalent character, 356, 357, 359 distribution of molecular speeds, 8
polarization, 356–358, 365 Joule, 3
water dimer, 357–359 Maxwell velocity distribution, 13
Hydrogen electrode, standard, 387, 389 mean free path, 10–12
mean speed, 9
I most probable speed, 9
Ice Ih (see Water) root mean square speed, 9
Ideal gas Kinetic reaction profile, 86
Boyle and Mariotte, 2 Kohn–Sham
empirical temperature scale, 3, 13, 65, 84, method, 159, 162, 163
91 orbitals, 159, 163
equation of state, 3, 13
Jacques Charles, 2 L
number density, 3, 5, 11, 12 Laplacian operator, 137, 190
Insulator, 23, 317 Latimer diagram, 390
Integral Lattice. See Bravais lattices
indefinite, 20, 44, 89 Lattice
line, 16, 21 energy, 308, 309, 381
path, 21 enthalpy, 308
Intermolecular interaction Legendre polynomials, associated, 134, 135
dispersion, 350, 354, 372, 377 Lewis
electrostatic, 58, 161, 350, 351, 365, 367, acid, 386, 387
377 base, 386, 387
induction, 350–352 complex, 386, 387
Interstice. See Packing spheres Linear momentum, 5, 117, 133
Interstice Line integral, 16, 21
cubic, 300, 301 Long-wavelength approximation, 175
octahedral, 299, 300
tetrahedral, 299, 300 M
Irreducible representation, 225, 226, 228, 229, Mariotte. See Ideal gas
231, 233, 234, 236, 257, 258 Mathematica codes, 8, 15, 16, 33, 35, 52, 69,
Irreversible cycle, 35, 36 71–76, 79, 82, 86, 103, 108, 110, 111,
Irreversible processes 118, 119, 127, 129, 135, 142, 145, 151,
chemical reaction, 38, 41, 48, 54 158, 184, 185, 191, 193, 195, 196,
gas expansion, 38, 368 198–201, 203, 224, 225, 229, 238, 241,
heat flow, 38, 56 264, 270, 280, 283, 286, 292, 294, 295,
matter diffusion, 38, 40, 41 302, 304, 305, 309, 315, 317, 319, 321,
rate of entropy production, 38, 62 323, 326, 327, 330–332, 338, 341, 362,
Isobaric heat capacity, 28, 66 378, 379, 385, 395–401
454 Index
Matrix diagonalization for CH4, 257, 263, 265

eigenvalues, 104, 105, 107, 110, 113, 117, for H2C=CH2, 264, 265
121 for HCCH, 266
eigenvectors, 104, 105, 107 Nernst equation, 64, 403
orthogonal transformation, 107 Neutralization reaction in water, 98
secular equation, 104 Nitrogen, liquid, 12
similarity transformation, 106, 107 Nucleophilic substitution reaction SN2, 94
Maxwell, velocity distribution, 7, 13
Mean free path, 9, 11, 12, 78, 82, 83 O
Mechanical work, 25, 27, 32, 35, 66, 78, 82 Operator(s)
Membrane potential commuting, 116, 205
electrochemical affinity, 60, 63, 76, 80 eigenfunctions of, 121–123, 207
electrochemical potential, 59, 62, 63, 77, 80 eigenvalues of, 120–122, 207, 225
ion-selective membrane, 60, 62, 63, 76, 80 exchange, 139
Metals, 308, 317, 334 Hamiltonian, 117, 136
Methane Hermitian, 120–122, 136, 165, 207
hybrid orbitals, 266, 289 Laplacian, 137
minimum basis set calculation, 289 linear, 116, 117, 120, 121
natural bond orbitals, 263, 265, 283 momentum, 118
Minimal basis set, 187, 257, 289 permutation, 139
Molecularity product of, 116, 132
bimolecular, 93 Optical activity
trimolecular, 93 specific rotation, 283
unimolecular, 93 Optimization, geometry, 288
Molecular speeds Orbital
distribution, 8, 69, 78, 82 exponent, 151, 187
Molecular vibrations fraction, s, 263, 266, 268
and symmetry, 231, 233, 236, 237, 347, population, gross, 259
348 Orthogonal functions, 136
for isomers of 1,2-dichloroethane, 230
for molecules with XY3 geometries, 233, P
236 Packing, 334
Moller–Plesset-2. See MP2 calculations coordination number, 296, 297, 299
Momentum, linear, 5, 139, 209 cubic close packing (ccp), 296, 298
Morse potential densities, 296, 299
anharmonic corrections, 130 hexagonal close packing (hcp), 295–298
MP2 calculations, 183 occupying interstices, 299
Mulliken population analysis of disks, 291, 292, 294, 319, 321, 323
atomic charge, 259 of spheres, 291, 293, 295–297, 299, 323,
density matrix, 256, 259, 289 334
gross orbital population, 259 regular tetrahedron of spheres, 293, 295,
matrix, 256–259, 289 299, 323, 359
minimal basis set calculation for CH4, 257 Particle in a one-dimensional box, 188, 209,
overlap matrix, 255, 259, 288, 289 242
population condensed to an atom, 259 Path-dependent quantity, 26
Mutual exclusion Pauli exclusion principle, 140, 141, 263, 283,
and symmetry, 233 288, 317, 334
of infrared and Raman active vibrations, Permutation operator. See Exchange operator
233 Perturbation theory
degenerate energy level, 168
N nondegenerate energy level, 164, 168
Natural variational perturbation method, 166
atomic orbitals, 347 Phase diagram, 363
bond orbitals, 260, 263–266, 347 Pi (p) bond, 265, 266
Index 455
Plane of reflection, 214, 215, 222, 246, 357 Reaction rate. See Rate of chemical reaction
Point groups, 245 Reduced mass, 124, 137, 207
abelian, 217, 225, 245, 247 Representation of point group
commutative, 217 character table, 224, 225
dichotomy diagram, 218 class, 224, 238
multiplication table, 224, 225, 241, 247 in x, y, z-basis, 224, 323
principal axis, 217, 225, 246 in x,y,z-basis, 225, 228
properties, 217, 245 irreducible, 225, 226, 228, 229, 231, 233,
Polarimeter, 282, 283, 288 245, 246
Polar plot, 135 totally symmetric, 225, 226, 229
Population analysis.See Mulliken population Resistivity, electrical
analysis; Natural bond orbitals for insulators, 316, 317
Potential energy surface for metallic conductors, 316
first order saddle point, 269 semiconductors, 316, 317, 319
Pressure, 2–5, 9, 12–16, 23–28, 30, 31, 37, 39, Reversible cycle, 33, 35, 36, 68, 76, 77, 79, 80
41, 42, 44–49, 52, 53, 55–57, 65, 70, Rigid rotor, two-particle, 190, 209, 243
71, 74, 76–83, 96, 123, 308, 355, 360, Rotation–reflection axis, 214–216
363, 364, 368, 369, 382, 388, 389 Rotations, molecular, 146
Prigogine, 37, 77, 81, 99, 111
Process dependent quantity, 26 S
Saddle point
R first- order, 269, 285, 289
Radial distribution function, 138 Schrödinger equation
Raman depolarization ratio, 236 for atoms, 143, 147
Raman scattering, 245, 246 for molecules, 146
anti-Stokes transition, 178 one-dimensional, 117, 119, 127
carbon tetrachloride spectrum, 178, 365, Second law
367, 402 Clausius’s inequality, 83
classical treatment, 125, 179 Secular equation, 105
depolarization ratio, 236, 237 Selection rules
discrete resonance, 182, 183, 230 and symmetry, 230, 231, 233–238, 241,
normal, 182 245, 246
preresonance, 182 for absorption and stimulated emission, 229
quantum-mechanical treatment, 180 for a harmonic oscillator, 229, 244
resonance, 178, 182 for a particle in a one-dimensional box,
Stokes transition, 178 229, 242
transition polarizability, 181, 182 for Raman scattering, 229
virtual state, 181, 182, 230 for two-particle rigid rotor, 191, 229
Rate of chemical reaction Self-consistent field, 150, 153, 205
experimental rate equation, 88, 94, 95, 112 Semiconductor
extent of chemical reaction, 41, 77, 81 doping, 318
first-order, 113 n-type, 319
half-life, 90, 112 p-type, 319
overall order, 88 Semiempirical calculations, 183, 207
partial orders, 88, 112 Set of basis functions, 144, 153, 184
pseudo-order rate constant, 88, 112 Sigma (r) bond, 214
pseudo-order rate equation, 88 Similarity transformation, 106, 107, 223, 224,
rate constant, 88, 90, 92, 111, 113 238, 239
second-order, 90, 97 Similar matrices, 106, 107, 223
zeroth-order, 91 Singlet, 142
Reaction coordinate, 288 Slater determinant, 141, 142, 148, 150, 151,
intrinsic, 270, 271 153, 159, 160, 162, 207, 255, 345
transition state, 94, 270, 271 Slater-type orbital, 150, 184, 187, 199
456 Index
Sodium chloride, 309, 310, 312, 331, 365, 367, T

373, 375 Temperature
Specific rotation, 283 absolute scale, 33, 35, 72
Spectroscopic transition, 176, 229 Celsius scale, 65, 84
allowed, 229, 243–246 Fahrenheit scale, 65
forbidden, 229, 244, 246 ideal gas empirical scale, 33
Speed. See kinetic model of gases Test charge, 252, 253, 288, 343
Spherical coordinates, 115, 130, 132, 134, 137, Thermodynamic state, 3, 23, 24, 31, 57, 371
138, 148, 190, 201 Thermodynamic system
Spherical harmonics, 115, 134, 137, 138, 148, closed, 23
195, 207, 243, 402 isolated, 23, 24
Spin-orbital, 140–142, 150, 151, 159, 207, 255, open, 23
345 Thermodynamic variables
Split-valence basis set, 187 extensive variables, 23, 43
Spontaneous emission of radiation, 177, 208 intensive variables, 3, 23, 24, 43, 44, 53,
Standard 77, 78, 81, 82
molar enthalpy of formation, 31 Time-dependent
reaction enthalpy, 31, 83 perturbation theory, 100, 170, 171, 173,
Standard basis sets, 115, 187 230
Standard electrode potential, 65, 387, 389, 390 Schrödinger equation, 170, 171
Standard hydrogen electrode potential, 388 Total differential, 19
Standard state, 30, 31, 45, 49, 51, 382, 386 Totally symmetric vibrations
State variables, 3, 23, 24, 28, 31 and Raman depolarization ratio, 236
Steepest descent method, 270 in XYZ3 molecules, 237
Stimulated emission of radiation, 177, 180, Translations, molecular, 124
208, 229 Translation vectors
STO-3G, 185, 187, 257, 263–265, 283, 289, for body-centered cubic, 298, 305, 306, 308
402 for face-centered cubic, 305, 306, 307, 331
Stoichiometric coefficients, 41, 77, 81, 87, 112 for hexagonal close packed, 307, 308, 320
Surface tension for simple cubic, 305, 306, 330
capillary action, 56 Triple point, 52, 54, 65, 363
definition, 55 Triplet, 143
liquid droplet in air, 55, 56
Symmetry and molecular vibrations U
isomers of 1,2-dichloroethene, 230 Uncertainty principle, 139, 181
XY3 molecules, 234–236 Unidimensional model, 315
Symmetry element, 246
axis of rotation, 214, 246 V
center of symmetry, 214, 217 Valence basis set
plane of reflection, 214, 215, 222 double-zeta, 187
rotation–reflection axis, 215 triple-zeta, 188, 267
Symmetry operation, 106, 213–217, 220, van der Waals equation, 79
222–228, 231, 245, 246, 258, 261, 278, critical temperature, 14, 71, 72, 79, 82
346 isotherm, 14, 16, 71, 79, 82, 83
System Maxwell construction, 15
closed, 25, 27, 39–41, 43, 47, 50, 52, 54, reduced variables, 16, 71, 82, 83
58, 76, 80, 365 van der Waals constants, 13, 15, 16, 72, 79,
isolated, 23, 24, 27, 38, 78, 82, 84 82
open, 78, 82 Variable
System of differential equations, 99, 102, 107 extensive, 23, 28, 49, 316
System of equations intensive, 23, 28, 40, 53, 316
homogeneous, 145, 152, 198 Variational derivative, 157, 206
underdetermined, 145 Variational method, 143, 163
Index 457
Vibrations self-consistent reaction field, 151

molecular, 123, 146, 230, 232, 233, 237 simple nonpolar solutes, 367, 368
symmetry species, 230, 233, 236, 237 sodium decanoate micelles, 1H NMR, 379
of XYZ3 molecules, 237 supercritical fluid, 52, 365
surfactant, 376, 377, 392
W thermodynamic functions for solutions, 16,
Water 371
acid ionization constant, 383, 386 titration curve, universal,, 385
acids and bases, 381–383, 386, 387 Water dimer
activity coefficient, mean, 373, 375 atomic charges, 357
amphipathic molecules, 376, 377, 381, 392 electrostatic potentials, 357–359
as solvent, 365, 373 hydrogen bond, 357, 358
autoionization, 382, 383 intermolecular vibrations, 358
boiling temperature, 65 polarizability, 358
buffer capacity, 391 Water molecule
change of concentration scales, 86, 88, 96, antisymmetric stretching, 348
112 atomic charges, 339, 344
coalescence of nonpolar solute cavities, bending, 358
371, 372 C2v character table, 225, 346, 348
critical micelle concentration (CMC), 377, dipole and quadrupole moments, 344
378, 380, 381, 392, 399 electric field streamlines, 343, 344
critical point, 15, 364 enthalpy of formation, 188, 339
Debye–Hückel limiting law, 374, 392 geometry, 267, 272, 288, 338, 376
Debye length, 374, 392 H to D isotopic substitution,, 271
electric permittivity, 365 natural bond orbitals, 263, 264, 347
gas hydrates, 360, 361, 368 normal vibrational modes, 348
hydrogen bond model, 266, 355–358, 360, orbitals, 184, 185, 314, 346,
361, 362, 365, 367, 368, 371, 372, 397, 347, 349
398 symmetric stretching, 348, 358, 402
hydrophobic interaction, 371, 372 thermal energy, 317, 318, 339
ice Ih, 359 vibrations (H2O, DHO), 348, 349, 402
ice polymorphism, 360 Wave equation, one-dimensional, 118, 192
ionic atmosphere, 374, 375 Wavenumber
ionic solutions, 372, 374, 375, 392 angular, 119, 174, 281
ionic strength, 375, 392 linear, 118
Luzar's model, 398 Well-behaved function, 118–120
melting temperature, 37, 65, 360 Work, mechanical, 25, 26
micelle, 377, 379, 380, 392, 398
model of equilibrium between micelles and X
monomers, 377 X-ray diffraction
nonpolar solutes, 367, 371 optical transform, 313, 314, 334
Ostwald coefficient, 368, 370, 392 pattern, 313, 334
permittivity, 367
phase diagram, 360, 363 Z
polarizable continuum model, 182, 183 Zero-point energy, 129, 147, 274
relative permittivity, 372 Zinc blende, 310, 311
relaxation processes, 377

Molecular Physical Chemistry: José J.C. Teixeira-Dias

Uploaded by

Copyright:

Available Formats

Molecular Physical Chemistry: José J.C. Teixeira-Dias

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

Molecular Physical Chemistry: José J.C. Teixeira-Dias

Uploaded by

Copyright:

Available Formats

José J.C.

ISBN 978-3-319-41092-0 ISBN 978-3-319-41093-7 (eBook)

Library of Congress Control Number: 2016947376

© Springer International Publishing Switzerland 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

This book contains chapters on thermodynamics, chemical kinetics, quantum

Aveiro, Portugal José J.C. Teixeira-Dias

1.12 Gibbs Energy. . . . . . . . . . . . . . . . . . . . . . . . ............ 47

3 The Schrödinger Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

3.1.4 Important Theorems . . . . . . . . . . . . . . . . . . . . . . . . 118

4 Molecular Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

5 Molecular Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

7.9 Ice Ih . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

Wolfram MathWorld (mathworld.wolfram.com)

© Springer International Publishing Switzerland 2017 1

commands. Some of these explanations present suggestions for the student to

1.1 Ideal Gas

V ¼ Tf2 ðpÞ ð1:2Þ

In 1811, Amedeo Avogadro (1776–1856) put forward a hypothesis according to

f1 ðTÞ ¼ constant nT ¼ nRT ð1:4Þ

A gas whose pressure p, volume V, temperature T, and amount of substance

Dividing both members of (1.7) by V gives

where qN is the number density, given by

1.2 Kinetic Model of Gases

Fig. 1.1 Molecules of a gas

(1818–1889), James Clerk Maxwell (1831–1879) and Ludwig Boltzmann (1844–

1.2.1 Pressure and Temperature

Fig. 1.2 In the time interval

mv2x DtAqN ð1:11Þ

mv2x AqN ð1:12Þ

Fig. 1.3 Molecule in elastic

The velocity axes are orthogonal, and so we can write

v2 ¼ v2x þ v2y þ v2z ð1:14Þ

For an isotropic space,

Substitution of (1.15) in the second member of (1.13) leads to

and we can write

Substitution of (1.18) in (1.8) leads to

This equation shows that temperature is proportional to the ensemble average of

1.2.2 Distribution of Velocities

PðvÞdvx dvy dvz ð1:20Þ

Fig. 1.4 Velocity space

(E1). Therefore, the probability density is

PðvÞ ¼ Pðvx ÞPðvy ÞPðvz Þ ð1:24Þ

Fig. 1.5 P(vx) at 100 and

For an isotropic space,

Pðvx Þ ¼ Pðvy Þ ¼ Pðvz Þ ð1:26Þ

Fig. 1.6 Distributions of

At a deﬁned temperature, the speed that corresponds to the maximum of f(v) is

At 273.15 K, the most probable speed for O2 is 377 m s−1.

At 273.15 K, the mean molecular speed for O2 is 425 m s−1.

For O2 at 273.15 K, the rms speed is equal to 461 m s−1 (E2).

1.2.3 Mean Free Path

Vint ¼ pd2 d ð1:34Þ

Substitution of (1.34) in (1.33) and then in (1.32) leads to

dd ¼ ðvrel Þrms Dt ð1:37Þ

It can be shown (see §2) that

dn vrms Dt pffiffiffi 1

Substitution of (1.42) in (1.39) leads to

dn vrms Dt pffiffiffi 1

dQp ¼ ðdU þ pdVÞp ¼ ðdU þ pdV þ VdpÞp ¼ ½dðU þ pVÞp ð1:91Þ