Infrared and Raman Spectra of Inorganic and Coordination Compounds
Infrared and Raman Spectra of Inorganic and Coordi...

Author:
Kazuo Nakamoto

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Infrared and Raman Spectra of Inorganic and Coordination Compounds

Infrared and Raman Spectra of Inorganic and Coordination Compounds Part A: Theory and Applications in Inorganic Chemistry Sixth Edition

Kazuo Nakamoto Wehr Professor Emeritus of Chemistry Marquette University

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright # 2009 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data is available.

ISBN 978-0-471-74339-2

Printed in the United States of America 10 9

8 7 6 5

4 3 2 1

Contents

PREFACE TO THE SIXTH EDITION

ix

ABBREVIATIONS

xi

Chapter 1. Theory of Normal Vibrations 1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 1.7. 1.8. 1.9. 1.10. 1.11. 1.12. 1.13. 1.14. 1.15. 1.16. 1.17.

1

Origin of Molecular Spectra / 1 Origin of Infrared and Raman Spectra / 5 Vibration of a Diatomic Molecule / 9 Normal Coordinates and Normal Vibrations / 15 Symmetry Elements and Point Groups / 21 Symmetry of Normal Vibrations and Selection Rules / 25 Introduction to Group Theory / 34 The Number of Normal Vibrations for Each Species / 39 Internal Coordinates / 46 Selection Rules for Infrared and Raman Spectra / 49 Structure Determination / 56 Principle of the GF Matrix Method / 58 Utilization of Symmetry Properties / 65 Potential Fields and Force Constants / 71 Solution of the Secular Equation / 75 Vibrational Frequencies of Isotopic Molecules / 77 Metal–Isotope Spectroscopy / 79 v

vi

CONTENTS

1.18. 1.19. 1.20. 1.21. 1.22. 1.23. 1.24. 1.25. 1.26. 1.27. 1.28. 1.29. 1.30. 1.31. 1.32.

Group Frequencies and Band Assignments / 82 Intensity of Infrared Absorption / 88 Depolarization of Raman Lines / 90 Intensity of Raman Scattering / 94 Principle of Resonance Raman Spectroscopy / 98 Resonance Raman Spectra / 101 Theoretical Calculation of Vibrational Frequencies / 106 Vibrational Spectra in Gaseous Phase and Inert Gas Matrices / 109 Matrix Cocondensation Reactions / 112 Symmetry in Crystals / 115 Vibrational Analysis of Crystals / 119 The Correlation Method / 124 Lattice Vibrations / 129 Polarized Spectra of Single Crystals / 133 Vibrational Analysis of Ceramic Superconductors / 136 References / 141

Chapter 2. Applications in Inorganic Chemistry 2.1. 2.2. 2.3. 2.4. 2.5. 2.6. 2.7. 2.8. 2.9. 2.10. 2.11. 2.12. 2.13. 2.14. 2.15. 2.16. 2.17. 2.18.

Diatomic Molecules / 149 Triatomic Molecules / 159 Pyramidal Four-Atom Molecules / 173 Planar Four-Atom Molecules / 180 Other Four-Atom Molecules / 187 Tetrahedral and Square–Planar Five-Atom Molecules / 192 Trigonal–Bipyramidal and Tetragonal–Pyramidal XY5 and Related Molecules / 213 Octahedral Molecules / 221 XY7 and XY8 Molecules / 237 X2Y4 and X2Y6 Molecules / 239 X2Y7, X2Y8, X2Y9, and X2Y10 Molecules / 244 Metal Cluster Compounds / 250 Compounds of Boron / 254 Compounds of Carbon / 258 Compounds of Silicon, Germanium, and Other Group IVB Elements / 276 Compounds of Nitrogen / 279 Compounds of Phosphorus and Other Group VB Elements / 285 Compounds of Sulfur and Selenium / 292

149

CONTENTS

vii

2.19. Compounds of Halogen / 296 References / 299 Appendixes

355

I. Point Groups and Their Character Tables / 355 II. Matrix Algebra / 368 III. General Formulas for Calculating the Number of Normal Vibrations in Each Species / 373 IV. Direct Products of Irreducible Representations / 377 V. Number of Infrared- and Raman-Active Stretching Vibrations for MXn Ym -Type Molecules / 378 VI. Derivation of Eq. 1.113 / 379 VII. The G and F Matrix Elements of Typical Molecules / 382 VIII. Group Frequency Charts / 388 IX. Correlation Tables / 393 X. Site Symmetry for the 230 Space Groups / 407 Index

415

Preface to the Sixth Edition

Since the fifth edition was published in 1996, a number of new developments have been made in the field of infrared and Raman spectra of inorganic and coordination compounds. The sixth edition is intended to emphasize new important developments as well as to catch up with the ever-increasing new literature. Major changes are described below. Part A. Chapter 1 (‘‘Theory of Normal Vibrations”) includes two new sections. Section 1.24 explains the procedure for calculating vibrational frequencies on the basis of density functional theory (DFT). The DFT method is currently used almost routinely to determine molecular structures and to calculate vibrational parameters. Section 1.26 describes new developments in matrix cocondensation techniques. More recently, a large number of novel inorganic and coordination compounds have been prepared by using this technique, and their structures have been determined and vibrational assignments have been made on the basis of results of DFT calculations. Chapter 2 (‘‘Applications in Inorganic Chemistry”) has been updated extensively, resulting in a total number of references of over 1800. In particular, sections on triangular X3- and tetrahedral X4-type molecules have been added as Secs. 2.2 and 2.5, respectively. In Sec. 2.8, the rotational–vibrational spectrum of the octahedral UF6 molecule is shown to demonstrate how an extremely small metal isotope shift by 235U/238U substitution (only 0.6040 cm 1) can be measured. Section 2.14 (‘‘Compounds of Carbon”) has been expanded to show significant applications of vibrational spectroscopy to the structural determination of fullerences, endohedral fullerenes, and carbon nanotubes. Vibrational data on a number of novel inorganic compounds prepared most recently have been added throughout Chapter 2. Part B. Chapter 3 (‘‘Applications in Coordination Chemistry”) contains two new Sections: Sec. 3.6 (‘‘Metallochlorins, Chlorophylls, and Metallophthalocyanines”) ix

x

PREFACE TO THE SIXTH EDITION

and Sec. 3.19 (‘‘Complexes of Carbon Dioxide”). The total number of references has approached 1700 because of substantial expansion of other sections such as Secs. 3.5, 3.18, 3.20, 3.22, and 3.28. Chapter 4 (‘‘Applications in Organometallic Chemistry”) includes new types of organometallic compounds obtained by matrix cocondensation techniques (Sec. 4.1). In Chapter 5 (‘‘Applications in Bioinorganic Chemistry”), a new section (Sec. 5.4) has been added, and several sections such as Secs. 5.3, 5.7, and 5.9 have been expanded to include many important new developments. I would like to express my sincere thanks to all who helped me in preparing this edition. Special thanks go to Prof. E. L. Varetti (University of La Plata, Argentina), Prof. S. Guha (University of Missouri, Columbia), and Prof. L. Andrews (University of Virginia) for their help in writing new sections in Chapter 1 of Part A. My thanks also go to all the authors and publishers who gave me permission to reproduce their figures in this and previous editions. Finally, I would like to thank the staff of John Raynor Science Library of Marquette University for their help in collecting new references. Milwaukee, Wisconsin March 2008

KAZUO NAKAMOTO

Abbreviations

Several different groups of acronyms and other abbreviations are used: 1. IR, infrared; R, Raman; RR, resonance Raman; p, polarized; dp, depolarized; ap, anomalous polarization; ia, inactive. 2. n, stretching; d, in-plane bending or deformation; rw, wagging; rr, rocking; rt, twisting; p, out-of-plane bending. Subscripts, a, s, and d denote antisymmetric, symmetric, and degenerate modes, respectively. Approximate normal modes of vibration corresponding to these vibrations are given in Figs. 1.25 and 1.26. 3. DFT, density functional theory; NCA, normal coordinate analysis; GVF, generalized valence force field; UBF, Urey–Bradley force field. 4. M, metal; L, ligand; X, halogen; R, alkyl group. 5. g, gas; l, liquid; s, solid; m or mat, matrix; sol’n or sl, solution; (gr) or (ex), ground or excited state. 6. Me, methyl; Et, ethyl; Pr, propyl; Bu, butyl; Ph, phenyl; Cp, cyclopentadienyl; OAc, acetate ion; py, pyridine; pic, pycoline; en, ethylenediamine. Abbreviations of other ligands are given when they appear in the text. In the tables of observed frequencies, values in parentheses are calculated or estimated values unless otherwise stated.

xi

Chapter

1

Theory of Normal Vibrations 1.1. ORIGIN OF MOLECULAR SPECTRA As a first approximation, the energy of the molecule can be separated into three additive components associated with (1) the motion of the electrons in the molecule,* (2) the vibrations of the constituent atoms, and (3) the rotation of the molecule as a whole: Etotal ¼ Eel þ Evib þ Erot

ð1:1Þ

The basis for this separation lies in the fact that electronic transitions occur on a much shorter timescale, and rotational transitions occur on a much longer timescale, than 10 vibrational transitions. The translational energy of the molecule may be ignored in this discussion because it is essentially not quantized. If a molecule is placed in an electromagnetic field (e.g., light), a transfer of energy from the field to the molecule will occur when Bohr’s frequency condition is satisfied: DE ¼ hn

ð1:2Þ

*Hereafter the word molecule may also represent an ion. Infrared and Raman Spectra of Inorganic and Coordination Compounds, Sixth Edition, Part A: Theory and Applications in Inorganic Chemistry, by Kazuo Nakamoto Copyright 2009 John Wiley & Sons, Inc.

1

2

THEORY OF NORMAL VIBRATIONS

where DE is the difference in energy between two quantized states, h is Planck’s constant (6.625 1027 erg s), and n is the frequency of the light. Here, the frequency is the number of electromagnetic waves in the distance that light travels in one second: c ð1:3Þ n¼ l where c is the velocity of light (3 1010 cm s1) and l is the wavelength of the electromagnetic wave. If l has the units of cm, n has dimensions of (cm s1)/cm ¼ s1, which is also called “Hertz (Hz).” The wavenumber ð~nÞ defined by 1 ~n ¼ ð1:4Þ l is most commonly used in vibrational spectroscopy. It has the dimension of cm1. By combining Eqs. 1.3 and 1.4, we obtain ~n ¼

1 n ¼ l c

or n ¼

c ¼ c~n l

ð1:5Þ

Although the dimensions of n and ~n differ from one another, it is convenient to use them interchangeably. Thus, an expression such as “a frequency shift of 5 cm1” is used throughout this book. Using Eq. 1.5, Bohr’s condition (Eq. 1.2) is written as DE ¼ hc~n

ð1:6Þ

Since h and c are known constants, DE can be expressed in units such as* 1ðcm 1 Þ ¼ 1:99 10 16 ðerg molecule 1 Þ ¼ 2:86 ðcal mol 1 Þ ¼ 1:24 10 4 ðeV molecule 1 Þ Suppose that DE ¼ E2 E1

ð1:7Þ

where E2 and E1 are the energies of the excited and ground states, respectively. Then, the molecule “absorbs” DE when it is excited from E1 to E2 and “emits” DE when it reverts from E2 to E1. Figure 1.1 shows the regions of the electromagnetic spectrum where DE is indicated in ~n, l, and n. In this book, we are concerned mainly with vibrational transitions which are observed in infrared (IR) or Raman (R) spectra. *Use conversion factors such as Avogadro’s number N0 ¼ 6:023 1023 ðmol 1 Þ 1 ðcalÞ ¼ 4:1846 107 ðergÞ 1 ðeVÞ ¼ 1:6021 10 12 ðerg molecule 1 Þ

ORIGIN OF MOLECULAR SPECTRA

3

Fig. 1.1. Regions of the electromagnetic spectrum and energy units.

These transitions appear in the 102 104 cm1 region, and originate from vibrations of nuclei constituting the molecule. Rotational transitions occur in the 1–102 cm1 region (microwave region) because rotational levels are relatively close to each other, whereas electronic transitions are observed in the 104–106 cm1 region (UV–visible region) because their energy levels are far apart. However, such division is somewhat arbitrary, for pure rotational spectra may appear in the far-infrared region if transitions to higher excited states are involved, and pure electronic transitions may appear in the near-infrared region if electronic levels are closely spaced. Figure 1.2 illustrates transitions of the three types mentioned for a diatomic molecule. As the figure shows, rotational intervals tend to increase as the rotational quantum number J increases, whereas vibrational intervals tend to decrease as the vibrational quantum number v increases. The dashed line below each electronic level indicates the “zero-point energy” that must exist even at a temperature of absolute zero as a result of Heisenberg’s uncertainty principle: 1 E0 ¼ hn ð1:8Þ 2 It should be emphasized that not all transitions between these levels are possible. To see whether the transition is “allowed” or “forbidden,” the relevant selection rule must be examined. This, in turn, is determined by the symmetry of the molecule. As expected from Fig. 1.2, electronic spectra are very complicated because they are accompanied by vibrational as well as rotational fine structure. The rotational fine structure in the electronic spectrum can be observed if a molecule is simple and the spectrum is measured in the gaseous state under high resolution. The vibrational fine structure of the electronic spectrum is easier to observe than the rotational fine structure, and can provide structural and bonding information about molecules in electronic excited states. Vibrational spectra are accompanied by rotational transitions. Figure 1.3 shows the rotational fine structure observed for the gaseous ammonia molecule. In most polyatomic molecules, however, such a rotational fine structure is not observed

4

THEORY OF NORMAL VIBRATIONS

1

υ” = 0

Zero point energy

Electronic excited state

Pure electronic transition

4

3

2 6 4 2 J” = 0 6 4 2 J” = 0

1

Pure rotational transition

Pure vibrational transition υ’ = 0

Zero point energy

Electronic ground state

Fig. 1.2. Energy level of a diatomic molecule (the actual spacings of electronic levels are much larger, and those of rotational levels are much smaller, than that shown in the figure).

Fig. 1.3. Rotational fine structure of gaseous NH3.

ORIGIN OF INFRARED AND RAMAN SPECTRA

5

because the rotational levels are closely spaced as a result of relatively large moments of inertia. Vibrational spectra obtained in solution do not exhibit rotational fine structure, since molecular collisions occur before a rotation is completed and the levels of the individual molecules are perturbed differently. The selection rule allows any transitions corresponding to Dv ¼ 1 if the molecule is assumed to be a harmonic oscillator (Sec. 1.3), Under ordinary conditions, however, only the fundamentals that originate in the transition from v ¼ 0 to v ¼ 1 in the electronic ground state can be observed. This is because the Maxwell–Boltzmann distribution law requires that the ratio of population at v ¼ 0 and v ¼ 1 states is given by R¼

Pðv ¼ 1Þ ¼ e DEv =kT Pðv ¼ 0Þ

ð1:9Þ

where DEv is the vibrational frequency (cm1) and kT ¼ 208 (cm1) at room temperature. In the case of H2, DEv ¼ 4160 cm1 and R ¼ 2.16 109. Thus, almost all molecules are at v ¼ 0. However, the population at v ¼ 1 increases as DEv becomes small. For example, R ¼ 0.36 for I2 (DEv ¼ 213 cm1) Then, about 27% of the molecules are at v ¼ 1 state, and the transition from v ¼ 1 to v ¼ 2 can be observed as a “hot band.” In addition to the harmonic oscillator selection rule, another restriction results from the symmetry of the molecule (Sec. 1.10). Thus, the number of allowed transitions in polyatomic molecules is greatly reduced. Overtones and combination bands* of these fundamentals are forbidden by the selection rule. However, they are weakly observed in the spectrum because of the anharmonicity of the vibration (Sec. 1.3). Since they are less important than the fundamentals, they will be discussed only when necessary. 1.2. ORIGIN OF INFRARED AND RAMAN SPECTRA As stated previously, vibrational transitions can be observed as infrared (IR) or Raman spectra.† However, the physical origins of these two spectra are markedly different. Infrared (absorption) spectra originate in photons in the infrared region that are absorbed by transitions between two vibrational levels of the molecule in the electronic ground state. On the other hand, Raman spectra have their origin in the electronic polarization caused by ultraviolet, visible, and near-IR light. If a molecule is irradiated by monochromatic light of frequency n (laser), then, because of electronic polarization induced in the molecule by this incident beam, the light of frequency n (“Rayleigh scattering”) as well as that of frequency n ni (“Raman scattering”) is scattered where ni represents a vibrational frequency of the molecule. Thus, Raman spectra are presented as shifts from the incident frequency in the ultraviolet, visible, and nearIR region. Figure 1.4 illustrates the difference between IR and Raman techniques. *Overtones represent some multiples of the fundamental, whereas combination bands arise from the sum or difference of two or more fundamentals. † Raman spectra were first observed by C. V. Raman [Indian J. Phys. 2, 387 (1928)] and C. V. Raman and K. S. Krishnan [Nature 121, 501 (1928)].

6

THEORY OF NORMAL VIBRATIONS

Fig. 1.4. Mechanisms of infrared absorption and Raman scattering.

Although Raman scattering is much weaker than Rayleigh scattering (by a factor of 103–105), it is still possible to observe the former by using a strong exciting source. In the past, the mercury lines at 435.8 nm (22.938 cm1) and 404.7 nm (24,705 cm1) from a low-pressure mercury arc were used to observe Raman scattering. However, the advent of lasers revolutionized Raman spectroscopy. Lasers provide strong, coherent monochromatic light in a wide range of wavelengths, as listed in Table 1.1. In the case of resonance Raman spectroscopy (Sec. 1.22), the exciting frequency is chosen so as to fall inside the electronic absorption band. The degree of resonance enhancement varies as a function of the exciting frequency and reaches a maximum when the exciting frequency coincides with that of the electronic absorption maximum. It is possible to change the exciting frequency continuously by pumping dye lasers with powerful gas or pulsed lasers. The origin of Raman spectra can be explained by an elementary classical theory. Consider a light wave of frequency n with an electric field strength E. Since E fluctuates at frequency n, we can write E ¼ E0 cos 2pnt

ð1:10Þ

where E0 is the amplitude and t the time. If a diatomic molecule is irradiated by this light, the dipole moment P given by P ¼ aE ¼ aE0 cos 2pnt

ð1:11Þ

ORIGIN OF INFRARED AND RAMAN SPECTRA

7

TABLE 1.1. Some Representative Laser Lines for Raman Spectroscopy Lasera

Mode

Gas lasers Ar–ion

CW

Kr–ion

CW

He–Ne He–Cd Nitrogen Excimer (XeCl)

CW CW Pulsed

Wavelength (nm) 488.0 (blue) 514.5 (green) 413.1 (violet) 530.9 (green/yellow) 647.1 (red) 632.8 (red) 441.6 (blue/violet) 337.1 (UV) 308 (UV)

~n ðcm 1 Þ 20491.8 19436.3 24207.2 18835.9 15453.6 15802.8 22644.9 29664.7 32467.5

Solid-state lasers CW or pulsed 1064 (near-IR) 9398.4 Nd:YAGb Liquid lasers A variety of dye solutions are pumped by strong CW or pulsed-laser sources; a wide range (440–800 nm) can be covered continuously by choosing proper organic dyes a

Acronym of light amplification by stimulated emission of radiation. Acronym of neodymium-doped yttrium aluminum garnet.

b

Source: For more information, see Nakamoto and collegues [21,26].

is induced. Here a is a proportionality constant and is called the polarizability. If the molecule is vibrating with frequency ni, the nuclear displacement q is written as q ¼ q0 cos 2pni t

ð1:12Þ

where q0 is the vibrational amplitude. For small amplitudes of vibration, a is a linear function of q. Thus, we can write ! qa a ¼ a0 þ q ð1:13Þ qq 0

Here, a0 is the polarizability at the equilibrium position, and (qa/qq)0 is the rate of change of a with respect to the change in q, evaluated at the equilibrium position. If we combine Eqs. 1.11–1.13, we have P ¼ aE0 cos 2pnt

! qa ¼ a0 E0 cos 2pnt þ q0 E0 cos 2pnt cos 2pni t qq 0

¼ a0 E0 cos 2pnt ! 1 qa þ q0 E0 fcos½2pðn þ ni Þt þ cos½2pðn ni Þtg 2 qq 0

ð1:14Þ

8

THEORY OF NORMAL VIBRATIONS

Resonance Raman

E1

ν + ν1

ν

ν

ν – ν1

Normal Raman

υ=1 υ=0

S

A

S

A

E0

Fig. 1.5. Mechanisms of normal and resonance Raman scattering. S and A denote Stokes and anti-Stokes scattering, respectively. The shaded areas indicate the broadening of rotationalvibrational levels in the liquid and solid states (Sec. 1.22).

According to classical theory, the first term describes an oscillating dipole that radiates light of frequency n (Rayleigh scattering). The second term gives the Raman scattering of frequencies n þ ni (anti-Stokes) and n ni (Stokes). If (qa/qq)0 is zero, the second term vanishes. Thus, the vibration is not Raman-active unless the polarizability changes during the vibration. Figure 1.5 illustrates the mechanisms of normal and resonance Raman (RR) scattering. In the former, the energy of the exciting line falls far below that required to excite the first electronic transition. In the latter, the energy of the exciting line coincides with that of an electronic transition.* If the photon is absorbed and then emitted during the process, it is called resonance fluorescence (RF). Although the conceptual difference between resonance Raman scattering and resonance fluorescence is subtle, there are several experimental differences which can be used to *If the exciting line is close to but not inside an electronic absorption band, the process is called preresonance Raman scattering.

VIBRATION OF A DIATOMIC MOLECULE

9

Fig. 1.6. Raman spectrum of CCI4 (488.0 nm excitation).

distinguish between these two phenomena. For example, in RF spectra all lines are depolarized, whereas in RR spectra some are polarized and others are depolarized. Additionally, RR bands tend to be broad and weak compared with RF bands [66,67]. In the case of Stokes lines, the molecule at v ¼ 0 is excited to the v ¼ 1 state by scattering light of frequency n ni. Anti-Stokes lines arise when the molecule initially in the v ¼ 1 state scatters radiation of frequency n þ n1 and reverts to the v ¼ 0 state. Since the population of molecules is larger at v ¼ 0 than at v ¼ 1 (Maxwell–Bottzmann distribution law), the Stokes lines are always stronger than the anti-Stokes lines. Thus, it is customary to measure Stokes lines in Raman spectroscopy. Figure 1.6 illustrates the Raman spectrum (below 500 cm1) of CCI4 excited by the blue line (488.0 nm) of an argon ion laser.

1.3. VIBRATION OF A DIATOMIC MOLECULE Through quantum mechanical considerations [2,7], the vibration of two nuclei in a diatomic molecule can be reduced to the motion of a single particle of mass m, whose displacement q from its equilibrium position is equal to the change of the internuclear distance. The mass m is called the reduced mass and is represented by 1 1 1 ¼ þ m m1 m2

ð1:15Þ

10

THEORY OF NORMAL VIBRATIONS

Fig. 1.7. Potential energy curves for a diatomic molecule: actual potential (solid line), parabolic potential (dashed line), and cubic potential (dotted line).

where m1 and m2 are the masses of the two nuclei. The kinetic energy is then 1 1 2 p T ¼ mq_ 2 ¼ 2 2m

ð1:16Þ

_ If a simple parabolic potential function such as where p is the conjugate momentum mq. that shown in Fig. 1.7 is assumed, the system represents a harmonic oscillator, and the potential energy is simply given by. 1 V ¼ Kq2 2

ð1:17Þ

Here K is the force constant for the vibration. Then the Schr€odinger wave equation becomes ! d2 c 8p2 m 1 2 þ 2 E Kq c ¼ 0 ð1:18Þ dq2 h 2 If this equation is solved with the condition that c must be single-valued, finite, and continuous, the eigenvalues are 1 1 Ev ¼ hn v þ ¼ hc~n v þ 2 2

ð1:19Þ

VIBRATION OF A DIATOMIC MOLECULE

with the frequency of vibration 1 n¼ 2p

sﬃﬃﬃﬃ sﬃﬃﬃﬃ K 1 K or ~n ¼ m 2pc m

11

ð1:20Þ

Here v is the vibrational quantum number, and it can have the values 0, 1, 2, 3, and so on. The corresponding eigenfunctions are pﬃﬃﬃ ða=pÞ1=4 aq2 =2 ﬃ e Hv ð aqÞ ð1:21Þ cv ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ v 2 v! pﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃ where a ¼ 2p mK =h ¼ 4p2 mn=h, and Hv ð aqÞ is a Hermite polynomial of the vth degree. Thus the eigenvalues and the corresponding eigenfunctions are 1 E0 ¼ hn; 2 3 E1 ¼ hn; 2

c0 ¼ ða=pÞ1=4 e aq =2 2

c1 ¼ ða=pÞ1=4 21=2 qe aq =2 2

ð1:22Þ

.. .

.. .

As Fig. 1.7 shows, actual potential curves can be approximated more exactly by adding a cubic term [2]: 1 V ¼ Kq2 Gq3 2

ðK GÞ

Then, the eigenvalues become 1 1 2 þ Ev ¼ hcwe v þ hcxe we v þ 2 2

ð1:23Þ

ð1:24Þ

where we is the wavenumber corrected for anharmonicity and xewe indicates the magnitude of anharmonicity. Table 2.1a (in Chapter 2) lists we and xewe for a number of diatomic molecules. Equation 1.24 shows that the energy levels of the anharmonic oscillator are not equidistant, and the separation decreases slowly as v increases. This anharmonicity is responsible for the appearance of overtones and combination vibrations, which are forbidden in the harmonic oscillator. The values of xe and xewe can be determined by observing a series of overtone bands in IR and Raman spectra. From Eq. 1.24, we obtain Ev E 0 ¼ vwe xe we ðv2 þ vÞ þ hc Then Fundamental: First overtone: Second overtone:

~n1 ¼ we 2xe we ~n2 ¼ 2we 6xe we ~n3 ¼ 3we 12xe we

12

THEORY OF NORMAL VIBRATIONS

For H35C1, these transitions are observed at 2885.9, 5668.1, and 8347.0 cm1, respectively, in IR spectrum [2]. Using these values, we find that we ¼ 2988:9 cm 1

and

xe we ¼ 52:05 cm 1

As will be shown in Sec. 1.23, a long series of overtone bands can be observed when Raman spectra of small molecules such as I2 and TiI4 are measured under rigorous resonance conditions. Anharmonicity constants can also be determined from the analysis of rotational fine structures of vibrational transitions [2]. Since the anharmonicity correction has not been made for most polyatomic molecules, in large part because of the complexity of the calculation, the frequencies given in Chapter 2 are not corrected for anharmonicity (except those given in Table 2.1a). According to Eq. 1.20, the wavenumber of the vibration in a diatomic molecule is given by sﬃﬃﬃﬃ 1 K ~n ¼ ð1:20Þ 2pc m A more exact expression is given by using the wavenumber corrected for anharmonicity: sﬃﬃﬃﬃ 1 K we ¼ ð1:25Þ 2pc m or K ¼ 4p2 c2 w2e m

ð1:26Þ

For HC1, we ¼ 2989 cm1 and m ¼ 0.9799 awu (where awu is the atomic weight unit). Thus, we obtain* K¼

4ð3:14Þ2 ð3 1010 Þ2 2 we m 6:025 1023

¼ ð5:8883 10 2 Þw2e m ¼ ð5:8883 10 2 Þð2989Þ2 ð0:9799Þ ¼ 5:16 105 ðdyn=cmÞ ¼ 5:16 ðmdyn=ÅÞ* Table 1.2 lists the observed frequencies ð~nÞ, wavenumbers corrected for anharmonicity (we), reduced masses (m), and force constants (K) for several series of diatomic *105 (dynes/cm) ¼ 105 (103 millidynes/108 Å) ¼ 1 (mdyn/Å) ¼ 102 N/m (SI unit).

13

VIBRATION OF A DIATOMIC MOLECULE

TABLE 1.2. Relationships between Vibrational Frequency, Reduced Mass, and Force Constant Molecule H2 HD D2 HF HCl HBr HI F2 Cl2 Br2 I2 N2 CO NO O2

Obs. ~n (cm1)

we (cm1)

4160 3632 2994 3962 2886 2558 2233 892 564 319 213 2331 2145 1877 1555

4395 3817 3118 4139 2989 2650 2310 — 565 323 215 2360 2170 1904 1580

m (awu)

˚) K (mdyn/A

0.5041 0.6719 1.0074 0.9573 0.9799 0.9956 1.002 9.5023 17.4814 39.958 63.466 7.004 6.8584 7.4688 8.000

5.73 5.77 5.77 9.65 5.16 4.12 3.12 4.45 3.19 2.46 1.76 22.9 19.0 15.8 11.8

molecules. In the first series, we decreases in the order H2 > HD > D2, because m increases in the same order, while K is almost constant (mass effect). In the second series, we decreses in the order HF > HC1 > HBr > HI, because K decreases in the same order, while m shows little change (force constant effect). In the third series, we decreases in the order F2 > Cl2 > Br2 > I2, because m increases and K decreases in the same order. In this case, both mass effect and force constant effect are operative. In the last series, we decreases in the order N2 > CO > NO > O2, mainly owing to the force constant effect. This may be attributed to the differences in bond order: NN > C O þ $ C ¼ O > N¼ O $ N ¼ O þ > O ¼ O $ O O þ More examples of similar series in diatomic molecules are found in Sec. 2.1. These simple rules, obtained for a diatomic molecule, are helpful in understanding the vibrational spectra of polyatomic molecules. Figure 1.8 indicates the relationship between the force constant and the dissociation energy for the three series listed in Table 1.2. In the series of hydrogen halides, the dissociation energy decreases almost linearly as the force constant decreases. Thus, the force constant may be used as a measure of the bond strength in this case. However, such a monotonic relationship does not hold for the other two series. This is not unexpected, because the force constant is a measure of the curvature of the potential well near the equilibrium position K¼

d2 V dq2

! ð1:27Þ q!0

14

THEORY OF NORMAL VIBRATIONS

Fig. 1.8. Relationships between force constants and dissociation energies in diatomic molecules.

whereas the dissociation energy De is given by the depth of the potential energy curve (Fig. 1.7). Thus, a large force constant means sharp curvature of the potential well near the bottom, but does not necessarily indicate a deep potential well. Usually, however, a larger force constant is an indication of a stronger bond if the nature of the bond is similar in a series. It is difficult to derive a general theoretical relationship between the force constant and the dissociation energy even for diatomic molecules. In the case of small molecules, attempts have been made to calculate the force constants by quantum mechanical methods. The principle of the method is to express the total electronic energy of a molecule as a function of nuclear displacements near the equilibrium position and to calculate its second derivatives, q2 V=qq2i , and so on for each displacement coordinate qi. In the past, ab initio calculations of force constants were made for small molecules such as HF, H2O, and NH3. The force constants thus obtained are in good agreement with those calculated from the analysis of vibrational spectra. More recent progress in computer technology has made it possible to extend this approach to more complex molecules (Sec. 1.24). There are several empirical relationships which relate the force constant to the bond distance. Badger’s rule [68] is given by K ¼ 1:86ðr dij Þ 3

ð1:28Þ

where r is the bond distance and dij has a fixed value for bonds between atoms of row i and j in the periodic table. Gordy’s rule [69] is expressed as !3=4 cA cB K ¼ aN þb ð1:29Þ d2

NORMAL COORDINATES AND NORMAL VIBRATIONS

15

where cA and cB are the electronegativities of atoms A and B constituting the bond, d is the bond distance, N is the bond order, and a and b are constants for certain broad classes of compounds. Herschbach and Laurie [70] modified Badger’s rule in the form r ¼ dij þ ðaij dij ÞK 1=3

ð1:30Þ

Here, aij and dij are constants for atoms of rows i and j in the periodic table. Another empirical relationship is found by plotting stretching frequencies against bond distances for a series of compounds having common bonds. Such plots are highly important in estimating the bond distance from the observed frequency. Typical examples are found in the OH O hydrogen-bonded compounds [71], carbon–oxygen and carbon–nitrogen bonded compounds [72], and molybdenum– oxygen [73] and vanadium–oxygen bonded compounds [74]. More examples are given in Sec. 2.2. 1.4. NORMAL COORDINATES AND NORMAL VIBRATIONS In diatomic molecules, the vibration of the nuclei occurs only along the line connecting two nuclei. In polyatomic molecules, however, the situation is much more complicated because all the nuclei perform their own harmonic oscillations. It can be shown, however, that any of these extremely complicated vibrations of the molecule may be represented as a superposition of a number of normal vibrations. Let the displacement of each nucleus be expressed in terms of rectangular coordinate systems with the origin of each system at the equilibrium position of each nucleus. Then the kinetic energy of an N-atom molecule would be expressed as 1X T¼ mN 2 N

"

dDxN dt

!2

dDyN þ dt

!2

dDzN þ dt

!2 # ð1:31Þ

If generalized coordinates such as q1 ¼

pﬃﬃﬃﬃﬃﬃ m1 Dx1 ;

q2 ¼

pﬃﬃﬃﬃﬃﬃ m1 Dy1 ;

q3 ¼

pﬃﬃﬃﬃﬃﬃ m1 Dz1

q4 ¼

pﬃﬃﬃﬃﬃﬃ m2 Dx2 ; . . .

ð1:32Þ

are used, the kinetic energy is simply written as

T¼

3N 1X q_ 2 2 i i

ð1:33Þ

16

THEORY OF NORMAL VIBRATIONS

The potential energy of the system is a complex function of all the coordinates involved. For small values of the displacements, it may be expanded in a Taylor’s series as ! ! 3N 3N X qV 1X q2 V Vðq1 ; q2 ; . . . ; q3N Þ ¼ V0 þ qi þ qi qj þ ð1:34Þ qqi 2 i;j qqi qqj i 0

0

where the derivatives are evaluated at qi ¼ 0, the equilibrium position. The constant term V0 can be taken as zero if the potential energy at qi ¼ 0 is taken as a standard. The (qV/qqi)0 terms also become zero, since V must be a minimum at qi ¼ 0. Thus, V may be represented by ! 3N 3N 1X q2 V 1X qi qj ¼ bij qi qj ð1:35Þ V¼ 2 i;j qqi qqj 2 i;j 0

neglecting higher-order terms. If the potential energy given by Eq. 1.35 did not include any cross-products such as qiqj, the problem could be solved directly by using Lagrange’s equation: ! d qT qV ¼ 0; þ dt qq_ i qqi

i ¼ 1; 2; . . . ; 3N

From Eqs. 1.33 and 1.35, Eq. 1.36 is written as X q€i þ bij qj ¼ 0; j ¼ 1; 2; . . . ; 3N

ð1:36Þ

ð1:37Þ

j

If bij ¼ 0 for i „ j, Eq. 1.37 becomes q€i þ bii qi ¼ 0

ð1:38Þ

and the solution is given by qi ¼

q0i sin

pﬃﬃﬃﬃﬃ bii t þ di

! ð1:39Þ

where q0i and di are the amplitude and the phase constant, respectively. Since, in general, this simplification is not applicable, the coordinates qi must be transformed into a set of new coordinates Qi through the relations X q1 ¼ B1i Qi q2 ¼

i X

B2i Qi

i

.. . qk ¼

X i

Bki Qi

ð1:40Þ

NORMAL COORDINATES AND NORMAL VIBRATIONS

17

The Qi are called normal coordinates for the system. By appropriate choice of the coefficients Bki, both the potential and the kinetic energies can be written as 1 X _2 Qi 2 i

ð1:41Þ

1X li Q2i 2 i

ð1:42Þ

T¼ V¼

without any cross-products. If Eqs. 1.41 and 1.42 are combined with Lagrange’s equation (Eq. 1.36), there results € þ li Qi ¼ 0 Q ð1:43Þ i The solution of this equation is given by Qi ¼

Q0i sin

pﬃﬃﬃﬃ li t þ di

! ð1:44Þ

and the frequency is ni ¼

1 pﬃﬃﬃﬃ li 2p

ð1:45Þ

Such a vibration is called a normal vibration. For the general N-atom molecule, it is obvious that the number of the normal vibrations is only 3N – 6, since six coordinates are required to describe the translational and rotational motion of the molecule as a whole. Linear molecules have 3N – 5 normal vibrations, as no rotational freedom exists around the molecular axis. Thus, the general form of the molecular vibration is a superposition of the 3N – 6 (or 3N – 5) normal vibrations given by Eq. 1.44. The physical meaning of the normal vibration may be demonstrated in the following way. As shown in Eq. 1.40, the original displacement coordinate is related to the normal coordinate by X qk ¼ Bki Qi ð1:40Þ i

Since all the normal vibrations are independent of each other, consideration may be limited to a special case in which only one normal vibration, subscripted by 1, is excited (i.e., Q01 „ 0; Q02 ¼ Q03 ¼ ¼ 0). Then, it follows from Eqs. 1.40 and 1.44 that pﬃﬃﬃﬃﬃ qk ¼ Bk1 Q1 ¼ Bk1 Q01 sinð l1 t þ d1 Þ pﬃﬃﬃﬃﬃ ¼ Ak1 sinð l1 t þ d1 Þ

ð1:46Þ

18

THEORY OF NORMAL VIBRATIONS

This relation holds for all k. Thus, it is seen that the excitation of one normal vibration of the system causes vibrations, given by Eq. 1.46, of all the nuclei in the system. In other words, in the normal vibration, all the nuclei move with the same frequency and in phase. This is true for any other normal vibration. Thus Eq. 1.46 may be written in the more general form pﬃﬃﬃ qk ¼ Ak sinð lt þ dÞ ð1:47Þ If Eq. 1.47 is combined with Eq. 1.37, there results lAk þ

X

bkj Aj ¼ 0

ð1:48Þ

j

This is a system of first-order simultaneous equations with respect to A. In order for all the As to be nonzero, we must solve b11 l b21 b31 . . .

b12 b22 l b32 .. .

b13 b23 b33 l .. .

... ... ... ¼ 0

ð1:49Þ

The order of this secular equation is equal to 3N. Suppose that one root, l1, is found for Eq. 1.49. If it is inserted in Eq. 1.48, Ak1, Ak2, . . . are obtained for all the nuclei. The same is true for the other roots of Eq. 1.49. Thus, the most general solution may be written as a superposition of all the normal vibrations: ! X pﬃﬃﬃﬃ 0 qk ¼ Bkl Ql sin ll t þ dl ð1:50Þ l

The general discussion developed above may be understood more easily if we apply it to a simple molecule such as CO2, which is constrained to move in only one direction. If the mass and the displacement of each atom are defined as follows:

the potential energy is given by V¼

i 1 h k ðDx1 Dx2 Þ2 þ ðDx2 Dx3 Þ2 2

ð1:51Þ

Considering that m1 ¼ m3, we find that the kinetic energy is written as T¼

1 1 m1 ðDx_ 21 þ Dx_ 23 Þ þ m2 Dx_ 22 2 2

ð1:52Þ

NORMAL COORDINATES AND NORMAL VIBRATIONS

19

Using the generalized coordinates defined by Eq. 1.32, we may rewrite these energies as "

q1 q2 pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ m1 m2

2V ¼ k

2T ¼

!2

q2 q3 þ pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ m2 m1

X

!2 #

q_ 2i

ð1:53Þ

ð1:54Þ

From comparison of Eq. 1.53 with Eq. 1.35, we obtain b11 ¼

k ; m1

b22 ¼

2k m2

k k b12 ¼ b21 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ; b23 ¼ b32 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2 m1 m2 b13 ¼ b31 ¼ 0;

b33 ¼

k m1

If these terms are inserted in Eq. 1.49, we obtain the following result: k m1 l k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2 0

k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2 2k l m2 k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2

k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0 m1 m2 k l m1 0

By solving this secular equation, we obtain three roots: l1 ¼

k ; m1

l2 ¼ km;

l3 ¼ 0

where m¼

2m1 þ m2 m1 m2

Equation 1.48 gives the following three equations: lA1 þ b11 A1 þ b12 A2 þ b13 A3 ¼ 0 lA2 þ b21 A1 þ b22 A2 þ b23 A3 ¼ 0 lA3 þ b31 A1 þ b32 A2 þ b33 A3 ¼ 0

ð1:55Þ

20

THEORY OF NORMAL VIBRATIONS

Using Eq. 1.47, we rewrite these as ðb11 lÞq1 þ b12 q2 þ b13 q3 ¼ 0 b21 q1 þ ðb22 lÞq2 þ b23 q3 ¼ 0 b31 q1 þ b32 q2 þ ðb33 lÞq3 ¼ 0 If l1 ¼ k/m1 is inserted in the simultaneous equations above, we obtain q1 ¼ q3 ;

q2 ¼ 0

Similar calculations give q1 ¼ q3 ;

q1 ¼ q3 ;

sﬃﬃﬃﬃﬃﬃ m1 q2 ¼ 2 q1 m2

q2 ¼

sﬃﬃﬃﬃﬃﬃ m2 q1 m1

for

l2 ¼ km

for l3 ¼ 0

The relative displacements are depicted in the following figure:

It is easy to see that l3 corresponds to the translational mode (Dx1 ¼ Dx2 ¼ Dx3). The inclusion of l3 could be avoided if we consider the restriction that the center of gravity does not move; m1(Dx1 þ Dx3) þ m2Dx2 ¼ 0. The relationships between the generalized coordinates and the normal coordinates are given by Eq. 1.40. In the present case, we have q1 ¼ B11 Q1 þ B12 Q2 þ B13 Q3 q2 ¼ B21 Q1 þ B22 Q2 þ B23 Q3 q3 ¼ B31 Q1 þ B32 Q2 þ B33 Q3 In the normal vibration whose normal coordinate is Q1, B11: B21: B31 gives the ratio of the displacements. From the previous calculation, it is ﬃ obvious that B11: B21: B31 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1:0:1. Similarly, B12 : B22 : B32 ¼ 1 : 2 m1 =m2 : 1 gives the ratio of the displacements in the normal vibration whose normal coordinate is Q2. Thus the mode

SYMMETRY ELEMENTS AND POINT GROUPS

21

of a normal vibration can be drawn if the normal coordinate is translated into a set of rectangular coordinates, as is shown above. So far, we have discussed only the vibrations whose displacements occur along the molecular axis. There are, however, two other normal vibrations in which the displacements occur in the direction perpendicular to the molecular axis. They are not treated here, since the calculation is not simple. It is clear that the method described above will become more complicated as a molecule becomes larger. In this respect, the GF matrix method described in Sec. 1.12 is important in the vibrational analysis of complex molecules. By using the normal coordinates, the Schr€ odinger wave equation for the system can be written as X q2 c i

n qQ2i

! 8p2 1X þ 2 E li Q2i cn ¼ 0 2 i h

ð1:56Þ

Since the normal coordinates are independent of each other, it is possible to write cn ¼ c1 ðQ1 Þc2 ðQ2 Þ

ð1:57Þ

and solve the simpler one-dimensional problem. If Eq. 1.57 is substituted in Eq. 1.56, there results d2 ci 8p2 þ 2 h dQ2i

! 1 2 Ei li Qi ci ¼ 0 2

ð1:58Þ

where E ¼ E1 þ E 2 þ with 0

1 1 Ei ¼ hni @vi þ A 2 vi ¼

1 pﬃﬃﬃﬃ li 2p

ð1:59Þ

1.5. SYMMETRY ELEMENTS AND POINT GROUPS [9–14] As noted before, polyatomic molecules have 3N – 6 or, if linear, 3N – 5 normal vibrations. For any given molecule, however, only vibrations that are permitted by

22

THEORY OF NORMAL VIBRATIONS

the selection rule for that molecule appear in the infrared and Raman spectra. Since the selection rule is determined by the symmetry of the molecule, this must first be studied. The spatial geometric arrangement of the nuclei constituting the molecule determines its symmetry. If a coordinate transformation (a reflection or a rotation or a combination of both) produces a configuration of the nuclei indistinguishable from the original one, this transformation is called a symmetry operation, and the molecule is said to have a corresponding symmetry element. Molecules may have the following symmetry elements. 1.5.1. Identity I This symmetry element is possessed by every molecule no matter how unsymmetric it is; the corresponding operation is to leave the molecule unchanged. The inclusion of this element is necessitated by mathematical reasons that will be discussed in Sec. 1.7. 1.5.2. A Plane of Symmetry If reflection of a molecule with respect to some plane produces a configuration indistinguishable from the original one, the plane is called a plane of symmetry. 1.5.3. A Center of Symmetry i If reflection at the center, that is, inversion, produces a configuration indistinguishable from the original one, the center is called a center of symmetry. This operation changes the signs of all the coordinates involved: xi ! xi ; yi ! yi ; zi ! zi . 1.5.4. A p-Fold Axis of Symmetry Cp* If rotation through an angle 360 /p about an axis produces a configuration indistinguishable from the original one, the axis is called a p-fold axis of symmetry Cp. For example, a twofold axis C2 implies that a rotation of 180 about the axis reproduces the original configuration. A molecule may have a two-, three-, four-, five-, or sixfold, or higher axis. A linear molecule has an infinite-fold (denoted by ¥-fold) axis of symmetry C¥ since a rotation of 360 /¥, that is, an infinitely small angle, transforms the molecule into one indistinguishable from the original. 1.5.5. A p-Fold Rotation–Reflection Axis Sp* If rotation by 360 /p about the axis, followed by reflection at a plane perpendicular to the axis, produces a configuration indistinguishable from the original one, the axis is called a p-fold rotation–reflection axis. A molecule may have a two-, three-, four-,

* The notation C n (or Sn ) is used to indicate that the Cp (or Sp) operation is carried out successively n times. p

p

SYMMETRY ELEMENTS AND POINT GROUPS

23

five-, or sixfold, or higher, rotation–reflection axis. A symmetrical linear molecule has an S¥ axis. It is easily seen that the presence of Sp always means the presence of Cp as well as s when p is odd. 1.5.6. Point Group A molecule may have more than one of these symmetry elements. Combination of more and more of these elements produces systems of higher and higher symmetry. Not all combinations of symmetry elements, however, are possible. For example, it is highly improbable that a molecule will have a C3 and C4 axis in the same direction because this requires the existence of a 12-fold axis in the molecule. It should also be noted that the presence of some symmetry elements often implies the presence of other elements. For example, if a molecule has two s planes at right angles to each other, the line of intersection of these two planes must be a C2 axis. A possible combination of symmetry operations whose axes intersect at a point is called a point group.* Theoretically, an infinite number of point groups exist, since there is no restriction on the order (p) of rotation axes that may exist in an isolated molecule. Practically, however, there are few molecules and ions that possess a rotation axis higher than C6. Thus most of the compounds discussed in this book belong to the following point groups: (1) Cp. Molecules having only a Cp and no other elements of symmetry: C1, C2, C3, and so on. (2) Cph. Molecules having a Cp and a sh perpendicular to it: C1hCs, C2h, C3h, and so on. (3) Cpv. Molecules having a Cp and psv through it: C1v Cs ; C2v ; C3v ; C4v ; . . . ; C¥v : (4) Dp. Molecules having a Cp and pC2 perpendicular to the Cp and at equal angles to one another: D1C2, D2V, D3, D4, and so on. (5) Dph. Molecules having a Cp, psn through it at angles of 360 /2p to one another, and a sh perpendicular to the Cp: D1h C2v ; D2h Vh ; D3h ; D4h ; D5h ; D6h ; . . . ; D¥h : (6) Dpd. Molecules having a Cp, pC2 perpendicular to it, and psd which go through the Cp and bisect the angles between two successive C2 axes: D2dVd, D3d, D4d, D5d, and so on. (7) Sp. Molecules having only a Sp (p even). For p odd, Sp is equivalent to Cp sh, for which other notations such as C3h are used: S2Ci, S4, S6, and so on. (8) Td. Molecules having three mutually perpendicular C2 axes, four C3 axes, and a sd through each pair of C3 axes: regular tetrahedral molecules. *In this respect, point groups differ from space groups, which involve translations and rotations about nonintersecting axes (see Sec. 1.27).

24

THEORY OF NORMAL VIBRATIONS

(9) Oh. Molecules having three mutually perpendicular C4 axes, four C3 axes, and a center of symmetry, i: regular octahedral and cubic molecules. (10) Ih. Molecules having 6 C5 axes, 10 C3 axes, 15 C2 axes, 15 s planes, and a center of symmetry. In total, such molecules possess 120 symmetry elements. One example is an icosahedron having 20 equilateral triangular faces, which is found in the B12 skeleton of the B12 H221 ion (Sec. 2.13). Another example is a regular dodecahedron having 12 regular pentagonal faces. Buckminsterfullerene, C60 (Sec. 2.14), also belongs to the Ih point group. It is a truncated icosahedron with 20 hexagonal and 12 pentagonal faces. Figure 1.9 illustrates the symmetry elements present in the point group, D4h. Complete listings of the symmetry elements for common point groups are found in the character tables included in Appendix I. Figure 1.10 illustrates examples of molecules belonging to some of these point groups. From the symmetry perspective, molecules

Fig. 1.9. Symmetry elements in D4h point group.

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

25

Fig. 1.10. Examples of molecules belonging to some point groups.

belonging to the C1, C2, C3, D2V, and D3 groups possess only Cp axes, and are thus optically active. 1.6. SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES Figures 1.11 and 1.12 illustrate the normal modes of vibration of CO2 and H2O molecules, respectively. In each normal vibration, the individual nuclei carry out a simple harmonic motion in the direction indicated by the arrow, and all the nuclei have

26

THEORY OF NORMAL VIBRATIONS

Fig. 1.10. (Continued)

the same frequency of oscillation (i.e., the frequency of the normal vibration) and are moving in the same phase. Furthermore, the relative lengths of the arrows indicate the relative velocities and the amplitudes for each nucleus.* The n2 vibrations in CO2 are worth comment, since they differ from the others in that two vibrations (n2a and n2b) have exactly the same frequency. Apparently, there are an infinite number of normal vibrations of this type, which differ only in their directions perpendicular to the molecular axis. Any of them, however, can be resolved into two vibrations such as n2a and n2b, which are perpendicular to each other. In this respect, the n2 vibrations in CO2 are called doubly degenerate vibrations. Doubly degenerate vibrations occur only when a molecule has an axis higher than twofold. Triply degenerate vibrations also occur in molecules having more than one C3 axis. To determine the symmetry of a normal vibration, it is necessary to begin by considering the kinetic and potential energies of the system. These were discussed in Sec. 1.4: *In this respect, all the normal modes of vibration shown in this book are only approximate.

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

27

Fig. 1.10. (Continued)

1X 2 Q_ i 2 i

ð1:60Þ

1X li Q2i 2 i

ð1:61Þ

T¼ V¼

Consider a case in which a molecule performs only one normal vibration, Qi. Then, 2 T ¼ 12 Q_ i and V ¼ 12 li Q2i . These energies must be invariant when a symmetry operation, R, changes Qi to RQi. Thus, we obtain 1 2 1 T ¼ Q_ i ¼ ðRQ_ i Þ2 2 2 1 1 V ¼ li Q2i ¼ li ðRQi Þ2 2 2

28

THEORY OF NORMAL VIBRATIONS

Fig. 1.10. (Continued)

Fig. 1.11. Normal modes of vibration in CO2 (þ and – denote vibrations going upward and downward, respectively, in the direction perpendicular to the paper plane).

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

29

For these relations to hold, it is necessary that ðRQi Þ2 ¼ Q2i

or

RQi ¼ Qi

ð1:62Þ

Thus, the normal coordinate must change either into itself or into its negative. If Qi ¼ RQi, the vibration is said to be symmetric. If Qi ¼ RQi, it is said to be antisymmetric. If the vibration is doubly degenerate, we have T¼

1 _2 1 2 Qia þ Q_ ib 2 2

1 1 V ¼ li ðQia Þ2 þ li ðQib Þ2 2 2 In this case, a relation such as ðRQia Þ2 þ ðRQib Þ2 ¼ Q2ia þ Q2ib

ð1:63Þ

must hold. As will be shown later, such a relationship is expressed more conveniently by using a matrix form:* #" " # " # Qia Qia A B R ¼ C D Qib Qib where the values of A, B, C, and D depend on the symmetry operation, R. In any case, the normal vibration must be either symmetric or antisymmetric or degenerate for each symmetry operation. The symmetry properties of the normal vibrations of the H2O molecule shown in Fig. 1.12 are classified as indicated in Table 1.3. Here, þ1 and 1 denote symmetric and antisymmetric, respectively. In the n1 and n2 vibrations, all the symmetry properties are preserved during the vibration. Therefore they are symmetric vibrations and are called, in particular, totally symmetric vibrations. In the n3 vibration, however, symmetry elements such as C2 and sv (xz) are lost. Thus, it is called a nonsymmetric vibration. If a molecule has a number of symmetry elements, the normal vibrations are classified according to the number and the kind of symmetry elements preserved during the vibration. To determine the activity of the vibrations in the infrared and Raman spectra, the selection rule must be applied to each normal vibration. As will be shown in Sec. 1.10, rigorous selection rules can be derived quantum mechanically, and applied to individual molecules using group theory. For small molecules, however, the IR and Raman activities may be determined by simple inspection of their normal modes. First, we consider the general rule which states that the vibration is IR-active if the dipole moment is changed during the vibration. It is obvious that the vibration of a homopolar diatomic molecule is *For matrix algebra, see Appendix II.

30

THEORY OF NORMAL VIBRATIONS

Fig. 1.12. Normal modes of vibration in H2O.

TABLE 1.3. C2v Q1, Q2 Q3 a

I þ1 þ1

C2(z)

sv(xz)a

sv(xz)a

þ1 1

þ1 1

þ1 þ1

sv ¼ vertical plane of symmetry.

not IR-active, whereas that of a heteropolar diatomic molecule is IR-active. In the case of CO2, shown in Fig. 1.11, it is readily seen that the n1 is not IR-active, whereas the n2 and n3 are IR-active. Figure 1.13 illustrates the changes in dipole moment during the three normal vibrations of H2O. It is readily seen that all the vibrations are IR-active because the magnitude or direction of the dipole moment is changed as indicated. To discuss Raman activity, we must consider the nature of polarizability introduced in Sec. 1.2. When a molecule interacts with the electric field of a laser beam, its electron cloud is distorted because the positively charged nuclei are attracted toward the negative pole, and the electrons toward the positive pole, as shown in Fig. 1.14. The charge separation produces an induced dipole moment (P) given by* P ¼ aE*

ð1:64Þ

*A more complete form of this equation is P ¼ aE þ 1 bE2 . . .. Here, b a and b is called the 2 hyperpolarizability. The second term becomes significant only when E is large (109 V cm1). In this case, we observe novel spectroscopic phenomena such as the hyper Raman effect, stimulated Raman effect, inverse Raman effect, and coherent anti-Stokes Raman scattering (CARS). For a discussion of nonlinear Raman spectroscopy, see Refs. 25 and 26.

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

31

Fig. 1.13. Changes in dipole moment for H2O during each normal vibration.

where E is the strength of the electric field and a is the polarizability. By resolving P, a, and E in the x, y, and z directions, we can write. Px ¼ ax Ex ;

Py ¼ a y Ey ;

P z ¼ a z Ez

ð1:65Þ

However, the actual relationship is more complicated since the direction of polarization may not coincide with the direction of the applied field. This is so because the direction of chemical bonds in the molecule also affects the direction of polarization. Thus, instead of Eq. 1.65, we have the relationship Px ¼ axx Ex þ axy Ey þ axz Ez Py ¼ ayx Ex þ ayy Ey þ ayz Ez

ð1:66Þ

Pz ¼ azx Ex þ azy Ey þ azz Ez In matrix form, Eq. 1.66 is written as "

Px Py Pz

# ¼

"a

xx

ayx azx

axy ayy azy

axz #" Ex # ayz Ey azz Ez

Fig. 1.14. Polarization of a diatomic molecule in an electric field.

ð1:67Þ

32

THEORY OF NORMAL VIBRATIONS

and the first matrix on the right-hand side is called the polarizability tensor. In normal Raman scattering, the tensor is symmetric; axy ¼ ayx, ayz ¼ azy, and axz ¼ azx. This is not so, however, in the case of resonance Raman scattering (Sec. 1.22). According to quantum mechanics, the vibration is Raman-active if one of these six components of the polarizability changes during the vibration. Thus, it is obvious that the vibration of a homopolar diatomic molecule is Raman-active but not IR-active, whereas the vibration of a heteropolar diatomic molecule is both IR- and Ramanactive. Changes in the polarizability tensor can be visualized if we draw a polarizability pﬃﬃﬃ ellipsoid by plotting 1= a in every direction from the origin. This gives a threedimensional surface such as shown below:

If we rotate such an ellipsoid so that its principal axes coincide with the molecular axes (X, Y, and Z), Eq. 1.67 is simplified to "

PX PY PZ

#

" ¼

aXX 0 0

0 aYY 0

0 0 aZZ

#"

EX EY EZ

# ð1:68Þ

Such axes are called principal axes of polarizability. In terms of the polarizability ellipsoid, the vibration is Raman-active if the polarizability ellipsoid changes in size, shape, or orientation during the vibration. As an example, Fig. 1.15 illustrates the polarizability ellipsoids for the three normal vibrations of H2O at the equilibrium (q ¼ 0) and two extreme configurations (q ¼ q). It is readily seen that both n1 and n2 are Raman-active because the size and the shape of the ellipsoid (axx, ayy, and azz) change during these vibrations. The n3 is also Raman-active because the orientation of the ellipsoid (ayz) changes during the vibration. Thus, all three normal vibrations of H2O are Raman-active. Figure 1.16 illustrates the changes in polarizability ellipsoids during the normal vibrations of CO2. It is readily seen that n1 is Raman-active because the size of the ellipsoid

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

33

Fig. 1.15. Changes in polarizability ellipsoid during normal vibrations of H2O.

changes during the vibration (axx, ayy, and azz change proportionately). Although the size and/or shape of the ellipsoid change during the n2 and n3 vibrations, they are identical in two extreme positions, as seen in Fig. 1.16. If we consider a limiting case where the nuclei undergo very small displacements, there is effectively no change in the polarizability. This is illustrated in Fig. 1.17. Thus, these two vibrations are not Raman-active. It should be noted that in CO2 the vibration symmetric with respect to the center of symmetry (n1) is Raman-active and not IR-active, whereas the vibrations antisymmetric with respect to the center of symmetry (n2 and n3) are IR-active but not Ramanactive. In a polyatomic molecule having a center of symmetry, the vibrations symmetric with respect to the center of symmetry (g vibrations*) are Raman-active

Fig. 1.16. Changes in polarizability ellipsoid during normal vibrations of CO2.

*The symbols g and u stand for gerade and ungerade (German), respectively.

34

THEORY OF NORMAL VIBRATIONS

Fig. 1.17. Changes in polarizability with respect to displacement coordinate during the n1 and n2,3 vibrations in CO2.

and not IR-active, but the vibrations antisymmetric with respect to the center of symmetry (u vibrations*) are IR-active and not Raman-active. This rule is called the “mutual exclusion rule.” It should be noted, however, that in polyatomic molecules having several symmetry elements in addition to the center of symmetry, the vibrations that should be active according to this rule may not necessarily be active, because of the presence of other symmetry elements. An example is seen in a square–planar XY4-type molecule of D4h symmetry, where the A2g vibrations are not Raman-active and the A1u, B1u, and B2u vibrations are not IR-active (see Sec. 2.6). 1.7. INTRODUCTION TO GROUP THEORY [9–14] In Sec. 1.5, the symmetry and the point group allocation of a given molecule were discussed. To understand the symmetry and selection rules of normal vibrations in polyatomic molecules, however, a knowledge of group theory is required. The minimum amount of group theory needed for this purpose is given here. Consider a pyramidal XY3 molecule (Fig. 1.18) for which the symmetry operations I, C3þ , C3 , s1, s2, and s3 are applicable. Here, C3þ and C3 denote rotation through 120 in the clockwise and counterclockwise directions, respectively; and s1, s2, and s3 indicate the symmetry planes that pass through X and Y1, X and Y2, and X and Y3, respectively. For simplicity, let these symmetry operations be denoted by I, A, B, C, D, and E, respectively. Other symmetry operations are possible, but each is equivalent to some one of the operations mentioned. For instance, a clockwise rotation through 240 is identical with operation B. It may also be shown that two successive applications of any one of these operations is equivalent to some single operation of the group mentioned. Let operation C be applied to the original figure. This interchanges Y2 and Y3. If operation A is applied to the resulting figure, the net result is the same as

*The symbols g and u stand for gerade and ungerade (German), respectively.

INTRODUCTION TO GROUP THEORY

35

Fig. 1.18. Pyramidal XY3 molecule.

application of the single operation D to the original figure. This is written as CA ¼ D. If all the possible multiplicative combinations are made, then a tabular display such as Table 1.4, in which the operation applied first is written across the top, is obtained. This is called the multiplication table of the group. It is seen that a group consisting of the mathematical elements (symmetry operations) I, A, B, C, D, and E satisfies the following conditions: (1) The product of any two elements in the set is another element in the set. (2) The set contains the identity operation that satisfies the relation IP ¼ PI ¼ P, where P is any element in the set. (3) The associative law holds for all the elements in the set, that is, (CB)A ¼ C (BA), for example. (4) Every element in the set has its reciprocal, X, which satisfies the relation XP ¼ PX ¼ I, where P is any element in the set. This reciprocal is usually denoted by P1. These are necessary and sufficient conditions for a set of elements to form a group. It is evident that operations I, A, B, C, D, and E form a group in this sense. It should be noted

TABLE 1.4.

I A B C D E

I

A

B

C

D

E

I A B C D E

A B I E C D

B I A D E C

C D E I A B

D E C B I A

E C D A B I

36

THEORY OF NORMAL VIBRATIONS

that the commutative law of multiplication does not necessarily hold. For example, Table 1.4 shows that CD „ DC. The six elements can be classified into three types of operation: the identity operation I, the rotations C3þ and C3 , and the reflections s1, s2, and s3. Each of these sets of operations is said to form a class. More precisely, two operations, P and Q, which satisfy the relation X1PX ¼ P or Q, where X is any operation of the group and X1 is its reciprocal, are said to belong to the same class. It can easily be shown that C3þ and C3 , for example, satisfy the relation. Thus the six elements of the point group C3v are usually abbreviated as I, 2C3, and 3sv. The relations between the elements of the group are shown in the multiplication table (Table 1.4). Such a tabulation of a group is, however, awkward to handle. The essential features of the table may be abstracted by replacing the elements by some analytical function that reproduces the multiplication table. Such an analytical expression may be composed of a simple integer, an exponential function, or a matrix. Any set of such expressions that satisfies the relations given by the multiplication table is called a representation of the group and is designated by G. The representations of the point group C3v discussed above are indicated in Table 1.5. It can easily be proved that each representation in the table satisfies the multiplication table. In addition to the three representations in Table 1.5, it is possible to write an infinite number of other representations of the group. If a set of six matrices of the type S1R(K)S is chosen, where R(K) is a representation of the element K given in Table 1.5, SðjSj „ 0Þ is any matrix of the same order as R, and S1 is the reciprocal of S, this set also satisfies the relations given by the multiplication table. The reason is obvious from the relation S 1 RðKÞSS 1 RðLÞS ¼ S 1 RðKÞRðLÞS ¼ S 1 RðKLÞS Such a transformation is called a similarity transformation. Thus, it is possible to make an infinite number of representations by means of similarity transformations. On the other hand, this statement suggests that a given representation may possibly be broken into simpler ones. If each representation of the symmetry element K is

TABLE 1.5. C3v

I

A

B

C

D

A1(G1) A2(G2)

1 1

1 1

1 1

1 1

1 1

E(G3)

pﬃﬃﬃ 1 0 1 3 ! B 2 2 C B C 1 0 B C B pﬃﬃﬃ C B C 0 1 1 3 @ A 2 2

pﬃﬃﬃ 1 0 1 3 B 2 2 C B C B C Bpﬃﬃﬃ C B 3 C 1 @ A 2 2

pﬃﬃﬃ 1 3 ! B2 2 B 1 0 B B pﬃﬃﬃ 0 1 B @ 3 1 2 2 0

E 1 1 pﬃﬃﬃ 1 1 0 1 3 C B2 2 C C B C C B C C B pﬃﬃﬃ C C B 3 C 1 A @ A 2 2

INTRODUCTION TO GROUP THEORY

37

transformed into the form

ð1:69Þ

by a similarity transformation, Q1(K), Q2(K), . . . are simpler representations. In such a case, R(K) is called reducible. If a representation cannot be simplified any further, it is said to be irreducible. The representations G1, G2, and G3 in Table 1.5 are all irreducible representations. It can be shown generally that the number of irreducible representations is equal to the number of classes. Thus, only three irreducible representations exist for the point group C3v. These representations are entirely independent of each other. Furthermore, the sum of the squares of the dimensions (l) of the irreducible representations of a group is always equal to the total number of the symmetry elements, namely, the order of the group (h). Thus X

l2i ¼ l21 þ l22 þ ¼ h

ð1:70Þ

In the point group C3v, it is seen that 12 þ 12 þ 22 ¼ 6 A point group is classified into species according to its irreducible representations. In the point group C3v, the species having the irreducible representations G1, G2, and G3 are called the A1, A2, and E species, respectively.* The sum of the diagonal elements of a matrix is called the character of the matrix and is denoted by c. It is to be noted in Table 1.5 that the character of each of the elements belonging to the same class is the same. Thus, using the character, Table 1.5 can be simplified to Table 1.6. Such a table is called the character table of the point group C3v. That the character of a matrix is not changed by a similarity transformation can be proved as follows. If a similarity transformation is expressed by T ¼ S1RS, then cT ¼

X

ðS 1 RSÞii ¼

i

¼

X j;k

dkj Rjk ¼

X

X

ðS 1 Þij Rjk Ski ¼

i;j;k

X

Ski ðS 1 Þij Rjk

j;k;i

Rkk ¼ cR

k

*For labeling of the irreducible representations (species), see Appendix I.

38

THEORY OF NORMAL VIBRATIONS

TABLE 1.6. Character Table of the Point Group C3v C3v

I

A1 (c1) A2 (c2) E (c3)

1 1 2

2C3(z) 1 1 1

3sv 1 1 0

where dkj is Kr€ onecker’s delta (0 for k „ j and 1 for k ¼ j). Thus, any reducible representation can be reduced to its irreducible representations by a similarity transformation that leaves the character unchanged. Therefore, the character of the reducible representation, c(K), is written as cðKÞ ¼

X

am cm ðKÞ

ð1:71Þ

m

where cm(K) is the character of Qm(K), and am is a positive integer that indicates the number of times that Qm(K) appears in the matrix of Eq. 1.69. Hereafter the character will be used rather than the corresponding representation because a 1 : 1 correspondence exists between these two, and the former is sufficient for vibrational problems. It is important to note that the following relation holds in Table 1.6: X

ci ðKÞcj ðKÞ ¼ hdij

ð1:72Þ

K

If Eq. 1.71 is multiplied by ci(K) on both sides, and the summation is taken over all the symmetry operations, then X

cðKÞci ðKÞ ¼

K

¼

XX K

m

m

K

XX

am cm ðKÞci ðKÞ am cm ðKÞci ðKÞ

For a fixed m, we have X

am cm ðKÞci ðKÞ ¼ am

K

X

cm ðKÞci ðKÞ ¼ am hdim

K

If we consider the sum of such a term over m, only the sum in which m ¼ i remains. Thus, we obtain X K

cðKÞcm ðKÞ ¼ ham

39

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

or am ¼

1X cðKÞcm ðKÞ h K

ð1:73Þ

This formula is written more conveniently as am ¼

1X ncðKÞcm ðKÞ h

ð1:74Þ

where n is the number of symmetry elements in any one class, and the summation is made over the different classes. As Sec. 1.8 will show, this formula is very useful in determining the number of normal vibrations belonging to each species.*

1.8. THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES As shown in Sec. 1.6, the 3N 6 (or 3N 5) normal vibrations of an N-atom molecule can be classified into various species according to their symmetry properties. The number of normal vibrations in each species can be calculated by using the general equations given in Appendix III. These equations were derived from consideration of the vibrational degrees of freedom contributed by each set of identical nuclei for each symmetry species [1]. As an example, let us consider the NH3 molecule belonging to the C3v point group. The general equations are as follows: A1 species: A2 species:

3m þ 2mv þ m0 1 3m þ mv 1

E species: 6m þ 3mv þ m0 2 Nðtotal number of atomsÞ ¼ 6m þ 3mv þ m0 From the definitions given in the footnotes of Appendix III, it is obvious that m ¼ 0, m0 ¼ 1, and mv ¼ 1 in this case. To check these numbers, we calculate the total number of atoms from the equation for N given above. Since the result is 4, these assigned numbers are correct. Then, the number of normal vibrations in each species can be calculated by inserting these numbers in the general equations given above: 2, 0, and 2, respectively, for the A1, A2, and E species. Since the E species is doubly degenerate, the total number of vibrations is counted as 6, which is expected from the 3N 6 rule. A more general method of finding the number of normal vibrations in each species can be developed by using group theory. The principle of the method is that all the *Since this equation is not applicable to the infinite point groups (C¥v and D¥h), several alternative approaches have been proposed (see Refs. 75 and 76).

40

THEORY OF NORMAL VIBRATIONS

representations are irreducible if normal coordinates are used as the basis for the representations. For example, the representations for the symmetry operations based on three normal coordinates, Q1, Q2, and Q3, which correspond to the n1, n2, and n3 vibrations in the H2O molecule of Fig. 1.12, are as follows: " Q1 # I Q2 Q3 "

Q1 sv ðxzÞ Q2 Q3

"1 ¼

0 0

#

" ¼

0 0 #" Q1 # 1 0 0 1 1 0 0 1 0 0

" Q1 #

Q2 ; Q3 0 0 1

#"

C2 ðzÞ Q2 Q3

# Q1 Q2 ; Q3

0 #" Q1 #

"1 0 ¼

"

Q1 sv ðyzÞ Q2 Q3

0 1 0 0 #

" ¼

0 1 1 0 0

0 0 1 0 0 1

Q2 Q3 #"

Q1 Q2 Q3

#

Let a representation be written with the 3N rectangular coordinates of an N-atom molecule as its basis. If it is decomposed into its irreducible components, the basis for these irreducible representations must be the normal coordinates, and the number of appearances of the same irreducible representation must be equal to the number of normal vibrations belonging to the species represented by this irreducible representation. As stated previously, however, the 3N rectangular coordinates involve six (or five) coordinates, which correspond to the translational and rotational motions of the molecule as a whole. Therefore, the representations that have such coordinates as their basis must be subtracted from the result obtained above. Use of the character of the representation, rather than the representation itself, yields the same result. For example, consider a pyramidal XY3 molecule that has six normal vibrations. At first, the representations for the various symmetry operations must be written with the 12 rectangular coordinates in Fig. 1.19 as their basis. Consider pure rotation Cpþ . If the

Fig. 1.19. Rectangular coordinates in a pyramidal XY3 molecule (Z axis is perpendicular to the paper plane).

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

41

clockwise rotation of the point (x,y,z) around the z axis by the angle y brings it to the point denoted by the coordinates (x0 ,y0,z0 ), the relations between these two sets of coordinates are given by. x0 ¼ x cosy þ y siny y0 ¼ x siny þ y cosy z0 ¼ z

ð1:75Þ

By using matrix notation,* this can be written as "

x0 y0 z0

# ¼

C0þ

" # " cosy siny x y ¼ siny cosy 0 0 z

0 0 1

#" # x y z

ð1:76Þ

Then the character of the matrix is given by cðCyþ Þ ¼ 1 þ 2cosy

ð1:77Þ

The same result is obtained for cðCy Þ. If this symmetry operation is applied to all the coordinates of the XY3 molecule, the result is

ð1:78Þ

where A denotes the small square matrix given by Eq. 1.76. Thus, the character of this representation is simply given by Eq. 1.77. It should be noted in Eq. 1.78 that only the small matrix A, related to the nuclei unchanged by the symmetry operation, appears as a diagonal element. Thus, a more general form of the character of the representation for rotation around the axis by y is

*For matrix algebra, see Appendix II.

42

THEORY OF NORMAL VIBRATIONS

cðRÞ ¼ NR ð1 þ 2 cosyÞ

ð1:79Þ

where NR is the number of nuclei unchanged by the rotation. In the present case, NR ¼ 1 and y ¼ 120 . Therefore, we obtain cðC3 Þ ¼ 0

ð1:80Þ

Identity (I) can be regarded as a special case of Eq. 1.79 in which NR ¼ 4 and y ¼ 0 . The character of the representation is cðIÞ ¼ 12

ð1:81Þ

Pure rotation and identity are called proper rotation. It is evident from Fig. 1.19 that a symmetry plane such as s1 changes the coordinates from (xi, yi, zi) to (xi, yi, zi). The corresponding representation is therefore written as " # " 1 x 0 s1 y ¼ 0 z

0 1 0

0 0 1

#" # x y z

ð1:82Þ

The result of such an operation on all the coordinates is

ð1:83Þ

where B denotes the small square matrix of Eq. 1.82. Thus, the character of this representation is calculated as 2 1 ¼ 2. It is noted again that the matrix on the diagonal is nonzero only for the nuclei unchanged by the operation. More generally, a reflection at a plane (s) is regarded as s ¼ i C2. Thus, the general form of Eq. 1.82 may be written as

43

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

"

1 0 0

0 1 0

0 0 1

#"

cosy siny 0

siny cosy 0

0 0 1

#

" ¼

cosy siny 0

siny cosy 0

0 0 1

#

Then cðsÞ ¼ ð1 þ 2 cosyÞ As a result, the character of the large matrix shown in Eq. 1.83 is given by cðRÞ ¼ NR ð1 þ 2 cosyÞ

ð1:84Þ

In the present case, NR ¼ 2 and y ¼ 180 . This gives cðsv Þ ¼ 2

ð1:85Þ

Symmetry operations such as i and Sp are regarded as i ¼ i I;

y ¼ 0

S3 ¼ i C6 ; S4 ¼ i C4;

y ¼ 60 y ¼ 90

S6 ¼ i C3 ;

y ¼ 120

Therefore, the characters of these symmetry operations can be calculated by Eq. 1.84 with the values of y defined above. Operations such as s, i, and Sp are called improper rotations. Thus, the character of the representation based on 12 rectangular coordinates is as follows: I 2C3 12 0

3sv 2

ð1:86Þ

To determine the number of normal vibrations belonging to each species, the c(R) thus obtained must be resolved into the ci(R) of the irreducible representations of each species in Table 1.6. First, however, the characters corresponding to the translational and rotational motions of the molecule must be subtracted from the result shown in Eq. 1.86. The characters for the translational motion of the molecule in the x, y, and z directions (denoted by Tx, Ty, and Tz) are the same as those obtained in Eqs. 1.79 and 1.84. They are as follows: ct ðRÞ ¼ ð1 þ 2 cosyÞ

ð1:87Þ

where the þ and signs are for proper and improper rotations, respectively. The characters for the rotations around the x, y, and z axes (denoted by Rx, Ry, and Rz) are

44

THEORY OF NORMAL VIBRATIONS

given by cr ðRÞ ¼ þ ð1 þ 2 cosyÞ

ð1:88Þ

for both proper and improper rotations. This is due to the fact that a rotation of the vectors in the plane perpendicular to the x, y, and z axes can be regarded as a rotation of the components of angular momentum, Mx,My, and Mz, about the given axes. If px, py, and pz are the components of linear momentum in the x, y, and z directions, the following relations hold: Mx ¼ ypz zpy My ¼ zpx xpz Mz ¼ xpy ypx Since (x,y,z) and (px,py ,pz) transform as shown in Eq. 1.76, it follows that " Cy

Mx My Mz

#

" ¼

cosy siny 0

siny cosy 0

0 0 1

#"

Mx My Mz

#

Then, a similar relation holds for Rx, Ry , and Rz: " Cy

Rx Ry Rz

#

" ¼

cosy siny 0

siny cosy 0

0 0 1

#"

Rx Ry Rz

#

Thus, the characters for the proper rotations are as given by Eq. 1.88. The same result is obtained for the improper rotation if the latter is regarded as i (proper rotation). Therefore, the character for the vibration is obtained from cv ðRÞ ¼ cðRÞ ct ðRÞ cr ðRÞ

ð1:89Þ

It is convenient to tabulate the foregoing calculations as in Table 1.7. By using the formula in Eq. 1.74 and the character of the irreducible representations in Table 1.6, am can be calculated as follows: am ðA1 Þ ¼ 16½ð1Þð6Þð1Þ þ ð2Þð0Þð1Þ þ ð3Þð2Þð1Þ ¼ 2 am ðA2 Þ ¼ 16½ð1Þð6Þð1Þ þ ð2Þð0Þð1Þ þ ð3Þð2Þð 1Þ ¼ 0 am ðEÞ ¼ 16½ð1Þð6Þð2Þ þ ð2Þð0Þð 1Þ þ ð3Þð2Þð0Þ ¼ 2

45

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

TABLE 1.7. Symmetry Operation:

I

Kind of Rotation: y:

0

120

180

cos y 1 þ 2 cos y NR c, NR(1 þ 2 cos y) ct , (1 þ 2 cos y) cr , þ (1 þ 2 cos y) cn , c ct cr

1 3 4 12 3 3 6

–1 –1 2 2 1 –1 2

3sv Improper

2C3 Proper

1 2

0 1 0 0 0 0

and cv ¼ 2cA1 þ 2cE

ð1:90Þ

In other words, the six normal vibrations of a pyramidal XY3 molecule are classified into two A1 and two E species. This procedure is applicable to any molecule. As another example, a similar calculation is shown in Table 1.8 for an octahedral XY6 molecule. By use of Eq. 1.74 and the character table in Appendix I, the am are obtained as am ðA1g Þ ¼481 ½ð1Þð15Þð1Þ þ ð8Þð0Þð1Þ þ ð6Þð1Þð1Þ þ ð6Þð1Þð1Þ þ ð3Þð 1Þð1Þ þ ð1Þð 3Þð1Þ þ ð6Þð 1Þð1Þ þ ð8Þð0Þð1Þ þ ð3Þð5Þð1Þ þ ð6Þð3Þð1Þ ¼1

TABLE 1.8. 00

Symmetry Operation

I

8C3

6C2 6C4 Proper

3C24 C2

S2i

6S4

Kind of Rotation: y:

0

120

180

90

180

0

90

cos y 1 þ 2 cos y NR c, NR(1 þ 2 cos y) ct, (1 þ 2 cos y) cr, þ (1 þ 2 cos y) cn, c ct cr

1 3 7 21 3 3 15

1 1 1 1 1 1 1

0 1 3 3 1 1 1

1 1 3 3 1 1 1

a

1 2

0 1 0 0 0 0

1 3 1 3 3 3 3

sh ¼ horizontal plane of symmetry; sd ¼ diagonal plane of symmetry.

0 1 1 1 1 1 1

a

8S6 3sh Improper 120

1 2

0 1 0 0 0 0

180 1 1 5 5 1 1 5

a

6sd

180 1 1 3 3 1 1 3

46

THEORY OF NORMAL VIBRATIONS

am ðA1u Þ ¼ 481 ½ð1Þð15Þð1Þ þ ð8Þð0Þð1Þ þ ð6Þð1Þð1Þ þ ð6Þð1Þð1Þ þ ð3Þð 1Þð1Þ þ ð1Þð 3Þð 1Þ þ ð6Þð 1Þð 1Þ þ ð8Þð0Þð 1Þ þ ð3Þð5Þð 1Þ þ ð6Þð3Þð 1Þ ¼0 .. . and therefore cv ¼ cA1g þ cEg þ 2cF1u þ cF2g þ cF2u

1.9. INTERNAL COORDINATES In Sec. 1.4, the potential and the kinetic energies were expressed in terms of rectangular coordinates. If, instead, these energies are expressed in terms of internal coordinates such as increments of the bond length and bond angle, the corresponding force constants have clearer physical meanings than do those expressed in terms of rectangular coordinates, since these force constants are characteristic of the bond stretching and the angle deformation involved. The number of internal coordinates must be equal to, or greater than, 3N 6 (or 3N 5), the degrees of vibrational freedom of an N-atom molecule. If more than 3N 6 (or 3N 5) coordinates are selected as the internal coordinates, this means that these coordinates are not independent of each other. Figure 1.20 illustrates the internal coordinates for various types of molecules. In linear XYZ (a), bent XY2 (b), and pyramidal XY3 (c) molecules, the number of internal coordinates is the same as the number of normal vibrations. In a nonplanar X2Y2 molecule (d) such as H2O2, the number of internal coordinates is the same as the number of vibrations if the twisting angle around the central bond (Dt) is considered. In a tetrahedral XY4 molecule (e), however, the number of internal coordinates exceeds the number of normal vibrations by one. This is due to the fact that the six angle coordinates around the central atom are not independent of each other, that is, they must satisfy the relation Da12 þ Da23 þ Da31 þ Da14 þ Da24 þ Da34 ¼ 0

ð1:91Þ

This is called a redundant condition. In a planar XY3 molecule (f), the number of internal coordinates is seven when the coordinate Dy, which represents the deviation from planarity, is considered. Since the number of vibrations is six, one redundant

INTERNAL COORDINATES

Fig. 1.20. Internal coordinates for various molecules.

47

48

THEORY OF NORMAL VIBRATIONS

condition such as

Da12 þ Da23 þ Da31 ¼ 0

ð1:92Þ

must be involved. Such redundant conditions always exist for the angle coordinates around the central atom. In an octahedral XY6 molecule (g), the number of internal coordinates exceeds the number of normal vibrations by three. This means that, of the 12 angle coordinates around the central atom, three redundant conditions are involved: Da12 þ Da26 þ Da64 þ Da41 ¼ 0 Da15 þ Da56 þ Da63 þ Da31 ¼ 0

ð1:93Þ

Da23 þ Da34 þ Da45 þ Da52 ¼ 0 The redundant conditions are more complex in ring compounds. For example, the number of internal coordinates in a triangular X3 molecule (h) exceeds the number of 0 vibrations by three. One of these redundant conditions (A1 species) is Da1 þ Da2 þ Da3 ¼ 0

ð1:94Þ

The other two redundant conditions (E0 species) involve bond stretching and angle deformation coordinates such as r ð2Dr1 Dr2 Dr3 Þ þ pﬃﬃﬃ ðDa1 þ Da2 2Da3 Þ ¼ 0 3 r ðDr2 Dr3 Þ pﬃﬃﬃ ðDa1 Da2 Þ ¼ 0 3

ð1:95Þ

where r is the equilibrium length of the XX bond. The redundant conditions mentioned above can be derived by using the method described in Sec. 1.12. The procedure for finding the number of normal vibrations in each species was described in Sec. 1.8. This procedure is, however, considerably simplified if internal coordinates are used. Again, consider a pyramidal XY3 molecule. Using the internal coordinates shown in Fig. 1.20c, we can write the representation for the C3þ operation as 2

Dr1 6 Dr2 6 6 Dr C3þ 6 3 6 Da12 4 Da 23 Da31

3

2

0 1 7 6 7 6 0 7 6 7¼6 0 7 6 5 6 40 0

0 0 1 0 0 0

1 0 0 0 0 0

0 0 0 0 1 0

0 0 0 0 0 1

3 0 2 Dr1 07 Dr 76 6 2 07 6 76 Dr3 Da12 17 76 4 0 5 Da23 Da31 0

3 7 7 7 7 7 5

ð1:96Þ

Thus cðC3þ Þ ¼ 0, as does cðC3 Þ. Similarly, c(I) ¼ 6 and c(sv) ¼ 2. This result is exactly the same as that obtained in Table 1.7 using rectangular coordinates. When

49

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

TABLE 1.9.

r

c (R) ca (R)

I

8C3

6C2

6C4

6 12

0 0

0 2

2 0

3C24 C

00

2

2 0

S2 i

6S4

0 0

0 0

8S6

3sh

6sd

0 0

4 4

2 2

using internal coordinates, however, the character of the representation is simply given by the number of internal coordinates unchanged by each symmetry operation. If this procedure is made separately for stretching (Dr) and bending (Da) coordinates, it is readily seen that cr ðRÞ ¼ cA1 þ cE ca ðRÞ ¼ cA1 þ cE

ð1:97Þ

Thus, it is found that both A1 and E species have one stretching and one bending vibration, respectively. No consideration of the translational and rotational motions is necessary if the internal coordinates are taken as the basis for the representation. Another example, for an octahedral XY6 molecule, is given in Table 1.9. Using Eq. 1.74 and the character table in Appendix I, we find that these characters are resolved into. cr ðRÞ ¼ cA1g þ cEg þ cF1u

ð1:98Þ

ca ðRÞ ¼ cA1g þ cEg þ cF1u þ cF2g þ cF2u

ð1:99Þ

Comparison of this result with that obtained in Sec. 1.8 immediately suggests that three redundant conditions are included in these bending vibrations (one in A1g and one in Eg). Therefore, ca(R) for genuine vibrations becomes ca ðRÞ ¼ cF1u þ cF2g þ cF2u

ð1:100Þ

Thus, it is concluded that six stretching and nine bending vibrations are distributed as indicated in Eqs. 1.98 and 1.100, respectively. Although the method given above is simpler than that of Sec. 1.8, caution must be exercised with respect to the bending vibrations whenever redundancy is involved. In such a case, comparison of the results obtained from both methods is useful in finding the species of redundancy. 1.10. SELECTION RULES FOR INFRARED AND RAMAN SPECTRA According to quantum mechanics [3], the selection rule for the infrared spectrum is determined by the following integral:

50

THEORY OF NORMAL VIBRATIONS

ð ½mv0 v00 ¼ c*v0 ðQa Þmcv00 ðQa ÞdQa

ð1:101Þ

Here, m is the dipole moment in the electronic ground state, c is the vibrational eigenfunction given by Eq. 1.21, and v0 and v00 are the vibrational quantum numbers before and after the transition, respectively. The activity of the normal vibration whose normal coordinate is Qa is being determined. By resolving the dipole moment into the three components in the x, y, and z directions, we obtain the result ð ½mx v0 v00 ¼ c*v0 ðQa Þmx cv00 ðQa ÞdQa ð ½my v0 v00 ¼ c*v0 ðQa Þmy cv00 ðQa ÞdQa ð ½mzv0 v00 ¼ c*v0 ðQa Þmz cv00 ðQa ÞdQa

ð1:102Þ

If one of these integrals is not zero, the normal vibration associated with Qa is infraredactive. If all the integrals are zero, the vibration is infrared-inactive. Similar to the case of infrared spectrum, the selection rule for the Raman spectrum is determined by the integral ð 0 00 ½av v ¼ c*v0 ðQa Þacv00 ðQa ÞdQa ð1:103Þ As shown in Sec. 1.6, a consists of six components, axx,ayy, azz, axy, ayz, and axz. Thus Eq. 1.103 may be resolved into six components:

ð1:104Þ

If one of these integrals is not zero, the normal vibration associated with Qa is Ramanactive. If all the integrals are zero, the vibration is Raman-inactive. As shown below, it is possible to determine whether the integrals of Eqs. 1.102 and 1.104 are zero or nonzero from a consideration of symmetry: 1.10.1. Selection Rules for Fundamentals Let us consider the fundamentals in which transitions occur from v0 ¼ 0 to v00 ¼ 1. It is evident from the form of the vibrational eigenfunction (Eq. 1.22) that c0(Qa) is invariant under any symmetry operation, whereas the symmetry of c1(Qa) is the same as that of Qa. Thus, the integral does not vanish when the symmetry of mx, for example, is the same as that of Qa. If the symmetry properties of mx and Qa differ in even one

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

51

symmetry element of the group, the integral becomes zero. In other words, for the integral to be nonzero, Qa must belong to the same species as mx. More generally, the normal vibration associated with Qa becomes infrared-active when at least one of the components of the dipole moment belongs to the same species as Qa. Similar conclusions are obtained for the Raman spectrum. Namely, the normal vibration associated with Qa becomes Raman-active when at least one of the components of the polarizability belongs to the same species as Qa. Since the species of the normal vibration can be determined by the methods described in Sections 1.8 and 1.9, it is necessary only to determine the species of the components of the dipole moment and polarizability of the molecule. This can be done as follows. The components of the dipole moment, mx, my, and mz, transform as do those of translational motion, Tx, Ty, and Tz, respectively. These were discussed in Sec. 1.8. Thus, the character of the dipole moment is given by Eq. 1.87, which is cu ðRÞ ¼ ð1 þ 2 cosyÞ

ð1:105Þ

where þ and have the same meaning as before. In a pyramidal XY3 molecule, Eq. 1.105 gives

cm ðRÞ

I

2C3

3

0

3sv 1

Using Eq. 1.74, we resolve this into A1 þ E. It is obvious that mz belongs to A1. Then, mx and my must belong to E. In fact, the pair, mx and my, transforms as follows: 2 I

mx my

"

¼

1

0

0

1

#

mx my

;

C3þ

mx

my

6 6 6 ¼6 6 4

s1

mx my

pﬃﬃﬃ 3 3 7 2 7 m 7 x 7 my 17 5 2

cðC3þ Þ ¼ 1

cðIÞ ¼ 2; "

1 2 pﬃﬃﬃ 3 2

#

" ¼

1 0 0 1

#"

mx my

#

cðs1 Þ ¼ 0 Thus, it is found that mz belongs to A1 and (mx, my) belongs to E.

52

THEORY OF NORMAL VIBRATIONS

The character of the representation of the polarizability is given by ca ðRÞ ¼ 2 cosyð1 þ 2 cosyÞ

ð1:106Þ

for both proper and improper rotations. This can be derived as follows. The polarizability in the x, y, and z directions is related to that in X, Y, and Z coordinates by " aXX

aXY

aXZ #

aYX aZX

aYY aZY

aYZ aZZ

"C ¼

Xx

CYx CZx

CXy CYy CZy

CXz #" axx CYz ayx CZz azx

axy ayy azy

axz #" CXx ayz CXy azz CXz

CYx CYy CYz

CZx CZy CZz

#

where CXx, CXy, and so forth denote the direction cosines between the two axes subscripted. If a rotation through y around the Z axis superimposes the X, Y, and Z axes on the x, y, and z axes, the preceding relation becomes "a

xx

Cy ayx azx

axy axz # " cosy siny 0 #" axx axy axz #" cosy ayy ayz ¼ siny cosy 0 ayx ayy ayz siny azy azz 0 0 1 azx azy azz 0

siny 0 cosy 0 0 1

#

This can be written as 3 2 axx cos2 y sin2 y 6 ayy 7 6 7 6 sin2 y 6 cos2 y 7 6 6 azz 7 6 0 0 7 6 Cy 6 6a 7¼ 6 6 6 xy 7 6 siny cosy siny cosy 7 6 4 axz 5 6 4 0 0 ayz 0 0 2

0 0

2 siny cosy 2 siny cosy

1 0

0 2 cos y 1

0

0

0

0

2

3 3 2 0 axx 7 0 7 6a 7 7 6 yy 7 7 6 0 0 7 7 6 azz 7 7 6 0 0 7 6 axy 7 7 7 4 axz 5 5 cosy siny ayz siny cosy 0 0

Thus, the character of this representation is given by Eq. 1.106. The same results are obtained for improper rotations if they are regarded as the product i (proper rotation). For a pyramidal XY3 molecule, Eq. 1.106 gives I ca ðRÞ 6

2C3 3sv 0

2

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

53

Using Eq. 1.74, this is resolved into 2A1 þ 2E. Again, it is immediately seen that* the component azz belongs to A1, and the pair azx and azy belongs to E since "

# zx zy

" # x ¼z ¼ A1 E ¼ E y

It is more convenient to consider the components axx þ ayy and axx ayy than axx and ayy. If a vector of unit length is considered, the relation x2 þ y2 þ z2 ¼ 1 holds. Since azz belongs to A1, axx þ ayy must belong to A1. Then, the pair axx ayy and axy must belong to E. As a result, the character table of the point group C3v, is completed as in Table 1.10. Thus, it is concluded that, in the point group C3v both the A1 and the E vibrations are infrared- as well as Raman-active, while the A2 vibrations are inactive. Complete character tables like Table 1.10 have already been worked out for all the point groups. Therefore, no elaborate treatment such as that described in this section is necessary in practice. Appendix I gives complete character tables for the point groups that appear frequently in this book. From these tables, the selection rules for the infrared and Raman spectra are obtained immediately: The vibration is infrared- or Raman-active if it belongs to the same species as one of the components of the dipole moment or polarizability, respectively. For example, the character table of the point group Oh signifies immediately that only the F1u vibrations are infrared-active and only the A1g, Eg, and F2g vibrations are Raman-active, for the components of the dipole moment or the polarizability belong to these species in this point group. It is to be noted in these character tables that (1) a totally symmetric vibration is Raman-active in any

*The quantum mechanical expression of the polarizability [77] is

axx ¼

2 X ðmx Þj0 v2j0 3h j v2j0 v2

axy ¼

2 X ðmx Þj0 ðmy Þj0 v2j0 3h j v2j0 v2

2

etc:

Here, (mx)j0, for example, is the induced dipole moment along the x axis caused by the 0 (ground state) ! j (excited state) electronic transition, vj0 is the frequency of the 0 ! j transition, and n is the exciting frequency. Thus, it is readily seen that the character of the polarizability components such as axx and axy is determined by considering the product of the characters of dipole moments such as mx and my.

54

THEORY OF NORMAL VIBRATIONS

TABLE 1.10. Character Table of the Point Group C3v C3v

I

2C3

3sv

A1 A2 E

þ1 þ1 þ2

þ1 þ1 1

þ1 1 0

a

mz

axx þ ayy, azz

(mx , my)a

(axz , ayz),a (axx ayy , axy)a

A doubly degenerate pair is represented by two terms in parentheses.

point group, and (2) the infrared- and Raman-active vibrations always belong to u and g types, respectively, in point groups having a center of symmetry. 1.10.2. Selection Rules for Combination and Overtone Bands As stated in Sec. 1.3, some combination and overtone bands appear weakly because actual vibrations are not harmonic and some of these nonfundamentals are allowed by symmetry selection rules. Selection rules for combination bands (ni nj) can be derived from the characters of the direct products of those of individual vibrations. Thus, we obtain cij ðRÞ ¼ ci ðRÞ cj ðRÞ

ð1:107Þ

As an example, consider a molecule of C3v symmetry. It is readily seen that

)

I

2C3

3sv

cA1 ðRÞ cE (R)

1 2

1 1

1 0

cA1 E ðRÞ

2

1

0

cA1 E ðRÞ ¼ cE ðRÞ Thus, a combination band between the A1 and E vibrations is IR- as well as Ramanactive. The activity of a combination band between two E vibrations can be determined by considering the direct product of their characters:

)

I

2C3

3sv

cE (R) cE (R)

2 2

1 1

0 0

cEE (R)

4

1

0

Using Eq. 1.74, this set of the characters can be resolved into cEE ðRÞ ¼ cA1 ðRÞ þ cA2 ðRÞ þ cE ðRÞ Since both A1 and E species are IR- and Raman-active, a combination band between two E vibrations is also IR- and Raman-active. It is convenient to apply the general rules of Appendix IV in determining the symmetry species of direct products.

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

55

Selection rules for overtones of nondegenerate vibrations can be obtained using the following relation: cn ðRÞ ¼ ½cðRÞn ð1:108Þ Here, cn(R) is the character of the (n 1)th overtone (n ¼ 2 for the first overtone). As an example, consider a molecule of C3v symmetry. The character of the first overtone of the A2 fundamental is calculated as follows:

)

I

2C3

3sv

cA2 ðRÞ cA2 ðRÞ

1 1

1 1

1 1

c2A2 ðRÞ

1

1

1

Namely, it is IR- as well as Raman-active because c2A2 ðRÞ ¼ cA1 ðRÞ. However, the second overtone of the A2 fundamental is IR- as well as Raman-inactive because c3A2 ðRÞ ¼ c2A2 ðRÞ cA2 ðRÞ ¼ cA2 ðRÞ (inactive). In general, odd overtones (A1) are IR- and Raman-active, whereas even overtones (A2) are inactive. It is obvious that all the overtones of the A1 fundamental are IR- and Raman-active. Selection rules for overtones of doubly degenerate vibrations (E species) are determined by ð1:109Þ cnE ðRÞ ¼ 12 cnE 1 ðRÞ cE ðRÞ þ cE ðRn Þ For the first overtone, this is written as h i c2E ðRÞ ¼ 12 fcE ðRÞg2 þ cE ðR2 Þ Here, cE(R2) is the character that corresponds to the operation R performed twice successively. Thus, one obtains cE ðI 2 Þ ¼ cE ðIÞ ¼ 2 cE ½ðC3þ Þ2 ¼ cE ðC3 Þ ¼ 1 cE ½ðsv Þ2 ¼ cE ðIÞ ¼ 2 Therefore, c2E ðRÞ can be calculated as follows: I

2C3

3sv

)

cE(R ) cE(R )

2 2

1 1

0 0

þ)

{cE(R )}2 cE(R2 )

4 2

1 1

0 2

{cE(R )}2þcE(R2 )

6

0

2

c2E ðRÞ

3

0

1

2)

c2E ðRÞ ¼ cA1 ðRÞ þ cE ðRÞ

56

THEORY OF NORMAL VIBRATIONS

Thus, the first overtone of the doubly degenerate vibration is IR- and Raman-active. The characters of overtones for triply degenerate vibrations are given by cnF ðRÞ ¼ 13f2cF ðRÞcnF 1 ðRÞ 12 cnF 2 ðRÞ½cF ðRÞ2 þ 12 cF ðR2 ÞcnF 2 ðRÞ þ cF ðRn Þg

ð1:110Þ

For more details, see Refs. [3,5], and [9].

1.11. STRUCTURE DETERMINATION Suppose that a molecule has several probable structures, each of which belongs to a different point group. Then the number of infrared- and Raman-active fundamentals should be different for each structure. Therefore, the most probable model can be selected by comparing the observed number of infrared- and Raman-active fundamentals with that predicted theoretically for each model. Consider the XeF4 molecule as an example. It may be tetrahedral or square-planar. By use of the methods described in the preceding sections, the number of infrared- or Raman-active fundamentals can be found easily for each structure. Tables 1.11a and 1.11b summarize the results. It is seen that the distinction of these two structures can be made by comparing the number of IR- and Raman-active FXeF bending modes; the tetrahedral structure predicts one IR and two Raman bands, whereas the square–planar structure predicts two IR and one Raman bands. In general, the XeF stretching vibrations appear above 500 cm1, whereas the FXeF bending vibrations are observed below 300 cm1. The IR and Raman spectra of XeF4 are shown in Fig. 2.17. The IR spectrum exhibits two bending bands at 291 and 123 cm1, whereas the Raman

TABLE 1.11a. Number of Fundamentals for Tetrahedral XeF4 Number of Fundamentals

XeF Stretching

FXeF Bending

R (p) ia R (dp) ia IR, R (dp)

1 0 1 0 2

1 0 0 0 1

0 0 1 0 1

IR R

2 4

1 2

1 2

Td

Activitya

A1 A2 E F1 F2 Total a

p, polarized; dp, depolarized (see Sec. 1.20).

57

STRUCTURE DETERMINATION

TABLE 1.11b. Number of Fundamentals for Square-Planar XeF4 D4h

Activity

Number of Fundamentals

XeF Stretching

FXeF Bending

A1g A1u A2g A2u B1g B1u B2g B2u Eg Eu

R (p) ia ia IR R (dp) ia R (dp) ia R (dp) IR

1 0 0 1 1 0 1 1 0 2

1 0 0 0 1 0 0 0 0 1

0 0 0 1 0 0 1 1 0 1

IR R

3 3

1 2

2 1

Total

spectrum shows one bending vibration at 218 cm1. Thus, the square–planar structure is preferable to the tetrahedral structure.* Another example is given by the XeF5 ion, which has the three possible structures shown in Fig. 1.21. The results of vibrational analysis for each are summarized in Table 1.12. It is seen that the numbers of IR-active vibrations are 5, 6, and 3 and those of Raman-active vibrations are 6, 9, and 3, respectively, for the D3h, C4v, and D5h structures, As discussed in Sec. 2.7.3, the XeF5 ion exhibits three IR bands (550–400, 290, and 274 cm1) and three Raman bands (502, 423, and 377 cm1). Thus, a pentagonal planar structure is preferable to the other two structures. The somewhat unusual structures thus obtained for XeF4 and XeF5 can be rationalized by the use of the valence shell electron-pair repulsion (VSEPR) theory (Sec. 2.6.3). This rather simple method has been used widely for the elucidation of molecular structure of inorganic and coordination compounds. In Chapter 2, the number of IRand Raman-active vibrations is compared for possible structures of XYn and other

Fig. 1.21. Three possible structures for the XeF5 ion.

*This conclusion may be drawn directly from observation of the mutual exclusion rule, which holds for D4h but not for Td.

58

THEORY OF NORMAL VIBRATIONS

TABLE 1.12. Number of IR- and Raman-Active Fundamentals for Three Possible Structures of the XeF5 Anion Type IR Raman

C4v

D3h 00

0

2A 2 þ 3E 0 0 00 2A 1 þ 3E þ E

3A1 þ 3E 3A1 þ 2B1 þ B2 þ 3E

D5h 00

0

A 2 þ 2E 1 0 0 A 1 þ 2E 2

molecules. Appendix V lists the number of IR- and Raman-active vibrations of MXnYm-type molecules. It should be noted, however, that this method does not give a clear-cut answer if the predicted numbers of infrared- and Raman-active fundamentals are similar for various probable structures. Furthermore, a practical difficulty arises in determining the number of fundamentals from the observed spectrum, since the intensities of overtone and combination bands are sometimes comparable to those of fundamentals when they appear as satellite bands of the fundamental. This is particularly true when overtone and combination bands are enhanced anomalously by Fermi resonance (accidental degeneracy). For example, the frequency of the first overtone of the n2 vibration of CO2 (667 cm1) is very close to that of the n1 vibration P (1337 cm1). Since these two vibrations belong to the same symmetry species ð gþ Þ, they interact with each other and give rise to two strong Raman lines at 1388 and 1286 cm1. Fermi resonances similar to the resonance observed for CO2 may occur for a number of other molecules. It is to be noted also that the number of observed bands depends on the resolving power of the instrument used. Finally, the molecular symmetry in the isolated state is not necessarily the same as that in the crystalline state (Sec. 1.27). Therefore, this method must be applied with caution to spectra obtained for compounds in the crystalline state. 1.12. PRINCIPLE OF THE GF MATRIX METHOD* As described in Sec. 1.4, the frequency of the normal vibration is determined by the kinetic and potential energies of the system. The kinetic energy is determined by the masses of the individual atoms and their geometrical arrangement in the molecule. On the other hand, the potential energy arises from interaction between the individual atoms and is described in terms of the force constants. Since the potential energy provides valuable information about the nature of interatomic forces, it is highly desirable to obtain the force constants from the observed frequencies. This is usually done by calculating the frequencies, assuming a suitable set of force constants. If the agreement between the calculated and observed frequencies is satisfactory, this particular set of the force constants is adopted as a representation of the potential energy of the system.

* For details, see Ref. 3. The term “normal coordinate analysis” is almost synonymous with the GF matrix method, since most of the normal coordinate calculations are carried out by using this method.

PRINCIPLE OF THE GF MATRIX METHOD

59

To calculate the vibrational frequencies, it is necessary first to express both the potential and the kinetic energies in terms of some common coordinates (Sec. 1.4). Internal coordinates (Sec. 1.9) are more suitable for this purpose than rectangular coordinates, since (1) force constants expressed in terms of internal coordinates have clearer physical meanings than those expressed in terms of rectangular coordinates, and (2) a set of internal coordinates does not involve translational and rotational motion of the molecule as a whole. Using the internal coordinates Ri we write the potential energy as 2V ¼ ~ RFR

ð1:111Þ

For a bent Y1XY2 molecule such as that in Fig. 1.20b, R is a column matrix of the form " R¼

Dr1 Dr2 Da

~R ¼ ½Dr1

Dr2

~R is its transpose:

#

Da

and F is a matrix whose components are the force constants " F¼

f11 f21 r1 f31

f12 f22 r2 f32

r1 f13 r2 f23 r1 r2 f33

#

"

F11 F21 F31

F12 F22 F32

F13 F23 F33

# ð1:112Þ*21

Here r1 and r2 are the equilibrium lengths of the XY1 and XY2 bonds, respectively. The kinetic energy is not easily expressed in terms of the same internal coordinates. Wilson [78] has shown, however, that the kinetic energy can be written as

~_ 1 R_ 2T ¼ RG

ð1:113Þ †

where G1 is the reciprocal of the G matrix, which will be defined later. If Eqs. 1.111 and 1.113 are combined with Lagrange’s equation ! d qT qV ¼0 ð1:36Þ þ dt qR_ k qRk

*Here f11 and f22 are the stretching force constants of the XY1 and XY2 bonds, respectively, and f33 is the bending force constant of the Y1XY2 angle. The other symbols represent interaction force constants between stretching and stretching or between stretching and bending vibrations. To make the dimensions of all the force constants the same, f13 (or f31), f23 (or f32), and f33 are multiplied by r1, r2, and r1r2, respectively. †Appendix VI gives the derivation of Eq. 1.113.

60

THEORY OF NORMAL VIBRATIONS

the following secular equation, which is similar to Eq. 1.49, is obtained: F11 ðG1 Þ l 11 F21 ðG1 Þ l 21 . ..

. . . . . . jF G 1 lj ¼ 0

F12 ðG1 Þ12 l F22 ðG1 Þ22 l .. .

ð1:114Þ

By multiplying by the determinant of G G11 G21 .. .

G12 G22 .. .

jGj

ð1:115Þ

from the left-hand side of Eq. 1.114, the following equation is obtained: P G1t Ft1 l P G2t Ft1 .. .

P P G1t Ft2 G2t Ft2 l .. .

jGF Elj ¼ 0

ð1:116Þ

Here, E is the unit matrix, and l is related to the wavenumber v~ by the relation l ¼ 4p2 c2 v~2 .* The order of the equation is equal to the number of internal coordinates used. The F matrix can be written by assuming a suitable set of force constants. If the G matrix is constructed by the following method, the vibrational frequencies are obtained by solving Eq. 1.116. The G matrix is defined as G ¼ BM1~ B

ð1:117Þ

1

..

.

Here M is a diagonal matrix whose components are mi, where mi is the reciprocal of the mass of the ith atom. For a bent XY2 molecule, we obtain 2 3 m1 6 m1 0 7 6 7 1 6 7 m M ¼6 1 7 4 5 0

m3

where m3 and m1 are the reciprocals of the masses of the X and Y atoms, respectively. The B matrix is defined as R ¼ BX ð1:118Þ where R and X are column matrices whose components are the internal and rectangular coordinates, respectively.

*Here l should not be confused with lw (wavelength).

PRINCIPLE OF THE GF MATRIX METHOD

61

Fig. 1.22. Relationship between internal and rectangular coordinates in (a) stretching and (b) bending vibrations of a bent XY2 molecule.

To write Eq. 1.118 for a bent XY2 molecule, we express the bond stretching (Dr1) in terms of the x and y coordinates shown in Fig. 1.22a. It is readily seen that Dr1 ¼ ðDx1 ÞðsÞ ðDy1 ÞðcÞ þ ðDx3 ÞðsÞ þ ðDy3 ÞðcÞ Here, s ¼ sin(a=2), c ¼ cos(a=2), and r is the equilibrium distance between the X and Y atoms. A similar expression is obtained for Dr2. Thus Dr2 ¼ ðDx2 ÞðsÞ ðDy2 ÞðcÞ ðDx3 ÞðsÞ þ ðDy3 ÞðcÞ For the bond bending (Da), consider the relationships illustrated in Fig. 1.22b. It is readily seen that rðDaÞ ¼ ðDx1 ÞðcÞ þ ðDy1 ÞðsÞ þ ðDx2 ÞðcÞ þ ðDy2 ÞðsÞ 2ðDy3 ÞðsÞ

62

THEORY OF NORMAL VIBRATIONS

If these equations are summarized in a matrix form, we obtain

ð1:119Þ

If unit vectors such as those in Fig. 1.23 are considered, Eq. 1.119 can be written in a more compact form using vector notationr: "

Dr1 Dr2 Da

#

" ¼

e31 0 p31 =r

0 e32 p32 =r

e31 e32 ðp31 þ p32 Þ=r

#"

q1 q2 q3

# ð1:120Þ

Here q1, q2, and q3 are the displacement vectors of atoms 1, 2, and 3, respectively. Thus Eq. 1.120 can be written simply as R ¼ Sq

ð1:121Þ

where the dot represents the scalar product of the two vectors. Here S is called the S matrix, and its components (S vector) can be written according to the following formulas: (1) bond stretching Dr1 ¼ D31 ¼ e31 q1 e31 q3

Fig. 1.23. Unit vectors in a bent XY2 molecule.

ð1:122Þ

PRINCIPLE OF THE GF MATRIX METHOD

63

and (2) angle bending: p31 q1 þ p32 q2 ðp31 þ p32 Þ q3 ð1:123Þ r It is seen that the S vector is oriented in the direction in which a given displacement of the ith atom will produce the greatest increase in Dr or Da. Formulas for obtaining the S vectors of other internal coordinates such as those of out-of-plane (Dy) and torsional (Dt) vibrations are also available [3]. By using the S matrix, Eq. 1.117 is written as Da ¼ Da132 ¼

G ¼ Sm 1~ S For a bent XY2 molecule, this becomes 2 0 e31 e31 6 e32 e32 G ¼ 40 2

p31 =r

e31 6 40 e31 2 6 6 6 6 ¼6 6 6 6 4

32

m1 76 54 0

p32 =r

ðp31 þ p32 Þ=r

0 e32

p31 =r p32 =r

e32

ðp31 þ p32 Þ=r

ðm3 þ m1 Þe231

m3 e31 e32 ðm3 m1 Þe232

ð1:124Þ

0 m1

0

0

3

3 0 7 0 5 m3

7 5 3 m1 m e31 p31 þ 3 e31 ðp31 þ p32 Þ 7 r r 7 7 m1 m3 e32 p32 þ e32 ðp31 þ p32 Þ 7 7 r r 7 7 7 m1 2 m1 2 m3 25 p þ p þ ðp þ p Þ 31 32 31 32 r2 r2 r2

Considering e31 e31 ¼ e32 e32 ¼ p31 p31 ¼ p32 p32 ¼ 1; e31 e32 ¼ cos a;

e31 p31 ¼ e32 p32 ¼ 0 e31 p32 ¼ e32 p31 ¼ sin a

ðp31 þ p32 Þ2 ¼ 2ð1 cos aÞ we find that the G matrix is calculated as follows: 3 2 m m3 þ m1 m3 cos a 3 sin a r 7 6 7 6 m3 7 6 7 6 m sin a þ m 3 1 G¼6 7 r 7 6 7 6 2m1 2m3 5 4 þ ð1 cos aÞ r2 r2

ð1:125Þ

64

THEORY OF NORMAL VIBRATIONS

If the G-matrix elements obtained are written for each combination of internal coordinates, there results GðDr1 ; Dr1 Þ ¼ m3 þ m1 GðDr2 ; Dr2 Þ ¼ m3 þ m1 GðDr1 ; Dr2 Þ ¼ m3 cos a 2m 2m GðDa; DaÞ ¼ 21 þ 23 ð1 cos aÞ r r GðDr1 ; DaÞ ¼

ð1:126Þ

m3 sin a r

m3 sin a r If such calculations are made for several types of molecules, it is immediately seen that the G-matrix elements themselves have many regularities. Decius [79] developed general formulas for writing G-matrix elements.* Some of them are as follows: GðDr2 ; DaÞ ¼

G2rr ¼ m1 þ m2

G1rr ¼ m1 cos f

G2rf ¼ r23 m2 sin f

G1rf

1 1

! ¼ ðr13 sin f213 cos c234 þ r14 sin f214 cos c243 Þm1

G3ff ¼ r212 m1 þ r223 m3 þ ðr212 þ r223 2r12 r23 cos fÞm2 G2

1 ¼ 1

ðr212 cos c314 Þm1 þ ½r12 r23 cos 123 r24 cos 124 Þr12 cos c314 þðsin 123 sin 124 sin2 c314 þ cos 324 cos c314 Þr23 r24 m2

*See also Refs. 3 and 80.

UTILIZATION OF SYMMETRY PROPERTIES

65

Fig. 1.24. Spherical angles involving atomic positions, a, b, g, and d.

Here, the atoms surrounded by a bold line circle are those common to both coordinates. The symbols m and r denote the reciprocals of mass and bond distance, respectively. The spherical angle cabg in Fig. 1.24 is defined as coscabg ¼

cos fadg cos fadb cos fbdg sin fadb sin fbdg

ð1:127Þ

The correspondence between the Decius formulas and the results obtained in Eq. 1.126 is evident. With the Decius formulas, the G-matrix elements of a pyramidal XY3 molecule have been calculated and are shown in Table 1.13.

1.13. UTILIZATION OF SYMMETRY PROPERTIES In view of the equivalence of the two XY bonds of a bent XY2 molecule, the F and G matrices obtained in Eqs. 1.112 and 1.125 are written as " F¼ 2

m3 þ m1

6 6 6 6 G ¼ 6 m3 cosa 6 6 m 4 3 sina r

f11 f12 rf13

m3 cosa m3 þ m1

m3 sina r

f12 f11 rf13

rf13 rf13 r 2 f33

#

3 m3 sina r 7 7 m3 7 7 sina 7 r 7 7 2m1 2m3 5 þ ð1 cosaÞ r2 r2

ð1:128Þ

ð1:129Þ

66

THEORY OF NORMAL VIBRATIONS

TABLE 1.13.

Dr1 Dr2 Dr3 Da23 Da31 Da12

Dr1

Dr2

Dr3

Da23

Da31

Da12

A — — — — —

B A — — — —

B B A — — —

C D D E — —

D C D F E —

D D C F F E

A ¼ G2rr ¼ mx þ my B ¼ G1rr ¼ mx cos a ! 1 2 cos að1 cos aÞmx ¼ C ¼ G1rf r sin a 1 mx 2 D ¼ G2f ¼ sin a r 2 E ¼ G3ff ¼ 2 my þ mx ð1 cos aÞ r! my 1 cos a m ð1 þ 3 cos aÞð1 cos aÞ þ 2x ¼ 2 F ¼ G2ff 1 þ cos a r 1 þ cos a r 1

Both of these matrices are of the form "

A C D

C A D

D D B

# ð1:130Þ

The appearance of the same elements is evidently due to the equivalence of the two internal coordinates, Dr1 and Dr2. Such symmetrically equivalent sets of internal coordinates are seen in many other molecules, such as those in Fig. 1.20. In these cases, it is possible to reduce the order of the F and G matrices (and hence the order of the secular equation resulting from them) by a coordinate transformation. Let the internal coordinates R be transformed by Rs ¼ UR Then, we obtain

~_ 1 R_ ¼ R ~_ s ~U1 G1 U1 R_ s 2T ¼ RG ~_ s G1 _s ¼R s R 2V ¼ ~ RFR ¼ ~ Rs ~ U1 FU1 Rs ¼~ Rs Fs Rs

ð1:131Þ

UTILIZATION OF SYMMETRY PROPERTIES

67

where G1 U1 G1 U 1 s ¼~

or

Fs ¼ ~ U1 FU1

Gs ¼ UG~ U

ð1:132Þ

If U is an orthogonal matrix ðU 1 ¼ ~ UÞ, Eq. 1.132 is written as Fs ¼ UF~ U

and

Gs ¼ UG~ U

ð1:133Þ

Both GF and GsFs give the same roots, since jGs Fs Elj ¼ jUG~ U~ U1 FU1 Elj ¼ jUGFU1 Elj

ð1:134Þ

¼ jUkGF ElkU1 j ¼ jGF Elj

If we choose a proper U matrix from symmetry consideration, it is possible to factor the original G and F matrices into smaller ones. This, in turn, reduces the order of the secular equation to be solved, thus facilitating their solution. These new coordinates Rs are called symmetry coordinates. The U matrix is constructed by using the equation Rs ¼ N

X

ci ðKÞKðDr1 Þ

ð1:135Þ

K

Here, K is a symmetry operation, and the summation is made over all symmetry operations. Also, ci(K) is the character of the representation to which Rs belongs. Called a generator, Dr1 is, by symmetry operation K, transformed into K(Dr1), which is another coordinate of the same symmetrically equivalent set. Finally, N is a normalizing factor. As an example, consider a bent XY2 molecule in which Dr1 and Dr2 are equivalent. Using Dr1 as a generator, we obtain I K(Dr1) cA1 ðKÞ cB2 ðKÞ

Dr1 1 1

C2(z)

s(xz)

s(yz)

Dr2 1 –1

Dr2 1 –1

Dr1 1 1

68

THEORY OF NORMAL VIBRATIONS

Thus RsA1 ¼ N

X

1 N ¼ pﬃﬃﬃ 2 2

cA1 ðKÞKðDr1 Þ ¼ 2NðDr1 þ Dr2 Þ sinceð2NÞ2 þ ð2NÞ2 ¼ 1

Then 1 RSA1 ¼ pﬃﬃﬃ ðDr1 þ Dr2 Þ 2

ð1:136Þ

1 RsB2 ¼ pﬃﬃﬃ ðDr1 Dr2 Þ 2

ð1:137Þ

Similarly,

The remaining internal coordinate, Da, belongs to the A1 species. Thus, the complete U matrix is written as 2 1 pﬃﬃﬃ 2 " Rs ðA Þ # 6 6 1 1 6 s R2 ðA1 Þ ¼ 6 6 0 6 1 Rs3 ðB2 Þ 4 pﬃﬃﬃ 2

1 pﬃﬃﬃ 2 0 1 pﬃﬃﬃ 2

3 0

7" 7 Dr1 # 7 17 7 Dr2 7 Da 05

ð1:138Þ

If the G and F matrices of type 1.130 are transformed by relations 1.133, where U is given by the matrix of Eq. 1.138, they become

ð1:139Þ

or, more explicitly

ð1:140Þ

UTILIZATION OF SYMMETRY PROPERTIES

69

ð1:141Þ

In a pyramidal XY3 molecule (Fig. 1.20c), Dr1, Dr2, and Dr3 are the equivalent set; so are Da23, Da31, and Da12. It is already known from Eq. 1.107 that one A1 and one E vibration are involved both in the stretching and in the bending vibrations. Using Dr1 as a generator, we obtain from Eq. 1.135

K(Dr1) cA1 ðKÞ cE(K)

I

C3þ

C3

s1

s2

s3

Dr1 1 2

Dr2 1 1

Dr3 1 1

Dr1 1 0

Dr3 1 0

Dr2 1 0

Then, we obtain 1 RsA1 ¼ pﬃﬃﬃ ðDr1 þ Dr2 þ Dr3 Þ 3

ð1:142Þ

1 RsE1 ¼ pﬃﬃﬃ ð2Dr1 Dr2 Dr3 Þ 6

ð1:143Þ

To find a coordinate that forms a degenerate pair with Eq. 1.143, we repeat the same procedure, using Dr2 and Dr3 as the generators. The results are RsE2 ¼ Nð2Dr2 Dr3 Dr1 Þ RsE2 ¼ Nð2Dr3 Dr1 Dr2 Þ However, these coordinates are not orthogonal to RsE1 (Eq. 1.143). If we take a linear combination, RsE2 þ RsE3 , we obtain Eq. 1.143. If we take s RE2 RsE3 , we obtain 1 RsE4 ¼ pﬃﬃﬃ ðDr2 Dr3 Þ 2

ð1:144Þ

70

THEORY OF NORMAL VIBRATIONS

Since Eqs. 1.143 and 1.144 are mutually orthogonal, these two coordinates are taken as a degenerate pair. Similar results are obtained for three angle-bending coordinates. Thus, the complete U matrix is written as pﬃﬃﬃ pﬃﬃﬃ 3 2 3 3 2 pﬃﬃﬃ Rs1 ðA1 Þ Dr1 1= 3 1= 3 1= 3 0 0 0 p ﬃﬃ ﬃ p ﬃﬃ ﬃ p ﬃﬃ ﬃ 7 6 6 Rs ðA1 Þ 7 6 0 0 0 1= 3 1= 3 1= 3 7 7 6 Dr2 7 7 6 6 2 pﬃﬃﬃ pﬃﬃﬃ 7 6 7 7 6 pﬃﬃﬃ 6 s 6 Dr 7 6 R3a ðEÞ 7 6 2= 6 1= 6 1= 6 0 0 0 7 76 3 7 7¼6 6 p ﬃﬃ ﬃ p ﬃﬃ ﬃ p ﬃﬃ ﬃ 7 6 6 Rs ðEÞ 7 6 0 0 0 2= 6 1= 6 1= 6 7 7 6 Da23 7 7 6 6 4a pﬃﬃﬃ pﬃﬃﬃ 7 6 7 7 6 6 s 4 R3b ðEÞ 5 4 0 1= 2 1= 2 0 0 0 5 4 Da31 5 pﬃﬃﬃ pﬃﬃﬃ Rs4b ðEÞ Da12 0 0 0 0 1= 2 1= 2 ð1:145Þ 2

The G matrix of a pyramidal XY3 molecule has already been calculated (see Table 1.14). By using Eq. 1.133, the new Gs matrix becomes

(1.146) Here A, B, and so forth denote the elements in Table 1.13. The F matrix transforms similarly. Therefore, it is necessary only to solve two quadratic equations for the A1 and E species. For the tetrahedral XY4 molecule shown in Fig. 1.20e, group theory (Secs. 1.8 and 1.9) predicts one A1 and one F2 stretching, and one E and one F2 bending, vibration. The U matrix for the four stretching coordinates is 2

Rs1 ðA1 Þ

3

2

6 Rs ðF Þ 7 6 6 2a 2 7 6 6 s 7¼6 4 R2b ðF2 Þ 5 4 Rs2c ðF2 Þ

1=2 1=2 pﬃﬃﬃ pﬃﬃﬃ 1= 6 1= 6 pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ 1= 12 1= 12 pﬃﬃﬃ pﬃﬃﬃ 1= 2 1= 2

1=2 pﬃﬃﬃ 2= 6 pﬃﬃﬃﬃﬃ 1= 12 0

32 3 Dr1 1=2 76 Dr 7 0 6 27 pﬃﬃﬃﬃﬃ 7 76 7 3= 12 54 Dr3 5 0

Dr4

ð1:147Þ

POTENTIAL FIELDS AND FORCE CONSTANTS

71

whereas the U matrix for the six bending coordinates becomes 2

pﬃﬃﬃ

1= 6 3 Rs1 ðA1 Þ 6 pﬃﬃﬃﬃﬃ 2= 12 6 Rs ðEÞ 7 6 7 6 6 2a 7 6 6 s 6 6 R2b ðEÞ 7 6 0 7 6 pﬃﬃﬃﬃﬃ 6 Rs ðF Þ 7 ¼ 6 6 3a 2 7 6 2= 12 7 6 pﬃﬃﬃ 6 s 4 R3b ðF2 Þ 5 6 6 1= 6 4 Rs3c ðF2 Þ 2

0

pﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃ 1= 6 1= 6 1= 6 1= 6 pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ 1= 12 1= 12 1= 12 1= 12 1=2 1=2 1=2 pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ 1= 12 1= 12 1= 12 pﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃ 1= 6 1= 6 1= 6 1=2

1=2

1=2

1=2 pﬃﬃﬃﬃﬃ 1= 12 pﬃﬃﬃ 1= 6 1=2

3 Da12 6 Da 7 6 23 7 7 6 6 Da31 7 7 6 6 7 6 Da14 7 7 6 4 Da24 5

pﬃﬃﬃ 3 1= 6 pﬃﬃﬃﬃﬃ 7 2= 12 7 7 7 7 0 7 pﬃﬃﬃﬃﬃ 7 2= 12 7 7 pﬃﬃﬃ 7 1= 6 7 5 0

2

ð1:148Þ

Da34 The symmetry coordinate Rs1 ðA1 Þ in Eq. 1.148 represents a redundant coordinate (see Eq. 1.91). In such a case, a coordinate transformation reduces the order of the matrix by one, since all the G-matrix elements related to this coordinate become zero. Conversely, this result provides a general method of finding redundant coordinates. Suppose that the elements of the G matrix are calculated in terms of internal coordinates such as those in Table 1.13. If a suitable combination of internal coordinates is made so that Sj Gij ¼ 0 (where j refers to all the equivalent internal coordinates), such a combination is a redundant coordinate. By using the U matrices in Eqs. 1.147 and 1.148, the problem of solving a tenth-order secular equation for a tetrahedral XY4 molecule is reduced to that of solving two first-order (A1 and E) and one quadratic (F2) equation. The normal modes of vibration of the tetrahedral XY4 molecule are shown in Fig. 2.12 of Chapter 2. It can be seen that the normal modes, n1 and n3, correspond to the symmetry coordinates Rs1 ðA1 Þ and Rs2b ðF2 Þ of Eq. 1.147, respectively, and that the normal modes, n2 and n4, correspond to the symmetry coordinates Rs2a ðEÞ and Rs3b ðF2 Þ of Eq. 1.148, respectively. 1.14. POTENTIAL FIELDS AND FORCE CONSTANTS Using Eqs. 1.111 and 1.128, we write the potential energy of a bent XY2 molecule as 2V ¼ f11 ðDr1 Þ2 þ f11 ðDr2 Þ2 þ f33 r 2 ðDaÞ2 þ 2f12 ðDr1 ÞðDr2 Þ þ 2f13 rðDr1 ÞðDaÞ þ 2f13 rðDr2 ÞðDaÞ

ð1:149Þ

72

THEORY OF NORMAL VIBRATIONS

This type of potential field is called a generalized valence force (GVF) field.* It consists of stretching and bending force constants, as well as the interaction force constants between them. When using such a potential field, four force constants are needed to describe the potential energy of a bent XY2 molecule. Since only three vibrations are observed in practice, it is impossible to determine all four force constants simultaneously. One method used to circumvent this difficulty is to calculate the vibrational frequencies of isotopic molecules (e.g., D2O and HDO for H2O), assuming the same set of force constants.† This method is satisfactory, however, only for simple molecules. As molecules become more complex, the number of interaction force constants in the GVF field becomes too large to allow any reliable evaluation. In another approach, Shimanouchi [82] introduced the Urey–Bradley force (UBF) field, which consists of stretching and bending force constants, as well as repulsive force constants between nonbonded atoms. The general form of the potential field is given by X 1 0 V¼ Ki ðDri Þ2 þ Ki ri ðDri Þ 2 i

!

! X 1 0 2 2 2 Hi ria þ ðDai Þ þ Hi ria ðDai Þ 2 i ! X 1 0 2 Fi ðDqi Þ þ Fi qi ðDqi Þ þ 2 i

ð1:150Þ

Here Dri, Dai, and Dqi, are the changes in the bond lengths, bond angles, and distances 0 0 0 between nonbonded atoms, respectively. The symbols Ki ; Ki ; Hi ; Hi and Fi ; Fi represent the stretching, bending, and repulsive force constants, respectively. Furthermore, ri, ria, and qi are the values of the distances at the equilibrium positions and are inserted 0 0 to make the force constants dimensionally similar. Linear force constants (Ki ; Hi and 0 Fi ) must be included since Dri, Dai, and Dqi are not completely independent of each other. Using the relation q2ij ¼ ri2 þ rj2 2ri rj cosaij

ð1:151Þ

*A potential field consisting of stretching and bending force constants only is called a simple valence force field. †In addition to isotope frequency shifts, mean amplitudes of vibration, Coriolis coupling constants, centrifugal distortion constants, and so forth may be used to refine the force constants of small molecules (see Ref. 81).

POTENTIAL FIELDS AND FORCE CONSTANTS

73

and considering that the first derivatives can be equated to zero in the equilibrium case, we can write the final form of the potential field as # " X 1X 2 0 2 ðtij Fij þ sij Fij Þ ðDri Þ2 V¼ Ki þ 2 i jð„iÞ

1X 0 pﬃﬃﬃﬃﬃﬃﬃ ðHij sij sji Fij þ tij tji Fij Þð ri rj Daij Þ2 2 i<j X 0 þ ð tij tji Fij þ sij sji Fij ÞðDri ÞðDrj Þ þ

i<j

þ

X i„j

rj ðtij sji Fij þ tji sij Fij Þ ri

!1=2

0

ð1:152Þ*

pﬃﬃﬃﬃﬃﬃﬃ ðDri Þð ri rj Daij Þ

Here sij ¼

ri rj cosaij qij

sji ¼

rj ri cosaij qij

tij ¼

rj sinaij qij

tji ¼

ri sinaij qij

ð1:153Þ

In a bent XY2 molecule, Eq. 1.152 becomes h i 0 0 V ¼ 12ðK þ t2 F þ s2 FÞ ðDr1 Þ2 þ ðDr2 Þ2 þ 12ðH s2 F þ t2 FÞðrDaÞ2 0

0

þ ðt2 F þ s2 FÞðDr1 ÞðDr2 Þ þ tsðF þ FÞðDr1 ÞðrDaÞ 0

þ tsðF þ FÞðDr2 ÞðrDaÞ where *In the case of tetrahedral molecules, a term X i„j„k

! k pﬃﬃﬃ rij rik ðrij Daij Þðrik Daik Þ 2

must be added, where k is called the internal tension.

ð1:154Þ

74

THEORY OF NORMAL VIBRATIONS

s¼

rð1 cosaÞ q

t¼

rsina q

Comparing Eqs. 1.154 and 1.149, we obtain the following relations between the force constants of the generalized valence force field and those of the Urey–Bradley force field: 0

f11 ¼ K þ t2 F þ s2 F 0

r 2 f33 ¼ ðH s2 F þ t2 FÞr 2 0

f12 ¼ t2 F þ s2 F

ð1:155Þ

0

rf13 ¼ tsðF þ FÞr Although the Urey–Bradley field has four force constants, F0 is usually taken as 1 10 F, on the assumption that the repulsive energy between nonbonded atoms is proportional to 1/r9.* Thus, only three force constants, K, H, and F, are needed to construct the F matrix. The orbital valence force (OVF) field developed by Heath and Linnett [83] is similar to the UBF field. The OVF field uses the angle (Db), which represents the distortion of the bond from the axis of the bonding orbital instead of the angle between two bonds (Da). The number of force constants in the Urey–Bradley field is, in general, much smaller than that in the generalized valence force field. In addition, the UBF field has the advantage that (1) the force constants have clearer physical meanings than those of the GVF field, and (2) they are often transferable from molecule to molecule. For example, the force constants obtained for SiCl4 and SiBr4 can be used for SiCl3Br, SiCl2Br2, and SiClBr3. Mizushima, Shimanouchi, and their coworkers [81] and Overend and Scherer [84] have given many examples that demonstrate the transferability of the force constants in the UBF field. This property of the Urey–Bradley force constants is highly useful in calculations for complex molecules. It should be mentioned, however, that ignorance of the interactions between nonneighboring stretching vibrations and between bending vibrations in the Urey–Bradley field sometimes causes difficulties in adjusting the force constants to fit the observed frequencies. In such a case, it is possible to improve the results by introducing more force constants [84,85].

*This assumption does not cause serious error in final results, since F0 is small in most cases.

SOLUTION OF THE SECULAR EQUATION

75

Evidently, the values of force constants depend on the force field initially assumed. Thus, a comparison of force constants between molecules should not be made unless they are obtained by using the same force field. The normal coordinate analysis developed in Secs. 1.12–1.14 has already been applied to a number of molecules of various structures. Appendix VII lists the G and F matrix elements for typical molecules. 1.15. SOLUTION OF THE SECULAR EQUATION Once the G and F matrices are obtained, the next step is to solve the matrix secular equation: jGF Elj ¼ 0

ð1:116Þ

In diatomic molecules, pﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ G ¼ G11 ¼ 1/m and F ¼ F11 ¼ K. Then l ¼ G11F11 and ~n ¼ l=2pc ¼ K=m=2pc (Eq. 1.20). If the units of mass and force constant are atomic weight and mdyn/A (or 105 dyn/cm), respectively,* l is related to ~nðcm 1 Þ by pﬃﬃﬃ ~n ¼ 1302:83 l or ~n l ¼ 0:58915 1000

!2 ð1:156Þ

As an example, for the HF molecule m ¼ 0.9573 and K ¼ 9.65 in these units. Then, from Eqs. 1.20 and 1.156, ~n is 4139 cm1. The F and G matrix elements of a bent XY2 molecule are given in Eqs. 1.140 and 1.141, respectively. The secular equation for the A1 species is quadratic: G F þ G12 F21 l jGF Elj ¼ 11 11 G21 F11 þ G22 F21

G11 F12 þ G12 F22 ¼0 G21 F12 þ G22 F22 l

ð1:157Þ

If this is expanded into an algebraic equation, the following result is obtained: 2 l2 ðG11 F11 þ G22 F22 þ 2G12 F12 Þl þ ðG11 G22 G212 ÞðF11 F22 F12 Þ¼0

ð1:158Þ

*Although the bond distance is involved in both the G and F matrices, it is canceled during multiplication of the G and F matrix elements. Therefore, any unit can be used for the bond distance.

76

THEORY OF NORMAL VIBRATIONS

For the H2O molecule, we obtain m1 ¼ mH ¼

1 ¼ 0:99206 1:008

m3 ¼ mO ¼

1 ¼ 0:06252 15:995

a ¼ 105

r ¼ 0:96ðAÞ;

sina ¼ sin 105 ¼ 0:96593 cosa ¼ cos 105 ¼ 0:25882 Then, the G matrix elements of Eq. 1.141 are G11 ¼ m1 þ m3 ð1 þ cosaÞ ¼ 1:03840 pﬃﬃﬃ 2 m sina ¼ 0:08896 G12 ¼ r 3 G22 ¼

1 ½2m1 þ 2m3 ð1 cosaÞ ¼ 2:32370 r2

If the force constants in terms of the generalized valence force field are selected as f11 ¼ 8:4280;

f12 ¼ 0:1050

f13 ¼ 0:2625;

f33 ¼ 0:7680

the F matrix elements of Eq. 1.140 are F11 ¼ f11 þ f12 ¼ 8:32300 pﬃﬃﬃ F12 ¼ 2rf13 ¼ 0:35638 F22 ¼ r 2 f33 ¼ 0:70779 Using these values, we find that Eq. 1.158 becomes l2 10:22389l þ 13:86234 ¼ 0 The solution of this equation gives l1 ¼ 8:61475;

l2 ¼ 1:60914

If these values are converted to ~n through Eq. 1.156, we obtain ~n1 ¼ 3824 cm 1 ;

~n2 ¼ 1653 cm 1

VIBRATIONAL FREQUENCIES OF ISOTOPIC MOLECULES

77

With the same set of force constants, the frequency of the B2 vibration is calculated as l3 ¼ G33 F33 ¼ ½m1 þ m3 ð1 cos aÞ ðf11 f12 Þ ¼ 9:13681 n~3 ¼ 3938 cm1 The observed frequencies corrected for anharmonicity are as follows: w1 ¼ 3825 cm1, w2 ¼ 1654 cm1, and w3 ¼ 3936 cm1. If the secular equation is third order, it gives rise to a cubic equation: l3 ðG11 F11 þ G22 F22 þ G33 F33 þ 2G12 F12 þ 2G13 F13 þ 2G23 F23 Þl2 ( G11 G12 F11 F12 G 12 G13 F12 F13 G11 G13 F11 F13 þ þ þ G21 G22 F21 F22 G22 G23 F22 F23 G21 G23 F21 F23 G11 G12 F11 F12 G12 G13 F12 F13 þ þ G31 G32 F31 F32 G32 G33 F32 F33 G G13 F11 F13 G21 G22 F21 F22 G22 G23 F22 F23 þ 11 þ þ G31 G33 F31 F33 G31 G32 F31 F32 G32 G33 F32 F33 ) G11 G12 G13 F11 F12 F13 G21 G23 F21 F23 þ l G21 G22 G33 F21 F23 F33 ¼ 0 ð1:159Þ G31 G33 F31 F33 G31 G32 G33 F31 F32 F33 Thus, it is possible to solve the secular equation by expanding it into an algebraic equation. If the order of the secular equation is higher than three, however, direct expansion such as that just shown becomes too cumbersome. There are several methods of calculating the coefficients of an algebraic equation using indirect expansion [3]. The use of an electronic computer greatly reduces the burden of calculation. Excellent programs written by Schachtschneider [86] and other workers are available for the vibrational analysis of polyatomic molecules. 1.16. VIBRATIONAL FREQUENCIES OF ISOTOPIC MOLECULES As stated in Sec. 1.14, the vibrational frequencies of isotopic molecules are very useful in refining a set of force constants in vibrational analysis. For large molecules, isotopic substitution is indispensable in making band assignments, since only vibrations involving the motion of the isotopic atom will be shifted by isotopic substitution. Two important rules hold for the vibrational frequencies of isotopic molecules. The first, called the product rule, can be derived as follows. Let l1, l2,. . ., ln be the roots of the secular equation |GF El| ¼ 0. Then l1 l2 ln ¼ jGkFj

ð1:160Þ

78

THEORY OF NORMAL VIBRATIONS

holds for a given molecule. Since the isotopic molecule has exactly the same |F| as that in Eq. 1.160, a similar relation 0

0

0

0

l 1 l 2 l n ¼ jG kFj holds for this molecule. It follows that l1 l2 ln jGj 0 0 0 ¼ 0 l 1 l 2 l n jG j

ð1:161Þ

Since ~n ¼

1 pﬃﬃﬃ l 2pc

Equation 1.161 can be written as ~n1~n2 ~nn 0 0 0 ¼ ~n 1~n 2 ~n n

sﬃﬃﬃﬃﬃﬃﬃﬃ jGj 0 jG j

ð1:162Þ

This rule has been confirmed by using pairs of molecules such as H2O and D2O and CH4 and CD4. The rule is also applicable to the product of vibrational frequencies belonging to a single symmetry species. A more general form of Eq. 1.162 is given by the Redlich-Teller product rule [1]: vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ! ! !t !dx ! dy ! dz u u m0 a m0 b ~n1~n2 ~nn M Ix Iy Iz 1 2 t 0 0 0 ¼ M0 m1 m2 I 0x I 0y I 0z ~n 1~n 2 ~n n

ð1:163Þ

Here m1, m2,. . . are the masses of the representative atoms of the various sets of equivalent nuclei (atoms represented by m,m0,mxy,. . . in the tables given in Appendix III); a, b,. . . are the coefficients of m,m0,mxy,. . .; M is the total mass of the molecule; t is the number of Tx,Ty,Tz in the symmetry type considered; Ix,Iy,Iz are the moments of inertia about the x, y, z axes, respectively, which go through the center of the mass; and dx,dy,dz are 1 to 0, depending on whether Rx,Ry,Rz belong to the symmetry type considered. A degenerate vibration is counted only once on both sides of the equation. Another useful rule in regard to the vibrational frequencies of isotopic molecules, called the sum rule, can be derived as follows. It is obvious from Eqs. 1.158 and 1.159 that l1 þ l2 þ þ l n ¼

X n

l¼

X i; j

Gij Fij

ð1:164Þ

79

METAL–ISOTOPE SPECTROSCOPY

Let sk denote SijGijFij for k different isotopic molecules, all of which have the same F matrix. If a suitable combination of molecules is taken, so that X

s1 þ s2 þ þ sk ¼ "

!

¼

þ

Gij Fij

X

X

!

!

X

þ

Gij

þ þ

Gij Fij !

1

X

2

! Gij Fij

!#

2

þ þ

Gij

1

X

Gij

!

k

X

Fij

k

¼0 then it follows that X

! þ

l 1

X

! þ þ

l

X

! ¼0

l

2

ð1:165Þ

k

This rule has been verified for such combinations as H2O, D2O, and HDO, where 2sðHDOÞ sðH2 OÞ sðD2 OÞ ¼ 0 Such relations between the frequencies of isotopic molecules are highly useful in making band assignments. 1.17. METAL–ISOTOPE SPECTROSCOPY [87,88] As a first approximation, vibrational spectra of coordination compounds (Chapter 3, in Part B) can be classifed into ligand vibrations that occur in the high-frequency region (4000–600 cm1) and metal–ligand vibrations that appear in the lowfrequency region (below 600 cm1). The former provide information about the effect of coordination on the electronic structure of the ligand, while the latter provide direct information about the structure of the coordination sphere and the nature of the metal–ligand bond. Since the main interest of coordination chemistry is the coordinate bond, it is the metal–ligand vibrations that have held the interest of inorganic vibrational spectroscopists. It is difficult, however, to make unequivocal assignments of metal–ligand vibrations since the interpretation of the low-frequency spectrum is complicated by the appearance of ligand vibrations as well as lattice vibrations in the case of solid-state spectra. Conventional methods that have been used to assign metal–ligand vibrations are (1) Comparison of spectra between a free ligand and its metal complex; the metal– ligand vibration should be absent in the spectrum of the free ligand. This method often fails to give a clear-cut assignment because some ligand vibrations activated by complex formation may appear in the same region as the metal–ligand vibrations.

80

THEORY OF NORMAL VIBRATIONS

(2) The metal–ligand vibration should be metal sensitive and be shifted by changing the metal or its oxidation state. This method is applicable only when a series of metal complexes have exactly the same structure, where only the central metal is different. Also, it does not provide definitive assignments since some ligand vibrations (such as chelate ring deformations) are also metal-sensitive. (3) The metal–ligand stretching vibration should appear in the same frequency region if the metal is the same and the ligands are similar. For example, the n(Zn-N) (n: stretching) of Zn(II) pyridine complexes are expected to be similar to those of Zn(II) a-picoline complexes. This method is applicable only when the metal–ligand vibration is known for one parent compound. (4) The metal–ligand vibration exhibits an isotope shift if the ligand is isotopically substituted. For example, the n(Ni-N) of [Ni(NH3)6]Cl2 at 334 cm1 is shifted to 318 cm1 on deuteration of the ammonia ligands. The observed shift (16 cm1) is in good agreement with that predicted theoretically for this mode. This method was used to assign the metal–ligand vibrations of chelate compounds such as oxamido (14N/15N) and acetylacetonato(l6O/l8O) complexes. However, isotopic substitution of the a-atom (atom directly bonded to the metal) causes shifts of not only metal–ligand vibrations but also of ligand vibrations involving the motion of the a-atom. Thus, this method alone cannot provide an unequivocal assignment of the metal–ligand vibration. (5) The frequency of a metal–ligand vibration may be predicted if the metal– ligand stretching and other force constants are known a priori. At present, this method is not practical since only a limited amount of information is available on the force constants of coordination compounds. It is obvious that none of the above methods is perfect in assigning metal–ligand vibrations. Furthermore, these methods encounter more difficulties as the structure of the complex (and hence the spectrum) becomes more complicated. Fortunately, the “metal isotope technique” which was developed in 1969 may be used to obtain reliable metal–ligand assignments [89]. Isotope pairs such as (H/D) and (16O/18O) had been used routinely by many spectroscopists. However, isotopic pairs of heavy metals such as (58Ni/62Ni) and (104Pd/110Pd) were not employed until 1969 when the first report on the assignments of the NiP vibrations of trans-Ni(PEt3)2X2 (X ¼ Cl and Br) was made. The delay in their use was probably due to two reasons: (1) It was thought that the magnitude of isotope shifts arising from metal isotope substitution might be too small to be of practical value. (2) Pure metal isotopes were too expensive to use routinely in the laboratory. Nakamoto and coworkers [87,88] have shown, however, that the magnitudes of metal isotope shifts are generally of the order of 2–10 cm1 for stretching modes and 0–2 cm1 for bending modes, and that the experimental error in measuring the frequency could be as small as 0.2 cm1 if proper precautions

METAL–ISOTOPE SPECTROSCOPY

81

are taken. They have also shown that this technique is financially feasible if the compounds are prepared on a milligram scale. Normally, the vibrational spectrum of a compound can be obtained with a sample less than 10 mg. Table 1.14 lists metal isotopes that are useful for vibrational studies of coordination compounds.

TABLE 1.14. Some Stable Metal Isotopesa Element

Inventory Form

Magnesium

Oxide

Silicon

Oxide

Titanium

Oxide

Chromium

Oxide

Iron

Oxide

Nickel

Copper

Metal Oxide Oxide Oxide

Zinc

Oxide

Germanium

Oxide

Zirconium

Oxide

Molybdenum

Metal/oxide

Ruthenium

Metal

Palladium

Metal

Tin

Oxide

Barium

Carbonate

a

Available at Oak Ridge National Laboratory.

Isotope

Natural Abundance (%)

Mg-24 Mg-25 Mg-26 Si-28 Si-30 Ti-46 Ti-48 Cr-50 Cr-52 Cr-53 Fe-54 Fe-56 Fe-57 Ni-58 Ni-60 Ni-62 Cu-63 Cu-65 Zn-64 Zn-66 Zn-68 Ge-72 Ge-74 Zr-90 Zr-94 Mo-92 Mo-95 Mo-97 Mo-100 Ru-99 Ru-102 Pd-104 Pd-108 Sn-116 Sn-120 Sn-124 Ba-135 Ba-138

78.99 10.00 11.01 92.23 3.10 8.0 73.8 4.345 83.79 9.50 5.9 91.72 2.1 68.27 26.1 3.59 69.17 30.83 48.6 27.9 18.8 27.4 36.5 51.45 17.38 14.84 15.92 9.55 9.63 12.7 31.6 11.4 26.46 14.53 32.59 5.79 6.593 71.70

Purity (%) 99.9 94 97 >99.8 >94 70–96 >99 >95 >99.7 >96 >96 >99.9 86–93 >99.9 >99 >96 >99.8 >99.6 >99.8 >98 97–99 >97 >98 97–99 >98 >97 >96 >92 >97 >98 >99 >95 >98 >95 >98 >94 >93 >99

82

THEORY OF NORMAL VIBRATIONS

It should be noted that the central atom of a highly symmetrical molecule (Td, Oh, etc.) does not move during the totally symmetric vibration. Thus, no metal–isotope shifts are expected in these cases [90]. When the central atom is coordinated by several different donor atoms, multiple isotope labeling is necessary to distinguish different coordinate bond-stretching vibrations. For example, complete assignments of bis (glycino)nickel(II) require 14N=15N and/or I6O=18O isotope shift data as well as 58 Ni=62Ni isotope shift data [91]. This book contains many other applications of metal–isotope techniques. These results show that metal–isotope data are indispensable not only in assigning the metal–ligand vibrations but also in refining metal–ligand stretching force constants in normal coordinate analysis [88]. The presence of vibrational coupling between metal–ligand and other vibrations can also be detected by combining metal–isotope data with normal coordinate calculations (Sec. 1.18) since both experimental and theoretical isotope shift values would be smaller when such couplings occur. The metal–isotope techniques become more important as the molecules become larger and complex. In biological molecules such as heme proteins, structural and bonding information about the active site (iron porphyrin) can be obtained through definitive assignments of coordinate bond-stretching vibrations around the iron atom. Using resonance Raman techniques (Sec. 1.22), it is possible to observe iron porphyrin and iron-axial ligand vibrations without interference from peptide chain vibrations. Thus, these vibrations can be assigned by comparing resonance Raman spectra of a natural heme protein with that of a 54Fe-reconstituted heme protein. Chapter 5 (in Part B) includes several examples of applications of metal-isotope techniques to bioinorganic compounds [hemoglobin (54Fe, 57Fe), oxy-hemocyanin (63Cu, 65Cu), etc.]. An example of a multiple labeling is seen in the case of cytochrome P450cam having the axial Fe–S linkage in addition to four Fe–N (porphyrin) bonds; Champion et al. [92] were able to assign its Fe–S stretching vibration (351 cm1) by combining 32S–34S and 54Fe–56Fe isotope shift data. 1.18. GROUP FREQUENCIES AND BAND ASSIGNMENTS From observation of the infrared spectra of a number of compounds having a common group of atoms, it is found that, regardless of the rest of the molecule, this common group absorbs over a narrow range of frequencies, called the group frequency. For example, the methyl group exhibits the antisymmetric stretching [na(CH3)], symmetric stretching [ns(CH3)], degenerate bending [dd(CH3)], symmetric bending [ds(CH3)], and rocking [rs(CH3)] vibrations in the ranges 3050– 2950, 2970–2860, 1480–1410, 1385–1250, and 1200–800 cm1, respectively. Figure 1.25 illustrates their vibrational modes together with the observed frequencies for CH3C1, The methylene group vibrations are also well known; the antisymmetric stretching [na(CH2)], symmetric stretching [ns(CH2)], scissoring [d(CH2)], wagging [rw(CH2)], twisting [rt,(CH2)], and rocking [rr(CH2)] vibrations appear

GROUP FREQUENCIES AND BAND ASSIGNMENTS

83

Fig. 1.25. Approximate normal modes of a CH3X-type molecule (frequencies are given for X¼Cl). The subscripts s, d, and r denote symmetric, degenerate, and rocking vibrations, respectively. Only the displacements of the H or X atoms are shown.

in the regions 3100–3000, 3060–2980, 1500–1350, 1400–1100, 1260–1030, and 1180–750 cm1, respectively. Figure 1.26 illustrates the normal modes of these vibrations together with the observed frequencies for CH2C12. The normal modes illustrated for CH3C1 and CH2C12 are applicable to vibrational assignments of coordination and organometallic compounds, which will be discussed in Chapters 3 and 4, respectively. Group frequencies have been found for a number of organic and inorganic groups, and they have been summarized as group frequency charts [39,40] which are highly useful in identifying the atomic groups from observed spectra. Group frequency charts for inorganic compounds are given in Appendix VIII as well as in Figs. 2.55 and 2.61. The concept of group frequency rests on the assumption that the vibrations of a particular group are relatively independent of those of the rest of the molecule. As stated in Sec. 1.4, however, all the nuclei of the molecule perform their harmonic oscillations in a normal vibration. Thus, an isolated vibration, which the group frequency would have to be, cannot be expected in polyatomic molecules. If, however, a group includes relatively light atoms such as hydrogen (OH, NH, NH2, CH, CH2, CH3, etc.) or relatively heavy atoms such as the halogens (CC1, CBr, CI, etc.), as compared to other atoms in the molecule, the idea of an isolated vibration may be justified, since the amplitudes (or velocities) of the harmonic oscillation of these atoms are relatively larger or smaller than those of the other atoms in the same molecule. Vibrations of groups having multiple bonds (CC, CN, C¼C, C¼N, C¼O, etc.) may

84

THEORY OF NORMAL VIBRATIONS

Fig. 1.26. Approximate normal modes of a CH2X2-type molecule (frequencies are given for X¼Cl). Only the displacements of the H or X atoms are shown.

also be relatively independent of the rest of the molecule if the groups do not belong to a conjugated system. If atoms of similar mass are connected by bonds of similar strength (force constant), the amplitude of oscillation is similar for each atom of the whole system. Therefore, it is not possible to isolate the group frequencies in a system like the following:

A similar situation may occur in a system in which resonance effects average out the single and multiple bonds by conjugation. Examples of this effect are seen for

GROUP FREQUENCIES AND BAND ASSIGNMENTS

85

the metal chelate compounds discussed in Chapter 3. When the group frequency approximation is permissible, the mode of vibration corresponding to this frequency can be inferred empirically from the band assignments obtained theoretically for simple molecules. If coupling between various group vibrations is serious, it is necessary to make a theoretical analysis for each individual compound, using a method like the following one. As stated in Sec. 1.4, the generalized coordinates are related to the normal coordinates by qk ¼

X

Bki Qi

ð1:40Þ

i

In matrix form, this is written as q ¼ Bq Q

ð1:166Þ

It can be shown [3] that the internal coordinates are also related to the normal coordinates by R ¼ LQ

ð1:167Þ

This is written more explicitly as R1 ¼ l11 Q1 þ l12 Q2 þ þ l1N QN R2 ¼ l21 Q1 þ l22 Q2 þ þ l2N QN .. .

.. .

Ri ¼ li1 Q1 þ li2 Q2 þ þ liN QN

ð1:168Þ

In a normal vibration in which the normal coordinate QN changes with frequency nN, all the internal coordinates, R1, R2, . . ., Ri, change with the same frequency. The amplitude of oscillation is, however, different for each internal coordinate. The relative ratio of the amplitudes of the internal coordinates in a normal vibration associated with QN is given by l1N : I2N : : IiN

ð1:169Þ

If one of these elements is relatively large compared to the others, the normal vibration is said to be predominantly due to the vibration caused by the change of this coordinate.

86

THEORY OF NORMAL VIBRATIONS

The ratio of l’s given by Eq. 1.169 can be obtained as a column matrix (or eigenvector) lN, which satisfies the relation [3] GFlN ¼ lN kN

ð1:170Þ

It consists of i elements, l1N ; l2N ; . . . ; liN , where i is the number of internal coordinates, and can be calculated if the G and F matrices are known. An assembly by columns of the l elements obtained for each l gives the relation GFL ¼ LL

ð1:171Þ

where L is a diagonal matrix whose elements consist of l values. As an example, calculate the L matrix of the H2O molecule, using the results obtained in Sec. 1.15. The G and F matrices for the A1 species are as follows: # # " " 8:32300 0:35638 1:03840 0:08896 ; F¼ G¼ 0:35638 0:70779 0:08896 2:32370 with l1 ¼ 8.61475 and l2 ¼ 1.60914. The GF product becomes #

" GF ¼

8:61090 0:08771

0:30710 1:61299

The L matrix can be calculated from Eq. 1.171: #"

" 8:61090 0:08771

0:30710 1:61299

# l11 l21

l12 l22

" ¼

#

#" l11 l21

l12 l22

8:61475 0

0 1:60914

However, this equation gives only the ratios l11: l21 and l12: l22 To determine their values, it is necessary to use the following normalization condition: L~ L¼G

ð1:172Þ30

Then the final result is "

# l11 l21

l12 l22

" ¼

1:01683 0:01274

0:06686 1:52432

#

*This equation can be derived as follows. According to Eq. 1.113, 2T ¼ ~ _ 1 R. _ On the other hand, RG _ ¼ LQ_ and ~R_ ¼ ~Q_ ~L. Thus 2T ¼ ~Q_ ~LG 1 L Q. _ Comparing this with 2T ¼ ~Q_ EQ_ (matrix Eq. 1.167 gives R ~ G–1 L ¼ E or L~L ¼ G. form of Eq. 1.41), we obtain L

GROUP FREQUENCIES AND BAND ASSIGNMENTS

87

This result indicates that, in the normal vibration Q1, the relative ratio of amplitudes of two internal coordinates, R1 (symmetric OH stretching) and R2 (HOH bending), is 1.0168: 0.0127. Therefore, this vibration (3824 cm1) is assigned to an almost pure OH stretching mode. The relative ratio of amplitudes for the Q2 vibration is 0.0669: 1.5243. Thus, this vibration is assigned to an almost pure HOH bending mode. In other cases, the l values do not provide the band assignments that are expected empirically. This occurs because the dimension of l for a stretching coordinate is different from that for a bending coordinate. The potential energy due to a normal vibration, QN is written as X 1 VðQN Þ ¼ Q2N Fij liN ljN 2 ij

ð1:173Þ

Individual FijliNljN terms in the summation represent the potential energy distribution (PED) in QN, which gives a better measure for making band assignments [93]. In general, the value of FijliNljN is large when i ¼ j. Therefore, the Fii l2iN terms are most important in determining the distribution of the potential energy. Thus, the ratios of the Fii l2iN terms provide a measure of the relative contribution of each internal coordinate Ri to the normal coordinate QN. If any Fii l2iN term is exceedingly large compared with the others, the vibration is assigned to the mode associated with Ri. If Fii l2iN and Fjj l2jN are relatively large compared with the others, the vibration is assigned to a mode associated with both Ri and Rj (coupled vibration). As an example, let us calculate the potential energy distribution for the H2O molecule. Using the F and L matrices obtained previously, we find that the ~ LFL matrix is calculated to be 2 2 3 l11 F11 þ l221 F22 þ 2l21 l11 F12 0 6 8:60551 7 0:00011 0:00923 6 2 7 2 4 l12 F11 þ l22 F22 þ 2l12 l22 F12 5 0 0:03721 1:644459 0:07264 Then, the potential energy distribution in each normal vibration ðFii l2iN Þ is given by l1

"

l2

R1 8:60551 0:03721 R2 0:00011 1:64459

#

*According to Eq. 1.111, the potential energy is written as 2V ¼ RFR. ~ Using Eq. 1.167, we can write this as ~ LFLQ. ~ 2V ¼ Q On the other hand, Eq. 1.42 can be written as 2V ¼ QLQ. A comparison of these two ~ expressions gives L ¼ LFL. If this is written for one normal vibration whose frequency is lN, we have X X ~lNi Fij ljN ¼ Fij liN ljN lN ¼ ij

ij

Then the potential energy due to this vibration is expressed by Eq. 1.173.

88

THEORY OF NORMAL VIBRATIONS

More conveniently, PED is expressed by calculating ðFii l2iN =SFii l2iN Þ 100 for each coordinate: "

l1

R1 99:99 R2

l2 2:21

#

0:01 97:79

In this case, the final results are the same whether the band assignments are based on the L matrix or on the potential energy distribution: Q1 is the symmetric OH stretching and Q2 is the HOH bending. In other cases, different results may be obtained, depending on which criterion is used for band assignments. In the example above, no serious coupling occurs between the OH stretching and HOH bending modes of H2O in the A1 species. This is because the vibrational frequencies of these two modes are far apart. In other cases, however, vibrational frequencies of two or more modes in the same symmetry species are relatively close to each other. Then, the liN values for these modes become comparable. A more rigorous method of determining the vibrational mode involves drawing the displacements of individual atoms in terms of rectangular coordinates. As in Eq. 1.167, the relationship between the rectangular and normal coordinates is given by X ¼ Lx Q

ð1:174Þ

The Lx matrix can be obtained from the relationship [94] ~ 1L Lx ¼ M 1 BG

(1.75) ð1:175Þ

The matrices on the right-hand sides have already been defined. Three-dimensional drawings of normal modes, such as those shown in Chapter 2, can be made from the Cartesian displacement calculations obtained above. However, hand plotting of these data is laborious and complicated. Use of computer plotting programs greatly facilitates this process [95]. 1.19. INTENSITY OF INFRARED ABSORPTION [16] The absorption of strictly monochromatic light (n) is expressed by the Lambert–Beer law: In ¼ I0;n e an pl

ð1:176Þ

*By combining Eqs. 1.118 and 1.174, we have R ¼ BX ¼ BLxQ. Since R ¼ LQ (Eq. 1.167), it follows ~_ X. _ In terms of internal coordinates, that LQ ¼ BLxQ or L ¼ BLx. The kinetic energy is written as 2T ¼ XM ~_ 1 R_ ¼ X ~_ B~G 1 BX. _ By comparing these two expressions, we have M ¼ B~G 1 B. it is written as 2T ¼ RG B G 1 BLx ¼ M 1 ~ Then we can write Lx ¼ M 1 MLx ¼ M 1 ~ B G 1 L.

INTENSITY OF INFRARED ABSORPTION

89

where In is the intensity of the light transmitted by a cell of length l containing a gas at pressure p, I0,n is the intensity of the incident light, and an is the absorption coefficient for unit pressure. The true integrated absorption coefficient A is defined by ! ð ð 1 I0;n A¼ an dn ¼ ln dn ð1:177Þ pl band In band where the integration is carried over the entire frequency region of a band. In practice, In and I0,v cannot be measured accurately, since no spectrophotometers have infinite resolving power. Therefore we measure instead the apparent intensity Tn: ð 0 Tn ¼ IðnÞgðn; n Þdn ð1:178Þ slit 0

where g(n, n ) is a function indicating the amount of light of frequency n when the spectrophotometer reading is set at n0 . Then the apparent integrated absorption coefficient B is defined by 1 B¼ pl

Ð 0 I0 ðnÞgðn; n Þdn 0 dn ln Ðslit 0 band slit IðnÞgðn; n Þdn

ð

ð1:179Þ

It can be shown that limpl ! 0 ðA BÞ ¼ 0

ð1:180Þ

if I0 and an are constant within the slit width used. (This condition is approximated by using a narrow slit.) In practice, we plot B/pl against pl, and extrapolate the curve to pl ! 0. To apply this method to gaseous molecules, it is necessary to broaden the vibrational–rotational bands by adding a high-pressure inert gas (pressure broadening). For liquids and solutions, p and a in the preceding equations are replaced by M (molar concentration) and e (molar absorption coefﬁcient), respectively. However, the extrapolation method just described is not applicable, since experimental errors in determining B values become too large at low concentration or at small cell length. The true integrated absorption coefficient of a liquid can be calculated if we assume that the shape of an absorption band is represented by the Lorentz equation and that the slit function is triangular [96]. Theoretically, the true integrated absorption coefficient AN of the Nth normal vibration is given by [3] " # qmy 2 np qmx 2 qmz 2 þ þ AN ¼ ð1:181Þ 3c qQN 0 qQN 0 qQN 0 where n is the number of molecules per cubic centimeter, and c is the velocity of light. As shown by Eq. 1.167, an internal coordinate Ri is related to a set of normal

90

THEORY OF NORMAL VIBRATIONS

coordinates by Ri ¼

X

ð1:182Þ

LiN QN

N

If the additivity of the bond dipole moment is assumed, it is possible to write ! ! X qm qm qRi ¼ qQN qRi qQN i ! ð1:183Þ X qm ¼ LiN qRi i Then Eq. 1.181 is written as " np AN ¼ 3c

X qm

np X ¼ 3c i

i

"

!2 x

qRi qmx qRi

þ

LiN !2 0

X qmy i

0

qmy þ qRi

!2 0

qRi

!2 þ

LiN

qmz þ qRi

X qm

0 !2 #

!2 # z

qRi

LiN 0

ð1:184Þ

ðLiN Þ2 0

This equation shows that the intensity of an infrared band depends on the values of the qm/qR terms as well as the L matrix elements. Equation 1.184 has been applied to relatively small molecules to calculate the qm/qR terms from the observed intensity and known LiN values [7]. However, the additivity of the bond dipole moment does not strictly hold, and the results obtained are often inconsistent and conflicting. Thus far, very few studies have been made on infrared intensities of large molecules because of these difficulties. As seen in Eq. 1.184, the IR band becomes stronger as the qm/qR term becomes larger if the LiN term does not vary significantly. In general, the more polar the bond, the larger the qm/qR term. Thus, the IR intensities of the stretching vibrations follow the general trends: nðOHÞ > nðNHÞ > nðCHÞ nðC ¼ OÞ > nðC ¼ NÞ > nðC ¼ CÞ It is also noted that the antisymmetric stretching vibration is always stronger than the symmetric stretching vibration because the qm/qR term is larger for the former than for the latter. This can be seen for functional groups such as These trends are entirely opposite to those found in Raman spectra (Sec. 1.21), and make both IR and Raman spectroscopy complementary. 1.20. DEPOLARIZATION OF RAMAN LINES As stated in Secs. 1.8 and 1.9, it is possible, by using group theory, to classify the normal vibration into various symmetry species. Experimentally, measurements of the

DEPOLARIZATION OF RAMAN LINES

91

Fig. 1.27. Experimental configuration for measuring depolarization ratios. The scrambler is placed after the analyzer because the monochromator gratings show different efficiencies for ? and | | directions.

infrared dichroism and polarization properties of Raman lines of an orientated crystal provide valuable information about the symmetry of normal vibrations (Sec. 1.31). Here, we consider the polarization properties of Raman lines in liquids and solutions in which molecules or ions take completely random orientations.* Suppose that we irradiate a molecule fixed at the origin of a space-fixed coordinate system with natural light from the positive-y direction, and observe the Raman scattering in the x direction as shown in Fig. 1.27. The incident light vector E may be resolved into two components, Ex and Ez, of equal magnitude (Ey ¼ 0). Both components give induced dipole moments, Px, Py, and Pz. However, only Py and Pz contribute to the scattering along the x axis, since an oscillating dipole cannot radiate in its own direction. Then, from Eq. 1.66, we have. Py ¼ ayx Ex þ ayz Ez Pz ¼ azx Ex þ azz Ez

ð1:185Þ ð1:186Þ

The intensity of the scattered light is proportional to the sum of squares of the individual aijEj terms. Thus, the ratio of the intensities in the y and z directions is rn ¼

Iy a2yx Ex2 þ a2yz Ez2 ¼ Iz a2zx Ex2 þ a2zz Ez2

ð1:187Þ

where rn is called the depolarization ratio for natural light (n). In a homogeneous liquid or gas, the molecules are randomly oriented, and we must consider the polarizability components averaged over all molecular *It is possible to obtain approximate depolarization ratios of fine powders where the molecules or ions take pseudorandom orientations (see Ref. 97).

92

THEORY OF NORMAL VIBRATIONS

(mean value) orientations. The results are expressed in terms of two quantities: a and g (anisotropy): 1 ¼ ðaxx þ ayy þ azz Þ a 3

g2 ¼

1h ðaxx ayy Þ2 þ ðayy azz Þ2 þ ðazz axx Þ2 þ 6ða2xy þ a2yz þ a2zx Þ 2

ð1:188Þ ð1:189Þ

These two quantities are invariant to any coordinate transformation. It can be shown [3] that the average values of the squares of aij are 1 ½45ð aÞ2 þ 4g2 45 1 ðaxy Þ2 ¼ ðayz Þ2 ¼ ðazx Þ2 ¼ g2 15 ðaxx Þ2 ¼ ðayy Þ2 ¼ ðazz Þ2 ¼

ð1:190Þ ð1:191Þ

Since Ex ¼ Ez ¼ E, Eq. 1.187 can be written as rn ¼

Iy 6g2 ¼ Iz 45ð aÞ2 þ 7g2

ð1:192Þ

The total intensity In is given by

1 2 2 ½45ð aÞ þ 13g E2 In ¼ Iy þ Iz ¼ const 45

ð1:193Þ

If the incident light is plane-polarized (e.g., a laser beam), with its electric vector in the z direction (Ex ¼ 0), Eq. 1.192 becomes rp ¼

Iy 3g2 ¼ Iz 45ð aÞ2 þ 4g2

ð1:194Þ

where rp is the depolarization ratio for polarized light (p). In this case, the total intensity is given by Ip ¼ Iy þ Iz ¼ const

1 ½45ð aÞ2 þ 7g2 E2 45

ð1:195Þ

The symmetry property of a normal vibration can be determined by measuring the depolarization ratio. From an inspection of character tables (Appendix I), it is obvious is nonzero only for totally symmetric vibrations. Then, Eq. 1.192 gives that a 0 rn < 67, and the Raman lines are said to be polarized. For all nontotally symmetric is zero, and rn ¼ 67. Then, the Raman lines are said to be depolarized. vibrations, a If the exciting line is plane polarized, these criteria must be changed according to

DEPOLARIZATION OF RAMAN LINES

93

Fig. 1.28. Raman spectra of CCI4 in two directions of polarization (488 nm excitation).

Eq. 1.194. Thus, 0 rp < 34 for totally symmetric vibrations, and rp ¼ 34 for nontotally symmetric vibrations. Figure 1.28 shows the Raman spectra of CCl4 (500–150 cm1) in two directions of polarization obtained with the 488 nm excitation. The three bands at 459, 314, and 218 cm1 give rp values of approximately 0.02, 0.75, and 0.75, respectively. Thus, it is concluded that the 459 cm1 band is polarized (A1), whereas the two bands at 314 (F2) and 218 (E) cm1 are depolarized. As stated in Sec. 1.6, the polarizability tensors are symmetric in normal Raman scattering. If the exciting frequency approaches that of an electronic absorption, some scattering tensors become antisymmetric,* and resonance Raman scattering can occur (Sec. 1.22). In this case, Eq. 1.194 must be written in a more general form [98] rp ¼

3gs þ 5ga 10go þ 4gs

ð1:196Þ

*A tensor is called antisymmetric if axx ¼ ayy ¼ azz ¼ 0 and axy ¼ ayx,ayz ¼ azy, and azx ¼ axz.

94

where

THEORY OF NORMAL VIBRATIONS

aÞ2 go ¼ 3ð h i 1 gs ¼ ðaxx ayy Þ2 þ ðayy azz Þ2 þ ðazz axx Þ2 3 i 1h þ ðaxy þ ayx Þ2 þ ðaxz þ azx Þ2 þ ðayz þ azy Þ2 2 h i 1 ga ¼ ðaxy ayx Þ2 þ ðaxz azx Þ2 þ ðayz azy Þ2 2

ð1:197Þ

If we define 3 g2s ¼ gs 2

and

3 g2as ¼ ga 2

ð1:198Þ

Eq. 1.196 can be written as rp ¼

3g2s þ 5g2as

ð1:199Þ

45ð aÞ2 þ 4g2s

In normal Raman scattering, g2s ¼ g2 and g2as ¼ 0. Then Eq. 1.199 is reduced to Eq. 1.194. The symmetry properties of resonance Raman lines can be predicted on the basis „ 0 and gas ¼ 0. Then, Eq. 20.15 of Eq. 1.199. For totally symmetric vibrations, a gives 0 rp < 34. Nontotally symmetric vibrations ð a ¼ 0Þ are classified into two types: those that have symmetric scattering tensors, and those that have antisymmetric scattering tensors. If the tensor is symmetric, gas ¼ 0 and gs „ 0. Then Eq. 1.199 gives rp ¼ 34 (depolarized). If the tensor is antisymmetric, gas „ 0 and gs ¼ 0. Then, Eq. 1.199 gives rp ¼ ¥ (inverse polarization). In the case of the D4h point group, the B1g and B2g representations belong to the former type, whereas the A2g representations belongs to the latter [99]. As will be shown in Sec. 1.22, Spiro and Strekas [98] observed for the first time A2g vibrations which exhibit rp values 3 4 < rp < ¥. For these vibrations, the term “anomalous polarization (ap)” is used instead of “inverse polarization (ip)” [100]. 1.21. INTENSITY OF RAMAN SCATTERING According to the quantum mechanical theory of light scattering, the intensity per unit solid angle of scattered light arising from a transition between states m and n is given by In

m

¼ constðn0 nmn Þ4

X rs

jðPrs Þmn j2

ð1:200Þ

INTENSITY OF RAMAN SCATTERING

95

Fig. 1.29. Energy-level diagram for Raman scattering.

where ðPrs Þmn

" # 1 X ðMr Þrn ðMs Þmr ðMr Þmr ðMs Þrn ¼ ðars Þmn E ¼ þ E h r nrm n0 nrn þ n0

ð1:201Þ

Here n0 is the frequency of the incident light: nrm, nrn, and nmn are the frequencies corresponding to the energy differences between subscripted states;Ð terms of the type (Ms)mr are the Cartesian components of transition moments such as C r ms Cm dt; and E is the electric vector of the incident light. It should be noted here that the states denoted by m, n, and r represent vibronic states cg ðx; QÞfgi ðQÞ; cg ðx; QÞfgj ðQÞ and ce ðx; QÞfev ðQÞ, respectively, where cg and ce are electronic ground- and excited-state wavefunctions, respectively, and fgi ; fgj , and fev vibrational functions. The symbols x and Q represent the electronic and nuclear coordinates, respectively. These notations are illustrated in Fig. 1.29. Finally, s and r denote x, y, and z components. Since the electric dipole operator acts only on the electronic wavefunctions, the (ars)mn term in Eq. 1.201 can be written in the form ðars Þmn ¼

ð 1 fgj ðars Þgg fgj dQ h

ð1:202Þ

where ðars Þgg ¼

X e

Ð

! Ð Ð Ð c g ms ce dt c e mr cg dt cg mr ce dt c e ms cg dt þ neg n0 neg þ n0

Here neg corresponds to the energy of a pure electronic transition between the ground and excited states.

96

THEORY OF NORMAL VIBRATIONS

To discuss the Raman scattering, we expand the (ars)gg term as a Taylor series with respect to the normal coordinate Q: " # qðars Þgg 0 ðars Þgg ¼ ðars Þgg þ Qþ ð1:203Þ qQ 0

Then, we write Eq. 1.202 as ðars Þmn

" # ð ð 1 1 qðars Þgg g g 0 fgj Qfgi dQ ¼ ðars Þgg fj fi dQ þ h h qQ

ð1:204Þ

0

The first term on the right-hand side is zero unless i ¼ j. This term is responsible for Rayleigh scattering. The second term determines the activity offundamental vibrations in Raman scattering; it vanishes for a harmonic oscillator unless j ¼ i 1 (Sec. 1.10). If we consider a Stokes transition, v ! v þ 1, Eq. 1.204 is written as [3] ðars Þv;v þ 1

" # sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 qðars Þgg ðv þ 1Þh ¼ h 8p2 mn qQ

ð1:205Þ

0

where m and n are the reduced mass and the Stokes frequency. Then Eq. 1.200 is written as " #2 E2 qðaps Þgg ðv þ 1Þh I ¼ constðn0 nÞ 2 8p2 mn h qQ 4

ð1:206Þ

0

In Sec. 1.20, we derived a classical equation for Raman intensity: !2 qa In ¼ const E2 qQ ( ) 1 ½45ð aÞ2 þ 13g2 E2 ¼ const 45

ð1:193Þ

By replacing the qa/qQ term of Eq. 1.206 with the term enclosed in braces in Eq. 1.193, we obtain ðv þ 1Þ E2 In ¼ constðn0 nÞ 8p2 mn h2 4

(

1 ½45ð aÞ2 þ 13g2 45

) ð1:207Þ

At room temperature, most of the scattering molecules are in the v ¼ 0 state, but some are in higher vibrational states. Using the Maxwell–Bolztmann distribution

INTENSITY OF RAMAN SCATTERING

97

law, we find that the fraction of molecules fv with vibrational quantum number v is given by e ½v þ ð1=2Þhn=kT fv ¼ P ½v þ ð1=2Þhn=kT ve

ð1:208Þ

P Then the total intensity is proportional to v fv ðv þ 1Þ, which is equal to (1 ehv/kT)1 (see Ref. 23). Hence we can rewrite Eq. 1.207 in the form In ¼ KI0

ðn0 nÞ4 ½45ð aÞ2 þ 13g2 mnð1 e hn=kT Þ

ð1:209Þ

Here I0 is the incident light intensity which is proportional to E2, and K summarizes all other constant terms. If the incident light is polarized, the form of Eq. 1.209 is slightly modified: Ir ¼ KI0

ðn0 nÞ4 ½45ð aÞ2 þ 7g2 mnð1 e hn=kT Þ

ð1:210Þ

As shown in Sec. 1.20, the degree of depolarization rp is rp ¼

3g2 45ð aÞ2 þ 4g2

or

g2 ¼

45ð aÞ2 rp 3 4rp

ð1:211Þ

Since rp ¼ 34 for nontotally symmetric vibrations, Eq. 1.211 holds only for totally symmetric vibrations. Then Eq. 1.210 is written as ! 1 þ rp ðn0 nÞ4 Ip ¼ K I0 ð a Þ2 mnð1 e hn=kT Þ 3 4rp 0

ð1:212Þ

In the case of a solution, the intensity is proportional to the molar concentration C. Then Eq. 1.212 is written as ! 1 þ rp Cðn0 nÞ4 Ip ¼ K I0 ð aÞ2 mnð1 e hn=kT Þ 3 4rp 00

ð1:213Þ

If we compare the intensities of totally symmetric vibrations (A1 mode) of two tetrahedral XY4-type molecules, the intensity ratio is given by I1 C1 ~n0 ~n1 ¼ I2 C2 ~n0 ~n2

!4

~n2 m2 ð1 e hc~n2 =kT Þð a1 Þ 2 ~n1 m1 ð1 e hc~n1 =kT Þð a2 Þ 2

ð1:214Þ

98

THEORY OF NORMAL VIBRATIONS

In this case, the rp term drops out, since g2 ¼ 0 for isotropic molecules such as tetrahedral XY4 and octahedral XY6 types. By using CCI4 as the standard, it is possible to determine the relative value of the qa/qQ term, which provides information about the degree of covalency and the bond order [23]. As seen above, the intensity of Raman scattering increases as the (qa/qQ)0 term becomes larger. The following general rules may reflect the trends in the (qa/qQ)0 values: (1) stretching vibrations produce stronger Raman bands than do bending vibrations; (2) symmetric vibrations produce stronger Raman bands than do antisymmetric vibrations, and totally symmetric “breathing” vibrations produce the most intense Raman bands; (3) stretching vibrations of covalent bonds produce Raman bands stronger than those of ionic bonds (among the covalent bonds, Raman intensities increase as the bond order increases; thus, the ratio of relative intensities of the CC, C¼C, and CC stretching vibrations is approximately 3:2:1); and (4) bonds involving heavier atoms produce stronger Raman bands than do those involving lighter atoms. For example, the n(SS) is stronger than the n(CC). Coordination compounds containing heavy metals and sulfur ligands are ideal for Raman studies. 1.22. PRINCIPLE OF RESONANCE RAMAN SPECTROSCOPY In normal Raman spectroscopy, the exciting frequency lies in the region where the compound has no electronic absorption band. In resonance Raman spectroscopy, the exciting frequency falls within the electronic band (Sec. 1.2). In the gaseous phase, this tends to cause resonance fluorescence since the rotational–vibrational levels are discrete. In the liquid and solid states, however, these levels are no longer discrete because of molecular collisions and/or intermolecular interactions. If such a broad vibronic bands is excited, it tends to give resonance Raman rather than resonance fluorescence spectra [101,102]. Resonance Raman spectroscopy is particularly suited to the study of biological macromolecules such as heme proteins because only a dilute solution (biological condition) is needed to observe the spectrum and only vibrations localized within the chromophoric group are enhanced when the exciting frequency approaches that of the relevant chromophore. This selectivity is highly important in studying the theoretical relationship between the electronic transition and the vibrations to be resonanceenhanced. The origin of resonance Raman enhancement is explained in terms of Eq. 1.201. In normal Raman spectroscopy, n0 is chosen in the region that is far from the electronic absorption. Then, nrm n0, and ars is independent of the exciting frequency n0. In resonance Raman spectroscopy, the denominator, nrm n0, becomes very small as n0 approaches nrm. Thus, the first term in the square brackets of Eq. 1.201 dominates all other terms and results in striking enhancement of Raman lines. However, Eq. 1.201 cannot account for the selectivity of resonance Raman enhancement since it is not specific about the states of the molecule. Albrecht [103] derived a more specific equation for the initial and final states of resonance Raman scattering by

PRINCIPLE OF RESONANCE RAMAN SPECTROSCOPY

99

introducing the Herzberg–Teller expansion of electronic wave functions into the Kramers–Heisenberg dispersion formula. The results are as follows: ðars Þgi;gj ¼ A þ B þ C

"

#

ð1:215Þ

0 X ðg0 jRs je0 Þðe0 jRr jg0 Þ X þ ðnonresonance termÞ hijvihvjji ð1:216Þ Eev Egi E0 e„g v # (" 0 0 XX X X ðg0 jRs je0 Þðe0 jha js0 Þðs0 jRr jg0 Þ þ ðnonresonance termÞ B¼ Eev Egi E0 e„g v s„e a

A¼

" # hi=vihv=Qa jji ðg0 jRs js0 Þðs0 jha je0 Þðe0 jRr jg0 Þ þ þ ðnonresonance termÞ Ee0 Es0 Eev Egi E0 hijQa jvihv=ji Ee0 Es0 C¼

0 0 X X XX

e„g t„g

v

) ð1:217Þ ("

a

ðe0 jRr jg0 Þðg0 jha jt0 Þðt0 jRs je0 Þ þ ðnonresonance termÞ Eev Egi E0

#

" # hijvihvjQa jji ðe0 jRr jt0 Þðt0 jha jg0 Þðg0 jRs je0 Þ þ þ ðnonresonance termÞ Eg0 Et0 Eev Egi E0 hijQa jvihvjji Eg0 Et0

) ð1:218Þ

The notations g, i, j, e, and v were explained in Sec. 1.21. Other notations are as follows: s, another excited electronic state; ha, the vibronic coupling operator qH=qQa ; H, and Qa, the electronic Hamiltonian and the ath normal coordinate of the electronic ground state, respectively; Egi and Eev, the energies of states gi and ev, respectively; |g0), |e0), and |s0), the electronic wavefunctions for the equilibrium nuclear positions of the ground and excited states; Ee0 and Es0 , the corresponding energies of the electronic states, e0 and s0, respectively; and E0, the energy of the exciting light. The nonresonance terms are similar to the preceding terms except that the denominator is (Eev Egj þ E0) instead of (Eev Egj E0) and that Rs and Rp (Cartesian components of the transition moment) in the numerator are interchanged. These terms can be neglected under the strict resonance condition since the resonance terms become very large. The C term is usually neglected because its components are denominated by Eg0 Et0, where t refers to an excited state that is much higher in energy than the first excited state. These notations are shown on the right-hand side of Fig. 1.29. In more rigorous expressions, the damping factor iG must be added to the

100

THEORY OF NORMAL VIBRATIONS

Fig. 1.30. Shift of equilibrium position caused by totally symmetric vibration.

denominators such as Eev Egj E0 so that the resonance term does not become infinity under rigorous resonance conditions [103]. For the A term to be nonzero, the g $ e electronic transition must be allowed, and the product of the integrals hi | vihv | ji (Franck–Condon factor) must be nonzero. The latter condition is satisfied if the equilibrium position is shifted by a transition (D „ 0 as shown in Fig. 1.30). Totally symmetric vibrations tend to be resonance-enhanced via the A term since they tend to shift the equilibrium position upon electronic excitation. If the equilibrium position is not shifted by the g $ e transition, the A term becomes zero since either one of the integrals, (i | v) or (v | j), becomes zero. This situation tends to occur for nontotally symmetric vibrations. In contrast to the A term, the B term resonance requires at least one more electronic excited state(s), which must be mixed with the e state via normal vibration, Qa. Namely, the integral (e0 | ha | s0) must be nonzero. The B-term resonance is significant only when these two excited states are closely located so that the denominator, Ee0 Es0 , is small. Other requirements are that the g $ e and g $ s transitions should be allowed and that the integrals hi | vi and hv | Qa | ji should not be zero. The B term can cause resonance enhancement of nontotally symmetric as well as totally symmetric vibrations, although it is mainly responsible for resonance enhancement of the former. As seen in Eqs. 1.216 and 1.217, both A and B terms involve summation over v, which is the vibrational quantum number at the electronic excited state. Since the harmonic oscillator selection rule (Dv ¼ 1) does not hold for a large v, overtones and combination bands may appear under resonance conditions. In fact, series of these nonfundamental vibrations have been observed in the case of A-term resonance (Sec. 1.23).

RESONANCE RAMAN SPECTRA

101

In Sec. 1.20, we have shown that the depolarization ratio for totally symmetric vibrations is in the range 0 rp < 34. For the A term to be nonzero, however, the term (g0 | Rs | e0)(e0 | Rr | g0) must be nonzero. Since g0 is totally symmetric, this condition is realized only when both c(Rs)c(e0) and c(Rr)c(e0) belong to the totally symmetric species. Thus, c(Rs) ¼ c(Rr), and only one of the diagonal terms of the polarizability tensor (axx, ayy, or azz) becomes nonzero. Then, it is readily seen from Eq. 1.199 that rp ¼ 13. This rule has been confirmed by many observations. If the electronic excited state is degenerate, two of the diagonal terms must be nonzero. In this case, rp ¼ 18. These rules hold for any point group with at least one C3 or higher axis, with the exception of the cubic groups in which the Cartesian coordinates transform as a single irreducible degenerate representation. 1.23. RESONANCE RAMAN SPECTRA Because of the advantages mentioned in the preceding section, resonance Raman (RR) spectroscopy has been applied to vibrational studies of a number of inorganic as well as organic compounds. It is currently possible to cover the whole range of electronic transitions continuously by using excitation lines from a variety of lasers and Raman shifters [26]. In particular, the availability of excitation lines in the UV region has made it possible to carry out UV resonance Raman (UVRR) spectroscopy [104]. In RR spectroscopy, the excitation line is chosen inside the electronic absorption band. This condition may cause thermal decomposition of the sample by local heating. To minimize thermal decomposition, several techniques have been developed. These include the rotating sample technique, the rotating (or oscillating) laser beam technique, and their combinations with low-temperature techniques [21,26]. It is always desirable to keep low laser power so that thermal decomposition is minimal. This will also minimize photodecomposition, which occurs depending on laser lines in some compounds. An example of the A-term resonance is given by I2, which has an absorption band near 500 nm and its fundamental at 210 cm1. Figure 1.31 shows the RR spectrum of I2 in solution obtained by Kiefer and Bernstein [105] (excitation at 514.5 nm). It shows a series of overtones up to the seventeenth. The vibrational energy of an anharmonic oscillator including the cubic term is expressed as 1 1 2 1 3 Ev ¼ hcwe v þ þ hcye we v þ hcce we v þ 2 2 2 The anharmonicity constants, cewe and yewe can be determined by plotting the ~nðobsÞ=v against the vibrational quantum number v, as shown in Fig. 1.32. The straight line gives cewe and the deviation from the straight line (at higher v) gives yewe The results (CCl4 solution) in cm1 are we ¼ 212:59 0:08;

ce we ¼ 0:62 0:03;

ye we ¼ 0:005 0:002

102

THEORY OF NORMAL VIBRATIONS

Fig. 1.31. Resonance Raman spectra of I2 in CCl4, CHCl3, and CS2 (514.5 nm excitation) [105].

As stated in the preceding section, the depolarization ratio (rp) is expected to be closed to 13 for totally symmetric vibrations. However, this value increases as the change in vibrational quantum number (Dv) increases. For example, it is 0.48 for the tenth overtone of I2 in CCl4 [100]. Carbontetrachloride 210

ν∼ υ

205 cm-1

200 1

5

10

υ

Fig. 1.32. Plot of v~=v vs. n for I2 in CCl4 [105].

15

RESONANCE RAMAN SPECTRA

103

Fig. 1.33. Resonance Raman spectrum of solid TiI4 (514.5 nm excitation) [106].

The second example is given by TiI4, which has the absorption maximum near 514 nm. Figure 1.33 shows the RR spectrum of solid TiI4 obtained by Clark and Mitchell [106] with 514.5 nm excitation. In this case, a series of overtones up to the twelfth has been observed. Using these data, they obtained the following constants (cm1): we ¼ 161:0 0:2

and X11 ¼ 0:11 0:03

Here, X11 is the anharmonicity constant corresponding to cewe of a diatomic molecule (Sec. 1.3). Theoretical treatments of Raman intensities of these overtone series under rigorous resonance conditions have been made by Nafie et al. [107]. There are many other examples of RR spectra of small molecules and ions that exhibit A-term resonance. Section 2.6 includes references on RR spectra of the CrO2 4 and MnO4 ions. Kiefer [101] reviewed RR studies of small molecules and ions. Hirakawa and Tsuboi [108] noted that among the totally symmetric modes, the mode that leads to the excited-state configuration is most strongly resonance enhanced. For example, the NH3 molecule is pyramidal in the ground state and planar in the excited state (216.8 nm above the ground state). Then, the symmetric bending mode near 950 cm1 is enhanced 10 times more than the symmetric stretching mode near 3300 cm1 when the excitation wavelength is changed from 514.5 to 351.1 nm. A typical example of the B-term resonance is given by metalloporphyrins and heme proteins. As shown in Fig. 1.34, Ni(OEP) (OEP:octaethylporphyrin) exhibits two electronic transitions referred to as the Q0 (or a) and B (or Soret) bands along with a vibronic side band (Q1 or b) in the 350–600 nm region. According to MO (molecular orbital) calculations on the porphyrin core of D4h, symmetry, Q0 and B transitions

104

THEORY OF NORMAL VIBRATIONS

Fig. 1.34. Structure, energy-level diagram, and electronic spectrum of Ni(OEP) [109].

result from strong interaction between the a1u(p) ! eg (p*) and a2u(p) ! eg (p*) transitions which have similar energies and the same excited-state symmetry (Eu). The transition dipoles add up for the strong B transition and nearly cancel out for the weak Q0 transition. This is an ideal situation for B-term resonance. According to Eq. 1.217, any normal modes which give nonzero values for the integral (e0 | ha | s0) are enhanced via the B term. These vibrations must belong to one of the symmetry species given by Eu Eu, that is, A1g þ B1g þ B2g þ A2g.* Figure 1.35 shows the RR spectra of Ni(OEP) with excitation near the B, Q1 and Q0 transitions as obtained by Spiro and coworkers [109]. As discussed in Sec. 1.20, polarization properties are expected to be A1g (p, polarized), B1g and B2g (dp, depolarized), and A2g (ap, anomalous polarization). These polarization properties, together with normal coordinate analysis, were used to make complete vibrational assignments of Ni(OEP) [110]. It is seen that the spectra obtained by excitation near the Q0 band (bottom traces) are dominated by the dp bands (B1g and B2g species), while those obtained by excitation between the Q0 and Q1 bands (middle traces) are dominated by the ap bands (A2g species). The spectra obtained by excitation near the B band (top traces) are dominated by the p bands (A1g species). In this

*For symmetry species of the direct products, see Appendix IV. Also note that the A1g vibrations are not effective in vibronic mixing.

RESONANCE RAMAN SPECTRA

105

Fig. 1.35. Resonance Raman spectra of Ni(OEP) obtained by excitation near the B, Q1, and Q0 bands [110].

case, the A-term resonance is much more important than the B-term resonance because of the very strong absorption strength of the B band. Mizutani and coworkers [110a] analyzed the resonance Raman spectra of Ni(OEP) obtained at the d–d excited state. As will be shown in Sec. 2.8, anomalously polarized bands have also been observed in the RR spectra of the IrCl26 ion, although their origin is different from the vibronic coupling mentioned above [111,112]. The majority of compounds studied thus far exhibit the A-term rather than the B-term resonance. A more complete study of resonance Raman spectra involves the observation of excitation profiles (Raman intensity plotted as a function of the excitation frequency for each mode), and the simulation of observed excitation profiles based on theoretical treatments of resonance Raman scattering [113].

106

THEORY OF NORMAL VIBRATIONS

1.24. THEORETICAL CALCULATION OF VIBRATIONAL FREQUENCIES The procedure for calculating harmonic vibrational frequencies and force constants by GF matrix method has been described in Sec. 1.12. In this method, both G (kinetic energy) and F (potential energy) matrices are expressed in terms of internal coordinates (R) such as increments of bond distances and bond angles. Then, the kinetic (T) and potential (V) energies are written as:

~_ 1 R_ 2T ¼ RG

ð1:113Þ

RFR 2V ¼ ~

ð1:111Þ

jGF Elj ¼ 0

ð1:116Þ

and the secular equation

is solved to obtain vibrational frequencies. Here, E is a unit matrix, and l is related to the wavenumber by the relation given by Eq. 1.156. The G matrix elements can be calculated from the known bond distances and angles, and the F matrix elements (force constants) are refined until the differences between the observed and calculated frequencies are minimized. Force constants thus obtained can be transferred to similar molecules to predict their vibrational frequencies. On the other hand, quantum mechanical methods utilize an entirely different approach to this problem. The principle of the method is to express the total electronic energy (E) of a molecule as a function of nuclear displacements near the equilibrium positions and to calculate its second derivatives with respect to nuclear displacements, namely, vibrational force constants. In the past, ab initio calculations of force constants were limited to small molecules [114] such as H2O and C2H2 because computational time required for large molecules was prohibitive. This problem has been solved largely by more recent progress in density functional theory (DFT) [115] coupled with commercially available computer software. Currently, the DFT method is used almost routinely to calculate vibrational parameters of a variety of inorganic, organic, coordination, and organometallic compounds. The advent of high-speed personal computers further facilitated this trend. Using a commercially available computer program such as “Gaussian 03,” DFT calculations of the H2O molecule proceed in the following order: First, molecular parameters (masses of constituent atoms, approximate bond distances and angles) are used as input data. As an example, let us assume that the OH distance is 0.96 A and the HOH angle is 109.5 . Then, the program calculates the cartesian coordinates of the three atoms by placing the origin at the center of gravity (CG) of the molecule. It also determines the point group of the molecule as C2v on the basis of the given geometry. Since the total electronic enegy (E) of the H2O molecule is a function of the nine Cartesian coordinates (xi, i where i ¼ 1, 2, . . ., 9), the equilibrium configuration that renders all the nine first derivatives (qE/qxi) zero can be determined (energy optimization). Accuracy of the results depends on the level of the approximation method

THEORETICAL CALCULATION OF VIBRATIONAL FREQUENCIES

107

TABLE 1.15. Calculated Cartesian Coordinates of H2O ˚) Coordinate (A Center Number 1 2 3

Atomic Number

X

8 1 1

0.0000 0.0000 0.0000

Y

Z

0.0000 0.8010 0.8010

0.1110 0.4440 0.4440

chosen in calculating E. Using 6-31G* basis set with B3LYP functional [116], for example, the optimal geometry was found to be: OH distance ¼ 0.9745 A and HOH angle ¼ 110.5670 These values were used to recalculate the Cartesian coordinates listed in Table 1.15 The next step is to calculate the second derivatives near the equilibrium positions [(qE2/qxi qxj)0,i,j ¼ 19], namely, the force constants in terms of Cartesian coordinates (fi,j), using analytical methods [116a]. The force constants thus obtained are converted to mass-weighted Cartesian coordinates (XM,i) by xM;1 ¼

pﬃﬃﬃﬃﬃﬃ m 1 x1 ;

pﬃﬃﬃﬃﬃﬃ m 1 x2 ; . . . ! q2 V qxMi qxMj

xM;2 ¼

fi;j fM;i;j ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ mi mj

ð1:219Þ ð1:220Þ

0

These fM,i,j terms are the elements of the force constant (Hessian) matrix. Using matrix expression, the relationship between mass-weighted Cartesian (XM) and Cartesian (X) coordinates is written as follows: XM ¼ M1=2 X

ð1:221Þ

pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ Here, M1/2 is a diagonal matrix whose elements are m1 ; . . . m9 and X is a column matrix, [x1,. . .,x9]. The product, M1/2X, gives a column matrix, XM, pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ ½ m1 x1 ; . . . ; m9 x9 . Equation (1.221) is rewritten as X ¼ M ð1=2Þ XM

ð1:222Þ

The relationship between the internal coordinate R and X (Eqs. 1.118 and 1.119) is now written as R ¼ BX ¼ BM ð1=2Þ XM

ð1:223Þ

~R ¼ ~XM M ð1=2Þ~B

ð1:224Þ

and its transpose is

where B is a rectangular (3 9) matrix shown by Eq. 1.119.

108

THEORY OF NORMAL VIBRATIONS

The kinetic energy (Eq. 1.113) in terms of XM is written as

~_ M X_ M 2T ¼ X

ð1:225Þ

and the potential energy (Eq. 1.111) in terms of the same coordinate, XM, is 2V ¼ ~ RFR R ¼ ~ XM M ð1=2Þ~ BFR BM ð1=2Þ XM ¼ ~ XM FXM XM

ð1:226Þ

where

or

FXM ¼ M ð1=2Þ~ BFR BM ð1=2Þ

ð1:227Þ

~BFR B ¼ M1=2 FXM M1=2

ð1:228Þ

Vibrational frequencies are calculated by solving the following equation: jFXM Elj ¼ 0

ð1:229Þ

In the case of H2O, we obtain nine frequencies, six of which are near zero. These six frequencies must be eliminated from the results because they correspond to the translational and rotational motions of the molecule as a whole (3N–6 rule). Thus, the remaining three frequencies correspond to the normal modes of the H2O molecule. It should be noted that vibrational frequencies are intrinsic of individual molecules and do not depend on the coordinate system chosen (R or XM). The next step is to convert the force constants (FXM) in terms of XM coordinates to those of internal coordinates (FR), which have clearer physical meanings. For this purpose, Eq. 1.228 is rewritten as [117–119] FR ¼ UG 1~ UBM1=2 FXM M1=2~ BUG 1~ U Here, U is an orthogonal matrix ðU ¼ ~ U matrix that is defined as follows:

1

ð1:230Þ

Þ that diagonalizes B~ B, and G is a diagonal

B~ BU ¼ UG

ð1:231Þ

In the case of H2O, B is a 3 9 matrix, whereas U, G, and FR are 3 3 matrices. To take advantage of symmetry properties, internal coordinates are further transformed to internal symmetry coordinates (RS) by the relation RS ¼ US R

ð1:131Þ

where U is an orthogonal matrix chosen by symmetry consideration. Then, the force constants in terms of internal symmetry coordinates ðFSR Þ are given by S

FSR ¼ US FR~ U

S

ð1:232Þ

VIBRATIONAL SPECTRA IN GASEOUS PHASE AND INERT GAS MATRICES

109

TABLE 1.16. Comparison of Observed and Calculated Frequencies (cm1) and Normal Modes of H2O Normal Vibration: a

Observed frequencies Calculated frequencies Normal modesb

X

n1(A1)

n2(A1)

n3(B2)

1595 1565.5

3657 3639.2

3756 3804.7

Y

Z

X

Y

Z

X

Y

Z

O(1) H(2) H(3)

0.00 0.00 0.08 0.00 0.38 0.60 0.00 0.38 0.60

0.00 0.00 0.04 0.00 0.62 0.35 0.00 0.62 0.35

0.00 0.07 0.00 0.00 0.58 0.40 0.00 0.58 0.40

Assignment

HOH bending

Symmetric OH stretch

Antisymmetric OH stretch

a

IR spectrum (Table 2.2d of Chapter 2). The normal modes shown in Fig. 1.12 are obtained by plotting these displacement coordinates.

b

Thus, FXM can be coverted to FSR using Eqs. 1.230 and 1.232. Several computer programs are available for this conversion [118]. Table 1.16 summarizes the results of DFT calculation of H2O. It is seen that the calculated frequencies are in good agreement with the observed values. In other cases, however, the former becomes higher than the latter because (1) the effect of anharmonicity is not included in the calculation and (2) systematic errors result from the use of finite basis sets and incomplete inclusion of electron correlation effects. In such cases, several empirical “scaling methods” are applied to the force constants [120] or frequencies [121] obtained by DFT calculations. Computer programs such as “Gaussian 03” can simultaneously calculate several other parameters including normal modes (shown in Table 1.15), IR intensities, Raman scattering intensities, and depolarization ratios. 1.25. VIBRATIONAL SPECTRA IN GASEOUS PHASE [122] AND INERT GAS MATRICES [29,30] As distinct from molecules in condensed phases, those in the gaseous phase are free from intermolecular interactions. If the molecules are relatively small, vibrational spectra of gases exhibit rotational fine structure (see Fig. 1.3) from which moments of inertia and hence internuclear distances and bond angles can be calculated [1,2]. Furthermore, detailed analysis of rotational fine structure provides information about the magnitude of rotation–vibration interaction (Coriolis coupling), centrifugal distortion, anharmonicity, and even nuclear spin statistics in some cases. In the past, infrared spectroscopy was the main tool in measuring gas-phase vibrational spectra. However, Raman spectroscopy is also playing a significant role because of the development of powerful laser sources and high-resolution spectrophotometers. For example, Clark and Rippon [123] measured gas-phase Raman spectra of Group IV tetrahalides, and calculated the Coriolis coupling constants from the observed band contours. For gas-phase Raman spectra of other inorganic compounds, see

110

THEORY OF NORMAL VIBRATIONS

Ref. [122]. Unfortunately, the majority of inorganic and coordination compounds exist as solids at room temperature. Although some of these compounds can be vaporized at high temperatures without decomposition, it is rather difficult to measure their spectra by the conventional method. Furthermore, high-temperature spectra are difficult to interpret because of the increased importance of rotational and vibrational hot bands. In 1954, Pimentel and his colleagues [124] developed the matrix isolation technique to study the infrared spectra of unstable and stable species. In this method, solute molecules and inert gas molecules such as Ar and N2 are mixed at a ratio of 1:500 or greater and deposited on an IR window such as a CsI crystal cooled to 10–15 K. Since the solute molecules trapped in an inert gas matrix are completely isolated from each other, the matrix isolation spectrum is similar to the gas-phase spectrum; no crystal field splittings and no lattice modes are observed. However, the former spectrum is simpler than the latter because, except for a few small hydride molecules, no rotational transitions are observed because of the rigidity of the matrix environment at low temperatures. The lack of rotational structure and intermolecular interactions results in a sharpening of the solute band so that evenvery closely located metal isotope peaks can be resolved in a matrix environment. This technique is also applicable to a compound which is not volatile at room temperature. For example, matrix isolation spectra of metal halides can be measured by vaporizing these compounds at high temperatures in a Knudsen cell and co-condensing their vapors with inert gas molecules on a cold window [125]. More recent developments of laser ablation techniques coupled with closed-cycle helium refrigerators greatly facilitated its application. Thus, matrix isolation technique has now been utilized to obtain structural and bonding information of a number of inorganic and coordination compounds. Some important features and applications are described in this and the following sections. 1.25.1. Vibrational Spectra of Radicals Highly reactive radicals can be produced in situ in inert gas matrices by photolysis and other techniques. Since these radicals are stabilized in matrix environments, their spectra can be measured by routine spectroscopic techniques. For example, the spectrum of the HOO radical [126] was obtained by measuring the spectrum of the photolysis product of a mixture of HI and O2 in an Ar matrix at 4 K. Chapter 2 lists the vibrational frequencies of many other radicals, such as CN, OF, and HSi, obtained by similar methods. 1.25.2. Vibrational Spectra of High-Temperature Species Alkali halide vapors produced at high temperatures consist mainly of monomers and dimers. The vibrational spectra of these salts at high temperatures are difficult to measure and difficult to interpret because of the presence of hot bands. The matrix isolation technique utilizing a Knudsen cell has solved this problem. The

VIBRATIONAL SPECTRA IN GASEOUS PHASE AND INERT GAS MATRICES

111

vibrational frequencies of some of these high-temperature species are listed in Chapter 2. 1.25.3. Isotope Splittings As stated before, it is possible to observe individual peaks due to heavy-metal isotopes in inert gas matrices since the bands are extremely sharp (half-band width, 1.5–1.0 cm1) under these conditions. Figure 1.36a shows the infrared spectrum of the n7 band (coupled vibration between CrC stretching and CrCO bending modes) of Cr(CO)6 in a N2 matrix [127]. The bottom curve shows a computer simulation using the measured isotope shift of 2.5 cm1 per atomic mass unit, a 1.2 cm1 half-band width, and the percentages of natural abundance of Cr isotopes: 50 Cr (4.31%), 52 Cr (83.76%), 53 Cr (9.55%), and 54 Cr (2.38%). The isotope splittings of SnCl4 and GeCl4 in Ar matrices are discussed in Sec. 2.6. These isotope frequencies are highly important in refining force constants in normal coordinate analysis.

Fig. 1.36. Matrix isolation IR spectra of Cr(CO)6 in N2 (a) and Ar (b) matrices. The bottom spectra were obtained by computer simulation [127].

112

THEORY OF NORMAL VIBRATIONS

1.25.4. Matrix Effect The vibrational frequencies of matrix-isolated molecules give slight shifts when the matrix gas is changed. This result suggests the presence of weak interaction between the solute and matrix molecules. In some cases, the spectra are complicated by the presence of more than one trapping site. For example, the infrared spectrum of Cr(CO)6 in an Ar matrix (Fig. 1.36b) is markedly different from that in a N2 matrix (Fig. 1.36a) [127]. The former spectrum can be interpreted by assuming two different sites in an Ar matrix. The computer-simulation spectrum (bottom curve) was obtained by assuming that these two sites are populated in a 2:1 ratio, the frequency separation of the corresponding peaks being 2 cm1. Thus, it is always desirable to obtain the matrix isolation spectra in several different environments.

1.26. MATRIX COCONDENSATION REACTIONS A number of unstable and transient species have been synthesized via matrix cocondensation reactions, and their structure and bonding have been studied by vibrational spectroscopy. The principle of the method is to cocondense two solute vapors (atom, salt, or molecule) diluted by an inert gas on an IR window (IR spectroscopy) or a metal plate (Raman spectroscopy) that is cooled to low temperature by a cryocooler. Solid compounds can be vaporized by conventional heating (Knudsen cell), laser ablation, or other techniques, and mixed with inert gases at proper ratios [128]. In general, the spectra of the cocondensation products thus obtained exhibit many peaks as a result of the mixed species produced. In order to make band assignments, the effects of changing the temperature, concentration (dilution ratios), and isotope substitution on the spectra must be studied. In some cases, theoretical calculations (Sec. 1.24) must be carried out to determine the structure and to make band assignments. Vibrational frequencies of many molecules and ions obtained by matrix cocondensation reactions are listed in Chapter 2. Matrix cocondensation reactions are classified into the following five types: 1. Atom-Atom Reaction. Liang et al. [129] studied the reaction: U þ S ! USn(n ¼ 1–3), Figure 1.37 shows the IR spectra of laser-ablated U atom vapor cocondensed with microwave-discharged S atom in Ar at 7 K. In (a) (32 S with higher concentration) and (b) (32 S with low concentration), the prominent band at 438.7 cm1 and a weak band at 449.8 cm1 show a consistent intensity ratio of 6 : 1, and are shifted to 427.8 and 437.4 cm1 respectively, by 32 S/34 S substitution (c). In the mixed 32 S þ 34 S experiments [traces] (d) and (e), both bands split into asymmetric triplets marked in (e). These results suggest that two equivalent S atoms are involved in this species. The bands at 438.7 and 449.8 cm1 were assigned to the antisymmetric (n3) and symmetric (n1) stretching modes, respectively, of the nonlinear U 32 S2 molecule. Although the analogous UO2 molecule is linear, the US2 molecule is bent because of more favorable U(6d)–S(3p) overlap. The weak band at 451.1 cm1, which is partly overlapped by the n1 mode of US2 in (a), is shifted to 439.2 cm1 by 32 S/34 S substitution (c), and was assigned to the diatomic US molecule. Another weak band at 458.2 cm1 in (a) shows a

113

MATRIX COCONDENSATION REACTIONS

US2 (v3) US (e) 1.0

Absorbance

US2 (v1) (d) U34S U34S2(v1)

US3

0.6

U34S2(v3) (c)

US2 (v1) 0.2

US2 (v3) US3

470

(b)

460

US 450 440 Wavenumbers (cm-1)

(a) 430

420

Fig. 1.37. Infrared spectra of laser-ablated U co-deposited with discharged isotopic S atoms in Ar at 7 K: (a) 32S with higher concentration, (b) 32S with lower concentration, (c) 34S, (d) 35/65 32 S þ 34S mixture, and (e) 50/50 32S þ 34S mixture [129].

triplet pattern in the mixed-isotope experiment shown in (d), and was assigned to the antisymmetric stretching vibration of the near-linear SUS unit of a T-shaped US3 molecule similar to UO3 All these structures and band assignments are supported by DFT calculations. 2. Atom–Molecule Reaction. Zhou and Andrews [130] studied the reaction of Sc þ CO ! Sc(CO)n(n ¼ 1–4) via cocondensation of laser-ablated Sc atoms with CO in excess Ar or Ne, and made band assignments based on the results of 13 CO and C18O isotope shifts, matrix annealing, broadband photolysis, and DFT calculations. Figure 1.38a shows the IR spectrum of the cocondensation product after one hour deposition at 10 K in Ar. The strong and sharp band at 1834.2 cm1 is shifted to 1792.9 and 1792.3 cm1 by 13 CO and C18O substitutions, respectively, and its intensity decreases upon annealng to 25 K (b) but increases on photolysis (c,d). Two weak bands near 1920 cm1 in (a) are also isotope-sensitive and become stronger on annealing to 25 K (b). These bands were assigned to the Sc(CO)þ cation produced by the reaction of laser-ablated Scþ ion with CO. In a Ne matrix, a similar experiment exhibits a sharp band at 1732.0 cm1 that was assigned to the Sc(CO) ion produced via electron capture by Sc(CO). The pair of bands at 1851.4 and 1716.3 cm1 in (b) were assigned to the antisymmetric and symmetric stretching vibrations of the bent Sc(CO)2 molecule while the strong and sharp band at 1775.5 cm1 in (c) was attributed to the linear Sc(CO)2 molecule. DFT calculations predict Sc(CO)3 to be trigonal pyramidal (C3v symmetry) in the ground state. The bands at 1968.0 and 1822.2 cm1 in (b) become weaker by photolysis (c,d) and stronger by annealing at 35 K (e), These bands were assigned to the nondegenerate symmetric and degenerate CO stretching

114

THEORY OF NORMAL VIBRATIONS

0.065

Sc(CO)3 0.060

Sc(CO)4

Sc(CO)2

(e)

0.055 0.050 0.045

Absorbance

(d) 0.040 0.035

Sc(CO)2

0.030

(c)

0.025

ScCO4

0.020 0.015

Sc(CO)2

Sc(CO)3

0.010

Sc(CO)3

Sc(CO)2 (b)

ScCO

0.005

(a)

0.000 1950

1900

1850 Wavenumber (cm-1)

1800

1750

Fig. 1.38. Infrared spectra of laser-ablated Sc atoms co-deposited with 0.5 % CO in excess Ar.: (a) after 1 hr sample deposition at 10 K, (b) after annealing to 25 K, (c) after 20 min l > 470 nm photolysis, (d) after 20 min full-arc photolysis, and (e) after annealing to 35 K [130].

vibrations, respectively, of the Sc(CO)3 molecule. The band at 1865.4 cm1 appeared last when the matrix was annealed to 35 K (e), and was tentatively assigned to a degenerate vibration of Sc(CO)4. Matrix cocondensation reactions of this type have been studied most extensively as shown in Part B, Sec. 3.18.6. 3. Salt–Molecule Reaction. Tevault and Nakamoto [131] carried out matrix cocondensation reactions of metal salts such as PbF2 with L(CO, NO and N2) in Ar, and confirmed the formation of PbF2–L adducts by observing the shifts of IR bands of both components. These spectra are shown in Fig. 3.67 (of Sec. 3.18.6). 4. Salt–Salt Reaction. Ault [132] obtained the IR spectrum of the PbF 3 ion via the reaction of the type, MF þ PbF2 ! M[PbF3] (M ¼ Na, K, Rb, and Cs). Both salts were premixed, and the mixture was heated in a single Knudsen cell at 300–325 C, which is roughly 200 C below that is needed to vaporize either PbF2 or MF. Figure 1.39 shows the IR spectra of PbF2 and its cocondensation products with MF in Ar matrices. It is seen that, except for M ¼ Na, the cocondensation products exhibit two prominent bands near 456 and 405 cm1 with no bands due to PbF2 Thus, these two bands were assigned to the A1 and E stretching vibrations, respectively, of the PbF3 ion of C3v symmetry. The band at 242 cm1 was observed weakly only in high-yield experiments, and tentatively assigned to a bending mode of the PbF3 ion. 5. Molecule–Molecule Reaction. The reaction Fe(TPP) þ O2 ! (TPP) Fe-O2 (TPP tetraphenylporphyrin) was carried out by Watanabe et al. [133] using 16 O2 and 18 O2

SYMMETRY IN CRYSTALS

115

PbF2

NaF/PbF2

ABSORPTION

KF/PbF2

RbF/PbF2

CsF/PbF2

560

520

480

440

400

360

320

ENERGY (cm–1) Fig. 1.39. Infrared spectra of the M þ PbF 3 isolated in Ar matrices. The top trace shows the spectrum of the parent PbF2 salt, while the next four spectra show the bands attributable to the vaporization products from the solid mixtures of NaF/PbF2, KF/PbF2, RbF/PbF2, and CsF/PbF2 in the 580 300 cm 1 region [132].

diluted in Ar. The IR spectra shown in Fig. 3.76 (in Sec. 3.21.4) indicate that two new bands appear at 1195 and 1108 cm1 which are due to the O2 stretching vibrations of the end-on and side-on type Fe(TPP)O2, respectively. Bajdor and Nakamoto [134] obtained a novel oxoferrylporphyrin [O¼Fe(IV)(TPP)] via laser photolysis of Fe (TPP)O2 in pure O2 matrices at 15 K. As seen in Fig. 1.82 ( Sec. 1.22.4 of Part B), it exhibits the Fe(IV)¼O stretching vibration at 852 cm1. 1.27. SYMMETRY IN CRYSTALS The symmetry elements and point groups of molecules and ions in the free state have been discussed in Sec. 1.5. For molecules and ions in crystals, however, it is necessary to consider some additional symmetry operations that characterize translational symmetries in the lattice. Addition of these translational operations results in the formation of the space groups that can be used to classify the symmetry of molecules and ions in crystals.

116

THEORY OF NORMAL VIBRATIONS

TABLE 1.17. Seven Crystallographic Systems and 32 Crystallographic Point Groups System

32 Crystallographic Point Groups

Axes and Angles

Triclinic

a a

b b

c g

I I C1 Ci

Monoclinic

a 90

b b

c 90

2/m 2 m or 2 C2 Cs C2h

Orthorthombic

a 90

b c 90 90

Tetragonal

a 90

b c 90 90

Trigonal aaa or (rhombohedral) aaaa

a 90

222 mm2 mmm D2 C2n D2h 4 4 C4 S 4 a c 3 3 90 120 C3 C3i

4/m 422 4mm 42m C4h D4 C4v D2d 32 D3

3m C3v

4/mmm D4h

3m D3d

Hexagonal

a 90

a c 90 120

6 6 C6 C3h

6/m 622 6mm 6m2 C6h D6 C6v D3h

Cubic

a 90

a c 90 90

23 m3 T Th

432 43m O Td

6/mmm D6h

m3m Oh

1.27.1. 32 Crystallographic Point Groups As discussed in Sec. 1.5, molecular symmetry can be described by point groups that are derived from the combinations of symmetry operations, I (identity), Cp (p-fold axis of symmetry), a (plane of symmetry), i (center of symmetry), and Sp (p-fold rotation–reflection axis), where p may range from 1 to ¥. In the case of crystal symmetry, p is limited to 1, 2, 3, 4, and 6 because of space-filling requirement in a crystal lattice. As a result, only the 32 crystallographic point groups listed in Table 1.17 are possible. Here, Hermann–Mauguin (HM) as well as Sch€onflies (S) notations are shown. In the HM notation, 1, 2, 3, 4, and 6 denote the principal axis of rotation, Cp, where p is 1, 2, 3, 4, and 6, respectively, and 1; 2; 3; 4, and 6 denote a combination of respective rotation about an axis and inversion through a point on the axis. Thus, 1 means the center of inversion (i), 2 indicates s on a mirror plane (m), and 3; 4; and 6 correspond to S6, S4, and S3, respectively. For example, the point group C2h, (I, C2, a, and i) can be written as 2m 1, which can be simplified to 2/m where the slash means the C2 axis is perpendicular to m. As shown in Table 1.17, these 32 crystallographic point groups which describe the symmetry of the unit cell can be classified into seven crystallographic systems according to the most simple choice of reference axes. 1.27.2. Symmetry of the Crystal Lattice A crystal has a lattice structure that is a repetition of identical units. Bravais has shown that only the 14 different types of lattices shown in Fig. 1.40 are possible. These

SYMMETRY IN CRYSTALS

117

Fig. 1.40. The 14 Bravais lattices belonging to the seven crystallographic systems [135].

Bravais lattices are also classified into the seven crystallographic systems mentioned above. Here, the lattice points are regarded as “points” or “small, indentical, perfect spheres,” thus representing the highest symmetry possible. In actual crystals, these points are occupied by unsymmetric but identical molecules or ions so that their high symmetry is lowered. In Fig. 1.40 P, C (or A or B), F, I, and R indicate primitive, basecentered, face-centered, body-centered, and rhombohedral lattices, respectively. The number of lattice points (LP) in each cell is summarized in Table 1.18. For crystal structures designated by P, crystallographic unit cells are identical with the Bravais unit cells. For those designated by other letters, crystallographic unit cells contain 2, 3, or 4 Bravais cells. In these cases, the number of molecules in the crystallographic unit

118

THEORY OF NORMAL VIBRATIONS

TABLE 1.18. Primitive and Centered Lattices Type of Crystal Structure Primitive (P) Base-centered (A, B, or C) Body-centered (I) Rhombohedral (R) Face-centered (F) a

Number of Lattice Points (LP) 1 2 2 3 or 1a 4

The crystallographic unit cell may have been divided by 3.

cell must be divided by the corresponding LP since only the consideration of one Bravais cell is sufficient to interpret the vibrational spectra of crystals. 1.27.3. Space Groups As stated above, the 32 point groups describe the symmetry of the unit cells, and the 14 Bravais lattices describe all possible arrangements of crystal lattices. To describe space symmetry, however, symmetry operations mentioned earlier are not enough. We must add three new symmetry operations: Translation. A translational operation along one lattice direction gives an identical point. Screw Axis. Rotation of a lattice point about an axis followed by translation gives an identical point. Namely, rotation by 2p/p followed by a translation of q=pðq ¼ 1; 2; . . . ; n 1Þ in the direction of the axis is written as pq. For example, 21 indicates twofold rotation followed by a translation of 12 a, 12 b, or 12 c, as shown in Fig. 1.41A. If the axis is threefold, the translation must be 13 or 23, and the symbols are 31 and 32, respectively. Glide Plane. Reflection in a plane followed by translation by 12 a, 12 b, 12 c, as shown in Fig. 1.41B. These planes are indicated by a, b, and c, respectively, n is used for diagonal translation of (b þ c)/2, (c þ a)/2, or (a þ b)/2.

Fig. 1.41. Screw axis and glide plane.

VIBRATIONAL ANALYSIS OF CRYSTALS

119

Addition of these symmetry operations results in 230 different combinations which are called “space groups.” Table 1.19 shows the distribution of space groups among the seven crystal systems [136]. Some of these space groups are rarely found in actual crystals. About half of the crystals belong to 13 space groups of the monoclinic system. Space groups can be described by using either Hermann–Mauguin (HM) or Sch€onflies 1 (S) notations. For example, P2/mð C2h Þ indicates a primitive lattice with m perpendicular to 2. The first symbol (capital letter) refers to the Bravais lattice (P, A, B, C, F, I, and R), and the nextsymbol indicates fundamental symmetry elements that lie along the special direction in the crystal. In the case of the monoclinic system, it is the twofold 2 axis. Thus, P21/mð C2h Þ is a primitive cell with m perpendicular to 21. These symbols indicate only the minimum symmetry elements that are necessary to distinguish the space groups. The differences among space groups can be found by comparing the “unit cell maps” given in the International Tables for X-Ray Crystallography [137] . 1.28. VIBRATIONAL ANALYSIS OF CRYSTALS [48–52] Becauseofintermolecularinteractions,thesymmetryofamoleculeis generally lowerin the crystalline state than in the gaseous (isolated) state. This change in symmetry may split the degenerate vibrations and activate infrared- (or Raman-) inactive vibrations. Additionally, intermolecular (interionic) interactions in the crystal lattice cause frequency shifts relative to the gaseous state. Finally, the spectra obtained in the crystalline state exhibit lattice modes—vibrations due to translatory and rotatory motions of a molecule in the crystalline lattice. Although their frequencies are usually lower than 300 cm1, they may appear in the high-frequency region as the combination bands with internal modes (see Fig. 2.18, for example). Thus, the vibrational spectra of crystals must be interpreted with caution, especially in the low-frequency region. To analyze the spectra of crystals, it is necessary to carry out a site group or factor group analysis, as described in the following subsection. 1.28.1. Subgroups and Correlation Tables For any point group, there are “subgroups” which consist of some, but not all, symmetry elements of the “parent group.” For example, the character table of the point group, C3v (Table 1.6) contains I, C3, and sv as the symmetry elements. Then, the subgroups of C3v are C3, consisting of I and C3, and Cs, consisting of I and sv only. The relationship between the symmetry species in the “parent group” and those in “subgroups” is given in a “correlation table.” Such correlation tables have already been worked out for all common point groups and are listed in Appendix IX. As an example, the correlation table for C3v is shown below: C3v

C3

Cs

A1 A2 E

A A E

A 00 A 0 00 A þA

0

120

HMb

1 1

2 m 2/m

222

mm2

mmm

4 4 4/m 422

4mm

42m

4/mmm

C1 Ci

C2(1–3) CS(1–4) C2h(1–6)

D2(1–9)

C2v(1–22)

D2h(1–28)

C4(1–6) S4(1–2) C4h(1–6) D4(1–10)

C4v(1–12)

D2d(1–12)

D4h(1–20)

Triclinic

Monoclinic

Orthorhombic

Tetragonal

System

Sa

Point Groups

TABLE 1.19. The 230 Space Groups

Pnnn Pbam Pnma Fmmm P41 I4 P42/m P4212 I422 P4bm I4mm P42c I4m2

P4 P4 P4/m P422 P43212 P4mm P42bc P42m P4n2 P4/mcc P42/mmc P42/ncm

P2221 I21212 Pmc21 Pna21 Ama2

P222 I222 Pmm2 Pba2 Abm2 Ima2 Pmmm Pcca Pbca Ccca

P4/mmm P4/ncc P42/ncm

P21 Pc P21/m

P2 Pm P2/m

P1 P1

P42/n P41212

P4/n P4122 I4122 P42cm I4cm 1m P42 I4c2 P4/nbm P42/mcm I4/mmm

P42nm I41md 1c P42 I42m P4/nnc P42/nbc I4/mcm

P43

Pban Pbcm Cmca Immm

Pma2 Cmm2 Fmm2

P212121

Cc P2/c

P42

Pccm Pccn Cmcm Fddd

Pcc2 Pnm2 Aba2

P21212

C2 Cm C2/m

Space Group

P21/c

P4/mbm P42/nnm I41/amd

P4cc I41cd P4m2 I42d

I4/m P4222

I4

Pmma Pnnm Cmmm Ibam

Pca21 Cmc21 Fdd2

C2221

P4/nmm P42/mnm

P4b2

P4c2 P4/mnc P42/mbc I41/acd

P42mc

P4322

Pmna Pbcn Cmma Imma

Pmn21 Amm2 Iba2

F222

P4nc

I41/a P42212

I41

Pnna Pmmn Cccm Ibca

Pnc2 Ccc2 Imm2

C222

C2/c

121

€nflies. Scho Hermann–Mauguin.

Source: Ref. [136].

b

a

F23 Pn3 P4232

23 m3 432

43m m3m

Td(1–6) Oh(1–10)

F43m Pn3n Im3m

P6/mcc

P6/mmm P23 Pm3 P432 I4132 P43m Pm3m Fd3c

6/mmm

T (1–5) Th(1–7) O(1–8) I43m Pm3n Ia3d

I23 Fm3 F432

P63/mcm

P6522 P63cm P62m

P63/m P6122 P6cc P6c2

Cubic

P65

P61

P6 P6 P6/m P622 P6mm P6m2

6 6 6/m 622 6mm 6m2

Ph3m

P43n

P213 Fd3 F4132

P63/mmc

P6222 P63mc P62c

P62

P3121 P31c P3c1

P3112 P3c1 P3m1

C6(1–6) C3h(1) C6h(1–2) D6(1–6) C6v(1–4) D3h(1–4) D6h(1–4)

Hexagonal

R3

Space Group P32

P31 R3 P321 P31m P31c

C3(1–4) C3i(1–2) D3(1–7) C3v(1–6) D3d(1–6)

Trigonal

P3 P3 P312 P3m1 P31m

3 3 32 3m 3m

S

HMb

Point Groups

System

a

F43c Fm3m

I213 Im3 I432

P6422

P64

P3212 R3m R3m

I43d Fm3c

Pa3 P4332

P6322

P63

P3221 R3c R3c

Fd3m

Ia3 P4132

R32

122

THEORY OF NORMAL VIBRATIONS

A pyramidal XY3 molecule such as NH3 belongs to the point group C3v. If one of the Yatoms is replaced by a Z atom, the resulting XY2Z molecule belongs to the point group Cs. As a result, the doubly degenerate vibration (E) splits into two vibrations (A0 þ A00 ). These correlation tables are highly important in predicting the effect of lowering of symmetry on molecular vibrations. Chapter 2 includes a number of examples of symmetry lowering by substitution of an atomic group by another group. 1.28.2. Site Group Analysis According to Halford [138], the vibrations of a molecule in the crystalline state are governed by a new selection rule derived from site symmetry—a local symmetry around the center of gravity of a molecule in a unit cell. The site symmetry can be found by using the following two conditions: (1) the site group must be a subgroup of both the space group of the crystal and the molecular point group of the isolated molecule, and (2) the number of equivalent sites must be equal to the number of molecules in the unit cell. Halford derived a complete table that lists possible site symmetries and the number of equivalent sites for 230 space groups. Suppose that the space group of the crystal, the number of molecules in the unit cell (Z), and the point group of the isolated molecule are known. Then, the site symmetry can be found from the Table of Site Symmetry for the 230 Space Groups, which was originally derived by Halford. Appendix X gives its modified version by Ferraro and Ziomek [9]. In general, the site symmetry is lower than the molecular symmetry in an isolated state. In some cases, it may be difficult to make an unambiguous choice of site symmetry by the method cited above. Then, Wyckoff ’s tables on crystallographic data [139] must be consulted. The vibrational spectra of calcite and aragonite crystals are markedly different, although both have the same composition (Sec. 2.4). This result can be explained if we consider the difference in site symmetry of the CO2 3 ion between these crystals. According to X-ray analysis, the space group of calcite is D63d and Z is 2 Appendix X gives D3 ð2Þ;

C3i ð2Þ;

C3 ð4Þ;

Ci ð6Þ;

C2 ð6Þ;

C1 ð12Þ

as possible site symmetries for space group D63d (the number in front of point group notation indicates the number of distinct sets of sites, and that in parentheses denotes the number of equivalent sites for each distinct set). Rule 2 eliminates all but D3(2) and C3i(2). Rule 1 eliminates the latter since C3i is not a subgroup of D3h. Thus, the site symmetry of the CO23 ion in calcite must be D3. On the other hand, the space group of aragonite is D16 2h and Z is 4. Appendix X gives 2Ci ð4Þ;

Cs ð4Þ;

C1 ð8Þ

Since Ci is not a subgroup of D3h, the site symmetry of the CO23 ion in aragonite must be Cs. Thus, the D3h, symmetry of the CO23 ion in an isolated state is lowered to D3 in

VIBRATIONAL ANALYSIS OF CRYSTALS

123

TABLE 1.20. Correlation Table for D3h, D3, C2v, and Cs n1

Point Group D3h D3 C2v Cs

0

A 1 ðRÞ A1(R) 0 A 1 ðI; RÞ A0 (I, R)

n2 00

A 2 ðIÞ A2(I) B1 (I, R) A00 (I, R)

n3 0

E (I, R) E (I, R) A1 (I, R) þ B2 (I, R) A0 (I, R) þ A0 (I, R)

n4 0

E (I, R) E (I, R) A1 (I, R) þ B2 (I, R) A0 (I, R) þ A0 (I, R)

calcite and to Cs in aragonite. Then, the selection rules are changed as shown in Table 1.20. There is no change in the selection rule in going from the free CO23 ion to calcite. In aragonite, however, n1 becomes infrared active, and n3 and n4 each split into two bands. The observed spectra of calcite and aragonite are in good agreement with these predictions (see Table 2.4b). 1.28.3. Factor Group Analysis A more complete analysis including lattice modes can be made by the method of factor group analysis developed by Bhagavantam and Venkatarayudu [140]. In this method, we consider all the normal vibrations for an entire Bravais cell. Figure 1.42 illustrates the Bravais cell of calcite, which consists of the following symmetry elements:

Fig. 1.42. The Bravais cell of calcite.

124

THEORY OF NORMAL VIBRATIONS

I, 2S6, 2S262C3 , S36 i; 3C2 , and 3sv (glide plane). These elements are exactly the same as those of the point group D3d, although the last element is a glide plane rather than a plane of symmetry in a single molecule. As mentioned in Sec. 1.27, it is possible to derive the 230 space groups by combining operations possessed by the 32 crystallographic point groups with operations such as pure translation, screw rotation (translation þ rotation), and glide plane reflection (translation þ reflection). If we regard the translations that carry a point in a unit cell into the equivalent point in another cell as identity, we define the 230 factor groups that are the subgroups of the corresponding space groups. In the case of calcite, the factor group consists of the symmetry elements described above, and is denoted by the same notation as that used for the space group ðD63d Þ. The site group discussed previously is a subgroup of a factor group. Since the Bravais cell contains 10 atoms, it has 3 10 – 3 ¼ 27 normal vibrations, excluding three translational motions of the cell as a whole.* These 27 vibrations can be classified into various symmetry species of the factor group D63d , using a procedure similar to that described in Sec. 1.8 for internal vibrations. First, we calculate the characters of representations corresponding to the entire freedom possessed by the Bravais cell [cR(N)], translational motions of the whole cell [cR(T)], translatory lattice modes [cR(T 0 )] rotatory lattice modes [cR(R0 )], and internal modes [cR(n)], using the equations given in Table 1.21. Then, each of these characters is resolved into the symmetry species of the point group, D3d. The final results show that three internal modes (A2u and two Eu), three translatory modes (A2u and two Eu), and two rotatory modes (A2u and Eu) are infrared-active, and three internal modes (A1g and two Eg) one translatory mode (Eg), and one rotatory mode (Eg) are Raman-active. As will be shown in the following sections, these predictions are in perfect agreement with the observed spectra. 1.29. THE CORRELATION METHOD In the preceding section, we described the application of factor group analysis to the calcite crystal. However, the correlation method developed largely by Fateley et al. [15] is simpler and gives the same results. In this method, intramolecular and lattice vibrations are classified under point groups of molecular symmetry, site symmetry, and factor group symmetry, and correlations are made using the correlation tables given in Appendix IX. In the following, we demonstrate its utility using calcite as an example. For more detailed discussions and applications, the reader should consult the books by Fateley et al. [15] and Ferraro and Ziomek [9]. 1.29.1. The CO23 Ion in the Free State As mentioned in Sec. 1.4, normal vibrations of a molecule can be described in terms of translational motions of the individual atoms along the x, y, and z axes. Thus, the *These three motions give acoustical modes (Sec. 1.30).

125

Key: p, total number of atoms in the Bravais cell. s, total number of molecules (ions) in the Bravais cell. n, total number of monoatomic molecules (ions) in the Bravais cell. NR(p), number of atoms unchanged by symmetry operation R. NR(s), number of molecules (ions) whose center of gravity is unchanged by symmetry operation R. NR(s n), NR(s) minus number of monoatomic molecules (ions) unchanged by symmetry operation R. cR(N) ¼ NR (p)[ (1 þ 2 cos y)], character of representation for entire freedom possessed by the Bravias cell. cR(T) ¼ (1 þ 2 cos y), character of representation for translation motions of the whole Bravais cell. cR(T0 ) ¼ {NR (s) 1} { (1 þ 2 cos y)} character of representation for translatory lattice modes. cR(R0 ) ¼ NR (s n) { (1 þ 2 cos y)}, character of representation for rotary lattice modes. cR(n) ¼ cR(N) cR(T) cR(T0 ) cR(R0 ), character of representation for internal modes. Note that þ and – signs are for proper and improper rotations, respectively. The symbol y should be taken as defined in Sec. 1.8.

TABLE 1.21. Factor Group Analysis of Calcite Crystal

126

THEORY OF NORMAL VIBRATIONS

Fig. 1.43. The x, y, and z axes chosen for planar CO32 ion.

normal modes of an N-atom molecule can be expressed by using 3N translational motions. Furthermore, the symmetry species of these normal modes must correlate with those of the translational motions of an atom located at a particular site within the molecule. Since the latter are known from the molecular structure, the former may be determined directly by using the correlation table. As an example, consider a planar CO23 ion for which the x, y, and z axes are chosen as shown in Fig. 1.43. It is readily seen that the C atom is situated at the D3h site, whereas the O atom is situated at a site where the local symmetry is only C2v (I, C2, sh, 00 and sv). The translational motions of the C atom under D3h symmetry are: TzðA 2 Þand (Tx, Ty)(E0 ). The translational motions of the O atom under C2v symmetry are: Tx(A1), Ty(B2) and Tz(B1). In Table 1.22, the symmetry species of these translational motions are connected by arrows to those of the whole ion using the correlation table. Then, the number of times a particular symmetry species occurs in the total representation is given by the number of arrows that terminate on that species. The number of normal vibrations of the CO23 ion in each symmetry species is given by subtracting those of the translational and rotational motions of the whole ion from the total representation. The results are shown at the bottom of Table 1.22. 0 00 Thus, the CO23 ion exhibits four normal vibrations; n1 (A 1 Raman-active), n2 (A 2 0 IR-active), n3(E0 , IR- and Raman-active), and n4 (E , IR- and Raman-active). The normal modes of these four vibrations are shown in Fig. 2.8. In Sec. 1.8, we derived the general method to classify normal vibrations into symmetry species based on group theory. As seen above, this procedure is greatly simplified if we use the correlation method. 1.29.2. Intramolecular Vibrations in the CO23 Ion in Calcite Applying the same principle as that used above, we can classify the normal vibrations of the CO23 ion in the calcite lattice by using the correlation method. As discussed in

THE CORRELATION METHOD

127

TABLE 1.22. Correlation Method for the CO2 3 Ion in the Free State

Sec. 1.28, calcite belongs to the space group D63d , and Z is 2. From the Bravais cell shown in Fig. 1.42, it is readily seen that the CO23 ion is at the D3 site, whereas the Ca2þ ion is at the C3i (S6) site. Table 1.23 shows the correlations among the molecular symmetry (D3h), site symmetry (D3), and factor group symmetry (D3d). Under the “Vib.” column, we list the number of vibrations belonging to each species of D3h symmetry (Table 1.22). Throughout the present correlations, the number of degrees of vibrational freedom for all the doubly degenerate species must be multiplied by 2 since there are two E0 modes (n3 and v4). Thus, it is 4 for the E0 species. The number under “f’” indicates “Vib.” times Z (¼2), which is the number of vibrations of the unit cell. In going from D3h to D3, no changes occur TABLE 1.23. Correlation between Molecular Symmetry, Site Symmetry, and Factor Group Symmetry for Intramolecular Vibrations of CO2 3 Ion in Calcite

128

THEORY OF NORMAL VIBRATIONS

TABLE 1.24. Correlation between Site Symmetry and Factor Group Symmetry for Lattice Vibrations of CO2 3 Ion in Calcite

except for changes in notations of symmetry species. Column C indicates the degeneracy, and column am shows the degree of vibrational freedom contributed by the corresponding molecular symmetry species. Finally, the species under D3 symmetry are connected to those of the factor group by the arrows using the correlation table. The last column, “a”, is the degree of vibrational freedom contributed by the corresponding site symmetry species to the factor group species. It should be noted that the number of degrees of vibrational freedom must be 12 throughout the described correlations above. Such bookkeeping must be carried out for every correlation. 1.29.3. Lattice Vibrations in Calcite Table 1.24 shows the correlation diagram for lattice vibrations of the CO23 ion. The variables “t” and “f” denote the degrees of translational freedom of the CO23 ion for each ion and for the Bravais cell, respectively. The same result is obtained for the rotatory lattice vibrations. Table 1.25 shows the correlation diagram for translatory TABLE 1.25. Correlation between Site Symmetry and Factory Group Symmetry for Lattice Vibrations of the Ca2þ Ion in Calcite

129

LATTICE VIBRATIONS

TABLE 1.26. Distribution of Normal Vibrations of Calcite as Obtained by Correlation Method Symmetry Species of Factor Group (D3d) A1g (Raman) A1u A2g A2u (IR) Eg (Raman) Eu (IR) Total

Translatory Lattice

Acoustical

Rotatory Lattice

Intramolecular

0 1 1 1 1 2

0 0 0 1 0 1

0 0 1 1 1 1

1 1 1 1 2 2

9

3

6

12

lattice vibrations of the Ca2þ ion. No rotatory lattice vibrations exist for single-atom ions such as the Ca2þ ion. The distribution of all the translatory lattice vibrations can be obtained by subtracting c (acoustical) from c (trans, CO23 ) þ c (trans, Ca2þ). 1.29.4. Summary Table 1.26 summarizes the results obtained in Tables 1.23–1.25. The total number of vibrations including the acoustical modes should be 30 since the Bravais cell contains 10 atoms. These results are in complete agreement with that obtained by factor group analysis (Table 1.21). 1.30. LATTICE VIBRATIONS [48] Consider a lattice consisting of two alternate layers of atoms; atoms 1 of mass M1 lie on one set of planes and atoms 2 of mass M2 lie on another set of planes, as shown in Fig. 1.44. For example, such an arrangement is found in the 111 plane of the TABLE 1.27. Correlation between Site Symmetry (C2v) and Factor Group Symmetry (D2h) for Orthorhombic form of the 123 Conductora

a

For f, t, and a, see Sec. 1.29. The complete correlation table is found in Appendix IX.

130

THEORY OF NORMAL VIBRATIONS

Fig. 1.44. Displacements of atoms 1 and 2 in a diatomic lattice.

NaCl crystal. Let a denote the distance between atoms 1 and 2. Then the length of the primitive cell is 2a. We consider waves that propagate in the direction shown by the arrow, and assume that each plane interacts only with its neighboring planes. If the force constants (f) are identical between all these neighboring planes, we have d 2 us M1 2 ¼ f ðvs þ vs 1 2us Þ ð1:233Þ dt M2

d 2 vs ¼ f ðus þ 1 þ us 2vs Þ dt2

ð1:234Þ

where us and vs denote the displacements of atoms 1 and 2 in the cell indexed by s. The solutions for these equations take the form of a traveling wave having different amplitudes u and v. Thus, we obtain us ¼ u exp½iðwt þ 2skaÞ vs ¼ v exp½iðwt þ 2skaÞ

ð1:235Þ ð1:236Þ

Here w is angular frequency, 2pn [in reciprocal seconds (s1)]. If these are substituted in Eqs. 1.233 and 1.234, respectively, we obtain ½M1 w2 2f u þ f ½1 þ expð 2ikaÞv ¼ 0

ð1:237Þ

f ½1 þ expð2ikaÞu þ ½M2 w2 2f v ¼ 0

ð1:238Þ

These homogeneous linear equations have a nontrivial solution if the following determinant is zero: M1 w2 2f f ½1 þ expð 2ikaÞ ð1:239Þ ¼0 f ½1 þ expð2ikaÞ M2 w2 2f By solving this equation, we obtain ! " !2 #1=2 1 1 1 1 4sin2 ka w ¼f þ þ þf M1 M2 M1 M2 M1 M2 ðoptical branchÞ 2

ð1:240Þ

LATTICE VIBRATIONS

1 M2

1/2

(M2f ( 2f D M

(

+

(

1 M1

Optical

B

(

(

2f

2

E

131

1/2

C

M1 > M2

1

ω

1/2

F A 0

Acoustical k

π/2a

Fig. 1.45. Dispersion curves for lattice vibrations.

! " !2 #1=2 1 1 1 1 4sin2 ka w ¼f þ þ f M1 M2 M1 M2 M1 M2 ðacoustical branchÞ 2

ð1:241Þ

The term k in Eqs. 1.235–1.241 is called the wavevector and indicates the phase difference between equivalent atoms in each unit cell. In the case of a one-dimensional lattice, |k| ¼ k. Thus, we use k rather than |k| in this case, and k can take any value between p/2a and þ p/2a. This region is called the first Brillouin zone. Figure 1.45 shows a plot of w versus k for the positive half of the first Brillouin zone. There are twovalues for each w that constitute the “optical” and “acoustical” branches in the dispersion curves. At the center of the first Brillouin zone (k ¼ 0), we have w ¼ 0 and

and

u M2 ¼ v M1

u¼v

ðacousticalÞ vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ! u u 1 1 t w ¼ 2f þ M1 M2

or M1 u þ M2 v ¼ 0

ð1:242Þ

ð1:243Þ

ðopticalÞ

At the end of the first Brillouin zone (k ¼ p/2a), we have rﬃﬃﬃﬃﬃﬃ 2f and u ¼ 0 ðopticalÞ w¼ M2 rﬃﬃﬃﬃﬃﬃ 2f w¼ and v ¼ 0 ðacousticalÞ M1

ð1:244Þ ð1:245Þ

points A, B, C, and D in Fig. 1.45 correspond to Eqs. 1.242, 1.243, 1.244, and 1.245, respectively.

132

THEORY OF NORMAL VIBRATIONS

To observe lattice vibrations in IR spectra, the momentum of the IR photon must be equal to that of the phonon.* The momentum of the photon (P) is given by P¼

h h=2p ¼ ¼ hQ l l=2p

ð1:246Þ

where Q ¼ 2p/l. On the other hand, the momentum of the phonon is given byhk [141]. Thus, the following relationship must hold: hQ ¼ hk

ð1:247Þ

Since lattice vibrations are observed in the low-frequency region (l ﬃ 103 cm), Q ¼ 2p/l ﬃ 103 cm. The k value at the end of the Brillouin zone is k ¼ p/2a ﬃ 108 cm1. Thus, the k value that corresponds to the IR photon for lattice vibration is much smaller than the k value at the end of the Brillouin zone. This result indicates that the optical transitions we observe in IR spectra occur practically at k ﬃ 0. A similar conclusion can be derived for Raman spectra of lattice vibrations. Figure 1.46 shows the modes of lattice vibrations corresponding to various points on the dispersion curve shown in Fig. 1.45. At point A (acoustical mode), all atoms move in the same direction (translational motion of the whole lattice) and its frequency is zero. This is seen in Fig. 1.46a. In a three-dimensional lattice, there are three such modes. Thus, we subtract 3 from our calculations in factor group analysis (Sec. 1.28). On the other hand, at point B (optical branch), the two atoms move in opposite directions, but the center of gravity of the unit cell remains unshifted (Fig. 1.46b). Furthermore, the equivalent atom in each lattice moves in phase. If the two atoms carry opposite electrical charges, such a motion produces an oscillating dipole moment that can interact with incident IR radiation. Thus, it is possible to observe it optically. It should molecule pﬃﬃﬃﬃﬃﬃﬃﬃ be noted that the frequency of a diatomic p ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ in the free state is w ¼ f =m, whereas that of a diatomic lattice is w ¼ 2f =m (m ¼ reduced mass). At point C (optical branch), the lighter atoms are moving back and forth against each other while the heavier atoms are fixed (Fig. 1.46c). At point D (acoustical branch), the situation is opposite to that of point C. The vibrational modes in the middle of the Brillouin zone (points E and F) are shown in Figs. 1.46e and 1.46f, respectively. Thus far, we have discussed the lattice vibrations of a one-dimensional chain. The treatment of the three-dimensional lattice is basically the same, although more complicated [48]. If the primitive cell contains s molecules, each of which consists of N atoms, there are 3 acoustical modes and 3Ns – 3 optical modes. The latter is grouped into (3N – 6)s internal modes and 6s – 3 lattice modes. The general forms of

*The lattice vibration causes elastic waves in crystals. The quantum of die lattice vibrational energy is called “phonon,” in analogy with the photon of the electromagnetic wave.

POLARIZED SPECTRA OF SINGLE CRYSTALS

133

Fig. 1.46. Wave motions corresponding to various points on the dispersion curves.

the dispersion curves of such a crystal are shown in Fig. 1.47. In this book, our interest is focused on vibrational analysis of 3Ns–3 optical modes at k ﬃ 0. Examples are found in diamond and graphite (Sec. 2.14.3) and quartz (Sec. 2.15.2). 1.31. POLARIZED SPECTRA OF SINGLE CRYSTALS In the preceding section, the 30 normal vibrations of the Bravais cell of calcite crystal have been classified into symmetry species of the factor group D3d. The results (Table 1.26) show that three intramolecular (A2u þ 2Eu), three translatory lattice (A2u þ 2Eu) and two rotatory lattice (A2u þ Eu) vibrations are IR-active, whereas three intramolecular (A1g þ 2Eg), one translatory lattice (Eg), and one rotatory lattice (Eg) vibrations are Raman-active. In order to classify the observed bands into these symmetry species, it is desirable to measure infrared dichroism and polarized Raman spectra using single crystals of calcite.

134

THEORY OF NORMAL VIBRATIONS

Fig. 1.47. General form of dispersion curves for a molecular crystal [48].

1.31.1. Infrared Dichroism Suppose that we irradiate a single crystal of calcite with polarized infrared radiation whose electric vector vibrates along the c axis (z direction) in Fig. 1.42. Then the infrared spectrum shown by the solid curve of Fig. 1.48 is obtained [142]. According to Table 1.21, only the A2u vibrations are activated under such conditions. Thus, the three bands observed at 885(n), 357(t), and 106(r) cm1 are assigned to the A2u species. The spectrum shown by the dotted curve is obtained if the direction of polarization is perpendicular to the c axis (x,y plane). In this case, only the Eu vibrations should be infrared-active. Therefore the five bands observed at 1484(n), 706(n), 330(t), 182(t), and 106(r) cm1 are assigned to the Eu species. Here, n, t, and r denote intramolecular, translatory lattice, and rotatory lattice modes, respectively. 1.31.2. Polarized Raman Spectra Polarized Raman spectra provide more information about the symmetry properties of normal vibrations than do polarized infrared spectra [143]. Again consider a single

Fig. 1.48. Infrared dichroism of calcite [142].

POLARIZED SPECTRA OF SINGLE CRYSTALS

135

Fig. 1.49. Schematic representation of experimental condition y(zz)x.

crystal of calcite. According to Table 1.21, the A1g vibrations become Raman-active if any one of the polarizability components, axx, ayy, and azz, is changed. Suppose that we irradiate a calcite crystal from the y direction, using polarized radiation whose electric vector vibrates parallel to the z axis (see Fig. 1.49), and observe the Raman scattering in the x direction with its polarization in the z direction. This condition is abbreviated as y(zz)x. In this case, Eq. 1.66 is simplified to Pz ¼ azzEz because Ex ¼ Ey ¼ 0 and Px ¼ Py ¼ 0. Since azz belongs to the A1g species, only the A1g vibrations are observed under this condition. Figure 1.50c illustrates the Raman spectrum obtained with this condition. Thus, the strong Raman line at 1088 cm1 (n) is assigned to the A1g species. Both the A1g and Eg vibrations are observed if the z(xx)y condition is used. The Raman spectrum (Fig. 1.50a) shows that five Raman lines [1088(n), 714(n), 283(r), 156(t), and 1434(n) cm1 (not shown)] are observed under this condition. Since the 1088 cm1 line belongs to the A1g species, the remaining four must belong to the Eg species. These assignments can also be confirmed by measuring Raman spectra using the y(xy)x and x (zx)y conditions (Figs. 1.50b and 1.50d) . 1.31.3. Normal Coordinate Analysis on the Bravais Cell In the discussion above, we have assigned several bands in the same symmetry species to the v, t, and r types. In general, the intramolecular (v) vibrations appear above 400 cm1, whereas the lattice vibrations appear below 400 cm1. However, more complete assignments can be made only via normal coordinate analysis on the entire Bravais cell [144]. Such calculations have been made by Nakagawa and Walter [145] on crystals of alkali–metal nitrates that are isomorphous with calcite. These workers employed four intramolecular and seven intermolecular force constants. The latter are in the range of 0.12–0.00 mdyn/A. Figure 1.51 illustrates the vibrational modes of the 18 (27 if E modes are counted as 2) optically active vibrations together with the corresponding frequencies of calcite.

136

THEORY OF NORMAL VIBRATIONS

Fig. 1.50. Polarized Raman spectra of calcite [143].

1.32. VIBRATIONAL ANALYSIS OF CERAMIC SUPERCONDUCTORS In 1987, Wu et al. [146] synthesized a ceramic superconductor of the composition, YBa2Cu3O7d, whose superconducting critical temperature (Tc) was above the boiling point of liquid nitrogen (77 K). Since then, IR and Raman spectra of this and related compounds have been studied extensively, and the results are reviewed by Ferraro and Maroni [147,148]. Here, we limit our discussion to the Raman spectra of the superconductor mentioned above and their significance in studying oxygen deficiency and the structural changes resulting from it.

VIBRATIONAL ANALYSIS OF CERAMIC SUPERCONDUCTORS

137

Fig. 1.51. Vibrational modes of calcite. The observed and calculated (in parentheses) are listed under each mode [145].

138

THEORY OF NORMAL VIBRATIONS

Fig. 1.52. The Bravais unit cells of the orthorhombic and tetragonal forms of YBa2Cu3O7d [149].

The superconductor, YBa2Cu3O7d (abbreviated as the 123 conductor), can be obtained by baking a mixture of Y2O3, and BaCO3, and CuO in a proper ratio. The product is normally a mixture of an orthorhombic form (0 < d < 1) which is superconducting and a tetragonal form (d ¼ 1) that is an insulator. The Tc increases as d approaches 0. Figure 1.52 shows the Bravais unit cells of the orthorhombic (Pmmm ¼ D12h ) and tetragonal (P4/mmm ¼ D14h ) forms. The former is a distorted, oxygen-deficient form of perovskite. The orthorhombic unit cell contains 13 atoms, and their possible site symmetries can be found from the tables of site symmetries (Appendix X) as follows: 8D2h ð1Þ;

12C2v ð2Þ;

6Cs ð4Þ;

C1 ð8Þ

It is seen in Fig. 1.52 that the three atoms, Y, O(1), and Cu(1), are at the D2h sites. These atoms contribute 3B1u þ 3B2u þ 3B3u vibrations since Tx, Ty, and Tz belong to the B3u, B2u, and B1u species, respectively, in the D2h point group. On the other hand, the 10 atoms, 2Ba, 2Cu(2), 2O(2), 2O(3), and 2O(4) are at the C2v sites. As shown in Table 1.27, each pair of these atoms possess six degrees of vibrational freedom (2A1þ 2B1þ2B2), which are split into Ag þ B2g þ B3g þ B1u þ B2u þ B3u under D2h symmetry. The number of optical modes at k ﬃ 0 in each species can be obtained by subtracting three acoustical modes (B1u þ B2u þ B3u) from the above counting. Table 1.28 summarizes the results. It is seen that the orthorhombic unit cell has 15 Raman-active modes (5Ag þ 5B2g þ 5B3g) and 21 IR-active modes (7B1u þ 7B2u þ 7B3u). It should be noted that the mutual exclusion rule holds in this case since the D2h point group has a center of symmetry. Although the same results can be obtained by using factor group analysis [150], the correlation method is simpler and straightforward. Similar calculations on the tetragonal unit cell (YBa2Cu3O6) show that 10 vibrations (4A1g þ B1g þ 5Eg) are Raman-active while 11 vibrations (5A2u þ 6Eu) are IR-active under D4h symmetry.

139

VIBRATIONAL ANALYSIS OF CERAMIC SUPERCONDUCTORS

TABLE 1.28. Vibrational Analysis for Orthorhombic Form of YBa2Cu3O7d Using the Correlation Method D2h

D2h Y, O(1), Cu (1)

C2v Ba, Cu(2), O(2), O(3), O(4)

Acoustical Vib.

Total Optical Vib.

Activity

Ag B1g B2g B3g Au B1u B2u B3u

0 0 0 0 0 3 3 3

5 0 5 5 0 5 5 5

0 0 0 0 0 1 1 1

5 0 5 5 0 7 7 7

Raman Raman Raman Raman Inactive IR IR IR

9

30

3

36

Total

Figure 1.53 shows the Raman spectrum of a polished sintered pellet of the 123 conductor (d ¼ 0.3) obtained by Ferraro et al. [151]. The five Ag modes appear strongly, and the most probable assignments for these bands are [152,153]. 492 cm1 445 cm1 336 cm1 145 cm1 116 cm1

Axial motion of the O(4) atoms O(2)Cu(2)O(3) bending with the two O atoms moving in phase O(2)Cu(2)O(3) bending with the two O atoms moving out of phase Axial motion of the Cu(2) atoms Axial motion of the Ba atoms

Fig. 1.53. Backscattered Raman spectrum of a polished sintered pellet of YBa2Cu3O7d [151].

140

THEORY OF NORMAL VIBRATIONS

These values fluctuate within 10 cm1 depending on variations in oxygen stoichiometry and crystalline disorder. The results of normal coordinate analysis [154] indicate considerable mixing among the vibrations represented by the Cartesian coordinates of individual atoms. If the 123 conductor is prepared in pure oxygen and the oxygen is removed quantitatively by heating the sample in argon, a series of samples having d ¼ 0, 0.2, 0.5, and 0.7 can be obtained. The Tc values of these samples were found to be 94, 77, 50, and 20 K, respectively. Thomsen et al. [155] prepared a series of such samples, and measured their Raman spectra. Figure 1.54 shows a plot of vibrational frequencies of the four Ag modes mentioned above against d values. It is seen that the two modes at 502 and 154 cm1 are softened and the two modes at 438 and 334 cm1 are hardened as the oxygen is removed from the sample. These results

Fig. 1.54. Dependence of four Raman frequencies on oxygen concentration at 4 K and 298 K [155].

REFERENCES

141

seem to suggest that the Tc of the 123 conductor is related to oxygen deficiency in the O(2)Cu(2)O(3) layer. The effect of changing the size of the cation (Y) [156] and isotopic substitution (16O/18O) [157] on these Raman bands has also been studied. Thus far, IR studies on the 123 conductor have been hindered by the difficulties in obtaining IR spectra from highly opaque samples and in making reliableb and assignments [147,148].

REFERENCES Theory of Molecular Vibrations 1. G. Herzberg, Molecular Spectra and Molecular Structure, Vol. II: Infrared and Raman Spectra of Polyatomic Molecules, Van Nostrand, Princeton, NJ, 1945. 2. G. Herzberg, Molecular Spectra and Molecular Structure, Vol. I: Spectra of Diatomic Molecules, Van Nostrand, Princeton, NJ, 1950. 3. E. B. Wilson, J. C. Decius, and P. C. Cross, Molecular Vibrations, McGraw-Hill, New York, 1955. 4. C. J. H. Schutte, The Theory of Molecular Spectroscopy, Vol. I: The Quantum Mechanics and Group Theory of Vibrating and Rotating Molecules, North Holland, Amsterdam, 1976. 5. L. A. Woodward, Introduction to the Theory of Molecular Vibrations and Vibrational Spectroscopy, Oxford University Press, London, 1972. 6. J. D. Craybeal, Molecular Spectroscopy, McGraw-Hill, New York, 1988. 7. J. M. Hollas, Modern Spectroscopy, 2nd ed., Wiley, New York, 1992. 8. M. Diem, Introduction to Modern Vibrational Spectroscopy, Wiley, New York, 1993.

Symmetry and Group Theory 9. J. R. Ferraro and J. S. Ziomek, Introductory Group Theory and Its Application to Molecular Structure, 2nd ed., Plenum Press, New York, 1975. 10. D. C. Harris and M. D. Bertolucci, Symmetry and Spectroscopy, Oxford University Press, London, 1978. 11. P. R. Bunker, Molecular Symmetry and Spectroscopy, Academic Press, New York, 1979. 12. F. A. Cotton, Chemical Application of Group Theory, 3rd ed., Wiley-Interscience, New York, 1990. 13. R. L. Carter, Molecular Symmetry and Group Theory, Wiley, New York, 1992. 14. S. F. A. Kettle, Symmetry and Structure, 2nd ed., Wiley, New York, 2000.

Correlation Method 15. W. G. Fateley, F. R. Dollish, N. T. McDevitt, and F. F. Bentley, Infrared and Raman Selection Rules for Molecular and Lattice Vibrations: The Correlation Method, WileyInterscience, New York, 1972.

142

THEORY OF NORMAL VIBRATIONS

Vibrational Intensities 16. W. B. Person and G. Zerbi, eds., Vibrational Intensities in Infrared and Raman Spectroscopy, Elsevier, Amsterdam, 1982.

Fourier Transform Infrared Spectroscopy 17. J. R. Ferraro and L. J. Basile, eds., Fourier Transform Infrared Spectroscopy, Vol. I, Academic Press, New York, 1978 to present. 18. P. G. Griffiths, and J. A. de Haseth, Fourier Transform Infrared Spectrometry, Wiley, New York, 1986. 19. J. R. Ferraro and K. Krishnan, eds., Practical Fourier Transform Infrared Spectroscopy, Academic Press, San Diego, CA, 1990. 20. H. C. Smith, Fundamentals of Fourier Transform Infrared Spectroscopy, CRC Press, Boca Raton, FL, 1996.

Raman Spectroscopy 21. D. P. Strommen and K. Nakamoto, Laboratory Raman Spectroscopy, Wiley, New York, 1984. 22. J. G. Grasselli and B. J. Bulkin, eds., Analytical Raman Spectroscopy, Wiley, New York, 1991. 23. H. A. Szymanski, eds., Raman Spectroscopy: Theory and Practice, Vol. 1, Plenum Press, New York, 1967; Vol. 2 1970. 24. J. A. Koningstein, Introduction to the Theory of the Raman Effect, D. Reidel, Dordrecht (Holland), 1973. 25. D. A. Long, Raman Spectroscopy, McGraw-Hill, New York, 1977. 26. J. R. Ferraro, K. Nakamoto, and C. W. Brown, Introductory Raman Specctroscopy, 2nd ed., Academic Press, San Diego, CA, 2003. 27. E. Smith, Modern Raman Spectroscopy, Wiley, New York, 2004.

Low-Temperature and Matrix Isolation Spectroscopy 28. B. Meyer, Low-Temperature Spectroscopy, Elsevier, Amsterdam, 1971. 29. H. E. Hallarn, ed., Vibrational Spectroscopy of Trapped Species, Wiley, New York, 1973. 30. M. Moskovits and G. A. Ozin, eds., Cryochemistry, Wiley, New York, 1976.

Time-Resolved Spectroscopy 31. G. H. Atkinson, Time-Resolved Vibrational Spectroscopy, Academic Press, New York, 1983. 32. M. D. Fayer, ed., Ultrafast Infrared and Raman Spectroscopy, Marcel Dekker, New York, 2001.

High-Pressure Spectroscopy 33. J. R. Ferraro, Vibrational Spectroscopy at High External Pressures—the Diamond Anvil Cell, Academic Press, New York, 1984.

REFERENCES

143

Vibrational Spectra of Inorganic, Coordination, and Organometallic Compounds 34. J. R. Ferraro, Low-Frequency Vibrations of Inorganic and Coordination Compounds, Plenum Press, New York, 1971. 35. L. H. Jones, Inorganic Spectroscopy, Vol. I, Marcel Dekker, New York, 1971. 36. R. A. Nyquist and R. O. Kagel, Infrared Spectra of Inorganic Compounds, Academic Press, New York, 1971. 37. S. D. Ross, Inorganic Infrared and Raman Spectra, McGraw-Hill, New York, 1972. 38. E. Maslowsky, Jr., Vibrational Spectra of Organometallic Compounds, Wiley, New York, 1977.

Vibrational Spectra of Organic Compounds 39. N. B. Colthup, L. H. Daly, and S. E. Wiberley, Introduction to Infrared and Raman Spectroscopy, 3rd ed., Academic Press, San Diego, CA, 1990. 40. D. Lin-Vien, N. B. Colthup, W. G. Fateley, and J. G. Grasselli, The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, San Diego, CA, 1991.

Vibrational Spectra of Biological Compounds 41. A. T. Tu, Raman Spectroscopy in Biology, Wiley, New York, 1982. 42. P. R. Carey, Biochemical Applications of Raman and Resonance Raman Spectroscopies, Academic Press, New York, 1982. 43. F. S. Parker, Application of Infrared, Raman and Resonance Raman Spectroscopy in Biochemistry, Plenum Press, New York, 1983. 44. T. G. Spiro, ed., Biological Applications of Raman Spectroscopy, Vols. 1–3, Wiley, New York, 1987–1988. 45. H.-U. Gremlich and B. Yan, eds., Infrared and Raman Spectroscopy of Biological Materials, Marcel Dekker, New York, 2001.

Vibrational Spectra of Adsorbed Species 46. A. T. Bell and M. L. Hair, eds., Vibrational Spectroscopies for Adsorbed Species, American Chemical Society, Washington, DC, 1980. 47. J. T. Yates, Jr. and T. E. Madey, eds., Vibrational Spectroscopy of Molecules on Surfaces, Plenum Press, New York, 1987.

Vibrational Spectra of Crystals and Minerals 48. G. Turrell, Infrared and Raman Spectra of Crystals, Academic Press, New York, 1972. 49. V. C. Farmer, The Infrared Spectra of Minerals, Mineralogical Society, London, 1974. 50. J. A. Gadsden, Infrared Spectra of Minerals and Related Inorganic Compounds, Butterworth, London, 1975. 51. C. Karr, ed., Infrared and Raman Spectroscopy of Lunar and Terrestrial Minerals, Academic Press, New York, 1975.

144

THEORY OF NORMAL VIBRATIONS

52. J. C. Decius and R. M. Hexter, Molecular Vibrations in Crystals, McGraw-Hill, New York, 1977.

Advances Series 53. Spectroscopic Properties of Inorganic and Organometallic Compounds, Vol. 1 to present, The Chemical Society, London. 54. Molecular Spectroscopy—Specialist Periodical Reports, Vol. 1 to present, The Chemical Society, London. 55. R. J. H. Clark and R. E. Hester, eds., Advances in Infrared and Raman Spectroscopy, Vol. 1 to present, Heyden, London. 56. J. Durig, ed., Vibrational Spectra and Structure, Vol. 1 to present, Elsevier, Amsterdam. 57. C. B. Moore, ed., Chemical and Biochemical Applications of Lasers, Vol. 1 to present, Academic Press, New York. 58. Structure and Bonding, Vol. 1 to present, Springer-Verlag, New York.

Index and Collection of Spectral Data 59. N. N. Greenwood, E. J. F. Ross, and B. P. Straughan, Index of Vibrational Spectra of Inorganic and Organometallic Compounds, Vols. 1–3, Butterworth, London, 1972– 1977. 60. B. Schrader, Raman/IR Atlas of Organic Compounds, VCH, New York, 1989. 61. Sadtler’s IR Handbook of Inorganic Chemicals, Bio-Rad Laboratories, Sadtler Division, Philadelphia, PA, 1996.

Handbooks and Encyclopedia 62. R. A. Ayquist, R. O. Kagel, C. I. Putzig, and M. A. Leugers, The Handbook of Infrared and Raman Spectra of Inorganic Compounds and Organic Salts, Vols. 1–4, Academic Press, San Diego, CA, 1996. 63. Encyclopedia of Spectroscopy and Spectrometry, 3-vol. set with online version Academic Press, San Diego, CA, 2000. 64. I. R. Lewis and H. G. M. Edwards, eds., Handbook of Raman Spectroscopy, Marcel Dekker, New York, 2001. 65. J. M. Chalmers and P. R. Griffiths, eds., Handbook of Vibrational Spectroscopy, Vols. 1– 5, Wiley, Chichester, UK, 2002.

General 66. 67. 68. 69. 70. 71. 72.

W. Holzer, W. F. Murphy, and H. J. Bernstein, J. Chem. Phys. 52, 399 (1970). C. F. Shaw, III, J. Chem. Educ. 58, 343 (1981). R. M. Badger, J. Chem. Phys. 2, 128 (1934). W. Gordy, J. Chem. Phys. 14, 305 (1946). D. R. Herschbach and V. W. Laurie, J. Chem. Phys. 35, 458 (1961). K. Nakamoto, M. Margoshes, and R. E. Rundle, J. Am. Chem. Soc. 77, 6480 (1955). E. M. Layton, Jr., R. D. Cross, and V. A. Fassel, J. Chem. Phys. 25, 135 (1956).

REFERENCES

73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.

89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106.

145

F. D. Hardcastle and I. E. Wachs, J. Raman Spectrosc. 21, 683 (1990). F. D. Hardcastle and I. E. Wachs, J. Phys. Chem. 95, 5031 (1991). D. P. Strommen and E. R. Lippincott, J. Chem. Educ. 49, 341 (1972). G. C. Lie, J. Chem. Educ. 56, 636 (1979). H. Eyring, J. Walter, and G. E. Kimball, Quantum Chemistry, 5th ed., Wiley, New York, 1949, p. 121. E. B. Wilson, J. Chem, Phys. 7, 1047 (1939); 9, 76 (1941). J. C. Decius, J. Chem. Phys. 16, 1025 (1948). T. Shimanouchi, J. Chem. Phys. 25, 660 (1956). T. Shimanouchi, “The Molecular Force Field,” in D. Henderson,ed., Physical Chemistry: An Advanced Treatise, Vol. 4, Academic Press, New York, 1970. T. Shimanouchi, J. Chem. Phys. 17, 245, 734, 848 (1949). D. F. Heath and J. W. Linnett, Trans. Faraday Soc. 44, 556, 873, 878, 884 (1948); 45, 264 (1949). J. Overend and J. R. Scherer, J. Chem. Phys. 32, 1289, 1296, 1720 (1960); 33, 446 (1960); 34, 547 (1961); 36, 3308 (1962). T. Shimanouchi, Pure Appl. Chem. 7, 131 (1963). J. H. Schachtschneider, “Vibrational Analysis of Polyatomic Molecules,” Pts. V and VI, Tech. Rept. 231-64 and 53-65, Shell Development Co., Emeryville, CA, 1964 and 1965. K. Nakamoto, Angew. Chem. 11, 666 (1972). N. Mohan, K. Nakamoto, and A. M€uller, “The Metal Isotope Effect on Molecular Vibrations,” in R. J. H. Clark and R. E. Hester,eds., Advances in Infrared and Raman Spectroscopy, Vol. 1, Heyden, London, 1976. K. Nakamoto, K. Shobatake, and B. Hutchinson, Chem. Commun. 1451 (1969). J. Takemoto and K. Nakamoto, Chem. Commun. 1017 (1970). J. R. Kincaid and K. Nakamoto, Spectrochim. Acta. 32A, 277 (1976). P. M. Champion, B. R. Stallard, G. C. Wagner, and I. C. Gunsalus, J. Am. Chem. Soc. 104, 5469 (1982). Y. Morino and K. Kuchitsu, J. Chem. Phys. 20, 1809 (1952). B. L. Crawford and W. H. Fletcher, J. Chem. Phys. 19, 141 (1951). P. LaBonville and J. M. Williams, Appl. Spectrosc. 25, 672 (1971). D. A. Ramsey, J. Am. Chem. Soc. 74, 72 (1952). D. P. Strommen and K. Nakamoto, Appl. Spectrosc. 37, 436 (1983). T. G. Spiro and T. C. Strekas, Proc. Natl Acad. Sci. USA. 69, 2622 (1972). W. M. McClain, J. Chem. Phys. 55, 2789 (1971). D. P. Strommen, J. Chem. Educ. 69, 803 (1992). W. Kiefer, Appl. Spectrosc. 28, 115 (1974). M. Mingardi and W. Siebrand, J. Chem. Phys. 62, 1074 (1975). A. C. Albrecht, J. Chem. Phys. 34, 1476 (1961). S. A. Asher, Anal. Chem. 65, 59A, 201A (1993). W. Kiefer and H. J. Bernstein, J. Raman Spectrosc. 1, 417 (1973). R. J. H. Clark and P. D. Mitchell, J. Am. Chem. Soc. 95, 8300 (1973).

146

THEORY OF NORMAL VIBRATIONS

L. A. Nafie, P. Stein, and W. L. Peticolas, Chem. Phys. Lett. 12, 131 (1971). Y. Hirakawa and M. Tsuboi, Science 188, 359 (1975). T. G. Spiro, R. S. Czernuszewicz, and X.-Y. Li, Coord. Chem. Rev. 100, 541 (1990). X.-Y. Li, R. S. Czernuszewicz, J. R. Kincaid, P. Stein, and T. G. Spiro, J. Phys. Chem. 94, 47 (1990). 110a. Y. Mizutani, Y. Uesugi, and T. Kitagawa, J. Chem. Phys. 111, 8950 (1999). 111. H. Hamaguchi, I. Harada, and T. Shimanouchi, Chem. Phys. Lett. 32, 103 (1975). 112. H. Hamaguchi, J. Chem. Phys. 69, 569 (1978). 113. F. Inagaki, M. Tasunai, and T. Miyazawa, J. Mol. Spectrosc. 50, 286 (1974). 114. P. Pulay, Mol. Phys. 17, 192 (1969); 18, 473 (1970); P. Pulay and W. Meyer, J. Mol. Spectrosc. 40, 59 (1971). 115. R. G. Parr and W. Yang, Density Functional Theory of Atoms and Molecules, Oxford University Press, Oxford, 1989; J. K. Labanowski and J. W. Andzelm, eds., Density Functional Methods in Chemistry, Springer, Berlin, 1990. S. M. Clark, R. L. Gabriel, and D. Ben-Amotz, J. Chem. Educ. 27, 654 (2000). 116. J. B. Foresman and A. Fisch, Exploring Chemistry with Electronic Structure Method, 2nd ed., Gaussian Inc., Pittsburgh, PA, 1996. 116a B. G. Johnson and M. J. Fisch, J. Chem. Phys. 100, 7429 (1994). 117. H. D. Stidham, A. C. Vlaservich, D. Y. Hsu, G. A. Guirgis, and J. R. Durig, J. Raman Spectrosc. 25, 751 (1994). 118. W. B. Collier, J. Chem. Phys. 88, 7295 (1988). 119. J. A. Boatz and M. S. Gordon, J. Phys. Chem. 93, 1819 (1989). 120. P. Pulay, J. Mol. Struct. 147, 293 (1995). 121. M. W. Wong, Chem. Phys. Lett. 256, 391 (1996). 122. G. A. Ozin, “Single Crystal and Gas Phase Raman Spectroscopy in Inorganic Chemistry,” in S. J. Lippard,ed., Progress in Inorganic Chemistry, Vol. 14, Wiley-Interscience, New York, 1971. 123. R. J. H. Clark and D. M. Rippon, J. Mol. Spectrosc. 44, 479 (1972). 124. E. Whittle, D. A. Dows, and G. C. Pimentel, J. Chem. Phys. 22, 1943 (1954). 125. M. J. Linevsky, J. Chem. Phys. 34, 587 (1961). 126. D. E. Milligan and M. E. Jacox, J. Chem. Phys. 38, 2627 (1963). 127. D. Tevault and K. Nakamoto, Inorg. Chem. 14, 2371 (1975). 128. S. P. Willson and L. Andrews, “Matrix Isolation Infrared Spectroscopy,” in J. M. Chalmers and P. R. Griffiths, eds., Handbook of Vibrational Spectroscopy, Vol.2, Wiley, London, 2002, p. 1342 . 129. B. Liang, L. Andrews, N. Ismail, and C. J. Marsden, Inorg. Chem. 41, 2811 (2002). 130. M. Zhou and L. Andrews, J. Phys. Chem. A 103, 2964 (1999). 131. D. Tevault and K. Nakamoto, Inorg. Chem. 15, 1282 (1976). 132. B. S. Ault, J. Phys. Chem. 85, 3083 (1981). 133. T. Watanabe, T. Ama, and K. Nakamoto, J. Phys. Chem. 88, 440 (1984). 134. K. Bajdor and K. Nakamoto, J. Am. Chem. Soc. 106, 3045 (1984). 135. G. Burns and A. M. Glazer, Space Groups for Solid State Scientists, Academic Press, New York, 1978, p. 81. 107. 108. 109. 110.

REFERENCES

147

136. J. M. Robertson, Organic Crystals and Molecules, Cornell University Press, Ithaca, NY, 1953, p. 44. 137. International Tables for X-Ray Crystallography, Knoch Press, Birmingham, UK, 1952. 138. R. S. Halford, J. Chem. Phys. 14, 8 (1946). 139. R. W. C. Wyckoff, Crystal Structures, Vols. I and II, Wiley-Interscience, New York, 1964. 140. S. Bhagavantam and T. Venkatarayudu, Proc. Indian Acad. Sci. 9A, 224 (1939).Theory of Groups and Its Application to Physical Problems, Andhra University, Waltair, India, 1951. 141. C. Kittel, Introduction to Solid State Physics, 5th ed., Wiley, New York, 1976, p. 118. 142. M. Tsuboi, Infrared Absorption Spectra, Vol. 6, Nankodo, Tokyo, 1958, p. 41. 143. S. P. Porto, J. A. Giordmaine, and T. C. Damen, Phys. Rev. 147, 608 (1966). 144. T. Shimanouchi, M. Tsuboi, and T. Miyazawa, J. Chem. Phys. 35, 1597 (1961). 145. I. Nakagawa and J. L. Walter, J. Chem. Phys. 51, 1389 (1969). 146. M. K. Wu, J. R. Ashburn, C. J. Torng, P. H. Hor, R. L. Meng, L. Gao, Z. J. Huang, Y. Q. Wang, and C. W. Chu, Phys. Rev. Lett. 55, 908 (1987). 147. J. R. Ferraro and V. A. Maroni, Appl Spectrosc. 44, 351 (1990). 148. V. A. Maroni and J. R. Ferraro, Practical Fourier Transform Infrared Spectrocopy, Academic Press, San Diego, CA, 1990, p. 1. 149. I. K. Schuller and J. D. Jorgensen, Mater. Res. Soc. Bull. 14(1), 27 (1989). 150. J. Hanuza, J. Klamut, R. Horyn, and B. Je_zowska-Trzebiatowska, J. Mol Struct. 193, 57 (1989). 151. J. R. Ferraro and V. A. Maroni, private communication. 152. M. Stavola, D. M. Krol, L. F. Schneemeyer, S. A. Sunshine, J. V. Waszczak, and S. G. Kosinski, Phys. Rev. B: Condens. Matter [3] 39, 287 (1989). 153. Y. Morioka, A. Tokiwa, M. Kikuchi, and Y. Syono, Solid State Commun. 67, 267 (1988). 154. F. E. Bates and J. E. Eldridge, Solid State Commun. 64, 1435 (1987). F. E. Bates, Phys. Rev. B39, 322 (1989). 155. C. Thomsen, R. Liu, M. Bauer, A. Wittlin, L. Genzel, M. Cardona, E. Sch€ onherr, W. Bauhofer, and W. K€onig, Solid State Commun. 65, 55 (1988). 156. M. Cardona, R. Liu, C. Thomsen, M. Bauer, L. Genzel, W. K€ onig, A. Wittlin, U. Amador, M. Barahona, F. Fernandez, C. Otero, and R. Saez, Solid State Commun. 65, 71 (1988). 157. B. Batlogg, R. J. Cava, A. Jarayaman, R. B. van Dover, G. A. Kourouklis, S. Sunshine, D. W. Murphy, L. W. Rupp, H. S. Chen, A. White, K. T. Short, A. M. Mujsce, and E. A. Rietman, Phys. Rev. Lett. 58, 2333 (1987).

Chapter

2

Applications in Inorganic Chemistry 2.1. DIATOMIC MOLECULES

*

As shown in Sec. 1.3, diatomic molecules have only one vibration along the chemical bond; its frequency in term of wavenumber ð~ nÞ is given by 1 ~ n¼ 2pc

sﬃﬃﬃﬃ K m

where K is the force constant, m the reduced mass and c is the velocity of light. In homopolar XX molecules (D¥h) thevibration is not infrared-active but is Raman-active, whereas it is both infrared- and Raman-active in heteropolar XY molecules (C¥n). 2.1.1. Vibrational Frequencies with Anharmonicity Corrections Tables 2.1a and 2.1b list a number of diatomic molecules and ions of the X2 and XY types, for which frequencies corrected for anharmonicity (oe) and anharmonicity constants (xeoe) are known. The force constants can be calculated directly from these oe values. In the X2 series, the oe values cover a wide range of frequencies, the highest being that of H2 (4395 cm1) and the lowest being that of Cs2 (42 cm1). Correspondingly, *

Hereafter, the word molecules is also used to represent ions and radicals.

Infrared and Raman Spectra of Inorganic and Coordination Compounds, Sixth Edition, Part A: Theory and Applications in Inorganic Chemistry, by Kazuo Nakamoto Copyright 2009 John Wiley & Sons, Inc.

149

150

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.1a. Harmonic Frequencies and Anharmonicity Constants of X2-Type Molecules (cm1)a,b Molecule H2 H2þ 7 Li2 23 Na2 39 K2 85 Rb2 133 Cs2 11 B2 Al2 12 C2 Si2 Ge2 Sn2 208 Pb2 14 N2 14 þ N2 15 N2 La2 Ce2 Pr2 Nd2 Gd2 Tb2 Lu2

oe 4395.2 4004.77 351.4 159.2 92.6 57.3 42.0 1051.3 297.5 1641.4 548.0 287.95 204.3 110.20 2358.54 2207.2 2278.79 236.0 245.4 244.9 148.0 138.7 137.6 121.6

xeoe

Ref.

Molecule

117.91 — 2.59 0.73 0.35 0.96 0.08 9.4 1.69 11.67 2.2 0.32 0.2 0.34 14.31 16.14 13.35

1 2 1 1 1 1 1 1 3 1 4 4a,4 4 5 6 1 6

31

0.9 — — 0.7 0.3 0.31 0.16

13 14 14 14 15 16 17

Hf2 Re2 (ex) Co2 Rh2 58 Ni2 Ag2

P2 As2 75 Asþ 2 Sb2 209 Bi2 16 O2 O2þ 32 S2 80 Se2 Te2 19 F2 35 Cl2 35 þ Cl2 79 Br2 127 I2 Xe2þ Y2 75

oe

xeoe

Ref.

780.4 429.4 (314.8) 269.9 172.7 1580.4 1876.4 725.68 391.8 251 916.93 559.71 645.3 325.43 214.6 124 184.4

2.80 1.12 (1.25) 0.59 0.32 12.07 16.53 2.85 1.06 0.55 11.32 2.70 2.90 1.10 0.61 0.5 0.30

1 1 1 1 1,7 1 1 1 1 1 8 9 1 10 1 11 12

>1 1.0 2.2 1.83 1.9 0.7

18 19 20 21 22 23

176.2 317.1 296.8 283.9 259.2 151.39

a The compounds are listed in the order of the periodic table beginning with the first atom: IA, IIA,. . ., IB, IIB, and so on. b Values in parentheses are uncertain. The oe and xeoe values are rounded off two decimals.

the xeoe value is largest for H2 (117.91 cm1) and smallest for Cs2 (0.08 cm1). Extensive studies on hydrides show that their frequencies span from 4395 cm1 (H2) to 891 cm1 (HCs). These tables also contain many isotopic frequencies involving light as well as heavy atoms.The effects of changing the mass and/or force constant in a series of diatomic molecules have been discussed in Sec. 1.3. Similar series are found in these tables. For example, we find that (all in units of cm1) Li2 ð351:4Þ > Na2 ð159:2Þ > K2 ð92:6Þ > Rb2 ð57:3Þ > Cs2 ð42:0Þ HBe ð2058:6Þ > HMg ð1495:3Þ > HCa ð1298:4Þ > HSr ð1206:9Þ > HBa ð1168:4Þ Across the periodic table, we find that HB ð 2366Þ < HC ð2860:4Þ < HN ð 3300Þ < HO ð3735:2Þ < HF ð4138:6Þ The frequency decreases on removal of bonding electrons: N2 ð2358:5Þ > Nþ 2 ð2207:2Þ;

As2 ð429:4Þ > Asþ 2 ð314:8Þ

DIATOMIC MOLECULES

151

TABLE 2.1b. Harmonic Frequencies and Anharmonicity Constants of XY-Type Molecules (cm1)a,b Molecule

oe

xeoe

Ref.

Molecule 138

HD DD H 6 Li H 7 Li D 6 Li D 7 Li H 23 Na D 23 Na H 39 K H 85 Rb D 85Rb H 87 Rb H Cs D 133 Cs H 9 Be H 24 Mg D 24 Mg H Ca H 88 Sr D 88 Sr

3817.1 3118.5 1420.12 1405.51 1074.33 1054.94 1171.76 845.97 985.67 937.10 666.66 936.98 891.25 632.80 2058.6 1495.26 1078.14 1298.40 1206.89 858.85

94.96 64.10 23.66 23.18 13.54 13.05 19.52 9.98 14.90 14.28 7.06 14.27 12.82 6.34 35.5 31.64 16.15 19.18 17.03 8.64

1 1 24–26 24,25 24,26 24 27,28 29 29,30 31 29 31 32 29 1 33 33 34 35 36

Ba H DBa HSc DSc HMn HCo HNi HCu H 107 Ag D 109 Ag H 197 Au H 64 Zn D 64 Zn H 110 Cd D 116 Cd H Hg DY D Cr H 11 B H 27 Al

D 27 Al H 69 Ga D 71 Ga H 115 In D 115 In H 205 Tl D 205 Tl H 12 C D 12 C H 28 Si H Ge H 120 Sn H 208 Pb H 14 N DN (H 14 N) HP H As H 209 Bi H 16 O

1211.77 1603.96 1142.77 1475.43 1048.60 1391.27 987.04 2860.4 2101.0 2042.5 1831.85 1655.49 1560.53 (3300) 2399.13 3191.5 2363.77 2076.93 1699.52 3735.2

15.06 28.42 14.42 25.16 12.70 23.10 11.67 64.11 34.7 35.67 32.86 28.83 28.79 — 42.11 85.6 43.91 39.22 31.93 82.81

49 50,51 44,51 52,53 44 54 44 55 1,56 57 58,59 60–62 63 1 64 65 64 66 67–69 1

D 16 O (D O) HS (H 32 S) H Se H 19 F D 19 F H 35 Cl D 35 Cl D 37 Cl H Br HI (H 132 Xe)þ 7 Li O 6 Li F 7 Li F 7 Li Cl 7 Li Br 7 Li I 23 Na K

106.6 536.1 366 302 258 426 281 213 212

0.46 3.83 2.05 1.50 1.08 2.4 1.30 0.80 0.70

1 80,81 82,83 82,84 82 85,86 82,83 82 1

Mg 35 Cl Mg 79 Br Mg 127 I 24 Mg 16 O Mg S 40 Ca 19 F Ca 35 Cl Ca 79 Br Ca 127 I

23

Na Rb Na 19 F 23 Na Cl 23 Na Br 23 Na I KF K Cl K Br KI 23

24 24

xeoe

Ref.

1168.43 829.84 1546.97 1141.27 1546.85 1858.79 2000 1940.4 1759.67 1250.68 2305.0 1615.72 1147.36 1461.13 1032.04 1387.1 1089.12 1183.20 (2366) 1682.4

14.61 7.37 — 12.38 27.60 — 40 37 33.93 17.21 43.12 59.62 28.72 61.99 29.29 83.01 9.85 15.60 (49) 29.11

37,38 36 39 39 40 41 42 42 43 44 1 45 36 1 36 1 46 47 1 48,49

2720.9 2723.5 2696,25 2647.1 2421.72 4138.55 2998.3 2991.0 2145.2 2141.97 2649.7 2309.5 2270.0 799.07 964.31 906.2 641.1 563.2 498.2 123.3

44.2 49.72 48.74 53.28 44.60 90.07 45.71 52.85 27.18 27.08 45.21 39.73 41.33 — 9.20 7.90 4.2 3.53 3.39 0.40

1,70 70 71 72 73 1 1 74 74,75 75 1 1 76 77 78 79,78 79 79 79 1

2.05 1.34 — 5.18 2.93 2.74 1.31 0.86 0.64

1 1 1 1,90,91 1 1,92 1 1 1

oe

465.4 373.8 [312] 785.1 525.2 587.1 369.8 285.3 242.0

(continued)

152

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.1b. (Continued (Continued ) Molecule

oe

xeoe

Ref.

Molecule

oe

xeoe

Ref.

— 1.90 0.92 1.63 0.75 (2.0) (1.2) 9.12 5.11 11.83 3.84

1 87,85 82 87 82 1 1,88 1 1 1 1,89

16

Ca O Sr 19 F Sr 35 Cl Sr 79 Br Sr 127 I Sr 16 O Ba 19 F 138 Ba 35 Cl Ba 79 Br Ba I 138 Ba O

732.03 500.1 302.3 216.5 173.9 653.5 468.9 279.3 193.8 152.16 669.73

4.83 2.21 0.95 0.51 0.42 4.0 1.79 0.89 0.42 0.27 2.02

93,94 1,95 1 1 1 1,96 1,97 1 1 98 99,90

3.95 2.45 2.23 2.99 1.20 3.70 4.07 1.23 6.3 4.61 3.45 4.41 4.9 4.22 — 6.5 3.01 1.4 0.9

1 1 1 1 1 1 1 1 1 1 1 100 1 101 102 1 1 1 1

55

Mn 16 O Fe 35 Cl Fe 16 O Co F Co Cl 193 Ir N Ni Cl 63 Cu 19 F 63 Cu 35 Cl 63 Cu 79 Br 63 Cu 127 I Cu 16 O 107 Ag 35 Cl 107 Ag 81Br 107 Ag I Ag O 197 Au 35 Cl Zn 19 F Zn 35 Cl

840.7 406.6 880 678.18 421.2 1126.18 419.2 622.7 416.9 314.1 264.8 628 343.6 247.7 206.2 493.21 382.8 (630) 390.56

4.89 1.2 5 2.74 0.74 6.29 1.04 3.95 1.57 0.87 0.71 3 1.16 0.68 0.43 4.10 1.30 (3.5) 1.55

1 1 1 103 1 104 1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1 105 1 106,107 1,108 1 105 109 1,110 1

27

Al 16 O Al S 69 Ga F 69 Ga 35 Cl 69 Ga 81 Br 69 Ga 127 I Ga As (gr) Ga As (ex) Ga O In As 115 In F 115 35 In Cl 115 81 In Br 115 127 In I In 16 O Tl 19 F Tl 35 Cl Tl 81Br Tl 127 I (CF)þ

978.23 623.13 622.37 365.0 263.0 216.4 215 152 767.7 182.7 535.36 317.41 221.0 177.1 703.1 475.0 287.5 192.18 150 1792.76

7.12 6.22 3.30 1.11 0.81 0.5 3 2.89 6.34 0.5 2.67 1.01 0.65 0.4 3.71 1.89 1.24 0.39 — 13.23

1,111 112 113 1 1 1 114 114 1 115 116 1 1 1 1 1,117 1 1 1 118

133

Cs Rb 85 Rb 19 F Rb Cl 132 Cs 19 F Cs Cl 133 Cs Br 133 Cs 127 I 9 Be 19 F 9 Be 35 Cl 9 Be 16 O 24 Mg 19 F

49.4 373.44 228 352.62 209 (194) 142 1265.5 846.65 1487.3 717.6

45

Sc 16 O Y 16 O 129 La 16 O Ce 16 O 141 Pr 16 O Gd 16 O Lu 16 O Yb Cl 48 Ti 35 Cl 48 Ti 16 O 90 Zr 16 O 180 Hf N V 16 O Cr F Cr Cl Cr 16 O 55 Mn F 55 Mn 35 Cl 55 Mn Br

971.6 852.5 811.6 865.0 818.9 841.0 841.7 293.6 456.4 1008.4 936.6 932.72 1012.7 664.10 396.66 898.8 618.8 384,9 289,7

Zn Br 64 Zn 127 I Cd 19 F Cd 35 Cl Cd Br Cd 127 I Hg 19 F Hg 35 Cl 202 Hg 81 Br Hg 127 I Hg Tl 11 19 B F 11 35 B Cl 11 81 B Br 11 14 B N 11 16 B O Al F 27 35 Al Cl 27 79 Al Br 27 127 Al I

(220) — 223.40 0,75 (535) — 330.55 1,2 230,0 0,50 178.56 0.63 490.8 4.05 292.60 1.60 186.39 0.98 125.65 1.09 26.9 0.69 1402.16 11.82 839.1 5.11 684.15 3.73 1514.6 12.3 1885.4 11,77 802.32 4.85 481.4 2.03 378.02 1.28 316.1 1.0

89

DIATOMIC MOLECULES

153

TABLE 2.1b. (Continued (Continued ) Molecule þ

oe

xeoe

Ref.

Molecule 28

(C Cl) 12 35 C Cl C Br 12 14 C N 12 31 C P 12 16 C O (gr) C O (ex) ð12 C 16 OÞþ 12 32 C S 12 C Se Si C (Si F)þ 28 19 Si F 28 Si 35 Cl ½28 Si 35 Clþ Si Br 28 Si 14 N 28 16 Si O (Si O)þ 28 32 Si S

1177.7 846 727.99 2068.7 1239.75 2170.23 1743.76 2214.24 1285.1 1036.0 964.77 1050.47 856.7 535.4 678.24 425.4 1151.7 1242.0 1162.18 749.5

6.65 1.0 3.95 13.14 6.86 13.46 14.57 15.16 6.5 4.8 5.60 4.95 4,7 2.20 — 1.52 6.56 6.05 6.97 2.56

119 1 120 1 1 1 121 1 1,122 1 123 124 1,125 1 126 1 1 1 127 1

Si Se Si Te (79Ge F)þ Ge F 74 Ge 35 Cl Ge Br 74 Ge 16 O 74 Ge 32 S 74 Ge 74 Ge 130 Te Sn 19 F Sn 35 Cl Sn Br Sn 16 O 120 Sn 32 S Sn Se Sn Te Pb 19 F Pb 35 Cl Pb 79 Br

Pb 127 I Pb 16 O 208 Pb 32 S Pb Se Pb Te 14 16 N O 14 NS 14 N Br 14 31 N P 14 75 N As 14 N Sb PF P Cl 31 16 P O P S. (As 35 Cl)þ 75 As 16 O Sb 209 Bi Sb 19 F Sb 35 Cl

160,53 721.8 428.1 277.6 211.8 1903.9 1220.0 693 1337.2 1068.0 942.0 846.7 551.4 1230.66 739.1 527.7 967.4 220.0 614.2 369.0

0.25 3.70 1.20 0.51 0.12 13.97 7.75 5.0 6.98 5.36 5.6 4.49 2.23 6.52 2.79 1.74 5.3 0.50 2.77 0.92

1 1 1 1 1 1,134 1 1 1,135 1 1 136 137 1 138 139 1,140 1 1,141 1,141

Sb 14 N Sb 16 O Bi N Bi P Bi As 209 19 Bi F 209 35 Bi Cl 209 79 Bi Br 209 127 Bi I Bi O OF 16 O Cl 16 O Br 16 OI 16 32 O S Se 16 O Te 16 O 19 35 F Cl 19 F Br Cl Br

610.26 384.2

3.14 1.47

150,151 1

127 79

IF 127 35

I Cl

28

I Br

oe

xeoe

Ref.

580.0 481.2 815.60 666.5 407.6 296.6 985.7 575.8 402.66 323.4 582.9 352.9 247.7 822.4 487.23 331.2 259.5 507.2 303.8 207.55

1.78 1.30 3.22 3.15 1.36 0.9 4.30 1.80 0.87 1.0 2.69 1.06 0.62 3.73 1.35 0.74 0.50 2.30 0.88 0.50

1 1 128 129 1 1 1,130 1,131 132 1 1 1 1 1 133 1 1 1 1 1

942.05 817.2 736.57 430.28 283.30 510.8 308.0 209.3 163.9 688.4 1053.45 (780) 713 (687) 1123.7 907.1 796.0 783.5 662.3 442.5

5.6 5.30 4.83 1.53 0.73 2.05 0.96 0.47 0.31 4.8 10.23 — 7 (5) 6.12 4.61 3.50 4.95 3.80 1.5

1 1 142 143 143 1,144 1 1 1 145 146 1 1,147 1 1 1 1 148 148 149

268.68

0.82

152

a The compounds are listed in the order of the periodic table begining with the first atom: IA, IIA,. . ., IB, IIB, and so on. b Values in parentheses are uncertain, and those in square brackets are observed frequencies without anharmonicity correction. The oe and xeoe values are rounded off to two decimals; (gr) and (ex) indicate the ground and excited states, respectively.

154

APPLICATIONS IN INORGANIC CHEMISTRY

while the opposite trend is found when antibonding electrons are removed: COþ ð2214:2Þ > CO ð2170:2Þ;

SiFþ ð1050:5Þ > SiF ð856:7Þ;

GeFþ ð815:6Þ > GeF ð666:5Þ Table 2.1b also presents two examples for which vibrational frequencies were reported for both ground and excited electronic states: CO ðX1 Sþ ; ground stateÞ; 2170:2 > CO ð3 P; excited stateÞ; 1743:8 GaAs ðX3 S ; ground stateÞ; 215 > GaAsð3 P0þ ; excited stateÞ; 152 Most of the oe and xeoe values listed in Tables 2.1a and 2.1b were obtained mainly by analysis of rotational–vibrational fine structures of infrared or Raman spectra in the gaseous phase [1]. Their values can also be obtained by the analysis of overtone progression in resonance Raman spectra (Sec. 1.22). Vibrational spectra of unstable (or metastable) molecules were obtained in inert gas matrices (Sec. 1.25). Anharmonicity corrections are limited largely to diatomic molecules because nine parameters are required even for nonlinear triatomic molecules. 2.1.2. Vibrational Frequencies without Anharmonicity Corrections Table 2.1c lists the observed frequencies of homopolar diatomic molecules, ions, and radicals. Combining these data with those given in Table 2.1a, the following frequency trends are immediately obvious: 2 Oþ 2 > O2 > O2 > O2 ;

Brþ 2 > Br2 ;

S2 > S 2;

Clþ 2 > Cl2 > Cl2 ;

Iþ 2 > I2 > I2

As mentioned earlier, these orders can be explained in terms of simple molecular orbital theory. Table 2.1d shows the relationship among various molecular parameters in the O2þ > O2 > O2 > O22þ series. It is seen that, as the bond order decreases, the bond distance increases, the bond energy decreases, and the vibrational frequency decreases. Although the force constant is not rigorously related to the bond energy (Sec. 1.3), there is an approximate linear relationship between these parameters. Thus, these frequency trends provide valuable information about the nature of diatomic ligands in coordination compounds (see Chapter 3 in Part B). The O2þ ion was found in compounds such as O2þ [MF6] (M ¼ As, Sb, etc.) and O2þ [M2F11] (M ¼ Nb, Ta, etc.) [171b], whereas the O2 ion was observed in a triangular MO2 type complex formed by the reaction of alkali metals with O2 in an Ar matrix [171c]. The n(O2) of simple metal superoxides (Mþ O2) and peroxides (M2þ O22) have been measured [171d]. Linear relationships are noted between n(S2) and SS bond distances for sulfur compounds [171e]. Table 2.1e lists vibrational frequencies of heteropolar diatomic molecules. It contains isotopic pairs such as (HO/DO), (6 Li 35 C1=7 Li 35 C1), and (28 Si 32 S= 28 Si 34 S).

DIATOMIC MOLECULES

155

TABLE 2.1c. Observed Frequencies of X2 Molecules (cm1) Molecule T2b C2 C22 Sn2 Pb2 P2 As2 16 O 2d O2 O22 S2 S2

Statea

~n

Liquid Gas Solidc Mat Mat Gas Gas Mat Mat Solid Mat Solide

2458 1757,8 1845 188 112 775 421 1555.6 1097 794, 738 718 623

Ref.

Molecule

153 154 155 156 156 157 157 158 159 160 161 162

Se2 F2

35

Cl 2 Cl 2 Br2þ Br2 I2þ I2 I2 K2 Ag2 Zn2 Cd2 35

State e

Solid Mat Mat Mat Sol’n Gas Sol’n Gas Mat Mat Mat Mat Mat

~n

Ref.

349 475 546 247 360 319 238 213 115 91.0 194 80 58

162 163 164 165 166 167 168 167 169 169 170 171 171

a

Mat ¼ inet gas matrix. T ¼ 3H (tritium). c Na salt. d 16 O17 O, 1528.4 cm1; e Doped in KCl. b

16

O18 O, 1508.3 cm1.

Charge effects on vibrational frequencies are seen for ðHOÞ > HO > ðHOÞþ ;

HS > ðHSÞþ ;

ðNOÞþ > NO > ðNOÞ > ðNOÞ2

Most of these frequencies were measured in inert gas matrices at low temperatures. For example, the SH radical was produced by photodecomposition of H2S in Ar or Kr matrices [182], and me FeH radical was produced via the reaction of Fe metal vapor with H2 in Ar matrices [187] (Sec. 3.26). 2.1.3. Dimers and Polymers The N4þ ion is linear and centrosymmetric in the ground state, and exhibits a strong IR absorption at 2237,6 cm1 in Ne matrices [252], and at 2234.51 cm1 in the gaseous state [253]. Surprisingly, stable AsF6 salt of the N5þ ion was synthesized, and its TABLE 2.1d. Relationship between Various Molecular Parameters in the O2 and Its Ions Species O2þ O2 O2 O22 a

Bond Order

Bond ˚ )a Distance (A

Bond Energy (Kcal/mole)

2.5 2.0 1.5 1.0

1.123 1.207 1.280 1.49

149.4 117.2 — 48.8

Ref. 171a. Na salt. c Solid-state splitting. b

~n (O2) (cm1) 1876 1580 1094 791/736b,c

force Constant ˚) (mdyn/A 16.59 11.76 5.67 2.76

156

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.1e. Observed Frequencies of XY Molecules (cm1) Molecule

Statea

~n

Ref.

Molecule

State

~n

Ref.

H Al H Si

DS (H 32 S)þ (D 32 S)þ HF H Cr

Mat Mat Gas Mat Gas Mat Mat Solid Gas Gas Solid Mat Gas Mat Gas Gas Gas Mat

1593 1967 1971.04 3131.6 2077.00 3548.2 2616.1 3637.4 3555.59 2956.37 2225 2607 2591 1847.8 2547.2 1829.1 3961.42 1548

172 173 174 175 176 177 177 178 179 180 181 182 183 183 184 184 185 186

D Cr H Fe H Rh 7 Li O Li F 6 35 Li Cl 7 35 Li Cl 6 37 Li Cl 7 37 Li Cl Li Br Li I Na F Na 35 Cl Mg F 26 Mg O Mg O 40 Ca O Ti O

Mat Mat Mat Mat Mat Gas Gas Gas Gas Mat Mat Mat Gas Mat Mat Gas Mat Mat

1112 1767.0 1920.6 745 885 686.1 643.0 683.3 640.1 510 433 515 364.7 738 815.4 774.74 707.0 1005

186 187 187 188 189 190 190 190 190 191 191 189 192 193 194 195 196 197

Ti N Zr O Nb N Nb O Ta O WO Fe C 56 Fe O Ru O Co O 58 Ni O Th N UN UO Pu N Pu O Zn O Cd O 11 BO BF

Mat Mat Mat Mat Mat Mat Gas Mat Mat Mat Mat Mat Mat Mat Mat Mat Mat Mat Gas Gas

1037 975 1002,5 968 1020 1050.9 852.14 873.1 834.4 846.4 825.7 934.6 995 820 855.7 820 802 719 1861.92 1378.7

198 197 199 199 200 201 202 203 204 203 203 205 206 207 208 209 210 210 211 212

Al O Al F Tl F 205 35 Tl Cl Tl Br Tl I CF CO CS (C N) CN 28 Si Ge 28 Si Sn 29 Si N 28 Si O 28 32 Si S 28 34 Si S 74 Ge O 74 Ge S

Mat Mat Mat Mat Mat Mat Mat Gas Mat Mat Sol’n Mat Mat Mat Gas Mat Gas Gas Mat Mat

975 793 441 284.7 179 143 1279 1286.13 2138.4 1274 2080 2046 422.5 360 1972.80 1225.9 749.65 739.30 973.4 566.6

213 214 215 216 215 215 217 218 219 220 221 222 223 223 224 225 226 226 227 228

74

Ge 80 Se Ge Te

Mat Mat

397.9 317.67

228 228

FO Cl N

Mat Mat

240 239

Sn O Sn S 120 Sn 80 Se Pb O Pb S 208 Pb 32 S Pb 80 Se (N O)þ NO

Mat Mat Mat Mat Mat Gas Mat Solid Mat

229 228 228 230 228 231 228 232 233

Cl O Br N Br O S Cl S Br SI Se Te (Cl F)þ 79 Br O

Mat Mat Mat Mat Mat Mat Mat Sol’n Gas

1028.7 818.5 825 850.7 691 729.9 617 518 443 313 819 723.4

HN H As HO DO (H O) (H O)þ (T O) HS

74

120

816.1 480.5 325.2 718.4 423.1 426.6 275.1 2273 1880

241 239 242 243 243 243 244 245 246

DIATOMIC MOLECULES

157

TABLE 2.1b. (Continued ) 2.1e. (Continued Molecule

(N O) (N O)2 PN SO (S Se) FN a

Statea Mat Mat Mat Mat Solid Mat

~n 1358–1374 886 1323 1101.4 464 1117

Ref. 234 235 236 237 238 239

Molecule

(Br O) Br F 79 Br 35 Cl IN IF I Cl I Br 79

State

~n

Ref.

Sol’n Gas Gas Mat Gas Gas Gas

618 669.9 444.3 590 610.2 381 265

247 248 248 249 250 251 251

Mat ¼ inert gas matrix.

V-shaped (C2v) structure was confirmed by X-ray analysis. Complete assignments of the IR and Raman spectra of the N5þ ion have been based on this structure [254,255]. Figure 2.1 shows the low-temperature Raman spectrum of solid N5þAsF6. A new allotropic form of nitrogen in which all nitrogen atoms are connected by single covalent bonds was produced at temperatures of >2000 K and pressures of >110 GPa [256]. The Raman spectra shown in Fig. 2.2 were obtained at 110 GPa using a diamond anvil cell. At 300 K, a sharp n(NN) peak is seen at 2400 cm1, which is replaced by a new peak at 840 cm1 when the temperature is raised to 1990 K. The latter frequency is close to the n(NN) of N2H4 The possibility of forming bicyclic N10 and fullerene-like N60 has been suggested on the basis of theoretical calculations [257]. In inert gas matrices, a variety of aggregates of diatomic molecules are produced, depending on the experimental conditions employed. For example, the Raman spectra of H2, HD, and D2 in Ar matrices suggest the formation of aggregates of more than six molecules after annealing above 35 K [258].

Fig. 2.1. Structure of the N5þ ion (a) and low-temperature Raman spectrum (b) of solid N5þAsF6. The bands between 700 and 250 cm1 are due to the AsF6 ion [254,255].

158

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.2. Raman spectra of polymeric nitrogen (cubic gauche structure) before and after heating at 110 GPa. The starting sample (300 K, bottom) is in the molecular phase as indicated by the peak at 2400 cm1. The bands at 1550–1300 cm1 are due to the stressed diamond adjacent to the sample. Laser heating to 1990 K (top) transforms it to the polymeric phase with a prominent peak at 840 cm1 [256]. 1GPa ¼ 104 atm.

In an Ar matrix, Kþ(O4) takes the trans structure:

and exhibits n(O¼O) (Ag), n(O¼O) (Bu), n(OO) (Ag), and d(O¼OO) at 1291.5, 993, 305, and 131 cm1, respectively [259,260]. The structure of the O4 molecule is predicted to be either quasisquare D4h or triangular D3h (similar to the CO32þ and NO3 ions) [262]. The IR spectra of the O4 ion in Ne matrices were interpreted as a mixture of a rectangular form (2949 and 1320 cm1) and a bent-trans form (2808 and 1164 cm1) [262a]. The [Cl4]þ ion in solid [Cl4]þ[IrF6] salt assumes a rectangular structure with

TRIATOMIC MOLECULES

159

two short ClCl (1.94 A) and two long ClCl (2.93 A) bonds. The Raman peaks at 578,175, and 241 cm1 were assigned to the n(ClCl), n(ClCl) and d((ClCl)), respectively [262b]. The Raman spectra of (O2)4 (D4h symmetry) at pressure of 32.8 and 33.4 GPa exhibit 4 A1g vibrations at 1600, 1380, 342, and 16 cm1 [262c]. Assignments of the IR spectra of (HF)2 (HF)(DF), and (DF)2 in Ar matrices have been based on the structure [263]

This work has been extended to higher-molecular-weight polymers, (HF)3 and (HF)4 [264]. Matrix IR spectra of (LiF)n (where n ¼ 2,3,4) were interpreted in terms of ring structures (rhombic D2h, D3h, and D4h for n ¼ 2, 3, and 4, respectively) [265]. The Raman spectrum of (NaI)2 in the gasous state at 1084 K supports D2h structure [266]. The matrix IR. spectra of (GeO)2 [267] and (KO)2 [268] suggest a rhomic structure (D2h) and a slightly out-of-plane rhombic structure (C2v), respectively. The nitric oxide dimer, (NO)2, in Ar matrices exists as the cis form (1862 and 1768 cm1) or the trans form (1740 cm1), with the n(NO) shown in brackets [268a]. The IR spectrum of the cis dimer in the gaseous state has been assigned [268b]. IR spectra of cis and trans isomers of (NO)2þ ion produced in Ne matrices exhibit the n(NO) at 1619 and 1424 cm1, respectively. The corresponding frequencies of the (NO)2 ion produced in Ne matrices are at 1225 (cis) and 1227 cml(trans) [268c]. In the gaseous phase, NO forms a weakly bound van der Waals dimer, and the intermolecular vibrations are observed at 429(da), 239(ds), 134 (intermolecular stretch) and 117 cm1(out-of-plane torsion) [268d]. 2.1.4. van der Waals Complexes Inert gases and hydrogen halides form very weakly interacting van der Waals complexes. As expected, the n(HF) of NeHF [269] and n(HCl) of Ne–HCl [270] are shifted only slightly (þ0.4722 andþ3024 cm1, respectively) from their frequencies in the free state. Vibrational frequencies of analogous Ar complexes are also reported [271–275]. IR spectra of van der Waals complexes such as COCH4 [276], COOCS [277] and COBF3 [278] are available. 2.2. TRIATOMIC MOLECULES The three normal modes of linear X3(D¥h)- and YXY(D¥h)-type molecules were shown in Fig. 1.11; n1 is Raman-active but not infrared-active, whereas n2 and n3 are infrared-active but not Raman-active (mutual exclusion rule). However, all three

160

APPLICATIONS IN INORGANIC CHEMISTRY

ν 1(Σ+)

X

Y

Z

ν (XY)

ν 2(II)

δd (XYZ)

ν 3(Σ+)

ν (YZ)

Fig. 2.3. Normal modes of vibration of linear XYZ molecules.

vibrations become infrared- as well as Raman-active in linear XYZ-type molecules (C¥v), shown in Fig. 2.3. The three normal modes of bent X3(C2v) and YXY(C2v)-type molecules were also shown in Fig. 1.12. In this case, all three vibrations are both infrared- and Raman-active. The same holds for bent XXY(Cs)- and XYZ(Cs)-type molecules. Section 1.12 describes the procedure for normal coordinate analysis of the bent YXY-type molecule. In the following, vibrational frequencies of a number of triatomic molecules are listed for each class of compounds. 2.2.1. XY2-Type Halides Table 2.2a lists the vibrational frequencies of XY2-type halides. Most of these data were obtained in inert gas matrices. It is seen that the majority of compounds follow the trend n3 > n1, and that exceptions occur for some bent molecules. Although the structures of these halides are classified either as linear or bent, it should be noted that the bond angles of the latter range from 95 to 170 [315]. Thus, some bent molecules are almost linear. In principle, the distinction between linear and bent structures can be made by IR/Raman selection rules. For example, the n1 is IR-active for the bent structure but IR-inactive for the linear structure. However, such a simple criterion may lead to conflicting results [319]. 0 The bond angle (a) of the YXY molecule can be calculated if the n 3 of the YX0 Y molecule is available. Here, X0 is an isotope of X. Using the G and F elements given in Appendix VII, the following equation can readily be derived: ~0 2 MX MX 0 þMY ð1cosaÞ n3 ¼ ~ n3 MX 0 MX þMY ð1cosaÞ Here M denotes the mass of the atom subscripted. Figure 2.4 shows the spectra of NiF2 in Ne and Ar matrices obtained by Hastie et al. [320]. Using the n3 values

TRIATOMIC MOLECULES

161

TABLE 2.2a. Vibrational Frequencies of XY2-Type Metal Halides (cm1) Compounda

Structure

BeF2 BeCl2 BeBr2 BeI2 MgF2 MgCl2 MgBr2 MgI2 40 CaF2 40 Ca 35 Cl2 CaI2 (g) 86 SrF2 88 Sr 35 Cl2 BaF2 BaCl2 69 GaCl2 InCl 2 CF2 35 C Cl2 CBr2 SiF2 28 35 Si Cl2 28 SiBr2 GeF2 GeCl2 SnF2 SnCl2 SnBr2 PbF2 PbCl2 PbBr2 (g) NF2 PF2 PCl2 PBr2 O35 Cl2 32 SF2 SCl2 SBr2 SI2 76 Se 19 F2 SeCl2 (sol’n) SeBr2 (sol’n) TeCl2 (g) KrF2 (g) XeF2 Xe35Cl2 63 CuF2 [CuCl2] (s) [CuBr2] (s)

Linear Linear Linear Linear Linear Linear Linear Linear Bent Linear Linear Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Linear Linear Linear Bent Linear Linear

~n1

~n2

(680) (390) (230) (160) 550 327 198 148 484.8 — — 441.5 269.3 389.6 255.2 373.0 328 1102 719.5 595.0 851.5 513 402.6 692 398 592.7 354 237 531.2 297 200 1069.6 834.0 452.0 369.0 641.97 838 518 405 368 701.8 372 261 377 449 497 — — 302 194

345 250 220 (175) 249 93 82 56 163.4 63.6 43.3 82.0 43.7 (64) — — 177 668 — — (345) — — 263 — 197 (120) 84 165 — 64 573.4 — — — — 355 208 — — — 168 110 125 233 213.2 — 183.0 108 81

~n3 1555 1135 1010 873 842 601 497 445 553.7 402.3 292.3 443.4 299.5 413.2 260.0 415.1 291 1222 745.7 640.5 864.6 502 399.5 663 373 570.9 334 223 507.2 321 — 930.7 843.5 524.8 410 686.59 817 526 418 376 674.8 346 221 — 596,580 555 314.1 743.9 407 326

Ref. 279 279 279 279 280 280 280 280 281 282 283 281 282 281 282 284 285 286 287 288 289 290 291 292 293 294 293 295 294 293 296 297 298 299 299 300 301 302 302 302 302a 303 303 304 305 306 307 308 308 308

(continued)

162

APPLICATIONS IN INORGANIC CHEMISTRY

2.2b. (Continued (Continued ) TABLE 2.2a. Compounda

[CuI2] (s) [AgCl2] (s) [AgBr2] (s) [AgI2] (s) [AuCl2] (s) [AuBr2] (s) [AuI2] (s) ZnF2 ZnCl2 ZnBr2 ZnI2 (g) CdF2 CdCl2 CdBr2 CdI2 HgF2 HgCl2 HgBr2 HgI2 TiF2 VF2 CrF2 CrCl2 MnF2 MnCl2 MnI2 (g) FeF2 FeCl2 CoF2 CoCl2 NiF2 NiCl2 NiBr2

Structure Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Bent Bent Linear Linear Linear Linear Linear Linear Linear Bent Linear Bent Linear Bent

~n1 148 268 170 133 329 209 158 596 352 223 163 555 (327) (205) (149) 568 (348) (219) (158) 665 — 668 — — — — — — — — — (350) —

~n2 65 88 61 49 120, 112 79, 75 67, 59 150 103,100 71 67.6 121 88 62 (50) 170 107 73 63 180 — 125 — 124.8 83 54.5 141.0 88 151.0 94.5 139.7 85 —

~n3

Ref.

279 333 253 215 350 254 210 754 503 404 337.5 660 419 319 269 642 405 294 237 766 733.2 749 493.5 700.1 476.8 324.2 731.3 493.2 723.5 493.4 779.6 520.6 331.9

309 309 309 309 310 310 310 311 312 312 283 311 313 313 313 311 313 313 313 314 315 316 317 315 317 283 315 317 315 317 315 317 318

a All data were obtained in inert gas matrices except those for which the physical state is indicated as g (gas), sol’n (solution), or s (solid).

obtained for isotopic NiF2, the FNiF angle was calculated to be 154–167 . The bond angle depends on the nature of the matrix gas employed. For example, NiCl2 is linear in Ar matrices, but bent (130 ) in N2 matrices [321]. Thus, structural data obtained in gas matrices must be interpreted with caution. As another example, the bond angle of CaF2 increases from 139 to 143 to 156 on changing the matrix gas from Kr to Ne to N2 [322]. The structure of the dimeric species (MX2)2 is known to be cyclic–planar:

TRIATOMIC MOLECULES

163

Fig. 2.4. IR spectra(n3) of NiF2 in Ne and Ar matrices. Matrix splitting, indicated by asterisks, is present in the Ne matrix spectrum [320].

although an exception is reported for (GeF2)2 [324]:

Mixed fluorides such as LiNaF2 and CaSrF4 take planar ring structures in inert gas matrices [325]:

2.2.2. XY2-Type Oxides, Sulfide, and Related Compounds Table 2.2b lists the vibrational frequencies of triatomic oxides, sulfides, selenides, and related compounds. Most of these data were obtained in inert gas matrices. Although the NO2 (nitrite) ion is bent, the NO2þ (nitronium) ion is linear in most compounds. Solid N2O5 consists of the NO2þ and NO3 ions, and the former is slightly

164

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.2b. Vibrational Frequencies of XY2-Type Oxides and Related Compounds (cm1) Compounda O Li O O Na O OBO Al O Al Ga O Ga O C O (g) 16 12 17 O C O (g) 16 12 18 O C O (g) (O C O)þ (O C O) (O C O)2 S C S (g) (S C S)þ (S C S) Se C Se O 120 Sn O (O N O) ONO (O N O)þ (s) OSO (O S O) O 80 Se O O Te O FOF Cl O Cl (s) (O Cl O) (s) O Cl O (O Cl O)þ (s) 16 79 O Br 16 O Br O Br (s) Br O Br (O Br O)þ (O Br O) OIO O Ce O O Tb O O Th O O Ru O O Pr O O 48 Tl O O Ta O 16 189 O W 16 O (ONiO)2 (s) OUO (O U O)2þ (s) O Pu O (N B N)3 (s) 14 N U 14 N 14 N U 15 N

Structure Bent Bent Linear Bent Bent Linear Linear Linear Linear Bent Bent Linear Linear Linear Linear Bent Bent Bent Linear Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Linear Linear Linear Linear Linear Linear Linear

~n1 — 1080.0 — 715 472 (1337)b 1333.37 1314.49 — 1253,8 1187 658 — — 369.1 — 1327 1325 1396 1147 985.1 992.0 831.7 925/915 630.7 790 944.8 1055 795.7 504 532.9 865 710 768.0 757.0 758.7 787.4 926 — — 971 1030.2 732.5 (765.4) 880 — 1078 (1008.3) 987.2

~n2 243.4 390.7 — (120) — 667 670.12 667.69 462.6 714.2 745 397 — — 313.1 — 806 752 570 517 495.5 372.5 294 461 296.4 400 448.7 520 317.0 197 — 375 320 — — — — — — — — 380 140 — 140 — 611 — —

~n3 730 332.8 1276 994 809.4 2349 2386.69 2378.45 1421.7 1658,3 1329 1533 1207.1 1159.4 1301.9 863.1 1286 1610 2360 1351 1041.9 965.6 848.3 821 670.8 840 1107.6 1295/1230 845.2 587 628.0 932 — 800.3 736.8 718.8 735.3 902.2 730.4 934.8 912 983.8 910 776.1 950 794.3 1702 1051.0 1040.7

Ref. 188 326 327 328 329 330 331 331 332 332 333 334 335 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 347 354 355 355 355 356 355 357 358 359 359a 360 361 362 363 364 364

TRIATOMIC MOLECULES

165

TABLE 2.2b. (Continued (Continued ) Compounda 11

3

(P B P) (s) (P Be P)4 (s) P Ga P (P Zn P)4 (s) (P Cd P)4 (s) (P Hg P)4 (s) (P Be P)4 (s) (As Be As)4 (s) (Sb Be Sb)4 (s) S Ti S S Zr S S Hf S a

Structure

~n1

~n2

Linear Linear Bent Linear Linear Linear Linear Linear Linear Bent Bent Bent

490 379 322 344 317 335 379 222 — 533.5 502.9 492.2

379/385 340 — 47 81 71 340 310 255 — — —

~n3 1076 860 220.9 415 355 341 860 775 700 577.8 504.6 483.2

Ref. 365 366 367 368 358 368 366 366 366 369 369 369

All data were obtained in inert gas matrices except for those denoted as g (gas) or s (solid). Fermi resonance with 2 n2 (Sec. 1.11).

b

bent, as indicated by the appearance of a strong d(ONO) band at 534 cm1 in the Raman spectrum [370]. Metal dioxides produced by sputtering techniques [371] in inert gas matrices take linear or bent OMO structures. Metal dinitrides produced by the same technique also take linear NMN (M ¼ U [372] and Pu [373]) or bent NMN (M ¼ Th [374]) structures. These structures are markedly different from those of molecular oxygen and nitrogen complexes of various metals produced by the conventional matrix cocondensation technique (Chapter 3 in Part B). The NpO22þ, PuO22þ, and AmO22þ ions are similar to the UO22þ ion it that they are linear and symmetric. They exhibit the n3 bands at 969, 962, and 939 cm1, respectively, in IR spectra [375]. The relationship between the U¼O stretching force constant and the U¼O distance was derived by Jones [376]. According to Bartlett and Cooney [377] the U¼O distance R (in pm ¼ 102 A) is given by R ¼ 9141ð~ n3 Þð2=3Þ þ 80:4 R ¼ 10650ð~ n1 Þð2=3Þ þ 57:5 where n is in units of cm1. Mizuoka and Ikeda [378] observed that the n3 of the (UIVO2)2þ ion at 895 cm1 in DMSO (dimethyl sulfoxide) solution is shifted to 770 cm1 when reduced electrochemically to the (UVO2)þ ion. Vibrational spectra of coordination compounds containing dioxo(O¼M¼O) groups are discussed in Chapter 3 (in Part B). AT-shaped structure of C2v symmetry was proposed for SiC2 in an Ar matrix at 8K [379]. Assignments of IR spectra of BC2 [380] and KO2 [381] prepared in Ar matrices were based on triangular structures of C2v symmetry. The linear CS2 molecule in a N2 matrix is converted to a cyclic structure [881.3 cm1(A1) and 520.9 cm1 (B2)] on photo irradiation (193 nm) [382].

166

APPLICATIONS IN INORGANIC CHEMISTRY

2.2.3. Triatomic Halogeno Compounds Table 2.2c lists the vibrational frequencies of triatomic halogeno compounds. The resonance Raman spectrum of the I3 ion gives a series of overtones of the n1 vibration [408,409]. The resonance Raman spectra of the I2Br and IBr2 ions and their complexes with amylose have been studied [410]. The same table also lists the vibrational frequencies of XHY-type (X,Y: halogens) compounds. All these species are linear except the ClHCl ion, which was found to be bent by an inelastic neutron scattering (INS) and Raman spectral study [402].

TABLE 2.2c. Vibrational Frequencies of Triatomic Halogeno Compounds (cm1) Compounda þ

FCIF FCIF (m) FCIF (s) FCICIþ (s) FBrFþ (s) FBrF (s) FIF (s) CIFCIþ (s) Cl 3 ðmÞ CIBrCl (sl) ClICIþ (s) ClICI (sl) ClIIþ (s) Brþ 3 ðsÞ Br 3 ðslÞ Br3 (m) BrIBrþ (s) BrII (s) BrIIþ (s) I 3 ðcÞ HFHþ (s) FHF (s) FHF (m) FHF (g) CIHF (s) CIHCI (s) CIHBr (s) ClHI (s) BrHF (s) BrHBr (s) BrHBr (m) IHF (s) IHI (m) F 3 ðmÞ a

Structureb b b l b b l l b b l b l b b l l l l b l b l l l l b l l l l l l l l

n~1c 807 500 510/478 744 706 460 445 529 253 278 371 269 360 293 164 190 256 117 258 116 2970 596/603 — 583.1 275 199 — — 220 126 168 180 (120.7) 461

n~2

n~3c

Ref.

387 242 — 299/293 362 236 206 293/258 — 135 147 127 (126) 124 53 — 124 84 — 56/42 1680 1233 1217 1286.0 863/823 660/602 508 485 740 1038 — 635 — —

830 578/570 636 535/528 702 450 462 586/593 340 225 364 226 197 297 191 — 256 168 198 125 3080 1450 1364 1331.2 2710 1670 1705 2200 2900 1420 670 3145 682.1 550

383 384 385 386 387 388 388a 389 390 391 392 391 393 394 391 395 393 396 393 396a 397 398 399 400 401 402 403 403 401 404 405 401 406 407

m ¼ insert gas matrix; sl ¼ solution; s ¼ solid. b ¼ bent; l ¼ linear. c For XYZ type, n1 and n3 correspond to n (XY) and n (YZ), respectively. b

TRIATOMIC MOLECULES

167

Ault and coworkers have utilized salt-molecule reactions to produce a number of novel triatomic and other anions in isolated environments [411]. For example, the linear symmetric [FHF] ion was produced in Ar matrices via codeposition of CsF vapor with HF gas diluted in Ar [399]: Csþ þ F þ HF ! Csþ ½F H F The reaction of CsF vapor with HC1 gas diluted in Ar yields a mixture of two types of the [FHC1] anions; in type I, the H atom is not equidistant from the F and Cl atoms, and its n3 [mainly n(HF)] is observed near 2500 cm1, whereas in type II, the H atom is symmetrically located and its n3, is at 933 cm1 [412].

The yield of each type is determined by the location of the cation in the ion pair, and the type I/type II ratio varies, depending on the cation used; smaller alkali metal cations favor type II, while larger cations prefer type I. These classifications were originally made by Evans and Lo, who observed crystalline HCl2 salts of types I and II depending on the cation employed [413]. 2.2.4. Bent XH2 Molecules Table 2.2d lists the vibrational frequencies of bent XH2-type molecules. The XH stretching frequencies are lower and the XH2 bending frequencies are higher in condensed phases than in the vapor phase because of hydrogen bonding in the former. This trend is also seen for H2O and H2S dissolved in organic solvents such as pyridine and dioxane [442,443]. In both liquid water and ice, H2O molecules interact extensively via OHO bonds. However, there are marked differences between the two phases. In the latter, H2O molecules are tetrahedrally hydrogen-bonded, and this local structure is repeated throughout the crystal. In liquid water, however, the OHO bond distance and angle vary locally, and the bond is sometimes broken. Thus, its vibrations cannot be described simply by using the three normal modes of the isolated H2O molecule. According to Walrafen et al. [444] an isosbestic point exists at 3403 cm1 in the Raman spectrum of liquid water obtained as a function of temperature, and the bands above and below this frequency are mainly due to non-hydrogen-bonded and hydrogenbonded species, respectively. In addition, liquid water exhibits librational and restricted translational modes that correspond to rotational and translational motions of the isolated molecule, respectively. The librations yield a broad contour at 1000– 330 cm1, while the restricted translations appear at 170 and 60 cm1 [445]. For more details, see the review by Walrafen [446].

168

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.2d. Vibrational Frequencies of Bent XH2-Type Molecules (cm1) Molecule

State

~n1

~n2

~n3

Ref.

CaH2 TiH2 ZrH2 VH2 CrH2 MoH2 NiH2 AlH2 GaH2 InH2 SiH2 GeH2 NH2

Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Gas Matrix Matrix Surface Solid Solid Matrix Gas Liquid Solid Gas Solid Gas Solid Gas Gas Solid Gas Matrix Solid Gas Solid Gas Gas Gas Gas

1267.0 1483.2 — 1532.4 1651.3 1752.7 2007 1766 1727.2 1548.6 1995.03 1887 — 3290 3270 2230 3182.7 3657 3450 3400 2727 2416 2671 2495, 2336 2657 — 1988 2615 2619.5 2532, 2523 1892 1843, 1835 2345 1687 1692.14 2065.27

— 496 625.03 529 — — — 760 740.1 607.4 998.62 920 1499 1610 1556 1080 1401.7 1595 1640 1620 1402 1490 1178 1210 1169 996 1023 1183 — 1186, 1171 934 857, 847 1034 741 — —

1192.0 1435.5 1861.0 1508.3 1614.9 1727.4 1969 1799 1799.5 1615.5 1992.82 1864 3220 3380 3323 2160 3219.5 3756 3615 3220 3707 3275 2788 2432 2764 2370 2104 2627 2632.6 2544 2000 1854 2358 1697 2351.19 2072.11

414 415 414 415 416 416 417 418 419 419 420 421 422 423 424 425 426,427 428 429 430 428 430 428 430 431 432 433 434 435 436 437 436 438 438,439 440 441

[NH2] [PH2] [OH2] þ 16 OH2 16

OHDa

16

OD2

18

OD2 OT2b SH2

SD2 SeH2 SeD2 80 SeHDa 130 TeH2 a b

Here n1 and n3 denote n (OD) and n (OH), respectively. T ¼ 3H (tritium).

Hornig et al. [430] assigned the spectrum of ice I as shown in Table 2.2d. They also assigned some librational and translational modes. Bertie and Whalley [447] studied the vibrational spectra of ice in other phases, and found the spectra to be consistent with reported crystal structures in each phase. For crystal water and aquo complexes, see Sec. 3.8. The vibrational frequencies of H2O, D2O, HDO, and their dimers in argon matrices have been reported [448]. Infrared studies on (H2O)n (16 O=18 O) (where n ¼ 1,2,>3) in N2 matrices [449] and (H2O)n (n ¼ 2–6) formed in droplets of liquid helium have been reported [450]. Devlin at al. [451] measured the IR spectra of ice consisting of large (H2O)n

169

TRIATOMIC MOLECULES

(n ¼ 100–64,000) clusters and studied the relationship between the frequency and intensity of the bending mode and the structure of the cluster. 2.2.5. X3-Type Molecules X3-type halogeno molecules are listed in Table 2.2c. Table 2.2e lists vibrational frequencies of other X3-type molecules. The O3 and O3 ion are bent (bond angle 120–110 ), and the S3, S3, and S32 ions arc all bent (103–115 ), with frequencies decreasing in this order. Detailed assignments of the IR spectra were made for six isotopomers of O3 containing 16 O and 18 O in liquid oxygen solution at 77 K [454]. The N6 radical anion was produced by the reaction of the N3 ion with N3 . The most probable structure, on the basis of theoretical calculations is a rectangular cyclic structure of D2h symmetry with two long NN bonds that connect two azidyl monomeric units via the terminal nitrogen atoms. This is also supported by the observation of an IR band at 1842 cml, which is intermediate between N3 and N6þ [469]. Figure 2.5 illustrates three normal modes of vibration of a triangular X3- type molecules of D3h symmetry. The n1 is a symmetric stretch (A10 , Raman-active) while a degenerate pair of n2a and n2b (E 0 , IR/Raman-active) involves stretching as well as bending coordinates (Sec. 1.9).

TABLE 2.2e. Vibrational Frequencies of X3-Type Molecules (cm1) Bent (C2v) and Linear (D¥h) Molecules Molecule

State

Structure

~n1

~n2

~n3

Ref.

(O3) 16 O3 (S3)2 S3 S3 N3 N3þ Rh3 Mo3

Solid Matrix Solid Solution Gas Solution Matrix Matrix Matrix

Bent Bent Bent Bent Bent Linear Linear Bent Bent

1020 1104.3 458 535 581 1343 1287 322.4 386.8

620 699.8 229 235.5 281 637.5 473.7 248 224.5

816 1039.6 479 571 680 2048 1657.5 259 236.2

452 453,454 455 456 457,458 459 460 461 462

Triangular (D3h) Molecules

Molecule

State

~n1 (A0 1)

~n2 (E 0 )

Ref.

Rb3 Cs3 Cr3 Fe3 Ag3 Ru3a Ta3a Nb3 Au3

Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix

53.9 39.5 432.2 249 119 303.4 251.2 334.9 172

38.3 24.4 302.0 — — — — 227.4 118

463 463 464 465 465 466 466 467 468

a

D3h or C2v.

170

APPLICATIONS IN INORGANIC CHEMISTRY

ν1(A1′)

ν2b(E ′)

ν2a(E ′)

Fig. 2.5. Normal modes of vibration of triangular X3 molecules.

2.2.6. XXY- and XYZ- Type Molecules Table 2.2f lists vibrational frequencies of XXY-type molecules. Tables 2.2g and 2.2h list those of linear and bent XYZ-type molecules, respectively. It should be noted that the description of vibrational modes such as n(XX), n(XY), and n(YZ) is only qualitative. Most of these frequencies were measured in inert gas matrices. Some of these vibrations split into two because of Fermi resonance, matrix effect, or crystal, field effects. The linear HArF molecule listed in Table 2.2g was obtained by photodissociation of HF in Ar matrices followed by annealing to increase the HArF concentration below 20 K. Figure 2.6 shows the IR spectra of H 40 ArF, D 40 ArF, and H 36 ArF in Ar matrices

TABLE 2.2f. Vibrational Frequencies of XXY-Type Molecules (cm1) Moleculea 6

6

Li Li O CCO (C C O) (N N H)þ (g) NNC N N O (g) (N N F)þ (s) O O H (g) O O D (g) OOF O O Cl O O Br OOS SSH S S O (g) 35 Cl 35 Cl O a

Structure

~n1b

~n2c

Linear Linear Linear Linear Linear Linear Linear Bent Bent Bent Bent Bent Bent Bent Bent Bent

— 1074 2082 2257.9 1235 2277 2371 1097.6 1121.5 1500 1477.8 1484.0 1005 595 679.1 374.2

118 381 656 686.8 394 596.5 391 1391.8 1020.2 586.4 432.4 — — 903 382 240.2

~n3d 1028.5 1978 1185 3334.0 2824 1300.3 1057 3436.2 2549.2 376.0 214.9 — 739.9 2460/2463 1166.5 961.9

All data were obtained in inert gas matrices except for those indicated as g (gas) or s (solid). n (XX). c d (XXY). d n (XY). b

Ref. 470 471 472 473 474 475 475a 476 476 477 478 479 480 481 482 483

TRIATOMIC MOLECULES

171

TABLE 2.2g. Vibrational Frequencies of Linear XYZ-Type Molecules (cm1) Molecule HCN DCN TCN HNC H 11 B 79 Br H 11 B O D 11 B O (H C S)þ H Si N H 40 Ar F H Kr F 19 11 32 F B S FCN FPO F P 32 S 35 Cl 12 C 14 N Cl C N 35 Cl 11 B 32 S 35 Cl 11 B O (ClCO)þ Cl C S 35 Cl C P Br C N ICN Li N C Na C N (C N O) NCO (N C O) (N C 32 S) (N C Se) (N C Te) (N S O) (P C S) PNO (O C S)þ (O C S) O C Te SCO S C Se S C Te Al O Si WCH Co C O 6

a b

State

~n1 (XY)

~n2 (d)

~n3 (YZ)

Ref.

Gas Matrix Gas Gas Matrix Gas Gas Matrix Gas Gas Matrix Matrix Matrix Gas Gas Gas Gas Gas Matrix Gas Gas Solid Matrix Gas Gas Matrix Gas Matrix Matrix Solid Matrix Solution Solid Solid Solid Solid Solid Matrix Matrix Matrix Matrix Gas Gas Solution Matrix Gas Matrix

3311 3306 2630 2460 3620 3813.4 2849,4 (2849) 2253.5 3141,7 2152.2 1969.5 1950 — 1077 819.6 803.3 747.8a 718 529.9 — 802.5 632.1 573.9 574 575 485 722.9 382 2096 1275 2155 2053 2070 2073 1283 1762 865.2 2071.1 1646.4 1965,3 859 1435 1347 534.9 1006 579.2

712 721 569 513 477 448.5 — 754 608.4 766.5 — 687.0 650 (457) 449 — (313.6) 381.9a 384/387b — — — — 303.4 342.5 349/351 304 121.7 170 471 487 630 486/471 424/416 366 528 — — — — — 520 (355) (377) — 660 424.9

2097 — 1925 1724 2029 2049.8 937.6 1817 1647.7 — 1162.2 435,7 415 1644.4 2290 1297.5 726.3 2249.3a — 1407.6 1972.2 2256 1189.3 1477.3 2200 — 2188 2080.5 2044 1106 1922 1282/1202b 748 558 450 999 747 1754.7 702.5 718.2 — 2062 506 423 1028.2 — 1957.5

484 485 484 485a 486 487 488 489 490 491 492 493, 493a 494 495 496 497 498 499 500 501 502 503 504 505 506 500 507 508 509 510 511 512 513 514 515 516 517 518 519 519 520 521 522 522 523 524 525

Corrected for anharmonicity. Fermi resonance between n3 and 2n2.

172

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.2h. Vibrational Frequencies of Bent XYZ-Type Molecules (cm1) Molecule

Statea

~n1 (XY)

HCO H 12 C F HNO H 14 N F HON HOF H O Cl H O Br HOI 7 16 28 Li O Si NOF NSF N S Cl ONF O N 35 Cl O N Br ONI OPF

Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Gas Gas Gas Gas Gas Matrix Gas Matrix Gas Matrix Matrix Gas Matrix Matrix Gas Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix

2483 — 3450 — 3467.2 3537.1 3578 3589 3597 668.4 1886.2 1372 1325 1843.5 1799.7 1799.0 1809 1297.5 1292,2 1272.8 1855 1880 1037.7 720.2 712 765.4 917 742 509.4 1327.3 328 295 — 1312.9 434 229.8

O P Cl OCF O C Cl O Cl F SPF S P Br FNO F N Cl 35 Cl C F 35 Cl 28 Si 16 O 35 Cl 32 S N Cl Sn Br Cl Pb Br Br O N Br 32 S N Br S F Pd 14 N 16 O a

~n2 (d) 1087 1405 1110 1432 — 1359.0 1239 1164 1103 249 734.9 366 273 765.8 595.8 542,0 470 — 416,0 491.6 626 281 310 313.6 — 519.6 (341) — — 403.8 — — 350 346.1 — 522.1/510.3

~n3 (YZ)

Ref.

1863 1181.5 1563 1000 1095.6 886.0 728 626 575 998.5 492.2 640 414 519.9 331.9 266.4 216 819.6 811.4 — 1018 570 596.9 791.4 372 1844.1 720 1146 1160.9 267.4 228 200 1820,0 226.2 765 1661.3

526 527 528 529 530 531 532 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 555 556 554 557 558

Matrix: inert gas matrix.

at 7.5 K obtained by Khriachtchev et al. [493, 493a]. Good agreement was obtained between the observed and theoretical. frequencies. This work was extended to HKrF [494]. Lapinski et al. [559] measured the IR spectra of 12 isotopomers of linear NNO involving 14 N=15 N and 16 O=17 O=18 O in N2 matrices, and assigned their fundamental, overtone, and combination bands on the basis of observed isotope shifts and DFT calculations. These 12 isotopomer bands in the n1 (NN stretch) region are marked by arrows in Fig. 2.7.

PYRAMIDAL FOUR-ATOM MOLECULES

Fig. 2.6. IR spectra of HArF in solid Ar showing the effects of [493,493a].

40

173

Ar=36 Ar and H/D substitutions

2.3. PYRAMIDAL FOUR-ATOM MOLECULES 2.3.1. XY3 Molecules (C3v) The four normal modes of vibrations of a pyramidal XY3 molecule are shown in Fig. 2.8. All four vibrations are both infrared- and Raman-active. The G and F matrix elements of the pyramidal XY3 molecule are given in Appendix VII. Table 2.3a lists the fundamental frequencies of XH3-type molecules. Several bands marked by an asterisk are split into two by inversion doubling. As is shown in Fig. 2.9, two configurations of the XH3 molecule are equally probable. If the potential barrier between them is small, the molecule may resonate between the two structures. As a result, each vibrational level splits into two levels (positive and negative) [560]. Transitions between levels of different signs are allowed in the infrared spectrum, whereas those between levels of the same sign are allowed in the Raman spectrum. The transition between the two levels at v ¼ 0 is also observed in the microwave region (~ n ¼ 0:79 cm1 ).

174

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.7. IR spectra (n1 region) of NNO in N2 matrices at 10 K. The labeling on the left indicates the isotopic composition of the mixture used in preparing the matrix. For example, 4568 denotes a mixture composed of equal fraction of 14N,15N,16O, 16O, and 18O. Thus, the peak due to 14 14 N, N,16O is indicated by the number 446 (bottom trace). This species exhibits the bands at 2235.6, 588.0, and 1291.2 cm1 in N2 matrices [559].

Splittings due to inversion doubling were also observed for the n4 of 14 NH3 [578], and and n1 and n3 of ND3 [579]. Optical isomers may be separated if the three Y groups are not identical and the inversion barrier is sufficiently high. As is seen in Table 2.3a n1 and n3 overlap or are close in most compounds. The presence of the hydronium (H3Oþ) ion in hydrated acids has been confirmed by observing its characteristic frequencies. For example, it was shown from infrared spectra that H2PtCl62H2O should be formulated as (H3O)2 [PtCl6] [580]. For normal

X Y1

Y3 Y2 ν1(A1) νs(XY)

ν2(A1) δs(YXY)

ν3a(E) νd(XY)

ν4a(E) δd(YXY)

Fig. 2.8. Normal modes of vibration of pyramidal XY3 molecules.

175

PYRAMIDAL FOUR-ATOM MOLECULES

TABLE 2.3a. Vibrational Frequencies of Pyramidal XH3 Molecules (cm1)

Solid

2318

15

Gas

3335

NT3

Gas

2016

PH3

Gas

2327

PD3 AsH3 AsD3 SbH3 SbD3 BiH3 ½OH3 þ SbCl 6 ½OH3 þ ClO 4 þ ½OH3 NO3 [OH3]þ

Gas Gas Gas Gas Gas Gas Sol’n Solid Solid Liquid SO2

1694 2122 1534 1891 1359 1733.3 3560 3285 2780 3385

~n2 (A1) 931:6 a 968:1 1060 748:6 a 749:0 815 926 a 961 647 990 a 992 730 906 660 782 561 726.7 1095 1175 1135 —

[OD3]þ

Liquid SO2

2490

—

[SD3]þ [SeH3]þ [CH3] Naþ SiH3 [GeH3]

Gas Solid Matrix Gas Liquid NH3

1827.2 2302 2760 — 1740

— 936 1092 727.9 809

Molecule

State

NH3

Gas

ND3

NH3

a

~n1 (A1)

Solid

3335:9 3337:5 3223

Gas

2419

a

~n3 (E)

~n4 (E)

Ref.

3414

1627.5

560

3378

1646

561

2555

1191.0

560

2500

1196

561

3335

1625

562

2163

1000

563

2421

1121

564

(1698) 2185 — 1894 1362 1734.5 3510 3100 2780 3470 3400 2660 2580 1838.6 2320 2805 — —

806 1005 714 831 593 751.2 1600 1577 1680 1700 1635 1255

564 564,565 564,566 567 567 568 569 570 671 572

— 1057 1384 — 886

573 574 575 576 577

572

Inversion doubling.

coordinate analysis of pyramidal XH3 molecules, see Refs. [581–583]. Infrared spectra of hydrated proton complexes Hþ(H2O)n (n ¼ 2–11), in water clusters were measured as Ar complexes at low temperatures, and band assignments were made based on their minimum-energy structures obtained by DFT calculation [583a]. Reaction of CH3I with Na (or K) atoms in N2 matrices at 20 K [575] produces an ion-pair complex CH3 Naþ (or CH3 Kþ), of C3v symmetry. The n(NaCH3) and n(KCH3) vibrations were observed at 298 and 280 cm1, respectively. Vibrational spectra of the (NH3)2 dimer in Ar matrices show that two NH3 molecules are bonded via a very weak hydrogen bond [584]. Table 2.3b lists the vibrational frequencies of pyramidal XY3 halogeno compounds. Clark and Rippon [601] have measured the Raman spectra of a number of these compounds in the gaseous phase. The compounds show a n1 > n3 and n2 > n4 trend, whereas the opposite trend holds for the neutral XH3 molecules listed in Table 2.3a. In some cases, the two stretching frequencies (n1 and n3) are too close to be distinguished empirically. This is also true for the two bending bands (n2 and n4).

176

APPLICATIONS IN INORGANIC CHEMISTRY

Y

Y Y

Y

X

X

Y

Y _ +

IR

IR

R

υ=1

R

_ + υ=0 Fig. 2.9. Inversion doubling of pyramidal XY3 molecules.

Figure 2.10 shows the Raman spectra of the SnX3 ions (X ¼ Cl, Br, and I) in solution obtained by Taylor [594]. In crystalline [As(Ph)4][SnX3], the symmetry of the SnX3 ion is lowered to Cs so that n3 and n4 each split into two bands [592]. Normal coordinate analyses on the SnX3 ion (X ¼ F,Cl,Br,I) [592] have been carried out. Table 2.3c lists the vibrational frequencies of pyramidal XO3-type compounds. Rocchiciolli [629] has measured the infrared spectra of a number of sulfites, selenites, chlorates, and bromates. Again, the n3 and n4 vibrations may split into two bands because of lowering of symmetry in the crystalline state. Although n2 > n4 holds in all cases, the order of two stretching frequencies (n1 and n3) depends on the nature of the central metal. It is to be noted that the SO3 and SO32 anions are pyramidal, while the SO3 molecule is planar (Table 2.4b). Figure 2.11 shows the ER spectra of KClO3 and KIO3 obtained in the crystalline state. 2.3.2. ZXY2 (Cs) and ZXYW (C1) Molecules Substitution of one of the Yatoms of a pyramidal XY3 molecule by a Z atom lowers the symmetry from C3v to Cs Then the degenerate vibrations split into two bands, and all six vibrations become infrared- and Raman-active. The relationship between C3v and Cs is shown in Table 2.3d. Table 2.3e lists the vibrational frequencies of pyramidal ZXY2 molecules. Vibrational spectra of the [SClnF3n]þ ions (n ¼ 0–3) [663] and PHnF3n (n ¼ 0–3) [664] have been assigned. The [XeO2F]þ ion takes a pseudo-tetrahedral structure owing to the presence of a sterically active lone-pair electron of the Xe atom [661]:

177

PYRAMIDAL FOUR-ATOM MOLECULES

Matrix cocondensation reactions of alkali halide molecules (MþX) with H2O diluted in Ar produce pyramidal [MOH2]þ ions that exhibit the n(OH2) and d(OH2) in the 3300–3100 and 700–400 cm1 regions, respectively [665]. These ions may serve as simple models of aquo complexes. TABLE 2.3b. Vibrational Frequencies of Pyramidal XY3 Halogeno Compounds (cm1) Molecule

State

~n1

~n2

~n3

~n4

Ref.

[InCl3] [InBr3] [Inl3] CF3 28 SiF3 [SiF3]

Solid Solid Solid Matrix Matrix Matrix

252 177 136 1084 832 770.3

102 74 78 703 406 401

97 46 40 600,500 290 268

585 585 585 586 587 588

Matrix Matrix Solid Solid Matrix Sol’n Sol’n Sol’n Matrix Sol’n Sol’n Sol’n Gas Sol’n Matrix Solid Gas Gas Gas Solid Gas Gas Melt Gas Matrix Gas Sol’n Gas Gas Gas Gas

470.2 388 303 520 509 297 205 162 456 249 176 137 1035 535 554.2 279 893.2 515.0 390.0 303 738.5 416.5 272 212.0 654 380.7 254 186.5 342 220 (162)

— — — 280 256 130 82 61 — — — — 649 347 365.2 146 486.5 258.3 159.9 111 336.8 192.5 128 89.6 259 150.8 101 74.0 123 77 59.7

185 149 110 1250 954 760.4 757.0 582.0 362 285 477 454 256 180 148 405 — 164 127 910 637 644 354 858.4 504.0 384.4 325 698.8 391.0 287 (201) 624 358.9 245 (147) 322 214 163.5

— — — 224 — 104 65 48 — — 58 3045 500 254 263 90 345.6 186.0 112 8 79 262.0 150.2 99 63.9 — 121 8 81 54.3 107 63 47.0

589 590 591 592 593 594 594 594 595 596 596 596 597 598 599 600 601,602 601 601 603 601,604 601,605 606 601 607 601 606,608 601,609 610 611 609

28 SiCl3 GeCl3 [GeCl3] [SnF3]

[SnCl3] [SnBr3] [SnI3] [PbF3] [PbCl3] [PbBr3] [PbI3] NF3 NCl3 14 NCl3 NI3 PF3 PCl3 PBr3 PI3 AsF3 AsCI3 AsBr3 AsI3 SbF3 SbCI3 SbBr3 SbI3 BiCI3 BiBr3 BiI3

(continued)

178

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.3b. (Continued (Continued ) Molecule þ

[SF3] [SCl3]þ [SBr3]þ [SeF3]þ [SeCl3]þ [SeBr3]þ [TeCl3]þ [TeBr3]þ [MnBr3]þ FeCl3 YCl3 CeCl3 NdCl3 SmCl3 GdCl3 DyCl3

State

~n1

~n2

~n3

~n4

Ref.

Melt Solid Solid Melt Sol’n Sol’n Sol’n Sol’n Solid Matrix Gas Gas Gas Gas Gas Gas

943 498 375 781 430 291 362 265 280 363.0 378 328 333 338 343 346

690 276 175 381 206 138 186 112 110 68.7 78 58 60 61 64 65

922 533, 521 429, 414 743 415 298 347 266 150 460.2 359 314 320 324 330 334

356 215, 208 128 275 172 108 (150) 92 80 113.8 58.6 44 45 46 46 47

612 613 614 612 615 606 616 606 617 618 619 620 620 620 620 620

Fig. 2.10. Raman spectra of the SnX3 ions in diethyl ether extracts from SnX2 in HX solutions. The arrows denote polarized bands [594].

PYRAMIDAL FOUR-ATOM MOLECULES

179

TABLE 2.3c. Vibrational Frequencies of Pyramidal XO3 Molecules (cm1) Molecule 2

[SO3] [SO3] [SeO3]2 [TeO3]2 [CIO3] 35

Cl 16 O3 [BrO3] [IO3] XeO3

State

~n1 (A1)

~n2 (A1)

Sol’n Matrix Sol’n Sol’n Sol’n Solid Matrix Sol’n Solid Sol’n Solid Sol’n

967 1000 807 758 933 939 905 805 810 805 796 780

620 604 432 364 608 614 567 418 428 358 348 344

~n3 (E) 933 1175 737 703 977 971 1081 805 790 775 745 833

~n4 (E) 469 511 374 326 477 489 476 358 361 320 306 317

Ref. 621 622 623 623 624 625 626 624,627 625 624 625 628

The ZXYW-type molecule belongs to the C1 point group, and all six vibrations are infrared- and Raman-active. The vibrational spectra of OSClBr [666] and [XSnYZ] (X,Y,Z: a halogen) have been reported [667]. For example, the Raman spectrum of the [ClSnBrl] ion in the solid state exhibits the n(SnCl), n (Sn-Br), and n (SnI) at 275, 193, and 155 cm1, respectively.

Fig. 2.11. IR spectra of KClO3 (solid line) and KIO3 (dotted line).

TABLE 2.3d. Relationship between C3v and Cs

180

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.3e. Vibrational Frequencies of Pyramidal ZXY2-Type Metal Halides (cm1) ~ n1 (A0 ) n (XZ)

~n2 (A0 ) n s (XY)

~n3 (A0 ) ds (YXZ)

~n4 (A0 ) d (YXY)

~n5 (A00 ) n a (YX)

~n6 (A00 ) da (YXZ)

Ref.

972 687 3234

500 — 1233

1307 1002 1564

888 695 3346

1424 1295 1241

630 631 632,633

ClNH2 DPH2 FPH2 C1PH2 FAsH2 ClPF2 BrPF2 FPCI2 FPBr2 IPF2 [FOH2]þ [BrOF2]þ OSF2 OSCl2 OSBr2 [FSH2] [FSO2]

3193 3279 891 679 674 — 795 511 649 545 459 838 824 375 865 1062 1323.5 1251 1121 853 594

3297 — 2304 2303 2108 864 858 525 398 851 3386 655 799.9 492 405 2500 1105

1534 892.9 1090 1096 984 (302) 391 203 123 413 1630 290 381.5 344 267 1181 500

634 635 636,637 637 638 639 639 639 639 640 641 642 643 644,645 646 647 648,649

288.2

260.1

650

[CISO2] [BrSO2] [ISO2] [FSeO2] OSeF2 OSeCl2 OSe (OH)2a FC1O2b FBrO2 [OCIF2]þ [FXeO2]þ

214 203 184 430 1049 995 831 1129.0 506 1333 594

172 115 (55) 283 362 161 430 550.0 305 513 281

526 530 530 324 282 279 336 407.1 394 384 321

3374 — 2310 2303 2117 852 849 525 423 846 3225 630 736.6 455 379 2447 1169 1183 1131.0 1309.6 1312 1308 1300 888 637 347 690 1292.2 953 695 931

1063 969.5 — — — 260 215 267 221 204 1261 315 397.3 284 223 1020 367

ClSO2

1056 1092.6 934 859 842 411 244 328 258 198 1067 365 531,9 194 120 987 378 393 497.7

(103) — — 238 253 255 364 360.1 271 404 249

651 651 651 652 653 654 655,656 657,658 659 660 661,662

HNF2 HNC12 FNH2

a b

1099.8 1098.2 1120 1117 1112 903 667 388 702 634.0 908 731 873

455.8

The OH group was assumed to be a single atom. Empirical assignment.

2.4. PLANAR FOUR-ATOM MOLECULES 2.4.1. XY3 Molecules (D3h) The four normal modes of vibration of planar XY3 molecules are shown in Fig. 2.12; n2, n3, and n4, are infrared-active, and n1, n3, and n4 are Raman-active. This case should be contrasted with pyramidal XY3 molecules, for which all four vibrations are both

PLANAR FOUR-ATOM MOLECULES

Y3

+

X

– + Y2

Y1 ν1(A′1) νs(XY)

181

+ ν2(A′′2) π (XY3)

ν3(E′)

ν4(E ′)

νd(XY)

δ d (YXY)

Fig. 2.12. Normal modes of vibration of planar XY3 molecules.

infrared- and Raman-active. Appendix VII lists the G and F matrix elements of the planar XY3 molecule. Table 2.4a lists the vibrational frequencies of planar XY3 molecules. As stated above, pyramidal and planar structures can be distinguished by the difference in selection rules. Thus, the pyramidal structure should exhibit two stretching bands (A1 and E) while the planar structure should show only one stretching band (E0 ) in IR spectra. The n3 and n4 frequencies are in the order AlH3 < GaH3 < InH3 [680]. MnF3 in the gaseous phase takes a planar structure of C2v symmetry due to the Jahn–Teller effect, and exhibits three stretching [759 (B2), 712 (A1), and 644 (A1)] and three bending [182 (B1), 186 (A1), and 180 (B2, calculated)] vibrations (all in cm1) [693]. Assignment of the IR spectrum of AlCl3NH3 has been based on normal coordinate analysis (staggered conformation, C3v symmetry) [694]. 11 BF3 in inert gas matrices takes a dimeric structure of C2h symmetry via two BFB bridges as supported by IR spectra [695]. Other dimeric species such as Al2F6 and Al2Cl6 are discussed in Sec. 2.10. Normal coordinate calculations on planar XY3 molecules have been carried out by many investigators [696,697]. Table 2.4b lists the vibrational frequencies of planar XO3-type compounds. Figure 2.13 shows the Raman spectra of KNO3 in the crystalline state and in aqueous solution. As discussed in Sec 1.26, the spectra of calcite and aragonite are markedly different because of the difference in crystal structure. More recent normal coordinate calculations [713] on the CO3 radical indicate a trans-Cs

182

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.4a. Vibrational Frequencies of Planar XY3 Molecules (cm1) Molecule 10

BH3 BF3 11 BC13 11 BBr3 11 BI3 A1H3 AID3 AlF3 AlCl3 AlBr3 AlI3 6 ½AlSb 3 GaH3 GaCI3 GaBr3 GaI3 [GaSb3]6 InH3 InCl3 InBr3 Inl3 [InAs3]6 TlBr3 CH3 [PS3] AuI3 [CdCl3] [CdBr3] [CdI3] [HgCl3] PrF3 VF3 CoF3 10

a

Statea

~n1 (A0 1)

~n2 (A00 2)

~n3 (E 0 )

~n4 (E 0 )

Ref.

Mat Gas Liquid Liquid Liquid Mat Mat Mat Mat Gas Gas Solid Mat Mat Gas Gas Solid Mat Mat Gas Gas Solid Sol’n Mat Melt Solid Sol’n Sol’n Sol’n Solid Mat Mat Mat

(2623) 888 472.7 278 192 — — — 393.5 228 156 132 1923.2 386.2 219, 237 147 128 1754.5 359.0 212 151 171 190 — 476 164 265 168 124 273 526,542 649 663

1132 719.3 — 374 — 697.8 513,9 286.2 — 107 77 181 717.4 — 95 63 110 613.2 — 74 56 99 125 730.3 540 — — — — 113 86 — —

2820 1505.8 950.7 802 691.8 1882.8 1376.5 909.4 618.8 450–500 370–410 293,318 — 469.3 — 275 192, 210 — 400.5 280 200–230 197, 215 220 — 695 148 287 184 161 263 458 732 732.2

1610 481.1 253.7 150 101.0 783.4 568.4 276.9 150 93 64 — 758.7 132 84 50 — 607.8 119 62 44 — 51 1383 242 106 90 58 51 100 99 — 162

668 669,670 671,672 671,673 671,673 674 675 676 677 678 678 679 680 677 677,681 677,681 679 680 677,682 678 678 683 684 685 686 687 688 688 688 689 690,691 692 692

Mat: inert gas matrix.

structure (a) rather than a three-membered ring C2v structure (b) suggested originally [714].

Crystals of LiNO3, NaNO3, and KNO3 assume the calcite structure (Sec. 1.31). Nakagawa and Walter [705] carried out normal coordinate analyses on the whole Bravais lattices of these crystals.

183

PLANAR FOUR-ATOM MOLECULES

TABLE 2.4b. Vibrational Frequencies of Planar XO3 and Related Compounds in the Crystalline State (cm1) Compound 10

La[ BO3 ] H3[BO3] Ca[CO3] (calcite) Ca[CO3] (aragonite) Ba[CS3] Ba[CSe3) Na[NO3] K[NO3] NO3 SO3a SeO3a a b

IR IR IR R IR R R IR IR R IR R IR IR R R

~n1 (A0 1)

~n2 (A00 2)

~n3 (E 0 )

~n4 (E 0 )

Ref.

939 1060 — 1087 1080 1084 510 290 — 1068 — 1049 1060 (1068) 1065 923

740.5 668, 648 879 — 866 852 516b 420 831 — 828 — 762 495 497.5 —

1330.0 1490–1428 1429–1492 1432 1504,1492 1460 920 802 1405 1385 1370 1390 1480 1391 1390 1219

606.2 545 706 714 711,706 704 314 185 692 724 695 716 380 529 530.2 305

698,699 706 701 701 701 701 702,703 703,704 705 705 705 705 706–708 709,710 711 712

Gaseous state. Infrared.

M€ uller et al. [715] have shown that the half-width of the n3 band of the NO3 ion at 1384 cm1 in KBr pellets is decreased by a factor of >5 when the ion is trapped in the IV V cavity of ionic carcerand compounds such as K10[HV12 V 6 O44 ðNO3 Þ2 ]14.5H2O. UV resonance Raman spectra of the NO3 ion in H2O and ethylene glycol are dominated by the n1, n3, and their overtones and combination bands. The large splitting of the n3

Fig. 2.13. Raman spectra of KNO3 in (a) the solid state and (b) aqueous solution.

184

APPLICATIONS IN INORGANIC CHEMISTRY

vibration ( 60 cm1) suggests the planar-to-pyramidal conversion by electronic excitation [716]. The IR spectrum of KNO3 in Ar matrices suggest bidentate coordination of the Kþ ion to two oxygen atoms (C2v symmetry) [717]. 28 Si 16 O3 in Ar matrices also takes a C2v structure with one shorter and two longer SiO bonds [718]. The spectra of anhydrous metal nitrates such as Zn(NO3)2 [719] and UO2(NO3)2 [720] can be interpreted in terms of C2v symmetry since the NO3 group is covalently bonded to the metal. Raman spectra of metal nitrates in the molten state [721,722] indicate that the degeneracy of the n3 vibration is lost and the Raman-inactive v2 vibration appears. Apparently, the D3h selection rule is violated because of cation– anion interaction. Like CO32 and NO3 ions, CS32 and CSe32 ions act as chelating ligands. For normal coordinate analyses of planar XO3 molecules, see Refs. 723 and 724. Some XY3-type halides take the unusual T-shaped structure of C2v symmetry shown below. This geometry is derived from a trigonal–bipyramidal structure in which two equatorial positions are occupied by two lone-pair electrons. Typical examples are ClF3 and BrF3. With the equatorial Y atom represented as Y0, the following assignments have been made for these molecules: [725] n(XY0 ), A1, 754 and 672; n(XY), B1 683.2 and 597; n(XY), A1, 523 and 547; d, A1 328 and 235; d, B1, 431 and 347; p, B2, 332 and 251.5 cm1 (for each mode, the former value is for ClF3 and the latter is for BrF3). The Raman spectrum of XeF4 in SbF5 exhibits two strong polarized bands at 643 and 584 cm1, which were assigned to the T-shaped XeF3þ ion in [XeF3][SbF6] [726]. XeOF2 assumes a similar T structure [727]. In an inert gas matrix, UO3 gives an infrared spectrum consistent with the T-shaped structure [728].

2.4.2. ZXY2 (C2v and ZXYW(Cs) Molecules If one of the Yatoms of a planar XY3 molecule is replaced by a Z atom, the symmetry is lowered to C2v. If two of the Yatoms are replaced by two different atoms, Wand Z, the symmetry is lowered to Cs. As a result, the selection rules are changed, as already shown in Table 1.18. In both cases, all six vibrations become active in infrared and Raman spectra. Table 2.4c lists the vibrational frequencies of planar ZXY2 and ZXYW molecules.

185

HBCl2 HAlCl2 HAlBr2 ClGaH2 HGaCl2 HGaBr2 ClInH2 HInCl2 [FCO2]a [(HO) CO2]a [HCO2] [(CH3) CO2]a Si¼CH2 Si¼CD2 O¼CF2 O¼CCl2 O¼CHF O¼CHI O¼CBr2 O¼CCIF O¼CBrCl

ZXYW (Cs)

ZXY2 (C2v)

XY3 (D3h)

786 473 454 620 464.3 445.5 415.7 469.7 — 835

1069 621

687 539 767.4 581.2 — — 512 667 547

2616 1966 1953 406.9 2015.3 1991.5 343.4 1846.9 883 960

2803 926

933 808 1930 1815.6 1837 1776.3 1828 1868 1828

2947 2185 965.6 568.3 2981 2930 425 776 517

1351 1413 — — 1243.7 837.4 1065 (561) 757 1095 806

1585 1556

1091 580 574 1978.1 437.3 333.0 1820.3 358.5 1749 1697

zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n4 ðB 2 Þ ~n2 ðA1 Þ n s ðXYÞ n a ðXYÞ 0 0 ~n4 ðA Þ ~n4 ðA Þ nðXZÞ nðXWÞ

~n6 (B1) p (ZXY2) ~n6 (A00 ) p

~n1 (A1) n (XZ) ~n1 (A0 ) n (XZ) 735 482 366 1964.6 414.3 287.5 1804.0 369.1 1316 1338

~n3 ðE Þ n d ðXYÞ

~n2 (A00 2) p (XY3)

~n1 (A0 1) n s (XY)

0

TABLE 2.4c. Vibrational Frequencies of Planar ZXY2 and ZXYW Molecules (cm1)

1273 1054 582.9 368.0 1343 1247.9 181 501 240

760 650

— — — 731.4 130 — 575.8 — — 712

0

263 216 619.9 441.8 663 — 350 415 372

1383 471

895 655 635 510.1 607.5 597.0 541.4 402.6 — 579

~n4 ðE Þ dd ðYXYÞ zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n5 ðB 2 Þ ~n3 ðA1 Þ ds ðZXYÞ da ðZXYÞ 0 0 ~n5 ðA Þ ~n3 ðA Þ dðZXYÞ dðZXWÞ

(continued)

740 740 741,742 743 744,745 746 747 747,747a 747

739 739

729 730 730 731 732,733 733 734,735 734 736,737 738

Ref.

186

1874 1063 1368 1120 1280 442 1240 638 739.1 580 902 555 370 1573 1862 1648 363.2

ZXYW (Cs)

O¼CBrF S¼CH2 S¼CF2 Se¼CCl2 Se¼CF2 [SeCS2] O¼SiCl2 S¼SiF2 S¼SiCl2 S¼GeCl2 (HO)¼NO2a FNO2 ClNO2 O¼NF2 [O¼NF2] þ [O¼NCl2] þ BrPO2

a

Only the ZXY2 skeletal vibrations are listed.

ZXY2 (C2v)

620 993 622 472 560 485 — 296 — — 767 740 652 (440) 715

~n6 (B1) p (ZXY2) ~n6 (A00 ) p

~n1 (A1) n (XZ) ~n1 (A0 ) n (XZ)

XY3 (D3h)

~n2 (A00 2) p (XY3)

~n1 (A0 1) n s (XY)

TABLE 2.4c. (Continued)

721 2970 787 500 710 433 501.1 996 — 404 1311 1308 1318 813 897 628 1146.5

1068 3028 1189 812 1208 925 637.5 969 609.9 440 1697 1800 1670 761 1163 735 1440.1

zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n4 ðB 2 Þ ~n2 ðA1 Þ n s ðXYÞ n a ðXYÞ 0 0 ~n4 ðA Þ ~n4 ðA Þ nðXZÞ nðXWÞ

0

~n3 ðE Þ n d ðXYÞ

398 1550 526 296 432 284 — 337 — — 660 810 787 705 569 220 513.5

0

335 1437 417 302 352 265 — 247 186 — 597 555 411 553 347 420 230

~n4 ðE Þ dd ðYXYÞ zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n5 ðB 2 Þ ~n3 ðA1 Þ ds ðZXYÞ da ðZXYÞ 0 0 ~n5 ðA Þ ~n3 ðA Þ dðZXYÞ dðZXWÞ 747 748 749,750 751 752 753 754 755 756 757 758 759,760 754,760 761 762,763 764 765

Ref.

OTHER FOUR-ATOM MOLECULES

187

Silanone produced by the silane–ozone photochemical reaction in Ar matrices gives rise to a band at 1202 cm1 in the IR spectrum that is assigned to the n(Si¼O) [766]:

The frequencies listed for the formate and acetate ions were obtained in aqueous solution. Vibrational spectra of metal salts of these ions are discussed in Sec. 3.8. Although not listed in this table, the IR spectra of binary mixed halides of boron [767] and aluminum [768] have also been reported. 2.5. OTHER FOUR-ATOM MOLECULES 2.5.1. X4 Molecules Tetrahedral X4-type molecules of Td symmetry exhibit three normal vibrations as shown in Fig. 2.14, and n1(A1) and n2(E) are Raman-active whereas n3(F2) is IR as well as Raman-active. Table 2.5a lists observed frequencies of tetrahedral X4 molecules. Comparison of vibrational frequencies of the X44 (X ¼ Si,Ge,Sn) series and the corresponding isoelectronic Y4 (Y ¼ P,As,Sb) series shows that the ratio, n(X44)/n (Y4) is approximately 0.77. On the basis of this criterion, Kliche et al. [770] predicted that the vibrational frequencies of the Pb44 ion are 115 (A1), 69 (E), and 93 (F2) cm1. On the other hand, the X4-type cation such as S42þ assumes a square–planar ring structure of D4h symmetry, and its 6 (3 4–6) vibrations are classified into A1g (R)þ B1g (R)þ B2g (R)þ B2u(inactive)þ Eu (IR). Assignments of these vibrations have been based on normal coordinate analysis [776a–776c]. Table 2.5b shows the normal modes and vibrational frequencies of square–planar X4 cations. 2.5.2. X2Y2 Molecules Molecules like O2H2 take the nonplanar C2 structure (twisted about the OO bond by 90 ), whereas N2F2 and [N2O2]2 exist in two forms: trans-planar (C2h) and

ν1(A1)

ν2(E)

ν3( F2)

Fig. 2.14. Normal modes of vibration of tetrahedral X4 molecules.

188

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.5a. Vibrational Frequencies of X4 Molecules (cm1) ~n1 (A1)

~n2 (E)

~n3 (F2)

Ref.

Si44

486

276

Sn44 P4 As4 Sb4 Bi4 Ta4

182 613.6 353 241.5 149.5 270.2

338 356 376 197 206 216 140 467.3 265.5 178.5 120.4 185.1

769,770

Ge44

282 305 , 154 165

Molecule

102 371.8 207.5 137.1 89.8 130.6

769

770 771,772 773 774 775 776

cis-planar (C2v). Figure 2.15 shows the six normal modes of vibration for these three structures. The selection rules for the C2v and C2 structures are different only in the n6 vibration, which is infrared-inactive and Raman depolarized in the planar model but infrared-active and Raman polarized in the nonplanar model. Table 2.5c lists the vibrational frequencies of X2Y2-type molecules. In N2 matrices, (NO)2 exists in three forms; cis, trans, and another form of uncertain structure. The cis form is most stable, and its ns and na are at 1870 and 1776 cm1 respectively [791]. The NO reacts with Lewis acids such as BF3 to form a red species at 77 K. Ohlsen and Laane [792] measured its resonance Raman spectrum and concluded that it is an asymmetric (NO)2 dimer having a cisplanar structure:

A possible mechanism for the formation of such a dimer in the presence of Lewis acids has been proposed. Although the N2O22 ion takes a cis or trans structure in the solid state, the N2O22 ion produced in Ar matrices exhibits an IR band at

TABLE 2.5b. Vibrational Frequencies of Square-Planar X4-Type Ions (cm1) Ion

S42þ

Se42þ

Te42þ

~ n1 (A1g)

583.6

321.3

219

~ n2 (B1g)

371.2

182.3

109

~ n3 (B2g)

598.1

312.3

219

~ n5 (Eu)

542

302

187

Source: Taken from Ref. 776b.

189

OTHER FOUR-ATOM MOLECULES

X1

X2 ν1

Y1

Y2

νs(XY)

ν2

ν3

ν(XX)

δs(YXX)

cis-Planar (C2υ)

A1 (p)

A1 (p)

A1 (p)

Nonplanar(C2)

A (p)

A (p)

A (p) +

ν4

ν5

νa(XY)

δa(XX)

–

–

+

ν6 δt (YXX)

cis-Planar

B2 (dp)

B2(dp)

A2(p)

Nonplanar

B (dp)

B (dp)

A (p)

Y2 X2

X1

Y1

ν1

ν2

ν3

νs(XY)

ν(XX)

δs (YXX)

Ag (p)

Ag (p)

trans–Planar (C2h) Ag (p)

+

ν4 Y1

νa(XY) Bu (dp)

– ν5 δa(YXX) Bu (dp)

+

ν6 ρl (XY) Au (dp)

Fig. 2.15. Normal modes of vibration of nonlinear X2Y2 molecules (p—polarized; dp—depolarized) [749].

1205 cm1, and its structure was suggested to be [793]

Diazene (N2H2) exists in both cis and trans forms:

190

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.5c. Vibrational Frequencies of X2Y2 Molecules (cm 1)a trans-X2Y2: cis-X2Y2: Twisted X2Y2:

C2h C2v C2

Assignment cis-[N2O2] trans-[N2O2]2 cis-N2F2 trans-N2F2 O2H2 O2D2 O2F2 S2H2 S2F2 S2Cl2 S2Br2 Se2Cl2 Se2Br2 2

Solid Soild Liquid Liquid Gas Gas Matrix Liquid Gas Gas Liquid Liquid Liquid

n1 Ag A1 A

n2 Ag A1 A

n3 Ag A1 A

n4 Bu B2 B

n5 Bu B2 B

n6 Au A2 A

ns (XY)

n (XY)

ds (YXX)

na (XY)

da (YXX)

p

Ref.

830 (1419) 896 1010 3607 2669 611 2509 717 466 365 367 265

1314 (1121) 1525 1522 1394 1029 1290 509 615 546 529 288 292

584 (396) 341 600 864 867 366 (868)b 320 202 172 130 107

1047 1031 952 990 3608 2661 624 2557c 681 457 351 367 265

330 371 737 423 1266 947 459 882 301 240 200 146 118

— 492 (550) 364 317 230 (202) — 183 92 66 87 50

777,778 779,780 781 782 783,784 783 785 786 787 788,789 790 789,790 790

a

Except for N2F2 and [N2O2]2 , all the molecules listed take the C2 or C2v structure. Solid. c Gas. b

and the observed frequencies are 3116, 3025, 1347, and 1304 cm 1 for the cis form, and 3109 and 1333 cm 1 for the trans form [794]. The IR spectra of (H2O2)2 dimer and its deuterated analogs have been measured in Ar matrices. Both experimental and theoretical studies suggest the cyclic structure shown below [794a]:

A strong vibrational coupling was observed between the two bonded OH stretching modes, and a relatively strong coupling was noted between the two HOO bending modes. The Raman spectra of Se2Cl2 and Se2Br2 in CS2 and CCl4 solutions indicate that the former takes the C2 structure, whereas the latter takes the C2v structure [795]. Although S2Cl2 does not dimerize at low temperatures, Se2Cl2 forms a dimer below 50 C. A new Raman band observed at 215 cm 1 was attributed to the dimer for which the following ring structure was proposed [796]:

OTHER FOUR-ATOM MOLECULES

191

Ketelaar et al. [797] found that the liquid mixture of CS2 and S2Br2, for example, exhibits IR spectra that show combination bands between the n3 of CS2 (1515 cm1) and a series of S2Br2 vibrations. Such simultaneous transitions seem to suggest the formation of an intermolecular complex, at least on the vibrational time scale. The Raman spectra of the [T12X2]2 ion (X ¼ Se or Te], which takes an almost planar diamond-shaped ring structure (C2v), have been assigned [798]. Si2H2 prepared in inert gas matrices assumes butterfly-like structure wth the SiSi bond located along the fold of the two SiHSi wing planes, and its IR spectrum was assigned on the basis of C2v symmetry [799]. The [C12O2]þ ion formed by the pp* charge transfer interaction of Cl2þ and O2 takes a planar trapezoid structure of C2v symmetry, and the v(OO) and v(ClCl) are at 1534 (A1) and 593 (A1), respectively, and v (CIO) are at 414 (B2) and 263 (A1) (all in cm1) [800]. Normal coordinate analyses of H2O2 [801] and S2X2 (X: a halogen) [802,790] have been carried out. Other (XY)2-type compounds include dimeric metal halides such as (LiF)2, which takes a planar ring structure (Sec. 2.1). 2.5.3. Planar WXYZ, XYZY, and XYYY Molecules (Cs) Planar four-atom molecules of the WXYZ, XYZY, and XYYY types have six normal modes of vibration, as shown in Fig. 2.16. All these vibrations are both infrared- and Raman-active. In HXYZ and HYZY molecules, the XYZ and YZY skeletons may be linear (HNCO, HSCN) or nonlinear (HONO, HNSO). In the latter case, the molecule may take a cis or trans structure. Table 2.5d lists the vibrational frequencies of molecules and ions belonging to these types. Normal coordinate analyses have been carried out for HN3 [831] and HONO [832]. Teles et al. [804] measured the IR spectra of all four possible isomers of HNCO in Ar matrices at 13 K, and compared the observed frequencies with those obtained by ab initio calculation:

192

APPLICATIONS IN INORGANIC CHEMISTRY

X

W

Y

Z

ν2(A′) ν(XY)

ν1(A′) ν(WX)

ν3(A′) ν(YZ)

ν4(A′) δ(WXY)

+

–

+

+ ν5(A′) δ(XYZ)

ν6(A′′) π

Fig. 2.16. Normal modes of vibration of nonlinear WXYZ molecules.

As shown above, fulminic acid is linear, while the remaining three acids assume bent structures. For a detailed discussion on the structures and IR spectra of these acids, see the review by Teles et al. [804]. Vibrational frequencies are reported for other four-atom molecules such as cisOSOO, which was produced by laser irradiation (193 nm) of the planar SO3 in Ar matrices at 12 K [833], and the HOOO radical, which is probably in cis-conformation in inert gas matrices [834]. Vibrational assignments have also been made for linear four-atom molecules containing CN groups; CNCN(isocyanogen) [835]. HCCN radical (cyanomethylene), [836] [FCNF] [837], and a pair of HBeCN and HBeNC that were formed by reacting Be atom with HCN in Ar matrices at 6–7 [838]. 2.6. TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES 2.6.1. Tetrahedral XY4 Molecules (Td) Figure 2.17 illustrates the four normal modes of vibration of a tetrahedral XY4 molecule. All four vibrations are Raman-active, whereas only n3 and n4 are infraredactive. Appendix VII lists the G and F matrix elements for such a molecule. Fundamental frequencies of XH4-type molecules are listed in Table 2.6a. The trends n3 > n1 and n2 > n4 hold for the majority of the compounds. The XH stretching frequencies are lowered whenever the XH4 ions form hydrogen bonds with counter-

193

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

TABLE 2.5d. Vibrational Frequencies of WXYZ and Related Molecules (cm1) Moleculea,b

n~1, n(WX) n~2, n(XY) n~3, n(YZ)

n~4, d(WXY) n~5, d(XYZ) n~6, p

Ref.

2085 2259.0

1092 —

1190 769.8

570 573.7

538 —

803 804,805

DNCO HNOO HOCN DOCN HCNO

3235 3516.8 3505.7 2606.9 3165.5 3569.6 2635,0 3317.2

2231,0 1092.3 1081.3 949.4 2192,7

— 1054.5 2286,3 2284,6 1244.1

578.6 1485.5 1227,9 1077,8 566.6

— 764.0 — — —

804 806 804,807 804,807 804

ClCNO (g)

—

2219

—

—

808

DCNO HONC DONC HNCO (film) FNCO ClNCO (g) BrNCO INCO (g) HNNN (g) FNNN (g) ClNNN (g) BrNNN (g) INNN (g) cis-HONO trans-HONO HOClO HNCS DNCS HCNS (g) DCNS (g) trans-HSCN ClSCN (sol’n) BrSCN (sol’n) I SCN (sol’n) cis-HNSO trans-HNSO trans-HNNO cis-ClONO trans-ClONO cis-BrONO trans-BrONO trans-HONS FNSO (liquid) ClNSO (liquid) BrNSO (liquid) INSO (liqid) cis-OSNO trans-OSNO

2612.7 3443,7 2545.2 3203

2063.2 628.4 623.1 2254

1350 1336 1218,5 2190.1 2190.3 13251194

475.4 — — — 536,9 538.2 —

418.7 1232,4 902.6 937793

— 361.2 357.3 612

— 379.3 362.1 649

804 804 804 809

861 607.7 505 462.3 3324 873 545 452 360 3397 3553 3527.1 3565 2623 3539 2645 2580.9 520 451 372 3309 3308 3254.0 640.0 658.0 862.6 835.9 3528.0 825 526 451 372 1156.1 1178.0

2172 2212.2 2196.0 2201.1 2150 2037 2075 2058 2055 847 796 591.5 1979 1938 1989 1944 — 678 676 700 1083 982 1628.9 — — 573.5 586.9 642.1 995 989 1000 1928 — —

(2160) 1306.6 1290.9 1298.1 1168 1090 1145 1150 1170 1630 1685 973.9 988 — 857 851 2182.3 2162 2157 2130 1249 1381 1296.2 1710.5 1754.9 1650.7 1723,4 969.5 1230 1221 1214 1247 1454,4 1459.0

529 — 137.4 — 1273 658 719 682 648 1317 1265 1176.9 577 548 615 549 965.9 — — — 900 878 1214.7 — — 420 391.2 1363.3 228 187 161 154 — —

695 707.7 691.1 667.7 527 241 223 — — 616 608 — 461 366 469 366 — 353 369 362 447 496 — — 850 — 197 476.5 600 672 624 602 — —

646 559.0 572.2 583.3 588 504 522 — 578 637 550 — — — 539 481 — — — — 755 651 746.5 — — 368,3 299.3 531 395 359 342 330 — —

810 811 812 811 813 813 813 813 813 814,815 814 816 817 817 818 818 819 820,811 820,811 820,811 821,822 823 824 825 825 826 826 827 828 828 828 828 829 829

(HNCN) (s) HNCO

(continued)

194

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.5d. (Continued ) Moleculea,b

n~1, n(WX)

n~2, n(XY) n~3, n(YZ)

n~4, d(WXY) n~5, d(XYZ) n~6, p

Ref.

cis-HOSO

3544

776

—

830

1166

—

—

a

Vibrational frequencies were measured in inert gas matrices except for those marked by s (solid), g (gas), sol’n (solution), liquid, and film. b Some assignments listed are only empirical.

Y4

X Y1

Y3

Y2 ν2(E) δd (YXY)

ν1(A1) νs(XY)

ν3(F2) νd (XY)

ν4(F2) δd (YXY)

Fig. 2.17. Normal modes of vibration of tetrahedral XY4 molecules.

TABLE 2.6a. Vibrational Frequencies of Tetrahedral XH4 Molecules (cm1) Molecule 10

[ BH4 ] [10 BD4 ] [AlH4] [AlD4] [GaH4] CH4 CD4 SiH4 SiH4 SiD4 GeH4 GeD4 SnH4 SnD4 PbH4 [14 NH4 ]þ [15 NH4 ]þ [ND4]þ [NT4]þ [PH4]þ [PD4]þ {AsH4}þ [AsD4]þ [SbH4]þ [SbD4]þ

~ n1 2270 1604 1757 1256 1807 2917 2085 2180 — (1545) 2106 1504 — — — 3040 — 2214 — 2295 1654 2321 1664 2051 1462

~n2 1208 856 772 549 — 1534 1092 970 — (689) 931 665 758 539 — 1680 (1646) 1215 976 1086, 1026 772, 725 1024 732 842 600

~n3 2250 1707 1678 1220 — 3019 2259 2183 2192 1597 2114 1522 1901 1368 1823 3145 3137 2346 2022 2366, 2272 1732 2341 1676 2061 1470

~n4 1093 827 760 or 766 560 or 556 — 1306 996 910 913 681 819 596 677 487 602 1400 1399 1065 913 974, 919 677 941 688 795 575

Ref. 839,840 839,840 841,842 841,842 841,842 843 844,845 843,846 847 846,848 843,849 843 850 850 851 843 852 843 852 853 853 854 854 854 854

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

195

ions. In the same family of the periodic table, the XH stretching frequency decreases as the mass of the X atom increases. Shirk and Shriver [841] noted, however, that the n1 frequency and the corresponding force constant show an unusual trend in group IIIB: [BH4] ~n1 (cm1) F11 (mdyn/e´)

2270 3.07

[GaH4] 1807 1.94

[AlH4] 1757 1.84

In a series of ammonium halide crystals, the n(NH4þ) becomes lower and the d(NH4þ) becomes higher as the NHX hydrogen bond becomes weaker in the order X ¼ F > Cl > Br > I. Figure 2.18 shows the infrared spectrum of NH4CI, measured by Hornig and others [855], who noted that the combination band between n4 (F2) and n6 (rotatory lattice mode) is observed for NH4F, NH4Cl, and NH4Br because the NH4þ ion does not rotate freely in these crystals. In NH4I (phase I), however, this band is not observed because the NH4þ ion rotates freely. Table 2.6b lists the vibrational frequencies of a number of tetrahalogeno compounds. The trends n3 > nx and n4 > n2 hold for most of the compounds. The latter trend is opposite to that found for MH4 compounds. The same table also shows the effect of changing the halogen. First, the MX stretching frequency decreases as the halogen is changed in the order F > Cl > Br > I. The average values, of n(MBr)/n(MCl) and n(MI)/n(MCl) calculated from all the compounds listed in Table 2.6b are approximately 0.76 and 0.62, respectively, for n3, and approximately 0.61 and 0.42, respectively, for n1. These values are useful when we assign the MX stretching bands of halogeno complexes (Sec. 3.23 in Part B). Also, the effect of changing the oxidation state on the MX stretching frequency is seen in pairs such as [FeX4] and [FeX4]2 (X ¼ Cl, Br) [894], the MX stretching frequency increases as the oxidation state of the metal becomes higher. The ratio n3(FeX4)/n3(FeX42) is 1.32 in this case. In the solid state, n3 and n4 may split into two or three bands because of the site effect. In some cases, the MX4 ions are distorted to a flattened tetrahedron (D2d) or a structure of lower symmetry (Cs) [892,912]. According to an X-ray analysis [913] the unit cell of [(CH3)2CHNH3]2 [CuCl4] contains two square-planar and four distorted tetrahedral [CuCU4]2 ions.

Fig. 2.18. IR spectrum of crystalline NH4Cl ( n5, n6—lattice modes).

196

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.6b. Vibrational Frequencies of Tetrahedral Halogeno Compounds (cm1) Molecule 2

[BcF4] [MgCL4]2 [MgBr4]2 [MgI4]2 [BF4] [BCl4] [BBr4] [BI4] [AlF4] [AlCl4] [AlBr4] [AlI4] [GaCl4] [GaBr4] [GaI4] [InCl4] [InBr4] [InI4] [TlCl4] [TlBr4] [TiI4] CF4 CCl4 CBr4 CI4 SiF4 SiCl4 SiBr4 SiI4 70 GeF4 GeCl4 GeBr4 GeI4 SnCl4 SnBr4 SnI4 PbCl4 [NF4]þ [NCl4]þ [PF4]þ [PCl4]þ [PBr4]þ [PI4]þ [AsF4]þ [AsCl4]þ [AsBr4]þ [AsI4]þ [SbCl4]þ [SbBr4]þ [CuCl4]2

~ n1

~n2

~n3

~n4

Ref.

547 252 150 107 777 406 243 170 622 348 212 146 343 210 145 321 197 139 303 186 130 908.4 460.0 267 178 800.8 423.1 246.7 168.1 735.0 396.9 235.7 158.7 369.1 222.1 147.7 331 848 635 906 458 266 193.5 745 422 244 183 395.6 234.4 —

255 100 60 42 360 192 118 83 210 119 98 51 120 71 52 89 55 42 83 54 — 434.5 214.2 123 90 264.2 145.2 84.8 62 202.9 125.0 74.7 60 95.2 59.4 42.4 90 443 430.3 275 178 106 71.0 213 156 88 72 121.5 76.2 77

800 330 290 259 1070 722 620 533 760 498 394 336 370 278 222 337 239 185 296 199 146 1283.0 792,765a 672 555 1029.6 616.5 494.0 405.5 806.9 459.1 332.0 264.1 408.2 284.0 210 352 1159 283.3 1167 662 512 410 829 500 349 319 450.7 304.9 267, 248

385 142 90 60 533 278 166 117 322 182 114 82 153 102 73 112 79 58 104 64 60 631.2 313.5 183 123 388.7 220.3 133.6 90.6 273.7 171.0 111.1 79.0 126.1 85.9 63.0 103 611 233.3 358 255 153 89.0 272 187 115 87 139.4 92.1 136, 118

856 857 857 857 858 859,860 859,861 862 863 864 865 866 867 868 869 870 871 869 872 872 873 874 874 875 876 874 874 874 877 878 874 874 877 874 874 874 879 880,882 881 882 883 884 885 886 887,888 889 890 891 891 892,893

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

197

TABLE 2.6b. (Continued ) Molecule 2

[CuBr4] [ZnCl4]2 [ZnBr4]2 [ZnI4]2 [CdCl4]2 [CdBr4]2 [CdI4]2 [HgCl4]2 [HgI4]2 [ScI4] TiF4 TiCl4 TiBr4 TiI4 ZrF4 ZrCl4 ZrBr4 ZrI4 HfCl4 HfBr4 HfI4 VCl4 CrF4 CrCl4 CrBr4 [MnCl4]2 [MnBr4]2 [MnI4]2 [FeCl4] [FeBr4] [FeCl4]2 [FeBr4]2 [Fe I4] [NiCl4]2 [Ni Br4]2 [Ni I4] 2 {CoCl4]2 [CoBr4]2 [Co I4]2 U F4 U Cl4 a

~ n1

~n2

~n3

~n4

Ref.

— 276 171 118 261 161 116 267 126 129 712 389 231.5 162 (725–600) 377 225.5 158 382 235.5 158 383 717 373 224 256 195 108 330 211 266 162 142 264 — 105 287 179 118 — —

— 80 — — 84 49 39 180 35 37 185 114 68.5 51 (200–150) 98 60 43 101.5 63 55 128 — 116 60 — 65 46 114 74 82 — 60 — — — 92 74 — (123) —

216, 174 277 204 164 249, 240 177 — 276 140 — 793 498 393 323 668 418 315 254 390 273 224 475 790 486 368 278, 301 209, 221 188, 193 378 307/292 286 219 252/235 294/280 228 191 320 243/249 202/194 537 337.4

85 126 91 — 98 75, 61 50 192 41 54 209 136 88 67 190 113 72 55 112 71 63 128, 150 201 126 71 120 89 56 136 94 119 84 73 119 81 — 143/126 101/90 56 114 71.7

892 894 894 894 895 895 895 896 897 898 899 900 900 900 901 900 900 900 900 900 900 902 903 904 904 894,905 894,905 894,905 894 906,894 894 894 907 892,908 892,908 892 909 909 892 910 911

Fermi resonance with (n1þ n4).

Willett et al. [914] demonstrated by using IR spectroscopy that distorted tetrahedral ions can be pressed to square–planar ions under high pressure. In nitromethane, (Et4N)2[CuCl4] exhibits two bands at 278 and 237 cm1, indicating the distortion in solution [915]. A solution of (Et3NH) [GaCl4] in 1,2-dichloroethylene exhibits three bands at 390,383, and 359 cm1 due to lowering of symmetry caused by NHCl hydrogen bonding [916]. The IR spectra of long-chain tertiary and quaternary

198

APPLICATIONS IN INORGANIC CHEMISTRY

ammonium salts of [GaCl4] and [GaBr4] ions in benzene solution show that the degree of distortion of these ions depends on the nature of the cation and the concentration [917]. Distortion of [MCI4]2 ions (M ¼ Fe,Co,Ni,Zn) is also reported for their cesium and rubidium salts [918]. Matrix isolation spectroscopy has provided structural information that is unique to XY4-type molecules in inert gas environments. For example, the symmetry of the anions in CsþBF4 and CsþPF4 formed via salt–molecule reactions in Ar matrices are lowered to C3v in the former, and to a Symmetry no higher than C2v, in the latter [920]. In inert gas matrices, the symmetry of UF4 is distorted to D2d [921]. Irradiation of CCl4, in Ar matrices at 12 K produced a new species, isotetrachloromethane, for which Maier et al. [922] proposed the following structure:

The v(CClt) are observed at 1019.7 and 929.1 cm1, whereas the v(CClb) and v(ClbCle) are assigned at 501.9 and 246.4 cm1 respectively. Jacox and Thompson [923]. presented IR evidence for the formation of the CO4 ion resulting from the reaction of CO2, O2, and excited Ne atoms in Ne matrices. The planar structure of Cs symmetry has been proposed:

Table 2.6b includes a number of data obtained by Clark et al. from their gas-phase Raman studies [874,900]. Some halides such as Til4 [924], Snl4 [877,925] and TiBr4 [900], and VCl4 [926] have strong electronic absorptions in the visible region, and their resonance Raman spectra have been measured in solution. As discussed in Sec. 1.23, these compounds exhibit a series of v1 overtones under rigorous resonance conditions. A tetrahedral MCl4 molecule in which M is isotopically pure and Cl is in natural abundance consists of five isotopic species because of mixing of the 35 Cl (75.4%) and 37 Cl (24.6%) isotopes. Table 2.6c lists their symmetries, percentages of natural abundance, and symmetry species of infrared-active modes corresponding to the n3 vibration of the Td molecule. It has been established [927] that these nine bands overlap partially to give a five-peak chlorine isotope pattern whose relative intensity is indicated by the vertical lines shown in Fig. 2.19b. If M is isotopically mixed, the spectrum is too complicated to assign by the conventional method. For example, tin is a mixture of 10 isotopes, none of which is predominant. Thus 50 bands are expected to appear in the n3 region of SnCl4. It is almost impossible to resolve all these peaks, even

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

199

TABLE 2.6c. Infrared-Active Vibrations of M35 Cln 37 Cl4n Type Molecules Species 35

M Cl4 M35 Cl37 3 Cl M35 Cl37 2 Cl2 M35 Cl 37 Cl3 M35 Cl4

Symmetry Td C3v C2v C3v Td

Abundance (%) 32.5 42.5 20.5 4.4 0.4

IR-Active Modes F2 A1, E A1, B1, B2 A1, E F2

in an inert gas matrix at 10 K. K€ oniger and M€ uller [928] therefore, prepared 116 SnCl4 35 116 and Sn Cl4 on a milligram scale and measured their infrared spectra in Ar matrices. As expected, the former gave a five-peak chlorine isotope pattern, whereas the latter showed a single peak at 409.8 cm1. This work was extended to GeCl4, which consists of two Cl and five Ge isotopes. In this case, 25 peaks are expected to appear in the n3 region. However, the observed spectrum (Fig. 2.19a) shows about 10 bands; K€oniger et al. [929], therefore, prepared 74 GeCl4 and Ge35 Cl4 and measured their spectra in Ar

Fig. 2.19. Matrix isolation IR (a) and computer simulation spectra (b) of GeCl4. Vertical lines in (b) show the five-peak chlorine isotope pattern of 74 GeCl4 .

200

APPLICATIONS IN INORGANIC CHEMISTRY

matrices. As expected, both compounds showed a five-peak spectrum. The ism (isotope shift per unit mass difference) values for Cl and Ge were found to be 3.8 and 1.2 cm1, respectively. Using these values, it was then possible to calculate the frequencies of all other isotopic molecules. Furthermore, the relative intensity of individual peaks is known from the relative concentration of each isotopic molecule. On the basis of this information, Tevault et al. [930] obtained a computer-simulation infrared spectrum of GeCl4 in natural abundance (Fig. 2.19). Normal coordinate analyses of tetrahedral XY4 molecules have been carried but by a number of investigators [931]. Thus far, Basile et al. [932] have made the most complete study. They calculated the force constants of 146 compounds by using GVF, UBF, and OVF fields (Sec. 1.14), and discussed several factors that influence the values of the XY stretching force constants. It has long been known that molecules such as SF4, SeF4, and TeF4 assume a distorted tetrahedral structure (C2v) derived from a trigonal–bipyramidal geometry with a lone-pair of elections occupying an equatorial position:

These molecules exhibit nine normal vibrations that are classified into 4A1þ A2 2B1þ 2B2. Here, all the vibrations are IR- as well as Raman-active except for A2 which is only Raman-active. Table 2.6d lists the vibrational frequencies of distorted XY4 molecules of this type. It should be noted that these molecules exhibit four stretching modes, two of which (2A1) are polarized in the Raman. Adams and Downs [939] carried out normal coordinate analyses on SeF4 and TeF4, and found that the axial bonds are weaker than the equatorial bonds. The vibrational spectra of tetraalkylammonium salts of [AsX4] (X ¼ Cl, Br), [BiX4] (X ¼ Cl,Br,I), and [SbX4] (X ¼ Cl,Br,I) have been studied in the solid state and in solution) [936]. Except for solid (Et4N)[AsCl4] and [(n-Bu)4N][SbI4], all these ions assume the distorted tetrahedral structure shown above. The [ClF4]þ, [BrF4]þ, and [IF4]þ ions were found in the following adducts [941]: CIF5 ðAsF5 Þ BrF5 ðSbF5 Þ2 IF5 ðSbF5 Þ

¼ ½CIF4 ½AsF6 ¼ ½BrF4 ½Sb2 F11 ¼ ½IF4 ½SbF6

It should be noted that SeCl4, SeBr4, TeCl4, and TeBr4 consist of the pyramidal XY3þ cation and the Y anion in the solid [615,616].

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

201

TABLE 2.6d. Vibrational Frequenciesa of Distorted Tetrahedral XY4 Molecules (cm1) C2v

[PF4] [SbF4] [SbCl4] [SbBr4] [Sb I4] SF4 SeF4 TeF4 TeCl4 [ClF4]þ [BrF4]þ {I F4]þ

n~1, A1

n~2, A1

n~3, A1

n~4, A1

n~5, A1

n~6, A1

n~7, A1

n~8, A1

n~9, A1

Ref.

798 584 339 228 169 892 739 695 — 800 723 728

422 404 296 190 — 558 551 572 — 571 606 614

446 305 147 — 114 532 362 333.2 — 385 385 345

210 161 — — — 228 200 151.5 — 250 219 263

— — — — — (437) — — — 475 — —

515 480 321 201 162 730 585 586.9 312 795 704 —

446 247 199 140 85 475 254 293.2 165 515 419 388

745 541 246 169 148 867 717 682.2 378 829 736 719

290 220 — — — 353 403 184.8 104 385 369 311

933,920 934,935 935 935,936 935,936 937 938 934,939 940 937 941 941

a Whereas n1 and n8 are stretching modes of equatorial bonds, n2 and n6 are stretching modes of axial bonds. For the normal modes of bending vibrations, see Ref. 942.

Table 2.6e lists the vibrational frequencies of tetrahedral MO42, MS42, and MSe2 -type compounds. The rule n3 > n1 and n4 > n2 hold for the majority of the compounds. Figure 2.20 shows the Raman spectra of K2SO4 in the solid state as well as in aqueous solution. It should be noted that n2 and n4 are often too close to be observed as separate bands in Raman spectra. Weinstock et al. [960] showed that, in Raman spectra, n2 should be stronger than n4, and that n4 is hidden by n2 in [MoO4]2 and [ReO4] ions. Baran et al. [974,975] have found several relationships between the n1/n3 ratio and the negative charge of the anion or the mass of the central atom in a series of oxoanions listed in Table 2.6e: (1) For a given central atom, the n1/n3 ratio increases as the negative charge of the anion increases (e.g., [MnO4] < [MnO4]2 < [MnO4]3). (2) For anions of the same negative charge with the central atom belonging same group of the periodic table, the n1/n3 ratio increases with the mass of the central atom (e.g., [PO4]3 < [AsO4]3). (3) For isoelectronic ions in which the mass of the central atom remains approximately constant, the n1/n3 ratio increases with the increasingly negative charge of the anion (e.g., [ReO4] < [WO4]2). These trends are very useful in making correct assignments of the n1 and n3 vibrations of tetraoxoanions [975]. The IR spectra of matrix-isolated M2(SO4) (M ¼ K,Rb,Cs) can be interpreted in terms of D2d symmetry and their n(MO) vibrations are observed at 262–220, 215– 190, and 195–148 cm1, respectively, for the K, Rb, and Cs salts [976].

202

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.6e. Vibrational Frequencies of Tetrahedral MO4-, MS4-, and MSe4-Type Compounds (cm1) ~n1

~n2

~n3

~n4

Ref.

880

372

627

943,944

495

353

—

945

[SiO4] [PO4]3

819 938

340 420

391

282

416

215 540 500 349 171 156 450 328 459 331 256 267 306 332 325 336 319 193.5 121 163 100 170 103 339 240.3 317 184 120 325 182 107 346 325 308 325 331 200 340 319 339 331

527 567 317 a 296 270 669 a 651 463 216 178 611 411 625 410 325 305.9 371 387 379 (336) 368 (193.5) (121) (163) (100) (170) (103) 375 363.9 (317) (184) 120 (325) (182) (107) 386 332 332 336 (331) (200) 322 336 312 336

946,947 948,949

[PS4]3

886 1000 a 805 956 1017 535 a 512 548 1012 a 988 878 419 380 1105 856 1119 878 853 879.2 770 846 800 804 780 470 365 421 316 399 277 863 775.8 837 472 340 838 455 309 902 820 778 912 920 486 790 921 845 804

Compund [BO4]

5

[CS4]

4

3

[PSe4]

3

[NO4]

843

[AsO4]3 [AsS4]3 [SbS4]3 [SO4]2 [SeO4]2 [ClO4] [BrO4] [IO4] XeO4 [TiO4]4 [ZrO4]4 [HfO4]4 [VO4]3 [VO4]4 [VS4]3 [VSe4]3 [NbS4]3 [NbSe4]3 [TaS4]3 [TaSe4]4 [CrO4]2 [CrO4]3 [MoO4]2 [MoS4]2 [MoSe4]2 [WO4]2 [WS4]2 [WSe4]2 [MnO4]2 [MnO4]2 [MnO4]3 [TcO4] [ReO4] [ReS4] [FeO4]2 RuO4 [RuO4] [RuO4]2

837 386 366 983 822 928 801 791 775.7 761 792 796 826 818 404.5 (232) 408 239 424 249 833 844.0 897 458 255 931 479 281 834 812 789 912 971 501 832 885.3 830 840

950 951 952b 953 953 953 946 954,946 953 955,956 957 958 959 959 959 960 959 961 961 961 961 961 961 954 955 960 962 963 960 962 961 964,965 959 966 960 960 967 959 968 959 959

203

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

TABLE 2.6b. (Continued ) 2.6e. (Continued Compund

~n1

~n2

~n3

~n4

Ref.

OsO4 XeO4 [CoO4]4 {B (OH)4] c [Al (OH)4] c [Zn (OH)4] c

965.2 776 670 754 615 470

333.1 267 320 379 310 300

960.1 878 633 945 (720) (570)

322.7 305 320 533 (310) (300)

969 969a 970 971,972 973 973

a b c

Splitting may suggest D2d symmetry in the crystalline state. Na3 [NO4]3 was obtained by the reaction of NaNO3 and Na2O. Only MO4 skeletal vibrations are listed for this ion.

The IR spectrum and normal coordinate analysis of the SO4 radical produced by the reaction of SO3 with atomic oxygen in inert gas matrices are suggestive of the Cs structure rather than the C2v structure [977]:

The IR spectra of H2SO4 and its deuterated derivatives in Ar matrices have been assigned on the basis of C2 symmetry ((HO)2SO2) [978]. The IR spectrum of the ClO4

Fig. 2.20. Raman spectra of K2SO4 in (a) the solid state and (b) aqueous solution.

204

APPLICATIONS IN INORGANIC CHEMISTRY

radical in inert gas matrices indicates a C3v structure with one long ClO bond as the C3 axis. The v(ClO) of this bond is at 874 cm1, which is much lower than those of the other three bonds (1234 and 1161 cm1) [979]. The IR/Raman spectra of MþClO4 salts in nonaqueous solvents suggest the formation of a unidentate ion pair for M ¼ Liþ, and a bidentate ion pair for M ¼ Naþ [980]. The RR spectra of highly colored ions such as CrO42 [981], MoS42 [982], VS43 [983], and MnO4 [964,981] have been measured. The n3(IR) bands of gaseous RuO4 [968] and XeO4 [984] exhibit complicated band contours consisting of individual isotope peaks of the central atoms. M€ uller and coworkers [974,985,986] reviewed the vibrational spectra of transition-metal chalcogen compounds. Basile et al. [932] carried out normal coordinate analysis on more than 60 compounds of these types. 2.6.2. Tetrahedral ZXY3, Z2XY2, and ZWXY2 Molecules If one of the Yatoms of an XY4 molecule is replaced by a Z atom, the symmetry of the molecule is lowered to C3v. If two Yatoms are replaced, the symmetry becomes C2v. In ZWXY2 and ZWXYU types, the symmetry is further lowered to Cs and C1, respectively. As a result, the selection rules are changed as shown in Table 2.6f. The number of infrared-active vibrations is six for ZXY3 and eight for Z2XY2. Table 2.6g lists the vibrational frequencies of ZXY3-type molecules. The SO stretching frequency of the [OSF3]þ ion in [OSF3]SbF6 (1536 cm1) is the highest that has been observed, and corresponds to a force constant of 14.7 mdyn/A [1006]. It is also interesting to note that the structure the [OXeF3]þ ion is Cs [726] whereas that of the [OXeF3] ion is C2v [727] as shown below:

TABLE 2.6f. Correlation Tablea for Td, C3v, C2v, and C1 Point Group

n1

n2

n3

Td C3v

A1 (R) A1 (IR, R)

E (R) E (IR, R)

C2v

A1 (IR, R)

A1 (IR, R)þ A2 (R)

F2 (IR, R) A1 (IR, R)þ E (IR, R) A1 (IR, R)þ B1 (IR, R)

C1

A (IR, R)

2A (IR, R)

3A (IR, R)

a

See Appendix IX.

n4 F2 (IR, R) A1 (IR, R)þ E (IR, R) A1 (IR, R)þ B1 (IR, R)þ B2 (IR, R) 3A (IR, R)

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

205

TABLE 2.6g. Vibrational Spectra of ZXY3 Molecules (cm1) C3v ZXY3

[HGaCl3] [OCF3] FCCl3 BrCC3 ICCl3 FSiH3 H28 SiF3 FSiCI3 FSiBr3 FSiI3 BrSiF3 79 BrSiH3 [SSiCl3] [SGeCl3] HGeCl3 ONF3 [HPC13]þ [ClPBr3]þ [BrPCl3]þ [HPF3]þ OPF3 OPCI3 OPBr3 [FPO3] SPF3 [OSF3]þ NSF3 ISCl3 [HSO3] [FSO3] [ClSO3] [SSO3] [SeSO3]2 [FSeO3] FC1O3 FBrO3 [NClO3]2 OVF3 OVC13 OVBr3 NVCl3 ONbCl3 [FCrO3] [ClCrO3] [BrCrO3] [OMoS3]2 [OMoSe3]2 [SMoO3]2 [SMoSe3]2

~ n1 (A1) n (XY3)

~n2 (A1) n (XZ)

~n3 (A1) d (XY3)

~n4 (E) n (XY3)

~n5 (E) d (XY3)

~n6 (E) rr (XY3)

Ref.

349.5 813 537.6 419.9 390 — 855.8 465 318 242 858 2198.8 395 356 418.4 743 528 285 399 948 873 486 340 1001 — 909 772.6 482 1038 1142 1042 995 995 896 1062.6 875.2 815 721.5 408 271 352 395 911 907 895 461 293 900 —

1960.4 1560 1080.4 730.7 684 990.9 2315.6 948 912 894 505 929.8 685 505 2155.7 1691 2481 587 582 2544 1415 1290 1261 794 985 1540, 1532 1522.9 293 2588 862 381 446 310 580,603 716.3 605.0 1256 1057.8 1035 1025 1033 997 635 438 260 862 858 472 471

167,9 595 351.3 246.5 224 875.0 425.3 239 163 115 288 430.8 220 163 181.8 528 208 149 217 480 473 267 173 534 696 535 524.6 242 629 571 601 669 645 392 549.7 (354) 594 257.8 165 120 — 106 338 295 230 183 120 318 121

320.6 960 847.8 785.2 755 — 997.8 640 520 424 940 2209.3 492 360 708.6 883 634 500 657 1090 990 581 488 1125 947 1063 814.6 493 1200 1302 1300 1123 1120 968,974 1317.9 974 870 806 504 400 430 448 955 954 950 470 355 846 342

609.3 422 242.8 290.2 284 962.2 843.6 282 226 194 338 996.4 — — 454 558 280 172 235 989 485 337 267 — 442 508 432.3 140 509 619 553 541 530 409 589.8 (376) 623 204.3 129 83 — 225 370 365 350 183 120 318 121

— 576 395.3 188.0 188 730.0 306.2 167 110 71 200 632.4 225 178 145.0 400 653 120 159 380 345 193 118 382 405 387 346.2 255 1123 424 312 335 — 301 403.9 (296) 457 308 249 212 — 110 261 209 (175) 263 188 239 —

987 988,989 990 990 991 992 993 994 994 994 994 995 996 996 997 998 999 1000 1000 1001,1001a 1002 1002 764,1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015,1016 1017,1018 1019 1020 1020 1020 1021 1021a 1022 1023 1024 1023 1025 1026 985

(continued)

206

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.6b. 2.6g. (Continued (Continued ) C3v ZXY3 2

[SeMoS3] [OWS3]2 [OWSe3]2 [SWSe3]2 FMnO3 ClMnO3 FTcO3 ClTcO3 FReO3 ClReO3 BrReO3 [NReO3]2 [SReO3] [NOsO3] [FOsO3]þ

~ n1 (A1) n (XY3)

~n2 (A1) n (XZ)

~n3 (A1) d (XY3)

~n4 (E) n (XY3)

~n5 (E) d (XY3)

~n6 (E) rr (XY3)

Ref.

349 474 292 468 903.6 889.9 962 948 1013.2 994 997 878 948 892.5 1002

458 878 878 281 715.5 458.9 696 451 701 436,427 350 1022 528 1026.2 745

— 182 (120) 108 339.0 305 317 300 305.5 291 195 315 322 310.0 333

473 451 312 311 950.6 951.6 951 932 978.3 963 963 830 906 872.0 992

150 182 (120) 150 380.0 365 347 342 346.9 337 332 273 322 299.7 372

183 264 194 108 264.0 200 231 197 234.2 192 168 380 (240) 371.5 238

1027 1023 1025 1027 1028,1029 1026,1029 1030 1031 1032 1032 1022 1033 1022 1034 1035

The first example of the trifluorosulfite anioin (OSF3) was obtained as the Me4Nþ salt, and its IR and Raman spectra assignments based on a theoretically predicted pseudo-trigonai–bipyramidal structure with two long SF bonds in the axial direction and one F, one O atom and a sterically active lone-pair electrons at the triangular corners of the equatorial plane [1036]. The Raman spectrum of the [NVO3]4 ion has been assigned as a pseudo-tetrahedral anion of Cs symmetry [1037]. Vibrational spectra have been reported for a number of mixed halogeno complexes. Some references are as follows: [AlClnBr4n] [1038], SiFnCl4n [1039], SiClnBr4n [1040], and [FeClnBr4n] [1041]. It is interesting to note that the SiFClBrI molecule exhibits the SiF, SiCl, SiBr, and SiI stretching bands at 910, 587, 486, and 333 cm1, respectively [1042]. The vibrational spectrum of OClF3 suggests a trigonal–bipyramidal structure similar to that of the [OXeF3]þ ion shown above [1043]. Table 2.6h lists the vibrational frequencies of tetrahedral Z2XY2 molecules. The vibrational spectrum of O2XeF2 can be interpreted on the basis of a trigonal– bipyramidal structure in which two F atoms are axial, and two O atoms and a pair of electrons are equatorial [1059]. The structures of [O2ClF2] [1057] and [O2BrF2] [1058] are similar to that of O2XeF2, but that of [O2ClF2]þ [1067] is pseudo-tetrahedral. The gas-phase Raman spectrum of Cl2TeBr2 at 310 C is indicative of the C1 symmetry shown below [1068]. Table 2.6i lists the vibrational frequencies of tetrahedral ZWXY2 molecules. Other references are as follows: SFPC12 [1075], FOPCl2 [1076], FOPB2 [1077], ClOPBr2 and BrOPCl2 [1003], FBrSO2 [1078], and FCIClO2 [1079].

207

þ

[F2NH2] [O2PH2] F2CI2 H2SiCl2 H2SiBr2 F2SiBr2 H2GeF2 H2GeCl2 H2GeBr2 H2GeI2 [Cl2PBr2]þ [O2PF2]þ O2SF2 O2SCl2 F2SCl2 [O2SeF2] [O2ClF2] [O2BrF2] O2XeF2 [O2VF2] [O2VCl2] [O2NbS2]3

Z2XY2

2637 2365 605 942 925 414 860 840 848 821 326 910 847.4 405 527 396 363 374 490 631 431 464

~n1 (A1) n (XY2)

1060 1046 1064 2221 2206 891 2155 2132 2122 2090 584 1179 1265.5 1182 592 859 1070 885 845 969 968 897

~n2 (A1) n (XZ2) 1543 1160 272 514 393 270 720 404 298 220 191 269 388.7 218 — 241 198 201 198 199 128 246

~n3 (A1) d (XY2) 528 470 112 188 122 115 (270) 163 104 96 132 567 552.0 560 — 445 559 429 333 365 323 356

~n4 (A1) d (XZ2)

TABLE 2.6h. Vibrational Frequencies of Z2XY2 Molecules (cm1)

— 930 251 710 688 187 (664) 648 105 628 150 — 390.8 388 — — 480 405 — — 200 —

~n5 (A2) rt (XY2) 2790 2308 740 868 828 540 814 772 757 706 518 962 884.1 362 533 — 510 448 578 664 436 514

~n6 (B1) n (XY2) 1176 820 200 566 456 241 720 420 324 294 173 492 545.0 380 — — 337 339 313 285 277 (297)

~n7 (B1) rw (XY2) 1036 1180 1110 2221 2232 974 2174 2155 2138 2110 616 1269 1497.0 1414 770 833 1221 912 902 960 958 872

~n8 (B2) n (XZ2) 1487 1093 278 602 — 257 596 533 492 451 201 528 538.2 282 — 304 337 314 313 265 194 (271)

~n9 (B2) rr (XY2)

(continued)

1044 1045 1046 1047 1048 1049 1050 1087 1050 1050 1022 1051 1052,1053 1054 1055 1056 1057 1058 1059 1060 1060 1061

Ref.

208

~n1 (A1) n (XY2)

727 475 399.0 262 473 283 454 282

Z2XY2

O2CrF2 O2CrCl2 O2CrBr2 O2MoBr2 2 ½O92 2 MoS2 [O2MoSe2]2 [O2WS2]2 [O2WSe2]2

TABLE 2.6b. 2.6h. (Continued (Continued )

1006 995 982.6 995 819 864 886 888

~n2 (A1) n (XZ2) 208 140 — 147 200 114 196 116

~n3 (A1) d (XY2) 364 356 (305) 373 307 339 310 319

~n4 (A1) d (XZ2) (259) (224) — — 267 251 280 235

~n5 (A2) rt (XY2) 789 500 403.5 338 506 353 442 329

~n6 (B1) n (XY2) 304 257 — 161 267 251 280 235

~n7 (B1) rw (XY2) 1016 1002 995.4 970 801 834 848 845

~n8 (B2) n (XZ2)

274 215 — 184 246 — 235 156

~n9 (B2) rr (XY2)

1062 1063 1064 1065 1066 1061 1061 1061

Ref.

209

b

X denotes the central atom. These assignments may be interchanged.

623 549 561 477 907 915 888 887 432 372 552 500 355 320 461 459

900 949 884 938 546 479 472 377 545 493 391 333 478b 473 360b 317b

OClPF2 SClPF2 OBrPF2 SBrPF2 OFPCl2 SFPCl2 OFPBr2 SFPBr2 OBrPCl2 SBrPCl2 OClPBr2 SClOBr2 [OSeMoS2]2 [OSeWS2]2 [OSMoSe2]2 [OSWSe2]2

a

~n2 (A0 ) n (XW)

~n1 (A0 ) ns (XY2)

ZWXY2a 1384 735 1380 719 1358 750 1337 713 1285 743 1275 729 869 879 865 882

~n3 (A0 ) n (XZ) (419) 394 (413) 389 205 193 133 129 242 (230) 130 121 190 190 — —

~n4 (A0 ) ds (XY2)

TABLE 2.6i. Vibrational Frequencies of ZWXY2 Molecules (cm1)

(274) 363 (240) 288 330 328 273 274 172 150 209 196 — — — —

~n5 (A0 ) ds (WXY) 412 209 316 175 382 268 304 218 285 206 291 190 273 265b — —

~n6 (A0 ) ds (ZXY2) 960 925 947 911 626 574 536 470 580 536 492 436 467b 458 320b 312b

~n7 (A00 ) na (XY2) 274 252 240 231 253 193 220 162 161 150 157 136 — — — —

~n8 (A00 ) d (ZXW) 419 613 413 297 374 317 290 254 327 230 271 205 273 255b — —

~n9 (A00 ) da (XY2)

639,1069 639,1069 639,1069 639,1069 639,1069 639,1069 639,1069 639,1069 1070 1070 1070 1070 1071,1072 1072,1073 1074 1074

Ref.

210

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.21. Normal modes of vibration of square–planar XY4 molecules.

2.6.3. Square–Planar XY4 Molecules (D4h) Figure 2.21 shows the seven normal modes of vibration of square–planar XY4 molecules. Vibrations n3, n6, and n7 are infrared-active, whereas n1, n2, and n4 are Raman-active. However, n5 is inactive both in infrared and Raman spectra. In the case of the [PdCl4]2 ion, the n5 was observed at 136 cm1 by inelastic neutron scattering (INS) spectroscopy [1088]. Table 2.6j lists the vibrational frequencies of some ions belonging to this group. Chen et al. [1088a] also reported the IR and Raman spectra of K2[MX4] (M ¼ Pt or Pd and X ¼ Cl or Br) and band assignments on the basis of normal coordinate analysis. XeF4 (Sec. 1.11) is an unusual example of a neutral molecule that takes a square– planar structure. This structure is predicted by the valence shell electron pair repulsion (VSEPR) theory [1089] which states that two lone-pairs in the valence shell occupy the axial positions because they exert greater mutual repulsion than

211

b

a

505 523 288 554.3 347 212 148 303 188 330 208 155

~n1 (A1g) ns (XY)

288 246 128 218 171 102 75 164 102 171 106 85

~n2 (B1g) d (XY2) 425 317 — 291 — — — 150 114 147 105 105

~n3 (A2u) p 417 449 261 524 324 196 110 275 172 312 194 142

~n4 (B2g) na (XY) 680–500 580–410 266 586 350 252b 192 321 243 313 227 180

~n6 (Eu) nd (XY) — (194) — (161) 179 110b 113 161 104 165 112 127

~n7 (Eu) dd (XY2)

For these molecules n5 is inactive. The designations B1g and B2g may be interchanged, depending on the definition of symmetry axes involved. From Ref. 1087.

[ClF4] [BrF4] [ICl4] XeF4 [AuCl4] [AuBr4] [AuI4] [PdCl4]2 [PdBr4]2 [PtCl4]2 [PtBr4]2 [Ptl4]2

XY4

TABLE 2.6j. Vibrational Frequencies of Square-Planar XY4 Molecules (cm1)a

1080 1080 1081 1082,1083 1084,1085 1084,1085 1084 1085,1086 1084,1086 1085,1086 1085,1086 1084,1086

Ref.

212

APPLICATIONS IN INORGANIC CHEMISTRY

single-bond pairs:

The structures of fluorides, oxides, and oxyfluorides of xenon discussed in other sections can also be rationalized on the basis of VSEPR theory. Figure 2.22

Fig. 2.22. (a) IR [1090] and (b) Raman [1082] spectra of XeF4 vapor.

TRIGONAL–BIPYRAMIDAL AND TETRAGONAL–PYRAMIDAL XY5 AND RELATED MOLECULES

213

shows the IR and Raman spectra of gaseous XeF4 obtained by Claassen et al. [1082,1090]. Bosworth and Clark [1091] measured the relative intensities of Raman-active fundamentals of some of these ions, and calculated their bond polarizability derivatives and bond anisotropies. They [1087] also measured the resonance Raman spectra of several [AuBr4] salts in the solid state, and observed progressions such as nn1 (n ¼ 1–9) and n2þ nn1 (n ¼ 1–5). For normal coordinate analyses of square–planar XY4 molecules, see Refs. [1081], [1092], and [1093].An empirical relationship between the antisymmetric MCl stretching frequency [na(MCl), cm1] and the MCl distance (RMCl, A) has been derived using IR and structural data for a large number of planar Pt(II) and Pd(II) complexes of MCI2L2, MC12LL0 , and MC1L3 types (L: unidentate ligand) ½na ðM ClÞ2 ¼

P ðRMCl 1:6Þ3

where P is 38988 for Pt(II) and 41440 for Pd(II) complexes [1094]. The IR and Raman spectra of the [AuCl4nBrn] ions (n ¼ 1–3) [1095] and Raman spectra of the square–planar tetrahydrido[PtH4]2 and related ions [[1095]a] are reported.

2.7. TRIGONAL–BIPYRAMIDAL AND TETRAGONAL–PYRAMIDAL XY5 AND RELATED MOLECULES An XY5 molecule may be a trigonal bipyramid (D3h) or a tetragonal pyramid (C4v). 00 If it is trigonal–bipyramidal, only two stretching vibrations (A2 and E0 ) are infraredactive. If it is tetragonal–pyramidal, three stretching vibrations (two A1 and E) are infrared-active. As discussed in Sec. 1.11, however, it is not always possible to make clear-cut distinctions of these structures based on the selection rules since practical difficulties arise in counting the number of fundamental vibrations in infrared and Raman spectra. 2.7.1. Trigonal–Bipyramidal XY5 Molecules (D3h) Figure 2.23 shows the eight normal vibrations of a trigonal–bipyramidal XY5 0 00 molecule. Six of these eight (A1 , E0 E00 ) are Raman-active, and five (A2 and E0 ) are infrared-active. Three stretching vibrations (n1, n2, and n5) are allowed in the Raman, whereas two (n3 and n5) are allowed in the infrared. Table 2.7a lists the observed frequencies and band assignments of trigonal-bipyramidal XY5 molecules. The IR spectrum of the ion-paired complex, Csþ[GeF5], in inert gas matrices has been reported [1116]. It takes a trigonal–bipyramidal structure perturbed by the presence of the Csþ ion in its vicinity.

214

APPLICATIONS IN INORGANIC CHEMISTRY

Y′ Y X Y

Y

Y′ ν1(A′1) νs(XY3)

ν5(E′1) νa(XY3)

ν2(A′1) νs(XY′2)

ν3(A′′2) νa(XY′2)

ν4(A′′2) π (XY3)

ν6(E′′) δ(XY3)

ν7(E′) δ(XY ′2)

ν8(E′′) ρr(XY′2)

Fig. 2.23. Normal modes of vibration of trigonal–bipyramidal XY5 molecules.

The majority of trigonal–bipyramidal compounds show the following frequency trend: v5 Equatorial stretch

>

v3 > Axial stretch

v1 Equatorial stretch

>

v2 Axial stretch

Although the IR spectrum of MoCl5 in the gaseous phase has been assigned based on a trigonal–bipyramidal structure [1107] the corresponding spectrum in N2 matrices suggests a tetragonal–pyramidal structure of C4v symmetry [1117]. This is supported by the observation that it shows only two prominent IR banos at 473 (A1) and 408 cm1 (E) in the v(MoCl) region. Normal coordinate analyses on trigonal–bipyramidal XY5 molecules have been carried out by several investigators [1118–1121]. These calculations show that equatorial bonds are stronger than axial bonds. Most neutral XY5 molecules are dimerized or polymerized in the condensed phases. The molecules MoCl5, [1122], NbCl5 [1113,1123], TaCl5 [1124], and WCl5 [1125] are dimeric in the liquid and solid states (D2h), whereas NbF5 and TaF5 are known to be tetrameric in the crystalline state [1126]. Some of these molecules are polymerized even in the gaseous phase. For example, SbF5 is monomeric (D3h) at 350 C but polymeric at 140 C in the gaseous phased [1127], and NbF5 and TaF5 are polymeric in the gaseous phase if the temperature is below 350 C. [1128]. Although PC15 exists as a D3h molecule in the gaseous and liquid states, it has an ionic structure consisting of [PC14]þ[PC16] units in the crystalline state, as

215

Gas Solid Solid Solid Sol’na Solid Gas Matrix Gas Gas Gas Matrix Gas

SbCl5 [SeO5]4 [CuCl5]3 [CdCl5]3 [TiCl5] [TiCl5] VF5 NbCl5 NbBr5 TaCl5 TaBr5 MoF5 MoCl5

b

Nonaqueous solution. May be assigned to n~7. c May be assigned to n~8.

a

Sol’n Sol’na Solid Solid Solid Gas Sol’na Gas Sol’na Gas

a

Phase

[SiF5] [SiCl5] [GeCl5] [SnCl5] [SnBr5] PF5 PCl5 AsF5 AsCl5 SbF5

Molecule

355 756 260 251 348 — 719 (349) 234 406 240 — 390

708 372 348 340 — 817 392 733 369 667

~n1

309 736 — — 302 — 608 (349) 178 324 182 — 313

519 — 236 — — 640 281 642 295 264

~n2

— 775 268 236 355 346 784 396 288 — — 683 —

785 395 310 314 208 944 443 784 385 —

~n3

— 674 — — 178 170 331 126 (93) — — — —

481 271 200 160 106 575 300 400 184 —

~n4 874 550 395 350 256 1026 580 809 437 716 710 400 344 170b 157b 411 385 810 444 315 — — 713 418

~n5

TABLE 2.7a. Vibrational Frequencies of Trigonal-Bipyramidal XY5 Molecules (cm1)

173 806 95c 98c 190 212 282 159 119 181 110 261 200

449 250 200 150 111 532 272 366 220 498

~n6

58 — — — 66 (83) (200) 99 67 54 70 112 100

— — — 66 — 300 102 123 83 90

~n7

120 355 — — 166 — 350 (139) 101 127 93 — 175

— — — 169 — 514 261 388 213 228

~n8

1107 1108 1109 1109 1110 1100 1111,1112 1113 1107,1114 1107 1107 1115 1107

1096 1097 1098 1099 1100 1101 1102,1103 1104 1105 1106

Ref.

216

APPLICATIONS IN INORGANIC CHEMISTRY

proved by Raman spectroscopy [1129]. The importance of the v7 vibration in the intramolecular conversion of pentacoordinate molecules has been discussed by Holmes [1130].

Vibrational spectra of mixed halogeno compounds such as PFnCl5n, [1131,1132] PF3X2(X ¼ Cl, Br), [1133], PHnF5n, [1134,1135], PH3F2, [1136,1137] AsCl3F2 [1138] and AsCl4F [1139] have been assigned. The structure of OSF4 is pseudo-trigonal–bipyramidal with two fluorines and one oxygen occupying the three equatorial positions. Thus, the IR and Raman spectra were assigned on the basis of C2v symmetry [1140]. Since the OPF4 ion is isoelectronic with OSF4, the structure is similar to that of OSF4, and assigment of its Raman spectrum is based on the same symmetry [1141]. The FOsO4 ion assumes a distorted trigonal–bipyramidal structure with three oxygen atoms occupying the equatorial positions, and the IR and Raman spectra were assigned under Cs symmetry [1142]:

Although the Raman spectrum of Csþ [trans-XeO2F3] was reported previously [1143], a later study showed that its structure is Csþ [XeO2F3nXeF2], in which the two oxygen atoms are cis to each other [1144]. In the [F(cis-OsO2F3)2]þ ion, a fluorine bridge connects two trigonal bipyramidal cations in which two oxygen atoms and one fluorine atom occupy the equatorial plane and two fluorine atoms are in axial positions. The Raman spectra of its AsF6 and Sb2F11 salts have been assigned on the basis of

TRIGONAL–BIPYRAMIDAL AND TETRAGONAL–PYRAMIDAL XY5 AND RELATED MOLECULES

217

such a dimeric structure [1145]. MO2F3SbF5 (M ¼ Te, Re) consists of infinite chains of alternating MO2F4 and SbF6 units in which the bridging fluorine atoms on the antimony are trans to each other. The Raman spectra of these and related compounds have been reported [1146]. 2.7.2. Tetragonal–Pyramidal XY5 and ZXY4 Molecules (C4v) Figure 2.24 shows the nine normal modes of vibration of a tetragonal–pyramidal ZXY4 molecule. Only A1 and E vibrations are infrared-active, whereas all nine vibrations belonging to the A1, B1, B2, and E species are Raman–active. Table 2.7b lists the vibrational frequencies of tetragonal–pyramidal XY5 and ZXY4 molecules. In the majority of XY5 molecules, the axial stretching frequency (n1) is higher than the equatorial stretching frequencies (n2, n4, and n7). This is opposite to the trend found for trigonal–bipyramidal XY5 molecules, discussed in the preceding section. For normal coordinate analyses on these compounds, see Refs. [1150], [1153], [1170], and [1173]. An adduct, XeF6:BiF5, is formulated as [XeF5]þ[BiF6] in the solid state; its [XeF5]þ cation is tetragonal–pyramidal [1157]. This should be contrasted to the [XeF5] anion which is pentagonal planar, as shown in the following section. In the OCrF4KrF2 adduct, KrF2 coordinates to the vacant axial position of OCrF4 through a

Z Y

Y X

Y

ν1(A1) ν (XZ)

Y

ν4(B1) νa(XY4)

ν7(E) νd(XY4)

ν2(A1) νs(XY4)

ν3(A1) π(ZXY4)

ν5(B1) δa(XY4)

ν6(B2) δs(XY4)

ν8(E) πd (XY4)

ν9(E) δd(XY4)

Fig. 2.24. Normal modes of vibration of tetragonal–pyramidal ZXY4 molecules.

218

294 557 445 796 666 616 363 723 682 698 740 660

837 1216

932

[OTeF4]2 [OClF4]

[OBrF4]

A1 ~n1

[InCl5] [SbF5]2 [SbCl5] [SF5] [SeF5] [TeF5] [TeCl5] ClF5 BrF5 IF5 NbF5 [XeF5]þ

2

XY5 or ZXY4

525

461 462

283 427 285 522 515 483 254 545 570 593 686 598

A1 ~n2

312

— 339

140 278 180 469 332 280 136 487 365 315 513 355

A1 ~n3

459

287 388 420 (435) 460 505 270 498 535 575 — 635 628 390 350

B1 ~n4

236

— —

193 — — 269 236 231 98 317 (281) (257) — 242

B1 ~n5

248

190 283

165 220 117 342 282 183 82 385 312 273 — 291

B2 ~n6

335 600,550 506 483

274 375,347 300 590 480 452 246 732 644 640 729 650

E ~n7

TABLE 2.7b. Vibrational Frequencies of Trigonal–Pyramidal XY5 and ZXY4 Molecules (cm1)

108 307 255 (435) 399 162 90 500 414 374 261 415 395 — 415,394 ) 421 407 398

E n~8

194 164

129 213

143 142 90 241 202 342 169 301 237 189 103 218

E n~9

1160,1161

1158 1159

1147 1148 1149 1150 1150 1148,1151 1151 1152–1154 1153,1155 1148,1153 1156 1157

Ref.

219

[OIF4] OXeF4 OCrF4 OMoF4 OMoCl4 [OMoCl4] OWF4 OReF4 [OReI4] [NRuCl4] [NRuBr4] [NOsCl4] [NOsBr4] ORuF4 OOsF4

XY5 or ZXY4

888 920 1028 1050 1015 1008 1055 1082 1014 1092 1088 1123 1119 1060 1080

A1 ~n1

TABLE 2.7b. (Continued )

533 567 686 714 450 354 733 685 169 346 224 358 162 685 695

A1 ~n2

273 285 277 267 143 184 248 — 85 197 156 184 122 — —

A1 ~n3 475 527 — — 400 327 631 — 183 304 187 352 156 — —

B1 ~n4 — (230) — — 148 158 328 — — 154 103 149 110 — —

B1 ~n5 214 233 — — 220 167 291 — — 172 128 174 120 — —

B2 ~n6 485 608 744 708 396 364 698 710 240 378 304 365 220 713/710 685

E ~n7 365 365 320 306 256 240 298 302 181 267 211 271 273 310 319

E n~8 124 161 271 238 172 114 236 245 — 163 98 132 98 — —

E n~9 1162 1153 1163,1164 1165–1167 1168,1169 1170 1166,1167 1171 1172 1170 1170 1170 1170 1171 1171

Ref.

220

APPLICATIONS IN INORGANIC CHEMISTRY

predominantly covalent bond [1174]:

As a result, the KrF2 vibrations are markedly shifted in frequency relative to those of free KrF2 [305]: n1

n2

n3

Free KrF2

449 233

596; 580

cm1

Bound KrF2

487 176

550; 542

cm1

2.7.3. Pentagonal–Planar XY5 Molecules (D5h) The first example of a pentagonal–planar XY5 type molecule was the XeF5 anion [1175]. This anion was obtained as the 1:1 adduct of XeF4 with N(CH3)4F, CsF or other alkali fluorides, and its structure was confirmed by X-ray analysis, IR/Raman, and NMR spectroscopy. Figure 2.25 shows the structure of the XeF5 anion, which can be derived by replacing the two axial fluorine ligands of the pentagonal–bipyramidal IF7 molecule with two sterically active lone-pair electrons. The difference in structure between this anion and the XeF5 cation discussed in the preceding section may be understood based on the VSERP theory [1089].

Fig. 2.25. Structures of the XeF5 and XeF3þ ions.

OCTAHEDRAL MOLECULES

221

The XeF5 anion has 12 normal vibrations, which are classified into 00 0 0 00 A1 ðRÞþA2 ðIRÞþ2E1 ðIRÞþ2E2 ðRÞþE2 (inactive) under D5h symmetry. Thus, only 00 0 0 0 three vibrations (A2 and 2E1 ) are IR-active and only three vibrations ðA 1 þ2E 2 Þ are Raman-active. In agreement with this prediction, the observed Raman spectrum of 0 (CH3)4N[XeF5] exhibits three bands at 502 (symmetric stretch, A1 ), 423 (asymmetric 0 0 1 stretch, E2 ), and 377 cm (in-plane bending, E2 ). As discussed in Sec. 1.11, the trigonal–bipyramidal and tetragonal-pyramidal structures can be ruled out since many more IR- and Raman-active bands are expected for these structures. The second example of a pentagonal planar XY5-type molecule is the IF52 ion, which is isoelectronic with the XeF5 ion. It was obtained as the N(CH3)4þ salt from the reaction of N(CH3)4IF4 and N(CH3)4F in CH3CN solution, and the IR and Raman spectra were assigned under D5h symmetry [1176]. 0

2.8. OCTAHEDRAL MOLECULES 2.8.1. Octahedral XY6 Molecules (Oh) Figure 2.26 illustrates the six normal modes of vibration of an octahedral XY6 molecule. Vibrations n1, n2, and n5 are Raman-active, whereas only n3 and n4 are infrared-active. Since n6 is inactive in both, its frequency is estimated by several

Y

Y

X

Y

Y Y

Y ν1(A1g) ν (XY)

ν4(F1u) δ (YXY)

ν2(Eg) ν (XY)

ν3(F1u) ν (XY)

ν5(F2g)

ν6(F2u)

δ (YXY)

δ (YXY)

Fig. 2.26. Normal modes of vibration of octahedral XY6 molecules.

222

APPLICATIONS IN INORGANIC CHEMISTRY

methods including the analysis of nonfundamental frequencies and rotational fine structures of vibration spectra. The G and F matrix elements of the octahedral XY6 molecule are listed in Appendix VII. Table 2.8a lists the vibrational frequencies of a number of hexahalogeno compounds. In general, the order of the stretching frequencies is n1 > n3 ( n2 or n1 < n3 ( n2, depending on the compound. The order of the bending frequencies is n4 > n5 > n6 in most cases. Parker et al. [1263] measured the IR, Raman, and INS (inelastic neutron scattering) spectra of the [FeH6]4 ion and its deuterated analog listed at the end of Table 2.8a. These hexahydrido complexes are unsual in that the v1(A1g) and v2(Eg) frequencies are very close, although they are well separated in other XY6 molecules. This result indicates that the stretching–stretching interaction force constant (frr in Appendix VII.6 is very small in the case of Fe–H(D) bonds. The v6 frequencies of these compounds listed in Table 2.8a were obtained by INS spectroscopy. Several trends are noted in the octahedral XY6 molecules listed in Table 2.8a, as described in Secs. 2.8.1.1–2.8.1.12. 2.8.1.1. Mass of Central Atom Within the same family of the periodic table, the stretching frequencies decrease as the mass of the central atom increases, for example [AlF6]3 v~1 ðcm1 Þ v~3 ðcm1 Þ

541 568

[GaF6]3 > >

[InF6]3 > >

535 481

497 447

[TiF6]3 > >

478 412

The trend in n1 directly reflects the trend in the stretching force constant (and bond strength) since the central atom is not moving in this mode. In n3, however, both X and Yatoms are moving, and the mass effect of the X atom cannot be ignored completely. 2.8.1.2. Oxidation State of Central Atom Across the periodic table, the stretching frequencies increase as the oxidation state of the central atom becomes higher. Thus we have [AlF6]3 1

v~1 ðcm Þ v~3 ðcm1 Þ

541 568

[PF6]

[SiF6]2 <

584 .578

[VF6]3 > >

533 511

As in many other cases, the higher the oxidation state, the higher the frequency. The bending frequencies do not exhibit clear-cut trends.

223

[AlF6] [GaF6]3 [InF6]3 [InCl6]3 [TlF6]3 [TlCl6]3 [TlBr6]3 [SiF6]2 [GeF6]2 [GeCl6]2 [SnF6]2 [SnCl6]2 [SnBr6]2 [SnI6]2 [PbCl6]2 [PF6] [PCl6] [AsF6] [AsCl6] [SbF6] [SbCl6] [SbCl6]3 [SbBr6] [SbBr6]3 [SbI6]3 [BiF6] [BiCl6]3 [BiBr6]3 [BiI6]3

3

Molecule

541 535 497 277 478 280 161 663 671 318 592 311 190 122 281 756 360 689 337 668 330 327 192 180 107 590 259 156 114

~n1

(400) (398) (395) 193 387 262 153 477 478 213 477 229 144 93 209 585,570 283 573 289 558 282 274 169 153 96 547 215 130 103

~n2 568 481 447 250 412 294 190,195 741 624 310 559 303 224 161 262 865,835 444 700 333 669 353 — 239,224 180 108 585 172 128 96

~n3

TABLE 2.8a. Vibrational Frequencies of Octahedral XY6 Molecules (cm1)

387 298 226 157 202 222,246 134,156 483 365,355 213 300 166 118 84 142 559,530 285 385 220 350 180 — 119 107 82 — 130 75 (59)

~n4 322 281 229 (149) 209 155 95 408 333/325 191 252 158 109 78 139 480 468 238 375 202 294 175 137 103,78 73 54 231/247 115 62 54

~n5 (228) (198) (162) — (148) (136) (80) — — — — — — — — — — (252) — — — — — — — — — — —

~n6a

(continued)

1177,1178 1178,1179 1177,1178 1180 1177,1178 1181 1181 1182 1183 1184 1182 1185 1186 1187 1185 1188 1184,1189 1182 1184 1182,1190 1191 1192 1193 1194,1195 1194,1195 1190,1196 1194,1195 1194,1195 1194,1195

Ref.

224

~n2

643 658 255 155 672 250 192 153 384 630 660 454 530 375 67 382 334 380 370 265 (440) 271 278 — (416) 237 144 (389) 257 142 538

~n1

775 708 299 176 698 298 264 174 525 679 658 568 595 495 119 476 443 473 491 295 618 320 322 192 589 327 194 572 325 197 676

Molecule

SF6 SeF6 [SeCl6]2 [SeBr6]2 TeF6 [TeCl6]2 [TeCl6]3 [TeBr6]2 [ClF6] [ClF6]þ [BrF6]þ [BrF6] [AuF6] [ScF6]3 [ScI6]3 [XF6]3 [LaF6]3 [GaF6]3 [YbF6]3 [CeCl6]2 [TiF6]2 [TiCl6]2 [TiCl6]3 [TiBr6]2 [ZrF6]2 [ZrCl6]2 [ZrBr6]2 [HfF6]2 [HfCl6]2 [HfBr6]2 [VF6]

TABLE 2.8a. (Continued )

939 780 280 221 752 250 230 198 — 890 — 400 — 458 — 160 130 140 156 268 615,660 316 304,290 244 537,522 290 223 448,490 275 189 646

~n3 614 437 160–140 129 325 136 146 — — 582 — 204,184 — 257 — 74 63 72 70 117 315,281 183 — 119 241,192 150 106 217,184 145 102 300

~n4 524 403 165 138 312 140 (135) 75 289 513 405 250 225 235 80 194 171 185 196 120 308,300 173 175 115 258,244 153 99 259,247 156 101 330

~n5 (347) (264) — — (197) — — — — — — (138) — — — — — — — (86) — — — — — — — — (80) — —

~n6a

1197–1199 1197,1198 1180 1200 1197,1198 1201 1170 1201,1202 1203 1204 1205 1206 1207 1208 898 1209 1209 1209 1209 1210 1211 1212 1213 1212 1214 1215 1216 1211 1185 1216 1217

Ref.

225

584 533 — 683 368 — — — — 692 378 — — — — (720) 286 741.8 685 356 329

305

770 437

382

341

[MoCl6]3

WF6 WCl6

[WCl6]

[WCl6]2

~n1

[VF6] [VF6]3 [VCl6]2 [NbF6] [NbCl6] [NbCl6]2 [NbBr6] [NbBr6] [NbI6] [TaF6] [TaCl6] [TaCl6]2 [TaBr6] [TaBr6]2 [Tal6] CrF6 [CrCl6]3 92 MoF6 [MoF6]2 [MoCl6] [MoCl6]2

2

Molecule

TABLE 2.8b. (Continued )

—

—

676 331

—

— — — 562 288 — — — — 581 298 — — — — (650) 237 652.0 598 — —

~n2

293

578 511 355,305 602 333 314 240 216 236 180 560 330 297 234 223 217 160 790 315 749.5 653 327 308 ) 268 286 302 711 373 332 312

~n3

166 150

157

258 160

273 292 — 244 162 165 — 112 70,66 240 158 160 — 109 80 (266) 199 265.7 250 162 168 167 187

~n4

—

168

321 182

150

— — — 280 183 — — — — 272 180 — — — — (309) 162 317 274 — 154

~n5

—

—

(127) —

—

— — — — — — — — — (192) — — — — — (110) 182 117 — — —

~n6a

(continued)

1230

1230

1197,1198 1185

1230

1217 1217 1218 1219,1220 1185 1220,1221 1222 1220,1221 1220,1223 1224 1185 1185 1222 1221 1223 1225,1226 1227 1228 1229 1230 1230

Ref.

226

[MnF6] TcF6 [ReF6]þ ReF6 [ReF6]2 [ReCl6] [ReCl6]2 [ReBr6]2 [FeF6]3 [NiF6]2 [PdF6]2 [PdCl6]2 [PdBr6]2 PtF6 [PtF6]2 [PtCl6]2 [PtBr6]2 [PtI6]2 [FeF6]3 RuF6 [RuCl6]2 OsF6 [OsCl6] [OsCl6]2 [OsBr6]2 [OsI6]2 RhF6 [RhCL6]2 IrF6 [IrCl6]2

2

Molecule

592 713 797 754 611 — 346 213 507 555 558 318 198 656 611 348 213 — 538 (675) — 731 375 345.3 210.6 152 (634) — 701.1 341.3

~n1

TABLE 2.8a. (Continued )

508 (639) 734 (671) 530 — (275) (174) — 512 276 289 176 (601) 576 318 190 — 374 (624) — (668) 302 245.2 169.2 121 (595) 320 647.3 289.4

~n2 620 748 783 715 535 318 313 217 456 648 615 346 253 705 571 342 243 186 — 735 346 720 325 326 227 170 724 335 715.6 —

~n3 335 275 353 257 249 161 172 118 — 332 — 200 130 273 281 183 146 46 — 275 188 268 168 176 122 91 283 — 267,5 —

~n4 308 (297) 359 (295) 221 — 159 104 273 307/298 212 178 100 (242) 210 171 137 — 253 (283) — (276) 183 160 100 80 (269) 253 265.2 159.6

~n5 — (145) — (147) (181) — — — — — — — — (211) (143) (88) — — — (186) — (205) — — — — (192) — 208.3c —

~n6a

1231 1198 1232 1198 1233 1234 1185,1235 1236 1179 1237 1238 1239 1240 1198 1241 1186 1239 1242 1243 1198,1244 1244 1198 1245 1246,1247 1246,1248 1249 1198,1244 1250 1251,1252 1253

Ref.

227

1342

[FeD6]4

1363

— 175.1 133 255 530 — 273 237 — 525 574 — 519 — 1878

~n2

1260

296 — 175 259 619 525 310 262 214 618 604 265 612 — 1746

~n3

a

pﬃﬃ The value of v6 can also be estimated by the relation, v6 ¼ v5 = 2 [1264,1265]. b Ground state. c Excited state. d INS frequency.

— 209.6 149 294 666 — 343 299 — 646 643 310 625 615 1873

~n1

[IrCl6] [IrBr6]2 [IrI6]3 [ThCl6]2 UF6 [UF6] [UCl6] [UCl6]2 [Ubr6] NpF6b NpF6c [NpCl6]2 PuF6b PuF6c [FeH6]4

3

Molecule

TABLE 2.8b. (Continued )

655d

200 — 87 — 184 173 122 114 87 198 191 117 200 202 899d

~n4 — 103.2 88 114 200 — 136 121 — 208 — 128 209 198 1019 1057 730

~n5

549d

— — — — — — — (80) — 169 138/149 — 177 172 836d

~n6a

1263

1242 1253 1254 1255 1256 1257 1257,1258 1255 1257 1259 1259 1260 1261 1262 1263

Ref.

228

APPLICATIONS IN INORGANIC CHEMISTRY

2.8.1.3. Halogen Series of series, for example

The effect of changing the halogen is seen in a number

[SnF6]2 v~1 ðcm1 Þ v~3 ðcm1 Þ

592 559

[SnCl]2 > >

[SnBr6]2 > >

311 303

190 224

[SnI6]2 > >

122 161

The stretching force constants also follow the same order. The ratios n(MBr)/n (MCl) and n(MI)/n(MCl) are about 0.61 and 0.42, respectively, for n1, and about 0.76 and 0.62, respectively, for n3. 2.8.1.4. Electronic Structure follows: v~3 ðcm1 Þ v~3 ðcm1 Þ

In the [MCl6]3 series, n3 and n4 change as

Cr3þ(d3)

Mn3þ (d 4)

Fe3þ(d 5)

In3þ

315 200

342 183

248 184

248 161

3 2 All these metal ions are in the high-spin state. For Fe3þ (t2g eg ), occupation of the antibonding orbitals lowers n3 drastically in comparison to the Cr3þ complex; its n3 is comparable to that of the In3þ complex, whose n3 is lowered because of the increased mass of the metal. On the other hand, the v3 of [MnCl6]3 is higher than that of [CrCl6]3 because the static Jahn–Teller effect of the Mn3þ ion causes a tetragonal distortion [1210].

2.8.1.5. Electronic Excited States Table 2.8a includes PuF6 and NpF6, for which vibrational frequencies have been determined both for the ground and excited states. The excited frequencies were obtained from the analysis of vibrational fine structure of electronic spectra. 2.8.1.6. Raman Intensities The Raman intensity of an XY6 molecule normally follows the order I(n1) > I(n2) > I(n5). Adams and Downs [1266] noted that I (n2)/I(n1) is 0.5 to 1 for [TeCl6]2 and [TeBr6]2, although it normally ranges from 0.05 to 0.1. Furthermore, they observed n3, which is not allowed in Raman spectra. From these and other items of evidence, they proposed that the Oh selection rule breaks down in [TeX6]2 because less symmetric electronic excited states perturb the Oh ground state. They also noted the distortion of [SbX6]3 ions to C3v symmetry from their Raman spectra in solution. Woodward and Creighton [1267] noted that I(n2) > I(n1) holds in the aqueous Raman spectra of Na2PtX6 (X ¼ Cl, Br) and Na2PdCl6, and attributed this unusual trend to the presence of six nonbonding d electrons in the valence shell. 2.8.1.7. Symmetry Lowering Table 2.8a contains XY6-type ions that exhibit splitting of degenerate vibrations due to lowering of symmetry in the crystalline state. For example, the n3 vibration of the ReF62 ion in K2 [ReF6] splits into two bands (565 and 543 cm1) due to trigonal distortion in the crystalline state. Similar splitting is observed for the v4 vibration (222 and 257 cm1) [1268,1269].

OCTAHEDRAL MOLECULES

229

More examples are found in the [ReCl6] (C4v) [1270] and [CuCl6]4(D2h) [1271] ions in the crystalline state. 2.8.1.8. Adduct Formation Hunt et al. [1272] obtained two types of 1:1 adducts by codepositing UF6 with HF in excess Ar at 12 K; UF6–HF exhibits a strong band at 3848 cm1, while the anti-hydrogen-bonded complex, UF6FH, shows a weak, broad band at 3903 cm1 in IR spectra. Similar experiments show, however, WF6HF exhibits a weak band at 3911 cm1, while WF6HF shows a strong band at 3884 cm1. These results indicate that, in hydrogen-bonded complexes, the order of proton affinity is WF6 < UF6, with HF serving as a proton donor and that, in antihydrogen-bonded complexes, the order of acidity is UF6 < WF6, with HF serving as a Lewis base. Relative yields of these two types are determined by the acidity/basicity of these fluorides. 2.8.1.9. Anomalous Polarization Preresonance Raman spectra of 17 XY6type metal halides have been measured by Bosworth and Clark [1192]. Hamaguchi et al. [1273] observed anomalous polarization (Sec. 1.23) for all the Raman bands of the IrCl62 ion measured under resonance condition. Figure 2.27 shows the polarized Raman spectra (488.0 nm excitation) of this ion in aqueous solution together with band assignments and depolarization ratios for each band. Since the Raman tensor for the n1, n2, and v5 vibrations cannot have an antisymmetric part, group theoretical arguments such as those used for heme proteins (Sec. 1.23) are not applicable to octahedral XY6 ions. A more detailed study by Hamaguchi [1274] shows that electronic degeneracy in the IrCl62 (5d5) ion induces an antisymmetric part of vibrational Raman tensors. Anomalous polarization was also observed for the [IrBr6]2 ion [1275]. 2.8.1.10. Jahn–Teller Effect Weinstock et al. [1276] noted that the combination bands (n1þ n3) and (n2þ n3) appear with similar frequencies, intensities, and shapes in the infrared Spectra of MoF6(d 0) and RhF6(d3). As shown in Fig. 2.28 however, (n2þ n3) was very broad and weak in TcF6(d1), ReF6(d1), and RuF6(d2), and OsF6(d2). This anomaly was attributed to a dynamic Jahn-Teller effect. The static Jahn–Teller effect does not seem to operate in these compounds since no splittings of the triply degenerate fundamentals were observed. Perhaps the most fascinating XY6-type molecule is XeF6. In their earlier work, Claassen et al. [1277] suggested the distortion of XeF6 from Oh symmetry since they observed two stretching bands in infrared and three stretching bands in Raman spectra. It was not possible, however to determine the precise structure of XeF6 until they [1278] carried out a detailed IR, Raman, and electronic spectral study of XeF6 vapor as a function of temperature. They were then able to show that XeF6 consists of the three electronic isomers shown in Fig. 2.29 and to explain subtle differences in spectra at different temperatures as a shift of equilibrium among these three isomers. 2.8.1.11. 235U / 238U Isotope Shift UF6 in natural abundance consists of 99.3% 238 UF6 and 0.7 % 235 UF6 . Since the isotope shifts between them is extremely small, it

230

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.27. Polarized Raman spectra of [(n-C4H9)4N]2 [IrCl6] in acetonitrile (488 nm excitation) [1273].

Fig. 2.28. Band profiles for (n1þ n3) and (n2þ n3) for the 4d transition series hexafluorides [1276].

OCTAHEDRAL MOLECULES

231

Fig. 2.29. Structures of three isomers of XeF6 and their IR-active stretching frequencies (cm1). The Xe atom at the center is not shown.

is difficult to observe the vibrational spectrum of the latter without serious overlapping of the former. Takami et al. [1279] have overcome this difficulty by measuring the spectrum of a UF6 (natural abundance)/Ar mixture as a supercooled vapor (below 100 K) using the cold-jet technique [1280]. Figure 2.30 shows the high-resolution IR spectrum of UF6 in the n3 region thus obtained [1279]. It exhibits sharp rotational fine structures [1281] without hot band contributions, and the rotational components of 238 UF6 and 235 UF6 are well separated; the former shows the Q-branch as well as R (0)–R(6) branches whereas the latter exhibits the Q-branch between R(5) and R(6) of the former. Although the isotope shift is estimated to be only 0.6040 cm1, Takeuchi et al. [1282] have found that UF6 gas in the supercooled state (

Infrared and Raman Spectra of Inorganic and Coordination Compounds Part A: Theory and Applications in Inorganic Chemistry Sixth Edition

Kazuo Nakamoto Wehr Professor Emeritus of Chemistry Marquette University

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright # 2009 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data is available.

ISBN 978-0-471-74339-2

Printed in the United States of America 10 9

8 7 6 5

4 3 2 1

Contents

PREFACE TO THE SIXTH EDITION

ix

ABBREVIATIONS

xi

Chapter 1. Theory of Normal Vibrations 1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 1.7. 1.8. 1.9. 1.10. 1.11. 1.12. 1.13. 1.14. 1.15. 1.16. 1.17.

1

Origin of Molecular Spectra / 1 Origin of Infrared and Raman Spectra / 5 Vibration of a Diatomic Molecule / 9 Normal Coordinates and Normal Vibrations / 15 Symmetry Elements and Point Groups / 21 Symmetry of Normal Vibrations and Selection Rules / 25 Introduction to Group Theory / 34 The Number of Normal Vibrations for Each Species / 39 Internal Coordinates / 46 Selection Rules for Infrared and Raman Spectra / 49 Structure Determination / 56 Principle of the GF Matrix Method / 58 Utilization of Symmetry Properties / 65 Potential Fields and Force Constants / 71 Solution of the Secular Equation / 75 Vibrational Frequencies of Isotopic Molecules / 77 Metal–Isotope Spectroscopy / 79 v

vi

CONTENTS

1.18. 1.19. 1.20. 1.21. 1.22. 1.23. 1.24. 1.25. 1.26. 1.27. 1.28. 1.29. 1.30. 1.31. 1.32.

Group Frequencies and Band Assignments / 82 Intensity of Infrared Absorption / 88 Depolarization of Raman Lines / 90 Intensity of Raman Scattering / 94 Principle of Resonance Raman Spectroscopy / 98 Resonance Raman Spectra / 101 Theoretical Calculation of Vibrational Frequencies / 106 Vibrational Spectra in Gaseous Phase and Inert Gas Matrices / 109 Matrix Cocondensation Reactions / 112 Symmetry in Crystals / 115 Vibrational Analysis of Crystals / 119 The Correlation Method / 124 Lattice Vibrations / 129 Polarized Spectra of Single Crystals / 133 Vibrational Analysis of Ceramic Superconductors / 136 References / 141

Chapter 2. Applications in Inorganic Chemistry 2.1. 2.2. 2.3. 2.4. 2.5. 2.6. 2.7. 2.8. 2.9. 2.10. 2.11. 2.12. 2.13. 2.14. 2.15. 2.16. 2.17. 2.18.

Diatomic Molecules / 149 Triatomic Molecules / 159 Pyramidal Four-Atom Molecules / 173 Planar Four-Atom Molecules / 180 Other Four-Atom Molecules / 187 Tetrahedral and Square–Planar Five-Atom Molecules / 192 Trigonal–Bipyramidal and Tetragonal–Pyramidal XY5 and Related Molecules / 213 Octahedral Molecules / 221 XY7 and XY8 Molecules / 237 X2Y4 and X2Y6 Molecules / 239 X2Y7, X2Y8, X2Y9, and X2Y10 Molecules / 244 Metal Cluster Compounds / 250 Compounds of Boron / 254 Compounds of Carbon / 258 Compounds of Silicon, Germanium, and Other Group IVB Elements / 276 Compounds of Nitrogen / 279 Compounds of Phosphorus and Other Group VB Elements / 285 Compounds of Sulfur and Selenium / 292

149

CONTENTS

vii

2.19. Compounds of Halogen / 296 References / 299 Appendixes

355

I. Point Groups and Their Character Tables / 355 II. Matrix Algebra / 368 III. General Formulas for Calculating the Number of Normal Vibrations in Each Species / 373 IV. Direct Products of Irreducible Representations / 377 V. Number of Infrared- and Raman-Active Stretching Vibrations for MXn Ym -Type Molecules / 378 VI. Derivation of Eq. 1.113 / 379 VII. The G and F Matrix Elements of Typical Molecules / 382 VIII. Group Frequency Charts / 388 IX. Correlation Tables / 393 X. Site Symmetry for the 230 Space Groups / 407 Index

415

Preface to the Sixth Edition

Since the fifth edition was published in 1996, a number of new developments have been made in the field of infrared and Raman spectra of inorganic and coordination compounds. The sixth edition is intended to emphasize new important developments as well as to catch up with the ever-increasing new literature. Major changes are described below. Part A. Chapter 1 (‘‘Theory of Normal Vibrations”) includes two new sections. Section 1.24 explains the procedure for calculating vibrational frequencies on the basis of density functional theory (DFT). The DFT method is currently used almost routinely to determine molecular structures and to calculate vibrational parameters. Section 1.26 describes new developments in matrix cocondensation techniques. More recently, a large number of novel inorganic and coordination compounds have been prepared by using this technique, and their structures have been determined and vibrational assignments have been made on the basis of results of DFT calculations. Chapter 2 (‘‘Applications in Inorganic Chemistry”) has been updated extensively, resulting in a total number of references of over 1800. In particular, sections on triangular X3- and tetrahedral X4-type molecules have been added as Secs. 2.2 and 2.5, respectively. In Sec. 2.8, the rotational–vibrational spectrum of the octahedral UF6 molecule is shown to demonstrate how an extremely small metal isotope shift by 235U/238U substitution (only 0.6040 cm 1) can be measured. Section 2.14 (‘‘Compounds of Carbon”) has been expanded to show significant applications of vibrational spectroscopy to the structural determination of fullerences, endohedral fullerenes, and carbon nanotubes. Vibrational data on a number of novel inorganic compounds prepared most recently have been added throughout Chapter 2. Part B. Chapter 3 (‘‘Applications in Coordination Chemistry”) contains two new Sections: Sec. 3.6 (‘‘Metallochlorins, Chlorophylls, and Metallophthalocyanines”) ix

x

PREFACE TO THE SIXTH EDITION

and Sec. 3.19 (‘‘Complexes of Carbon Dioxide”). The total number of references has approached 1700 because of substantial expansion of other sections such as Secs. 3.5, 3.18, 3.20, 3.22, and 3.28. Chapter 4 (‘‘Applications in Organometallic Chemistry”) includes new types of organometallic compounds obtained by matrix cocondensation techniques (Sec. 4.1). In Chapter 5 (‘‘Applications in Bioinorganic Chemistry”), a new section (Sec. 5.4) has been added, and several sections such as Secs. 5.3, 5.7, and 5.9 have been expanded to include many important new developments. I would like to express my sincere thanks to all who helped me in preparing this edition. Special thanks go to Prof. E. L. Varetti (University of La Plata, Argentina), Prof. S. Guha (University of Missouri, Columbia), and Prof. L. Andrews (University of Virginia) for their help in writing new sections in Chapter 1 of Part A. My thanks also go to all the authors and publishers who gave me permission to reproduce their figures in this and previous editions. Finally, I would like to thank the staff of John Raynor Science Library of Marquette University for their help in collecting new references. Milwaukee, Wisconsin March 2008

KAZUO NAKAMOTO

Abbreviations

Several different groups of acronyms and other abbreviations are used: 1. IR, infrared; R, Raman; RR, resonance Raman; p, polarized; dp, depolarized; ap, anomalous polarization; ia, inactive. 2. n, stretching; d, in-plane bending or deformation; rw, wagging; rr, rocking; rt, twisting; p, out-of-plane bending. Subscripts, a, s, and d denote antisymmetric, symmetric, and degenerate modes, respectively. Approximate normal modes of vibration corresponding to these vibrations are given in Figs. 1.25 and 1.26. 3. DFT, density functional theory; NCA, normal coordinate analysis; GVF, generalized valence force field; UBF, Urey–Bradley force field. 4. M, metal; L, ligand; X, halogen; R, alkyl group. 5. g, gas; l, liquid; s, solid; m or mat, matrix; sol’n or sl, solution; (gr) or (ex), ground or excited state. 6. Me, methyl; Et, ethyl; Pr, propyl; Bu, butyl; Ph, phenyl; Cp, cyclopentadienyl; OAc, acetate ion; py, pyridine; pic, pycoline; en, ethylenediamine. Abbreviations of other ligands are given when they appear in the text. In the tables of observed frequencies, values in parentheses are calculated or estimated values unless otherwise stated.

xi

Chapter

1

Theory of Normal Vibrations 1.1. ORIGIN OF MOLECULAR SPECTRA As a first approximation, the energy of the molecule can be separated into three additive components associated with (1) the motion of the electrons in the molecule,* (2) the vibrations of the constituent atoms, and (3) the rotation of the molecule as a whole: Etotal ¼ Eel þ Evib þ Erot

ð1:1Þ

The basis for this separation lies in the fact that electronic transitions occur on a much shorter timescale, and rotational transitions occur on a much longer timescale, than 10 vibrational transitions. The translational energy of the molecule may be ignored in this discussion because it is essentially not quantized. If a molecule is placed in an electromagnetic field (e.g., light), a transfer of energy from the field to the molecule will occur when Bohr’s frequency condition is satisfied: DE ¼ hn

ð1:2Þ

*Hereafter the word molecule may also represent an ion. Infrared and Raman Spectra of Inorganic and Coordination Compounds, Sixth Edition, Part A: Theory and Applications in Inorganic Chemistry, by Kazuo Nakamoto Copyright 2009 John Wiley & Sons, Inc.

1

2

THEORY OF NORMAL VIBRATIONS

where DE is the difference in energy between two quantized states, h is Planck’s constant (6.625 1027 erg s), and n is the frequency of the light. Here, the frequency is the number of electromagnetic waves in the distance that light travels in one second: c ð1:3Þ n¼ l where c is the velocity of light (3 1010 cm s1) and l is the wavelength of the electromagnetic wave. If l has the units of cm, n has dimensions of (cm s1)/cm ¼ s1, which is also called “Hertz (Hz).” The wavenumber ð~nÞ defined by 1 ~n ¼ ð1:4Þ l is most commonly used in vibrational spectroscopy. It has the dimension of cm1. By combining Eqs. 1.3 and 1.4, we obtain ~n ¼

1 n ¼ l c

or n ¼

c ¼ c~n l

ð1:5Þ

Although the dimensions of n and ~n differ from one another, it is convenient to use them interchangeably. Thus, an expression such as “a frequency shift of 5 cm1” is used throughout this book. Using Eq. 1.5, Bohr’s condition (Eq. 1.2) is written as DE ¼ hc~n

ð1:6Þ

Since h and c are known constants, DE can be expressed in units such as* 1ðcm 1 Þ ¼ 1:99 10 16 ðerg molecule 1 Þ ¼ 2:86 ðcal mol 1 Þ ¼ 1:24 10 4 ðeV molecule 1 Þ Suppose that DE ¼ E2 E1

ð1:7Þ

where E2 and E1 are the energies of the excited and ground states, respectively. Then, the molecule “absorbs” DE when it is excited from E1 to E2 and “emits” DE when it reverts from E2 to E1. Figure 1.1 shows the regions of the electromagnetic spectrum where DE is indicated in ~n, l, and n. In this book, we are concerned mainly with vibrational transitions which are observed in infrared (IR) or Raman (R) spectra. *Use conversion factors such as Avogadro’s number N0 ¼ 6:023 1023 ðmol 1 Þ 1 ðcalÞ ¼ 4:1846 107 ðergÞ 1 ðeVÞ ¼ 1:6021 10 12 ðerg molecule 1 Þ

ORIGIN OF MOLECULAR SPECTRA

3

Fig. 1.1. Regions of the electromagnetic spectrum and energy units.

These transitions appear in the 102 104 cm1 region, and originate from vibrations of nuclei constituting the molecule. Rotational transitions occur in the 1–102 cm1 region (microwave region) because rotational levels are relatively close to each other, whereas electronic transitions are observed in the 104–106 cm1 region (UV–visible region) because their energy levels are far apart. However, such division is somewhat arbitrary, for pure rotational spectra may appear in the far-infrared region if transitions to higher excited states are involved, and pure electronic transitions may appear in the near-infrared region if electronic levels are closely spaced. Figure 1.2 illustrates transitions of the three types mentioned for a diatomic molecule. As the figure shows, rotational intervals tend to increase as the rotational quantum number J increases, whereas vibrational intervals tend to decrease as the vibrational quantum number v increases. The dashed line below each electronic level indicates the “zero-point energy” that must exist even at a temperature of absolute zero as a result of Heisenberg’s uncertainty principle: 1 E0 ¼ hn ð1:8Þ 2 It should be emphasized that not all transitions between these levels are possible. To see whether the transition is “allowed” or “forbidden,” the relevant selection rule must be examined. This, in turn, is determined by the symmetry of the molecule. As expected from Fig. 1.2, electronic spectra are very complicated because they are accompanied by vibrational as well as rotational fine structure. The rotational fine structure in the electronic spectrum can be observed if a molecule is simple and the spectrum is measured in the gaseous state under high resolution. The vibrational fine structure of the electronic spectrum is easier to observe than the rotational fine structure, and can provide structural and bonding information about molecules in electronic excited states. Vibrational spectra are accompanied by rotational transitions. Figure 1.3 shows the rotational fine structure observed for the gaseous ammonia molecule. In most polyatomic molecules, however, such a rotational fine structure is not observed

4

THEORY OF NORMAL VIBRATIONS

1

υ” = 0

Zero point energy

Electronic excited state

Pure electronic transition

4

3

2 6 4 2 J” = 0 6 4 2 J” = 0

1

Pure rotational transition

Pure vibrational transition υ’ = 0

Zero point energy

Electronic ground state

Fig. 1.2. Energy level of a diatomic molecule (the actual spacings of electronic levels are much larger, and those of rotational levels are much smaller, than that shown in the figure).

Fig. 1.3. Rotational fine structure of gaseous NH3.

ORIGIN OF INFRARED AND RAMAN SPECTRA

5

because the rotational levels are closely spaced as a result of relatively large moments of inertia. Vibrational spectra obtained in solution do not exhibit rotational fine structure, since molecular collisions occur before a rotation is completed and the levels of the individual molecules are perturbed differently. The selection rule allows any transitions corresponding to Dv ¼ 1 if the molecule is assumed to be a harmonic oscillator (Sec. 1.3), Under ordinary conditions, however, only the fundamentals that originate in the transition from v ¼ 0 to v ¼ 1 in the electronic ground state can be observed. This is because the Maxwell–Boltzmann distribution law requires that the ratio of population at v ¼ 0 and v ¼ 1 states is given by R¼

Pðv ¼ 1Þ ¼ e DEv =kT Pðv ¼ 0Þ

ð1:9Þ

where DEv is the vibrational frequency (cm1) and kT ¼ 208 (cm1) at room temperature. In the case of H2, DEv ¼ 4160 cm1 and R ¼ 2.16 109. Thus, almost all molecules are at v ¼ 0. However, the population at v ¼ 1 increases as DEv becomes small. For example, R ¼ 0.36 for I2 (DEv ¼ 213 cm1) Then, about 27% of the molecules are at v ¼ 1 state, and the transition from v ¼ 1 to v ¼ 2 can be observed as a “hot band.” In addition to the harmonic oscillator selection rule, another restriction results from the symmetry of the molecule (Sec. 1.10). Thus, the number of allowed transitions in polyatomic molecules is greatly reduced. Overtones and combination bands* of these fundamentals are forbidden by the selection rule. However, they are weakly observed in the spectrum because of the anharmonicity of the vibration (Sec. 1.3). Since they are less important than the fundamentals, they will be discussed only when necessary. 1.2. ORIGIN OF INFRARED AND RAMAN SPECTRA As stated previously, vibrational transitions can be observed as infrared (IR) or Raman spectra.† However, the physical origins of these two spectra are markedly different. Infrared (absorption) spectra originate in photons in the infrared region that are absorbed by transitions between two vibrational levels of the molecule in the electronic ground state. On the other hand, Raman spectra have their origin in the electronic polarization caused by ultraviolet, visible, and near-IR light. If a molecule is irradiated by monochromatic light of frequency n (laser), then, because of electronic polarization induced in the molecule by this incident beam, the light of frequency n (“Rayleigh scattering”) as well as that of frequency n ni (“Raman scattering”) is scattered where ni represents a vibrational frequency of the molecule. Thus, Raman spectra are presented as shifts from the incident frequency in the ultraviolet, visible, and nearIR region. Figure 1.4 illustrates the difference between IR and Raman techniques. *Overtones represent some multiples of the fundamental, whereas combination bands arise from the sum or difference of two or more fundamentals. † Raman spectra were first observed by C. V. Raman [Indian J. Phys. 2, 387 (1928)] and C. V. Raman and K. S. Krishnan [Nature 121, 501 (1928)].

6

THEORY OF NORMAL VIBRATIONS

Fig. 1.4. Mechanisms of infrared absorption and Raman scattering.

Although Raman scattering is much weaker than Rayleigh scattering (by a factor of 103–105), it is still possible to observe the former by using a strong exciting source. In the past, the mercury lines at 435.8 nm (22.938 cm1) and 404.7 nm (24,705 cm1) from a low-pressure mercury arc were used to observe Raman scattering. However, the advent of lasers revolutionized Raman spectroscopy. Lasers provide strong, coherent monochromatic light in a wide range of wavelengths, as listed in Table 1.1. In the case of resonance Raman spectroscopy (Sec. 1.22), the exciting frequency is chosen so as to fall inside the electronic absorption band. The degree of resonance enhancement varies as a function of the exciting frequency and reaches a maximum when the exciting frequency coincides with that of the electronic absorption maximum. It is possible to change the exciting frequency continuously by pumping dye lasers with powerful gas or pulsed lasers. The origin of Raman spectra can be explained by an elementary classical theory. Consider a light wave of frequency n with an electric field strength E. Since E fluctuates at frequency n, we can write E ¼ E0 cos 2pnt

ð1:10Þ

where E0 is the amplitude and t the time. If a diatomic molecule is irradiated by this light, the dipole moment P given by P ¼ aE ¼ aE0 cos 2pnt

ð1:11Þ

ORIGIN OF INFRARED AND RAMAN SPECTRA

7

TABLE 1.1. Some Representative Laser Lines for Raman Spectroscopy Lasera

Mode

Gas lasers Ar–ion

CW

Kr–ion

CW

He–Ne He–Cd Nitrogen Excimer (XeCl)

CW CW Pulsed

Wavelength (nm) 488.0 (blue) 514.5 (green) 413.1 (violet) 530.9 (green/yellow) 647.1 (red) 632.8 (red) 441.6 (blue/violet) 337.1 (UV) 308 (UV)

~n ðcm 1 Þ 20491.8 19436.3 24207.2 18835.9 15453.6 15802.8 22644.9 29664.7 32467.5

Solid-state lasers CW or pulsed 1064 (near-IR) 9398.4 Nd:YAGb Liquid lasers A variety of dye solutions are pumped by strong CW or pulsed-laser sources; a wide range (440–800 nm) can be covered continuously by choosing proper organic dyes a

Acronym of light amplification by stimulated emission of radiation. Acronym of neodymium-doped yttrium aluminum garnet.

b

Source: For more information, see Nakamoto and collegues [21,26].

is induced. Here a is a proportionality constant and is called the polarizability. If the molecule is vibrating with frequency ni, the nuclear displacement q is written as q ¼ q0 cos 2pni t

ð1:12Þ

where q0 is the vibrational amplitude. For small amplitudes of vibration, a is a linear function of q. Thus, we can write ! qa a ¼ a0 þ q ð1:13Þ qq 0

Here, a0 is the polarizability at the equilibrium position, and (qa/qq)0 is the rate of change of a with respect to the change in q, evaluated at the equilibrium position. If we combine Eqs. 1.11–1.13, we have P ¼ aE0 cos 2pnt

! qa ¼ a0 E0 cos 2pnt þ q0 E0 cos 2pnt cos 2pni t qq 0

¼ a0 E0 cos 2pnt ! 1 qa þ q0 E0 fcos½2pðn þ ni Þt þ cos½2pðn ni Þtg 2 qq 0

ð1:14Þ

8

THEORY OF NORMAL VIBRATIONS

Resonance Raman

E1

ν + ν1

ν

ν

ν – ν1

Normal Raman

υ=1 υ=0

S

A

S

A

E0

Fig. 1.5. Mechanisms of normal and resonance Raman scattering. S and A denote Stokes and anti-Stokes scattering, respectively. The shaded areas indicate the broadening of rotationalvibrational levels in the liquid and solid states (Sec. 1.22).

According to classical theory, the first term describes an oscillating dipole that radiates light of frequency n (Rayleigh scattering). The second term gives the Raman scattering of frequencies n þ ni (anti-Stokes) and n ni (Stokes). If (qa/qq)0 is zero, the second term vanishes. Thus, the vibration is not Raman-active unless the polarizability changes during the vibration. Figure 1.5 illustrates the mechanisms of normal and resonance Raman (RR) scattering. In the former, the energy of the exciting line falls far below that required to excite the first electronic transition. In the latter, the energy of the exciting line coincides with that of an electronic transition.* If the photon is absorbed and then emitted during the process, it is called resonance fluorescence (RF). Although the conceptual difference between resonance Raman scattering and resonance fluorescence is subtle, there are several experimental differences which can be used to *If the exciting line is close to but not inside an electronic absorption band, the process is called preresonance Raman scattering.

VIBRATION OF A DIATOMIC MOLECULE

9

Fig. 1.6. Raman spectrum of CCI4 (488.0 nm excitation).

distinguish between these two phenomena. For example, in RF spectra all lines are depolarized, whereas in RR spectra some are polarized and others are depolarized. Additionally, RR bands tend to be broad and weak compared with RF bands [66,67]. In the case of Stokes lines, the molecule at v ¼ 0 is excited to the v ¼ 1 state by scattering light of frequency n ni. Anti-Stokes lines arise when the molecule initially in the v ¼ 1 state scatters radiation of frequency n þ n1 and reverts to the v ¼ 0 state. Since the population of molecules is larger at v ¼ 0 than at v ¼ 1 (Maxwell–Bottzmann distribution law), the Stokes lines are always stronger than the anti-Stokes lines. Thus, it is customary to measure Stokes lines in Raman spectroscopy. Figure 1.6 illustrates the Raman spectrum (below 500 cm1) of CCI4 excited by the blue line (488.0 nm) of an argon ion laser.

1.3. VIBRATION OF A DIATOMIC MOLECULE Through quantum mechanical considerations [2,7], the vibration of two nuclei in a diatomic molecule can be reduced to the motion of a single particle of mass m, whose displacement q from its equilibrium position is equal to the change of the internuclear distance. The mass m is called the reduced mass and is represented by 1 1 1 ¼ þ m m1 m2

ð1:15Þ

10

THEORY OF NORMAL VIBRATIONS

Fig. 1.7. Potential energy curves for a diatomic molecule: actual potential (solid line), parabolic potential (dashed line), and cubic potential (dotted line).

where m1 and m2 are the masses of the two nuclei. The kinetic energy is then 1 1 2 p T ¼ mq_ 2 ¼ 2 2m

ð1:16Þ

_ If a simple parabolic potential function such as where p is the conjugate momentum mq. that shown in Fig. 1.7 is assumed, the system represents a harmonic oscillator, and the potential energy is simply given by. 1 V ¼ Kq2 2

ð1:17Þ

Here K is the force constant for the vibration. Then the Schr€odinger wave equation becomes ! d2 c 8p2 m 1 2 þ 2 E Kq c ¼ 0 ð1:18Þ dq2 h 2 If this equation is solved with the condition that c must be single-valued, finite, and continuous, the eigenvalues are 1 1 Ev ¼ hn v þ ¼ hc~n v þ 2 2

ð1:19Þ

VIBRATION OF A DIATOMIC MOLECULE

with the frequency of vibration 1 n¼ 2p

sﬃﬃﬃﬃ sﬃﬃﬃﬃ K 1 K or ~n ¼ m 2pc m

11

ð1:20Þ

Here v is the vibrational quantum number, and it can have the values 0, 1, 2, 3, and so on. The corresponding eigenfunctions are pﬃﬃﬃ ða=pÞ1=4 aq2 =2 ﬃ e Hv ð aqÞ ð1:21Þ cv ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ v 2 v! pﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃ where a ¼ 2p mK =h ¼ 4p2 mn=h, and Hv ð aqÞ is a Hermite polynomial of the vth degree. Thus the eigenvalues and the corresponding eigenfunctions are 1 E0 ¼ hn; 2 3 E1 ¼ hn; 2

c0 ¼ ða=pÞ1=4 e aq =2 2

c1 ¼ ða=pÞ1=4 21=2 qe aq =2 2

ð1:22Þ

.. .

.. .

As Fig. 1.7 shows, actual potential curves can be approximated more exactly by adding a cubic term [2]: 1 V ¼ Kq2 Gq3 2

ðK GÞ

Then, the eigenvalues become 1 1 2 þ Ev ¼ hcwe v þ hcxe we v þ 2 2

ð1:23Þ

ð1:24Þ

where we is the wavenumber corrected for anharmonicity and xewe indicates the magnitude of anharmonicity. Table 2.1a (in Chapter 2) lists we and xewe for a number of diatomic molecules. Equation 1.24 shows that the energy levels of the anharmonic oscillator are not equidistant, and the separation decreases slowly as v increases. This anharmonicity is responsible for the appearance of overtones and combination vibrations, which are forbidden in the harmonic oscillator. The values of xe and xewe can be determined by observing a series of overtone bands in IR and Raman spectra. From Eq. 1.24, we obtain Ev E 0 ¼ vwe xe we ðv2 þ vÞ þ hc Then Fundamental: First overtone: Second overtone:

~n1 ¼ we 2xe we ~n2 ¼ 2we 6xe we ~n3 ¼ 3we 12xe we

12

THEORY OF NORMAL VIBRATIONS

For H35C1, these transitions are observed at 2885.9, 5668.1, and 8347.0 cm1, respectively, in IR spectrum [2]. Using these values, we find that we ¼ 2988:9 cm 1

and

xe we ¼ 52:05 cm 1

As will be shown in Sec. 1.23, a long series of overtone bands can be observed when Raman spectra of small molecules such as I2 and TiI4 are measured under rigorous resonance conditions. Anharmonicity constants can also be determined from the analysis of rotational fine structures of vibrational transitions [2]. Since the anharmonicity correction has not been made for most polyatomic molecules, in large part because of the complexity of the calculation, the frequencies given in Chapter 2 are not corrected for anharmonicity (except those given in Table 2.1a). According to Eq. 1.20, the wavenumber of the vibration in a diatomic molecule is given by sﬃﬃﬃﬃ 1 K ~n ¼ ð1:20Þ 2pc m A more exact expression is given by using the wavenumber corrected for anharmonicity: sﬃﬃﬃﬃ 1 K we ¼ ð1:25Þ 2pc m or K ¼ 4p2 c2 w2e m

ð1:26Þ

For HC1, we ¼ 2989 cm1 and m ¼ 0.9799 awu (where awu is the atomic weight unit). Thus, we obtain* K¼

4ð3:14Þ2 ð3 1010 Þ2 2 we m 6:025 1023

¼ ð5:8883 10 2 Þw2e m ¼ ð5:8883 10 2 Þð2989Þ2 ð0:9799Þ ¼ 5:16 105 ðdyn=cmÞ ¼ 5:16 ðmdyn=ÅÞ* Table 1.2 lists the observed frequencies ð~nÞ, wavenumbers corrected for anharmonicity (we), reduced masses (m), and force constants (K) for several series of diatomic *105 (dynes/cm) ¼ 105 (103 millidynes/108 Å) ¼ 1 (mdyn/Å) ¼ 102 N/m (SI unit).

13

VIBRATION OF A DIATOMIC MOLECULE

TABLE 1.2. Relationships between Vibrational Frequency, Reduced Mass, and Force Constant Molecule H2 HD D2 HF HCl HBr HI F2 Cl2 Br2 I2 N2 CO NO O2

Obs. ~n (cm1)

we (cm1)

4160 3632 2994 3962 2886 2558 2233 892 564 319 213 2331 2145 1877 1555

4395 3817 3118 4139 2989 2650 2310 — 565 323 215 2360 2170 1904 1580

m (awu)

˚) K (mdyn/A

0.5041 0.6719 1.0074 0.9573 0.9799 0.9956 1.002 9.5023 17.4814 39.958 63.466 7.004 6.8584 7.4688 8.000

5.73 5.77 5.77 9.65 5.16 4.12 3.12 4.45 3.19 2.46 1.76 22.9 19.0 15.8 11.8

molecules. In the first series, we decreases in the order H2 > HD > D2, because m increases in the same order, while K is almost constant (mass effect). In the second series, we decreses in the order HF > HC1 > HBr > HI, because K decreases in the same order, while m shows little change (force constant effect). In the third series, we decreases in the order F2 > Cl2 > Br2 > I2, because m increases and K decreases in the same order. In this case, both mass effect and force constant effect are operative. In the last series, we decreases in the order N2 > CO > NO > O2, mainly owing to the force constant effect. This may be attributed to the differences in bond order: NN > C O þ $ C ¼ O > N¼ O $ N ¼ O þ > O ¼ O $ O O þ More examples of similar series in diatomic molecules are found in Sec. 2.1. These simple rules, obtained for a diatomic molecule, are helpful in understanding the vibrational spectra of polyatomic molecules. Figure 1.8 indicates the relationship between the force constant and the dissociation energy for the three series listed in Table 1.2. In the series of hydrogen halides, the dissociation energy decreases almost linearly as the force constant decreases. Thus, the force constant may be used as a measure of the bond strength in this case. However, such a monotonic relationship does not hold for the other two series. This is not unexpected, because the force constant is a measure of the curvature of the potential well near the equilibrium position K¼

d2 V dq2

! ð1:27Þ q!0

14

THEORY OF NORMAL VIBRATIONS

Fig. 1.8. Relationships between force constants and dissociation energies in diatomic molecules.

whereas the dissociation energy De is given by the depth of the potential energy curve (Fig. 1.7). Thus, a large force constant means sharp curvature of the potential well near the bottom, but does not necessarily indicate a deep potential well. Usually, however, a larger force constant is an indication of a stronger bond if the nature of the bond is similar in a series. It is difficult to derive a general theoretical relationship between the force constant and the dissociation energy even for diatomic molecules. In the case of small molecules, attempts have been made to calculate the force constants by quantum mechanical methods. The principle of the method is to express the total electronic energy of a molecule as a function of nuclear displacements near the equilibrium position and to calculate its second derivatives, q2 V=qq2i , and so on for each displacement coordinate qi. In the past, ab initio calculations of force constants were made for small molecules such as HF, H2O, and NH3. The force constants thus obtained are in good agreement with those calculated from the analysis of vibrational spectra. More recent progress in computer technology has made it possible to extend this approach to more complex molecules (Sec. 1.24). There are several empirical relationships which relate the force constant to the bond distance. Badger’s rule [68] is given by K ¼ 1:86ðr dij Þ 3

ð1:28Þ

where r is the bond distance and dij has a fixed value for bonds between atoms of row i and j in the periodic table. Gordy’s rule [69] is expressed as !3=4 cA cB K ¼ aN þb ð1:29Þ d2

NORMAL COORDINATES AND NORMAL VIBRATIONS

15

where cA and cB are the electronegativities of atoms A and B constituting the bond, d is the bond distance, N is the bond order, and a and b are constants for certain broad classes of compounds. Herschbach and Laurie [70] modified Badger’s rule in the form r ¼ dij þ ðaij dij ÞK 1=3

ð1:30Þ

Here, aij and dij are constants for atoms of rows i and j in the periodic table. Another empirical relationship is found by plotting stretching frequencies against bond distances for a series of compounds having common bonds. Such plots are highly important in estimating the bond distance from the observed frequency. Typical examples are found in the OH O hydrogen-bonded compounds [71], carbon–oxygen and carbon–nitrogen bonded compounds [72], and molybdenum– oxygen [73] and vanadium–oxygen bonded compounds [74]. More examples are given in Sec. 2.2. 1.4. NORMAL COORDINATES AND NORMAL VIBRATIONS In diatomic molecules, the vibration of the nuclei occurs only along the line connecting two nuclei. In polyatomic molecules, however, the situation is much more complicated because all the nuclei perform their own harmonic oscillations. It can be shown, however, that any of these extremely complicated vibrations of the molecule may be represented as a superposition of a number of normal vibrations. Let the displacement of each nucleus be expressed in terms of rectangular coordinate systems with the origin of each system at the equilibrium position of each nucleus. Then the kinetic energy of an N-atom molecule would be expressed as 1X T¼ mN 2 N

"

dDxN dt

!2

dDyN þ dt

!2

dDzN þ dt

!2 # ð1:31Þ

If generalized coordinates such as q1 ¼

pﬃﬃﬃﬃﬃﬃ m1 Dx1 ;

q2 ¼

pﬃﬃﬃﬃﬃﬃ m1 Dy1 ;

q3 ¼

pﬃﬃﬃﬃﬃﬃ m1 Dz1

q4 ¼

pﬃﬃﬃﬃﬃﬃ m2 Dx2 ; . . .

ð1:32Þ

are used, the kinetic energy is simply written as

T¼

3N 1X q_ 2 2 i i

ð1:33Þ

16

THEORY OF NORMAL VIBRATIONS

The potential energy of the system is a complex function of all the coordinates involved. For small values of the displacements, it may be expanded in a Taylor’s series as ! ! 3N 3N X qV 1X q2 V Vðq1 ; q2 ; . . . ; q3N Þ ¼ V0 þ qi þ qi qj þ ð1:34Þ qqi 2 i;j qqi qqj i 0

0

where the derivatives are evaluated at qi ¼ 0, the equilibrium position. The constant term V0 can be taken as zero if the potential energy at qi ¼ 0 is taken as a standard. The (qV/qqi)0 terms also become zero, since V must be a minimum at qi ¼ 0. Thus, V may be represented by ! 3N 3N 1X q2 V 1X qi qj ¼ bij qi qj ð1:35Þ V¼ 2 i;j qqi qqj 2 i;j 0

neglecting higher-order terms. If the potential energy given by Eq. 1.35 did not include any cross-products such as qiqj, the problem could be solved directly by using Lagrange’s equation: ! d qT qV ¼ 0; þ dt qq_ i qqi

i ¼ 1; 2; . . . ; 3N

From Eqs. 1.33 and 1.35, Eq. 1.36 is written as X q€i þ bij qj ¼ 0; j ¼ 1; 2; . . . ; 3N

ð1:36Þ

ð1:37Þ

j

If bij ¼ 0 for i „ j, Eq. 1.37 becomes q€i þ bii qi ¼ 0

ð1:38Þ

and the solution is given by qi ¼

q0i sin

pﬃﬃﬃﬃﬃ bii t þ di

! ð1:39Þ

where q0i and di are the amplitude and the phase constant, respectively. Since, in general, this simplification is not applicable, the coordinates qi must be transformed into a set of new coordinates Qi through the relations X q1 ¼ B1i Qi q2 ¼

i X

B2i Qi

i

.. . qk ¼

X i

Bki Qi

ð1:40Þ

NORMAL COORDINATES AND NORMAL VIBRATIONS

17

The Qi are called normal coordinates for the system. By appropriate choice of the coefficients Bki, both the potential and the kinetic energies can be written as 1 X _2 Qi 2 i

ð1:41Þ

1X li Q2i 2 i

ð1:42Þ

T¼ V¼

without any cross-products. If Eqs. 1.41 and 1.42 are combined with Lagrange’s equation (Eq. 1.36), there results € þ li Qi ¼ 0 Q ð1:43Þ i The solution of this equation is given by Qi ¼

Q0i sin

pﬃﬃﬃﬃ li t þ di

! ð1:44Þ

and the frequency is ni ¼

1 pﬃﬃﬃﬃ li 2p

ð1:45Þ

Such a vibration is called a normal vibration. For the general N-atom molecule, it is obvious that the number of the normal vibrations is only 3N – 6, since six coordinates are required to describe the translational and rotational motion of the molecule as a whole. Linear molecules have 3N – 5 normal vibrations, as no rotational freedom exists around the molecular axis. Thus, the general form of the molecular vibration is a superposition of the 3N – 6 (or 3N – 5) normal vibrations given by Eq. 1.44. The physical meaning of the normal vibration may be demonstrated in the following way. As shown in Eq. 1.40, the original displacement coordinate is related to the normal coordinate by X qk ¼ Bki Qi ð1:40Þ i

Since all the normal vibrations are independent of each other, consideration may be limited to a special case in which only one normal vibration, subscripted by 1, is excited (i.e., Q01 „ 0; Q02 ¼ Q03 ¼ ¼ 0). Then, it follows from Eqs. 1.40 and 1.44 that pﬃﬃﬃﬃﬃ qk ¼ Bk1 Q1 ¼ Bk1 Q01 sinð l1 t þ d1 Þ pﬃﬃﬃﬃﬃ ¼ Ak1 sinð l1 t þ d1 Þ

ð1:46Þ

18

THEORY OF NORMAL VIBRATIONS

This relation holds for all k. Thus, it is seen that the excitation of one normal vibration of the system causes vibrations, given by Eq. 1.46, of all the nuclei in the system. In other words, in the normal vibration, all the nuclei move with the same frequency and in phase. This is true for any other normal vibration. Thus Eq. 1.46 may be written in the more general form pﬃﬃﬃ qk ¼ Ak sinð lt þ dÞ ð1:47Þ If Eq. 1.47 is combined with Eq. 1.37, there results lAk þ

X

bkj Aj ¼ 0

ð1:48Þ

j

This is a system of first-order simultaneous equations with respect to A. In order for all the As to be nonzero, we must solve b11 l b21 b31 . . .

b12 b22 l b32 .. .

b13 b23 b33 l .. .

... ... ... ¼ 0

ð1:49Þ

The order of this secular equation is equal to 3N. Suppose that one root, l1, is found for Eq. 1.49. If it is inserted in Eq. 1.48, Ak1, Ak2, . . . are obtained for all the nuclei. The same is true for the other roots of Eq. 1.49. Thus, the most general solution may be written as a superposition of all the normal vibrations: ! X pﬃﬃﬃﬃ 0 qk ¼ Bkl Ql sin ll t þ dl ð1:50Þ l

The general discussion developed above may be understood more easily if we apply it to a simple molecule such as CO2, which is constrained to move in only one direction. If the mass and the displacement of each atom are defined as follows:

the potential energy is given by V¼

i 1 h k ðDx1 Dx2 Þ2 þ ðDx2 Dx3 Þ2 2

ð1:51Þ

Considering that m1 ¼ m3, we find that the kinetic energy is written as T¼

1 1 m1 ðDx_ 21 þ Dx_ 23 Þ þ m2 Dx_ 22 2 2

ð1:52Þ

NORMAL COORDINATES AND NORMAL VIBRATIONS

19

Using the generalized coordinates defined by Eq. 1.32, we may rewrite these energies as "

q1 q2 pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ m1 m2

2V ¼ k

2T ¼

!2

q2 q3 þ pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ m2 m1

X

!2 #

q_ 2i

ð1:53Þ

ð1:54Þ

From comparison of Eq. 1.53 with Eq. 1.35, we obtain b11 ¼

k ; m1

b22 ¼

2k m2

k k b12 ¼ b21 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ; b23 ¼ b32 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2 m1 m2 b13 ¼ b31 ¼ 0;

b33 ¼

k m1

If these terms are inserted in Eq. 1.49, we obtain the following result: k m1 l k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2 0

k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2 2k l m2 k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m1 m2

k pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0 m1 m2 k l m1 0

By solving this secular equation, we obtain three roots: l1 ¼

k ; m1

l2 ¼ km;

l3 ¼ 0

where m¼

2m1 þ m2 m1 m2

Equation 1.48 gives the following three equations: lA1 þ b11 A1 þ b12 A2 þ b13 A3 ¼ 0 lA2 þ b21 A1 þ b22 A2 þ b23 A3 ¼ 0 lA3 þ b31 A1 þ b32 A2 þ b33 A3 ¼ 0

ð1:55Þ

20

THEORY OF NORMAL VIBRATIONS

Using Eq. 1.47, we rewrite these as ðb11 lÞq1 þ b12 q2 þ b13 q3 ¼ 0 b21 q1 þ ðb22 lÞq2 þ b23 q3 ¼ 0 b31 q1 þ b32 q2 þ ðb33 lÞq3 ¼ 0 If l1 ¼ k/m1 is inserted in the simultaneous equations above, we obtain q1 ¼ q3 ;

q2 ¼ 0

Similar calculations give q1 ¼ q3 ;

q1 ¼ q3 ;

sﬃﬃﬃﬃﬃﬃ m1 q2 ¼ 2 q1 m2

q2 ¼

sﬃﬃﬃﬃﬃﬃ m2 q1 m1

for

l2 ¼ km

for l3 ¼ 0

The relative displacements are depicted in the following figure:

It is easy to see that l3 corresponds to the translational mode (Dx1 ¼ Dx2 ¼ Dx3). The inclusion of l3 could be avoided if we consider the restriction that the center of gravity does not move; m1(Dx1 þ Dx3) þ m2Dx2 ¼ 0. The relationships between the generalized coordinates and the normal coordinates are given by Eq. 1.40. In the present case, we have q1 ¼ B11 Q1 þ B12 Q2 þ B13 Q3 q2 ¼ B21 Q1 þ B22 Q2 þ B23 Q3 q3 ¼ B31 Q1 þ B32 Q2 þ B33 Q3 In the normal vibration whose normal coordinate is Q1, B11: B21: B31 gives the ratio of the displacements. From the previous calculation, it is ﬃ obvious that B11: B21: B31 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1:0:1. Similarly, B12 : B22 : B32 ¼ 1 : 2 m1 =m2 : 1 gives the ratio of the displacements in the normal vibration whose normal coordinate is Q2. Thus the mode

SYMMETRY ELEMENTS AND POINT GROUPS

21

of a normal vibration can be drawn if the normal coordinate is translated into a set of rectangular coordinates, as is shown above. So far, we have discussed only the vibrations whose displacements occur along the molecular axis. There are, however, two other normal vibrations in which the displacements occur in the direction perpendicular to the molecular axis. They are not treated here, since the calculation is not simple. It is clear that the method described above will become more complicated as a molecule becomes larger. In this respect, the GF matrix method described in Sec. 1.12 is important in the vibrational analysis of complex molecules. By using the normal coordinates, the Schr€ odinger wave equation for the system can be written as X q2 c i

n qQ2i

! 8p2 1X þ 2 E li Q2i cn ¼ 0 2 i h

ð1:56Þ

Since the normal coordinates are independent of each other, it is possible to write cn ¼ c1 ðQ1 Þc2 ðQ2 Þ

ð1:57Þ

and solve the simpler one-dimensional problem. If Eq. 1.57 is substituted in Eq. 1.56, there results d2 ci 8p2 þ 2 h dQ2i

! 1 2 Ei li Qi ci ¼ 0 2

ð1:58Þ

where E ¼ E1 þ E 2 þ with 0

1 1 Ei ¼ hni @vi þ A 2 vi ¼

1 pﬃﬃﬃﬃ li 2p

ð1:59Þ

1.5. SYMMETRY ELEMENTS AND POINT GROUPS [9–14] As noted before, polyatomic molecules have 3N – 6 or, if linear, 3N – 5 normal vibrations. For any given molecule, however, only vibrations that are permitted by

22

THEORY OF NORMAL VIBRATIONS

the selection rule for that molecule appear in the infrared and Raman spectra. Since the selection rule is determined by the symmetry of the molecule, this must first be studied. The spatial geometric arrangement of the nuclei constituting the molecule determines its symmetry. If a coordinate transformation (a reflection or a rotation or a combination of both) produces a configuration of the nuclei indistinguishable from the original one, this transformation is called a symmetry operation, and the molecule is said to have a corresponding symmetry element. Molecules may have the following symmetry elements. 1.5.1. Identity I This symmetry element is possessed by every molecule no matter how unsymmetric it is; the corresponding operation is to leave the molecule unchanged. The inclusion of this element is necessitated by mathematical reasons that will be discussed in Sec. 1.7. 1.5.2. A Plane of Symmetry If reflection of a molecule with respect to some plane produces a configuration indistinguishable from the original one, the plane is called a plane of symmetry. 1.5.3. A Center of Symmetry i If reflection at the center, that is, inversion, produces a configuration indistinguishable from the original one, the center is called a center of symmetry. This operation changes the signs of all the coordinates involved: xi ! xi ; yi ! yi ; zi ! zi . 1.5.4. A p-Fold Axis of Symmetry Cp* If rotation through an angle 360 /p about an axis produces a configuration indistinguishable from the original one, the axis is called a p-fold axis of symmetry Cp. For example, a twofold axis C2 implies that a rotation of 180 about the axis reproduces the original configuration. A molecule may have a two-, three-, four-, five-, or sixfold, or higher axis. A linear molecule has an infinite-fold (denoted by ¥-fold) axis of symmetry C¥ since a rotation of 360 /¥, that is, an infinitely small angle, transforms the molecule into one indistinguishable from the original. 1.5.5. A p-Fold Rotation–Reflection Axis Sp* If rotation by 360 /p about the axis, followed by reflection at a plane perpendicular to the axis, produces a configuration indistinguishable from the original one, the axis is called a p-fold rotation–reflection axis. A molecule may have a two-, three-, four-,

* The notation C n (or Sn ) is used to indicate that the Cp (or Sp) operation is carried out successively n times. p

p

SYMMETRY ELEMENTS AND POINT GROUPS

23

five-, or sixfold, or higher, rotation–reflection axis. A symmetrical linear molecule has an S¥ axis. It is easily seen that the presence of Sp always means the presence of Cp as well as s when p is odd. 1.5.6. Point Group A molecule may have more than one of these symmetry elements. Combination of more and more of these elements produces systems of higher and higher symmetry. Not all combinations of symmetry elements, however, are possible. For example, it is highly improbable that a molecule will have a C3 and C4 axis in the same direction because this requires the existence of a 12-fold axis in the molecule. It should also be noted that the presence of some symmetry elements often implies the presence of other elements. For example, if a molecule has two s planes at right angles to each other, the line of intersection of these two planes must be a C2 axis. A possible combination of symmetry operations whose axes intersect at a point is called a point group.* Theoretically, an infinite number of point groups exist, since there is no restriction on the order (p) of rotation axes that may exist in an isolated molecule. Practically, however, there are few molecules and ions that possess a rotation axis higher than C6. Thus most of the compounds discussed in this book belong to the following point groups: (1) Cp. Molecules having only a Cp and no other elements of symmetry: C1, C2, C3, and so on. (2) Cph. Molecules having a Cp and a sh perpendicular to it: C1hCs, C2h, C3h, and so on. (3) Cpv. Molecules having a Cp and psv through it: C1v Cs ; C2v ; C3v ; C4v ; . . . ; C¥v : (4) Dp. Molecules having a Cp and pC2 perpendicular to the Cp and at equal angles to one another: D1C2, D2V, D3, D4, and so on. (5) Dph. Molecules having a Cp, psn through it at angles of 360 /2p to one another, and a sh perpendicular to the Cp: D1h C2v ; D2h Vh ; D3h ; D4h ; D5h ; D6h ; . . . ; D¥h : (6) Dpd. Molecules having a Cp, pC2 perpendicular to it, and psd which go through the Cp and bisect the angles between two successive C2 axes: D2dVd, D3d, D4d, D5d, and so on. (7) Sp. Molecules having only a Sp (p even). For p odd, Sp is equivalent to Cp sh, for which other notations such as C3h are used: S2Ci, S4, S6, and so on. (8) Td. Molecules having three mutually perpendicular C2 axes, four C3 axes, and a sd through each pair of C3 axes: regular tetrahedral molecules. *In this respect, point groups differ from space groups, which involve translations and rotations about nonintersecting axes (see Sec. 1.27).

24

THEORY OF NORMAL VIBRATIONS

(9) Oh. Molecules having three mutually perpendicular C4 axes, four C3 axes, and a center of symmetry, i: regular octahedral and cubic molecules. (10) Ih. Molecules having 6 C5 axes, 10 C3 axes, 15 C2 axes, 15 s planes, and a center of symmetry. In total, such molecules possess 120 symmetry elements. One example is an icosahedron having 20 equilateral triangular faces, which is found in the B12 skeleton of the B12 H221 ion (Sec. 2.13). Another example is a regular dodecahedron having 12 regular pentagonal faces. Buckminsterfullerene, C60 (Sec. 2.14), also belongs to the Ih point group. It is a truncated icosahedron with 20 hexagonal and 12 pentagonal faces. Figure 1.9 illustrates the symmetry elements present in the point group, D4h. Complete listings of the symmetry elements for common point groups are found in the character tables included in Appendix I. Figure 1.10 illustrates examples of molecules belonging to some of these point groups. From the symmetry perspective, molecules

Fig. 1.9. Symmetry elements in D4h point group.

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

25

Fig. 1.10. Examples of molecules belonging to some point groups.

belonging to the C1, C2, C3, D2V, and D3 groups possess only Cp axes, and are thus optically active. 1.6. SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES Figures 1.11 and 1.12 illustrate the normal modes of vibration of CO2 and H2O molecules, respectively. In each normal vibration, the individual nuclei carry out a simple harmonic motion in the direction indicated by the arrow, and all the nuclei have

26

THEORY OF NORMAL VIBRATIONS

Fig. 1.10. (Continued)

the same frequency of oscillation (i.e., the frequency of the normal vibration) and are moving in the same phase. Furthermore, the relative lengths of the arrows indicate the relative velocities and the amplitudes for each nucleus.* The n2 vibrations in CO2 are worth comment, since they differ from the others in that two vibrations (n2a and n2b) have exactly the same frequency. Apparently, there are an infinite number of normal vibrations of this type, which differ only in their directions perpendicular to the molecular axis. Any of them, however, can be resolved into two vibrations such as n2a and n2b, which are perpendicular to each other. In this respect, the n2 vibrations in CO2 are called doubly degenerate vibrations. Doubly degenerate vibrations occur only when a molecule has an axis higher than twofold. Triply degenerate vibrations also occur in molecules having more than one C3 axis. To determine the symmetry of a normal vibration, it is necessary to begin by considering the kinetic and potential energies of the system. These were discussed in Sec. 1.4: *In this respect, all the normal modes of vibration shown in this book are only approximate.

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

27

Fig. 1.10. (Continued)

1X 2 Q_ i 2 i

ð1:60Þ

1X li Q2i 2 i

ð1:61Þ

T¼ V¼

Consider a case in which a molecule performs only one normal vibration, Qi. Then, 2 T ¼ 12 Q_ i and V ¼ 12 li Q2i . These energies must be invariant when a symmetry operation, R, changes Qi to RQi. Thus, we obtain 1 2 1 T ¼ Q_ i ¼ ðRQ_ i Þ2 2 2 1 1 V ¼ li Q2i ¼ li ðRQi Þ2 2 2

28

THEORY OF NORMAL VIBRATIONS

Fig. 1.10. (Continued)

Fig. 1.11. Normal modes of vibration in CO2 (þ and – denote vibrations going upward and downward, respectively, in the direction perpendicular to the paper plane).

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

29

For these relations to hold, it is necessary that ðRQi Þ2 ¼ Q2i

or

RQi ¼ Qi

ð1:62Þ

Thus, the normal coordinate must change either into itself or into its negative. If Qi ¼ RQi, the vibration is said to be symmetric. If Qi ¼ RQi, it is said to be antisymmetric. If the vibration is doubly degenerate, we have T¼

1 _2 1 2 Qia þ Q_ ib 2 2

1 1 V ¼ li ðQia Þ2 þ li ðQib Þ2 2 2 In this case, a relation such as ðRQia Þ2 þ ðRQib Þ2 ¼ Q2ia þ Q2ib

ð1:63Þ

must hold. As will be shown later, such a relationship is expressed more conveniently by using a matrix form:* #" " # " # Qia Qia A B R ¼ C D Qib Qib where the values of A, B, C, and D depend on the symmetry operation, R. In any case, the normal vibration must be either symmetric or antisymmetric or degenerate for each symmetry operation. The symmetry properties of the normal vibrations of the H2O molecule shown in Fig. 1.12 are classified as indicated in Table 1.3. Here, þ1 and 1 denote symmetric and antisymmetric, respectively. In the n1 and n2 vibrations, all the symmetry properties are preserved during the vibration. Therefore they are symmetric vibrations and are called, in particular, totally symmetric vibrations. In the n3 vibration, however, symmetry elements such as C2 and sv (xz) are lost. Thus, it is called a nonsymmetric vibration. If a molecule has a number of symmetry elements, the normal vibrations are classified according to the number and the kind of symmetry elements preserved during the vibration. To determine the activity of the vibrations in the infrared and Raman spectra, the selection rule must be applied to each normal vibration. As will be shown in Sec. 1.10, rigorous selection rules can be derived quantum mechanically, and applied to individual molecules using group theory. For small molecules, however, the IR and Raman activities may be determined by simple inspection of their normal modes. First, we consider the general rule which states that the vibration is IR-active if the dipole moment is changed during the vibration. It is obvious that the vibration of a homopolar diatomic molecule is *For matrix algebra, see Appendix II.

30

THEORY OF NORMAL VIBRATIONS

Fig. 1.12. Normal modes of vibration in H2O.

TABLE 1.3. C2v Q1, Q2 Q3 a

I þ1 þ1

C2(z)

sv(xz)a

sv(xz)a

þ1 1

þ1 1

þ1 þ1

sv ¼ vertical plane of symmetry.

not IR-active, whereas that of a heteropolar diatomic molecule is IR-active. In the case of CO2, shown in Fig. 1.11, it is readily seen that the n1 is not IR-active, whereas the n2 and n3 are IR-active. Figure 1.13 illustrates the changes in dipole moment during the three normal vibrations of H2O. It is readily seen that all the vibrations are IR-active because the magnitude or direction of the dipole moment is changed as indicated. To discuss Raman activity, we must consider the nature of polarizability introduced in Sec. 1.2. When a molecule interacts with the electric field of a laser beam, its electron cloud is distorted because the positively charged nuclei are attracted toward the negative pole, and the electrons toward the positive pole, as shown in Fig. 1.14. The charge separation produces an induced dipole moment (P) given by* P ¼ aE*

ð1:64Þ

*A more complete form of this equation is P ¼ aE þ 1 bE2 . . .. Here, b a and b is called the 2 hyperpolarizability. The second term becomes significant only when E is large (109 V cm1). In this case, we observe novel spectroscopic phenomena such as the hyper Raman effect, stimulated Raman effect, inverse Raman effect, and coherent anti-Stokes Raman scattering (CARS). For a discussion of nonlinear Raman spectroscopy, see Refs. 25 and 26.

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

31

Fig. 1.13. Changes in dipole moment for H2O during each normal vibration.

where E is the strength of the electric field and a is the polarizability. By resolving P, a, and E in the x, y, and z directions, we can write. Px ¼ ax Ex ;

Py ¼ a y Ey ;

P z ¼ a z Ez

ð1:65Þ

However, the actual relationship is more complicated since the direction of polarization may not coincide with the direction of the applied field. This is so because the direction of chemical bonds in the molecule also affects the direction of polarization. Thus, instead of Eq. 1.65, we have the relationship Px ¼ axx Ex þ axy Ey þ axz Ez Py ¼ ayx Ex þ ayy Ey þ ayz Ez

ð1:66Þ

Pz ¼ azx Ex þ azy Ey þ azz Ez In matrix form, Eq. 1.66 is written as "

Px Py Pz

# ¼

"a

xx

ayx azx

axy ayy azy

axz #" Ex # ayz Ey azz Ez

Fig. 1.14. Polarization of a diatomic molecule in an electric field.

ð1:67Þ

32

THEORY OF NORMAL VIBRATIONS

and the first matrix on the right-hand side is called the polarizability tensor. In normal Raman scattering, the tensor is symmetric; axy ¼ ayx, ayz ¼ azy, and axz ¼ azx. This is not so, however, in the case of resonance Raman scattering (Sec. 1.22). According to quantum mechanics, the vibration is Raman-active if one of these six components of the polarizability changes during the vibration. Thus, it is obvious that the vibration of a homopolar diatomic molecule is Raman-active but not IR-active, whereas the vibration of a heteropolar diatomic molecule is both IR- and Ramanactive. Changes in the polarizability tensor can be visualized if we draw a polarizability pﬃﬃﬃ ellipsoid by plotting 1= a in every direction from the origin. This gives a threedimensional surface such as shown below:

If we rotate such an ellipsoid so that its principal axes coincide with the molecular axes (X, Y, and Z), Eq. 1.67 is simplified to "

PX PY PZ

#

" ¼

aXX 0 0

0 aYY 0

0 0 aZZ

#"

EX EY EZ

# ð1:68Þ

Such axes are called principal axes of polarizability. In terms of the polarizability ellipsoid, the vibration is Raman-active if the polarizability ellipsoid changes in size, shape, or orientation during the vibration. As an example, Fig. 1.15 illustrates the polarizability ellipsoids for the three normal vibrations of H2O at the equilibrium (q ¼ 0) and two extreme configurations (q ¼ q). It is readily seen that both n1 and n2 are Raman-active because the size and the shape of the ellipsoid (axx, ayy, and azz) change during these vibrations. The n3 is also Raman-active because the orientation of the ellipsoid (ayz) changes during the vibration. Thus, all three normal vibrations of H2O are Raman-active. Figure 1.16 illustrates the changes in polarizability ellipsoids during the normal vibrations of CO2. It is readily seen that n1 is Raman-active because the size of the ellipsoid

SYMMETRY OF NORMAL VIBRATIONS AND SELECTION RULES

33

Fig. 1.15. Changes in polarizability ellipsoid during normal vibrations of H2O.

changes during the vibration (axx, ayy, and azz change proportionately). Although the size and/or shape of the ellipsoid change during the n2 and n3 vibrations, they are identical in two extreme positions, as seen in Fig. 1.16. If we consider a limiting case where the nuclei undergo very small displacements, there is effectively no change in the polarizability. This is illustrated in Fig. 1.17. Thus, these two vibrations are not Raman-active. It should be noted that in CO2 the vibration symmetric with respect to the center of symmetry (n1) is Raman-active and not IR-active, whereas the vibrations antisymmetric with respect to the center of symmetry (n2 and n3) are IR-active but not Ramanactive. In a polyatomic molecule having a center of symmetry, the vibrations symmetric with respect to the center of symmetry (g vibrations*) are Raman-active

Fig. 1.16. Changes in polarizability ellipsoid during normal vibrations of CO2.

*The symbols g and u stand for gerade and ungerade (German), respectively.

34

THEORY OF NORMAL VIBRATIONS

Fig. 1.17. Changes in polarizability with respect to displacement coordinate during the n1 and n2,3 vibrations in CO2.

and not IR-active, but the vibrations antisymmetric with respect to the center of symmetry (u vibrations*) are IR-active and not Raman-active. This rule is called the “mutual exclusion rule.” It should be noted, however, that in polyatomic molecules having several symmetry elements in addition to the center of symmetry, the vibrations that should be active according to this rule may not necessarily be active, because of the presence of other symmetry elements. An example is seen in a square–planar XY4-type molecule of D4h symmetry, where the A2g vibrations are not Raman-active and the A1u, B1u, and B2u vibrations are not IR-active (see Sec. 2.6). 1.7. INTRODUCTION TO GROUP THEORY [9–14] In Sec. 1.5, the symmetry and the point group allocation of a given molecule were discussed. To understand the symmetry and selection rules of normal vibrations in polyatomic molecules, however, a knowledge of group theory is required. The minimum amount of group theory needed for this purpose is given here. Consider a pyramidal XY3 molecule (Fig. 1.18) for which the symmetry operations I, C3þ , C3 , s1, s2, and s3 are applicable. Here, C3þ and C3 denote rotation through 120 in the clockwise and counterclockwise directions, respectively; and s1, s2, and s3 indicate the symmetry planes that pass through X and Y1, X and Y2, and X and Y3, respectively. For simplicity, let these symmetry operations be denoted by I, A, B, C, D, and E, respectively. Other symmetry operations are possible, but each is equivalent to some one of the operations mentioned. For instance, a clockwise rotation through 240 is identical with operation B. It may also be shown that two successive applications of any one of these operations is equivalent to some single operation of the group mentioned. Let operation C be applied to the original figure. This interchanges Y2 and Y3. If operation A is applied to the resulting figure, the net result is the same as

*The symbols g and u stand for gerade and ungerade (German), respectively.

INTRODUCTION TO GROUP THEORY

35

Fig. 1.18. Pyramidal XY3 molecule.

application of the single operation D to the original figure. This is written as CA ¼ D. If all the possible multiplicative combinations are made, then a tabular display such as Table 1.4, in which the operation applied first is written across the top, is obtained. This is called the multiplication table of the group. It is seen that a group consisting of the mathematical elements (symmetry operations) I, A, B, C, D, and E satisfies the following conditions: (1) The product of any two elements in the set is another element in the set. (2) The set contains the identity operation that satisfies the relation IP ¼ PI ¼ P, where P is any element in the set. (3) The associative law holds for all the elements in the set, that is, (CB)A ¼ C (BA), for example. (4) Every element in the set has its reciprocal, X, which satisfies the relation XP ¼ PX ¼ I, where P is any element in the set. This reciprocal is usually denoted by P1. These are necessary and sufficient conditions for a set of elements to form a group. It is evident that operations I, A, B, C, D, and E form a group in this sense. It should be noted

TABLE 1.4.

I A B C D E

I

A

B

C

D

E

I A B C D E

A B I E C D

B I A D E C

C D E I A B

D E C B I A

E C D A B I

36

THEORY OF NORMAL VIBRATIONS

that the commutative law of multiplication does not necessarily hold. For example, Table 1.4 shows that CD „ DC. The six elements can be classified into three types of operation: the identity operation I, the rotations C3þ and C3 , and the reflections s1, s2, and s3. Each of these sets of operations is said to form a class. More precisely, two operations, P and Q, which satisfy the relation X1PX ¼ P or Q, where X is any operation of the group and X1 is its reciprocal, are said to belong to the same class. It can easily be shown that C3þ and C3 , for example, satisfy the relation. Thus the six elements of the point group C3v are usually abbreviated as I, 2C3, and 3sv. The relations between the elements of the group are shown in the multiplication table (Table 1.4). Such a tabulation of a group is, however, awkward to handle. The essential features of the table may be abstracted by replacing the elements by some analytical function that reproduces the multiplication table. Such an analytical expression may be composed of a simple integer, an exponential function, or a matrix. Any set of such expressions that satisfies the relations given by the multiplication table is called a representation of the group and is designated by G. The representations of the point group C3v discussed above are indicated in Table 1.5. It can easily be proved that each representation in the table satisfies the multiplication table. In addition to the three representations in Table 1.5, it is possible to write an infinite number of other representations of the group. If a set of six matrices of the type S1R(K)S is chosen, where R(K) is a representation of the element K given in Table 1.5, SðjSj „ 0Þ is any matrix of the same order as R, and S1 is the reciprocal of S, this set also satisfies the relations given by the multiplication table. The reason is obvious from the relation S 1 RðKÞSS 1 RðLÞS ¼ S 1 RðKÞRðLÞS ¼ S 1 RðKLÞS Such a transformation is called a similarity transformation. Thus, it is possible to make an infinite number of representations by means of similarity transformations. On the other hand, this statement suggests that a given representation may possibly be broken into simpler ones. If each representation of the symmetry element K is

TABLE 1.5. C3v

I

A

B

C

D

A1(G1) A2(G2)

1 1

1 1

1 1

1 1

1 1

E(G3)

pﬃﬃﬃ 1 0 1 3 ! B 2 2 C B C 1 0 B C B pﬃﬃﬃ C B C 0 1 1 3 @ A 2 2

pﬃﬃﬃ 1 0 1 3 B 2 2 C B C B C Bpﬃﬃﬃ C B 3 C 1 @ A 2 2

pﬃﬃﬃ 1 3 ! B2 2 B 1 0 B B pﬃﬃﬃ 0 1 B @ 3 1 2 2 0

E 1 1 pﬃﬃﬃ 1 1 0 1 3 C B2 2 C C B C C B C C B pﬃﬃﬃ C C B 3 C 1 A @ A 2 2

INTRODUCTION TO GROUP THEORY

37

transformed into the form

ð1:69Þ

by a similarity transformation, Q1(K), Q2(K), . . . are simpler representations. In such a case, R(K) is called reducible. If a representation cannot be simplified any further, it is said to be irreducible. The representations G1, G2, and G3 in Table 1.5 are all irreducible representations. It can be shown generally that the number of irreducible representations is equal to the number of classes. Thus, only three irreducible representations exist for the point group C3v. These representations are entirely independent of each other. Furthermore, the sum of the squares of the dimensions (l) of the irreducible representations of a group is always equal to the total number of the symmetry elements, namely, the order of the group (h). Thus X

l2i ¼ l21 þ l22 þ ¼ h

ð1:70Þ

In the point group C3v, it is seen that 12 þ 12 þ 22 ¼ 6 A point group is classified into species according to its irreducible representations. In the point group C3v, the species having the irreducible representations G1, G2, and G3 are called the A1, A2, and E species, respectively.* The sum of the diagonal elements of a matrix is called the character of the matrix and is denoted by c. It is to be noted in Table 1.5 that the character of each of the elements belonging to the same class is the same. Thus, using the character, Table 1.5 can be simplified to Table 1.6. Such a table is called the character table of the point group C3v. That the character of a matrix is not changed by a similarity transformation can be proved as follows. If a similarity transformation is expressed by T ¼ S1RS, then cT ¼

X

ðS 1 RSÞii ¼

i

¼

X j;k

dkj Rjk ¼

X

X

ðS 1 Þij Rjk Ski ¼

i;j;k

X

Ski ðS 1 Þij Rjk

j;k;i

Rkk ¼ cR

k

*For labeling of the irreducible representations (species), see Appendix I.

38

THEORY OF NORMAL VIBRATIONS

TABLE 1.6. Character Table of the Point Group C3v C3v

I

A1 (c1) A2 (c2) E (c3)

1 1 2

2C3(z) 1 1 1

3sv 1 1 0

where dkj is Kr€ onecker’s delta (0 for k „ j and 1 for k ¼ j). Thus, any reducible representation can be reduced to its irreducible representations by a similarity transformation that leaves the character unchanged. Therefore, the character of the reducible representation, c(K), is written as cðKÞ ¼

X

am cm ðKÞ

ð1:71Þ

m

where cm(K) is the character of Qm(K), and am is a positive integer that indicates the number of times that Qm(K) appears in the matrix of Eq. 1.69. Hereafter the character will be used rather than the corresponding representation because a 1 : 1 correspondence exists between these two, and the former is sufficient for vibrational problems. It is important to note that the following relation holds in Table 1.6: X

ci ðKÞcj ðKÞ ¼ hdij

ð1:72Þ

K

If Eq. 1.71 is multiplied by ci(K) on both sides, and the summation is taken over all the symmetry operations, then X

cðKÞci ðKÞ ¼

K

¼

XX K

m

m

K

XX

am cm ðKÞci ðKÞ am cm ðKÞci ðKÞ

For a fixed m, we have X

am cm ðKÞci ðKÞ ¼ am

K

X

cm ðKÞci ðKÞ ¼ am hdim

K

If we consider the sum of such a term over m, only the sum in which m ¼ i remains. Thus, we obtain X K

cðKÞcm ðKÞ ¼ ham

39

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

or am ¼

1X cðKÞcm ðKÞ h K

ð1:73Þ

This formula is written more conveniently as am ¼

1X ncðKÞcm ðKÞ h

ð1:74Þ

where n is the number of symmetry elements in any one class, and the summation is made over the different classes. As Sec. 1.8 will show, this formula is very useful in determining the number of normal vibrations belonging to each species.*

1.8. THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES As shown in Sec. 1.6, the 3N 6 (or 3N 5) normal vibrations of an N-atom molecule can be classified into various species according to their symmetry properties. The number of normal vibrations in each species can be calculated by using the general equations given in Appendix III. These equations were derived from consideration of the vibrational degrees of freedom contributed by each set of identical nuclei for each symmetry species [1]. As an example, let us consider the NH3 molecule belonging to the C3v point group. The general equations are as follows: A1 species: A2 species:

3m þ 2mv þ m0 1 3m þ mv 1

E species: 6m þ 3mv þ m0 2 Nðtotal number of atomsÞ ¼ 6m þ 3mv þ m0 From the definitions given in the footnotes of Appendix III, it is obvious that m ¼ 0, m0 ¼ 1, and mv ¼ 1 in this case. To check these numbers, we calculate the total number of atoms from the equation for N given above. Since the result is 4, these assigned numbers are correct. Then, the number of normal vibrations in each species can be calculated by inserting these numbers in the general equations given above: 2, 0, and 2, respectively, for the A1, A2, and E species. Since the E species is doubly degenerate, the total number of vibrations is counted as 6, which is expected from the 3N 6 rule. A more general method of finding the number of normal vibrations in each species can be developed by using group theory. The principle of the method is that all the *Since this equation is not applicable to the infinite point groups (C¥v and D¥h), several alternative approaches have been proposed (see Refs. 75 and 76).

40

THEORY OF NORMAL VIBRATIONS

representations are irreducible if normal coordinates are used as the basis for the representations. For example, the representations for the symmetry operations based on three normal coordinates, Q1, Q2, and Q3, which correspond to the n1, n2, and n3 vibrations in the H2O molecule of Fig. 1.12, are as follows: " Q1 # I Q2 Q3 "

Q1 sv ðxzÞ Q2 Q3

"1 ¼

0 0

#

" ¼

0 0 #" Q1 # 1 0 0 1 1 0 0 1 0 0

" Q1 #

Q2 ; Q3 0 0 1

#"

C2 ðzÞ Q2 Q3

# Q1 Q2 ; Q3

0 #" Q1 #

"1 0 ¼

"

Q1 sv ðyzÞ Q2 Q3

0 1 0 0 #

" ¼

0 1 1 0 0

0 0 1 0 0 1

Q2 Q3 #"

Q1 Q2 Q3

#

Let a representation be written with the 3N rectangular coordinates of an N-atom molecule as its basis. If it is decomposed into its irreducible components, the basis for these irreducible representations must be the normal coordinates, and the number of appearances of the same irreducible representation must be equal to the number of normal vibrations belonging to the species represented by this irreducible representation. As stated previously, however, the 3N rectangular coordinates involve six (or five) coordinates, which correspond to the translational and rotational motions of the molecule as a whole. Therefore, the representations that have such coordinates as their basis must be subtracted from the result obtained above. Use of the character of the representation, rather than the representation itself, yields the same result. For example, consider a pyramidal XY3 molecule that has six normal vibrations. At first, the representations for the various symmetry operations must be written with the 12 rectangular coordinates in Fig. 1.19 as their basis. Consider pure rotation Cpþ . If the

Fig. 1.19. Rectangular coordinates in a pyramidal XY3 molecule (Z axis is perpendicular to the paper plane).

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

41

clockwise rotation of the point (x,y,z) around the z axis by the angle y brings it to the point denoted by the coordinates (x0 ,y0,z0 ), the relations between these two sets of coordinates are given by. x0 ¼ x cosy þ y siny y0 ¼ x siny þ y cosy z0 ¼ z

ð1:75Þ

By using matrix notation,* this can be written as "

x0 y0 z0

# ¼

C0þ

" # " cosy siny x y ¼ siny cosy 0 0 z

0 0 1

#" # x y z

ð1:76Þ

Then the character of the matrix is given by cðCyþ Þ ¼ 1 þ 2cosy

ð1:77Þ

The same result is obtained for cðCy Þ. If this symmetry operation is applied to all the coordinates of the XY3 molecule, the result is

ð1:78Þ

where A denotes the small square matrix given by Eq. 1.76. Thus, the character of this representation is simply given by Eq. 1.77. It should be noted in Eq. 1.78 that only the small matrix A, related to the nuclei unchanged by the symmetry operation, appears as a diagonal element. Thus, a more general form of the character of the representation for rotation around the axis by y is

*For matrix algebra, see Appendix II.

42

THEORY OF NORMAL VIBRATIONS

cðRÞ ¼ NR ð1 þ 2 cosyÞ

ð1:79Þ

where NR is the number of nuclei unchanged by the rotation. In the present case, NR ¼ 1 and y ¼ 120 . Therefore, we obtain cðC3 Þ ¼ 0

ð1:80Þ

Identity (I) can be regarded as a special case of Eq. 1.79 in which NR ¼ 4 and y ¼ 0 . The character of the representation is cðIÞ ¼ 12

ð1:81Þ

Pure rotation and identity are called proper rotation. It is evident from Fig. 1.19 that a symmetry plane such as s1 changes the coordinates from (xi, yi, zi) to (xi, yi, zi). The corresponding representation is therefore written as " # " 1 x 0 s1 y ¼ 0 z

0 1 0

0 0 1

#" # x y z

ð1:82Þ

The result of such an operation on all the coordinates is

ð1:83Þ

where B denotes the small square matrix of Eq. 1.82. Thus, the character of this representation is calculated as 2 1 ¼ 2. It is noted again that the matrix on the diagonal is nonzero only for the nuclei unchanged by the operation. More generally, a reflection at a plane (s) is regarded as s ¼ i C2. Thus, the general form of Eq. 1.82 may be written as

43

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

"

1 0 0

0 1 0

0 0 1

#"

cosy siny 0

siny cosy 0

0 0 1

#

" ¼

cosy siny 0

siny cosy 0

0 0 1

#

Then cðsÞ ¼ ð1 þ 2 cosyÞ As a result, the character of the large matrix shown in Eq. 1.83 is given by cðRÞ ¼ NR ð1 þ 2 cosyÞ

ð1:84Þ

In the present case, NR ¼ 2 and y ¼ 180 . This gives cðsv Þ ¼ 2

ð1:85Þ

Symmetry operations such as i and Sp are regarded as i ¼ i I;

y ¼ 0

S3 ¼ i C6 ; S4 ¼ i C4;

y ¼ 60 y ¼ 90

S6 ¼ i C3 ;

y ¼ 120

Therefore, the characters of these symmetry operations can be calculated by Eq. 1.84 with the values of y defined above. Operations such as s, i, and Sp are called improper rotations. Thus, the character of the representation based on 12 rectangular coordinates is as follows: I 2C3 12 0

3sv 2

ð1:86Þ

To determine the number of normal vibrations belonging to each species, the c(R) thus obtained must be resolved into the ci(R) of the irreducible representations of each species in Table 1.6. First, however, the characters corresponding to the translational and rotational motions of the molecule must be subtracted from the result shown in Eq. 1.86. The characters for the translational motion of the molecule in the x, y, and z directions (denoted by Tx, Ty, and Tz) are the same as those obtained in Eqs. 1.79 and 1.84. They are as follows: ct ðRÞ ¼ ð1 þ 2 cosyÞ

ð1:87Þ

where the þ and signs are for proper and improper rotations, respectively. The characters for the rotations around the x, y, and z axes (denoted by Rx, Ry, and Rz) are

44

THEORY OF NORMAL VIBRATIONS

given by cr ðRÞ ¼ þ ð1 þ 2 cosyÞ

ð1:88Þ

for both proper and improper rotations. This is due to the fact that a rotation of the vectors in the plane perpendicular to the x, y, and z axes can be regarded as a rotation of the components of angular momentum, Mx,My, and Mz, about the given axes. If px, py, and pz are the components of linear momentum in the x, y, and z directions, the following relations hold: Mx ¼ ypz zpy My ¼ zpx xpz Mz ¼ xpy ypx Since (x,y,z) and (px,py ,pz) transform as shown in Eq. 1.76, it follows that " Cy

Mx My Mz

#

" ¼

cosy siny 0

siny cosy 0

0 0 1

#"

Mx My Mz

#

Then, a similar relation holds for Rx, Ry , and Rz: " Cy

Rx Ry Rz

#

" ¼

cosy siny 0

siny cosy 0

0 0 1

#"

Rx Ry Rz

#

Thus, the characters for the proper rotations are as given by Eq. 1.88. The same result is obtained for the improper rotation if the latter is regarded as i (proper rotation). Therefore, the character for the vibration is obtained from cv ðRÞ ¼ cðRÞ ct ðRÞ cr ðRÞ

ð1:89Þ

It is convenient to tabulate the foregoing calculations as in Table 1.7. By using the formula in Eq. 1.74 and the character of the irreducible representations in Table 1.6, am can be calculated as follows: am ðA1 Þ ¼ 16½ð1Þð6Þð1Þ þ ð2Þð0Þð1Þ þ ð3Þð2Þð1Þ ¼ 2 am ðA2 Þ ¼ 16½ð1Þð6Þð1Þ þ ð2Þð0Þð1Þ þ ð3Þð2Þð 1Þ ¼ 0 am ðEÞ ¼ 16½ð1Þð6Þð2Þ þ ð2Þð0Þð 1Þ þ ð3Þð2Þð0Þ ¼ 2

45

THE NUMBER OF NORMAL VIBRATIONS FOR EACH SPECIES

TABLE 1.7. Symmetry Operation:

I

Kind of Rotation: y:

0

120

180

cos y 1 þ 2 cos y NR c, NR(1 þ 2 cos y) ct , (1 þ 2 cos y) cr , þ (1 þ 2 cos y) cn , c ct cr

1 3 4 12 3 3 6

–1 –1 2 2 1 –1 2

3sv Improper

2C3 Proper

1 2

0 1 0 0 0 0

and cv ¼ 2cA1 þ 2cE

ð1:90Þ

In other words, the six normal vibrations of a pyramidal XY3 molecule are classified into two A1 and two E species. This procedure is applicable to any molecule. As another example, a similar calculation is shown in Table 1.8 for an octahedral XY6 molecule. By use of Eq. 1.74 and the character table in Appendix I, the am are obtained as am ðA1g Þ ¼481 ½ð1Þð15Þð1Þ þ ð8Þð0Þð1Þ þ ð6Þð1Þð1Þ þ ð6Þð1Þð1Þ þ ð3Þð 1Þð1Þ þ ð1Þð 3Þð1Þ þ ð6Þð 1Þð1Þ þ ð8Þð0Þð1Þ þ ð3Þð5Þð1Þ þ ð6Þð3Þð1Þ ¼1

TABLE 1.8. 00

Symmetry Operation

I

8C3

6C2 6C4 Proper

3C24 C2

S2i

6S4

Kind of Rotation: y:

0

120

180

90

180

0

90

cos y 1 þ 2 cos y NR c, NR(1 þ 2 cos y) ct, (1 þ 2 cos y) cr, þ (1 þ 2 cos y) cn, c ct cr

1 3 7 21 3 3 15

1 1 1 1 1 1 1

0 1 3 3 1 1 1

1 1 3 3 1 1 1

a

1 2

0 1 0 0 0 0

1 3 1 3 3 3 3

sh ¼ horizontal plane of symmetry; sd ¼ diagonal plane of symmetry.

0 1 1 1 1 1 1

a

8S6 3sh Improper 120

1 2

0 1 0 0 0 0

180 1 1 5 5 1 1 5

a

6sd

180 1 1 3 3 1 1 3

46

THEORY OF NORMAL VIBRATIONS

am ðA1u Þ ¼ 481 ½ð1Þð15Þð1Þ þ ð8Þð0Þð1Þ þ ð6Þð1Þð1Þ þ ð6Þð1Þð1Þ þ ð3Þð 1Þð1Þ þ ð1Þð 3Þð 1Þ þ ð6Þð 1Þð 1Þ þ ð8Þð0Þð 1Þ þ ð3Þð5Þð 1Þ þ ð6Þð3Þð 1Þ ¼0 .. . and therefore cv ¼ cA1g þ cEg þ 2cF1u þ cF2g þ cF2u

1.9. INTERNAL COORDINATES In Sec. 1.4, the potential and the kinetic energies were expressed in terms of rectangular coordinates. If, instead, these energies are expressed in terms of internal coordinates such as increments of the bond length and bond angle, the corresponding force constants have clearer physical meanings than do those expressed in terms of rectangular coordinates, since these force constants are characteristic of the bond stretching and the angle deformation involved. The number of internal coordinates must be equal to, or greater than, 3N 6 (or 3N 5), the degrees of vibrational freedom of an N-atom molecule. If more than 3N 6 (or 3N 5) coordinates are selected as the internal coordinates, this means that these coordinates are not independent of each other. Figure 1.20 illustrates the internal coordinates for various types of molecules. In linear XYZ (a), bent XY2 (b), and pyramidal XY3 (c) molecules, the number of internal coordinates is the same as the number of normal vibrations. In a nonplanar X2Y2 molecule (d) such as H2O2, the number of internal coordinates is the same as the number of vibrations if the twisting angle around the central bond (Dt) is considered. In a tetrahedral XY4 molecule (e), however, the number of internal coordinates exceeds the number of normal vibrations by one. This is due to the fact that the six angle coordinates around the central atom are not independent of each other, that is, they must satisfy the relation Da12 þ Da23 þ Da31 þ Da14 þ Da24 þ Da34 ¼ 0

ð1:91Þ

This is called a redundant condition. In a planar XY3 molecule (f), the number of internal coordinates is seven when the coordinate Dy, which represents the deviation from planarity, is considered. Since the number of vibrations is six, one redundant

INTERNAL COORDINATES

Fig. 1.20. Internal coordinates for various molecules.

47

48

THEORY OF NORMAL VIBRATIONS

condition such as

Da12 þ Da23 þ Da31 ¼ 0

ð1:92Þ

must be involved. Such redundant conditions always exist for the angle coordinates around the central atom. In an octahedral XY6 molecule (g), the number of internal coordinates exceeds the number of normal vibrations by three. This means that, of the 12 angle coordinates around the central atom, three redundant conditions are involved: Da12 þ Da26 þ Da64 þ Da41 ¼ 0 Da15 þ Da56 þ Da63 þ Da31 ¼ 0

ð1:93Þ

Da23 þ Da34 þ Da45 þ Da52 ¼ 0 The redundant conditions are more complex in ring compounds. For example, the number of internal coordinates in a triangular X3 molecule (h) exceeds the number of 0 vibrations by three. One of these redundant conditions (A1 species) is Da1 þ Da2 þ Da3 ¼ 0

ð1:94Þ

The other two redundant conditions (E0 species) involve bond stretching and angle deformation coordinates such as r ð2Dr1 Dr2 Dr3 Þ þ pﬃﬃﬃ ðDa1 þ Da2 2Da3 Þ ¼ 0 3 r ðDr2 Dr3 Þ pﬃﬃﬃ ðDa1 Da2 Þ ¼ 0 3

ð1:95Þ

where r is the equilibrium length of the XX bond. The redundant conditions mentioned above can be derived by using the method described in Sec. 1.12. The procedure for finding the number of normal vibrations in each species was described in Sec. 1.8. This procedure is, however, considerably simplified if internal coordinates are used. Again, consider a pyramidal XY3 molecule. Using the internal coordinates shown in Fig. 1.20c, we can write the representation for the C3þ operation as 2

Dr1 6 Dr2 6 6 Dr C3þ 6 3 6 Da12 4 Da 23 Da31

3

2

0 1 7 6 7 6 0 7 6 7¼6 0 7 6 5 6 40 0

0 0 1 0 0 0

1 0 0 0 0 0

0 0 0 0 1 0

0 0 0 0 0 1

3 0 2 Dr1 07 Dr 76 6 2 07 6 76 Dr3 Da12 17 76 4 0 5 Da23 Da31 0

3 7 7 7 7 7 5

ð1:96Þ

Thus cðC3þ Þ ¼ 0, as does cðC3 Þ. Similarly, c(I) ¼ 6 and c(sv) ¼ 2. This result is exactly the same as that obtained in Table 1.7 using rectangular coordinates. When

49

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

TABLE 1.9.

r

c (R) ca (R)

I

8C3

6C2

6C4

6 12

0 0

0 2

2 0

3C24 C

00

2

2 0

S2 i

6S4

0 0

0 0

8S6

3sh

6sd

0 0

4 4

2 2

using internal coordinates, however, the character of the representation is simply given by the number of internal coordinates unchanged by each symmetry operation. If this procedure is made separately for stretching (Dr) and bending (Da) coordinates, it is readily seen that cr ðRÞ ¼ cA1 þ cE ca ðRÞ ¼ cA1 þ cE

ð1:97Þ

Thus, it is found that both A1 and E species have one stretching and one bending vibration, respectively. No consideration of the translational and rotational motions is necessary if the internal coordinates are taken as the basis for the representation. Another example, for an octahedral XY6 molecule, is given in Table 1.9. Using Eq. 1.74 and the character table in Appendix I, we find that these characters are resolved into. cr ðRÞ ¼ cA1g þ cEg þ cF1u

ð1:98Þ

ca ðRÞ ¼ cA1g þ cEg þ cF1u þ cF2g þ cF2u

ð1:99Þ

Comparison of this result with that obtained in Sec. 1.8 immediately suggests that three redundant conditions are included in these bending vibrations (one in A1g and one in Eg). Therefore, ca(R) for genuine vibrations becomes ca ðRÞ ¼ cF1u þ cF2g þ cF2u

ð1:100Þ

Thus, it is concluded that six stretching and nine bending vibrations are distributed as indicated in Eqs. 1.98 and 1.100, respectively. Although the method given above is simpler than that of Sec. 1.8, caution must be exercised with respect to the bending vibrations whenever redundancy is involved. In such a case, comparison of the results obtained from both methods is useful in finding the species of redundancy. 1.10. SELECTION RULES FOR INFRARED AND RAMAN SPECTRA According to quantum mechanics [3], the selection rule for the infrared spectrum is determined by the following integral:

50

THEORY OF NORMAL VIBRATIONS

ð ½mv0 v00 ¼ c*v0 ðQa Þmcv00 ðQa ÞdQa

ð1:101Þ

Here, m is the dipole moment in the electronic ground state, c is the vibrational eigenfunction given by Eq. 1.21, and v0 and v00 are the vibrational quantum numbers before and after the transition, respectively. The activity of the normal vibration whose normal coordinate is Qa is being determined. By resolving the dipole moment into the three components in the x, y, and z directions, we obtain the result ð ½mx v0 v00 ¼ c*v0 ðQa Þmx cv00 ðQa ÞdQa ð ½my v0 v00 ¼ c*v0 ðQa Þmy cv00 ðQa ÞdQa ð ½mzv0 v00 ¼ c*v0 ðQa Þmz cv00 ðQa ÞdQa

ð1:102Þ

If one of these integrals is not zero, the normal vibration associated with Qa is infraredactive. If all the integrals are zero, the vibration is infrared-inactive. Similar to the case of infrared spectrum, the selection rule for the Raman spectrum is determined by the integral ð 0 00 ½av v ¼ c*v0 ðQa Þacv00 ðQa ÞdQa ð1:103Þ As shown in Sec. 1.6, a consists of six components, axx,ayy, azz, axy, ayz, and axz. Thus Eq. 1.103 may be resolved into six components:

ð1:104Þ

If one of these integrals is not zero, the normal vibration associated with Qa is Ramanactive. If all the integrals are zero, the vibration is Raman-inactive. As shown below, it is possible to determine whether the integrals of Eqs. 1.102 and 1.104 are zero or nonzero from a consideration of symmetry: 1.10.1. Selection Rules for Fundamentals Let us consider the fundamentals in which transitions occur from v0 ¼ 0 to v00 ¼ 1. It is evident from the form of the vibrational eigenfunction (Eq. 1.22) that c0(Qa) is invariant under any symmetry operation, whereas the symmetry of c1(Qa) is the same as that of Qa. Thus, the integral does not vanish when the symmetry of mx, for example, is the same as that of Qa. If the symmetry properties of mx and Qa differ in even one

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

51

symmetry element of the group, the integral becomes zero. In other words, for the integral to be nonzero, Qa must belong to the same species as mx. More generally, the normal vibration associated with Qa becomes infrared-active when at least one of the components of the dipole moment belongs to the same species as Qa. Similar conclusions are obtained for the Raman spectrum. Namely, the normal vibration associated with Qa becomes Raman-active when at least one of the components of the polarizability belongs to the same species as Qa. Since the species of the normal vibration can be determined by the methods described in Sections 1.8 and 1.9, it is necessary only to determine the species of the components of the dipole moment and polarizability of the molecule. This can be done as follows. The components of the dipole moment, mx, my, and mz, transform as do those of translational motion, Tx, Ty, and Tz, respectively. These were discussed in Sec. 1.8. Thus, the character of the dipole moment is given by Eq. 1.87, which is cu ðRÞ ¼ ð1 þ 2 cosyÞ

ð1:105Þ

where þ and have the same meaning as before. In a pyramidal XY3 molecule, Eq. 1.105 gives

cm ðRÞ

I

2C3

3

0

3sv 1

Using Eq. 1.74, we resolve this into A1 þ E. It is obvious that mz belongs to A1. Then, mx and my must belong to E. In fact, the pair, mx and my, transforms as follows: 2 I

mx my

"

¼

1

0

0

1

#

mx my

;

C3þ

mx

my

6 6 6 ¼6 6 4

s1

mx my

pﬃﬃﬃ 3 3 7 2 7 m 7 x 7 my 17 5 2

cðC3þ Þ ¼ 1

cðIÞ ¼ 2; "

1 2 pﬃﬃﬃ 3 2

#

" ¼

1 0 0 1

#"

mx my

#

cðs1 Þ ¼ 0 Thus, it is found that mz belongs to A1 and (mx, my) belongs to E.

52

THEORY OF NORMAL VIBRATIONS

The character of the representation of the polarizability is given by ca ðRÞ ¼ 2 cosyð1 þ 2 cosyÞ

ð1:106Þ

for both proper and improper rotations. This can be derived as follows. The polarizability in the x, y, and z directions is related to that in X, Y, and Z coordinates by " aXX

aXY

aXZ #

aYX aZX

aYY aZY

aYZ aZZ

"C ¼

Xx

CYx CZx

CXy CYy CZy

CXz #" axx CYz ayx CZz azx

axy ayy azy

axz #" CXx ayz CXy azz CXz

CYx CYy CYz

CZx CZy CZz

#

where CXx, CXy, and so forth denote the direction cosines between the two axes subscripted. If a rotation through y around the Z axis superimposes the X, Y, and Z axes on the x, y, and z axes, the preceding relation becomes "a

xx

Cy ayx azx

axy axz # " cosy siny 0 #" axx axy axz #" cosy ayy ayz ¼ siny cosy 0 ayx ayy ayz siny azy azz 0 0 1 azx azy azz 0

siny 0 cosy 0 0 1

#

This can be written as 3 2 axx cos2 y sin2 y 6 ayy 7 6 7 6 sin2 y 6 cos2 y 7 6 6 azz 7 6 0 0 7 6 Cy 6 6a 7¼ 6 6 6 xy 7 6 siny cosy siny cosy 7 6 4 axz 5 6 4 0 0 ayz 0 0 2

0 0

2 siny cosy 2 siny cosy

1 0

0 2 cos y 1

0

0

0

0

2

3 3 2 0 axx 7 0 7 6a 7 7 6 yy 7 7 6 0 0 7 7 6 azz 7 7 6 0 0 7 6 axy 7 7 7 4 axz 5 5 cosy siny ayz siny cosy 0 0

Thus, the character of this representation is given by Eq. 1.106. The same results are obtained for improper rotations if they are regarded as the product i (proper rotation). For a pyramidal XY3 molecule, Eq. 1.106 gives I ca ðRÞ 6

2C3 3sv 0

2

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

53

Using Eq. 1.74, this is resolved into 2A1 þ 2E. Again, it is immediately seen that* the component azz belongs to A1, and the pair azx and azy belongs to E since "

# zx zy

" # x ¼z ¼ A1 E ¼ E y

It is more convenient to consider the components axx þ ayy and axx ayy than axx and ayy. If a vector of unit length is considered, the relation x2 þ y2 þ z2 ¼ 1 holds. Since azz belongs to A1, axx þ ayy must belong to A1. Then, the pair axx ayy and axy must belong to E. As a result, the character table of the point group C3v, is completed as in Table 1.10. Thus, it is concluded that, in the point group C3v both the A1 and the E vibrations are infrared- as well as Raman-active, while the A2 vibrations are inactive. Complete character tables like Table 1.10 have already been worked out for all the point groups. Therefore, no elaborate treatment such as that described in this section is necessary in practice. Appendix I gives complete character tables for the point groups that appear frequently in this book. From these tables, the selection rules for the infrared and Raman spectra are obtained immediately: The vibration is infrared- or Raman-active if it belongs to the same species as one of the components of the dipole moment or polarizability, respectively. For example, the character table of the point group Oh signifies immediately that only the F1u vibrations are infrared-active and only the A1g, Eg, and F2g vibrations are Raman-active, for the components of the dipole moment or the polarizability belong to these species in this point group. It is to be noted in these character tables that (1) a totally symmetric vibration is Raman-active in any

*The quantum mechanical expression of the polarizability [77] is

axx ¼

2 X ðmx Þj0 v2j0 3h j v2j0 v2

axy ¼

2 X ðmx Þj0 ðmy Þj0 v2j0 3h j v2j0 v2

2

etc:

Here, (mx)j0, for example, is the induced dipole moment along the x axis caused by the 0 (ground state) ! j (excited state) electronic transition, vj0 is the frequency of the 0 ! j transition, and n is the exciting frequency. Thus, it is readily seen that the character of the polarizability components such as axx and axy is determined by considering the product of the characters of dipole moments such as mx and my.

54

THEORY OF NORMAL VIBRATIONS

TABLE 1.10. Character Table of the Point Group C3v C3v

I

2C3

3sv

A1 A2 E

þ1 þ1 þ2

þ1 þ1 1

þ1 1 0

a

mz

axx þ ayy, azz

(mx , my)a

(axz , ayz),a (axx ayy , axy)a

A doubly degenerate pair is represented by two terms in parentheses.

point group, and (2) the infrared- and Raman-active vibrations always belong to u and g types, respectively, in point groups having a center of symmetry. 1.10.2. Selection Rules for Combination and Overtone Bands As stated in Sec. 1.3, some combination and overtone bands appear weakly because actual vibrations are not harmonic and some of these nonfundamentals are allowed by symmetry selection rules. Selection rules for combination bands (ni nj) can be derived from the characters of the direct products of those of individual vibrations. Thus, we obtain cij ðRÞ ¼ ci ðRÞ cj ðRÞ

ð1:107Þ

As an example, consider a molecule of C3v symmetry. It is readily seen that

)

I

2C3

3sv

cA1 ðRÞ cE (R)

1 2

1 1

1 0

cA1 E ðRÞ

2

1

0

cA1 E ðRÞ ¼ cE ðRÞ Thus, a combination band between the A1 and E vibrations is IR- as well as Ramanactive. The activity of a combination band between two E vibrations can be determined by considering the direct product of their characters:

)

I

2C3

3sv

cE (R) cE (R)

2 2

1 1

0 0

cEE (R)

4

1

0

Using Eq. 1.74, this set of the characters can be resolved into cEE ðRÞ ¼ cA1 ðRÞ þ cA2 ðRÞ þ cE ðRÞ Since both A1 and E species are IR- and Raman-active, a combination band between two E vibrations is also IR- and Raman-active. It is convenient to apply the general rules of Appendix IV in determining the symmetry species of direct products.

SELECTION RULES FOR INFRARED AND RAMAN SPECTRA

55

Selection rules for overtones of nondegenerate vibrations can be obtained using the following relation: cn ðRÞ ¼ ½cðRÞn ð1:108Þ Here, cn(R) is the character of the (n 1)th overtone (n ¼ 2 for the first overtone). As an example, consider a molecule of C3v symmetry. The character of the first overtone of the A2 fundamental is calculated as follows:

)

I

2C3

3sv

cA2 ðRÞ cA2 ðRÞ

1 1

1 1

1 1

c2A2 ðRÞ

1

1

1

Namely, it is IR- as well as Raman-active because c2A2 ðRÞ ¼ cA1 ðRÞ. However, the second overtone of the A2 fundamental is IR- as well as Raman-inactive because c3A2 ðRÞ ¼ c2A2 ðRÞ cA2 ðRÞ ¼ cA2 ðRÞ (inactive). In general, odd overtones (A1) are IR- and Raman-active, whereas even overtones (A2) are inactive. It is obvious that all the overtones of the A1 fundamental are IR- and Raman-active. Selection rules for overtones of doubly degenerate vibrations (E species) are determined by ð1:109Þ cnE ðRÞ ¼ 12 cnE 1 ðRÞ cE ðRÞ þ cE ðRn Þ For the first overtone, this is written as h i c2E ðRÞ ¼ 12 fcE ðRÞg2 þ cE ðR2 Þ Here, cE(R2) is the character that corresponds to the operation R performed twice successively. Thus, one obtains cE ðI 2 Þ ¼ cE ðIÞ ¼ 2 cE ½ðC3þ Þ2 ¼ cE ðC3 Þ ¼ 1 cE ½ðsv Þ2 ¼ cE ðIÞ ¼ 2 Therefore, c2E ðRÞ can be calculated as follows: I

2C3

3sv

)

cE(R ) cE(R )

2 2

1 1

0 0

þ)

{cE(R )}2 cE(R2 )

4 2

1 1

0 2

{cE(R )}2þcE(R2 )

6

0

2

c2E ðRÞ

3

0

1

2)

c2E ðRÞ ¼ cA1 ðRÞ þ cE ðRÞ

56

THEORY OF NORMAL VIBRATIONS

Thus, the first overtone of the doubly degenerate vibration is IR- and Raman-active. The characters of overtones for triply degenerate vibrations are given by cnF ðRÞ ¼ 13f2cF ðRÞcnF 1 ðRÞ 12 cnF 2 ðRÞ½cF ðRÞ2 þ 12 cF ðR2 ÞcnF 2 ðRÞ þ cF ðRn Þg

ð1:110Þ

For more details, see Refs. [3,5], and [9].

1.11. STRUCTURE DETERMINATION Suppose that a molecule has several probable structures, each of which belongs to a different point group. Then the number of infrared- and Raman-active fundamentals should be different for each structure. Therefore, the most probable model can be selected by comparing the observed number of infrared- and Raman-active fundamentals with that predicted theoretically for each model. Consider the XeF4 molecule as an example. It may be tetrahedral or square-planar. By use of the methods described in the preceding sections, the number of infrared- or Raman-active fundamentals can be found easily for each structure. Tables 1.11a and 1.11b summarize the results. It is seen that the distinction of these two structures can be made by comparing the number of IR- and Raman-active FXeF bending modes; the tetrahedral structure predicts one IR and two Raman bands, whereas the square–planar structure predicts two IR and one Raman bands. In general, the XeF stretching vibrations appear above 500 cm1, whereas the FXeF bending vibrations are observed below 300 cm1. The IR and Raman spectra of XeF4 are shown in Fig. 2.17. The IR spectrum exhibits two bending bands at 291 and 123 cm1, whereas the Raman

TABLE 1.11a. Number of Fundamentals for Tetrahedral XeF4 Number of Fundamentals

XeF Stretching

FXeF Bending

R (p) ia R (dp) ia IR, R (dp)

1 0 1 0 2

1 0 0 0 1

0 0 1 0 1

IR R

2 4

1 2

1 2

Td

Activitya

A1 A2 E F1 F2 Total a

p, polarized; dp, depolarized (see Sec. 1.20).

57

STRUCTURE DETERMINATION

TABLE 1.11b. Number of Fundamentals for Square-Planar XeF4 D4h

Activity

Number of Fundamentals

XeF Stretching

FXeF Bending

A1g A1u A2g A2u B1g B1u B2g B2u Eg Eu

R (p) ia ia IR R (dp) ia R (dp) ia R (dp) IR

1 0 0 1 1 0 1 1 0 2

1 0 0 0 1 0 0 0 0 1

0 0 0 1 0 0 1 1 0 1

IR R

3 3

1 2

2 1

Total

spectrum shows one bending vibration at 218 cm1. Thus, the square–planar structure is preferable to the tetrahedral structure.* Another example is given by the XeF5 ion, which has the three possible structures shown in Fig. 1.21. The results of vibrational analysis for each are summarized in Table 1.12. It is seen that the numbers of IR-active vibrations are 5, 6, and 3 and those of Raman-active vibrations are 6, 9, and 3, respectively, for the D3h, C4v, and D5h structures, As discussed in Sec. 2.7.3, the XeF5 ion exhibits three IR bands (550–400, 290, and 274 cm1) and three Raman bands (502, 423, and 377 cm1). Thus, a pentagonal planar structure is preferable to the other two structures. The somewhat unusual structures thus obtained for XeF4 and XeF5 can be rationalized by the use of the valence shell electron-pair repulsion (VSEPR) theory (Sec. 2.6.3). This rather simple method has been used widely for the elucidation of molecular structure of inorganic and coordination compounds. In Chapter 2, the number of IRand Raman-active vibrations is compared for possible structures of XYn and other

Fig. 1.21. Three possible structures for the XeF5 ion.

*This conclusion may be drawn directly from observation of the mutual exclusion rule, which holds for D4h but not for Td.

58

THEORY OF NORMAL VIBRATIONS

TABLE 1.12. Number of IR- and Raman-Active Fundamentals for Three Possible Structures of the XeF5 Anion Type IR Raman

C4v

D3h 00

0

2A 2 þ 3E 0 0 00 2A 1 þ 3E þ E

3A1 þ 3E 3A1 þ 2B1 þ B2 þ 3E

D5h 00

0

A 2 þ 2E 1 0 0 A 1 þ 2E 2

molecules. Appendix V lists the number of IR- and Raman-active vibrations of MXnYm-type molecules. It should be noted, however, that this method does not give a clear-cut answer if the predicted numbers of infrared- and Raman-active fundamentals are similar for various probable structures. Furthermore, a practical difficulty arises in determining the number of fundamentals from the observed spectrum, since the intensities of overtone and combination bands are sometimes comparable to those of fundamentals when they appear as satellite bands of the fundamental. This is particularly true when overtone and combination bands are enhanced anomalously by Fermi resonance (accidental degeneracy). For example, the frequency of the first overtone of the n2 vibration of CO2 (667 cm1) is very close to that of the n1 vibration P (1337 cm1). Since these two vibrations belong to the same symmetry species ð gþ Þ, they interact with each other and give rise to two strong Raman lines at 1388 and 1286 cm1. Fermi resonances similar to the resonance observed for CO2 may occur for a number of other molecules. It is to be noted also that the number of observed bands depends on the resolving power of the instrument used. Finally, the molecular symmetry in the isolated state is not necessarily the same as that in the crystalline state (Sec. 1.27). Therefore, this method must be applied with caution to spectra obtained for compounds in the crystalline state. 1.12. PRINCIPLE OF THE GF MATRIX METHOD* As described in Sec. 1.4, the frequency of the normal vibration is determined by the kinetic and potential energies of the system. The kinetic energy is determined by the masses of the individual atoms and their geometrical arrangement in the molecule. On the other hand, the potential energy arises from interaction between the individual atoms and is described in terms of the force constants. Since the potential energy provides valuable information about the nature of interatomic forces, it is highly desirable to obtain the force constants from the observed frequencies. This is usually done by calculating the frequencies, assuming a suitable set of force constants. If the agreement between the calculated and observed frequencies is satisfactory, this particular set of the force constants is adopted as a representation of the potential energy of the system.

* For details, see Ref. 3. The term “normal coordinate analysis” is almost synonymous with the GF matrix method, since most of the normal coordinate calculations are carried out by using this method.

PRINCIPLE OF THE GF MATRIX METHOD

59

To calculate the vibrational frequencies, it is necessary first to express both the potential and the kinetic energies in terms of some common coordinates (Sec. 1.4). Internal coordinates (Sec. 1.9) are more suitable for this purpose than rectangular coordinates, since (1) force constants expressed in terms of internal coordinates have clearer physical meanings than those expressed in terms of rectangular coordinates, and (2) a set of internal coordinates does not involve translational and rotational motion of the molecule as a whole. Using the internal coordinates Ri we write the potential energy as 2V ¼ ~ RFR

ð1:111Þ

For a bent Y1XY2 molecule such as that in Fig. 1.20b, R is a column matrix of the form " R¼

Dr1 Dr2 Da

~R ¼ ½Dr1

Dr2

~R is its transpose:

#

Da

and F is a matrix whose components are the force constants " F¼

f11 f21 r1 f31

f12 f22 r2 f32

r1 f13 r2 f23 r1 r2 f33

#

"

F11 F21 F31

F12 F22 F32

F13 F23 F33

# ð1:112Þ*21

Here r1 and r2 are the equilibrium lengths of the XY1 and XY2 bonds, respectively. The kinetic energy is not easily expressed in terms of the same internal coordinates. Wilson [78] has shown, however, that the kinetic energy can be written as

~_ 1 R_ 2T ¼ RG

ð1:113Þ †

where G1 is the reciprocal of the G matrix, which will be defined later. If Eqs. 1.111 and 1.113 are combined with Lagrange’s equation ! d qT qV ¼0 ð1:36Þ þ dt qR_ k qRk

*Here f11 and f22 are the stretching force constants of the XY1 and XY2 bonds, respectively, and f33 is the bending force constant of the Y1XY2 angle. The other symbols represent interaction force constants between stretching and stretching or between stretching and bending vibrations. To make the dimensions of all the force constants the same, f13 (or f31), f23 (or f32), and f33 are multiplied by r1, r2, and r1r2, respectively. †Appendix VI gives the derivation of Eq. 1.113.

60

THEORY OF NORMAL VIBRATIONS

the following secular equation, which is similar to Eq. 1.49, is obtained: F11 ðG1 Þ l 11 F21 ðG1 Þ l 21 . ..

. . . . . . jF G 1 lj ¼ 0

F12 ðG1 Þ12 l F22 ðG1 Þ22 l .. .

ð1:114Þ

By multiplying by the determinant of G G11 G21 .. .

G12 G22 .. .

jGj

ð1:115Þ

from the left-hand side of Eq. 1.114, the following equation is obtained: P G1t Ft1 l P G2t Ft1 .. .

P P G1t Ft2 G2t Ft2 l .. .

jGF Elj ¼ 0

ð1:116Þ

Here, E is the unit matrix, and l is related to the wavenumber v~ by the relation l ¼ 4p2 c2 v~2 .* The order of the equation is equal to the number of internal coordinates used. The F matrix can be written by assuming a suitable set of force constants. If the G matrix is constructed by the following method, the vibrational frequencies are obtained by solving Eq. 1.116. The G matrix is defined as G ¼ BM1~ B

ð1:117Þ

1

..

.

Here M is a diagonal matrix whose components are mi, where mi is the reciprocal of the mass of the ith atom. For a bent XY2 molecule, we obtain 2 3 m1 6 m1 0 7 6 7 1 6 7 m M ¼6 1 7 4 5 0

m3

where m3 and m1 are the reciprocals of the masses of the X and Y atoms, respectively. The B matrix is defined as R ¼ BX ð1:118Þ where R and X are column matrices whose components are the internal and rectangular coordinates, respectively.

*Here l should not be confused with lw (wavelength).

PRINCIPLE OF THE GF MATRIX METHOD

61

Fig. 1.22. Relationship between internal and rectangular coordinates in (a) stretching and (b) bending vibrations of a bent XY2 molecule.

To write Eq. 1.118 for a bent XY2 molecule, we express the bond stretching (Dr1) in terms of the x and y coordinates shown in Fig. 1.22a. It is readily seen that Dr1 ¼ ðDx1 ÞðsÞ ðDy1 ÞðcÞ þ ðDx3 ÞðsÞ þ ðDy3 ÞðcÞ Here, s ¼ sin(a=2), c ¼ cos(a=2), and r is the equilibrium distance between the X and Y atoms. A similar expression is obtained for Dr2. Thus Dr2 ¼ ðDx2 ÞðsÞ ðDy2 ÞðcÞ ðDx3 ÞðsÞ þ ðDy3 ÞðcÞ For the bond bending (Da), consider the relationships illustrated in Fig. 1.22b. It is readily seen that rðDaÞ ¼ ðDx1 ÞðcÞ þ ðDy1 ÞðsÞ þ ðDx2 ÞðcÞ þ ðDy2 ÞðsÞ 2ðDy3 ÞðsÞ

62

THEORY OF NORMAL VIBRATIONS

If these equations are summarized in a matrix form, we obtain

ð1:119Þ

If unit vectors such as those in Fig. 1.23 are considered, Eq. 1.119 can be written in a more compact form using vector notationr: "

Dr1 Dr2 Da

#

" ¼

e31 0 p31 =r

0 e32 p32 =r

e31 e32 ðp31 þ p32 Þ=r

#"

q1 q2 q3

# ð1:120Þ

Here q1, q2, and q3 are the displacement vectors of atoms 1, 2, and 3, respectively. Thus Eq. 1.120 can be written simply as R ¼ Sq

ð1:121Þ

where the dot represents the scalar product of the two vectors. Here S is called the S matrix, and its components (S vector) can be written according to the following formulas: (1) bond stretching Dr1 ¼ D31 ¼ e31 q1 e31 q3

Fig. 1.23. Unit vectors in a bent XY2 molecule.

ð1:122Þ

PRINCIPLE OF THE GF MATRIX METHOD

63

and (2) angle bending: p31 q1 þ p32 q2 ðp31 þ p32 Þ q3 ð1:123Þ r It is seen that the S vector is oriented in the direction in which a given displacement of the ith atom will produce the greatest increase in Dr or Da. Formulas for obtaining the S vectors of other internal coordinates such as those of out-of-plane (Dy) and torsional (Dt) vibrations are also available [3]. By using the S matrix, Eq. 1.117 is written as Da ¼ Da132 ¼

G ¼ Sm 1~ S For a bent XY2 molecule, this becomes 2 0 e31 e31 6 e32 e32 G ¼ 40 2

p31 =r

e31 6 40 e31 2 6 6 6 6 ¼6 6 6 6 4

32

m1 76 54 0

p32 =r

ðp31 þ p32 Þ=r

0 e32

p31 =r p32 =r

e32

ðp31 þ p32 Þ=r

ðm3 þ m1 Þe231

m3 e31 e32 ðm3 m1 Þe232

ð1:124Þ

0 m1

0

0

3

3 0 7 0 5 m3

7 5 3 m1 m e31 p31 þ 3 e31 ðp31 þ p32 Þ 7 r r 7 7 m1 m3 e32 p32 þ e32 ðp31 þ p32 Þ 7 7 r r 7 7 7 m1 2 m1 2 m3 25 p þ p þ ðp þ p Þ 31 32 31 32 r2 r2 r2

Considering e31 e31 ¼ e32 e32 ¼ p31 p31 ¼ p32 p32 ¼ 1; e31 e32 ¼ cos a;

e31 p31 ¼ e32 p32 ¼ 0 e31 p32 ¼ e32 p31 ¼ sin a

ðp31 þ p32 Þ2 ¼ 2ð1 cos aÞ we find that the G matrix is calculated as follows: 3 2 m m3 þ m1 m3 cos a 3 sin a r 7 6 7 6 m3 7 6 7 6 m sin a þ m 3 1 G¼6 7 r 7 6 7 6 2m1 2m3 5 4 þ ð1 cos aÞ r2 r2

ð1:125Þ

64

THEORY OF NORMAL VIBRATIONS

If the G-matrix elements obtained are written for each combination of internal coordinates, there results GðDr1 ; Dr1 Þ ¼ m3 þ m1 GðDr2 ; Dr2 Þ ¼ m3 þ m1 GðDr1 ; Dr2 Þ ¼ m3 cos a 2m 2m GðDa; DaÞ ¼ 21 þ 23 ð1 cos aÞ r r GðDr1 ; DaÞ ¼

ð1:126Þ

m3 sin a r

m3 sin a r If such calculations are made for several types of molecules, it is immediately seen that the G-matrix elements themselves have many regularities. Decius [79] developed general formulas for writing G-matrix elements.* Some of them are as follows: GðDr2 ; DaÞ ¼

G2rr ¼ m1 þ m2

G1rr ¼ m1 cos f

G2rf ¼ r23 m2 sin f

G1rf

1 1

! ¼ ðr13 sin f213 cos c234 þ r14 sin f214 cos c243 Þm1

G3ff ¼ r212 m1 þ r223 m3 þ ðr212 þ r223 2r12 r23 cos fÞm2 G2

1 ¼ 1

ðr212 cos c314 Þm1 þ ½r12 r23 cos 123 r24 cos 124 Þr12 cos c314 þðsin 123 sin 124 sin2 c314 þ cos 324 cos c314 Þr23 r24 m2

*See also Refs. 3 and 80.

UTILIZATION OF SYMMETRY PROPERTIES

65

Fig. 1.24. Spherical angles involving atomic positions, a, b, g, and d.

Here, the atoms surrounded by a bold line circle are those common to both coordinates. The symbols m and r denote the reciprocals of mass and bond distance, respectively. The spherical angle cabg in Fig. 1.24 is defined as coscabg ¼

cos fadg cos fadb cos fbdg sin fadb sin fbdg

ð1:127Þ

The correspondence between the Decius formulas and the results obtained in Eq. 1.126 is evident. With the Decius formulas, the G-matrix elements of a pyramidal XY3 molecule have been calculated and are shown in Table 1.13.

1.13. UTILIZATION OF SYMMETRY PROPERTIES In view of the equivalence of the two XY bonds of a bent XY2 molecule, the F and G matrices obtained in Eqs. 1.112 and 1.125 are written as " F¼ 2

m3 þ m1

6 6 6 6 G ¼ 6 m3 cosa 6 6 m 4 3 sina r

f11 f12 rf13

m3 cosa m3 þ m1

m3 sina r

f12 f11 rf13

rf13 rf13 r 2 f33

#

3 m3 sina r 7 7 m3 7 7 sina 7 r 7 7 2m1 2m3 5 þ ð1 cosaÞ r2 r2

ð1:128Þ

ð1:129Þ

66

THEORY OF NORMAL VIBRATIONS

TABLE 1.13.

Dr1 Dr2 Dr3 Da23 Da31 Da12

Dr1

Dr2

Dr3

Da23

Da31

Da12

A — — — — —

B A — — — —

B B A — — —

C D D E — —

D C D F E —

D D C F F E

A ¼ G2rr ¼ mx þ my B ¼ G1rr ¼ mx cos a ! 1 2 cos að1 cos aÞmx ¼ C ¼ G1rf r sin a 1 mx 2 D ¼ G2f ¼ sin a r 2 E ¼ G3ff ¼ 2 my þ mx ð1 cos aÞ r! my 1 cos a m ð1 þ 3 cos aÞð1 cos aÞ þ 2x ¼ 2 F ¼ G2ff 1 þ cos a r 1 þ cos a r 1

Both of these matrices are of the form "

A C D

C A D

D D B

# ð1:130Þ

The appearance of the same elements is evidently due to the equivalence of the two internal coordinates, Dr1 and Dr2. Such symmetrically equivalent sets of internal coordinates are seen in many other molecules, such as those in Fig. 1.20. In these cases, it is possible to reduce the order of the F and G matrices (and hence the order of the secular equation resulting from them) by a coordinate transformation. Let the internal coordinates R be transformed by Rs ¼ UR Then, we obtain

~_ 1 R_ ¼ R ~_ s ~U1 G1 U1 R_ s 2T ¼ RG ~_ s G1 _s ¼R s R 2V ¼ ~ RFR ¼ ~ Rs ~ U1 FU1 Rs ¼~ Rs Fs Rs

ð1:131Þ

UTILIZATION OF SYMMETRY PROPERTIES

67

where G1 U1 G1 U 1 s ¼~

or

Fs ¼ ~ U1 FU1

Gs ¼ UG~ U

ð1:132Þ

If U is an orthogonal matrix ðU 1 ¼ ~ UÞ, Eq. 1.132 is written as Fs ¼ UF~ U

and

Gs ¼ UG~ U

ð1:133Þ

Both GF and GsFs give the same roots, since jGs Fs Elj ¼ jUG~ U~ U1 FU1 Elj ¼ jUGFU1 Elj

ð1:134Þ

¼ jUkGF ElkU1 j ¼ jGF Elj

If we choose a proper U matrix from symmetry consideration, it is possible to factor the original G and F matrices into smaller ones. This, in turn, reduces the order of the secular equation to be solved, thus facilitating their solution. These new coordinates Rs are called symmetry coordinates. The U matrix is constructed by using the equation Rs ¼ N

X

ci ðKÞKðDr1 Þ

ð1:135Þ

K

Here, K is a symmetry operation, and the summation is made over all symmetry operations. Also, ci(K) is the character of the representation to which Rs belongs. Called a generator, Dr1 is, by symmetry operation K, transformed into K(Dr1), which is another coordinate of the same symmetrically equivalent set. Finally, N is a normalizing factor. As an example, consider a bent XY2 molecule in which Dr1 and Dr2 are equivalent. Using Dr1 as a generator, we obtain I K(Dr1) cA1 ðKÞ cB2 ðKÞ

Dr1 1 1

C2(z)

s(xz)

s(yz)

Dr2 1 –1

Dr2 1 –1

Dr1 1 1

68

THEORY OF NORMAL VIBRATIONS

Thus RsA1 ¼ N

X

1 N ¼ pﬃﬃﬃ 2 2

cA1 ðKÞKðDr1 Þ ¼ 2NðDr1 þ Dr2 Þ sinceð2NÞ2 þ ð2NÞ2 ¼ 1

Then 1 RSA1 ¼ pﬃﬃﬃ ðDr1 þ Dr2 Þ 2

ð1:136Þ

1 RsB2 ¼ pﬃﬃﬃ ðDr1 Dr2 Þ 2

ð1:137Þ

Similarly,

The remaining internal coordinate, Da, belongs to the A1 species. Thus, the complete U matrix is written as 2 1 pﬃﬃﬃ 2 " Rs ðA Þ # 6 6 1 1 6 s R2 ðA1 Þ ¼ 6 6 0 6 1 Rs3 ðB2 Þ 4 pﬃﬃﬃ 2

1 pﬃﬃﬃ 2 0 1 pﬃﬃﬃ 2

3 0

7" 7 Dr1 # 7 17 7 Dr2 7 Da 05

ð1:138Þ

If the G and F matrices of type 1.130 are transformed by relations 1.133, where U is given by the matrix of Eq. 1.138, they become

ð1:139Þ

or, more explicitly

ð1:140Þ

UTILIZATION OF SYMMETRY PROPERTIES

69

ð1:141Þ

In a pyramidal XY3 molecule (Fig. 1.20c), Dr1, Dr2, and Dr3 are the equivalent set; so are Da23, Da31, and Da12. It is already known from Eq. 1.107 that one A1 and one E vibration are involved both in the stretching and in the bending vibrations. Using Dr1 as a generator, we obtain from Eq. 1.135

K(Dr1) cA1 ðKÞ cE(K)

I

C3þ

C3

s1

s2

s3

Dr1 1 2

Dr2 1 1

Dr3 1 1

Dr1 1 0

Dr3 1 0

Dr2 1 0

Then, we obtain 1 RsA1 ¼ pﬃﬃﬃ ðDr1 þ Dr2 þ Dr3 Þ 3

ð1:142Þ

1 RsE1 ¼ pﬃﬃﬃ ð2Dr1 Dr2 Dr3 Þ 6

ð1:143Þ

To find a coordinate that forms a degenerate pair with Eq. 1.143, we repeat the same procedure, using Dr2 and Dr3 as the generators. The results are RsE2 ¼ Nð2Dr2 Dr3 Dr1 Þ RsE2 ¼ Nð2Dr3 Dr1 Dr2 Þ However, these coordinates are not orthogonal to RsE1 (Eq. 1.143). If we take a linear combination, RsE2 þ RsE3 , we obtain Eq. 1.143. If we take s RE2 RsE3 , we obtain 1 RsE4 ¼ pﬃﬃﬃ ðDr2 Dr3 Þ 2

ð1:144Þ

70

THEORY OF NORMAL VIBRATIONS

Since Eqs. 1.143 and 1.144 are mutually orthogonal, these two coordinates are taken as a degenerate pair. Similar results are obtained for three angle-bending coordinates. Thus, the complete U matrix is written as pﬃﬃﬃ pﬃﬃﬃ 3 2 3 3 2 pﬃﬃﬃ Rs1 ðA1 Þ Dr1 1= 3 1= 3 1= 3 0 0 0 p ﬃﬃ ﬃ p ﬃﬃ ﬃ p ﬃﬃ ﬃ 7 6 6 Rs ðA1 Þ 7 6 0 0 0 1= 3 1= 3 1= 3 7 7 6 Dr2 7 7 6 6 2 pﬃﬃﬃ pﬃﬃﬃ 7 6 7 7 6 pﬃﬃﬃ 6 s 6 Dr 7 6 R3a ðEÞ 7 6 2= 6 1= 6 1= 6 0 0 0 7 76 3 7 7¼6 6 p ﬃﬃ ﬃ p ﬃﬃ ﬃ p ﬃﬃ ﬃ 7 6 6 Rs ðEÞ 7 6 0 0 0 2= 6 1= 6 1= 6 7 7 6 Da23 7 7 6 6 4a pﬃﬃﬃ pﬃﬃﬃ 7 6 7 7 6 6 s 4 R3b ðEÞ 5 4 0 1= 2 1= 2 0 0 0 5 4 Da31 5 pﬃﬃﬃ pﬃﬃﬃ Rs4b ðEÞ Da12 0 0 0 0 1= 2 1= 2 ð1:145Þ 2

The G matrix of a pyramidal XY3 molecule has already been calculated (see Table 1.14). By using Eq. 1.133, the new Gs matrix becomes

(1.146) Here A, B, and so forth denote the elements in Table 1.13. The F matrix transforms similarly. Therefore, it is necessary only to solve two quadratic equations for the A1 and E species. For the tetrahedral XY4 molecule shown in Fig. 1.20e, group theory (Secs. 1.8 and 1.9) predicts one A1 and one F2 stretching, and one E and one F2 bending, vibration. The U matrix for the four stretching coordinates is 2

Rs1 ðA1 Þ

3

2

6 Rs ðF Þ 7 6 6 2a 2 7 6 6 s 7¼6 4 R2b ðF2 Þ 5 4 Rs2c ðF2 Þ

1=2 1=2 pﬃﬃﬃ pﬃﬃﬃ 1= 6 1= 6 pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ 1= 12 1= 12 pﬃﬃﬃ pﬃﬃﬃ 1= 2 1= 2

1=2 pﬃﬃﬃ 2= 6 pﬃﬃﬃﬃﬃ 1= 12 0

32 3 Dr1 1=2 76 Dr 7 0 6 27 pﬃﬃﬃﬃﬃ 7 76 7 3= 12 54 Dr3 5 0

Dr4

ð1:147Þ

POTENTIAL FIELDS AND FORCE CONSTANTS

71

whereas the U matrix for the six bending coordinates becomes 2

pﬃﬃﬃ

1= 6 3 Rs1 ðA1 Þ 6 pﬃﬃﬃﬃﬃ 2= 12 6 Rs ðEÞ 7 6 7 6 6 2a 7 6 6 s 6 6 R2b ðEÞ 7 6 0 7 6 pﬃﬃﬃﬃﬃ 6 Rs ðF Þ 7 ¼ 6 6 3a 2 7 6 2= 12 7 6 pﬃﬃﬃ 6 s 4 R3b ðF2 Þ 5 6 6 1= 6 4 Rs3c ðF2 Þ 2

0

pﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃ 1= 6 1= 6 1= 6 1= 6 pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ 1= 12 1= 12 1= 12 1= 12 1=2 1=2 1=2 pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ 1= 12 1= 12 1= 12 pﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃ 1= 6 1= 6 1= 6 1=2

1=2

1=2

1=2 pﬃﬃﬃﬃﬃ 1= 12 pﬃﬃﬃ 1= 6 1=2

3 Da12 6 Da 7 6 23 7 7 6 6 Da31 7 7 6 6 7 6 Da14 7 7 6 4 Da24 5

pﬃﬃﬃ 3 1= 6 pﬃﬃﬃﬃﬃ 7 2= 12 7 7 7 7 0 7 pﬃﬃﬃﬃﬃ 7 2= 12 7 7 pﬃﬃﬃ 7 1= 6 7 5 0

2

ð1:148Þ

Da34 The symmetry coordinate Rs1 ðA1 Þ in Eq. 1.148 represents a redundant coordinate (see Eq. 1.91). In such a case, a coordinate transformation reduces the order of the matrix by one, since all the G-matrix elements related to this coordinate become zero. Conversely, this result provides a general method of finding redundant coordinates. Suppose that the elements of the G matrix are calculated in terms of internal coordinates such as those in Table 1.13. If a suitable combination of internal coordinates is made so that Sj Gij ¼ 0 (where j refers to all the equivalent internal coordinates), such a combination is a redundant coordinate. By using the U matrices in Eqs. 1.147 and 1.148, the problem of solving a tenth-order secular equation for a tetrahedral XY4 molecule is reduced to that of solving two first-order (A1 and E) and one quadratic (F2) equation. The normal modes of vibration of the tetrahedral XY4 molecule are shown in Fig. 2.12 of Chapter 2. It can be seen that the normal modes, n1 and n3, correspond to the symmetry coordinates Rs1 ðA1 Þ and Rs2b ðF2 Þ of Eq. 1.147, respectively, and that the normal modes, n2 and n4, correspond to the symmetry coordinates Rs2a ðEÞ and Rs3b ðF2 Þ of Eq. 1.148, respectively. 1.14. POTENTIAL FIELDS AND FORCE CONSTANTS Using Eqs. 1.111 and 1.128, we write the potential energy of a bent XY2 molecule as 2V ¼ f11 ðDr1 Þ2 þ f11 ðDr2 Þ2 þ f33 r 2 ðDaÞ2 þ 2f12 ðDr1 ÞðDr2 Þ þ 2f13 rðDr1 ÞðDaÞ þ 2f13 rðDr2 ÞðDaÞ

ð1:149Þ

72

THEORY OF NORMAL VIBRATIONS

This type of potential field is called a generalized valence force (GVF) field.* It consists of stretching and bending force constants, as well as the interaction force constants between them. When using such a potential field, four force constants are needed to describe the potential energy of a bent XY2 molecule. Since only three vibrations are observed in practice, it is impossible to determine all four force constants simultaneously. One method used to circumvent this difficulty is to calculate the vibrational frequencies of isotopic molecules (e.g., D2O and HDO for H2O), assuming the same set of force constants.† This method is satisfactory, however, only for simple molecules. As molecules become more complex, the number of interaction force constants in the GVF field becomes too large to allow any reliable evaluation. In another approach, Shimanouchi [82] introduced the Urey–Bradley force (UBF) field, which consists of stretching and bending force constants, as well as repulsive force constants between nonbonded atoms. The general form of the potential field is given by X 1 0 V¼ Ki ðDri Þ2 þ Ki ri ðDri Þ 2 i

!

! X 1 0 2 2 2 Hi ria þ ðDai Þ þ Hi ria ðDai Þ 2 i ! X 1 0 2 Fi ðDqi Þ þ Fi qi ðDqi Þ þ 2 i

ð1:150Þ

Here Dri, Dai, and Dqi, are the changes in the bond lengths, bond angles, and distances 0 0 0 between nonbonded atoms, respectively. The symbols Ki ; Ki ; Hi ; Hi and Fi ; Fi represent the stretching, bending, and repulsive force constants, respectively. Furthermore, ri, ria, and qi are the values of the distances at the equilibrium positions and are inserted 0 0 to make the force constants dimensionally similar. Linear force constants (Ki ; Hi and 0 Fi ) must be included since Dri, Dai, and Dqi are not completely independent of each other. Using the relation q2ij ¼ ri2 þ rj2 2ri rj cosaij

ð1:151Þ

*A potential field consisting of stretching and bending force constants only is called a simple valence force field. †In addition to isotope frequency shifts, mean amplitudes of vibration, Coriolis coupling constants, centrifugal distortion constants, and so forth may be used to refine the force constants of small molecules (see Ref. 81).

POTENTIAL FIELDS AND FORCE CONSTANTS

73

and considering that the first derivatives can be equated to zero in the equilibrium case, we can write the final form of the potential field as # " X 1X 2 0 2 ðtij Fij þ sij Fij Þ ðDri Þ2 V¼ Ki þ 2 i jð„iÞ

1X 0 pﬃﬃﬃﬃﬃﬃﬃ ðHij sij sji Fij þ tij tji Fij Þð ri rj Daij Þ2 2 i<j X 0 þ ð tij tji Fij þ sij sji Fij ÞðDri ÞðDrj Þ þ

i<j

þ

X i„j

rj ðtij sji Fij þ tji sij Fij Þ ri

!1=2

0

ð1:152Þ*

pﬃﬃﬃﬃﬃﬃﬃ ðDri Þð ri rj Daij Þ

Here sij ¼

ri rj cosaij qij

sji ¼

rj ri cosaij qij

tij ¼

rj sinaij qij

tji ¼

ri sinaij qij

ð1:153Þ

In a bent XY2 molecule, Eq. 1.152 becomes h i 0 0 V ¼ 12ðK þ t2 F þ s2 FÞ ðDr1 Þ2 þ ðDr2 Þ2 þ 12ðH s2 F þ t2 FÞðrDaÞ2 0

0

þ ðt2 F þ s2 FÞðDr1 ÞðDr2 Þ þ tsðF þ FÞðDr1 ÞðrDaÞ 0

þ tsðF þ FÞðDr2 ÞðrDaÞ where *In the case of tetrahedral molecules, a term X i„j„k

! k pﬃﬃﬃ rij rik ðrij Daij Þðrik Daik Þ 2

must be added, where k is called the internal tension.

ð1:154Þ

74

THEORY OF NORMAL VIBRATIONS

s¼

rð1 cosaÞ q

t¼

rsina q

Comparing Eqs. 1.154 and 1.149, we obtain the following relations between the force constants of the generalized valence force field and those of the Urey–Bradley force field: 0

f11 ¼ K þ t2 F þ s2 F 0

r 2 f33 ¼ ðH s2 F þ t2 FÞr 2 0

f12 ¼ t2 F þ s2 F

ð1:155Þ

0

rf13 ¼ tsðF þ FÞr Although the Urey–Bradley field has four force constants, F0 is usually taken as 1 10 F, on the assumption that the repulsive energy between nonbonded atoms is proportional to 1/r9.* Thus, only three force constants, K, H, and F, are needed to construct the F matrix. The orbital valence force (OVF) field developed by Heath and Linnett [83] is similar to the UBF field. The OVF field uses the angle (Db), which represents the distortion of the bond from the axis of the bonding orbital instead of the angle between two bonds (Da). The number of force constants in the Urey–Bradley field is, in general, much smaller than that in the generalized valence force field. In addition, the UBF field has the advantage that (1) the force constants have clearer physical meanings than those of the GVF field, and (2) they are often transferable from molecule to molecule. For example, the force constants obtained for SiCl4 and SiBr4 can be used for SiCl3Br, SiCl2Br2, and SiClBr3. Mizushima, Shimanouchi, and their coworkers [81] and Overend and Scherer [84] have given many examples that demonstrate the transferability of the force constants in the UBF field. This property of the Urey–Bradley force constants is highly useful in calculations for complex molecules. It should be mentioned, however, that ignorance of the interactions between nonneighboring stretching vibrations and between bending vibrations in the Urey–Bradley field sometimes causes difficulties in adjusting the force constants to fit the observed frequencies. In such a case, it is possible to improve the results by introducing more force constants [84,85].

*This assumption does not cause serious error in final results, since F0 is small in most cases.

SOLUTION OF THE SECULAR EQUATION

75

Evidently, the values of force constants depend on the force field initially assumed. Thus, a comparison of force constants between molecules should not be made unless they are obtained by using the same force field. The normal coordinate analysis developed in Secs. 1.12–1.14 has already been applied to a number of molecules of various structures. Appendix VII lists the G and F matrix elements for typical molecules. 1.15. SOLUTION OF THE SECULAR EQUATION Once the G and F matrices are obtained, the next step is to solve the matrix secular equation: jGF Elj ¼ 0

ð1:116Þ

In diatomic molecules, pﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ G ¼ G11 ¼ 1/m and F ¼ F11 ¼ K. Then l ¼ G11F11 and ~n ¼ l=2pc ¼ K=m=2pc (Eq. 1.20). If the units of mass and force constant are atomic weight and mdyn/A (or 105 dyn/cm), respectively,* l is related to ~nðcm 1 Þ by pﬃﬃﬃ ~n ¼ 1302:83 l or ~n l ¼ 0:58915 1000

!2 ð1:156Þ

As an example, for the HF molecule m ¼ 0.9573 and K ¼ 9.65 in these units. Then, from Eqs. 1.20 and 1.156, ~n is 4139 cm1. The F and G matrix elements of a bent XY2 molecule are given in Eqs. 1.140 and 1.141, respectively. The secular equation for the A1 species is quadratic: G F þ G12 F21 l jGF Elj ¼ 11 11 G21 F11 þ G22 F21

G11 F12 þ G12 F22 ¼0 G21 F12 þ G22 F22 l

ð1:157Þ

If this is expanded into an algebraic equation, the following result is obtained: 2 l2 ðG11 F11 þ G22 F22 þ 2G12 F12 Þl þ ðG11 G22 G212 ÞðF11 F22 F12 Þ¼0

ð1:158Þ

*Although the bond distance is involved in both the G and F matrices, it is canceled during multiplication of the G and F matrix elements. Therefore, any unit can be used for the bond distance.

76

THEORY OF NORMAL VIBRATIONS

For the H2O molecule, we obtain m1 ¼ mH ¼

1 ¼ 0:99206 1:008

m3 ¼ mO ¼

1 ¼ 0:06252 15:995

a ¼ 105

r ¼ 0:96ðAÞ;

sina ¼ sin 105 ¼ 0:96593 cosa ¼ cos 105 ¼ 0:25882 Then, the G matrix elements of Eq. 1.141 are G11 ¼ m1 þ m3 ð1 þ cosaÞ ¼ 1:03840 pﬃﬃﬃ 2 m sina ¼ 0:08896 G12 ¼ r 3 G22 ¼

1 ½2m1 þ 2m3 ð1 cosaÞ ¼ 2:32370 r2

If the force constants in terms of the generalized valence force field are selected as f11 ¼ 8:4280;

f12 ¼ 0:1050

f13 ¼ 0:2625;

f33 ¼ 0:7680

the F matrix elements of Eq. 1.140 are F11 ¼ f11 þ f12 ¼ 8:32300 pﬃﬃﬃ F12 ¼ 2rf13 ¼ 0:35638 F22 ¼ r 2 f33 ¼ 0:70779 Using these values, we find that Eq. 1.158 becomes l2 10:22389l þ 13:86234 ¼ 0 The solution of this equation gives l1 ¼ 8:61475;

l2 ¼ 1:60914

If these values are converted to ~n through Eq. 1.156, we obtain ~n1 ¼ 3824 cm 1 ;

~n2 ¼ 1653 cm 1

VIBRATIONAL FREQUENCIES OF ISOTOPIC MOLECULES

77

With the same set of force constants, the frequency of the B2 vibration is calculated as l3 ¼ G33 F33 ¼ ½m1 þ m3 ð1 cos aÞ ðf11 f12 Þ ¼ 9:13681 n~3 ¼ 3938 cm1 The observed frequencies corrected for anharmonicity are as follows: w1 ¼ 3825 cm1, w2 ¼ 1654 cm1, and w3 ¼ 3936 cm1. If the secular equation is third order, it gives rise to a cubic equation: l3 ðG11 F11 þ G22 F22 þ G33 F33 þ 2G12 F12 þ 2G13 F13 þ 2G23 F23 Þl2 ( G11 G12 F11 F12 G 12 G13 F12 F13 G11 G13 F11 F13 þ þ þ G21 G22 F21 F22 G22 G23 F22 F23 G21 G23 F21 F23 G11 G12 F11 F12 G12 G13 F12 F13 þ þ G31 G32 F31 F32 G32 G33 F32 F33 G G13 F11 F13 G21 G22 F21 F22 G22 G23 F22 F23 þ 11 þ þ G31 G33 F31 F33 G31 G32 F31 F32 G32 G33 F32 F33 ) G11 G12 G13 F11 F12 F13 G21 G23 F21 F23 þ l G21 G22 G33 F21 F23 F33 ¼ 0 ð1:159Þ G31 G33 F31 F33 G31 G32 G33 F31 F32 F33 Thus, it is possible to solve the secular equation by expanding it into an algebraic equation. If the order of the secular equation is higher than three, however, direct expansion such as that just shown becomes too cumbersome. There are several methods of calculating the coefficients of an algebraic equation using indirect expansion [3]. The use of an electronic computer greatly reduces the burden of calculation. Excellent programs written by Schachtschneider [86] and other workers are available for the vibrational analysis of polyatomic molecules. 1.16. VIBRATIONAL FREQUENCIES OF ISOTOPIC MOLECULES As stated in Sec. 1.14, the vibrational frequencies of isotopic molecules are very useful in refining a set of force constants in vibrational analysis. For large molecules, isotopic substitution is indispensable in making band assignments, since only vibrations involving the motion of the isotopic atom will be shifted by isotopic substitution. Two important rules hold for the vibrational frequencies of isotopic molecules. The first, called the product rule, can be derived as follows. Let l1, l2,. . ., ln be the roots of the secular equation |GF El| ¼ 0. Then l1 l2 ln ¼ jGkFj

ð1:160Þ

78

THEORY OF NORMAL VIBRATIONS

holds for a given molecule. Since the isotopic molecule has exactly the same |F| as that in Eq. 1.160, a similar relation 0

0

0

0

l 1 l 2 l n ¼ jG kFj holds for this molecule. It follows that l1 l2 ln jGj 0 0 0 ¼ 0 l 1 l 2 l n jG j

ð1:161Þ

Since ~n ¼

1 pﬃﬃﬃ l 2pc

Equation 1.161 can be written as ~n1~n2 ~nn 0 0 0 ¼ ~n 1~n 2 ~n n

sﬃﬃﬃﬃﬃﬃﬃﬃ jGj 0 jG j

ð1:162Þ

This rule has been confirmed by using pairs of molecules such as H2O and D2O and CH4 and CD4. The rule is also applicable to the product of vibrational frequencies belonging to a single symmetry species. A more general form of Eq. 1.162 is given by the Redlich-Teller product rule [1]: vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ! ! !t !dx ! dy ! dz u u m0 a m0 b ~n1~n2 ~nn M Ix Iy Iz 1 2 t 0 0 0 ¼ M0 m1 m2 I 0x I 0y I 0z ~n 1~n 2 ~n n

ð1:163Þ

Here m1, m2,. . . are the masses of the representative atoms of the various sets of equivalent nuclei (atoms represented by m,m0,mxy,. . . in the tables given in Appendix III); a, b,. . . are the coefficients of m,m0,mxy,. . .; M is the total mass of the molecule; t is the number of Tx,Ty,Tz in the symmetry type considered; Ix,Iy,Iz are the moments of inertia about the x, y, z axes, respectively, which go through the center of the mass; and dx,dy,dz are 1 to 0, depending on whether Rx,Ry,Rz belong to the symmetry type considered. A degenerate vibration is counted only once on both sides of the equation. Another useful rule in regard to the vibrational frequencies of isotopic molecules, called the sum rule, can be derived as follows. It is obvious from Eqs. 1.158 and 1.159 that l1 þ l2 þ þ l n ¼

X n

l¼

X i; j

Gij Fij

ð1:164Þ

79

METAL–ISOTOPE SPECTROSCOPY

Let sk denote SijGijFij for k different isotopic molecules, all of which have the same F matrix. If a suitable combination of molecules is taken, so that X

s1 þ s2 þ þ sk ¼ "

!

¼

þ

Gij Fij

X

X

!

!

X

þ

Gij

þ þ

Gij Fij !

1

X

2

! Gij Fij

!#

2

þ þ

Gij

1

X

Gij

!

k

X

Fij

k

¼0 then it follows that X

! þ

l 1

X

! þ þ

l

X

! ¼0

l

2

ð1:165Þ

k

This rule has been verified for such combinations as H2O, D2O, and HDO, where 2sðHDOÞ sðH2 OÞ sðD2 OÞ ¼ 0 Such relations between the frequencies of isotopic molecules are highly useful in making band assignments. 1.17. METAL–ISOTOPE SPECTROSCOPY [87,88] As a first approximation, vibrational spectra of coordination compounds (Chapter 3, in Part B) can be classifed into ligand vibrations that occur in the high-frequency region (4000–600 cm1) and metal–ligand vibrations that appear in the lowfrequency region (below 600 cm1). The former provide information about the effect of coordination on the electronic structure of the ligand, while the latter provide direct information about the structure of the coordination sphere and the nature of the metal–ligand bond. Since the main interest of coordination chemistry is the coordinate bond, it is the metal–ligand vibrations that have held the interest of inorganic vibrational spectroscopists. It is difficult, however, to make unequivocal assignments of metal–ligand vibrations since the interpretation of the low-frequency spectrum is complicated by the appearance of ligand vibrations as well as lattice vibrations in the case of solid-state spectra. Conventional methods that have been used to assign metal–ligand vibrations are (1) Comparison of spectra between a free ligand and its metal complex; the metal– ligand vibration should be absent in the spectrum of the free ligand. This method often fails to give a clear-cut assignment because some ligand vibrations activated by complex formation may appear in the same region as the metal–ligand vibrations.

80

THEORY OF NORMAL VIBRATIONS

(2) The metal–ligand vibration should be metal sensitive and be shifted by changing the metal or its oxidation state. This method is applicable only when a series of metal complexes have exactly the same structure, where only the central metal is different. Also, it does not provide definitive assignments since some ligand vibrations (such as chelate ring deformations) are also metal-sensitive. (3) The metal–ligand stretching vibration should appear in the same frequency region if the metal is the same and the ligands are similar. For example, the n(Zn-N) (n: stretching) of Zn(II) pyridine complexes are expected to be similar to those of Zn(II) a-picoline complexes. This method is applicable only when the metal–ligand vibration is known for one parent compound. (4) The metal–ligand vibration exhibits an isotope shift if the ligand is isotopically substituted. For example, the n(Ni-N) of [Ni(NH3)6]Cl2 at 334 cm1 is shifted to 318 cm1 on deuteration of the ammonia ligands. The observed shift (16 cm1) is in good agreement with that predicted theoretically for this mode. This method was used to assign the metal–ligand vibrations of chelate compounds such as oxamido (14N/15N) and acetylacetonato(l6O/l8O) complexes. However, isotopic substitution of the a-atom (atom directly bonded to the metal) causes shifts of not only metal–ligand vibrations but also of ligand vibrations involving the motion of the a-atom. Thus, this method alone cannot provide an unequivocal assignment of the metal–ligand vibration. (5) The frequency of a metal–ligand vibration may be predicted if the metal– ligand stretching and other force constants are known a priori. At present, this method is not practical since only a limited amount of information is available on the force constants of coordination compounds. It is obvious that none of the above methods is perfect in assigning metal–ligand vibrations. Furthermore, these methods encounter more difficulties as the structure of the complex (and hence the spectrum) becomes more complicated. Fortunately, the “metal isotope technique” which was developed in 1969 may be used to obtain reliable metal–ligand assignments [89]. Isotope pairs such as (H/D) and (16O/18O) had been used routinely by many spectroscopists. However, isotopic pairs of heavy metals such as (58Ni/62Ni) and (104Pd/110Pd) were not employed until 1969 when the first report on the assignments of the NiP vibrations of trans-Ni(PEt3)2X2 (X ¼ Cl and Br) was made. The delay in their use was probably due to two reasons: (1) It was thought that the magnitude of isotope shifts arising from metal isotope substitution might be too small to be of practical value. (2) Pure metal isotopes were too expensive to use routinely in the laboratory. Nakamoto and coworkers [87,88] have shown, however, that the magnitudes of metal isotope shifts are generally of the order of 2–10 cm1 for stretching modes and 0–2 cm1 for bending modes, and that the experimental error in measuring the frequency could be as small as 0.2 cm1 if proper precautions

METAL–ISOTOPE SPECTROSCOPY

81

are taken. They have also shown that this technique is financially feasible if the compounds are prepared on a milligram scale. Normally, the vibrational spectrum of a compound can be obtained with a sample less than 10 mg. Table 1.14 lists metal isotopes that are useful for vibrational studies of coordination compounds.

TABLE 1.14. Some Stable Metal Isotopesa Element

Inventory Form

Magnesium

Oxide

Silicon

Oxide

Titanium

Oxide

Chromium

Oxide

Iron

Oxide

Nickel

Copper

Metal Oxide Oxide Oxide

Zinc

Oxide

Germanium

Oxide

Zirconium

Oxide

Molybdenum

Metal/oxide

Ruthenium

Metal

Palladium

Metal

Tin

Oxide

Barium

Carbonate

a

Available at Oak Ridge National Laboratory.

Isotope

Natural Abundance (%)

Mg-24 Mg-25 Mg-26 Si-28 Si-30 Ti-46 Ti-48 Cr-50 Cr-52 Cr-53 Fe-54 Fe-56 Fe-57 Ni-58 Ni-60 Ni-62 Cu-63 Cu-65 Zn-64 Zn-66 Zn-68 Ge-72 Ge-74 Zr-90 Zr-94 Mo-92 Mo-95 Mo-97 Mo-100 Ru-99 Ru-102 Pd-104 Pd-108 Sn-116 Sn-120 Sn-124 Ba-135 Ba-138

78.99 10.00 11.01 92.23 3.10 8.0 73.8 4.345 83.79 9.50 5.9 91.72 2.1 68.27 26.1 3.59 69.17 30.83 48.6 27.9 18.8 27.4 36.5 51.45 17.38 14.84 15.92 9.55 9.63 12.7 31.6 11.4 26.46 14.53 32.59 5.79 6.593 71.70

Purity (%) 99.9 94 97 >99.8 >94 70–96 >99 >95 >99.7 >96 >96 >99.9 86–93 >99.9 >99 >96 >99.8 >99.6 >99.8 >98 97–99 >97 >98 97–99 >98 >97 >96 >92 >97 >98 >99 >95 >98 >95 >98 >94 >93 >99

82

THEORY OF NORMAL VIBRATIONS

It should be noted that the central atom of a highly symmetrical molecule (Td, Oh, etc.) does not move during the totally symmetric vibration. Thus, no metal–isotope shifts are expected in these cases [90]. When the central atom is coordinated by several different donor atoms, multiple isotope labeling is necessary to distinguish different coordinate bond-stretching vibrations. For example, complete assignments of bis (glycino)nickel(II) require 14N=15N and/or I6O=18O isotope shift data as well as 58 Ni=62Ni isotope shift data [91]. This book contains many other applications of metal–isotope techniques. These results show that metal–isotope data are indispensable not only in assigning the metal–ligand vibrations but also in refining metal–ligand stretching force constants in normal coordinate analysis [88]. The presence of vibrational coupling between metal–ligand and other vibrations can also be detected by combining metal–isotope data with normal coordinate calculations (Sec. 1.18) since both experimental and theoretical isotope shift values would be smaller when such couplings occur. The metal–isotope techniques become more important as the molecules become larger and complex. In biological molecules such as heme proteins, structural and bonding information about the active site (iron porphyrin) can be obtained through definitive assignments of coordinate bond-stretching vibrations around the iron atom. Using resonance Raman techniques (Sec. 1.22), it is possible to observe iron porphyrin and iron-axial ligand vibrations without interference from peptide chain vibrations. Thus, these vibrations can be assigned by comparing resonance Raman spectra of a natural heme protein with that of a 54Fe-reconstituted heme protein. Chapter 5 (in Part B) includes several examples of applications of metal-isotope techniques to bioinorganic compounds [hemoglobin (54Fe, 57Fe), oxy-hemocyanin (63Cu, 65Cu), etc.]. An example of a multiple labeling is seen in the case of cytochrome P450cam having the axial Fe–S linkage in addition to four Fe–N (porphyrin) bonds; Champion et al. [92] were able to assign its Fe–S stretching vibration (351 cm1) by combining 32S–34S and 54Fe–56Fe isotope shift data. 1.18. GROUP FREQUENCIES AND BAND ASSIGNMENTS From observation of the infrared spectra of a number of compounds having a common group of atoms, it is found that, regardless of the rest of the molecule, this common group absorbs over a narrow range of frequencies, called the group frequency. For example, the methyl group exhibits the antisymmetric stretching [na(CH3)], symmetric stretching [ns(CH3)], degenerate bending [dd(CH3)], symmetric bending [ds(CH3)], and rocking [rs(CH3)] vibrations in the ranges 3050– 2950, 2970–2860, 1480–1410, 1385–1250, and 1200–800 cm1, respectively. Figure 1.25 illustrates their vibrational modes together with the observed frequencies for CH3C1, The methylene group vibrations are also well known; the antisymmetric stretching [na(CH2)], symmetric stretching [ns(CH2)], scissoring [d(CH2)], wagging [rw(CH2)], twisting [rt,(CH2)], and rocking [rr(CH2)] vibrations appear

GROUP FREQUENCIES AND BAND ASSIGNMENTS

83

Fig. 1.25. Approximate normal modes of a CH3X-type molecule (frequencies are given for X¼Cl). The subscripts s, d, and r denote symmetric, degenerate, and rocking vibrations, respectively. Only the displacements of the H or X atoms are shown.

in the regions 3100–3000, 3060–2980, 1500–1350, 1400–1100, 1260–1030, and 1180–750 cm1, respectively. Figure 1.26 illustrates the normal modes of these vibrations together with the observed frequencies for CH2C12. The normal modes illustrated for CH3C1 and CH2C12 are applicable to vibrational assignments of coordination and organometallic compounds, which will be discussed in Chapters 3 and 4, respectively. Group frequencies have been found for a number of organic and inorganic groups, and they have been summarized as group frequency charts [39,40] which are highly useful in identifying the atomic groups from observed spectra. Group frequency charts for inorganic compounds are given in Appendix VIII as well as in Figs. 2.55 and 2.61. The concept of group frequency rests on the assumption that the vibrations of a particular group are relatively independent of those of the rest of the molecule. As stated in Sec. 1.4, however, all the nuclei of the molecule perform their harmonic oscillations in a normal vibration. Thus, an isolated vibration, which the group frequency would have to be, cannot be expected in polyatomic molecules. If, however, a group includes relatively light atoms such as hydrogen (OH, NH, NH2, CH, CH2, CH3, etc.) or relatively heavy atoms such as the halogens (CC1, CBr, CI, etc.), as compared to other atoms in the molecule, the idea of an isolated vibration may be justified, since the amplitudes (or velocities) of the harmonic oscillation of these atoms are relatively larger or smaller than those of the other atoms in the same molecule. Vibrations of groups having multiple bonds (CC, CN, C¼C, C¼N, C¼O, etc.) may

84

THEORY OF NORMAL VIBRATIONS

Fig. 1.26. Approximate normal modes of a CH2X2-type molecule (frequencies are given for X¼Cl). Only the displacements of the H or X atoms are shown.

also be relatively independent of the rest of the molecule if the groups do not belong to a conjugated system. If atoms of similar mass are connected by bonds of similar strength (force constant), the amplitude of oscillation is similar for each atom of the whole system. Therefore, it is not possible to isolate the group frequencies in a system like the following:

A similar situation may occur in a system in which resonance effects average out the single and multiple bonds by conjugation. Examples of this effect are seen for

GROUP FREQUENCIES AND BAND ASSIGNMENTS

85

the metal chelate compounds discussed in Chapter 3. When the group frequency approximation is permissible, the mode of vibration corresponding to this frequency can be inferred empirically from the band assignments obtained theoretically for simple molecules. If coupling between various group vibrations is serious, it is necessary to make a theoretical analysis for each individual compound, using a method like the following one. As stated in Sec. 1.4, the generalized coordinates are related to the normal coordinates by qk ¼

X

Bki Qi

ð1:40Þ

i

In matrix form, this is written as q ¼ Bq Q

ð1:166Þ

It can be shown [3] that the internal coordinates are also related to the normal coordinates by R ¼ LQ

ð1:167Þ

This is written more explicitly as R1 ¼ l11 Q1 þ l12 Q2 þ þ l1N QN R2 ¼ l21 Q1 þ l22 Q2 þ þ l2N QN .. .

.. .

Ri ¼ li1 Q1 þ li2 Q2 þ þ liN QN

ð1:168Þ

In a normal vibration in which the normal coordinate QN changes with frequency nN, all the internal coordinates, R1, R2, . . ., Ri, change with the same frequency. The amplitude of oscillation is, however, different for each internal coordinate. The relative ratio of the amplitudes of the internal coordinates in a normal vibration associated with QN is given by l1N : I2N : : IiN

ð1:169Þ

If one of these elements is relatively large compared to the others, the normal vibration is said to be predominantly due to the vibration caused by the change of this coordinate.

86

THEORY OF NORMAL VIBRATIONS

The ratio of l’s given by Eq. 1.169 can be obtained as a column matrix (or eigenvector) lN, which satisfies the relation [3] GFlN ¼ lN kN

ð1:170Þ

It consists of i elements, l1N ; l2N ; . . . ; liN , where i is the number of internal coordinates, and can be calculated if the G and F matrices are known. An assembly by columns of the l elements obtained for each l gives the relation GFL ¼ LL

ð1:171Þ

where L is a diagonal matrix whose elements consist of l values. As an example, calculate the L matrix of the H2O molecule, using the results obtained in Sec. 1.15. The G and F matrices for the A1 species are as follows: # # " " 8:32300 0:35638 1:03840 0:08896 ; F¼ G¼ 0:35638 0:70779 0:08896 2:32370 with l1 ¼ 8.61475 and l2 ¼ 1.60914. The GF product becomes #

" GF ¼

8:61090 0:08771

0:30710 1:61299

The L matrix can be calculated from Eq. 1.171: #"

" 8:61090 0:08771

0:30710 1:61299

# l11 l21

l12 l22

" ¼

#

#" l11 l21

l12 l22

8:61475 0

0 1:60914

However, this equation gives only the ratios l11: l21 and l12: l22 To determine their values, it is necessary to use the following normalization condition: L~ L¼G

ð1:172Þ30

Then the final result is "

# l11 l21

l12 l22

" ¼

1:01683 0:01274

0:06686 1:52432

#

*This equation can be derived as follows. According to Eq. 1.113, 2T ¼ ~ _ 1 R. _ On the other hand, RG _ ¼ LQ_ and ~R_ ¼ ~Q_ ~L. Thus 2T ¼ ~Q_ ~LG 1 L Q. _ Comparing this with 2T ¼ ~Q_ EQ_ (matrix Eq. 1.167 gives R ~ G–1 L ¼ E or L~L ¼ G. form of Eq. 1.41), we obtain L

GROUP FREQUENCIES AND BAND ASSIGNMENTS

87

This result indicates that, in the normal vibration Q1, the relative ratio of amplitudes of two internal coordinates, R1 (symmetric OH stretching) and R2 (HOH bending), is 1.0168: 0.0127. Therefore, this vibration (3824 cm1) is assigned to an almost pure OH stretching mode. The relative ratio of amplitudes for the Q2 vibration is 0.0669: 1.5243. Thus, this vibration is assigned to an almost pure HOH bending mode. In other cases, the l values do not provide the band assignments that are expected empirically. This occurs because the dimension of l for a stretching coordinate is different from that for a bending coordinate. The potential energy due to a normal vibration, QN is written as X 1 VðQN Þ ¼ Q2N Fij liN ljN 2 ij

ð1:173Þ

Individual FijliNljN terms in the summation represent the potential energy distribution (PED) in QN, which gives a better measure for making band assignments [93]. In general, the value of FijliNljN is large when i ¼ j. Therefore, the Fii l2iN terms are most important in determining the distribution of the potential energy. Thus, the ratios of the Fii l2iN terms provide a measure of the relative contribution of each internal coordinate Ri to the normal coordinate QN. If any Fii l2iN term is exceedingly large compared with the others, the vibration is assigned to the mode associated with Ri. If Fii l2iN and Fjj l2jN are relatively large compared with the others, the vibration is assigned to a mode associated with both Ri and Rj (coupled vibration). As an example, let us calculate the potential energy distribution for the H2O molecule. Using the F and L matrices obtained previously, we find that the ~ LFL matrix is calculated to be 2 2 3 l11 F11 þ l221 F22 þ 2l21 l11 F12 0 6 8:60551 7 0:00011 0:00923 6 2 7 2 4 l12 F11 þ l22 F22 þ 2l12 l22 F12 5 0 0:03721 1:644459 0:07264 Then, the potential energy distribution in each normal vibration ðFii l2iN Þ is given by l1

"

l2

R1 8:60551 0:03721 R2 0:00011 1:64459

#

*According to Eq. 1.111, the potential energy is written as 2V ¼ RFR. ~ Using Eq. 1.167, we can write this as ~ LFLQ. ~ 2V ¼ Q On the other hand, Eq. 1.42 can be written as 2V ¼ QLQ. A comparison of these two ~ expressions gives L ¼ LFL. If this is written for one normal vibration whose frequency is lN, we have X X ~lNi Fij ljN ¼ Fij liN ljN lN ¼ ij

ij

Then the potential energy due to this vibration is expressed by Eq. 1.173.

88

THEORY OF NORMAL VIBRATIONS

More conveniently, PED is expressed by calculating ðFii l2iN =SFii l2iN Þ 100 for each coordinate: "

l1

R1 99:99 R2

l2 2:21

#

0:01 97:79

In this case, the final results are the same whether the band assignments are based on the L matrix or on the potential energy distribution: Q1 is the symmetric OH stretching and Q2 is the HOH bending. In other cases, different results may be obtained, depending on which criterion is used for band assignments. In the example above, no serious coupling occurs between the OH stretching and HOH bending modes of H2O in the A1 species. This is because the vibrational frequencies of these two modes are far apart. In other cases, however, vibrational frequencies of two or more modes in the same symmetry species are relatively close to each other. Then, the liN values for these modes become comparable. A more rigorous method of determining the vibrational mode involves drawing the displacements of individual atoms in terms of rectangular coordinates. As in Eq. 1.167, the relationship between the rectangular and normal coordinates is given by X ¼ Lx Q

ð1:174Þ

The Lx matrix can be obtained from the relationship [94] ~ 1L Lx ¼ M 1 BG

(1.75) ð1:175Þ

The matrices on the right-hand sides have already been defined. Three-dimensional drawings of normal modes, such as those shown in Chapter 2, can be made from the Cartesian displacement calculations obtained above. However, hand plotting of these data is laborious and complicated. Use of computer plotting programs greatly facilitates this process [95]. 1.19. INTENSITY OF INFRARED ABSORPTION [16] The absorption of strictly monochromatic light (n) is expressed by the Lambert–Beer law: In ¼ I0;n e an pl

ð1:176Þ

*By combining Eqs. 1.118 and 1.174, we have R ¼ BX ¼ BLxQ. Since R ¼ LQ (Eq. 1.167), it follows ~_ X. _ In terms of internal coordinates, that LQ ¼ BLxQ or L ¼ BLx. The kinetic energy is written as 2T ¼ XM ~_ 1 R_ ¼ X ~_ B~G 1 BX. _ By comparing these two expressions, we have M ¼ B~G 1 B. it is written as 2T ¼ RG B G 1 BLx ¼ M 1 ~ Then we can write Lx ¼ M 1 MLx ¼ M 1 ~ B G 1 L.

INTENSITY OF INFRARED ABSORPTION

89

where In is the intensity of the light transmitted by a cell of length l containing a gas at pressure p, I0,n is the intensity of the incident light, and an is the absorption coefficient for unit pressure. The true integrated absorption coefficient A is defined by ! ð ð 1 I0;n A¼ an dn ¼ ln dn ð1:177Þ pl band In band where the integration is carried over the entire frequency region of a band. In practice, In and I0,v cannot be measured accurately, since no spectrophotometers have infinite resolving power. Therefore we measure instead the apparent intensity Tn: ð 0 Tn ¼ IðnÞgðn; n Þdn ð1:178Þ slit 0

where g(n, n ) is a function indicating the amount of light of frequency n when the spectrophotometer reading is set at n0 . Then the apparent integrated absorption coefficient B is defined by 1 B¼ pl

Ð 0 I0 ðnÞgðn; n Þdn 0 dn ln Ðslit 0 band slit IðnÞgðn; n Þdn

ð

ð1:179Þ

It can be shown that limpl ! 0 ðA BÞ ¼ 0

ð1:180Þ

if I0 and an are constant within the slit width used. (This condition is approximated by using a narrow slit.) In practice, we plot B/pl against pl, and extrapolate the curve to pl ! 0. To apply this method to gaseous molecules, it is necessary to broaden the vibrational–rotational bands by adding a high-pressure inert gas (pressure broadening). For liquids and solutions, p and a in the preceding equations are replaced by M (molar concentration) and e (molar absorption coefﬁcient), respectively. However, the extrapolation method just described is not applicable, since experimental errors in determining B values become too large at low concentration or at small cell length. The true integrated absorption coefficient of a liquid can be calculated if we assume that the shape of an absorption band is represented by the Lorentz equation and that the slit function is triangular [96]. Theoretically, the true integrated absorption coefficient AN of the Nth normal vibration is given by [3] " # qmy 2 np qmx 2 qmz 2 þ þ AN ¼ ð1:181Þ 3c qQN 0 qQN 0 qQN 0 where n is the number of molecules per cubic centimeter, and c is the velocity of light. As shown by Eq. 1.167, an internal coordinate Ri is related to a set of normal

90

THEORY OF NORMAL VIBRATIONS

coordinates by Ri ¼

X

ð1:182Þ

LiN QN

N

If the additivity of the bond dipole moment is assumed, it is possible to write ! ! X qm qm qRi ¼ qQN qRi qQN i ! ð1:183Þ X qm ¼ LiN qRi i Then Eq. 1.181 is written as " np AN ¼ 3c

X qm

np X ¼ 3c i

i

"

!2 x

qRi qmx qRi

þ

LiN !2 0

X qmy i

0

qmy þ qRi

!2 0

qRi

!2 þ

LiN

qmz þ qRi

X qm

0 !2 #

!2 # z

qRi

LiN 0

ð1:184Þ

ðLiN Þ2 0

This equation shows that the intensity of an infrared band depends on the values of the qm/qR terms as well as the L matrix elements. Equation 1.184 has been applied to relatively small molecules to calculate the qm/qR terms from the observed intensity and known LiN values [7]. However, the additivity of the bond dipole moment does not strictly hold, and the results obtained are often inconsistent and conflicting. Thus far, very few studies have been made on infrared intensities of large molecules because of these difficulties. As seen in Eq. 1.184, the IR band becomes stronger as the qm/qR term becomes larger if the LiN term does not vary significantly. In general, the more polar the bond, the larger the qm/qR term. Thus, the IR intensities of the stretching vibrations follow the general trends: nðOHÞ > nðNHÞ > nðCHÞ nðC ¼ OÞ > nðC ¼ NÞ > nðC ¼ CÞ It is also noted that the antisymmetric stretching vibration is always stronger than the symmetric stretching vibration because the qm/qR term is larger for the former than for the latter. This can be seen for functional groups such as These trends are entirely opposite to those found in Raman spectra (Sec. 1.21), and make both IR and Raman spectroscopy complementary. 1.20. DEPOLARIZATION OF RAMAN LINES As stated in Secs. 1.8 and 1.9, it is possible, by using group theory, to classify the normal vibration into various symmetry species. Experimentally, measurements of the

DEPOLARIZATION OF RAMAN LINES

91

Fig. 1.27. Experimental configuration for measuring depolarization ratios. The scrambler is placed after the analyzer because the monochromator gratings show different efficiencies for ? and | | directions.

infrared dichroism and polarization properties of Raman lines of an orientated crystal provide valuable information about the symmetry of normal vibrations (Sec. 1.31). Here, we consider the polarization properties of Raman lines in liquids and solutions in which molecules or ions take completely random orientations.* Suppose that we irradiate a molecule fixed at the origin of a space-fixed coordinate system with natural light from the positive-y direction, and observe the Raman scattering in the x direction as shown in Fig. 1.27. The incident light vector E may be resolved into two components, Ex and Ez, of equal magnitude (Ey ¼ 0). Both components give induced dipole moments, Px, Py, and Pz. However, only Py and Pz contribute to the scattering along the x axis, since an oscillating dipole cannot radiate in its own direction. Then, from Eq. 1.66, we have. Py ¼ ayx Ex þ ayz Ez Pz ¼ azx Ex þ azz Ez

ð1:185Þ ð1:186Þ

The intensity of the scattered light is proportional to the sum of squares of the individual aijEj terms. Thus, the ratio of the intensities in the y and z directions is rn ¼

Iy a2yx Ex2 þ a2yz Ez2 ¼ Iz a2zx Ex2 þ a2zz Ez2

ð1:187Þ

where rn is called the depolarization ratio for natural light (n). In a homogeneous liquid or gas, the molecules are randomly oriented, and we must consider the polarizability components averaged over all molecular *It is possible to obtain approximate depolarization ratios of fine powders where the molecules or ions take pseudorandom orientations (see Ref. 97).

92

THEORY OF NORMAL VIBRATIONS

(mean value) orientations. The results are expressed in terms of two quantities: a and g (anisotropy): 1 ¼ ðaxx þ ayy þ azz Þ a 3

g2 ¼

1h ðaxx ayy Þ2 þ ðayy azz Þ2 þ ðazz axx Þ2 þ 6ða2xy þ a2yz þ a2zx Þ 2

ð1:188Þ ð1:189Þ

These two quantities are invariant to any coordinate transformation. It can be shown [3] that the average values of the squares of aij are 1 ½45ð aÞ2 þ 4g2 45 1 ðaxy Þ2 ¼ ðayz Þ2 ¼ ðazx Þ2 ¼ g2 15 ðaxx Þ2 ¼ ðayy Þ2 ¼ ðazz Þ2 ¼

ð1:190Þ ð1:191Þ

Since Ex ¼ Ez ¼ E, Eq. 1.187 can be written as rn ¼

Iy 6g2 ¼ Iz 45ð aÞ2 þ 7g2

ð1:192Þ

The total intensity In is given by

1 2 2 ½45ð aÞ þ 13g E2 In ¼ Iy þ Iz ¼ const 45

ð1:193Þ

If the incident light is plane-polarized (e.g., a laser beam), with its electric vector in the z direction (Ex ¼ 0), Eq. 1.192 becomes rp ¼

Iy 3g2 ¼ Iz 45ð aÞ2 þ 4g2

ð1:194Þ

where rp is the depolarization ratio for polarized light (p). In this case, the total intensity is given by Ip ¼ Iy þ Iz ¼ const

1 ½45ð aÞ2 þ 7g2 E2 45

ð1:195Þ

The symmetry property of a normal vibration can be determined by measuring the depolarization ratio. From an inspection of character tables (Appendix I), it is obvious is nonzero only for totally symmetric vibrations. Then, Eq. 1.192 gives that a 0 rn < 67, and the Raman lines are said to be polarized. For all nontotally symmetric is zero, and rn ¼ 67. Then, the Raman lines are said to be depolarized. vibrations, a If the exciting line is plane polarized, these criteria must be changed according to

DEPOLARIZATION OF RAMAN LINES

93

Fig. 1.28. Raman spectra of CCI4 in two directions of polarization (488 nm excitation).

Eq. 1.194. Thus, 0 rp < 34 for totally symmetric vibrations, and rp ¼ 34 for nontotally symmetric vibrations. Figure 1.28 shows the Raman spectra of CCl4 (500–150 cm1) in two directions of polarization obtained with the 488 nm excitation. The three bands at 459, 314, and 218 cm1 give rp values of approximately 0.02, 0.75, and 0.75, respectively. Thus, it is concluded that the 459 cm1 band is polarized (A1), whereas the two bands at 314 (F2) and 218 (E) cm1 are depolarized. As stated in Sec. 1.6, the polarizability tensors are symmetric in normal Raman scattering. If the exciting frequency approaches that of an electronic absorption, some scattering tensors become antisymmetric,* and resonance Raman scattering can occur (Sec. 1.22). In this case, Eq. 1.194 must be written in a more general form [98] rp ¼

3gs þ 5ga 10go þ 4gs

ð1:196Þ

*A tensor is called antisymmetric if axx ¼ ayy ¼ azz ¼ 0 and axy ¼ ayx,ayz ¼ azy, and azx ¼ axz.

94

where

THEORY OF NORMAL VIBRATIONS

aÞ2 go ¼ 3ð h i 1 gs ¼ ðaxx ayy Þ2 þ ðayy azz Þ2 þ ðazz axx Þ2 3 i 1h þ ðaxy þ ayx Þ2 þ ðaxz þ azx Þ2 þ ðayz þ azy Þ2 2 h i 1 ga ¼ ðaxy ayx Þ2 þ ðaxz azx Þ2 þ ðayz azy Þ2 2

ð1:197Þ

If we define 3 g2s ¼ gs 2

and

3 g2as ¼ ga 2

ð1:198Þ

Eq. 1.196 can be written as rp ¼

3g2s þ 5g2as

ð1:199Þ

45ð aÞ2 þ 4g2s

In normal Raman scattering, g2s ¼ g2 and g2as ¼ 0. Then Eq. 1.199 is reduced to Eq. 1.194. The symmetry properties of resonance Raman lines can be predicted on the basis „ 0 and gas ¼ 0. Then, Eq. 20.15 of Eq. 1.199. For totally symmetric vibrations, a gives 0 rp < 34. Nontotally symmetric vibrations ð a ¼ 0Þ are classified into two types: those that have symmetric scattering tensors, and those that have antisymmetric scattering tensors. If the tensor is symmetric, gas ¼ 0 and gs „ 0. Then Eq. 1.199 gives rp ¼ 34 (depolarized). If the tensor is antisymmetric, gas „ 0 and gs ¼ 0. Then, Eq. 1.199 gives rp ¼ ¥ (inverse polarization). In the case of the D4h point group, the B1g and B2g representations belong to the former type, whereas the A2g representations belongs to the latter [99]. As will be shown in Sec. 1.22, Spiro and Strekas [98] observed for the first time A2g vibrations which exhibit rp values 3 4 < rp < ¥. For these vibrations, the term “anomalous polarization (ap)” is used instead of “inverse polarization (ip)” [100]. 1.21. INTENSITY OF RAMAN SCATTERING According to the quantum mechanical theory of light scattering, the intensity per unit solid angle of scattered light arising from a transition between states m and n is given by In

m

¼ constðn0 nmn Þ4

X rs

jðPrs Þmn j2

ð1:200Þ

INTENSITY OF RAMAN SCATTERING

95

Fig. 1.29. Energy-level diagram for Raman scattering.

where ðPrs Þmn

" # 1 X ðMr Þrn ðMs Þmr ðMr Þmr ðMs Þrn ¼ ðars Þmn E ¼ þ E h r nrm n0 nrn þ n0

ð1:201Þ

Here n0 is the frequency of the incident light: nrm, nrn, and nmn are the frequencies corresponding to the energy differences between subscripted states;Ð terms of the type (Ms)mr are the Cartesian components of transition moments such as C r ms Cm dt; and E is the electric vector of the incident light. It should be noted here that the states denoted by m, n, and r represent vibronic states cg ðx; QÞfgi ðQÞ; cg ðx; QÞfgj ðQÞ and ce ðx; QÞfev ðQÞ, respectively, where cg and ce are electronic ground- and excited-state wavefunctions, respectively, and fgi ; fgj , and fev vibrational functions. The symbols x and Q represent the electronic and nuclear coordinates, respectively. These notations are illustrated in Fig. 1.29. Finally, s and r denote x, y, and z components. Since the electric dipole operator acts only on the electronic wavefunctions, the (ars)mn term in Eq. 1.201 can be written in the form ðars Þmn ¼

ð 1 fgj ðars Þgg fgj dQ h

ð1:202Þ

where ðars Þgg ¼

X e

Ð

! Ð Ð Ð c g ms ce dt c e mr cg dt cg mr ce dt c e ms cg dt þ neg n0 neg þ n0

Here neg corresponds to the energy of a pure electronic transition between the ground and excited states.

96

THEORY OF NORMAL VIBRATIONS

To discuss the Raman scattering, we expand the (ars)gg term as a Taylor series with respect to the normal coordinate Q: " # qðars Þgg 0 ðars Þgg ¼ ðars Þgg þ Qþ ð1:203Þ qQ 0

Then, we write Eq. 1.202 as ðars Þmn

" # ð ð 1 1 qðars Þgg g g 0 fgj Qfgi dQ ¼ ðars Þgg fj fi dQ þ h h qQ

ð1:204Þ

0

The first term on the right-hand side is zero unless i ¼ j. This term is responsible for Rayleigh scattering. The second term determines the activity offundamental vibrations in Raman scattering; it vanishes for a harmonic oscillator unless j ¼ i 1 (Sec. 1.10). If we consider a Stokes transition, v ! v þ 1, Eq. 1.204 is written as [3] ðars Þv;v þ 1

" # sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 qðars Þgg ðv þ 1Þh ¼ h 8p2 mn qQ

ð1:205Þ

0

where m and n are the reduced mass and the Stokes frequency. Then Eq. 1.200 is written as " #2 E2 qðaps Þgg ðv þ 1Þh I ¼ constðn0 nÞ 2 8p2 mn h qQ 4

ð1:206Þ

0

In Sec. 1.20, we derived a classical equation for Raman intensity: !2 qa In ¼ const E2 qQ ( ) 1 ½45ð aÞ2 þ 13g2 E2 ¼ const 45

ð1:193Þ

By replacing the qa/qQ term of Eq. 1.206 with the term enclosed in braces in Eq. 1.193, we obtain ðv þ 1Þ E2 In ¼ constðn0 nÞ 8p2 mn h2 4

(

1 ½45ð aÞ2 þ 13g2 45

) ð1:207Þ

At room temperature, most of the scattering molecules are in the v ¼ 0 state, but some are in higher vibrational states. Using the Maxwell–Bolztmann distribution

INTENSITY OF RAMAN SCATTERING

97

law, we find that the fraction of molecules fv with vibrational quantum number v is given by e ½v þ ð1=2Þhn=kT fv ¼ P ½v þ ð1=2Þhn=kT ve

ð1:208Þ

P Then the total intensity is proportional to v fv ðv þ 1Þ, which is equal to (1 ehv/kT)1 (see Ref. 23). Hence we can rewrite Eq. 1.207 in the form In ¼ KI0

ðn0 nÞ4 ½45ð aÞ2 þ 13g2 mnð1 e hn=kT Þ

ð1:209Þ

Here I0 is the incident light intensity which is proportional to E2, and K summarizes all other constant terms. If the incident light is polarized, the form of Eq. 1.209 is slightly modified: Ir ¼ KI0

ðn0 nÞ4 ½45ð aÞ2 þ 7g2 mnð1 e hn=kT Þ

ð1:210Þ

As shown in Sec. 1.20, the degree of depolarization rp is rp ¼

3g2 45ð aÞ2 þ 4g2

or

g2 ¼

45ð aÞ2 rp 3 4rp

ð1:211Þ

Since rp ¼ 34 for nontotally symmetric vibrations, Eq. 1.211 holds only for totally symmetric vibrations. Then Eq. 1.210 is written as ! 1 þ rp ðn0 nÞ4 Ip ¼ K I0 ð a Þ2 mnð1 e hn=kT Þ 3 4rp 0

ð1:212Þ

In the case of a solution, the intensity is proportional to the molar concentration C. Then Eq. 1.212 is written as ! 1 þ rp Cðn0 nÞ4 Ip ¼ K I0 ð aÞ2 mnð1 e hn=kT Þ 3 4rp 00

ð1:213Þ

If we compare the intensities of totally symmetric vibrations (A1 mode) of two tetrahedral XY4-type molecules, the intensity ratio is given by I1 C1 ~n0 ~n1 ¼ I2 C2 ~n0 ~n2

!4

~n2 m2 ð1 e hc~n2 =kT Þð a1 Þ 2 ~n1 m1 ð1 e hc~n1 =kT Þð a2 Þ 2

ð1:214Þ

98

THEORY OF NORMAL VIBRATIONS

In this case, the rp term drops out, since g2 ¼ 0 for isotropic molecules such as tetrahedral XY4 and octahedral XY6 types. By using CCI4 as the standard, it is possible to determine the relative value of the qa/qQ term, which provides information about the degree of covalency and the bond order [23]. As seen above, the intensity of Raman scattering increases as the (qa/qQ)0 term becomes larger. The following general rules may reflect the trends in the (qa/qQ)0 values: (1) stretching vibrations produce stronger Raman bands than do bending vibrations; (2) symmetric vibrations produce stronger Raman bands than do antisymmetric vibrations, and totally symmetric “breathing” vibrations produce the most intense Raman bands; (3) stretching vibrations of covalent bonds produce Raman bands stronger than those of ionic bonds (among the covalent bonds, Raman intensities increase as the bond order increases; thus, the ratio of relative intensities of the CC, C¼C, and CC stretching vibrations is approximately 3:2:1); and (4) bonds involving heavier atoms produce stronger Raman bands than do those involving lighter atoms. For example, the n(SS) is stronger than the n(CC). Coordination compounds containing heavy metals and sulfur ligands are ideal for Raman studies. 1.22. PRINCIPLE OF RESONANCE RAMAN SPECTROSCOPY In normal Raman spectroscopy, the exciting frequency lies in the region where the compound has no electronic absorption band. In resonance Raman spectroscopy, the exciting frequency falls within the electronic band (Sec. 1.2). In the gaseous phase, this tends to cause resonance fluorescence since the rotational–vibrational levels are discrete. In the liquid and solid states, however, these levels are no longer discrete because of molecular collisions and/or intermolecular interactions. If such a broad vibronic bands is excited, it tends to give resonance Raman rather than resonance fluorescence spectra [101,102]. Resonance Raman spectroscopy is particularly suited to the study of biological macromolecules such as heme proteins because only a dilute solution (biological condition) is needed to observe the spectrum and only vibrations localized within the chromophoric group are enhanced when the exciting frequency approaches that of the relevant chromophore. This selectivity is highly important in studying the theoretical relationship between the electronic transition and the vibrations to be resonanceenhanced. The origin of resonance Raman enhancement is explained in terms of Eq. 1.201. In normal Raman spectroscopy, n0 is chosen in the region that is far from the electronic absorption. Then, nrm n0, and ars is independent of the exciting frequency n0. In resonance Raman spectroscopy, the denominator, nrm n0, becomes very small as n0 approaches nrm. Thus, the first term in the square brackets of Eq. 1.201 dominates all other terms and results in striking enhancement of Raman lines. However, Eq. 1.201 cannot account for the selectivity of resonance Raman enhancement since it is not specific about the states of the molecule. Albrecht [103] derived a more specific equation for the initial and final states of resonance Raman scattering by

PRINCIPLE OF RESONANCE RAMAN SPECTROSCOPY

99

introducing the Herzberg–Teller expansion of electronic wave functions into the Kramers–Heisenberg dispersion formula. The results are as follows: ðars Þgi;gj ¼ A þ B þ C

"

#

ð1:215Þ

0 X ðg0 jRs je0 Þðe0 jRr jg0 Þ X þ ðnonresonance termÞ hijvihvjji ð1:216Þ Eev Egi E0 e„g v # (" 0 0 XX X X ðg0 jRs je0 Þðe0 jha js0 Þðs0 jRr jg0 Þ þ ðnonresonance termÞ B¼ Eev Egi E0 e„g v s„e a

A¼

" # hi=vihv=Qa jji ðg0 jRs js0 Þðs0 jha je0 Þðe0 jRr jg0 Þ þ þ ðnonresonance termÞ Ee0 Es0 Eev Egi E0 hijQa jvihv=ji Ee0 Es0 C¼

0 0 X X XX

e„g t„g

v

) ð1:217Þ ("

a

ðe0 jRr jg0 Þðg0 jha jt0 Þðt0 jRs je0 Þ þ ðnonresonance termÞ Eev Egi E0

#

" # hijvihvjQa jji ðe0 jRr jt0 Þðt0 jha jg0 Þðg0 jRs je0 Þ þ þ ðnonresonance termÞ Eg0 Et0 Eev Egi E0 hijQa jvihvjji Eg0 Et0

) ð1:218Þ

The notations g, i, j, e, and v were explained in Sec. 1.21. Other notations are as follows: s, another excited electronic state; ha, the vibronic coupling operator qH=qQa ; H, and Qa, the electronic Hamiltonian and the ath normal coordinate of the electronic ground state, respectively; Egi and Eev, the energies of states gi and ev, respectively; |g0), |e0), and |s0), the electronic wavefunctions for the equilibrium nuclear positions of the ground and excited states; Ee0 and Es0 , the corresponding energies of the electronic states, e0 and s0, respectively; and E0, the energy of the exciting light. The nonresonance terms are similar to the preceding terms except that the denominator is (Eev Egj þ E0) instead of (Eev Egj E0) and that Rs and Rp (Cartesian components of the transition moment) in the numerator are interchanged. These terms can be neglected under the strict resonance condition since the resonance terms become very large. The C term is usually neglected because its components are denominated by Eg0 Et0, where t refers to an excited state that is much higher in energy than the first excited state. These notations are shown on the right-hand side of Fig. 1.29. In more rigorous expressions, the damping factor iG must be added to the

100

THEORY OF NORMAL VIBRATIONS

Fig. 1.30. Shift of equilibrium position caused by totally symmetric vibration.

denominators such as Eev Egj E0 so that the resonance term does not become infinity under rigorous resonance conditions [103]. For the A term to be nonzero, the g $ e electronic transition must be allowed, and the product of the integrals hi | vihv | ji (Franck–Condon factor) must be nonzero. The latter condition is satisfied if the equilibrium position is shifted by a transition (D „ 0 as shown in Fig. 1.30). Totally symmetric vibrations tend to be resonance-enhanced via the A term since they tend to shift the equilibrium position upon electronic excitation. If the equilibrium position is not shifted by the g $ e transition, the A term becomes zero since either one of the integrals, (i | v) or (v | j), becomes zero. This situation tends to occur for nontotally symmetric vibrations. In contrast to the A term, the B term resonance requires at least one more electronic excited state(s), which must be mixed with the e state via normal vibration, Qa. Namely, the integral (e0 | ha | s0) must be nonzero. The B-term resonance is significant only when these two excited states are closely located so that the denominator, Ee0 Es0 , is small. Other requirements are that the g $ e and g $ s transitions should be allowed and that the integrals hi | vi and hv | Qa | ji should not be zero. The B term can cause resonance enhancement of nontotally symmetric as well as totally symmetric vibrations, although it is mainly responsible for resonance enhancement of the former. As seen in Eqs. 1.216 and 1.217, both A and B terms involve summation over v, which is the vibrational quantum number at the electronic excited state. Since the harmonic oscillator selection rule (Dv ¼ 1) does not hold for a large v, overtones and combination bands may appear under resonance conditions. In fact, series of these nonfundamental vibrations have been observed in the case of A-term resonance (Sec. 1.23).

RESONANCE RAMAN SPECTRA

101

In Sec. 1.20, we have shown that the depolarization ratio for totally symmetric vibrations is in the range 0 rp < 34. For the A term to be nonzero, however, the term (g0 | Rs | e0)(e0 | Rr | g0) must be nonzero. Since g0 is totally symmetric, this condition is realized only when both c(Rs)c(e0) and c(Rr)c(e0) belong to the totally symmetric species. Thus, c(Rs) ¼ c(Rr), and only one of the diagonal terms of the polarizability tensor (axx, ayy, or azz) becomes nonzero. Then, it is readily seen from Eq. 1.199 that rp ¼ 13. This rule has been confirmed by many observations. If the electronic excited state is degenerate, two of the diagonal terms must be nonzero. In this case, rp ¼ 18. These rules hold for any point group with at least one C3 or higher axis, with the exception of the cubic groups in which the Cartesian coordinates transform as a single irreducible degenerate representation. 1.23. RESONANCE RAMAN SPECTRA Because of the advantages mentioned in the preceding section, resonance Raman (RR) spectroscopy has been applied to vibrational studies of a number of inorganic as well as organic compounds. It is currently possible to cover the whole range of electronic transitions continuously by using excitation lines from a variety of lasers and Raman shifters [26]. In particular, the availability of excitation lines in the UV region has made it possible to carry out UV resonance Raman (UVRR) spectroscopy [104]. In RR spectroscopy, the excitation line is chosen inside the electronic absorption band. This condition may cause thermal decomposition of the sample by local heating. To minimize thermal decomposition, several techniques have been developed. These include the rotating sample technique, the rotating (or oscillating) laser beam technique, and their combinations with low-temperature techniques [21,26]. It is always desirable to keep low laser power so that thermal decomposition is minimal. This will also minimize photodecomposition, which occurs depending on laser lines in some compounds. An example of the A-term resonance is given by I2, which has an absorption band near 500 nm and its fundamental at 210 cm1. Figure 1.31 shows the RR spectrum of I2 in solution obtained by Kiefer and Bernstein [105] (excitation at 514.5 nm). It shows a series of overtones up to the seventeenth. The vibrational energy of an anharmonic oscillator including the cubic term is expressed as 1 1 2 1 3 Ev ¼ hcwe v þ þ hcye we v þ hcce we v þ 2 2 2 The anharmonicity constants, cewe and yewe can be determined by plotting the ~nðobsÞ=v against the vibrational quantum number v, as shown in Fig. 1.32. The straight line gives cewe and the deviation from the straight line (at higher v) gives yewe The results (CCl4 solution) in cm1 are we ¼ 212:59 0:08;

ce we ¼ 0:62 0:03;

ye we ¼ 0:005 0:002

102

THEORY OF NORMAL VIBRATIONS

Fig. 1.31. Resonance Raman spectra of I2 in CCl4, CHCl3, and CS2 (514.5 nm excitation) [105].

As stated in the preceding section, the depolarization ratio (rp) is expected to be closed to 13 for totally symmetric vibrations. However, this value increases as the change in vibrational quantum number (Dv) increases. For example, it is 0.48 for the tenth overtone of I2 in CCl4 [100]. Carbontetrachloride 210

ν∼ υ

205 cm-1

200 1

5

10

υ

Fig. 1.32. Plot of v~=v vs. n for I2 in CCl4 [105].

15

RESONANCE RAMAN SPECTRA

103

Fig. 1.33. Resonance Raman spectrum of solid TiI4 (514.5 nm excitation) [106].

The second example is given by TiI4, which has the absorption maximum near 514 nm. Figure 1.33 shows the RR spectrum of solid TiI4 obtained by Clark and Mitchell [106] with 514.5 nm excitation. In this case, a series of overtones up to the twelfth has been observed. Using these data, they obtained the following constants (cm1): we ¼ 161:0 0:2

and X11 ¼ 0:11 0:03

Here, X11 is the anharmonicity constant corresponding to cewe of a diatomic molecule (Sec. 1.3). Theoretical treatments of Raman intensities of these overtone series under rigorous resonance conditions have been made by Nafie et al. [107]. There are many other examples of RR spectra of small molecules and ions that exhibit A-term resonance. Section 2.6 includes references on RR spectra of the CrO2 4 and MnO4 ions. Kiefer [101] reviewed RR studies of small molecules and ions. Hirakawa and Tsuboi [108] noted that among the totally symmetric modes, the mode that leads to the excited-state configuration is most strongly resonance enhanced. For example, the NH3 molecule is pyramidal in the ground state and planar in the excited state (216.8 nm above the ground state). Then, the symmetric bending mode near 950 cm1 is enhanced 10 times more than the symmetric stretching mode near 3300 cm1 when the excitation wavelength is changed from 514.5 to 351.1 nm. A typical example of the B-term resonance is given by metalloporphyrins and heme proteins. As shown in Fig. 1.34, Ni(OEP) (OEP:octaethylporphyrin) exhibits two electronic transitions referred to as the Q0 (or a) and B (or Soret) bands along with a vibronic side band (Q1 or b) in the 350–600 nm region. According to MO (molecular orbital) calculations on the porphyrin core of D4h, symmetry, Q0 and B transitions

104

THEORY OF NORMAL VIBRATIONS

Fig. 1.34. Structure, energy-level diagram, and electronic spectrum of Ni(OEP) [109].

result from strong interaction between the a1u(p) ! eg (p*) and a2u(p) ! eg (p*) transitions which have similar energies and the same excited-state symmetry (Eu). The transition dipoles add up for the strong B transition and nearly cancel out for the weak Q0 transition. This is an ideal situation for B-term resonance. According to Eq. 1.217, any normal modes which give nonzero values for the integral (e0 | ha | s0) are enhanced via the B term. These vibrations must belong to one of the symmetry species given by Eu Eu, that is, A1g þ B1g þ B2g þ A2g.* Figure 1.35 shows the RR spectra of Ni(OEP) with excitation near the B, Q1 and Q0 transitions as obtained by Spiro and coworkers [109]. As discussed in Sec. 1.20, polarization properties are expected to be A1g (p, polarized), B1g and B2g (dp, depolarized), and A2g (ap, anomalous polarization). These polarization properties, together with normal coordinate analysis, were used to make complete vibrational assignments of Ni(OEP) [110]. It is seen that the spectra obtained by excitation near the Q0 band (bottom traces) are dominated by the dp bands (B1g and B2g species), while those obtained by excitation between the Q0 and Q1 bands (middle traces) are dominated by the ap bands (A2g species). The spectra obtained by excitation near the B band (top traces) are dominated by the p bands (A1g species). In this

*For symmetry species of the direct products, see Appendix IV. Also note that the A1g vibrations are not effective in vibronic mixing.

RESONANCE RAMAN SPECTRA

105

Fig. 1.35. Resonance Raman spectra of Ni(OEP) obtained by excitation near the B, Q1, and Q0 bands [110].

case, the A-term resonance is much more important than the B-term resonance because of the very strong absorption strength of the B band. Mizutani and coworkers [110a] analyzed the resonance Raman spectra of Ni(OEP) obtained at the d–d excited state. As will be shown in Sec. 2.8, anomalously polarized bands have also been observed in the RR spectra of the IrCl26 ion, although their origin is different from the vibronic coupling mentioned above [111,112]. The majority of compounds studied thus far exhibit the A-term rather than the B-term resonance. A more complete study of resonance Raman spectra involves the observation of excitation profiles (Raman intensity plotted as a function of the excitation frequency for each mode), and the simulation of observed excitation profiles based on theoretical treatments of resonance Raman scattering [113].

106

THEORY OF NORMAL VIBRATIONS

1.24. THEORETICAL CALCULATION OF VIBRATIONAL FREQUENCIES The procedure for calculating harmonic vibrational frequencies and force constants by GF matrix method has been described in Sec. 1.12. In this method, both G (kinetic energy) and F (potential energy) matrices are expressed in terms of internal coordinates (R) such as increments of bond distances and bond angles. Then, the kinetic (T) and potential (V) energies are written as:

~_ 1 R_ 2T ¼ RG

ð1:113Þ

RFR 2V ¼ ~

ð1:111Þ

jGF Elj ¼ 0

ð1:116Þ

and the secular equation

is solved to obtain vibrational frequencies. Here, E is a unit matrix, and l is related to the wavenumber by the relation given by Eq. 1.156. The G matrix elements can be calculated from the known bond distances and angles, and the F matrix elements (force constants) are refined until the differences between the observed and calculated frequencies are minimized. Force constants thus obtained can be transferred to similar molecules to predict their vibrational frequencies. On the other hand, quantum mechanical methods utilize an entirely different approach to this problem. The principle of the method is to express the total electronic energy (E) of a molecule as a function of nuclear displacements near the equilibrium positions and to calculate its second derivatives with respect to nuclear displacements, namely, vibrational force constants. In the past, ab initio calculations of force constants were limited to small molecules [114] such as H2O and C2H2 because computational time required for large molecules was prohibitive. This problem has been solved largely by more recent progress in density functional theory (DFT) [115] coupled with commercially available computer software. Currently, the DFT method is used almost routinely to calculate vibrational parameters of a variety of inorganic, organic, coordination, and organometallic compounds. The advent of high-speed personal computers further facilitated this trend. Using a commercially available computer program such as “Gaussian 03,” DFT calculations of the H2O molecule proceed in the following order: First, molecular parameters (masses of constituent atoms, approximate bond distances and angles) are used as input data. As an example, let us assume that the OH distance is 0.96 A and the HOH angle is 109.5 . Then, the program calculates the cartesian coordinates of the three atoms by placing the origin at the center of gravity (CG) of the molecule. It also determines the point group of the molecule as C2v on the basis of the given geometry. Since the total electronic enegy (E) of the H2O molecule is a function of the nine Cartesian coordinates (xi, i where i ¼ 1, 2, . . ., 9), the equilibrium configuration that renders all the nine first derivatives (qE/qxi) zero can be determined (energy optimization). Accuracy of the results depends on the level of the approximation method

THEORETICAL CALCULATION OF VIBRATIONAL FREQUENCIES

107

TABLE 1.15. Calculated Cartesian Coordinates of H2O ˚) Coordinate (A Center Number 1 2 3

Atomic Number

X

8 1 1

0.0000 0.0000 0.0000

Y

Z

0.0000 0.8010 0.8010

0.1110 0.4440 0.4440

chosen in calculating E. Using 6-31G* basis set with B3LYP functional [116], for example, the optimal geometry was found to be: OH distance ¼ 0.9745 A and HOH angle ¼ 110.5670 These values were used to recalculate the Cartesian coordinates listed in Table 1.15 The next step is to calculate the second derivatives near the equilibrium positions [(qE2/qxi qxj)0,i,j ¼ 19], namely, the force constants in terms of Cartesian coordinates (fi,j), using analytical methods [116a]. The force constants thus obtained are converted to mass-weighted Cartesian coordinates (XM,i) by xM;1 ¼

pﬃﬃﬃﬃﬃﬃ m 1 x1 ;

pﬃﬃﬃﬃﬃﬃ m 1 x2 ; . . . ! q2 V qxMi qxMj

xM;2 ¼

fi;j fM;i;j ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ mi mj

ð1:219Þ ð1:220Þ

0

These fM,i,j terms are the elements of the force constant (Hessian) matrix. Using matrix expression, the relationship between mass-weighted Cartesian (XM) and Cartesian (X) coordinates is written as follows: XM ¼ M1=2 X

ð1:221Þ

pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ Here, M1/2 is a diagonal matrix whose elements are m1 ; . . . m9 and X is a column matrix, [x1,. . .,x9]. The product, M1/2X, gives a column matrix, XM, pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ ½ m1 x1 ; . . . ; m9 x9 . Equation (1.221) is rewritten as X ¼ M ð1=2Þ XM

ð1:222Þ

The relationship between the internal coordinate R and X (Eqs. 1.118 and 1.119) is now written as R ¼ BX ¼ BM ð1=2Þ XM

ð1:223Þ

~R ¼ ~XM M ð1=2Þ~B

ð1:224Þ

and its transpose is

where B is a rectangular (3 9) matrix shown by Eq. 1.119.

108

THEORY OF NORMAL VIBRATIONS

The kinetic energy (Eq. 1.113) in terms of XM is written as

~_ M X_ M 2T ¼ X

ð1:225Þ

and the potential energy (Eq. 1.111) in terms of the same coordinate, XM, is 2V ¼ ~ RFR R ¼ ~ XM M ð1=2Þ~ BFR BM ð1=2Þ XM ¼ ~ XM FXM XM

ð1:226Þ

where

or

FXM ¼ M ð1=2Þ~ BFR BM ð1=2Þ

ð1:227Þ

~BFR B ¼ M1=2 FXM M1=2

ð1:228Þ

Vibrational frequencies are calculated by solving the following equation: jFXM Elj ¼ 0

ð1:229Þ

In the case of H2O, we obtain nine frequencies, six of which are near zero. These six frequencies must be eliminated from the results because they correspond to the translational and rotational motions of the molecule as a whole (3N–6 rule). Thus, the remaining three frequencies correspond to the normal modes of the H2O molecule. It should be noted that vibrational frequencies are intrinsic of individual molecules and do not depend on the coordinate system chosen (R or XM). The next step is to convert the force constants (FXM) in terms of XM coordinates to those of internal coordinates (FR), which have clearer physical meanings. For this purpose, Eq. 1.228 is rewritten as [117–119] FR ¼ UG 1~ UBM1=2 FXM M1=2~ BUG 1~ U Here, U is an orthogonal matrix ðU ¼ ~ U matrix that is defined as follows:

1

ð1:230Þ

Þ that diagonalizes B~ B, and G is a diagonal

B~ BU ¼ UG

ð1:231Þ

In the case of H2O, B is a 3 9 matrix, whereas U, G, and FR are 3 3 matrices. To take advantage of symmetry properties, internal coordinates are further transformed to internal symmetry coordinates (RS) by the relation RS ¼ US R

ð1:131Þ

where U is an orthogonal matrix chosen by symmetry consideration. Then, the force constants in terms of internal symmetry coordinates ðFSR Þ are given by S

FSR ¼ US FR~ U

S

ð1:232Þ

VIBRATIONAL SPECTRA IN GASEOUS PHASE AND INERT GAS MATRICES

109

TABLE 1.16. Comparison of Observed and Calculated Frequencies (cm1) and Normal Modes of H2O Normal Vibration: a

Observed frequencies Calculated frequencies Normal modesb

X

n1(A1)

n2(A1)

n3(B2)

1595 1565.5

3657 3639.2

3756 3804.7

Y

Z

X

Y

Z

X

Y

Z

O(1) H(2) H(3)

0.00 0.00 0.08 0.00 0.38 0.60 0.00 0.38 0.60

0.00 0.00 0.04 0.00 0.62 0.35 0.00 0.62 0.35

0.00 0.07 0.00 0.00 0.58 0.40 0.00 0.58 0.40

Assignment

HOH bending

Symmetric OH stretch

Antisymmetric OH stretch

a

IR spectrum (Table 2.2d of Chapter 2). The normal modes shown in Fig. 1.12 are obtained by plotting these displacement coordinates.

b

Thus, FXM can be coverted to FSR using Eqs. 1.230 and 1.232. Several computer programs are available for this conversion [118]. Table 1.16 summarizes the results of DFT calculation of H2O. It is seen that the calculated frequencies are in good agreement with the observed values. In other cases, however, the former becomes higher than the latter because (1) the effect of anharmonicity is not included in the calculation and (2) systematic errors result from the use of finite basis sets and incomplete inclusion of electron correlation effects. In such cases, several empirical “scaling methods” are applied to the force constants [120] or frequencies [121] obtained by DFT calculations. Computer programs such as “Gaussian 03” can simultaneously calculate several other parameters including normal modes (shown in Table 1.15), IR intensities, Raman scattering intensities, and depolarization ratios. 1.25. VIBRATIONAL SPECTRA IN GASEOUS PHASE [122] AND INERT GAS MATRICES [29,30] As distinct from molecules in condensed phases, those in the gaseous phase are free from intermolecular interactions. If the molecules are relatively small, vibrational spectra of gases exhibit rotational fine structure (see Fig. 1.3) from which moments of inertia and hence internuclear distances and bond angles can be calculated [1,2]. Furthermore, detailed analysis of rotational fine structure provides information about the magnitude of rotation–vibration interaction (Coriolis coupling), centrifugal distortion, anharmonicity, and even nuclear spin statistics in some cases. In the past, infrared spectroscopy was the main tool in measuring gas-phase vibrational spectra. However, Raman spectroscopy is also playing a significant role because of the development of powerful laser sources and high-resolution spectrophotometers. For example, Clark and Rippon [123] measured gas-phase Raman spectra of Group IV tetrahalides, and calculated the Coriolis coupling constants from the observed band contours. For gas-phase Raman spectra of other inorganic compounds, see

110

THEORY OF NORMAL VIBRATIONS

Ref. [122]. Unfortunately, the majority of inorganic and coordination compounds exist as solids at room temperature. Although some of these compounds can be vaporized at high temperatures without decomposition, it is rather difficult to measure their spectra by the conventional method. Furthermore, high-temperature spectra are difficult to interpret because of the increased importance of rotational and vibrational hot bands. In 1954, Pimentel and his colleagues [124] developed the matrix isolation technique to study the infrared spectra of unstable and stable species. In this method, solute molecules and inert gas molecules such as Ar and N2 are mixed at a ratio of 1:500 or greater and deposited on an IR window such as a CsI crystal cooled to 10–15 K. Since the solute molecules trapped in an inert gas matrix are completely isolated from each other, the matrix isolation spectrum is similar to the gas-phase spectrum; no crystal field splittings and no lattice modes are observed. However, the former spectrum is simpler than the latter because, except for a few small hydride molecules, no rotational transitions are observed because of the rigidity of the matrix environment at low temperatures. The lack of rotational structure and intermolecular interactions results in a sharpening of the solute band so that evenvery closely located metal isotope peaks can be resolved in a matrix environment. This technique is also applicable to a compound which is not volatile at room temperature. For example, matrix isolation spectra of metal halides can be measured by vaporizing these compounds at high temperatures in a Knudsen cell and co-condensing their vapors with inert gas molecules on a cold window [125]. More recent developments of laser ablation techniques coupled with closed-cycle helium refrigerators greatly facilitated its application. Thus, matrix isolation technique has now been utilized to obtain structural and bonding information of a number of inorganic and coordination compounds. Some important features and applications are described in this and the following sections. 1.25.1. Vibrational Spectra of Radicals Highly reactive radicals can be produced in situ in inert gas matrices by photolysis and other techniques. Since these radicals are stabilized in matrix environments, their spectra can be measured by routine spectroscopic techniques. For example, the spectrum of the HOO radical [126] was obtained by measuring the spectrum of the photolysis product of a mixture of HI and O2 in an Ar matrix at 4 K. Chapter 2 lists the vibrational frequencies of many other radicals, such as CN, OF, and HSi, obtained by similar methods. 1.25.2. Vibrational Spectra of High-Temperature Species Alkali halide vapors produced at high temperatures consist mainly of monomers and dimers. The vibrational spectra of these salts at high temperatures are difficult to measure and difficult to interpret because of the presence of hot bands. The matrix isolation technique utilizing a Knudsen cell has solved this problem. The

VIBRATIONAL SPECTRA IN GASEOUS PHASE AND INERT GAS MATRICES

111

vibrational frequencies of some of these high-temperature species are listed in Chapter 2. 1.25.3. Isotope Splittings As stated before, it is possible to observe individual peaks due to heavy-metal isotopes in inert gas matrices since the bands are extremely sharp (half-band width, 1.5–1.0 cm1) under these conditions. Figure 1.36a shows the infrared spectrum of the n7 band (coupled vibration between CrC stretching and CrCO bending modes) of Cr(CO)6 in a N2 matrix [127]. The bottom curve shows a computer simulation using the measured isotope shift of 2.5 cm1 per atomic mass unit, a 1.2 cm1 half-band width, and the percentages of natural abundance of Cr isotopes: 50 Cr (4.31%), 52 Cr (83.76%), 53 Cr (9.55%), and 54 Cr (2.38%). The isotope splittings of SnCl4 and GeCl4 in Ar matrices are discussed in Sec. 2.6. These isotope frequencies are highly important in refining force constants in normal coordinate analysis.

Fig. 1.36. Matrix isolation IR spectra of Cr(CO)6 in N2 (a) and Ar (b) matrices. The bottom spectra were obtained by computer simulation [127].

112

THEORY OF NORMAL VIBRATIONS

1.25.4. Matrix Effect The vibrational frequencies of matrix-isolated molecules give slight shifts when the matrix gas is changed. This result suggests the presence of weak interaction between the solute and matrix molecules. In some cases, the spectra are complicated by the presence of more than one trapping site. For example, the infrared spectrum of Cr(CO)6 in an Ar matrix (Fig. 1.36b) is markedly different from that in a N2 matrix (Fig. 1.36a) [127]. The former spectrum can be interpreted by assuming two different sites in an Ar matrix. The computer-simulation spectrum (bottom curve) was obtained by assuming that these two sites are populated in a 2:1 ratio, the frequency separation of the corresponding peaks being 2 cm1. Thus, it is always desirable to obtain the matrix isolation spectra in several different environments.

1.26. MATRIX COCONDENSATION REACTIONS A number of unstable and transient species have been synthesized via matrix cocondensation reactions, and their structure and bonding have been studied by vibrational spectroscopy. The principle of the method is to cocondense two solute vapors (atom, salt, or molecule) diluted by an inert gas on an IR window (IR spectroscopy) or a metal plate (Raman spectroscopy) that is cooled to low temperature by a cryocooler. Solid compounds can be vaporized by conventional heating (Knudsen cell), laser ablation, or other techniques, and mixed with inert gases at proper ratios [128]. In general, the spectra of the cocondensation products thus obtained exhibit many peaks as a result of the mixed species produced. In order to make band assignments, the effects of changing the temperature, concentration (dilution ratios), and isotope substitution on the spectra must be studied. In some cases, theoretical calculations (Sec. 1.24) must be carried out to determine the structure and to make band assignments. Vibrational frequencies of many molecules and ions obtained by matrix cocondensation reactions are listed in Chapter 2. Matrix cocondensation reactions are classified into the following five types: 1. Atom-Atom Reaction. Liang et al. [129] studied the reaction: U þ S ! USn(n ¼ 1–3), Figure 1.37 shows the IR spectra of laser-ablated U atom vapor cocondensed with microwave-discharged S atom in Ar at 7 K. In (a) (32 S with higher concentration) and (b) (32 S with low concentration), the prominent band at 438.7 cm1 and a weak band at 449.8 cm1 show a consistent intensity ratio of 6 : 1, and are shifted to 427.8 and 437.4 cm1 respectively, by 32 S/34 S substitution (c). In the mixed 32 S þ 34 S experiments [traces] (d) and (e), both bands split into asymmetric triplets marked in (e). These results suggest that two equivalent S atoms are involved in this species. The bands at 438.7 and 449.8 cm1 were assigned to the antisymmetric (n3) and symmetric (n1) stretching modes, respectively, of the nonlinear U 32 S2 molecule. Although the analogous UO2 molecule is linear, the US2 molecule is bent because of more favorable U(6d)–S(3p) overlap. The weak band at 451.1 cm1, which is partly overlapped by the n1 mode of US2 in (a), is shifted to 439.2 cm1 by 32 S/34 S substitution (c), and was assigned to the diatomic US molecule. Another weak band at 458.2 cm1 in (a) shows a

113

MATRIX COCONDENSATION REACTIONS

US2 (v3) US (e) 1.0

Absorbance

US2 (v1) (d) U34S U34S2(v1)

US3

0.6

U34S2(v3) (c)

US2 (v1) 0.2

US2 (v3) US3

470

(b)

460

US 450 440 Wavenumbers (cm-1)

(a) 430

420

Fig. 1.37. Infrared spectra of laser-ablated U co-deposited with discharged isotopic S atoms in Ar at 7 K: (a) 32S with higher concentration, (b) 32S with lower concentration, (c) 34S, (d) 35/65 32 S þ 34S mixture, and (e) 50/50 32S þ 34S mixture [129].

triplet pattern in the mixed-isotope experiment shown in (d), and was assigned to the antisymmetric stretching vibration of the near-linear SUS unit of a T-shaped US3 molecule similar to UO3 All these structures and band assignments are supported by DFT calculations. 2. Atom–Molecule Reaction. Zhou and Andrews [130] studied the reaction of Sc þ CO ! Sc(CO)n(n ¼ 1–4) via cocondensation of laser-ablated Sc atoms with CO in excess Ar or Ne, and made band assignments based on the results of 13 CO and C18O isotope shifts, matrix annealing, broadband photolysis, and DFT calculations. Figure 1.38a shows the IR spectrum of the cocondensation product after one hour deposition at 10 K in Ar. The strong and sharp band at 1834.2 cm1 is shifted to 1792.9 and 1792.3 cm1 by 13 CO and C18O substitutions, respectively, and its intensity decreases upon annealng to 25 K (b) but increases on photolysis (c,d). Two weak bands near 1920 cm1 in (a) are also isotope-sensitive and become stronger on annealing to 25 K (b). These bands were assigned to the Sc(CO)þ cation produced by the reaction of laser-ablated Scþ ion with CO. In a Ne matrix, a similar experiment exhibits a sharp band at 1732.0 cm1 that was assigned to the Sc(CO) ion produced via electron capture by Sc(CO). The pair of bands at 1851.4 and 1716.3 cm1 in (b) were assigned to the antisymmetric and symmetric stretching vibrations of the bent Sc(CO)2 molecule while the strong and sharp band at 1775.5 cm1 in (c) was attributed to the linear Sc(CO)2 molecule. DFT calculations predict Sc(CO)3 to be trigonal pyramidal (C3v symmetry) in the ground state. The bands at 1968.0 and 1822.2 cm1 in (b) become weaker by photolysis (c,d) and stronger by annealing at 35 K (e), These bands were assigned to the nondegenerate symmetric and degenerate CO stretching

114

THEORY OF NORMAL VIBRATIONS

0.065

Sc(CO)3 0.060

Sc(CO)4

Sc(CO)2

(e)

0.055 0.050 0.045

Absorbance

(d) 0.040 0.035

Sc(CO)2

0.030

(c)

0.025

ScCO4

0.020 0.015

Sc(CO)2

Sc(CO)3

0.010

Sc(CO)3

Sc(CO)2 (b)

ScCO

0.005

(a)

0.000 1950

1900

1850 Wavenumber (cm-1)

1800

1750

Fig. 1.38. Infrared spectra of laser-ablated Sc atoms co-deposited with 0.5 % CO in excess Ar.: (a) after 1 hr sample deposition at 10 K, (b) after annealing to 25 K, (c) after 20 min l > 470 nm photolysis, (d) after 20 min full-arc photolysis, and (e) after annealing to 35 K [130].

vibrations, respectively, of the Sc(CO)3 molecule. The band at 1865.4 cm1 appeared last when the matrix was annealed to 35 K (e), and was tentatively assigned to a degenerate vibration of Sc(CO)4. Matrix cocondensation reactions of this type have been studied most extensively as shown in Part B, Sec. 3.18.6. 3. Salt–Molecule Reaction. Tevault and Nakamoto [131] carried out matrix cocondensation reactions of metal salts such as PbF2 with L(CO, NO and N2) in Ar, and confirmed the formation of PbF2–L adducts by observing the shifts of IR bands of both components. These spectra are shown in Fig. 3.67 (of Sec. 3.18.6). 4. Salt–Salt Reaction. Ault [132] obtained the IR spectrum of the PbF 3 ion via the reaction of the type, MF þ PbF2 ! M[PbF3] (M ¼ Na, K, Rb, and Cs). Both salts were premixed, and the mixture was heated in a single Knudsen cell at 300–325 C, which is roughly 200 C below that is needed to vaporize either PbF2 or MF. Figure 1.39 shows the IR spectra of PbF2 and its cocondensation products with MF in Ar matrices. It is seen that, except for M ¼ Na, the cocondensation products exhibit two prominent bands near 456 and 405 cm1 with no bands due to PbF2 Thus, these two bands were assigned to the A1 and E stretching vibrations, respectively, of the PbF3 ion of C3v symmetry. The band at 242 cm1 was observed weakly only in high-yield experiments, and tentatively assigned to a bending mode of the PbF3 ion. 5. Molecule–Molecule Reaction. The reaction Fe(TPP) þ O2 ! (TPP) Fe-O2 (TPP tetraphenylporphyrin) was carried out by Watanabe et al. [133] using 16 O2 and 18 O2

SYMMETRY IN CRYSTALS

115

PbF2

NaF/PbF2

ABSORPTION

KF/PbF2

RbF/PbF2

CsF/PbF2

560

520

480

440

400

360

320

ENERGY (cm–1) Fig. 1.39. Infrared spectra of the M þ PbF 3 isolated in Ar matrices. The top trace shows the spectrum of the parent PbF2 salt, while the next four spectra show the bands attributable to the vaporization products from the solid mixtures of NaF/PbF2, KF/PbF2, RbF/PbF2, and CsF/PbF2 in the 580 300 cm 1 region [132].

diluted in Ar. The IR spectra shown in Fig. 3.76 (in Sec. 3.21.4) indicate that two new bands appear at 1195 and 1108 cm1 which are due to the O2 stretching vibrations of the end-on and side-on type Fe(TPP)O2, respectively. Bajdor and Nakamoto [134] obtained a novel oxoferrylporphyrin [O¼Fe(IV)(TPP)] via laser photolysis of Fe (TPP)O2 in pure O2 matrices at 15 K. As seen in Fig. 1.82 ( Sec. 1.22.4 of Part B), it exhibits the Fe(IV)¼O stretching vibration at 852 cm1. 1.27. SYMMETRY IN CRYSTALS The symmetry elements and point groups of molecules and ions in the free state have been discussed in Sec. 1.5. For molecules and ions in crystals, however, it is necessary to consider some additional symmetry operations that characterize translational symmetries in the lattice. Addition of these translational operations results in the formation of the space groups that can be used to classify the symmetry of molecules and ions in crystals.

116

THEORY OF NORMAL VIBRATIONS

TABLE 1.17. Seven Crystallographic Systems and 32 Crystallographic Point Groups System

32 Crystallographic Point Groups

Axes and Angles

Triclinic

a a

b b

c g

I I C1 Ci

Monoclinic

a 90

b b

c 90

2/m 2 m or 2 C2 Cs C2h

Orthorthombic

a 90

b c 90 90

Tetragonal

a 90

b c 90 90

Trigonal aaa or (rhombohedral) aaaa

a 90

222 mm2 mmm D2 C2n D2h 4 4 C4 S 4 a c 3 3 90 120 C3 C3i

4/m 422 4mm 42m C4h D4 C4v D2d 32 D3

3m C3v

4/mmm D4h

3m D3d

Hexagonal

a 90

a c 90 120

6 6 C6 C3h

6/m 622 6mm 6m2 C6h D6 C6v D3h

Cubic

a 90

a c 90 90

23 m3 T Th

432 43m O Td

6/mmm D6h

m3m Oh

1.27.1. 32 Crystallographic Point Groups As discussed in Sec. 1.5, molecular symmetry can be described by point groups that are derived from the combinations of symmetry operations, I (identity), Cp (p-fold axis of symmetry), a (plane of symmetry), i (center of symmetry), and Sp (p-fold rotation–reflection axis), where p may range from 1 to ¥. In the case of crystal symmetry, p is limited to 1, 2, 3, 4, and 6 because of space-filling requirement in a crystal lattice. As a result, only the 32 crystallographic point groups listed in Table 1.17 are possible. Here, Hermann–Mauguin (HM) as well as Sch€onflies (S) notations are shown. In the HM notation, 1, 2, 3, 4, and 6 denote the principal axis of rotation, Cp, where p is 1, 2, 3, 4, and 6, respectively, and 1; 2; 3; 4, and 6 denote a combination of respective rotation about an axis and inversion through a point on the axis. Thus, 1 means the center of inversion (i), 2 indicates s on a mirror plane (m), and 3; 4; and 6 correspond to S6, S4, and S3, respectively. For example, the point group C2h, (I, C2, a, and i) can be written as 2m 1, which can be simplified to 2/m where the slash means the C2 axis is perpendicular to m. As shown in Table 1.17, these 32 crystallographic point groups which describe the symmetry of the unit cell can be classified into seven crystallographic systems according to the most simple choice of reference axes. 1.27.2. Symmetry of the Crystal Lattice A crystal has a lattice structure that is a repetition of identical units. Bravais has shown that only the 14 different types of lattices shown in Fig. 1.40 are possible. These

SYMMETRY IN CRYSTALS

117

Fig. 1.40. The 14 Bravais lattices belonging to the seven crystallographic systems [135].

Bravais lattices are also classified into the seven crystallographic systems mentioned above. Here, the lattice points are regarded as “points” or “small, indentical, perfect spheres,” thus representing the highest symmetry possible. In actual crystals, these points are occupied by unsymmetric but identical molecules or ions so that their high symmetry is lowered. In Fig. 1.40 P, C (or A or B), F, I, and R indicate primitive, basecentered, face-centered, body-centered, and rhombohedral lattices, respectively. The number of lattice points (LP) in each cell is summarized in Table 1.18. For crystal structures designated by P, crystallographic unit cells are identical with the Bravais unit cells. For those designated by other letters, crystallographic unit cells contain 2, 3, or 4 Bravais cells. In these cases, the number of molecules in the crystallographic unit

118

THEORY OF NORMAL VIBRATIONS

TABLE 1.18. Primitive and Centered Lattices Type of Crystal Structure Primitive (P) Base-centered (A, B, or C) Body-centered (I) Rhombohedral (R) Face-centered (F) a

Number of Lattice Points (LP) 1 2 2 3 or 1a 4

The crystallographic unit cell may have been divided by 3.

cell must be divided by the corresponding LP since only the consideration of one Bravais cell is sufficient to interpret the vibrational spectra of crystals. 1.27.3. Space Groups As stated above, the 32 point groups describe the symmetry of the unit cells, and the 14 Bravais lattices describe all possible arrangements of crystal lattices. To describe space symmetry, however, symmetry operations mentioned earlier are not enough. We must add three new symmetry operations: Translation. A translational operation along one lattice direction gives an identical point. Screw Axis. Rotation of a lattice point about an axis followed by translation gives an identical point. Namely, rotation by 2p/p followed by a translation of q=pðq ¼ 1; 2; . . . ; n 1Þ in the direction of the axis is written as pq. For example, 21 indicates twofold rotation followed by a translation of 12 a, 12 b, or 12 c, as shown in Fig. 1.41A. If the axis is threefold, the translation must be 13 or 23, and the symbols are 31 and 32, respectively. Glide Plane. Reflection in a plane followed by translation by 12 a, 12 b, 12 c, as shown in Fig. 1.41B. These planes are indicated by a, b, and c, respectively, n is used for diagonal translation of (b þ c)/2, (c þ a)/2, or (a þ b)/2.

Fig. 1.41. Screw axis and glide plane.

VIBRATIONAL ANALYSIS OF CRYSTALS

119

Addition of these symmetry operations results in 230 different combinations which are called “space groups.” Table 1.19 shows the distribution of space groups among the seven crystal systems [136]. Some of these space groups are rarely found in actual crystals. About half of the crystals belong to 13 space groups of the monoclinic system. Space groups can be described by using either Hermann–Mauguin (HM) or Sch€onflies 1 (S) notations. For example, P2/mð C2h Þ indicates a primitive lattice with m perpendicular to 2. The first symbol (capital letter) refers to the Bravais lattice (P, A, B, C, F, I, and R), and the nextsymbol indicates fundamental symmetry elements that lie along the special direction in the crystal. In the case of the monoclinic system, it is the twofold 2 axis. Thus, P21/mð C2h Þ is a primitive cell with m perpendicular to 21. These symbols indicate only the minimum symmetry elements that are necessary to distinguish the space groups. The differences among space groups can be found by comparing the “unit cell maps” given in the International Tables for X-Ray Crystallography [137] . 1.28. VIBRATIONAL ANALYSIS OF CRYSTALS [48–52] Becauseofintermolecularinteractions,thesymmetryofamoleculeis generally lowerin the crystalline state than in the gaseous (isolated) state. This change in symmetry may split the degenerate vibrations and activate infrared- (or Raman-) inactive vibrations. Additionally, intermolecular (interionic) interactions in the crystal lattice cause frequency shifts relative to the gaseous state. Finally, the spectra obtained in the crystalline state exhibit lattice modes—vibrations due to translatory and rotatory motions of a molecule in the crystalline lattice. Although their frequencies are usually lower than 300 cm1, they may appear in the high-frequency region as the combination bands with internal modes (see Fig. 2.18, for example). Thus, the vibrational spectra of crystals must be interpreted with caution, especially in the low-frequency region. To analyze the spectra of crystals, it is necessary to carry out a site group or factor group analysis, as described in the following subsection. 1.28.1. Subgroups and Correlation Tables For any point group, there are “subgroups” which consist of some, but not all, symmetry elements of the “parent group.” For example, the character table of the point group, C3v (Table 1.6) contains I, C3, and sv as the symmetry elements. Then, the subgroups of C3v are C3, consisting of I and C3, and Cs, consisting of I and sv only. The relationship between the symmetry species in the “parent group” and those in “subgroups” is given in a “correlation table.” Such correlation tables have already been worked out for all common point groups and are listed in Appendix IX. As an example, the correlation table for C3v is shown below: C3v

C3

Cs

A1 A2 E

A A E

A 00 A 0 00 A þA

0

120

HMb

1 1

2 m 2/m

222

mm2

mmm

4 4 4/m 422

4mm

42m

4/mmm

C1 Ci

C2(1–3) CS(1–4) C2h(1–6)

D2(1–9)

C2v(1–22)

D2h(1–28)

C4(1–6) S4(1–2) C4h(1–6) D4(1–10)

C4v(1–12)

D2d(1–12)

D4h(1–20)

Triclinic

Monoclinic

Orthorhombic

Tetragonal

System

Sa

Point Groups

TABLE 1.19. The 230 Space Groups

Pnnn Pbam Pnma Fmmm P41 I4 P42/m P4212 I422 P4bm I4mm P42c I4m2

P4 P4 P4/m P422 P43212 P4mm P42bc P42m P4n2 P4/mcc P42/mmc P42/ncm

P2221 I21212 Pmc21 Pna21 Ama2

P222 I222 Pmm2 Pba2 Abm2 Ima2 Pmmm Pcca Pbca Ccca

P4/mmm P4/ncc P42/ncm

P21 Pc P21/m

P2 Pm P2/m

P1 P1

P42/n P41212

P4/n P4122 I4122 P42cm I4cm 1m P42 I4c2 P4/nbm P42/mcm I4/mmm

P42nm I41md 1c P42 I42m P4/nnc P42/nbc I4/mcm

P43

Pban Pbcm Cmca Immm

Pma2 Cmm2 Fmm2

P212121

Cc P2/c

P42

Pccm Pccn Cmcm Fddd

Pcc2 Pnm2 Aba2

P21212

C2 Cm C2/m

Space Group

P21/c

P4/mbm P42/nnm I41/amd

P4cc I41cd P4m2 I42d

I4/m P4222

I4

Pmma Pnnm Cmmm Ibam

Pca21 Cmc21 Fdd2

C2221

P4/nmm P42/mnm

P4b2

P4c2 P4/mnc P42/mbc I41/acd

P42mc

P4322

Pmna Pbcn Cmma Imma

Pmn21 Amm2 Iba2

F222

P4nc

I41/a P42212

I41

Pnna Pmmn Cccm Ibca

Pnc2 Ccc2 Imm2

C222

C2/c

121

€nflies. Scho Hermann–Mauguin.

Source: Ref. [136].

b

a

F23 Pn3 P4232

23 m3 432

43m m3m

Td(1–6) Oh(1–10)

F43m Pn3n Im3m

P6/mcc

P6/mmm P23 Pm3 P432 I4132 P43m Pm3m Fd3c

6/mmm

T (1–5) Th(1–7) O(1–8) I43m Pm3n Ia3d

I23 Fm3 F432

P63/mcm

P6522 P63cm P62m

P63/m P6122 P6cc P6c2

Cubic

P65

P61

P6 P6 P6/m P622 P6mm P6m2

6 6 6/m 622 6mm 6m2

Ph3m

P43n

P213 Fd3 F4132

P63/mmc

P6222 P63mc P62c

P62

P3121 P31c P3c1

P3112 P3c1 P3m1

C6(1–6) C3h(1) C6h(1–2) D6(1–6) C6v(1–4) D3h(1–4) D6h(1–4)

Hexagonal

R3

Space Group P32

P31 R3 P321 P31m P31c

C3(1–4) C3i(1–2) D3(1–7) C3v(1–6) D3d(1–6)

Trigonal

P3 P3 P312 P3m1 P31m

3 3 32 3m 3m

S

HMb

Point Groups

System

a

F43c Fm3m

I213 Im3 I432

P6422

P64

P3212 R3m R3m

I43d Fm3c

Pa3 P4332

P6322

P63

P3221 R3c R3c

Fd3m

Ia3 P4132

R32

122

THEORY OF NORMAL VIBRATIONS

A pyramidal XY3 molecule such as NH3 belongs to the point group C3v. If one of the Yatoms is replaced by a Z atom, the resulting XY2Z molecule belongs to the point group Cs. As a result, the doubly degenerate vibration (E) splits into two vibrations (A0 þ A00 ). These correlation tables are highly important in predicting the effect of lowering of symmetry on molecular vibrations. Chapter 2 includes a number of examples of symmetry lowering by substitution of an atomic group by another group. 1.28.2. Site Group Analysis According to Halford [138], the vibrations of a molecule in the crystalline state are governed by a new selection rule derived from site symmetry—a local symmetry around the center of gravity of a molecule in a unit cell. The site symmetry can be found by using the following two conditions: (1) the site group must be a subgroup of both the space group of the crystal and the molecular point group of the isolated molecule, and (2) the number of equivalent sites must be equal to the number of molecules in the unit cell. Halford derived a complete table that lists possible site symmetries and the number of equivalent sites for 230 space groups. Suppose that the space group of the crystal, the number of molecules in the unit cell (Z), and the point group of the isolated molecule are known. Then, the site symmetry can be found from the Table of Site Symmetry for the 230 Space Groups, which was originally derived by Halford. Appendix X gives its modified version by Ferraro and Ziomek [9]. In general, the site symmetry is lower than the molecular symmetry in an isolated state. In some cases, it may be difficult to make an unambiguous choice of site symmetry by the method cited above. Then, Wyckoff ’s tables on crystallographic data [139] must be consulted. The vibrational spectra of calcite and aragonite crystals are markedly different, although both have the same composition (Sec. 2.4). This result can be explained if we consider the difference in site symmetry of the CO2 3 ion between these crystals. According to X-ray analysis, the space group of calcite is D63d and Z is 2 Appendix X gives D3 ð2Þ;

C3i ð2Þ;

C3 ð4Þ;

Ci ð6Þ;

C2 ð6Þ;

C1 ð12Þ

as possible site symmetries for space group D63d (the number in front of point group notation indicates the number of distinct sets of sites, and that in parentheses denotes the number of equivalent sites for each distinct set). Rule 2 eliminates all but D3(2) and C3i(2). Rule 1 eliminates the latter since C3i is not a subgroup of D3h. Thus, the site symmetry of the CO23 ion in calcite must be D3. On the other hand, the space group of aragonite is D16 2h and Z is 4. Appendix X gives 2Ci ð4Þ;

Cs ð4Þ;

C1 ð8Þ

Since Ci is not a subgroup of D3h, the site symmetry of the CO23 ion in aragonite must be Cs. Thus, the D3h, symmetry of the CO23 ion in an isolated state is lowered to D3 in

VIBRATIONAL ANALYSIS OF CRYSTALS

123

TABLE 1.20. Correlation Table for D3h, D3, C2v, and Cs n1

Point Group D3h D3 C2v Cs

0

A 1 ðRÞ A1(R) 0 A 1 ðI; RÞ A0 (I, R)

n2 00

A 2 ðIÞ A2(I) B1 (I, R) A00 (I, R)

n3 0

E (I, R) E (I, R) A1 (I, R) þ B2 (I, R) A0 (I, R) þ A0 (I, R)

n4 0

E (I, R) E (I, R) A1 (I, R) þ B2 (I, R) A0 (I, R) þ A0 (I, R)

calcite and to Cs in aragonite. Then, the selection rules are changed as shown in Table 1.20. There is no change in the selection rule in going from the free CO23 ion to calcite. In aragonite, however, n1 becomes infrared active, and n3 and n4 each split into two bands. The observed spectra of calcite and aragonite are in good agreement with these predictions (see Table 2.4b). 1.28.3. Factor Group Analysis A more complete analysis including lattice modes can be made by the method of factor group analysis developed by Bhagavantam and Venkatarayudu [140]. In this method, we consider all the normal vibrations for an entire Bravais cell. Figure 1.42 illustrates the Bravais cell of calcite, which consists of the following symmetry elements:

Fig. 1.42. The Bravais cell of calcite.

124

THEORY OF NORMAL VIBRATIONS

I, 2S6, 2S262C3 , S36 i; 3C2 , and 3sv (glide plane). These elements are exactly the same as those of the point group D3d, although the last element is a glide plane rather than a plane of symmetry in a single molecule. As mentioned in Sec. 1.27, it is possible to derive the 230 space groups by combining operations possessed by the 32 crystallographic point groups with operations such as pure translation, screw rotation (translation þ rotation), and glide plane reflection (translation þ reflection). If we regard the translations that carry a point in a unit cell into the equivalent point in another cell as identity, we define the 230 factor groups that are the subgroups of the corresponding space groups. In the case of calcite, the factor group consists of the symmetry elements described above, and is denoted by the same notation as that used for the space group ðD63d Þ. The site group discussed previously is a subgroup of a factor group. Since the Bravais cell contains 10 atoms, it has 3 10 – 3 ¼ 27 normal vibrations, excluding three translational motions of the cell as a whole.* These 27 vibrations can be classified into various symmetry species of the factor group D63d , using a procedure similar to that described in Sec. 1.8 for internal vibrations. First, we calculate the characters of representations corresponding to the entire freedom possessed by the Bravais cell [cR(N)], translational motions of the whole cell [cR(T)], translatory lattice modes [cR(T 0 )] rotatory lattice modes [cR(R0 )], and internal modes [cR(n)], using the equations given in Table 1.21. Then, each of these characters is resolved into the symmetry species of the point group, D3d. The final results show that three internal modes (A2u and two Eu), three translatory modes (A2u and two Eu), and two rotatory modes (A2u and Eu) are infrared-active, and three internal modes (A1g and two Eg) one translatory mode (Eg), and one rotatory mode (Eg) are Raman-active. As will be shown in the following sections, these predictions are in perfect agreement with the observed spectra. 1.29. THE CORRELATION METHOD In the preceding section, we described the application of factor group analysis to the calcite crystal. However, the correlation method developed largely by Fateley et al. [15] is simpler and gives the same results. In this method, intramolecular and lattice vibrations are classified under point groups of molecular symmetry, site symmetry, and factor group symmetry, and correlations are made using the correlation tables given in Appendix IX. In the following, we demonstrate its utility using calcite as an example. For more detailed discussions and applications, the reader should consult the books by Fateley et al. [15] and Ferraro and Ziomek [9]. 1.29.1. The CO23 Ion in the Free State As mentioned in Sec. 1.4, normal vibrations of a molecule can be described in terms of translational motions of the individual atoms along the x, y, and z axes. Thus, the *These three motions give acoustical modes (Sec. 1.30).

125

Key: p, total number of atoms in the Bravais cell. s, total number of molecules (ions) in the Bravais cell. n, total number of monoatomic molecules (ions) in the Bravais cell. NR(p), number of atoms unchanged by symmetry operation R. NR(s), number of molecules (ions) whose center of gravity is unchanged by symmetry operation R. NR(s n), NR(s) minus number of monoatomic molecules (ions) unchanged by symmetry operation R. cR(N) ¼ NR (p)[ (1 þ 2 cos y)], character of representation for entire freedom possessed by the Bravias cell. cR(T) ¼ (1 þ 2 cos y), character of representation for translation motions of the whole Bravais cell. cR(T0 ) ¼ {NR (s) 1} { (1 þ 2 cos y)} character of representation for translatory lattice modes. cR(R0 ) ¼ NR (s n) { (1 þ 2 cos y)}, character of representation for rotary lattice modes. cR(n) ¼ cR(N) cR(T) cR(T0 ) cR(R0 ), character of representation for internal modes. Note that þ and – signs are for proper and improper rotations, respectively. The symbol y should be taken as defined in Sec. 1.8.

TABLE 1.21. Factor Group Analysis of Calcite Crystal

126

THEORY OF NORMAL VIBRATIONS

Fig. 1.43. The x, y, and z axes chosen for planar CO32 ion.

normal modes of an N-atom molecule can be expressed by using 3N translational motions. Furthermore, the symmetry species of these normal modes must correlate with those of the translational motions of an atom located at a particular site within the molecule. Since the latter are known from the molecular structure, the former may be determined directly by using the correlation table. As an example, consider a planar CO23 ion for which the x, y, and z axes are chosen as shown in Fig. 1.43. It is readily seen that the C atom is situated at the D3h site, whereas the O atom is situated at a site where the local symmetry is only C2v (I, C2, sh, 00 and sv). The translational motions of the C atom under D3h symmetry are: TzðA 2 Þand (Tx, Ty)(E0 ). The translational motions of the O atom under C2v symmetry are: Tx(A1), Ty(B2) and Tz(B1). In Table 1.22, the symmetry species of these translational motions are connected by arrows to those of the whole ion using the correlation table. Then, the number of times a particular symmetry species occurs in the total representation is given by the number of arrows that terminate on that species. The number of normal vibrations of the CO23 ion in each symmetry species is given by subtracting those of the translational and rotational motions of the whole ion from the total representation. The results are shown at the bottom of Table 1.22. 0 00 Thus, the CO23 ion exhibits four normal vibrations; n1 (A 1 Raman-active), n2 (A 2 0 IR-active), n3(E0 , IR- and Raman-active), and n4 (E , IR- and Raman-active). The normal modes of these four vibrations are shown in Fig. 2.8. In Sec. 1.8, we derived the general method to classify normal vibrations into symmetry species based on group theory. As seen above, this procedure is greatly simplified if we use the correlation method. 1.29.2. Intramolecular Vibrations in the CO23 Ion in Calcite Applying the same principle as that used above, we can classify the normal vibrations of the CO23 ion in the calcite lattice by using the correlation method. As discussed in

THE CORRELATION METHOD

127

TABLE 1.22. Correlation Method for the CO2 3 Ion in the Free State

Sec. 1.28, calcite belongs to the space group D63d , and Z is 2. From the Bravais cell shown in Fig. 1.42, it is readily seen that the CO23 ion is at the D3 site, whereas the Ca2þ ion is at the C3i (S6) site. Table 1.23 shows the correlations among the molecular symmetry (D3h), site symmetry (D3), and factor group symmetry (D3d). Under the “Vib.” column, we list the number of vibrations belonging to each species of D3h symmetry (Table 1.22). Throughout the present correlations, the number of degrees of vibrational freedom for all the doubly degenerate species must be multiplied by 2 since there are two E0 modes (n3 and v4). Thus, it is 4 for the E0 species. The number under “f’” indicates “Vib.” times Z (¼2), which is the number of vibrations of the unit cell. In going from D3h to D3, no changes occur TABLE 1.23. Correlation between Molecular Symmetry, Site Symmetry, and Factor Group Symmetry for Intramolecular Vibrations of CO2 3 Ion in Calcite

128

THEORY OF NORMAL VIBRATIONS

TABLE 1.24. Correlation between Site Symmetry and Factor Group Symmetry for Lattice Vibrations of CO2 3 Ion in Calcite

except for changes in notations of symmetry species. Column C indicates the degeneracy, and column am shows the degree of vibrational freedom contributed by the corresponding molecular symmetry species. Finally, the species under D3 symmetry are connected to those of the factor group by the arrows using the correlation table. The last column, “a”, is the degree of vibrational freedom contributed by the corresponding site symmetry species to the factor group species. It should be noted that the number of degrees of vibrational freedom must be 12 throughout the described correlations above. Such bookkeeping must be carried out for every correlation. 1.29.3. Lattice Vibrations in Calcite Table 1.24 shows the correlation diagram for lattice vibrations of the CO23 ion. The variables “t” and “f” denote the degrees of translational freedom of the CO23 ion for each ion and for the Bravais cell, respectively. The same result is obtained for the rotatory lattice vibrations. Table 1.25 shows the correlation diagram for translatory TABLE 1.25. Correlation between Site Symmetry and Factory Group Symmetry for Lattice Vibrations of the Ca2þ Ion in Calcite

129

LATTICE VIBRATIONS

TABLE 1.26. Distribution of Normal Vibrations of Calcite as Obtained by Correlation Method Symmetry Species of Factor Group (D3d) A1g (Raman) A1u A2g A2u (IR) Eg (Raman) Eu (IR) Total

Translatory Lattice

Acoustical

Rotatory Lattice

Intramolecular

0 1 1 1 1 2

0 0 0 1 0 1

0 0 1 1 1 1

1 1 1 1 2 2

9

3

6

12

lattice vibrations of the Ca2þ ion. No rotatory lattice vibrations exist for single-atom ions such as the Ca2þ ion. The distribution of all the translatory lattice vibrations can be obtained by subtracting c (acoustical) from c (trans, CO23 ) þ c (trans, Ca2þ). 1.29.4. Summary Table 1.26 summarizes the results obtained in Tables 1.23–1.25. The total number of vibrations including the acoustical modes should be 30 since the Bravais cell contains 10 atoms. These results are in complete agreement with that obtained by factor group analysis (Table 1.21). 1.30. LATTICE VIBRATIONS [48] Consider a lattice consisting of two alternate layers of atoms; atoms 1 of mass M1 lie on one set of planes and atoms 2 of mass M2 lie on another set of planes, as shown in Fig. 1.44. For example, such an arrangement is found in the 111 plane of the TABLE 1.27. Correlation between Site Symmetry (C2v) and Factor Group Symmetry (D2h) for Orthorhombic form of the 123 Conductora

a

For f, t, and a, see Sec. 1.29. The complete correlation table is found in Appendix IX.

130

THEORY OF NORMAL VIBRATIONS

Fig. 1.44. Displacements of atoms 1 and 2 in a diatomic lattice.

NaCl crystal. Let a denote the distance between atoms 1 and 2. Then the length of the primitive cell is 2a. We consider waves that propagate in the direction shown by the arrow, and assume that each plane interacts only with its neighboring planes. If the force constants (f) are identical between all these neighboring planes, we have d 2 us M1 2 ¼ f ðvs þ vs 1 2us Þ ð1:233Þ dt M2

d 2 vs ¼ f ðus þ 1 þ us 2vs Þ dt2

ð1:234Þ

where us and vs denote the displacements of atoms 1 and 2 in the cell indexed by s. The solutions for these equations take the form of a traveling wave having different amplitudes u and v. Thus, we obtain us ¼ u exp½iðwt þ 2skaÞ vs ¼ v exp½iðwt þ 2skaÞ

ð1:235Þ ð1:236Þ

Here w is angular frequency, 2pn [in reciprocal seconds (s1)]. If these are substituted in Eqs. 1.233 and 1.234, respectively, we obtain ½M1 w2 2f u þ f ½1 þ expð 2ikaÞv ¼ 0

ð1:237Þ

f ½1 þ expð2ikaÞu þ ½M2 w2 2f v ¼ 0

ð1:238Þ

These homogeneous linear equations have a nontrivial solution if the following determinant is zero: M1 w2 2f f ½1 þ expð 2ikaÞ ð1:239Þ ¼0 f ½1 þ expð2ikaÞ M2 w2 2f By solving this equation, we obtain ! " !2 #1=2 1 1 1 1 4sin2 ka w ¼f þ þ þf M1 M2 M1 M2 M1 M2 ðoptical branchÞ 2

ð1:240Þ

LATTICE VIBRATIONS

1 M2

1/2

(M2f ( 2f D M

(

+

(

1 M1

Optical

B

(

(

2f

2

E

131

1/2

C

M1 > M2

1

ω

1/2

F A 0

Acoustical k

π/2a

Fig. 1.45. Dispersion curves for lattice vibrations.

! " !2 #1=2 1 1 1 1 4sin2 ka w ¼f þ þ f M1 M2 M1 M2 M1 M2 ðacoustical branchÞ 2

ð1:241Þ

The term k in Eqs. 1.235–1.241 is called the wavevector and indicates the phase difference between equivalent atoms in each unit cell. In the case of a one-dimensional lattice, |k| ¼ k. Thus, we use k rather than |k| in this case, and k can take any value between p/2a and þ p/2a. This region is called the first Brillouin zone. Figure 1.45 shows a plot of w versus k for the positive half of the first Brillouin zone. There are twovalues for each w that constitute the “optical” and “acoustical” branches in the dispersion curves. At the center of the first Brillouin zone (k ¼ 0), we have w ¼ 0 and

and

u M2 ¼ v M1

u¼v

ðacousticalÞ vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ! u u 1 1 t w ¼ 2f þ M1 M2

or M1 u þ M2 v ¼ 0

ð1:242Þ

ð1:243Þ

ðopticalÞ

At the end of the first Brillouin zone (k ¼ p/2a), we have rﬃﬃﬃﬃﬃﬃ 2f and u ¼ 0 ðopticalÞ w¼ M2 rﬃﬃﬃﬃﬃﬃ 2f w¼ and v ¼ 0 ðacousticalÞ M1

ð1:244Þ ð1:245Þ

points A, B, C, and D in Fig. 1.45 correspond to Eqs. 1.242, 1.243, 1.244, and 1.245, respectively.

132

THEORY OF NORMAL VIBRATIONS

To observe lattice vibrations in IR spectra, the momentum of the IR photon must be equal to that of the phonon.* The momentum of the photon (P) is given by P¼

h h=2p ¼ ¼ hQ l l=2p

ð1:246Þ

where Q ¼ 2p/l. On the other hand, the momentum of the phonon is given byhk [141]. Thus, the following relationship must hold: hQ ¼ hk

ð1:247Þ

Since lattice vibrations are observed in the low-frequency region (l ﬃ 103 cm), Q ¼ 2p/l ﬃ 103 cm. The k value at the end of the Brillouin zone is k ¼ p/2a ﬃ 108 cm1. Thus, the k value that corresponds to the IR photon for lattice vibration is much smaller than the k value at the end of the Brillouin zone. This result indicates that the optical transitions we observe in IR spectra occur practically at k ﬃ 0. A similar conclusion can be derived for Raman spectra of lattice vibrations. Figure 1.46 shows the modes of lattice vibrations corresponding to various points on the dispersion curve shown in Fig. 1.45. At point A (acoustical mode), all atoms move in the same direction (translational motion of the whole lattice) and its frequency is zero. This is seen in Fig. 1.46a. In a three-dimensional lattice, there are three such modes. Thus, we subtract 3 from our calculations in factor group analysis (Sec. 1.28). On the other hand, at point B (optical branch), the two atoms move in opposite directions, but the center of gravity of the unit cell remains unshifted (Fig. 1.46b). Furthermore, the equivalent atom in each lattice moves in phase. If the two atoms carry opposite electrical charges, such a motion produces an oscillating dipole moment that can interact with incident IR radiation. Thus, it is possible to observe it optically. It should molecule pﬃﬃﬃﬃﬃﬃﬃﬃ be noted that the frequency of a diatomic p ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ in the free state is w ¼ f =m, whereas that of a diatomic lattice is w ¼ 2f =m (m ¼ reduced mass). At point C (optical branch), the lighter atoms are moving back and forth against each other while the heavier atoms are fixed (Fig. 1.46c). At point D (acoustical branch), the situation is opposite to that of point C. The vibrational modes in the middle of the Brillouin zone (points E and F) are shown in Figs. 1.46e and 1.46f, respectively. Thus far, we have discussed the lattice vibrations of a one-dimensional chain. The treatment of the three-dimensional lattice is basically the same, although more complicated [48]. If the primitive cell contains s molecules, each of which consists of N atoms, there are 3 acoustical modes and 3Ns – 3 optical modes. The latter is grouped into (3N – 6)s internal modes and 6s – 3 lattice modes. The general forms of

*The lattice vibration causes elastic waves in crystals. The quantum of die lattice vibrational energy is called “phonon,” in analogy with the photon of the electromagnetic wave.

POLARIZED SPECTRA OF SINGLE CRYSTALS

133

Fig. 1.46. Wave motions corresponding to various points on the dispersion curves.

the dispersion curves of such a crystal are shown in Fig. 1.47. In this book, our interest is focused on vibrational analysis of 3Ns–3 optical modes at k ﬃ 0. Examples are found in diamond and graphite (Sec. 2.14.3) and quartz (Sec. 2.15.2). 1.31. POLARIZED SPECTRA OF SINGLE CRYSTALS In the preceding section, the 30 normal vibrations of the Bravais cell of calcite crystal have been classified into symmetry species of the factor group D3d. The results (Table 1.26) show that three intramolecular (A2u þ 2Eu), three translatory lattice (A2u þ 2Eu) and two rotatory lattice (A2u þ Eu) vibrations are IR-active, whereas three intramolecular (A1g þ 2Eg), one translatory lattice (Eg), and one rotatory lattice (Eg) vibrations are Raman-active. In order to classify the observed bands into these symmetry species, it is desirable to measure infrared dichroism and polarized Raman spectra using single crystals of calcite.

134

THEORY OF NORMAL VIBRATIONS

Fig. 1.47. General form of dispersion curves for a molecular crystal [48].

1.31.1. Infrared Dichroism Suppose that we irradiate a single crystal of calcite with polarized infrared radiation whose electric vector vibrates along the c axis (z direction) in Fig. 1.42. Then the infrared spectrum shown by the solid curve of Fig. 1.48 is obtained [142]. According to Table 1.21, only the A2u vibrations are activated under such conditions. Thus, the three bands observed at 885(n), 357(t), and 106(r) cm1 are assigned to the A2u species. The spectrum shown by the dotted curve is obtained if the direction of polarization is perpendicular to the c axis (x,y plane). In this case, only the Eu vibrations should be infrared-active. Therefore the five bands observed at 1484(n), 706(n), 330(t), 182(t), and 106(r) cm1 are assigned to the Eu species. Here, n, t, and r denote intramolecular, translatory lattice, and rotatory lattice modes, respectively. 1.31.2. Polarized Raman Spectra Polarized Raman spectra provide more information about the symmetry properties of normal vibrations than do polarized infrared spectra [143]. Again consider a single

Fig. 1.48. Infrared dichroism of calcite [142].

POLARIZED SPECTRA OF SINGLE CRYSTALS

135

Fig. 1.49. Schematic representation of experimental condition y(zz)x.

crystal of calcite. According to Table 1.21, the A1g vibrations become Raman-active if any one of the polarizability components, axx, ayy, and azz, is changed. Suppose that we irradiate a calcite crystal from the y direction, using polarized radiation whose electric vector vibrates parallel to the z axis (see Fig. 1.49), and observe the Raman scattering in the x direction with its polarization in the z direction. This condition is abbreviated as y(zz)x. In this case, Eq. 1.66 is simplified to Pz ¼ azzEz because Ex ¼ Ey ¼ 0 and Px ¼ Py ¼ 0. Since azz belongs to the A1g species, only the A1g vibrations are observed under this condition. Figure 1.50c illustrates the Raman spectrum obtained with this condition. Thus, the strong Raman line at 1088 cm1 (n) is assigned to the A1g species. Both the A1g and Eg vibrations are observed if the z(xx)y condition is used. The Raman spectrum (Fig. 1.50a) shows that five Raman lines [1088(n), 714(n), 283(r), 156(t), and 1434(n) cm1 (not shown)] are observed under this condition. Since the 1088 cm1 line belongs to the A1g species, the remaining four must belong to the Eg species. These assignments can also be confirmed by measuring Raman spectra using the y(xy)x and x (zx)y conditions (Figs. 1.50b and 1.50d) . 1.31.3. Normal Coordinate Analysis on the Bravais Cell In the discussion above, we have assigned several bands in the same symmetry species to the v, t, and r types. In general, the intramolecular (v) vibrations appear above 400 cm1, whereas the lattice vibrations appear below 400 cm1. However, more complete assignments can be made only via normal coordinate analysis on the entire Bravais cell [144]. Such calculations have been made by Nakagawa and Walter [145] on crystals of alkali–metal nitrates that are isomorphous with calcite. These workers employed four intramolecular and seven intermolecular force constants. The latter are in the range of 0.12–0.00 mdyn/A. Figure 1.51 illustrates the vibrational modes of the 18 (27 if E modes are counted as 2) optically active vibrations together with the corresponding frequencies of calcite.

136

THEORY OF NORMAL VIBRATIONS

Fig. 1.50. Polarized Raman spectra of calcite [143].

1.32. VIBRATIONAL ANALYSIS OF CERAMIC SUPERCONDUCTORS In 1987, Wu et al. [146] synthesized a ceramic superconductor of the composition, YBa2Cu3O7d, whose superconducting critical temperature (Tc) was above the boiling point of liquid nitrogen (77 K). Since then, IR and Raman spectra of this and related compounds have been studied extensively, and the results are reviewed by Ferraro and Maroni [147,148]. Here, we limit our discussion to the Raman spectra of the superconductor mentioned above and their significance in studying oxygen deficiency and the structural changes resulting from it.

VIBRATIONAL ANALYSIS OF CERAMIC SUPERCONDUCTORS

137

Fig. 1.51. Vibrational modes of calcite. The observed and calculated (in parentheses) are listed under each mode [145].

138

THEORY OF NORMAL VIBRATIONS

Fig. 1.52. The Bravais unit cells of the orthorhombic and tetragonal forms of YBa2Cu3O7d [149].

The superconductor, YBa2Cu3O7d (abbreviated as the 123 conductor), can be obtained by baking a mixture of Y2O3, and BaCO3, and CuO in a proper ratio. The product is normally a mixture of an orthorhombic form (0 < d < 1) which is superconducting and a tetragonal form (d ¼ 1) that is an insulator. The Tc increases as d approaches 0. Figure 1.52 shows the Bravais unit cells of the orthorhombic (Pmmm ¼ D12h ) and tetragonal (P4/mmm ¼ D14h ) forms. The former is a distorted, oxygen-deficient form of perovskite. The orthorhombic unit cell contains 13 atoms, and their possible site symmetries can be found from the tables of site symmetries (Appendix X) as follows: 8D2h ð1Þ;

12C2v ð2Þ;

6Cs ð4Þ;

C1 ð8Þ

It is seen in Fig. 1.52 that the three atoms, Y, O(1), and Cu(1), are at the D2h sites. These atoms contribute 3B1u þ 3B2u þ 3B3u vibrations since Tx, Ty, and Tz belong to the B3u, B2u, and B1u species, respectively, in the D2h point group. On the other hand, the 10 atoms, 2Ba, 2Cu(2), 2O(2), 2O(3), and 2O(4) are at the C2v sites. As shown in Table 1.27, each pair of these atoms possess six degrees of vibrational freedom (2A1þ 2B1þ2B2), which are split into Ag þ B2g þ B3g þ B1u þ B2u þ B3u under D2h symmetry. The number of optical modes at k ﬃ 0 in each species can be obtained by subtracting three acoustical modes (B1u þ B2u þ B3u) from the above counting. Table 1.28 summarizes the results. It is seen that the orthorhombic unit cell has 15 Raman-active modes (5Ag þ 5B2g þ 5B3g) and 21 IR-active modes (7B1u þ 7B2u þ 7B3u). It should be noted that the mutual exclusion rule holds in this case since the D2h point group has a center of symmetry. Although the same results can be obtained by using factor group analysis [150], the correlation method is simpler and straightforward. Similar calculations on the tetragonal unit cell (YBa2Cu3O6) show that 10 vibrations (4A1g þ B1g þ 5Eg) are Raman-active while 11 vibrations (5A2u þ 6Eu) are IR-active under D4h symmetry.

139

VIBRATIONAL ANALYSIS OF CERAMIC SUPERCONDUCTORS

TABLE 1.28. Vibrational Analysis for Orthorhombic Form of YBa2Cu3O7d Using the Correlation Method D2h

D2h Y, O(1), Cu (1)

C2v Ba, Cu(2), O(2), O(3), O(4)

Acoustical Vib.

Total Optical Vib.

Activity

Ag B1g B2g B3g Au B1u B2u B3u

0 0 0 0 0 3 3 3

5 0 5 5 0 5 5 5

0 0 0 0 0 1 1 1

5 0 5 5 0 7 7 7

Raman Raman Raman Raman Inactive IR IR IR

9

30

3

36

Total

Figure 1.53 shows the Raman spectrum of a polished sintered pellet of the 123 conductor (d ¼ 0.3) obtained by Ferraro et al. [151]. The five Ag modes appear strongly, and the most probable assignments for these bands are [152,153]. 492 cm1 445 cm1 336 cm1 145 cm1 116 cm1

Axial motion of the O(4) atoms O(2)Cu(2)O(3) bending with the two O atoms moving in phase O(2)Cu(2)O(3) bending with the two O atoms moving out of phase Axial motion of the Cu(2) atoms Axial motion of the Ba atoms

Fig. 1.53. Backscattered Raman spectrum of a polished sintered pellet of YBa2Cu3O7d [151].

140

THEORY OF NORMAL VIBRATIONS

These values fluctuate within 10 cm1 depending on variations in oxygen stoichiometry and crystalline disorder. The results of normal coordinate analysis [154] indicate considerable mixing among the vibrations represented by the Cartesian coordinates of individual atoms. If the 123 conductor is prepared in pure oxygen and the oxygen is removed quantitatively by heating the sample in argon, a series of samples having d ¼ 0, 0.2, 0.5, and 0.7 can be obtained. The Tc values of these samples were found to be 94, 77, 50, and 20 K, respectively. Thomsen et al. [155] prepared a series of such samples, and measured their Raman spectra. Figure 1.54 shows a plot of vibrational frequencies of the four Ag modes mentioned above against d values. It is seen that the two modes at 502 and 154 cm1 are softened and the two modes at 438 and 334 cm1 are hardened as the oxygen is removed from the sample. These results

Fig. 1.54. Dependence of four Raman frequencies on oxygen concentration at 4 K and 298 K [155].

REFERENCES

141

seem to suggest that the Tc of the 123 conductor is related to oxygen deficiency in the O(2)Cu(2)O(3) layer. The effect of changing the size of the cation (Y) [156] and isotopic substitution (16O/18O) [157] on these Raman bands has also been studied. Thus far, IR studies on the 123 conductor have been hindered by the difficulties in obtaining IR spectra from highly opaque samples and in making reliableb and assignments [147,148].

REFERENCES Theory of Molecular Vibrations 1. G. Herzberg, Molecular Spectra and Molecular Structure, Vol. II: Infrared and Raman Spectra of Polyatomic Molecules, Van Nostrand, Princeton, NJ, 1945. 2. G. Herzberg, Molecular Spectra and Molecular Structure, Vol. I: Spectra of Diatomic Molecules, Van Nostrand, Princeton, NJ, 1950. 3. E. B. Wilson, J. C. Decius, and P. C. Cross, Molecular Vibrations, McGraw-Hill, New York, 1955. 4. C. J. H. Schutte, The Theory of Molecular Spectroscopy, Vol. I: The Quantum Mechanics and Group Theory of Vibrating and Rotating Molecules, North Holland, Amsterdam, 1976. 5. L. A. Woodward, Introduction to the Theory of Molecular Vibrations and Vibrational Spectroscopy, Oxford University Press, London, 1972. 6. J. D. Craybeal, Molecular Spectroscopy, McGraw-Hill, New York, 1988. 7. J. M. Hollas, Modern Spectroscopy, 2nd ed., Wiley, New York, 1992. 8. M. Diem, Introduction to Modern Vibrational Spectroscopy, Wiley, New York, 1993.

Symmetry and Group Theory 9. J. R. Ferraro and J. S. Ziomek, Introductory Group Theory and Its Application to Molecular Structure, 2nd ed., Plenum Press, New York, 1975. 10. D. C. Harris and M. D. Bertolucci, Symmetry and Spectroscopy, Oxford University Press, London, 1978. 11. P. R. Bunker, Molecular Symmetry and Spectroscopy, Academic Press, New York, 1979. 12. F. A. Cotton, Chemical Application of Group Theory, 3rd ed., Wiley-Interscience, New York, 1990. 13. R. L. Carter, Molecular Symmetry and Group Theory, Wiley, New York, 1992. 14. S. F. A. Kettle, Symmetry and Structure, 2nd ed., Wiley, New York, 2000.

Correlation Method 15. W. G. Fateley, F. R. Dollish, N. T. McDevitt, and F. F. Bentley, Infrared and Raman Selection Rules for Molecular and Lattice Vibrations: The Correlation Method, WileyInterscience, New York, 1972.

142

THEORY OF NORMAL VIBRATIONS

Vibrational Intensities 16. W. B. Person and G. Zerbi, eds., Vibrational Intensities in Infrared and Raman Spectroscopy, Elsevier, Amsterdam, 1982.

Fourier Transform Infrared Spectroscopy 17. J. R. Ferraro and L. J. Basile, eds., Fourier Transform Infrared Spectroscopy, Vol. I, Academic Press, New York, 1978 to present. 18. P. G. Griffiths, and J. A. de Haseth, Fourier Transform Infrared Spectrometry, Wiley, New York, 1986. 19. J. R. Ferraro and K. Krishnan, eds., Practical Fourier Transform Infrared Spectroscopy, Academic Press, San Diego, CA, 1990. 20. H. C. Smith, Fundamentals of Fourier Transform Infrared Spectroscopy, CRC Press, Boca Raton, FL, 1996.

Raman Spectroscopy 21. D. P. Strommen and K. Nakamoto, Laboratory Raman Spectroscopy, Wiley, New York, 1984. 22. J. G. Grasselli and B. J. Bulkin, eds., Analytical Raman Spectroscopy, Wiley, New York, 1991. 23. H. A. Szymanski, eds., Raman Spectroscopy: Theory and Practice, Vol. 1, Plenum Press, New York, 1967; Vol. 2 1970. 24. J. A. Koningstein, Introduction to the Theory of the Raman Effect, D. Reidel, Dordrecht (Holland), 1973. 25. D. A. Long, Raman Spectroscopy, McGraw-Hill, New York, 1977. 26. J. R. Ferraro, K. Nakamoto, and C. W. Brown, Introductory Raman Specctroscopy, 2nd ed., Academic Press, San Diego, CA, 2003. 27. E. Smith, Modern Raman Spectroscopy, Wiley, New York, 2004.

Low-Temperature and Matrix Isolation Spectroscopy 28. B. Meyer, Low-Temperature Spectroscopy, Elsevier, Amsterdam, 1971. 29. H. E. Hallarn, ed., Vibrational Spectroscopy of Trapped Species, Wiley, New York, 1973. 30. M. Moskovits and G. A. Ozin, eds., Cryochemistry, Wiley, New York, 1976.

Time-Resolved Spectroscopy 31. G. H. Atkinson, Time-Resolved Vibrational Spectroscopy, Academic Press, New York, 1983. 32. M. D. Fayer, ed., Ultrafast Infrared and Raman Spectroscopy, Marcel Dekker, New York, 2001.

High-Pressure Spectroscopy 33. J. R. Ferraro, Vibrational Spectroscopy at High External Pressures—the Diamond Anvil Cell, Academic Press, New York, 1984.

REFERENCES

143

Vibrational Spectra of Inorganic, Coordination, and Organometallic Compounds 34. J. R. Ferraro, Low-Frequency Vibrations of Inorganic and Coordination Compounds, Plenum Press, New York, 1971. 35. L. H. Jones, Inorganic Spectroscopy, Vol. I, Marcel Dekker, New York, 1971. 36. R. A. Nyquist and R. O. Kagel, Infrared Spectra of Inorganic Compounds, Academic Press, New York, 1971. 37. S. D. Ross, Inorganic Infrared and Raman Spectra, McGraw-Hill, New York, 1972. 38. E. Maslowsky, Jr., Vibrational Spectra of Organometallic Compounds, Wiley, New York, 1977.

Vibrational Spectra of Organic Compounds 39. N. B. Colthup, L. H. Daly, and S. E. Wiberley, Introduction to Infrared and Raman Spectroscopy, 3rd ed., Academic Press, San Diego, CA, 1990. 40. D. Lin-Vien, N. B. Colthup, W. G. Fateley, and J. G. Grasselli, The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press, San Diego, CA, 1991.

Vibrational Spectra of Biological Compounds 41. A. T. Tu, Raman Spectroscopy in Biology, Wiley, New York, 1982. 42. P. R. Carey, Biochemical Applications of Raman and Resonance Raman Spectroscopies, Academic Press, New York, 1982. 43. F. S. Parker, Application of Infrared, Raman and Resonance Raman Spectroscopy in Biochemistry, Plenum Press, New York, 1983. 44. T. G. Spiro, ed., Biological Applications of Raman Spectroscopy, Vols. 1–3, Wiley, New York, 1987–1988. 45. H.-U. Gremlich and B. Yan, eds., Infrared and Raman Spectroscopy of Biological Materials, Marcel Dekker, New York, 2001.

Vibrational Spectra of Adsorbed Species 46. A. T. Bell and M. L. Hair, eds., Vibrational Spectroscopies for Adsorbed Species, American Chemical Society, Washington, DC, 1980. 47. J. T. Yates, Jr. and T. E. Madey, eds., Vibrational Spectroscopy of Molecules on Surfaces, Plenum Press, New York, 1987.

Vibrational Spectra of Crystals and Minerals 48. G. Turrell, Infrared and Raman Spectra of Crystals, Academic Press, New York, 1972. 49. V. C. Farmer, The Infrared Spectra of Minerals, Mineralogical Society, London, 1974. 50. J. A. Gadsden, Infrared Spectra of Minerals and Related Inorganic Compounds, Butterworth, London, 1975. 51. C. Karr, ed., Infrared and Raman Spectroscopy of Lunar and Terrestrial Minerals, Academic Press, New York, 1975.

144

THEORY OF NORMAL VIBRATIONS

52. J. C. Decius and R. M. Hexter, Molecular Vibrations in Crystals, McGraw-Hill, New York, 1977.

Advances Series 53. Spectroscopic Properties of Inorganic and Organometallic Compounds, Vol. 1 to present, The Chemical Society, London. 54. Molecular Spectroscopy—Specialist Periodical Reports, Vol. 1 to present, The Chemical Society, London. 55. R. J. H. Clark and R. E. Hester, eds., Advances in Infrared and Raman Spectroscopy, Vol. 1 to present, Heyden, London. 56. J. Durig, ed., Vibrational Spectra and Structure, Vol. 1 to present, Elsevier, Amsterdam. 57. C. B. Moore, ed., Chemical and Biochemical Applications of Lasers, Vol. 1 to present, Academic Press, New York. 58. Structure and Bonding, Vol. 1 to present, Springer-Verlag, New York.

Index and Collection of Spectral Data 59. N. N. Greenwood, E. J. F. Ross, and B. P. Straughan, Index of Vibrational Spectra of Inorganic and Organometallic Compounds, Vols. 1–3, Butterworth, London, 1972– 1977. 60. B. Schrader, Raman/IR Atlas of Organic Compounds, VCH, New York, 1989. 61. Sadtler’s IR Handbook of Inorganic Chemicals, Bio-Rad Laboratories, Sadtler Division, Philadelphia, PA, 1996.

Handbooks and Encyclopedia 62. R. A. Ayquist, R. O. Kagel, C. I. Putzig, and M. A. Leugers, The Handbook of Infrared and Raman Spectra of Inorganic Compounds and Organic Salts, Vols. 1–4, Academic Press, San Diego, CA, 1996. 63. Encyclopedia of Spectroscopy and Spectrometry, 3-vol. set with online version Academic Press, San Diego, CA, 2000. 64. I. R. Lewis and H. G. M. Edwards, eds., Handbook of Raman Spectroscopy, Marcel Dekker, New York, 2001. 65. J. M. Chalmers and P. R. Griffiths, eds., Handbook of Vibrational Spectroscopy, Vols. 1– 5, Wiley, Chichester, UK, 2002.

General 66. 67. 68. 69. 70. 71. 72.

W. Holzer, W. F. Murphy, and H. J. Bernstein, J. Chem. Phys. 52, 399 (1970). C. F. Shaw, III, J. Chem. Educ. 58, 343 (1981). R. M. Badger, J. Chem. Phys. 2, 128 (1934). W. Gordy, J. Chem. Phys. 14, 305 (1946). D. R. Herschbach and V. W. Laurie, J. Chem. Phys. 35, 458 (1961). K. Nakamoto, M. Margoshes, and R. E. Rundle, J. Am. Chem. Soc. 77, 6480 (1955). E. M. Layton, Jr., R. D. Cross, and V. A. Fassel, J. Chem. Phys. 25, 135 (1956).

REFERENCES

73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.

89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106.

145

F. D. Hardcastle and I. E. Wachs, J. Raman Spectrosc. 21, 683 (1990). F. D. Hardcastle and I. E. Wachs, J. Phys. Chem. 95, 5031 (1991). D. P. Strommen and E. R. Lippincott, J. Chem. Educ. 49, 341 (1972). G. C. Lie, J. Chem. Educ. 56, 636 (1979). H. Eyring, J. Walter, and G. E. Kimball, Quantum Chemistry, 5th ed., Wiley, New York, 1949, p. 121. E. B. Wilson, J. Chem, Phys. 7, 1047 (1939); 9, 76 (1941). J. C. Decius, J. Chem. Phys. 16, 1025 (1948). T. Shimanouchi, J. Chem. Phys. 25, 660 (1956). T. Shimanouchi, “The Molecular Force Field,” in D. Henderson,ed., Physical Chemistry: An Advanced Treatise, Vol. 4, Academic Press, New York, 1970. T. Shimanouchi, J. Chem. Phys. 17, 245, 734, 848 (1949). D. F. Heath and J. W. Linnett, Trans. Faraday Soc. 44, 556, 873, 878, 884 (1948); 45, 264 (1949). J. Overend and J. R. Scherer, J. Chem. Phys. 32, 1289, 1296, 1720 (1960); 33, 446 (1960); 34, 547 (1961); 36, 3308 (1962). T. Shimanouchi, Pure Appl. Chem. 7, 131 (1963). J. H. Schachtschneider, “Vibrational Analysis of Polyatomic Molecules,” Pts. V and VI, Tech. Rept. 231-64 and 53-65, Shell Development Co., Emeryville, CA, 1964 and 1965. K. Nakamoto, Angew. Chem. 11, 666 (1972). N. Mohan, K. Nakamoto, and A. M€uller, “The Metal Isotope Effect on Molecular Vibrations,” in R. J. H. Clark and R. E. Hester,eds., Advances in Infrared and Raman Spectroscopy, Vol. 1, Heyden, London, 1976. K. Nakamoto, K. Shobatake, and B. Hutchinson, Chem. Commun. 1451 (1969). J. Takemoto and K. Nakamoto, Chem. Commun. 1017 (1970). J. R. Kincaid and K. Nakamoto, Spectrochim. Acta. 32A, 277 (1976). P. M. Champion, B. R. Stallard, G. C. Wagner, and I. C. Gunsalus, J. Am. Chem. Soc. 104, 5469 (1982). Y. Morino and K. Kuchitsu, J. Chem. Phys. 20, 1809 (1952). B. L. Crawford and W. H. Fletcher, J. Chem. Phys. 19, 141 (1951). P. LaBonville and J. M. Williams, Appl. Spectrosc. 25, 672 (1971). D. A. Ramsey, J. Am. Chem. Soc. 74, 72 (1952). D. P. Strommen and K. Nakamoto, Appl. Spectrosc. 37, 436 (1983). T. G. Spiro and T. C. Strekas, Proc. Natl Acad. Sci. USA. 69, 2622 (1972). W. M. McClain, J. Chem. Phys. 55, 2789 (1971). D. P. Strommen, J. Chem. Educ. 69, 803 (1992). W. Kiefer, Appl. Spectrosc. 28, 115 (1974). M. Mingardi and W. Siebrand, J. Chem. Phys. 62, 1074 (1975). A. C. Albrecht, J. Chem. Phys. 34, 1476 (1961). S. A. Asher, Anal. Chem. 65, 59A, 201A (1993). W. Kiefer and H. J. Bernstein, J. Raman Spectrosc. 1, 417 (1973). R. J. H. Clark and P. D. Mitchell, J. Am. Chem. Soc. 95, 8300 (1973).

146

THEORY OF NORMAL VIBRATIONS

L. A. Nafie, P. Stein, and W. L. Peticolas, Chem. Phys. Lett. 12, 131 (1971). Y. Hirakawa and M. Tsuboi, Science 188, 359 (1975). T. G. Spiro, R. S. Czernuszewicz, and X.-Y. Li, Coord. Chem. Rev. 100, 541 (1990). X.-Y. Li, R. S. Czernuszewicz, J. R. Kincaid, P. Stein, and T. G. Spiro, J. Phys. Chem. 94, 47 (1990). 110a. Y. Mizutani, Y. Uesugi, and T. Kitagawa, J. Chem. Phys. 111, 8950 (1999). 111. H. Hamaguchi, I. Harada, and T. Shimanouchi, Chem. Phys. Lett. 32, 103 (1975). 112. H. Hamaguchi, J. Chem. Phys. 69, 569 (1978). 113. F. Inagaki, M. Tasunai, and T. Miyazawa, J. Mol. Spectrosc. 50, 286 (1974). 114. P. Pulay, Mol. Phys. 17, 192 (1969); 18, 473 (1970); P. Pulay and W. Meyer, J. Mol. Spectrosc. 40, 59 (1971). 115. R. G. Parr and W. Yang, Density Functional Theory of Atoms and Molecules, Oxford University Press, Oxford, 1989; J. K. Labanowski and J. W. Andzelm, eds., Density Functional Methods in Chemistry, Springer, Berlin, 1990. S. M. Clark, R. L. Gabriel, and D. Ben-Amotz, J. Chem. Educ. 27, 654 (2000). 116. J. B. Foresman and A. Fisch, Exploring Chemistry with Electronic Structure Method, 2nd ed., Gaussian Inc., Pittsburgh, PA, 1996. 116a B. G. Johnson and M. J. Fisch, J. Chem. Phys. 100, 7429 (1994). 117. H. D. Stidham, A. C. Vlaservich, D. Y. Hsu, G. A. Guirgis, and J. R. Durig, J. Raman Spectrosc. 25, 751 (1994). 118. W. B. Collier, J. Chem. Phys. 88, 7295 (1988). 119. J. A. Boatz and M. S. Gordon, J. Phys. Chem. 93, 1819 (1989). 120. P. Pulay, J. Mol. Struct. 147, 293 (1995). 121. M. W. Wong, Chem. Phys. Lett. 256, 391 (1996). 122. G. A. Ozin, “Single Crystal and Gas Phase Raman Spectroscopy in Inorganic Chemistry,” in S. J. Lippard,ed., Progress in Inorganic Chemistry, Vol. 14, Wiley-Interscience, New York, 1971. 123. R. J. H. Clark and D. M. Rippon, J. Mol. Spectrosc. 44, 479 (1972). 124. E. Whittle, D. A. Dows, and G. C. Pimentel, J. Chem. Phys. 22, 1943 (1954). 125. M. J. Linevsky, J. Chem. Phys. 34, 587 (1961). 126. D. E. Milligan and M. E. Jacox, J. Chem. Phys. 38, 2627 (1963). 127. D. Tevault and K. Nakamoto, Inorg. Chem. 14, 2371 (1975). 128. S. P. Willson and L. Andrews, “Matrix Isolation Infrared Spectroscopy,” in J. M. Chalmers and P. R. Griffiths, eds., Handbook of Vibrational Spectroscopy, Vol.2, Wiley, London, 2002, p. 1342 . 129. B. Liang, L. Andrews, N. Ismail, and C. J. Marsden, Inorg. Chem. 41, 2811 (2002). 130. M. Zhou and L. Andrews, J. Phys. Chem. A 103, 2964 (1999). 131. D. Tevault and K. Nakamoto, Inorg. Chem. 15, 1282 (1976). 132. B. S. Ault, J. Phys. Chem. 85, 3083 (1981). 133. T. Watanabe, T. Ama, and K. Nakamoto, J. Phys. Chem. 88, 440 (1984). 134. K. Bajdor and K. Nakamoto, J. Am. Chem. Soc. 106, 3045 (1984). 135. G. Burns and A. M. Glazer, Space Groups for Solid State Scientists, Academic Press, New York, 1978, p. 81. 107. 108. 109. 110.

REFERENCES

147

136. J. M. Robertson, Organic Crystals and Molecules, Cornell University Press, Ithaca, NY, 1953, p. 44. 137. International Tables for X-Ray Crystallography, Knoch Press, Birmingham, UK, 1952. 138. R. S. Halford, J. Chem. Phys. 14, 8 (1946). 139. R. W. C. Wyckoff, Crystal Structures, Vols. I and II, Wiley-Interscience, New York, 1964. 140. S. Bhagavantam and T. Venkatarayudu, Proc. Indian Acad. Sci. 9A, 224 (1939).Theory of Groups and Its Application to Physical Problems, Andhra University, Waltair, India, 1951. 141. C. Kittel, Introduction to Solid State Physics, 5th ed., Wiley, New York, 1976, p. 118. 142. M. Tsuboi, Infrared Absorption Spectra, Vol. 6, Nankodo, Tokyo, 1958, p. 41. 143. S. P. Porto, J. A. Giordmaine, and T. C. Damen, Phys. Rev. 147, 608 (1966). 144. T. Shimanouchi, M. Tsuboi, and T. Miyazawa, J. Chem. Phys. 35, 1597 (1961). 145. I. Nakagawa and J. L. Walter, J. Chem. Phys. 51, 1389 (1969). 146. M. K. Wu, J. R. Ashburn, C. J. Torng, P. H. Hor, R. L. Meng, L. Gao, Z. J. Huang, Y. Q. Wang, and C. W. Chu, Phys. Rev. Lett. 55, 908 (1987). 147. J. R. Ferraro and V. A. Maroni, Appl Spectrosc. 44, 351 (1990). 148. V. A. Maroni and J. R. Ferraro, Practical Fourier Transform Infrared Spectrocopy, Academic Press, San Diego, CA, 1990, p. 1. 149. I. K. Schuller and J. D. Jorgensen, Mater. Res. Soc. Bull. 14(1), 27 (1989). 150. J. Hanuza, J. Klamut, R. Horyn, and B. Je_zowska-Trzebiatowska, J. Mol Struct. 193, 57 (1989). 151. J. R. Ferraro and V. A. Maroni, private communication. 152. M. Stavola, D. M. Krol, L. F. Schneemeyer, S. A. Sunshine, J. V. Waszczak, and S. G. Kosinski, Phys. Rev. B: Condens. Matter [3] 39, 287 (1989). 153. Y. Morioka, A. Tokiwa, M. Kikuchi, and Y. Syono, Solid State Commun. 67, 267 (1988). 154. F. E. Bates and J. E. Eldridge, Solid State Commun. 64, 1435 (1987). F. E. Bates, Phys. Rev. B39, 322 (1989). 155. C. Thomsen, R. Liu, M. Bauer, A. Wittlin, L. Genzel, M. Cardona, E. Sch€ onherr, W. Bauhofer, and W. K€onig, Solid State Commun. 65, 55 (1988). 156. M. Cardona, R. Liu, C. Thomsen, M. Bauer, L. Genzel, W. K€ onig, A. Wittlin, U. Amador, M. Barahona, F. Fernandez, C. Otero, and R. Saez, Solid State Commun. 65, 71 (1988). 157. B. Batlogg, R. J. Cava, A. Jarayaman, R. B. van Dover, G. A. Kourouklis, S. Sunshine, D. W. Murphy, L. W. Rupp, H. S. Chen, A. White, K. T. Short, A. M. Mujsce, and E. A. Rietman, Phys. Rev. Lett. 58, 2333 (1987).

Chapter

2

Applications in Inorganic Chemistry 2.1. DIATOMIC MOLECULES

*

As shown in Sec. 1.3, diatomic molecules have only one vibration along the chemical bond; its frequency in term of wavenumber ð~ nÞ is given by 1 ~ n¼ 2pc

sﬃﬃﬃﬃ K m

where K is the force constant, m the reduced mass and c is the velocity of light. In homopolar XX molecules (D¥h) thevibration is not infrared-active but is Raman-active, whereas it is both infrared- and Raman-active in heteropolar XY molecules (C¥n). 2.1.1. Vibrational Frequencies with Anharmonicity Corrections Tables 2.1a and 2.1b list a number of diatomic molecules and ions of the X2 and XY types, for which frequencies corrected for anharmonicity (oe) and anharmonicity constants (xeoe) are known. The force constants can be calculated directly from these oe values. In the X2 series, the oe values cover a wide range of frequencies, the highest being that of H2 (4395 cm1) and the lowest being that of Cs2 (42 cm1). Correspondingly, *

Hereafter, the word molecules is also used to represent ions and radicals.

Infrared and Raman Spectra of Inorganic and Coordination Compounds, Sixth Edition, Part A: Theory and Applications in Inorganic Chemistry, by Kazuo Nakamoto Copyright 2009 John Wiley & Sons, Inc.

149

150

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.1a. Harmonic Frequencies and Anharmonicity Constants of X2-Type Molecules (cm1)a,b Molecule H2 H2þ 7 Li2 23 Na2 39 K2 85 Rb2 133 Cs2 11 B2 Al2 12 C2 Si2 Ge2 Sn2 208 Pb2 14 N2 14 þ N2 15 N2 La2 Ce2 Pr2 Nd2 Gd2 Tb2 Lu2

oe 4395.2 4004.77 351.4 159.2 92.6 57.3 42.0 1051.3 297.5 1641.4 548.0 287.95 204.3 110.20 2358.54 2207.2 2278.79 236.0 245.4 244.9 148.0 138.7 137.6 121.6

xeoe

Ref.

Molecule

117.91 — 2.59 0.73 0.35 0.96 0.08 9.4 1.69 11.67 2.2 0.32 0.2 0.34 14.31 16.14 13.35

1 2 1 1 1 1 1 1 3 1 4 4a,4 4 5 6 1 6

31

0.9 — — 0.7 0.3 0.31 0.16

13 14 14 14 15 16 17

Hf2 Re2 (ex) Co2 Rh2 58 Ni2 Ag2

P2 As2 75 Asþ 2 Sb2 209 Bi2 16 O2 O2þ 32 S2 80 Se2 Te2 19 F2 35 Cl2 35 þ Cl2 79 Br2 127 I2 Xe2þ Y2 75

oe

xeoe

Ref.

780.4 429.4 (314.8) 269.9 172.7 1580.4 1876.4 725.68 391.8 251 916.93 559.71 645.3 325.43 214.6 124 184.4

2.80 1.12 (1.25) 0.59 0.32 12.07 16.53 2.85 1.06 0.55 11.32 2.70 2.90 1.10 0.61 0.5 0.30

1 1 1 1 1,7 1 1 1 1 1 8 9 1 10 1 11 12

>1 1.0 2.2 1.83 1.9 0.7

18 19 20 21 22 23

176.2 317.1 296.8 283.9 259.2 151.39

a The compounds are listed in the order of the periodic table beginning with the first atom: IA, IIA,. . ., IB, IIB, and so on. b Values in parentheses are uncertain. The oe and xeoe values are rounded off two decimals.

the xeoe value is largest for H2 (117.91 cm1) and smallest for Cs2 (0.08 cm1). Extensive studies on hydrides show that their frequencies span from 4395 cm1 (H2) to 891 cm1 (HCs). These tables also contain many isotopic frequencies involving light as well as heavy atoms.The effects of changing the mass and/or force constant in a series of diatomic molecules have been discussed in Sec. 1.3. Similar series are found in these tables. For example, we find that (all in units of cm1) Li2 ð351:4Þ > Na2 ð159:2Þ > K2 ð92:6Þ > Rb2 ð57:3Þ > Cs2 ð42:0Þ HBe ð2058:6Þ > HMg ð1495:3Þ > HCa ð1298:4Þ > HSr ð1206:9Þ > HBa ð1168:4Þ Across the periodic table, we find that HB ð 2366Þ < HC ð2860:4Þ < HN ð 3300Þ < HO ð3735:2Þ < HF ð4138:6Þ The frequency decreases on removal of bonding electrons: N2 ð2358:5Þ > Nþ 2 ð2207:2Þ;

As2 ð429:4Þ > Asþ 2 ð314:8Þ

DIATOMIC MOLECULES

151

TABLE 2.1b. Harmonic Frequencies and Anharmonicity Constants of XY-Type Molecules (cm1)a,b Molecule

oe

xeoe

Ref.

Molecule 138

HD DD H 6 Li H 7 Li D 6 Li D 7 Li H 23 Na D 23 Na H 39 K H 85 Rb D 85Rb H 87 Rb H Cs D 133 Cs H 9 Be H 24 Mg D 24 Mg H Ca H 88 Sr D 88 Sr

3817.1 3118.5 1420.12 1405.51 1074.33 1054.94 1171.76 845.97 985.67 937.10 666.66 936.98 891.25 632.80 2058.6 1495.26 1078.14 1298.40 1206.89 858.85

94.96 64.10 23.66 23.18 13.54 13.05 19.52 9.98 14.90 14.28 7.06 14.27 12.82 6.34 35.5 31.64 16.15 19.18 17.03 8.64

1 1 24–26 24,25 24,26 24 27,28 29 29,30 31 29 31 32 29 1 33 33 34 35 36

Ba H DBa HSc DSc HMn HCo HNi HCu H 107 Ag D 109 Ag H 197 Au H 64 Zn D 64 Zn H 110 Cd D 116 Cd H Hg DY D Cr H 11 B H 27 Al

D 27 Al H 69 Ga D 71 Ga H 115 In D 115 In H 205 Tl D 205 Tl H 12 C D 12 C H 28 Si H Ge H 120 Sn H 208 Pb H 14 N DN (H 14 N) HP H As H 209 Bi H 16 O

1211.77 1603.96 1142.77 1475.43 1048.60 1391.27 987.04 2860.4 2101.0 2042.5 1831.85 1655.49 1560.53 (3300) 2399.13 3191.5 2363.77 2076.93 1699.52 3735.2

15.06 28.42 14.42 25.16 12.70 23.10 11.67 64.11 34.7 35.67 32.86 28.83 28.79 — 42.11 85.6 43.91 39.22 31.93 82.81

49 50,51 44,51 52,53 44 54 44 55 1,56 57 58,59 60–62 63 1 64 65 64 66 67–69 1

D 16 O (D O) HS (H 32 S) H Se H 19 F D 19 F H 35 Cl D 35 Cl D 37 Cl H Br HI (H 132 Xe)þ 7 Li O 6 Li F 7 Li F 7 Li Cl 7 Li Br 7 Li I 23 Na K

106.6 536.1 366 302 258 426 281 213 212

0.46 3.83 2.05 1.50 1.08 2.4 1.30 0.80 0.70

1 80,81 82,83 82,84 82 85,86 82,83 82 1

Mg 35 Cl Mg 79 Br Mg 127 I 24 Mg 16 O Mg S 40 Ca 19 F Ca 35 Cl Ca 79 Br Ca 127 I

23

Na Rb Na 19 F 23 Na Cl 23 Na Br 23 Na I KF K Cl K Br KI 23

24 24

xeoe

Ref.

1168.43 829.84 1546.97 1141.27 1546.85 1858.79 2000 1940.4 1759.67 1250.68 2305.0 1615.72 1147.36 1461.13 1032.04 1387.1 1089.12 1183.20 (2366) 1682.4

14.61 7.37 — 12.38 27.60 — 40 37 33.93 17.21 43.12 59.62 28.72 61.99 29.29 83.01 9.85 15.60 (49) 29.11

37,38 36 39 39 40 41 42 42 43 44 1 45 36 1 36 1 46 47 1 48,49

2720.9 2723.5 2696,25 2647.1 2421.72 4138.55 2998.3 2991.0 2145.2 2141.97 2649.7 2309.5 2270.0 799.07 964.31 906.2 641.1 563.2 498.2 123.3

44.2 49.72 48.74 53.28 44.60 90.07 45.71 52.85 27.18 27.08 45.21 39.73 41.33 — 9.20 7.90 4.2 3.53 3.39 0.40

1,70 70 71 72 73 1 1 74 74,75 75 1 1 76 77 78 79,78 79 79 79 1

2.05 1.34 — 5.18 2.93 2.74 1.31 0.86 0.64

1 1 1 1,90,91 1 1,92 1 1 1

oe

465.4 373.8 [312] 785.1 525.2 587.1 369.8 285.3 242.0

(continued)

152

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.1b. (Continued (Continued ) Molecule

oe

xeoe

Ref.

Molecule

oe

xeoe

Ref.

— 1.90 0.92 1.63 0.75 (2.0) (1.2) 9.12 5.11 11.83 3.84

1 87,85 82 87 82 1 1,88 1 1 1 1,89

16

Ca O Sr 19 F Sr 35 Cl Sr 79 Br Sr 127 I Sr 16 O Ba 19 F 138 Ba 35 Cl Ba 79 Br Ba I 138 Ba O

732.03 500.1 302.3 216.5 173.9 653.5 468.9 279.3 193.8 152.16 669.73

4.83 2.21 0.95 0.51 0.42 4.0 1.79 0.89 0.42 0.27 2.02

93,94 1,95 1 1 1 1,96 1,97 1 1 98 99,90

3.95 2.45 2.23 2.99 1.20 3.70 4.07 1.23 6.3 4.61 3.45 4.41 4.9 4.22 — 6.5 3.01 1.4 0.9

1 1 1 1 1 1 1 1 1 1 1 100 1 101 102 1 1 1 1

55

Mn 16 O Fe 35 Cl Fe 16 O Co F Co Cl 193 Ir N Ni Cl 63 Cu 19 F 63 Cu 35 Cl 63 Cu 79 Br 63 Cu 127 I Cu 16 O 107 Ag 35 Cl 107 Ag 81Br 107 Ag I Ag O 197 Au 35 Cl Zn 19 F Zn 35 Cl

840.7 406.6 880 678.18 421.2 1126.18 419.2 622.7 416.9 314.1 264.8 628 343.6 247.7 206.2 493.21 382.8 (630) 390.56

4.89 1.2 5 2.74 0.74 6.29 1.04 3.95 1.57 0.87 0.71 3 1.16 0.68 0.43 4.10 1.30 (3.5) 1.55

1 1 1 103 1 104 1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1 105 1 106,107 1,108 1 105 109 1,110 1

27

Al 16 O Al S 69 Ga F 69 Ga 35 Cl 69 Ga 81 Br 69 Ga 127 I Ga As (gr) Ga As (ex) Ga O In As 115 In F 115 35 In Cl 115 81 In Br 115 127 In I In 16 O Tl 19 F Tl 35 Cl Tl 81Br Tl 127 I (CF)þ

978.23 623.13 622.37 365.0 263.0 216.4 215 152 767.7 182.7 535.36 317.41 221.0 177.1 703.1 475.0 287.5 192.18 150 1792.76

7.12 6.22 3.30 1.11 0.81 0.5 3 2.89 6.34 0.5 2.67 1.01 0.65 0.4 3.71 1.89 1.24 0.39 — 13.23

1,111 112 113 1 1 1 114 114 1 115 116 1 1 1 1 1,117 1 1 1 118

133

Cs Rb 85 Rb 19 F Rb Cl 132 Cs 19 F Cs Cl 133 Cs Br 133 Cs 127 I 9 Be 19 F 9 Be 35 Cl 9 Be 16 O 24 Mg 19 F

49.4 373.44 228 352.62 209 (194) 142 1265.5 846.65 1487.3 717.6

45

Sc 16 O Y 16 O 129 La 16 O Ce 16 O 141 Pr 16 O Gd 16 O Lu 16 O Yb Cl 48 Ti 35 Cl 48 Ti 16 O 90 Zr 16 O 180 Hf N V 16 O Cr F Cr Cl Cr 16 O 55 Mn F 55 Mn 35 Cl 55 Mn Br

971.6 852.5 811.6 865.0 818.9 841.0 841.7 293.6 456.4 1008.4 936.6 932.72 1012.7 664.10 396.66 898.8 618.8 384,9 289,7

Zn Br 64 Zn 127 I Cd 19 F Cd 35 Cl Cd Br Cd 127 I Hg 19 F Hg 35 Cl 202 Hg 81 Br Hg 127 I Hg Tl 11 19 B F 11 35 B Cl 11 81 B Br 11 14 B N 11 16 B O Al F 27 35 Al Cl 27 79 Al Br 27 127 Al I

(220) — 223.40 0,75 (535) — 330.55 1,2 230,0 0,50 178.56 0.63 490.8 4.05 292.60 1.60 186.39 0.98 125.65 1.09 26.9 0.69 1402.16 11.82 839.1 5.11 684.15 3.73 1514.6 12.3 1885.4 11,77 802.32 4.85 481.4 2.03 378.02 1.28 316.1 1.0

89

DIATOMIC MOLECULES

153

TABLE 2.1b. (Continued (Continued ) Molecule þ

oe

xeoe

Ref.

Molecule 28

(C Cl) 12 35 C Cl C Br 12 14 C N 12 31 C P 12 16 C O (gr) C O (ex) ð12 C 16 OÞþ 12 32 C S 12 C Se Si C (Si F)þ 28 19 Si F 28 Si 35 Cl ½28 Si 35 Clþ Si Br 28 Si 14 N 28 16 Si O (Si O)þ 28 32 Si S

1177.7 846 727.99 2068.7 1239.75 2170.23 1743.76 2214.24 1285.1 1036.0 964.77 1050.47 856.7 535.4 678.24 425.4 1151.7 1242.0 1162.18 749.5

6.65 1.0 3.95 13.14 6.86 13.46 14.57 15.16 6.5 4.8 5.60 4.95 4,7 2.20 — 1.52 6.56 6.05 6.97 2.56

119 1 120 1 1 1 121 1 1,122 1 123 124 1,125 1 126 1 1 1 127 1

Si Se Si Te (79Ge F)þ Ge F 74 Ge 35 Cl Ge Br 74 Ge 16 O 74 Ge 32 S 74 Ge 74 Ge 130 Te Sn 19 F Sn 35 Cl Sn Br Sn 16 O 120 Sn 32 S Sn Se Sn Te Pb 19 F Pb 35 Cl Pb 79 Br

Pb 127 I Pb 16 O 208 Pb 32 S Pb Se Pb Te 14 16 N O 14 NS 14 N Br 14 31 N P 14 75 N As 14 N Sb PF P Cl 31 16 P O P S. (As 35 Cl)þ 75 As 16 O Sb 209 Bi Sb 19 F Sb 35 Cl

160,53 721.8 428.1 277.6 211.8 1903.9 1220.0 693 1337.2 1068.0 942.0 846.7 551.4 1230.66 739.1 527.7 967.4 220.0 614.2 369.0

0.25 3.70 1.20 0.51 0.12 13.97 7.75 5.0 6.98 5.36 5.6 4.49 2.23 6.52 2.79 1.74 5.3 0.50 2.77 0.92

1 1 1 1 1 1,134 1 1 1,135 1 1 136 137 1 138 139 1,140 1 1,141 1,141

Sb 14 N Sb 16 O Bi N Bi P Bi As 209 19 Bi F 209 35 Bi Cl 209 79 Bi Br 209 127 Bi I Bi O OF 16 O Cl 16 O Br 16 OI 16 32 O S Se 16 O Te 16 O 19 35 F Cl 19 F Br Cl Br

610.26 384.2

3.14 1.47

150,151 1

127 79

IF 127 35

I Cl

28

I Br

oe

xeoe

Ref.

580.0 481.2 815.60 666.5 407.6 296.6 985.7 575.8 402.66 323.4 582.9 352.9 247.7 822.4 487.23 331.2 259.5 507.2 303.8 207.55

1.78 1.30 3.22 3.15 1.36 0.9 4.30 1.80 0.87 1.0 2.69 1.06 0.62 3.73 1.35 0.74 0.50 2.30 0.88 0.50

1 1 128 129 1 1 1,130 1,131 132 1 1 1 1 1 133 1 1 1 1 1

942.05 817.2 736.57 430.28 283.30 510.8 308.0 209.3 163.9 688.4 1053.45 (780) 713 (687) 1123.7 907.1 796.0 783.5 662.3 442.5

5.6 5.30 4.83 1.53 0.73 2.05 0.96 0.47 0.31 4.8 10.23 — 7 (5) 6.12 4.61 3.50 4.95 3.80 1.5

1 1 142 143 143 1,144 1 1 1 145 146 1 1,147 1 1 1 1 148 148 149

268.68

0.82

152

a The compounds are listed in the order of the periodic table begining with the first atom: IA, IIA,. . ., IB, IIB, and so on. b Values in parentheses are uncertain, and those in square brackets are observed frequencies without anharmonicity correction. The oe and xeoe values are rounded off to two decimals; (gr) and (ex) indicate the ground and excited states, respectively.

154

APPLICATIONS IN INORGANIC CHEMISTRY

while the opposite trend is found when antibonding electrons are removed: COþ ð2214:2Þ > CO ð2170:2Þ;

SiFþ ð1050:5Þ > SiF ð856:7Þ;

GeFþ ð815:6Þ > GeF ð666:5Þ Table 2.1b also presents two examples for which vibrational frequencies were reported for both ground and excited electronic states: CO ðX1 Sþ ; ground stateÞ; 2170:2 > CO ð3 P; excited stateÞ; 1743:8 GaAs ðX3 S ; ground stateÞ; 215 > GaAsð3 P0þ ; excited stateÞ; 152 Most of the oe and xeoe values listed in Tables 2.1a and 2.1b were obtained mainly by analysis of rotational–vibrational fine structures of infrared or Raman spectra in the gaseous phase [1]. Their values can also be obtained by the analysis of overtone progression in resonance Raman spectra (Sec. 1.22). Vibrational spectra of unstable (or metastable) molecules were obtained in inert gas matrices (Sec. 1.25). Anharmonicity corrections are limited largely to diatomic molecules because nine parameters are required even for nonlinear triatomic molecules. 2.1.2. Vibrational Frequencies without Anharmonicity Corrections Table 2.1c lists the observed frequencies of homopolar diatomic molecules, ions, and radicals. Combining these data with those given in Table 2.1a, the following frequency trends are immediately obvious: 2 Oþ 2 > O2 > O2 > O2 ;

Brþ 2 > Br2 ;

S2 > S 2;

Clþ 2 > Cl2 > Cl2 ;

Iþ 2 > I2 > I2

As mentioned earlier, these orders can be explained in terms of simple molecular orbital theory. Table 2.1d shows the relationship among various molecular parameters in the O2þ > O2 > O2 > O22þ series. It is seen that, as the bond order decreases, the bond distance increases, the bond energy decreases, and the vibrational frequency decreases. Although the force constant is not rigorously related to the bond energy (Sec. 1.3), there is an approximate linear relationship between these parameters. Thus, these frequency trends provide valuable information about the nature of diatomic ligands in coordination compounds (see Chapter 3 in Part B). The O2þ ion was found in compounds such as O2þ [MF6] (M ¼ As, Sb, etc.) and O2þ [M2F11] (M ¼ Nb, Ta, etc.) [171b], whereas the O2 ion was observed in a triangular MO2 type complex formed by the reaction of alkali metals with O2 in an Ar matrix [171c]. The n(O2) of simple metal superoxides (Mþ O2) and peroxides (M2þ O22) have been measured [171d]. Linear relationships are noted between n(S2) and SS bond distances for sulfur compounds [171e]. Table 2.1e lists vibrational frequencies of heteropolar diatomic molecules. It contains isotopic pairs such as (HO/DO), (6 Li 35 C1=7 Li 35 C1), and (28 Si 32 S= 28 Si 34 S).

DIATOMIC MOLECULES

155

TABLE 2.1c. Observed Frequencies of X2 Molecules (cm1) Molecule T2b C2 C22 Sn2 Pb2 P2 As2 16 O 2d O2 O22 S2 S2

Statea

~n

Liquid Gas Solidc Mat Mat Gas Gas Mat Mat Solid Mat Solide

2458 1757,8 1845 188 112 775 421 1555.6 1097 794, 738 718 623

Ref.

Molecule

153 154 155 156 156 157 157 158 159 160 161 162

Se2 F2

35

Cl 2 Cl 2 Br2þ Br2 I2þ I2 I2 K2 Ag2 Zn2 Cd2 35

State e

Solid Mat Mat Mat Sol’n Gas Sol’n Gas Mat Mat Mat Mat Mat

~n

Ref.

349 475 546 247 360 319 238 213 115 91.0 194 80 58

162 163 164 165 166 167 168 167 169 169 170 171 171

a

Mat ¼ inet gas matrix. T ¼ 3H (tritium). c Na salt. d 16 O17 O, 1528.4 cm1; e Doped in KCl. b

16

O18 O, 1508.3 cm1.

Charge effects on vibrational frequencies are seen for ðHOÞ > HO > ðHOÞþ ;

HS > ðHSÞþ ;

ðNOÞþ > NO > ðNOÞ > ðNOÞ2

Most of these frequencies were measured in inert gas matrices at low temperatures. For example, the SH radical was produced by photodecomposition of H2S in Ar or Kr matrices [182], and me FeH radical was produced via the reaction of Fe metal vapor with H2 in Ar matrices [187] (Sec. 3.26). 2.1.3. Dimers and Polymers The N4þ ion is linear and centrosymmetric in the ground state, and exhibits a strong IR absorption at 2237,6 cm1 in Ne matrices [252], and at 2234.51 cm1 in the gaseous state [253]. Surprisingly, stable AsF6 salt of the N5þ ion was synthesized, and its TABLE 2.1d. Relationship between Various Molecular Parameters in the O2 and Its Ions Species O2þ O2 O2 O22 a

Bond Order

Bond ˚ )a Distance (A

Bond Energy (Kcal/mole)

2.5 2.0 1.5 1.0

1.123 1.207 1.280 1.49

149.4 117.2 — 48.8

Ref. 171a. Na salt. c Solid-state splitting. b

~n (O2) (cm1) 1876 1580 1094 791/736b,c

force Constant ˚) (mdyn/A 16.59 11.76 5.67 2.76

156

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.1e. Observed Frequencies of XY Molecules (cm1) Molecule

Statea

~n

Ref.

Molecule

State

~n

Ref.

H Al H Si

DS (H 32 S)þ (D 32 S)þ HF H Cr

Mat Mat Gas Mat Gas Mat Mat Solid Gas Gas Solid Mat Gas Mat Gas Gas Gas Mat

1593 1967 1971.04 3131.6 2077.00 3548.2 2616.1 3637.4 3555.59 2956.37 2225 2607 2591 1847.8 2547.2 1829.1 3961.42 1548

172 173 174 175 176 177 177 178 179 180 181 182 183 183 184 184 185 186

D Cr H Fe H Rh 7 Li O Li F 6 35 Li Cl 7 35 Li Cl 6 37 Li Cl 7 37 Li Cl Li Br Li I Na F Na 35 Cl Mg F 26 Mg O Mg O 40 Ca O Ti O

Mat Mat Mat Mat Mat Gas Gas Gas Gas Mat Mat Mat Gas Mat Mat Gas Mat Mat

1112 1767.0 1920.6 745 885 686.1 643.0 683.3 640.1 510 433 515 364.7 738 815.4 774.74 707.0 1005

186 187 187 188 189 190 190 190 190 191 191 189 192 193 194 195 196 197

Ti N Zr O Nb N Nb O Ta O WO Fe C 56 Fe O Ru O Co O 58 Ni O Th N UN UO Pu N Pu O Zn O Cd O 11 BO BF

Mat Mat Mat Mat Mat Mat Gas Mat Mat Mat Mat Mat Mat Mat Mat Mat Mat Mat Gas Gas

1037 975 1002,5 968 1020 1050.9 852.14 873.1 834.4 846.4 825.7 934.6 995 820 855.7 820 802 719 1861.92 1378.7

198 197 199 199 200 201 202 203 204 203 203 205 206 207 208 209 210 210 211 212

Al O Al F Tl F 205 35 Tl Cl Tl Br Tl I CF CO CS (C N) CN 28 Si Ge 28 Si Sn 29 Si N 28 Si O 28 32 Si S 28 34 Si S 74 Ge O 74 Ge S

Mat Mat Mat Mat Mat Mat Mat Gas Mat Mat Sol’n Mat Mat Mat Gas Mat Gas Gas Mat Mat

975 793 441 284.7 179 143 1279 1286.13 2138.4 1274 2080 2046 422.5 360 1972.80 1225.9 749.65 739.30 973.4 566.6

213 214 215 216 215 215 217 218 219 220 221 222 223 223 224 225 226 226 227 228

74

Ge 80 Se Ge Te

Mat Mat

397.9 317.67

228 228

FO Cl N

Mat Mat

240 239

Sn O Sn S 120 Sn 80 Se Pb O Pb S 208 Pb 32 S Pb 80 Se (N O)þ NO

Mat Mat Mat Mat Mat Gas Mat Solid Mat

229 228 228 230 228 231 228 232 233

Cl O Br N Br O S Cl S Br SI Se Te (Cl F)þ 79 Br O

Mat Mat Mat Mat Mat Mat Mat Sol’n Gas

1028.7 818.5 825 850.7 691 729.9 617 518 443 313 819 723.4

HN H As HO DO (H O) (H O)þ (T O) HS

74

120

816.1 480.5 325.2 718.4 423.1 426.6 275.1 2273 1880

241 239 242 243 243 243 244 245 246

DIATOMIC MOLECULES

157

TABLE 2.1b. (Continued ) 2.1e. (Continued Molecule

(N O) (N O)2 PN SO (S Se) FN a

Statea Mat Mat Mat Mat Solid Mat

~n 1358–1374 886 1323 1101.4 464 1117

Ref. 234 235 236 237 238 239

Molecule

(Br O) Br F 79 Br 35 Cl IN IF I Cl I Br 79

State

~n

Ref.

Sol’n Gas Gas Mat Gas Gas Gas

618 669.9 444.3 590 610.2 381 265

247 248 248 249 250 251 251

Mat ¼ inert gas matrix.

V-shaped (C2v) structure was confirmed by X-ray analysis. Complete assignments of the IR and Raman spectra of the N5þ ion have been based on this structure [254,255]. Figure 2.1 shows the low-temperature Raman spectrum of solid N5þAsF6. A new allotropic form of nitrogen in which all nitrogen atoms are connected by single covalent bonds was produced at temperatures of >2000 K and pressures of >110 GPa [256]. The Raman spectra shown in Fig. 2.2 were obtained at 110 GPa using a diamond anvil cell. At 300 K, a sharp n(NN) peak is seen at 2400 cm1, which is replaced by a new peak at 840 cm1 when the temperature is raised to 1990 K. The latter frequency is close to the n(NN) of N2H4 The possibility of forming bicyclic N10 and fullerene-like N60 has been suggested on the basis of theoretical calculations [257]. In inert gas matrices, a variety of aggregates of diatomic molecules are produced, depending on the experimental conditions employed. For example, the Raman spectra of H2, HD, and D2 in Ar matrices suggest the formation of aggregates of more than six molecules after annealing above 35 K [258].

Fig. 2.1. Structure of the N5þ ion (a) and low-temperature Raman spectrum (b) of solid N5þAsF6. The bands between 700 and 250 cm1 are due to the AsF6 ion [254,255].

158

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.2. Raman spectra of polymeric nitrogen (cubic gauche structure) before and after heating at 110 GPa. The starting sample (300 K, bottom) is in the molecular phase as indicated by the peak at 2400 cm1. The bands at 1550–1300 cm1 are due to the stressed diamond adjacent to the sample. Laser heating to 1990 K (top) transforms it to the polymeric phase with a prominent peak at 840 cm1 [256]. 1GPa ¼ 104 atm.

In an Ar matrix, Kþ(O4) takes the trans structure:

and exhibits n(O¼O) (Ag), n(O¼O) (Bu), n(OO) (Ag), and d(O¼OO) at 1291.5, 993, 305, and 131 cm1, respectively [259,260]. The structure of the O4 molecule is predicted to be either quasisquare D4h or triangular D3h (similar to the CO32þ and NO3 ions) [262]. The IR spectra of the O4 ion in Ne matrices were interpreted as a mixture of a rectangular form (2949 and 1320 cm1) and a bent-trans form (2808 and 1164 cm1) [262a]. The [Cl4]þ ion in solid [Cl4]þ[IrF6] salt assumes a rectangular structure with

TRIATOMIC MOLECULES

159

two short ClCl (1.94 A) and two long ClCl (2.93 A) bonds. The Raman peaks at 578,175, and 241 cm1 were assigned to the n(ClCl), n(ClCl) and d((ClCl)), respectively [262b]. The Raman spectra of (O2)4 (D4h symmetry) at pressure of 32.8 and 33.4 GPa exhibit 4 A1g vibrations at 1600, 1380, 342, and 16 cm1 [262c]. Assignments of the IR spectra of (HF)2 (HF)(DF), and (DF)2 in Ar matrices have been based on the structure [263]

This work has been extended to higher-molecular-weight polymers, (HF)3 and (HF)4 [264]. Matrix IR spectra of (LiF)n (where n ¼ 2,3,4) were interpreted in terms of ring structures (rhombic D2h, D3h, and D4h for n ¼ 2, 3, and 4, respectively) [265]. The Raman spectrum of (NaI)2 in the gasous state at 1084 K supports D2h structure [266]. The matrix IR. spectra of (GeO)2 [267] and (KO)2 [268] suggest a rhomic structure (D2h) and a slightly out-of-plane rhombic structure (C2v), respectively. The nitric oxide dimer, (NO)2, in Ar matrices exists as the cis form (1862 and 1768 cm1) or the trans form (1740 cm1), with the n(NO) shown in brackets [268a]. The IR spectrum of the cis dimer in the gaseous state has been assigned [268b]. IR spectra of cis and trans isomers of (NO)2þ ion produced in Ne matrices exhibit the n(NO) at 1619 and 1424 cm1, respectively. The corresponding frequencies of the (NO)2 ion produced in Ne matrices are at 1225 (cis) and 1227 cml(trans) [268c]. In the gaseous phase, NO forms a weakly bound van der Waals dimer, and the intermolecular vibrations are observed at 429(da), 239(ds), 134 (intermolecular stretch) and 117 cm1(out-of-plane torsion) [268d]. 2.1.4. van der Waals Complexes Inert gases and hydrogen halides form very weakly interacting van der Waals complexes. As expected, the n(HF) of NeHF [269] and n(HCl) of Ne–HCl [270] are shifted only slightly (þ0.4722 andþ3024 cm1, respectively) from their frequencies in the free state. Vibrational frequencies of analogous Ar complexes are also reported [271–275]. IR spectra of van der Waals complexes such as COCH4 [276], COOCS [277] and COBF3 [278] are available. 2.2. TRIATOMIC MOLECULES The three normal modes of linear X3(D¥h)- and YXY(D¥h)-type molecules were shown in Fig. 1.11; n1 is Raman-active but not infrared-active, whereas n2 and n3 are infrared-active but not Raman-active (mutual exclusion rule). However, all three

160

APPLICATIONS IN INORGANIC CHEMISTRY

ν 1(Σ+)

X

Y

Z

ν (XY)

ν 2(II)

δd (XYZ)

ν 3(Σ+)

ν (YZ)

Fig. 2.3. Normal modes of vibration of linear XYZ molecules.

vibrations become infrared- as well as Raman-active in linear XYZ-type molecules (C¥v), shown in Fig. 2.3. The three normal modes of bent X3(C2v) and YXY(C2v)-type molecules were also shown in Fig. 1.12. In this case, all three vibrations are both infrared- and Raman-active. The same holds for bent XXY(Cs)- and XYZ(Cs)-type molecules. Section 1.12 describes the procedure for normal coordinate analysis of the bent YXY-type molecule. In the following, vibrational frequencies of a number of triatomic molecules are listed for each class of compounds. 2.2.1. XY2-Type Halides Table 2.2a lists the vibrational frequencies of XY2-type halides. Most of these data were obtained in inert gas matrices. It is seen that the majority of compounds follow the trend n3 > n1, and that exceptions occur for some bent molecules. Although the structures of these halides are classified either as linear or bent, it should be noted that the bond angles of the latter range from 95 to 170 [315]. Thus, some bent molecules are almost linear. In principle, the distinction between linear and bent structures can be made by IR/Raman selection rules. For example, the n1 is IR-active for the bent structure but IR-inactive for the linear structure. However, such a simple criterion may lead to conflicting results [319]. 0 The bond angle (a) of the YXY molecule can be calculated if the n 3 of the YX0 Y molecule is available. Here, X0 is an isotope of X. Using the G and F elements given in Appendix VII, the following equation can readily be derived: ~0 2 MX MX 0 þMY ð1cosaÞ n3 ¼ ~ n3 MX 0 MX þMY ð1cosaÞ Here M denotes the mass of the atom subscripted. Figure 2.4 shows the spectra of NiF2 in Ne and Ar matrices obtained by Hastie et al. [320]. Using the n3 values

TRIATOMIC MOLECULES

161

TABLE 2.2a. Vibrational Frequencies of XY2-Type Metal Halides (cm1) Compounda

Structure

BeF2 BeCl2 BeBr2 BeI2 MgF2 MgCl2 MgBr2 MgI2 40 CaF2 40 Ca 35 Cl2 CaI2 (g) 86 SrF2 88 Sr 35 Cl2 BaF2 BaCl2 69 GaCl2 InCl 2 CF2 35 C Cl2 CBr2 SiF2 28 35 Si Cl2 28 SiBr2 GeF2 GeCl2 SnF2 SnCl2 SnBr2 PbF2 PbCl2 PbBr2 (g) NF2 PF2 PCl2 PBr2 O35 Cl2 32 SF2 SCl2 SBr2 SI2 76 Se 19 F2 SeCl2 (sol’n) SeBr2 (sol’n) TeCl2 (g) KrF2 (g) XeF2 Xe35Cl2 63 CuF2 [CuCl2] (s) [CuBr2] (s)

Linear Linear Linear Linear Linear Linear Linear Linear Bent Linear Linear Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Linear Linear Linear Bent Linear Linear

~n1

~n2

(680) (390) (230) (160) 550 327 198 148 484.8 — — 441.5 269.3 389.6 255.2 373.0 328 1102 719.5 595.0 851.5 513 402.6 692 398 592.7 354 237 531.2 297 200 1069.6 834.0 452.0 369.0 641.97 838 518 405 368 701.8 372 261 377 449 497 — — 302 194

345 250 220 (175) 249 93 82 56 163.4 63.6 43.3 82.0 43.7 (64) — — 177 668 — — (345) — — 263 — 197 (120) 84 165 — 64 573.4 — — — — 355 208 — — — 168 110 125 233 213.2 — 183.0 108 81

~n3 1555 1135 1010 873 842 601 497 445 553.7 402.3 292.3 443.4 299.5 413.2 260.0 415.1 291 1222 745.7 640.5 864.6 502 399.5 663 373 570.9 334 223 507.2 321 — 930.7 843.5 524.8 410 686.59 817 526 418 376 674.8 346 221 — 596,580 555 314.1 743.9 407 326

Ref. 279 279 279 279 280 280 280 280 281 282 283 281 282 281 282 284 285 286 287 288 289 290 291 292 293 294 293 295 294 293 296 297 298 299 299 300 301 302 302 302 302a 303 303 304 305 306 307 308 308 308

(continued)

162

APPLICATIONS IN INORGANIC CHEMISTRY

2.2b. (Continued (Continued ) TABLE 2.2a. Compounda

[CuI2] (s) [AgCl2] (s) [AgBr2] (s) [AgI2] (s) [AuCl2] (s) [AuBr2] (s) [AuI2] (s) ZnF2 ZnCl2 ZnBr2 ZnI2 (g) CdF2 CdCl2 CdBr2 CdI2 HgF2 HgCl2 HgBr2 HgI2 TiF2 VF2 CrF2 CrCl2 MnF2 MnCl2 MnI2 (g) FeF2 FeCl2 CoF2 CoCl2 NiF2 NiCl2 NiBr2

Structure Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Linear Bent Bent Linear Linear Linear Linear Linear Linear Linear Bent Linear Bent Linear Bent

~n1 148 268 170 133 329 209 158 596 352 223 163 555 (327) (205) (149) 568 (348) (219) (158) 665 — 668 — — — — — — — — — (350) —

~n2 65 88 61 49 120, 112 79, 75 67, 59 150 103,100 71 67.6 121 88 62 (50) 170 107 73 63 180 — 125 — 124.8 83 54.5 141.0 88 151.0 94.5 139.7 85 —

~n3

Ref.

279 333 253 215 350 254 210 754 503 404 337.5 660 419 319 269 642 405 294 237 766 733.2 749 493.5 700.1 476.8 324.2 731.3 493.2 723.5 493.4 779.6 520.6 331.9

309 309 309 309 310 310 310 311 312 312 283 311 313 313 313 311 313 313 313 314 315 316 317 315 317 283 315 317 315 317 315 317 318

a All data were obtained in inert gas matrices except those for which the physical state is indicated as g (gas), sol’n (solution), or s (solid).

obtained for isotopic NiF2, the FNiF angle was calculated to be 154–167 . The bond angle depends on the nature of the matrix gas employed. For example, NiCl2 is linear in Ar matrices, but bent (130 ) in N2 matrices [321]. Thus, structural data obtained in gas matrices must be interpreted with caution. As another example, the bond angle of CaF2 increases from 139 to 143 to 156 on changing the matrix gas from Kr to Ne to N2 [322]. The structure of the dimeric species (MX2)2 is known to be cyclic–planar:

TRIATOMIC MOLECULES

163

Fig. 2.4. IR spectra(n3) of NiF2 in Ne and Ar matrices. Matrix splitting, indicated by asterisks, is present in the Ne matrix spectrum [320].

although an exception is reported for (GeF2)2 [324]:

Mixed fluorides such as LiNaF2 and CaSrF4 take planar ring structures in inert gas matrices [325]:

2.2.2. XY2-Type Oxides, Sulfide, and Related Compounds Table 2.2b lists the vibrational frequencies of triatomic oxides, sulfides, selenides, and related compounds. Most of these data were obtained in inert gas matrices. Although the NO2 (nitrite) ion is bent, the NO2þ (nitronium) ion is linear in most compounds. Solid N2O5 consists of the NO2þ and NO3 ions, and the former is slightly

164

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.2b. Vibrational Frequencies of XY2-Type Oxides and Related Compounds (cm1) Compounda O Li O O Na O OBO Al O Al Ga O Ga O C O (g) 16 12 17 O C O (g) 16 12 18 O C O (g) (O C O)þ (O C O) (O C O)2 S C S (g) (S C S)þ (S C S) Se C Se O 120 Sn O (O N O) ONO (O N O)þ (s) OSO (O S O) O 80 Se O O Te O FOF Cl O Cl (s) (O Cl O) (s) O Cl O (O Cl O)þ (s) 16 79 O Br 16 O Br O Br (s) Br O Br (O Br O)þ (O Br O) OIO O Ce O O Tb O O Th O O Ru O O Pr O O 48 Tl O O Ta O 16 189 O W 16 O (ONiO)2 (s) OUO (O U O)2þ (s) O Pu O (N B N)3 (s) 14 N U 14 N 14 N U 15 N

Structure Bent Bent Linear Bent Bent Linear Linear Linear Linear Bent Bent Linear Linear Linear Linear Bent Bent Bent Linear Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Bent Linear Linear Linear Linear Linear Linear Linear

~n1 — 1080.0 — 715 472 (1337)b 1333.37 1314.49 — 1253,8 1187 658 — — 369.1 — 1327 1325 1396 1147 985.1 992.0 831.7 925/915 630.7 790 944.8 1055 795.7 504 532.9 865 710 768.0 757.0 758.7 787.4 926 — — 971 1030.2 732.5 (765.4) 880 — 1078 (1008.3) 987.2

~n2 243.4 390.7 — (120) — 667 670.12 667.69 462.6 714.2 745 397 — — 313.1 — 806 752 570 517 495.5 372.5 294 461 296.4 400 448.7 520 317.0 197 — 375 320 — — — — — — — — 380 140 — 140 — 611 — —

~n3 730 332.8 1276 994 809.4 2349 2386.69 2378.45 1421.7 1658,3 1329 1533 1207.1 1159.4 1301.9 863.1 1286 1610 2360 1351 1041.9 965.6 848.3 821 670.8 840 1107.6 1295/1230 845.2 587 628.0 932 — 800.3 736.8 718.8 735.3 902.2 730.4 934.8 912 983.8 910 776.1 950 794.3 1702 1051.0 1040.7

Ref. 188 326 327 328 329 330 331 331 332 332 333 334 335 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 347 354 355 355 355 356 355 357 358 359 359a 360 361 362 363 364 364

TRIATOMIC MOLECULES

165

TABLE 2.2b. (Continued (Continued ) Compounda 11

3

(P B P) (s) (P Be P)4 (s) P Ga P (P Zn P)4 (s) (P Cd P)4 (s) (P Hg P)4 (s) (P Be P)4 (s) (As Be As)4 (s) (Sb Be Sb)4 (s) S Ti S S Zr S S Hf S a

Structure

~n1

~n2

Linear Linear Bent Linear Linear Linear Linear Linear Linear Bent Bent Bent

490 379 322 344 317 335 379 222 — 533.5 502.9 492.2

379/385 340 — 47 81 71 340 310 255 — — —

~n3 1076 860 220.9 415 355 341 860 775 700 577.8 504.6 483.2

Ref. 365 366 367 368 358 368 366 366 366 369 369 369

All data were obtained in inert gas matrices except for those denoted as g (gas) or s (solid). Fermi resonance with 2 n2 (Sec. 1.11).

b

bent, as indicated by the appearance of a strong d(ONO) band at 534 cm1 in the Raman spectrum [370]. Metal dioxides produced by sputtering techniques [371] in inert gas matrices take linear or bent OMO structures. Metal dinitrides produced by the same technique also take linear NMN (M ¼ U [372] and Pu [373]) or bent NMN (M ¼ Th [374]) structures. These structures are markedly different from those of molecular oxygen and nitrogen complexes of various metals produced by the conventional matrix cocondensation technique (Chapter 3 in Part B). The NpO22þ, PuO22þ, and AmO22þ ions are similar to the UO22þ ion it that they are linear and symmetric. They exhibit the n3 bands at 969, 962, and 939 cm1, respectively, in IR spectra [375]. The relationship between the U¼O stretching force constant and the U¼O distance was derived by Jones [376]. According to Bartlett and Cooney [377] the U¼O distance R (in pm ¼ 102 A) is given by R ¼ 9141ð~ n3 Þð2=3Þ þ 80:4 R ¼ 10650ð~ n1 Þð2=3Þ þ 57:5 where n is in units of cm1. Mizuoka and Ikeda [378] observed that the n3 of the (UIVO2)2þ ion at 895 cm1 in DMSO (dimethyl sulfoxide) solution is shifted to 770 cm1 when reduced electrochemically to the (UVO2)þ ion. Vibrational spectra of coordination compounds containing dioxo(O¼M¼O) groups are discussed in Chapter 3 (in Part B). AT-shaped structure of C2v symmetry was proposed for SiC2 in an Ar matrix at 8K [379]. Assignments of IR spectra of BC2 [380] and KO2 [381] prepared in Ar matrices were based on triangular structures of C2v symmetry. The linear CS2 molecule in a N2 matrix is converted to a cyclic structure [881.3 cm1(A1) and 520.9 cm1 (B2)] on photo irradiation (193 nm) [382].

166

APPLICATIONS IN INORGANIC CHEMISTRY

2.2.3. Triatomic Halogeno Compounds Table 2.2c lists the vibrational frequencies of triatomic halogeno compounds. The resonance Raman spectrum of the I3 ion gives a series of overtones of the n1 vibration [408,409]. The resonance Raman spectra of the I2Br and IBr2 ions and their complexes with amylose have been studied [410]. The same table also lists the vibrational frequencies of XHY-type (X,Y: halogens) compounds. All these species are linear except the ClHCl ion, which was found to be bent by an inelastic neutron scattering (INS) and Raman spectral study [402].

TABLE 2.2c. Vibrational Frequencies of Triatomic Halogeno Compounds (cm1) Compounda þ

FCIF FCIF (m) FCIF (s) FCICIþ (s) FBrFþ (s) FBrF (s) FIF (s) CIFCIþ (s) Cl 3 ðmÞ CIBrCl (sl) ClICIþ (s) ClICI (sl) ClIIþ (s) Brþ 3 ðsÞ Br 3 ðslÞ Br3 (m) BrIBrþ (s) BrII (s) BrIIþ (s) I 3 ðcÞ HFHþ (s) FHF (s) FHF (m) FHF (g) CIHF (s) CIHCI (s) CIHBr (s) ClHI (s) BrHF (s) BrHBr (s) BrHBr (m) IHF (s) IHI (m) F 3 ðmÞ a

Structureb b b l b b l l b b l b l b b l l l l b l b l l l l b l l l l l l l l

n~1c 807 500 510/478 744 706 460 445 529 253 278 371 269 360 293 164 190 256 117 258 116 2970 596/603 — 583.1 275 199 — — 220 126 168 180 (120.7) 461

n~2

n~3c

Ref.

387 242 — 299/293 362 236 206 293/258 — 135 147 127 (126) 124 53 — 124 84 — 56/42 1680 1233 1217 1286.0 863/823 660/602 508 485 740 1038 — 635 — —

830 578/570 636 535/528 702 450 462 586/593 340 225 364 226 197 297 191 — 256 168 198 125 3080 1450 1364 1331.2 2710 1670 1705 2200 2900 1420 670 3145 682.1 550

383 384 385 386 387 388 388a 389 390 391 392 391 393 394 391 395 393 396 393 396a 397 398 399 400 401 402 403 403 401 404 405 401 406 407

m ¼ insert gas matrix; sl ¼ solution; s ¼ solid. b ¼ bent; l ¼ linear. c For XYZ type, n1 and n3 correspond to n (XY) and n (YZ), respectively. b

TRIATOMIC MOLECULES

167

Ault and coworkers have utilized salt-molecule reactions to produce a number of novel triatomic and other anions in isolated environments [411]. For example, the linear symmetric [FHF] ion was produced in Ar matrices via codeposition of CsF vapor with HF gas diluted in Ar [399]: Csþ þ F þ HF ! Csþ ½F H F The reaction of CsF vapor with HC1 gas diluted in Ar yields a mixture of two types of the [FHC1] anions; in type I, the H atom is not equidistant from the F and Cl atoms, and its n3 [mainly n(HF)] is observed near 2500 cm1, whereas in type II, the H atom is symmetrically located and its n3, is at 933 cm1 [412].

The yield of each type is determined by the location of the cation in the ion pair, and the type I/type II ratio varies, depending on the cation used; smaller alkali metal cations favor type II, while larger cations prefer type I. These classifications were originally made by Evans and Lo, who observed crystalline HCl2 salts of types I and II depending on the cation employed [413]. 2.2.4. Bent XH2 Molecules Table 2.2d lists the vibrational frequencies of bent XH2-type molecules. The XH stretching frequencies are lower and the XH2 bending frequencies are higher in condensed phases than in the vapor phase because of hydrogen bonding in the former. This trend is also seen for H2O and H2S dissolved in organic solvents such as pyridine and dioxane [442,443]. In both liquid water and ice, H2O molecules interact extensively via OHO bonds. However, there are marked differences between the two phases. In the latter, H2O molecules are tetrahedrally hydrogen-bonded, and this local structure is repeated throughout the crystal. In liquid water, however, the OHO bond distance and angle vary locally, and the bond is sometimes broken. Thus, its vibrations cannot be described simply by using the three normal modes of the isolated H2O molecule. According to Walrafen et al. [444] an isosbestic point exists at 3403 cm1 in the Raman spectrum of liquid water obtained as a function of temperature, and the bands above and below this frequency are mainly due to non-hydrogen-bonded and hydrogenbonded species, respectively. In addition, liquid water exhibits librational and restricted translational modes that correspond to rotational and translational motions of the isolated molecule, respectively. The librations yield a broad contour at 1000– 330 cm1, while the restricted translations appear at 170 and 60 cm1 [445]. For more details, see the review by Walrafen [446].

168

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.2d. Vibrational Frequencies of Bent XH2-Type Molecules (cm1) Molecule

State

~n1

~n2

~n3

Ref.

CaH2 TiH2 ZrH2 VH2 CrH2 MoH2 NiH2 AlH2 GaH2 InH2 SiH2 GeH2 NH2

Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Gas Matrix Matrix Surface Solid Solid Matrix Gas Liquid Solid Gas Solid Gas Solid Gas Gas Solid Gas Matrix Solid Gas Solid Gas Gas Gas Gas

1267.0 1483.2 — 1532.4 1651.3 1752.7 2007 1766 1727.2 1548.6 1995.03 1887 — 3290 3270 2230 3182.7 3657 3450 3400 2727 2416 2671 2495, 2336 2657 — 1988 2615 2619.5 2532, 2523 1892 1843, 1835 2345 1687 1692.14 2065.27

— 496 625.03 529 — — — 760 740.1 607.4 998.62 920 1499 1610 1556 1080 1401.7 1595 1640 1620 1402 1490 1178 1210 1169 996 1023 1183 — 1186, 1171 934 857, 847 1034 741 — —

1192.0 1435.5 1861.0 1508.3 1614.9 1727.4 1969 1799 1799.5 1615.5 1992.82 1864 3220 3380 3323 2160 3219.5 3756 3615 3220 3707 3275 2788 2432 2764 2370 2104 2627 2632.6 2544 2000 1854 2358 1697 2351.19 2072.11

414 415 414 415 416 416 417 418 419 419 420 421 422 423 424 425 426,427 428 429 430 428 430 428 430 431 432 433 434 435 436 437 436 438 438,439 440 441

[NH2] [PH2] [OH2] þ 16 OH2 16

OHDa

16

OD2

18

OD2 OT2b SH2

SD2 SeH2 SeD2 80 SeHDa 130 TeH2 a b

Here n1 and n3 denote n (OD) and n (OH), respectively. T ¼ 3H (tritium).

Hornig et al. [430] assigned the spectrum of ice I as shown in Table 2.2d. They also assigned some librational and translational modes. Bertie and Whalley [447] studied the vibrational spectra of ice in other phases, and found the spectra to be consistent with reported crystal structures in each phase. For crystal water and aquo complexes, see Sec. 3.8. The vibrational frequencies of H2O, D2O, HDO, and their dimers in argon matrices have been reported [448]. Infrared studies on (H2O)n (16 O=18 O) (where n ¼ 1,2,>3) in N2 matrices [449] and (H2O)n (n ¼ 2–6) formed in droplets of liquid helium have been reported [450]. Devlin at al. [451] measured the IR spectra of ice consisting of large (H2O)n

169

TRIATOMIC MOLECULES

(n ¼ 100–64,000) clusters and studied the relationship between the frequency and intensity of the bending mode and the structure of the cluster. 2.2.5. X3-Type Molecules X3-type halogeno molecules are listed in Table 2.2c. Table 2.2e lists vibrational frequencies of other X3-type molecules. The O3 and O3 ion are bent (bond angle 120–110 ), and the S3, S3, and S32 ions arc all bent (103–115 ), with frequencies decreasing in this order. Detailed assignments of the IR spectra were made for six isotopomers of O3 containing 16 O and 18 O in liquid oxygen solution at 77 K [454]. The N6 radical anion was produced by the reaction of the N3 ion with N3 . The most probable structure, on the basis of theoretical calculations is a rectangular cyclic structure of D2h symmetry with two long NN bonds that connect two azidyl monomeric units via the terminal nitrogen atoms. This is also supported by the observation of an IR band at 1842 cml, which is intermediate between N3 and N6þ [469]. Figure 2.5 illustrates three normal modes of vibration of a triangular X3- type molecules of D3h symmetry. The n1 is a symmetric stretch (A10 , Raman-active) while a degenerate pair of n2a and n2b (E 0 , IR/Raman-active) involves stretching as well as bending coordinates (Sec. 1.9).

TABLE 2.2e. Vibrational Frequencies of X3-Type Molecules (cm1) Bent (C2v) and Linear (D¥h) Molecules Molecule

State

Structure

~n1

~n2

~n3

Ref.

(O3) 16 O3 (S3)2 S3 S3 N3 N3þ Rh3 Mo3

Solid Matrix Solid Solution Gas Solution Matrix Matrix Matrix

Bent Bent Bent Bent Bent Linear Linear Bent Bent

1020 1104.3 458 535 581 1343 1287 322.4 386.8

620 699.8 229 235.5 281 637.5 473.7 248 224.5

816 1039.6 479 571 680 2048 1657.5 259 236.2

452 453,454 455 456 457,458 459 460 461 462

Triangular (D3h) Molecules

Molecule

State

~n1 (A0 1)

~n2 (E 0 )

Ref.

Rb3 Cs3 Cr3 Fe3 Ag3 Ru3a Ta3a Nb3 Au3

Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix

53.9 39.5 432.2 249 119 303.4 251.2 334.9 172

38.3 24.4 302.0 — — — — 227.4 118

463 463 464 465 465 466 466 467 468

a

D3h or C2v.

170

APPLICATIONS IN INORGANIC CHEMISTRY

ν1(A1′)

ν2b(E ′)

ν2a(E ′)

Fig. 2.5. Normal modes of vibration of triangular X3 molecules.

2.2.6. XXY- and XYZ- Type Molecules Table 2.2f lists vibrational frequencies of XXY-type molecules. Tables 2.2g and 2.2h list those of linear and bent XYZ-type molecules, respectively. It should be noted that the description of vibrational modes such as n(XX), n(XY), and n(YZ) is only qualitative. Most of these frequencies were measured in inert gas matrices. Some of these vibrations split into two because of Fermi resonance, matrix effect, or crystal, field effects. The linear HArF molecule listed in Table 2.2g was obtained by photodissociation of HF in Ar matrices followed by annealing to increase the HArF concentration below 20 K. Figure 2.6 shows the IR spectra of H 40 ArF, D 40 ArF, and H 36 ArF in Ar matrices

TABLE 2.2f. Vibrational Frequencies of XXY-Type Molecules (cm1) Moleculea 6

6

Li Li O CCO (C C O) (N N H)þ (g) NNC N N O (g) (N N F)þ (s) O O H (g) O O D (g) OOF O O Cl O O Br OOS SSH S S O (g) 35 Cl 35 Cl O a

Structure

~n1b

~n2c

Linear Linear Linear Linear Linear Linear Linear Bent Bent Bent Bent Bent Bent Bent Bent Bent

— 1074 2082 2257.9 1235 2277 2371 1097.6 1121.5 1500 1477.8 1484.0 1005 595 679.1 374.2

118 381 656 686.8 394 596.5 391 1391.8 1020.2 586.4 432.4 — — 903 382 240.2

~n3d 1028.5 1978 1185 3334.0 2824 1300.3 1057 3436.2 2549.2 376.0 214.9 — 739.9 2460/2463 1166.5 961.9

All data were obtained in inert gas matrices except for those indicated as g (gas) or s (solid). n (XX). c d (XXY). d n (XY). b

Ref. 470 471 472 473 474 475 475a 476 476 477 478 479 480 481 482 483

TRIATOMIC MOLECULES

171

TABLE 2.2g. Vibrational Frequencies of Linear XYZ-Type Molecules (cm1) Molecule HCN DCN TCN HNC H 11 B 79 Br H 11 B O D 11 B O (H C S)þ H Si N H 40 Ar F H Kr F 19 11 32 F B S FCN FPO F P 32 S 35 Cl 12 C 14 N Cl C N 35 Cl 11 B 32 S 35 Cl 11 B O (ClCO)þ Cl C S 35 Cl C P Br C N ICN Li N C Na C N (C N O) NCO (N C O) (N C 32 S) (N C Se) (N C Te) (N S O) (P C S) PNO (O C S)þ (O C S) O C Te SCO S C Se S C Te Al O Si WCH Co C O 6

a b

State

~n1 (XY)

~n2 (d)

~n3 (YZ)

Ref.

Gas Matrix Gas Gas Matrix Gas Gas Matrix Gas Gas Matrix Matrix Matrix Gas Gas Gas Gas Gas Matrix Gas Gas Solid Matrix Gas Gas Matrix Gas Matrix Matrix Solid Matrix Solution Solid Solid Solid Solid Solid Matrix Matrix Matrix Matrix Gas Gas Solution Matrix Gas Matrix

3311 3306 2630 2460 3620 3813.4 2849,4 (2849) 2253.5 3141,7 2152.2 1969.5 1950 — 1077 819.6 803.3 747.8a 718 529.9 — 802.5 632.1 573.9 574 575 485 722.9 382 2096 1275 2155 2053 2070 2073 1283 1762 865.2 2071.1 1646.4 1965,3 859 1435 1347 534.9 1006 579.2

712 721 569 513 477 448.5 — 754 608.4 766.5 — 687.0 650 (457) 449 — (313.6) 381.9a 384/387b — — — — 303.4 342.5 349/351 304 121.7 170 471 487 630 486/471 424/416 366 528 — — — — — 520 (355) (377) — 660 424.9

2097 — 1925 1724 2029 2049.8 937.6 1817 1647.7 — 1162.2 435,7 415 1644.4 2290 1297.5 726.3 2249.3a — 1407.6 1972.2 2256 1189.3 1477.3 2200 — 2188 2080.5 2044 1106 1922 1282/1202b 748 558 450 999 747 1754.7 702.5 718.2 — 2062 506 423 1028.2 — 1957.5

484 485 484 485a 486 487 488 489 490 491 492 493, 493a 494 495 496 497 498 499 500 501 502 503 504 505 506 500 507 508 509 510 511 512 513 514 515 516 517 518 519 519 520 521 522 522 523 524 525

Corrected for anharmonicity. Fermi resonance between n3 and 2n2.

172

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.2h. Vibrational Frequencies of Bent XYZ-Type Molecules (cm1) Molecule

Statea

~n1 (XY)

HCO H 12 C F HNO H 14 N F HON HOF H O Cl H O Br HOI 7 16 28 Li O Si NOF NSF N S Cl ONF O N 35 Cl O N Br ONI OPF

Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Gas Gas Gas Gas Gas Matrix Gas Matrix Gas Matrix Matrix Gas Matrix Matrix Gas Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix Matrix

2483 — 3450 — 3467.2 3537.1 3578 3589 3597 668.4 1886.2 1372 1325 1843.5 1799.7 1799.0 1809 1297.5 1292,2 1272.8 1855 1880 1037.7 720.2 712 765.4 917 742 509.4 1327.3 328 295 — 1312.9 434 229.8

O P Cl OCF O C Cl O Cl F SPF S P Br FNO F N Cl 35 Cl C F 35 Cl 28 Si 16 O 35 Cl 32 S N Cl Sn Br Cl Pb Br Br O N Br 32 S N Br S F Pd 14 N 16 O a

~n2 (d) 1087 1405 1110 1432 — 1359.0 1239 1164 1103 249 734.9 366 273 765.8 595.8 542,0 470 — 416,0 491.6 626 281 310 313.6 — 519.6 (341) — — 403.8 — — 350 346.1 — 522.1/510.3

~n3 (YZ)

Ref.

1863 1181.5 1563 1000 1095.6 886.0 728 626 575 998.5 492.2 640 414 519.9 331.9 266.4 216 819.6 811.4 — 1018 570 596.9 791.4 372 1844.1 720 1146 1160.9 267.4 228 200 1820,0 226.2 765 1661.3

526 527 528 529 530 531 532 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 555 556 554 557 558

Matrix: inert gas matrix.

at 7.5 K obtained by Khriachtchev et al. [493, 493a]. Good agreement was obtained between the observed and theoretical. frequencies. This work was extended to HKrF [494]. Lapinski et al. [559] measured the IR spectra of 12 isotopomers of linear NNO involving 14 N=15 N and 16 O=17 O=18 O in N2 matrices, and assigned their fundamental, overtone, and combination bands on the basis of observed isotope shifts and DFT calculations. These 12 isotopomer bands in the n1 (NN stretch) region are marked by arrows in Fig. 2.7.

PYRAMIDAL FOUR-ATOM MOLECULES

Fig. 2.6. IR spectra of HArF in solid Ar showing the effects of [493,493a].

40

173

Ar=36 Ar and H/D substitutions

2.3. PYRAMIDAL FOUR-ATOM MOLECULES 2.3.1. XY3 Molecules (C3v) The four normal modes of vibrations of a pyramidal XY3 molecule are shown in Fig. 2.8. All four vibrations are both infrared- and Raman-active. The G and F matrix elements of the pyramidal XY3 molecule are given in Appendix VII. Table 2.3a lists the fundamental frequencies of XH3-type molecules. Several bands marked by an asterisk are split into two by inversion doubling. As is shown in Fig. 2.9, two configurations of the XH3 molecule are equally probable. If the potential barrier between them is small, the molecule may resonate between the two structures. As a result, each vibrational level splits into two levels (positive and negative) [560]. Transitions between levels of different signs are allowed in the infrared spectrum, whereas those between levels of the same sign are allowed in the Raman spectrum. The transition between the two levels at v ¼ 0 is also observed in the microwave region (~ n ¼ 0:79 cm1 ).

174

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.7. IR spectra (n1 region) of NNO in N2 matrices at 10 K. The labeling on the left indicates the isotopic composition of the mixture used in preparing the matrix. For example, 4568 denotes a mixture composed of equal fraction of 14N,15N,16O, 16O, and 18O. Thus, the peak due to 14 14 N, N,16O is indicated by the number 446 (bottom trace). This species exhibits the bands at 2235.6, 588.0, and 1291.2 cm1 in N2 matrices [559].

Splittings due to inversion doubling were also observed for the n4 of 14 NH3 [578], and and n1 and n3 of ND3 [579]. Optical isomers may be separated if the three Y groups are not identical and the inversion barrier is sufficiently high. As is seen in Table 2.3a n1 and n3 overlap or are close in most compounds. The presence of the hydronium (H3Oþ) ion in hydrated acids has been confirmed by observing its characteristic frequencies. For example, it was shown from infrared spectra that H2PtCl62H2O should be formulated as (H3O)2 [PtCl6] [580]. For normal

X Y1

Y3 Y2 ν1(A1) νs(XY)

ν2(A1) δs(YXY)

ν3a(E) νd(XY)

ν4a(E) δd(YXY)

Fig. 2.8. Normal modes of vibration of pyramidal XY3 molecules.

175

PYRAMIDAL FOUR-ATOM MOLECULES

TABLE 2.3a. Vibrational Frequencies of Pyramidal XH3 Molecules (cm1)

Solid

2318

15

Gas

3335

NT3

Gas

2016

PH3

Gas

2327

PD3 AsH3 AsD3 SbH3 SbD3 BiH3 ½OH3 þ SbCl 6 ½OH3 þ ClO 4 þ ½OH3 NO3 [OH3]þ

Gas Gas Gas Gas Gas Gas Sol’n Solid Solid Liquid SO2

1694 2122 1534 1891 1359 1733.3 3560 3285 2780 3385

~n2 (A1) 931:6 a 968:1 1060 748:6 a 749:0 815 926 a 961 647 990 a 992 730 906 660 782 561 726.7 1095 1175 1135 —

[OD3]þ

Liquid SO2

2490

—

[SD3]þ [SeH3]þ [CH3] Naþ SiH3 [GeH3]

Gas Solid Matrix Gas Liquid NH3

1827.2 2302 2760 — 1740

— 936 1092 727.9 809

Molecule

State

NH3

Gas

ND3

NH3

a

~n1 (A1)

Solid

3335:9 3337:5 3223

Gas

2419

a

~n3 (E)

~n4 (E)

Ref.

3414

1627.5

560

3378

1646

561

2555

1191.0

560

2500

1196

561

3335

1625

562

2163

1000

563

2421

1121

564

(1698) 2185 — 1894 1362 1734.5 3510 3100 2780 3470 3400 2660 2580 1838.6 2320 2805 — —

806 1005 714 831 593 751.2 1600 1577 1680 1700 1635 1255

564 564,565 564,566 567 567 568 569 570 671 572

— 1057 1384 — 886

573 574 575 576 577

572

Inversion doubling.

coordinate analysis of pyramidal XH3 molecules, see Refs. [581–583]. Infrared spectra of hydrated proton complexes Hþ(H2O)n (n ¼ 2–11), in water clusters were measured as Ar complexes at low temperatures, and band assignments were made based on their minimum-energy structures obtained by DFT calculation [583a]. Reaction of CH3I with Na (or K) atoms in N2 matrices at 20 K [575] produces an ion-pair complex CH3 Naþ (or CH3 Kþ), of C3v symmetry. The n(NaCH3) and n(KCH3) vibrations were observed at 298 and 280 cm1, respectively. Vibrational spectra of the (NH3)2 dimer in Ar matrices show that two NH3 molecules are bonded via a very weak hydrogen bond [584]. Table 2.3b lists the vibrational frequencies of pyramidal XY3 halogeno compounds. Clark and Rippon [601] have measured the Raman spectra of a number of these compounds in the gaseous phase. The compounds show a n1 > n3 and n2 > n4 trend, whereas the opposite trend holds for the neutral XH3 molecules listed in Table 2.3a. In some cases, the two stretching frequencies (n1 and n3) are too close to be distinguished empirically. This is also true for the two bending bands (n2 and n4).

176

APPLICATIONS IN INORGANIC CHEMISTRY

Y

Y Y

Y

X

X

Y

Y _ +

IR

IR

R

υ=1

R

_ + υ=0 Fig. 2.9. Inversion doubling of pyramidal XY3 molecules.

Figure 2.10 shows the Raman spectra of the SnX3 ions (X ¼ Cl, Br, and I) in solution obtained by Taylor [594]. In crystalline [As(Ph)4][SnX3], the symmetry of the SnX3 ion is lowered to Cs so that n3 and n4 each split into two bands [592]. Normal coordinate analyses on the SnX3 ion (X ¼ F,Cl,Br,I) [592] have been carried out. Table 2.3c lists the vibrational frequencies of pyramidal XO3-type compounds. Rocchiciolli [629] has measured the infrared spectra of a number of sulfites, selenites, chlorates, and bromates. Again, the n3 and n4 vibrations may split into two bands because of lowering of symmetry in the crystalline state. Although n2 > n4 holds in all cases, the order of two stretching frequencies (n1 and n3) depends on the nature of the central metal. It is to be noted that the SO3 and SO32 anions are pyramidal, while the SO3 molecule is planar (Table 2.4b). Figure 2.11 shows the ER spectra of KClO3 and KIO3 obtained in the crystalline state. 2.3.2. ZXY2 (Cs) and ZXYW (C1) Molecules Substitution of one of the Yatoms of a pyramidal XY3 molecule by a Z atom lowers the symmetry from C3v to Cs Then the degenerate vibrations split into two bands, and all six vibrations become infrared- and Raman-active. The relationship between C3v and Cs is shown in Table 2.3d. Table 2.3e lists the vibrational frequencies of pyramidal ZXY2 molecules. Vibrational spectra of the [SClnF3n]þ ions (n ¼ 0–3) [663] and PHnF3n (n ¼ 0–3) [664] have been assigned. The [XeO2F]þ ion takes a pseudo-tetrahedral structure owing to the presence of a sterically active lone-pair electron of the Xe atom [661]:

177

PYRAMIDAL FOUR-ATOM MOLECULES

Matrix cocondensation reactions of alkali halide molecules (MþX) with H2O diluted in Ar produce pyramidal [MOH2]þ ions that exhibit the n(OH2) and d(OH2) in the 3300–3100 and 700–400 cm1 regions, respectively [665]. These ions may serve as simple models of aquo complexes. TABLE 2.3b. Vibrational Frequencies of Pyramidal XY3 Halogeno Compounds (cm1) Molecule

State

~n1

~n2

~n3

~n4

Ref.

[InCl3] [InBr3] [Inl3] CF3 28 SiF3 [SiF3]

Solid Solid Solid Matrix Matrix Matrix

252 177 136 1084 832 770.3

102 74 78 703 406 401

97 46 40 600,500 290 268

585 585 585 586 587 588

Matrix Matrix Solid Solid Matrix Sol’n Sol’n Sol’n Matrix Sol’n Sol’n Sol’n Gas Sol’n Matrix Solid Gas Gas Gas Solid Gas Gas Melt Gas Matrix Gas Sol’n Gas Gas Gas Gas

470.2 388 303 520 509 297 205 162 456 249 176 137 1035 535 554.2 279 893.2 515.0 390.0 303 738.5 416.5 272 212.0 654 380.7 254 186.5 342 220 (162)

— — — 280 256 130 82 61 — — — — 649 347 365.2 146 486.5 258.3 159.9 111 336.8 192.5 128 89.6 259 150.8 101 74.0 123 77 59.7

185 149 110 1250 954 760.4 757.0 582.0 362 285 477 454 256 180 148 405 — 164 127 910 637 644 354 858.4 504.0 384.4 325 698.8 391.0 287 (201) 624 358.9 245 (147) 322 214 163.5

— — — 224 — 104 65 48 — — 58 3045 500 254 263 90 345.6 186.0 112 8 79 262.0 150.2 99 63.9 — 121 8 81 54.3 107 63 47.0

589 590 591 592 593 594 594 594 595 596 596 596 597 598 599 600 601,602 601 601 603 601,604 601,605 606 601 607 601 606,608 601,609 610 611 609

28 SiCl3 GeCl3 [GeCl3] [SnF3]

[SnCl3] [SnBr3] [SnI3] [PbF3] [PbCl3] [PbBr3] [PbI3] NF3 NCl3 14 NCl3 NI3 PF3 PCl3 PBr3 PI3 AsF3 AsCI3 AsBr3 AsI3 SbF3 SbCI3 SbBr3 SbI3 BiCI3 BiBr3 BiI3

(continued)

178

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.3b. (Continued (Continued ) Molecule þ

[SF3] [SCl3]þ [SBr3]þ [SeF3]þ [SeCl3]þ [SeBr3]þ [TeCl3]þ [TeBr3]þ [MnBr3]þ FeCl3 YCl3 CeCl3 NdCl3 SmCl3 GdCl3 DyCl3

State

~n1

~n2

~n3

~n4

Ref.

Melt Solid Solid Melt Sol’n Sol’n Sol’n Sol’n Solid Matrix Gas Gas Gas Gas Gas Gas

943 498 375 781 430 291 362 265 280 363.0 378 328 333 338 343 346

690 276 175 381 206 138 186 112 110 68.7 78 58 60 61 64 65

922 533, 521 429, 414 743 415 298 347 266 150 460.2 359 314 320 324 330 334

356 215, 208 128 275 172 108 (150) 92 80 113.8 58.6 44 45 46 46 47

612 613 614 612 615 606 616 606 617 618 619 620 620 620 620 620

Fig. 2.10. Raman spectra of the SnX3 ions in diethyl ether extracts from SnX2 in HX solutions. The arrows denote polarized bands [594].

PYRAMIDAL FOUR-ATOM MOLECULES

179

TABLE 2.3c. Vibrational Frequencies of Pyramidal XO3 Molecules (cm1) Molecule 2

[SO3] [SO3] [SeO3]2 [TeO3]2 [CIO3] 35

Cl 16 O3 [BrO3] [IO3] XeO3

State

~n1 (A1)

~n2 (A1)

Sol’n Matrix Sol’n Sol’n Sol’n Solid Matrix Sol’n Solid Sol’n Solid Sol’n

967 1000 807 758 933 939 905 805 810 805 796 780

620 604 432 364 608 614 567 418 428 358 348 344

~n3 (E) 933 1175 737 703 977 971 1081 805 790 775 745 833

~n4 (E) 469 511 374 326 477 489 476 358 361 320 306 317

Ref. 621 622 623 623 624 625 626 624,627 625 624 625 628

The ZXYW-type molecule belongs to the C1 point group, and all six vibrations are infrared- and Raman-active. The vibrational spectra of OSClBr [666] and [XSnYZ] (X,Y,Z: a halogen) have been reported [667]. For example, the Raman spectrum of the [ClSnBrl] ion in the solid state exhibits the n(SnCl), n (Sn-Br), and n (SnI) at 275, 193, and 155 cm1, respectively.

Fig. 2.11. IR spectra of KClO3 (solid line) and KIO3 (dotted line).

TABLE 2.3d. Relationship between C3v and Cs

180

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.3e. Vibrational Frequencies of Pyramidal ZXY2-Type Metal Halides (cm1) ~ n1 (A0 ) n (XZ)

~n2 (A0 ) n s (XY)

~n3 (A0 ) ds (YXZ)

~n4 (A0 ) d (YXY)

~n5 (A00 ) n a (YX)

~n6 (A00 ) da (YXZ)

Ref.

972 687 3234

500 — 1233

1307 1002 1564

888 695 3346

1424 1295 1241

630 631 632,633

ClNH2 DPH2 FPH2 C1PH2 FAsH2 ClPF2 BrPF2 FPCI2 FPBr2 IPF2 [FOH2]þ [BrOF2]þ OSF2 OSCl2 OSBr2 [FSH2] [FSO2]

3193 3279 891 679 674 — 795 511 649 545 459 838 824 375 865 1062 1323.5 1251 1121 853 594

3297 — 2304 2303 2108 864 858 525 398 851 3386 655 799.9 492 405 2500 1105

1534 892.9 1090 1096 984 (302) 391 203 123 413 1630 290 381.5 344 267 1181 500

634 635 636,637 637 638 639 639 639 639 640 641 642 643 644,645 646 647 648,649

288.2

260.1

650

[CISO2] [BrSO2] [ISO2] [FSeO2] OSeF2 OSeCl2 OSe (OH)2a FC1O2b FBrO2 [OCIF2]þ [FXeO2]þ

214 203 184 430 1049 995 831 1129.0 506 1333 594

172 115 (55) 283 362 161 430 550.0 305 513 281

526 530 530 324 282 279 336 407.1 394 384 321

3374 — 2310 2303 2117 852 849 525 423 846 3225 630 736.6 455 379 2447 1169 1183 1131.0 1309.6 1312 1308 1300 888 637 347 690 1292.2 953 695 931

1063 969.5 — — — 260 215 267 221 204 1261 315 397.3 284 223 1020 367

ClSO2

1056 1092.6 934 859 842 411 244 328 258 198 1067 365 531,9 194 120 987 378 393 497.7

(103) — — 238 253 255 364 360.1 271 404 249

651 651 651 652 653 654 655,656 657,658 659 660 661,662

HNF2 HNC12 FNH2

a b

1099.8 1098.2 1120 1117 1112 903 667 388 702 634.0 908 731 873

455.8

The OH group was assumed to be a single atom. Empirical assignment.

2.4. PLANAR FOUR-ATOM MOLECULES 2.4.1. XY3 Molecules (D3h) The four normal modes of vibration of planar XY3 molecules are shown in Fig. 2.12; n2, n3, and n4, are infrared-active, and n1, n3, and n4 are Raman-active. This case should be contrasted with pyramidal XY3 molecules, for which all four vibrations are both

PLANAR FOUR-ATOM MOLECULES

Y3

+

X

– + Y2

Y1 ν1(A′1) νs(XY)

181

+ ν2(A′′2) π (XY3)

ν3(E′)

ν4(E ′)

νd(XY)

δ d (YXY)

Fig. 2.12. Normal modes of vibration of planar XY3 molecules.

infrared- and Raman-active. Appendix VII lists the G and F matrix elements of the planar XY3 molecule. Table 2.4a lists the vibrational frequencies of planar XY3 molecules. As stated above, pyramidal and planar structures can be distinguished by the difference in selection rules. Thus, the pyramidal structure should exhibit two stretching bands (A1 and E) while the planar structure should show only one stretching band (E0 ) in IR spectra. The n3 and n4 frequencies are in the order AlH3 < GaH3 < InH3 [680]. MnF3 in the gaseous phase takes a planar structure of C2v symmetry due to the Jahn–Teller effect, and exhibits three stretching [759 (B2), 712 (A1), and 644 (A1)] and three bending [182 (B1), 186 (A1), and 180 (B2, calculated)] vibrations (all in cm1) [693]. Assignment of the IR spectrum of AlCl3NH3 has been based on normal coordinate analysis (staggered conformation, C3v symmetry) [694]. 11 BF3 in inert gas matrices takes a dimeric structure of C2h symmetry via two BFB bridges as supported by IR spectra [695]. Other dimeric species such as Al2F6 and Al2Cl6 are discussed in Sec. 2.10. Normal coordinate calculations on planar XY3 molecules have been carried out by many investigators [696,697]. Table 2.4b lists the vibrational frequencies of planar XO3-type compounds. Figure 2.13 shows the Raman spectra of KNO3 in the crystalline state and in aqueous solution. As discussed in Sec 1.26, the spectra of calcite and aragonite are markedly different because of the difference in crystal structure. More recent normal coordinate calculations [713] on the CO3 radical indicate a trans-Cs

182

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.4a. Vibrational Frequencies of Planar XY3 Molecules (cm1) Molecule 10

BH3 BF3 11 BC13 11 BBr3 11 BI3 A1H3 AID3 AlF3 AlCl3 AlBr3 AlI3 6 ½AlSb 3 GaH3 GaCI3 GaBr3 GaI3 [GaSb3]6 InH3 InCl3 InBr3 Inl3 [InAs3]6 TlBr3 CH3 [PS3] AuI3 [CdCl3] [CdBr3] [CdI3] [HgCl3] PrF3 VF3 CoF3 10

a

Statea

~n1 (A0 1)

~n2 (A00 2)

~n3 (E 0 )

~n4 (E 0 )

Ref.

Mat Gas Liquid Liquid Liquid Mat Mat Mat Mat Gas Gas Solid Mat Mat Gas Gas Solid Mat Mat Gas Gas Solid Sol’n Mat Melt Solid Sol’n Sol’n Sol’n Solid Mat Mat Mat

(2623) 888 472.7 278 192 — — — 393.5 228 156 132 1923.2 386.2 219, 237 147 128 1754.5 359.0 212 151 171 190 — 476 164 265 168 124 273 526,542 649 663

1132 719.3 — 374 — 697.8 513,9 286.2 — 107 77 181 717.4 — 95 63 110 613.2 — 74 56 99 125 730.3 540 — — — — 113 86 — —

2820 1505.8 950.7 802 691.8 1882.8 1376.5 909.4 618.8 450–500 370–410 293,318 — 469.3 — 275 192, 210 — 400.5 280 200–230 197, 215 220 — 695 148 287 184 161 263 458 732 732.2

1610 481.1 253.7 150 101.0 783.4 568.4 276.9 150 93 64 — 758.7 132 84 50 — 607.8 119 62 44 — 51 1383 242 106 90 58 51 100 99 — 162

668 669,670 671,672 671,673 671,673 674 675 676 677 678 678 679 680 677 677,681 677,681 679 680 677,682 678 678 683 684 685 686 687 688 688 688 689 690,691 692 692

Mat: inert gas matrix.

structure (a) rather than a three-membered ring C2v structure (b) suggested originally [714].

Crystals of LiNO3, NaNO3, and KNO3 assume the calcite structure (Sec. 1.31). Nakagawa and Walter [705] carried out normal coordinate analyses on the whole Bravais lattices of these crystals.

183

PLANAR FOUR-ATOM MOLECULES

TABLE 2.4b. Vibrational Frequencies of Planar XO3 and Related Compounds in the Crystalline State (cm1) Compound 10

La[ BO3 ] H3[BO3] Ca[CO3] (calcite) Ca[CO3] (aragonite) Ba[CS3] Ba[CSe3) Na[NO3] K[NO3] NO3 SO3a SeO3a a b

IR IR IR R IR R R IR IR R IR R IR IR R R

~n1 (A0 1)

~n2 (A00 2)

~n3 (E 0 )

~n4 (E 0 )

Ref.

939 1060 — 1087 1080 1084 510 290 — 1068 — 1049 1060 (1068) 1065 923

740.5 668, 648 879 — 866 852 516b 420 831 — 828 — 762 495 497.5 —

1330.0 1490–1428 1429–1492 1432 1504,1492 1460 920 802 1405 1385 1370 1390 1480 1391 1390 1219

606.2 545 706 714 711,706 704 314 185 692 724 695 716 380 529 530.2 305

698,699 706 701 701 701 701 702,703 703,704 705 705 705 705 706–708 709,710 711 712

Gaseous state. Infrared.

M€ uller et al. [715] have shown that the half-width of the n3 band of the NO3 ion at 1384 cm1 in KBr pellets is decreased by a factor of >5 when the ion is trapped in the IV V cavity of ionic carcerand compounds such as K10[HV12 V 6 O44 ðNO3 Þ2 ]14.5H2O. UV resonance Raman spectra of the NO3 ion in H2O and ethylene glycol are dominated by the n1, n3, and their overtones and combination bands. The large splitting of the n3

Fig. 2.13. Raman spectra of KNO3 in (a) the solid state and (b) aqueous solution.

184

APPLICATIONS IN INORGANIC CHEMISTRY

vibration ( 60 cm1) suggests the planar-to-pyramidal conversion by electronic excitation [716]. The IR spectrum of KNO3 in Ar matrices suggest bidentate coordination of the Kþ ion to two oxygen atoms (C2v symmetry) [717]. 28 Si 16 O3 in Ar matrices also takes a C2v structure with one shorter and two longer SiO bonds [718]. The spectra of anhydrous metal nitrates such as Zn(NO3)2 [719] and UO2(NO3)2 [720] can be interpreted in terms of C2v symmetry since the NO3 group is covalently bonded to the metal. Raman spectra of metal nitrates in the molten state [721,722] indicate that the degeneracy of the n3 vibration is lost and the Raman-inactive v2 vibration appears. Apparently, the D3h selection rule is violated because of cation– anion interaction. Like CO32 and NO3 ions, CS32 and CSe32 ions act as chelating ligands. For normal coordinate analyses of planar XO3 molecules, see Refs. 723 and 724. Some XY3-type halides take the unusual T-shaped structure of C2v symmetry shown below. This geometry is derived from a trigonal–bipyramidal structure in which two equatorial positions are occupied by two lone-pair electrons. Typical examples are ClF3 and BrF3. With the equatorial Y atom represented as Y0, the following assignments have been made for these molecules: [725] n(XY0 ), A1, 754 and 672; n(XY), B1 683.2 and 597; n(XY), A1, 523 and 547; d, A1 328 and 235; d, B1, 431 and 347; p, B2, 332 and 251.5 cm1 (for each mode, the former value is for ClF3 and the latter is for BrF3). The Raman spectrum of XeF4 in SbF5 exhibits two strong polarized bands at 643 and 584 cm1, which were assigned to the T-shaped XeF3þ ion in [XeF3][SbF6] [726]. XeOF2 assumes a similar T structure [727]. In an inert gas matrix, UO3 gives an infrared spectrum consistent with the T-shaped structure [728].

2.4.2. ZXY2 (C2v and ZXYW(Cs) Molecules If one of the Yatoms of a planar XY3 molecule is replaced by a Z atom, the symmetry is lowered to C2v. If two of the Yatoms are replaced by two different atoms, Wand Z, the symmetry is lowered to Cs. As a result, the selection rules are changed, as already shown in Table 1.18. In both cases, all six vibrations become active in infrared and Raman spectra. Table 2.4c lists the vibrational frequencies of planar ZXY2 and ZXYW molecules.

185

HBCl2 HAlCl2 HAlBr2 ClGaH2 HGaCl2 HGaBr2 ClInH2 HInCl2 [FCO2]a [(HO) CO2]a [HCO2] [(CH3) CO2]a Si¼CH2 Si¼CD2 O¼CF2 O¼CCl2 O¼CHF O¼CHI O¼CBr2 O¼CCIF O¼CBrCl

ZXYW (Cs)

ZXY2 (C2v)

XY3 (D3h)

786 473 454 620 464.3 445.5 415.7 469.7 — 835

1069 621

687 539 767.4 581.2 — — 512 667 547

2616 1966 1953 406.9 2015.3 1991.5 343.4 1846.9 883 960

2803 926

933 808 1930 1815.6 1837 1776.3 1828 1868 1828

2947 2185 965.6 568.3 2981 2930 425 776 517

1351 1413 — — 1243.7 837.4 1065 (561) 757 1095 806

1585 1556

1091 580 574 1978.1 437.3 333.0 1820.3 358.5 1749 1697

zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n4 ðB 2 Þ ~n2 ðA1 Þ n s ðXYÞ n a ðXYÞ 0 0 ~n4 ðA Þ ~n4 ðA Þ nðXZÞ nðXWÞ

~n6 (B1) p (ZXY2) ~n6 (A00 ) p

~n1 (A1) n (XZ) ~n1 (A0 ) n (XZ) 735 482 366 1964.6 414.3 287.5 1804.0 369.1 1316 1338

~n3 ðE Þ n d ðXYÞ

~n2 (A00 2) p (XY3)

~n1 (A0 1) n s (XY)

0

TABLE 2.4c. Vibrational Frequencies of Planar ZXY2 and ZXYW Molecules (cm1)

1273 1054 582.9 368.0 1343 1247.9 181 501 240

760 650

— — — 731.4 130 — 575.8 — — 712

0

263 216 619.9 441.8 663 — 350 415 372

1383 471

895 655 635 510.1 607.5 597.0 541.4 402.6 — 579

~n4 ðE Þ dd ðYXYÞ zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n5 ðB 2 Þ ~n3 ðA1 Þ ds ðZXYÞ da ðZXYÞ 0 0 ~n5 ðA Þ ~n3 ðA Þ dðZXYÞ dðZXWÞ

(continued)

740 740 741,742 743 744,745 746 747 747,747a 747

739 739

729 730 730 731 732,733 733 734,735 734 736,737 738

Ref.

186

1874 1063 1368 1120 1280 442 1240 638 739.1 580 902 555 370 1573 1862 1648 363.2

ZXYW (Cs)

O¼CBrF S¼CH2 S¼CF2 Se¼CCl2 Se¼CF2 [SeCS2] O¼SiCl2 S¼SiF2 S¼SiCl2 S¼GeCl2 (HO)¼NO2a FNO2 ClNO2 O¼NF2 [O¼NF2] þ [O¼NCl2] þ BrPO2

a

Only the ZXY2 skeletal vibrations are listed.

ZXY2 (C2v)

620 993 622 472 560 485 — 296 — — 767 740 652 (440) 715

~n6 (B1) p (ZXY2) ~n6 (A00 ) p

~n1 (A1) n (XZ) ~n1 (A0 ) n (XZ)

XY3 (D3h)

~n2 (A00 2) p (XY3)

~n1 (A0 1) n s (XY)

TABLE 2.4c. (Continued)

721 2970 787 500 710 433 501.1 996 — 404 1311 1308 1318 813 897 628 1146.5

1068 3028 1189 812 1208 925 637.5 969 609.9 440 1697 1800 1670 761 1163 735 1440.1

zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n4 ðB 2 Þ ~n2 ðA1 Þ n s ðXYÞ n a ðXYÞ 0 0 ~n4 ðA Þ ~n4 ðA Þ nðXZÞ nðXWÞ

0

~n3 ðE Þ n d ðXYÞ

398 1550 526 296 432 284 — 337 — — 660 810 787 705 569 220 513.5

0

335 1437 417 302 352 265 — 247 186 — 597 555 411 553 347 420 230

~n4 ðE Þ dd ðYXYÞ zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{ ~n5 ðB 2 Þ ~n3 ðA1 Þ ds ðZXYÞ da ðZXYÞ 0 0 ~n5 ðA Þ ~n3 ðA Þ dðZXYÞ dðZXWÞ 747 748 749,750 751 752 753 754 755 756 757 758 759,760 754,760 761 762,763 764 765

Ref.

OTHER FOUR-ATOM MOLECULES

187

Silanone produced by the silane–ozone photochemical reaction in Ar matrices gives rise to a band at 1202 cm1 in the IR spectrum that is assigned to the n(Si¼O) [766]:

The frequencies listed for the formate and acetate ions were obtained in aqueous solution. Vibrational spectra of metal salts of these ions are discussed in Sec. 3.8. Although not listed in this table, the IR spectra of binary mixed halides of boron [767] and aluminum [768] have also been reported. 2.5. OTHER FOUR-ATOM MOLECULES 2.5.1. X4 Molecules Tetrahedral X4-type molecules of Td symmetry exhibit three normal vibrations as shown in Fig. 2.14, and n1(A1) and n2(E) are Raman-active whereas n3(F2) is IR as well as Raman-active. Table 2.5a lists observed frequencies of tetrahedral X4 molecules. Comparison of vibrational frequencies of the X44 (X ¼ Si,Ge,Sn) series and the corresponding isoelectronic Y4 (Y ¼ P,As,Sb) series shows that the ratio, n(X44)/n (Y4) is approximately 0.77. On the basis of this criterion, Kliche et al. [770] predicted that the vibrational frequencies of the Pb44 ion are 115 (A1), 69 (E), and 93 (F2) cm1. On the other hand, the X4-type cation such as S42þ assumes a square–planar ring structure of D4h symmetry, and its 6 (3 4–6) vibrations are classified into A1g (R)þ B1g (R)þ B2g (R)þ B2u(inactive)þ Eu (IR). Assignments of these vibrations have been based on normal coordinate analysis [776a–776c]. Table 2.5b shows the normal modes and vibrational frequencies of square–planar X4 cations. 2.5.2. X2Y2 Molecules Molecules like O2H2 take the nonplanar C2 structure (twisted about the OO bond by 90 ), whereas N2F2 and [N2O2]2 exist in two forms: trans-planar (C2h) and

ν1(A1)

ν2(E)

ν3( F2)

Fig. 2.14. Normal modes of vibration of tetrahedral X4 molecules.

188

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.5a. Vibrational Frequencies of X4 Molecules (cm1) ~n1 (A1)

~n2 (E)

~n3 (F2)

Ref.

Si44

486

276

Sn44 P4 As4 Sb4 Bi4 Ta4

182 613.6 353 241.5 149.5 270.2

338 356 376 197 206 216 140 467.3 265.5 178.5 120.4 185.1

769,770

Ge44

282 305 , 154 165

Molecule

102 371.8 207.5 137.1 89.8 130.6

769

770 771,772 773 774 775 776

cis-planar (C2v). Figure 2.15 shows the six normal modes of vibration for these three structures. The selection rules for the C2v and C2 structures are different only in the n6 vibration, which is infrared-inactive and Raman depolarized in the planar model but infrared-active and Raman polarized in the nonplanar model. Table 2.5c lists the vibrational frequencies of X2Y2-type molecules. In N2 matrices, (NO)2 exists in three forms; cis, trans, and another form of uncertain structure. The cis form is most stable, and its ns and na are at 1870 and 1776 cm1 respectively [791]. The NO reacts with Lewis acids such as BF3 to form a red species at 77 K. Ohlsen and Laane [792] measured its resonance Raman spectrum and concluded that it is an asymmetric (NO)2 dimer having a cisplanar structure:

A possible mechanism for the formation of such a dimer in the presence of Lewis acids has been proposed. Although the N2O22 ion takes a cis or trans structure in the solid state, the N2O22 ion produced in Ar matrices exhibits an IR band at

TABLE 2.5b. Vibrational Frequencies of Square-Planar X4-Type Ions (cm1) Ion

S42þ

Se42þ

Te42þ

~ n1 (A1g)

583.6

321.3

219

~ n2 (B1g)

371.2

182.3

109

~ n3 (B2g)

598.1

312.3

219

~ n5 (Eu)

542

302

187

Source: Taken from Ref. 776b.

189

OTHER FOUR-ATOM MOLECULES

X1

X2 ν1

Y1

Y2

νs(XY)

ν2

ν3

ν(XX)

δs(YXX)

cis-Planar (C2υ)

A1 (p)

A1 (p)

A1 (p)

Nonplanar(C2)

A (p)

A (p)

A (p) +

ν4

ν5

νa(XY)

δa(XX)

–

–

+

ν6 δt (YXX)

cis-Planar

B2 (dp)

B2(dp)

A2(p)

Nonplanar

B (dp)

B (dp)

A (p)

Y2 X2

X1

Y1

ν1

ν2

ν3

νs(XY)

ν(XX)

δs (YXX)

Ag (p)

Ag (p)

trans–Planar (C2h) Ag (p)

+

ν4 Y1

νa(XY) Bu (dp)

– ν5 δa(YXX) Bu (dp)

+

ν6 ρl (XY) Au (dp)

Fig. 2.15. Normal modes of vibration of nonlinear X2Y2 molecules (p—polarized; dp—depolarized) [749].

1205 cm1, and its structure was suggested to be [793]

Diazene (N2H2) exists in both cis and trans forms:

190

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.5c. Vibrational Frequencies of X2Y2 Molecules (cm 1)a trans-X2Y2: cis-X2Y2: Twisted X2Y2:

C2h C2v C2

Assignment cis-[N2O2] trans-[N2O2]2 cis-N2F2 trans-N2F2 O2H2 O2D2 O2F2 S2H2 S2F2 S2Cl2 S2Br2 Se2Cl2 Se2Br2 2

Solid Soild Liquid Liquid Gas Gas Matrix Liquid Gas Gas Liquid Liquid Liquid

n1 Ag A1 A

n2 Ag A1 A

n3 Ag A1 A

n4 Bu B2 B

n5 Bu B2 B

n6 Au A2 A

ns (XY)

n (XY)

ds (YXX)

na (XY)

da (YXX)

p

Ref.

830 (1419) 896 1010 3607 2669 611 2509 717 466 365 367 265

1314 (1121) 1525 1522 1394 1029 1290 509 615 546 529 288 292

584 (396) 341 600 864 867 366 (868)b 320 202 172 130 107

1047 1031 952 990 3608 2661 624 2557c 681 457 351 367 265

330 371 737 423 1266 947 459 882 301 240 200 146 118

— 492 (550) 364 317 230 (202) — 183 92 66 87 50

777,778 779,780 781 782 783,784 783 785 786 787 788,789 790 789,790 790

a

Except for N2F2 and [N2O2]2 , all the molecules listed take the C2 or C2v structure. Solid. c Gas. b

and the observed frequencies are 3116, 3025, 1347, and 1304 cm 1 for the cis form, and 3109 and 1333 cm 1 for the trans form [794]. The IR spectra of (H2O2)2 dimer and its deuterated analogs have been measured in Ar matrices. Both experimental and theoretical studies suggest the cyclic structure shown below [794a]:

A strong vibrational coupling was observed between the two bonded OH stretching modes, and a relatively strong coupling was noted between the two HOO bending modes. The Raman spectra of Se2Cl2 and Se2Br2 in CS2 and CCl4 solutions indicate that the former takes the C2 structure, whereas the latter takes the C2v structure [795]. Although S2Cl2 does not dimerize at low temperatures, Se2Cl2 forms a dimer below 50 C. A new Raman band observed at 215 cm 1 was attributed to the dimer for which the following ring structure was proposed [796]:

OTHER FOUR-ATOM MOLECULES

191

Ketelaar et al. [797] found that the liquid mixture of CS2 and S2Br2, for example, exhibits IR spectra that show combination bands between the n3 of CS2 (1515 cm1) and a series of S2Br2 vibrations. Such simultaneous transitions seem to suggest the formation of an intermolecular complex, at least on the vibrational time scale. The Raman spectra of the [T12X2]2 ion (X ¼ Se or Te], which takes an almost planar diamond-shaped ring structure (C2v), have been assigned [798]. Si2H2 prepared in inert gas matrices assumes butterfly-like structure wth the SiSi bond located along the fold of the two SiHSi wing planes, and its IR spectrum was assigned on the basis of C2v symmetry [799]. The [C12O2]þ ion formed by the pp* charge transfer interaction of Cl2þ and O2 takes a planar trapezoid structure of C2v symmetry, and the v(OO) and v(ClCl) are at 1534 (A1) and 593 (A1), respectively, and v (CIO) are at 414 (B2) and 263 (A1) (all in cm1) [800]. Normal coordinate analyses of H2O2 [801] and S2X2 (X: a halogen) [802,790] have been carried out. Other (XY)2-type compounds include dimeric metal halides such as (LiF)2, which takes a planar ring structure (Sec. 2.1). 2.5.3. Planar WXYZ, XYZY, and XYYY Molecules (Cs) Planar four-atom molecules of the WXYZ, XYZY, and XYYY types have six normal modes of vibration, as shown in Fig. 2.16. All these vibrations are both infrared- and Raman-active. In HXYZ and HYZY molecules, the XYZ and YZY skeletons may be linear (HNCO, HSCN) or nonlinear (HONO, HNSO). In the latter case, the molecule may take a cis or trans structure. Table 2.5d lists the vibrational frequencies of molecules and ions belonging to these types. Normal coordinate analyses have been carried out for HN3 [831] and HONO [832]. Teles et al. [804] measured the IR spectra of all four possible isomers of HNCO in Ar matrices at 13 K, and compared the observed frequencies with those obtained by ab initio calculation:

192

APPLICATIONS IN INORGANIC CHEMISTRY

X

W

Y

Z

ν2(A′) ν(XY)

ν1(A′) ν(WX)

ν3(A′) ν(YZ)

ν4(A′) δ(WXY)

+

–

+

+ ν5(A′) δ(XYZ)

ν6(A′′) π

Fig. 2.16. Normal modes of vibration of nonlinear WXYZ molecules.

As shown above, fulminic acid is linear, while the remaining three acids assume bent structures. For a detailed discussion on the structures and IR spectra of these acids, see the review by Teles et al. [804]. Vibrational frequencies are reported for other four-atom molecules such as cisOSOO, which was produced by laser irradiation (193 nm) of the planar SO3 in Ar matrices at 12 K [833], and the HOOO radical, which is probably in cis-conformation in inert gas matrices [834]. Vibrational assignments have also been made for linear four-atom molecules containing CN groups; CNCN(isocyanogen) [835]. HCCN radical (cyanomethylene), [836] [FCNF] [837], and a pair of HBeCN and HBeNC that were formed by reacting Be atom with HCN in Ar matrices at 6–7 [838]. 2.6. TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES 2.6.1. Tetrahedral XY4 Molecules (Td) Figure 2.17 illustrates the four normal modes of vibration of a tetrahedral XY4 molecule. All four vibrations are Raman-active, whereas only n3 and n4 are infraredactive. Appendix VII lists the G and F matrix elements for such a molecule. Fundamental frequencies of XH4-type molecules are listed in Table 2.6a. The trends n3 > n1 and n2 > n4 hold for the majority of the compounds. The XH stretching frequencies are lowered whenever the XH4 ions form hydrogen bonds with counter-

193

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

TABLE 2.5d. Vibrational Frequencies of WXYZ and Related Molecules (cm1) Moleculea,b

n~1, n(WX) n~2, n(XY) n~3, n(YZ)

n~4, d(WXY) n~5, d(XYZ) n~6, p

Ref.

2085 2259.0

1092 —

1190 769.8

570 573.7

538 —

803 804,805

DNCO HNOO HOCN DOCN HCNO

3235 3516.8 3505.7 2606.9 3165.5 3569.6 2635,0 3317.2

2231,0 1092.3 1081.3 949.4 2192,7

— 1054.5 2286,3 2284,6 1244.1

578.6 1485.5 1227,9 1077,8 566.6

— 764.0 — — —

804 806 804,807 804,807 804

ClCNO (g)

—

2219

—

—

808

DCNO HONC DONC HNCO (film) FNCO ClNCO (g) BrNCO INCO (g) HNNN (g) FNNN (g) ClNNN (g) BrNNN (g) INNN (g) cis-HONO trans-HONO HOClO HNCS DNCS HCNS (g) DCNS (g) trans-HSCN ClSCN (sol’n) BrSCN (sol’n) I SCN (sol’n) cis-HNSO trans-HNSO trans-HNNO cis-ClONO trans-ClONO cis-BrONO trans-BrONO trans-HONS FNSO (liquid) ClNSO (liquid) BrNSO (liquid) INSO (liqid) cis-OSNO trans-OSNO

2612.7 3443,7 2545.2 3203

2063.2 628.4 623.1 2254

1350 1336 1218,5 2190.1 2190.3 13251194

475.4 — — — 536,9 538.2 —

418.7 1232,4 902.6 937793

— 361.2 357.3 612

— 379.3 362.1 649

804 804 804 809

861 607.7 505 462.3 3324 873 545 452 360 3397 3553 3527.1 3565 2623 3539 2645 2580.9 520 451 372 3309 3308 3254.0 640.0 658.0 862.6 835.9 3528.0 825 526 451 372 1156.1 1178.0

2172 2212.2 2196.0 2201.1 2150 2037 2075 2058 2055 847 796 591.5 1979 1938 1989 1944 — 678 676 700 1083 982 1628.9 — — 573.5 586.9 642.1 995 989 1000 1928 — —

(2160) 1306.6 1290.9 1298.1 1168 1090 1145 1150 1170 1630 1685 973.9 988 — 857 851 2182.3 2162 2157 2130 1249 1381 1296.2 1710.5 1754.9 1650.7 1723,4 969.5 1230 1221 1214 1247 1454,4 1459.0

529 — 137.4 — 1273 658 719 682 648 1317 1265 1176.9 577 548 615 549 965.9 — — — 900 878 1214.7 — — 420 391.2 1363.3 228 187 161 154 — —

695 707.7 691.1 667.7 527 241 223 — — 616 608 — 461 366 469 366 — 353 369 362 447 496 — — 850 — 197 476.5 600 672 624 602 — —

646 559.0 572.2 583.3 588 504 522 — 578 637 550 — — — 539 481 — — — — 755 651 746.5 — — 368,3 299.3 531 395 359 342 330 — —

810 811 812 811 813 813 813 813 813 814,815 814 816 817 817 818 818 819 820,811 820,811 820,811 821,822 823 824 825 825 826 826 827 828 828 828 828 829 829

(HNCN) (s) HNCO

(continued)

194

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.5d. (Continued ) Moleculea,b

n~1, n(WX)

n~2, n(XY) n~3, n(YZ)

n~4, d(WXY) n~5, d(XYZ) n~6, p

Ref.

cis-HOSO

3544

776

—

830

1166

—

—

a

Vibrational frequencies were measured in inert gas matrices except for those marked by s (solid), g (gas), sol’n (solution), liquid, and film. b Some assignments listed are only empirical.

Y4

X Y1

Y3

Y2 ν2(E) δd (YXY)

ν1(A1) νs(XY)

ν3(F2) νd (XY)

ν4(F2) δd (YXY)

Fig. 2.17. Normal modes of vibration of tetrahedral XY4 molecules.

TABLE 2.6a. Vibrational Frequencies of Tetrahedral XH4 Molecules (cm1) Molecule 10

[ BH4 ] [10 BD4 ] [AlH4] [AlD4] [GaH4] CH4 CD4 SiH4 SiH4 SiD4 GeH4 GeD4 SnH4 SnD4 PbH4 [14 NH4 ]þ [15 NH4 ]þ [ND4]þ [NT4]þ [PH4]þ [PD4]þ {AsH4}þ [AsD4]þ [SbH4]þ [SbD4]þ

~ n1 2270 1604 1757 1256 1807 2917 2085 2180 — (1545) 2106 1504 — — — 3040 — 2214 — 2295 1654 2321 1664 2051 1462

~n2 1208 856 772 549 — 1534 1092 970 — (689) 931 665 758 539 — 1680 (1646) 1215 976 1086, 1026 772, 725 1024 732 842 600

~n3 2250 1707 1678 1220 — 3019 2259 2183 2192 1597 2114 1522 1901 1368 1823 3145 3137 2346 2022 2366, 2272 1732 2341 1676 2061 1470

~n4 1093 827 760 or 766 560 or 556 — 1306 996 910 913 681 819 596 677 487 602 1400 1399 1065 913 974, 919 677 941 688 795 575

Ref. 839,840 839,840 841,842 841,842 841,842 843 844,845 843,846 847 846,848 843,849 843 850 850 851 843 852 843 852 853 853 854 854 854 854

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

195

ions. In the same family of the periodic table, the XH stretching frequency decreases as the mass of the X atom increases. Shirk and Shriver [841] noted, however, that the n1 frequency and the corresponding force constant show an unusual trend in group IIIB: [BH4] ~n1 (cm1) F11 (mdyn/e´)

2270 3.07

[GaH4] 1807 1.94

[AlH4] 1757 1.84

In a series of ammonium halide crystals, the n(NH4þ) becomes lower and the d(NH4þ) becomes higher as the NHX hydrogen bond becomes weaker in the order X ¼ F > Cl > Br > I. Figure 2.18 shows the infrared spectrum of NH4CI, measured by Hornig and others [855], who noted that the combination band between n4 (F2) and n6 (rotatory lattice mode) is observed for NH4F, NH4Cl, and NH4Br because the NH4þ ion does not rotate freely in these crystals. In NH4I (phase I), however, this band is not observed because the NH4þ ion rotates freely. Table 2.6b lists the vibrational frequencies of a number of tetrahalogeno compounds. The trends n3 > nx and n4 > n2 hold for most of the compounds. The latter trend is opposite to that found for MH4 compounds. The same table also shows the effect of changing the halogen. First, the MX stretching frequency decreases as the halogen is changed in the order F > Cl > Br > I. The average values, of n(MBr)/n(MCl) and n(MI)/n(MCl) calculated from all the compounds listed in Table 2.6b are approximately 0.76 and 0.62, respectively, for n3, and approximately 0.61 and 0.42, respectively, for n1. These values are useful when we assign the MX stretching bands of halogeno complexes (Sec. 3.23 in Part B). Also, the effect of changing the oxidation state on the MX stretching frequency is seen in pairs such as [FeX4] and [FeX4]2 (X ¼ Cl, Br) [894], the MX stretching frequency increases as the oxidation state of the metal becomes higher. The ratio n3(FeX4)/n3(FeX42) is 1.32 in this case. In the solid state, n3 and n4 may split into two or three bands because of the site effect. In some cases, the MX4 ions are distorted to a flattened tetrahedron (D2d) or a structure of lower symmetry (Cs) [892,912]. According to an X-ray analysis [913] the unit cell of [(CH3)2CHNH3]2 [CuCl4] contains two square-planar and four distorted tetrahedral [CuCU4]2 ions.

Fig. 2.18. IR spectrum of crystalline NH4Cl ( n5, n6—lattice modes).

196

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.6b. Vibrational Frequencies of Tetrahedral Halogeno Compounds (cm1) Molecule 2

[BcF4] [MgCL4]2 [MgBr4]2 [MgI4]2 [BF4] [BCl4] [BBr4] [BI4] [AlF4] [AlCl4] [AlBr4] [AlI4] [GaCl4] [GaBr4] [GaI4] [InCl4] [InBr4] [InI4] [TlCl4] [TlBr4] [TiI4] CF4 CCl4 CBr4 CI4 SiF4 SiCl4 SiBr4 SiI4 70 GeF4 GeCl4 GeBr4 GeI4 SnCl4 SnBr4 SnI4 PbCl4 [NF4]þ [NCl4]þ [PF4]þ [PCl4]þ [PBr4]þ [PI4]þ [AsF4]þ [AsCl4]þ [AsBr4]þ [AsI4]þ [SbCl4]þ [SbBr4]þ [CuCl4]2

~ n1

~n2

~n3

~n4

Ref.

547 252 150 107 777 406 243 170 622 348 212 146 343 210 145 321 197 139 303 186 130 908.4 460.0 267 178 800.8 423.1 246.7 168.1 735.0 396.9 235.7 158.7 369.1 222.1 147.7 331 848 635 906 458 266 193.5 745 422 244 183 395.6 234.4 —

255 100 60 42 360 192 118 83 210 119 98 51 120 71 52 89 55 42 83 54 — 434.5 214.2 123 90 264.2 145.2 84.8 62 202.9 125.0 74.7 60 95.2 59.4 42.4 90 443 430.3 275 178 106 71.0 213 156 88 72 121.5 76.2 77

800 330 290 259 1070 722 620 533 760 498 394 336 370 278 222 337 239 185 296 199 146 1283.0 792,765a 672 555 1029.6 616.5 494.0 405.5 806.9 459.1 332.0 264.1 408.2 284.0 210 352 1159 283.3 1167 662 512 410 829 500 349 319 450.7 304.9 267, 248

385 142 90 60 533 278 166 117 322 182 114 82 153 102 73 112 79 58 104 64 60 631.2 313.5 183 123 388.7 220.3 133.6 90.6 273.7 171.0 111.1 79.0 126.1 85.9 63.0 103 611 233.3 358 255 153 89.0 272 187 115 87 139.4 92.1 136, 118

856 857 857 857 858 859,860 859,861 862 863 864 865 866 867 868 869 870 871 869 872 872 873 874 874 875 876 874 874 874 877 878 874 874 877 874 874 874 879 880,882 881 882 883 884 885 886 887,888 889 890 891 891 892,893

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

197

TABLE 2.6b. (Continued ) Molecule 2

[CuBr4] [ZnCl4]2 [ZnBr4]2 [ZnI4]2 [CdCl4]2 [CdBr4]2 [CdI4]2 [HgCl4]2 [HgI4]2 [ScI4] TiF4 TiCl4 TiBr4 TiI4 ZrF4 ZrCl4 ZrBr4 ZrI4 HfCl4 HfBr4 HfI4 VCl4 CrF4 CrCl4 CrBr4 [MnCl4]2 [MnBr4]2 [MnI4]2 [FeCl4] [FeBr4] [FeCl4]2 [FeBr4]2 [Fe I4] [NiCl4]2 [Ni Br4]2 [Ni I4] 2 {CoCl4]2 [CoBr4]2 [Co I4]2 U F4 U Cl4 a

~ n1

~n2

~n3

~n4

Ref.

— 276 171 118 261 161 116 267 126 129 712 389 231.5 162 (725–600) 377 225.5 158 382 235.5 158 383 717 373 224 256 195 108 330 211 266 162 142 264 — 105 287 179 118 — —

— 80 — — 84 49 39 180 35 37 185 114 68.5 51 (200–150) 98 60 43 101.5 63 55 128 — 116 60 — 65 46 114 74 82 — 60 — — — 92 74 — (123) —

216, 174 277 204 164 249, 240 177 — 276 140 — 793 498 393 323 668 418 315 254 390 273 224 475 790 486 368 278, 301 209, 221 188, 193 378 307/292 286 219 252/235 294/280 228 191 320 243/249 202/194 537 337.4

85 126 91 — 98 75, 61 50 192 41 54 209 136 88 67 190 113 72 55 112 71 63 128, 150 201 126 71 120 89 56 136 94 119 84 73 119 81 — 143/126 101/90 56 114 71.7

892 894 894 894 895 895 895 896 897 898 899 900 900 900 901 900 900 900 900 900 900 902 903 904 904 894,905 894,905 894,905 894 906,894 894 894 907 892,908 892,908 892 909 909 892 910 911

Fermi resonance with (n1þ n4).

Willett et al. [914] demonstrated by using IR spectroscopy that distorted tetrahedral ions can be pressed to square–planar ions under high pressure. In nitromethane, (Et4N)2[CuCl4] exhibits two bands at 278 and 237 cm1, indicating the distortion in solution [915]. A solution of (Et3NH) [GaCl4] in 1,2-dichloroethylene exhibits three bands at 390,383, and 359 cm1 due to lowering of symmetry caused by NHCl hydrogen bonding [916]. The IR spectra of long-chain tertiary and quaternary

198

APPLICATIONS IN INORGANIC CHEMISTRY

ammonium salts of [GaCl4] and [GaBr4] ions in benzene solution show that the degree of distortion of these ions depends on the nature of the cation and the concentration [917]. Distortion of [MCI4]2 ions (M ¼ Fe,Co,Ni,Zn) is also reported for their cesium and rubidium salts [918]. Matrix isolation spectroscopy has provided structural information that is unique to XY4-type molecules in inert gas environments. For example, the symmetry of the anions in CsþBF4 and CsþPF4 formed via salt–molecule reactions in Ar matrices are lowered to C3v in the former, and to a Symmetry no higher than C2v, in the latter [920]. In inert gas matrices, the symmetry of UF4 is distorted to D2d [921]. Irradiation of CCl4, in Ar matrices at 12 K produced a new species, isotetrachloromethane, for which Maier et al. [922] proposed the following structure:

The v(CClt) are observed at 1019.7 and 929.1 cm1, whereas the v(CClb) and v(ClbCle) are assigned at 501.9 and 246.4 cm1 respectively. Jacox and Thompson [923]. presented IR evidence for the formation of the CO4 ion resulting from the reaction of CO2, O2, and excited Ne atoms in Ne matrices. The planar structure of Cs symmetry has been proposed:

Table 2.6b includes a number of data obtained by Clark et al. from their gas-phase Raman studies [874,900]. Some halides such as Til4 [924], Snl4 [877,925] and TiBr4 [900], and VCl4 [926] have strong electronic absorptions in the visible region, and their resonance Raman spectra have been measured in solution. As discussed in Sec. 1.23, these compounds exhibit a series of v1 overtones under rigorous resonance conditions. A tetrahedral MCl4 molecule in which M is isotopically pure and Cl is in natural abundance consists of five isotopic species because of mixing of the 35 Cl (75.4%) and 37 Cl (24.6%) isotopes. Table 2.6c lists their symmetries, percentages of natural abundance, and symmetry species of infrared-active modes corresponding to the n3 vibration of the Td molecule. It has been established [927] that these nine bands overlap partially to give a five-peak chlorine isotope pattern whose relative intensity is indicated by the vertical lines shown in Fig. 2.19b. If M is isotopically mixed, the spectrum is too complicated to assign by the conventional method. For example, tin is a mixture of 10 isotopes, none of which is predominant. Thus 50 bands are expected to appear in the n3 region of SnCl4. It is almost impossible to resolve all these peaks, even

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

199

TABLE 2.6c. Infrared-Active Vibrations of M35 Cln 37 Cl4n Type Molecules Species 35

M Cl4 M35 Cl37 3 Cl M35 Cl37 2 Cl2 M35 Cl 37 Cl3 M35 Cl4

Symmetry Td C3v C2v C3v Td

Abundance (%) 32.5 42.5 20.5 4.4 0.4

IR-Active Modes F2 A1, E A1, B1, B2 A1, E F2

in an inert gas matrix at 10 K. K€ oniger and M€ uller [928] therefore, prepared 116 SnCl4 35 116 and Sn Cl4 on a milligram scale and measured their infrared spectra in Ar matrices. As expected, the former gave a five-peak chlorine isotope pattern, whereas the latter showed a single peak at 409.8 cm1. This work was extended to GeCl4, which consists of two Cl and five Ge isotopes. In this case, 25 peaks are expected to appear in the n3 region. However, the observed spectrum (Fig. 2.19a) shows about 10 bands; K€oniger et al. [929], therefore, prepared 74 GeCl4 and Ge35 Cl4 and measured their spectra in Ar

Fig. 2.19. Matrix isolation IR (a) and computer simulation spectra (b) of GeCl4. Vertical lines in (b) show the five-peak chlorine isotope pattern of 74 GeCl4 .

200

APPLICATIONS IN INORGANIC CHEMISTRY

matrices. As expected, both compounds showed a five-peak spectrum. The ism (isotope shift per unit mass difference) values for Cl and Ge were found to be 3.8 and 1.2 cm1, respectively. Using these values, it was then possible to calculate the frequencies of all other isotopic molecules. Furthermore, the relative intensity of individual peaks is known from the relative concentration of each isotopic molecule. On the basis of this information, Tevault et al. [930] obtained a computer-simulation infrared spectrum of GeCl4 in natural abundance (Fig. 2.19). Normal coordinate analyses of tetrahedral XY4 molecules have been carried but by a number of investigators [931]. Thus far, Basile et al. [932] have made the most complete study. They calculated the force constants of 146 compounds by using GVF, UBF, and OVF fields (Sec. 1.14), and discussed several factors that influence the values of the XY stretching force constants. It has long been known that molecules such as SF4, SeF4, and TeF4 assume a distorted tetrahedral structure (C2v) derived from a trigonal–bipyramidal geometry with a lone-pair of elections occupying an equatorial position:

These molecules exhibit nine normal vibrations that are classified into 4A1þ A2 2B1þ 2B2. Here, all the vibrations are IR- as well as Raman-active except for A2 which is only Raman-active. Table 2.6d lists the vibrational frequencies of distorted XY4 molecules of this type. It should be noted that these molecules exhibit four stretching modes, two of which (2A1) are polarized in the Raman. Adams and Downs [939] carried out normal coordinate analyses on SeF4 and TeF4, and found that the axial bonds are weaker than the equatorial bonds. The vibrational spectra of tetraalkylammonium salts of [AsX4] (X ¼ Cl, Br), [BiX4] (X ¼ Cl,Br,I), and [SbX4] (X ¼ Cl,Br,I) have been studied in the solid state and in solution) [936]. Except for solid (Et4N)[AsCl4] and [(n-Bu)4N][SbI4], all these ions assume the distorted tetrahedral structure shown above. The [ClF4]þ, [BrF4]þ, and [IF4]þ ions were found in the following adducts [941]: CIF5 ðAsF5 Þ BrF5 ðSbF5 Þ2 IF5 ðSbF5 Þ

¼ ½CIF4 ½AsF6 ¼ ½BrF4 ½Sb2 F11 ¼ ½IF4 ½SbF6

It should be noted that SeCl4, SeBr4, TeCl4, and TeBr4 consist of the pyramidal XY3þ cation and the Y anion in the solid [615,616].

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

201

TABLE 2.6d. Vibrational Frequenciesa of Distorted Tetrahedral XY4 Molecules (cm1) C2v

[PF4] [SbF4] [SbCl4] [SbBr4] [Sb I4] SF4 SeF4 TeF4 TeCl4 [ClF4]þ [BrF4]þ {I F4]þ

n~1, A1

n~2, A1

n~3, A1

n~4, A1

n~5, A1

n~6, A1

n~7, A1

n~8, A1

n~9, A1

Ref.

798 584 339 228 169 892 739 695 — 800 723 728

422 404 296 190 — 558 551 572 — 571 606 614

446 305 147 — 114 532 362 333.2 — 385 385 345

210 161 — — — 228 200 151.5 — 250 219 263

— — — — — (437) — — — 475 — —

515 480 321 201 162 730 585 586.9 312 795 704 —

446 247 199 140 85 475 254 293.2 165 515 419 388

745 541 246 169 148 867 717 682.2 378 829 736 719

290 220 — — — 353 403 184.8 104 385 369 311

933,920 934,935 935 935,936 935,936 937 938 934,939 940 937 941 941

a Whereas n1 and n8 are stretching modes of equatorial bonds, n2 and n6 are stretching modes of axial bonds. For the normal modes of bending vibrations, see Ref. 942.

Table 2.6e lists the vibrational frequencies of tetrahedral MO42, MS42, and MSe2 -type compounds. The rule n3 > n1 and n4 > n2 hold for the majority of the compounds. Figure 2.20 shows the Raman spectra of K2SO4 in the solid state as well as in aqueous solution. It should be noted that n2 and n4 are often too close to be observed as separate bands in Raman spectra. Weinstock et al. [960] showed that, in Raman spectra, n2 should be stronger than n4, and that n4 is hidden by n2 in [MoO4]2 and [ReO4] ions. Baran et al. [974,975] have found several relationships between the n1/n3 ratio and the negative charge of the anion or the mass of the central atom in a series of oxoanions listed in Table 2.6e: (1) For a given central atom, the n1/n3 ratio increases as the negative charge of the anion increases (e.g., [MnO4] < [MnO4]2 < [MnO4]3). (2) For anions of the same negative charge with the central atom belonging same group of the periodic table, the n1/n3 ratio increases with the mass of the central atom (e.g., [PO4]3 < [AsO4]3). (3) For isoelectronic ions in which the mass of the central atom remains approximately constant, the n1/n3 ratio increases with the increasingly negative charge of the anion (e.g., [ReO4] < [WO4]2). These trends are very useful in making correct assignments of the n1 and n3 vibrations of tetraoxoanions [975]. The IR spectra of matrix-isolated M2(SO4) (M ¼ K,Rb,Cs) can be interpreted in terms of D2d symmetry and their n(MO) vibrations are observed at 262–220, 215– 190, and 195–148 cm1, respectively, for the K, Rb, and Cs salts [976].

202

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.6e. Vibrational Frequencies of Tetrahedral MO4-, MS4-, and MSe4-Type Compounds (cm1) ~n1

~n2

~n3

~n4

Ref.

880

372

627

943,944

495

353

—

945

[SiO4] [PO4]3

819 938

340 420

391

282

416

215 540 500 349 171 156 450 328 459 331 256 267 306 332 325 336 319 193.5 121 163 100 170 103 339 240.3 317 184 120 325 182 107 346 325 308 325 331 200 340 319 339 331

527 567 317 a 296 270 669 a 651 463 216 178 611 411 625 410 325 305.9 371 387 379 (336) 368 (193.5) (121) (163) (100) (170) (103) 375 363.9 (317) (184) 120 (325) (182) (107) 386 332 332 336 (331) (200) 322 336 312 336

946,947 948,949

[PS4]3

886 1000 a 805 956 1017 535 a 512 548 1012 a 988 878 419 380 1105 856 1119 878 853 879.2 770 846 800 804 780 470 365 421 316 399 277 863 775.8 837 472 340 838 455 309 902 820 778 912 920 486 790 921 845 804

Compund [BO4]

5

[CS4]

4

3

[PSe4]

3

[NO4]

843

[AsO4]3 [AsS4]3 [SbS4]3 [SO4]2 [SeO4]2 [ClO4] [BrO4] [IO4] XeO4 [TiO4]4 [ZrO4]4 [HfO4]4 [VO4]3 [VO4]4 [VS4]3 [VSe4]3 [NbS4]3 [NbSe4]3 [TaS4]3 [TaSe4]4 [CrO4]2 [CrO4]3 [MoO4]2 [MoS4]2 [MoSe4]2 [WO4]2 [WS4]2 [WSe4]2 [MnO4]2 [MnO4]2 [MnO4]3 [TcO4] [ReO4] [ReS4] [FeO4]2 RuO4 [RuO4] [RuO4]2

837 386 366 983 822 928 801 791 775.7 761 792 796 826 818 404.5 (232) 408 239 424 249 833 844.0 897 458 255 931 479 281 834 812 789 912 971 501 832 885.3 830 840

950 951 952b 953 953 953 946 954,946 953 955,956 957 958 959 959 959 960 959 961 961 961 961 961 961 954 955 960 962 963 960 962 961 964,965 959 966 960 960 967 959 968 959 959

203

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

TABLE 2.6b. (Continued ) 2.6e. (Continued Compund

~n1

~n2

~n3

~n4

Ref.

OsO4 XeO4 [CoO4]4 {B (OH)4] c [Al (OH)4] c [Zn (OH)4] c

965.2 776 670 754 615 470

333.1 267 320 379 310 300

960.1 878 633 945 (720) (570)

322.7 305 320 533 (310) (300)

969 969a 970 971,972 973 973

a b c

Splitting may suggest D2d symmetry in the crystalline state. Na3 [NO4]3 was obtained by the reaction of NaNO3 and Na2O. Only MO4 skeletal vibrations are listed for this ion.

The IR spectrum and normal coordinate analysis of the SO4 radical produced by the reaction of SO3 with atomic oxygen in inert gas matrices are suggestive of the Cs structure rather than the C2v structure [977]:

The IR spectra of H2SO4 and its deuterated derivatives in Ar matrices have been assigned on the basis of C2 symmetry ((HO)2SO2) [978]. The IR spectrum of the ClO4

Fig. 2.20. Raman spectra of K2SO4 in (a) the solid state and (b) aqueous solution.

204

APPLICATIONS IN INORGANIC CHEMISTRY

radical in inert gas matrices indicates a C3v structure with one long ClO bond as the C3 axis. The v(ClO) of this bond is at 874 cm1, which is much lower than those of the other three bonds (1234 and 1161 cm1) [979]. The IR/Raman spectra of MþClO4 salts in nonaqueous solvents suggest the formation of a unidentate ion pair for M ¼ Liþ, and a bidentate ion pair for M ¼ Naþ [980]. The RR spectra of highly colored ions such as CrO42 [981], MoS42 [982], VS43 [983], and MnO4 [964,981] have been measured. The n3(IR) bands of gaseous RuO4 [968] and XeO4 [984] exhibit complicated band contours consisting of individual isotope peaks of the central atoms. M€ uller and coworkers [974,985,986] reviewed the vibrational spectra of transition-metal chalcogen compounds. Basile et al. [932] carried out normal coordinate analysis on more than 60 compounds of these types. 2.6.2. Tetrahedral ZXY3, Z2XY2, and ZWXY2 Molecules If one of the Yatoms of an XY4 molecule is replaced by a Z atom, the symmetry of the molecule is lowered to C3v. If two Yatoms are replaced, the symmetry becomes C2v. In ZWXY2 and ZWXYU types, the symmetry is further lowered to Cs and C1, respectively. As a result, the selection rules are changed as shown in Table 2.6f. The number of infrared-active vibrations is six for ZXY3 and eight for Z2XY2. Table 2.6g lists the vibrational frequencies of ZXY3-type molecules. The SO stretching frequency of the [OSF3]þ ion in [OSF3]SbF6 (1536 cm1) is the highest that has been observed, and corresponds to a force constant of 14.7 mdyn/A [1006]. It is also interesting to note that the structure the [OXeF3]þ ion is Cs [726] whereas that of the [OXeF3] ion is C2v [727] as shown below:

TABLE 2.6f. Correlation Tablea for Td, C3v, C2v, and C1 Point Group

n1

n2

n3

Td C3v

A1 (R) A1 (IR, R)

E (R) E (IR, R)

C2v

A1 (IR, R)

A1 (IR, R)þ A2 (R)

F2 (IR, R) A1 (IR, R)þ E (IR, R) A1 (IR, R)þ B1 (IR, R)

C1

A (IR, R)

2A (IR, R)

3A (IR, R)

a

See Appendix IX.

n4 F2 (IR, R) A1 (IR, R)þ E (IR, R) A1 (IR, R)þ B1 (IR, R)þ B2 (IR, R) 3A (IR, R)

TETRAHEDRAL AND SQUARE–PLANAR FIVE-ATOM MOLECULES

205

TABLE 2.6g. Vibrational Spectra of ZXY3 Molecules (cm1) C3v ZXY3

[HGaCl3] [OCF3] FCCl3 BrCC3 ICCl3 FSiH3 H28 SiF3 FSiCI3 FSiBr3 FSiI3 BrSiF3 79 BrSiH3 [SSiCl3] [SGeCl3] HGeCl3 ONF3 [HPC13]þ [ClPBr3]þ [BrPCl3]þ [HPF3]þ OPF3 OPCI3 OPBr3 [FPO3] SPF3 [OSF3]þ NSF3 ISCl3 [HSO3] [FSO3] [ClSO3] [SSO3] [SeSO3]2 [FSeO3] FC1O3 FBrO3 [NClO3]2 OVF3 OVC13 OVBr3 NVCl3 ONbCl3 [FCrO3] [ClCrO3] [BrCrO3] [OMoS3]2 [OMoSe3]2 [SMoO3]2 [SMoSe3]2

~ n1 (A1) n (XY3)

~n2 (A1) n (XZ)

~n3 (A1) d (XY3)

~n4 (E) n (XY3)

~n5 (E) d (XY3)

~n6 (E) rr (XY3)

Ref.

349.5 813 537.6 419.9 390 — 855.8 465 318 242 858 2198.8 395 356 418.4 743 528 285 399 948 873 486 340 1001 — 909 772.6 482 1038 1142 1042 995 995 896 1062.6 875.2 815 721.5 408 271 352 395 911 907 895 461 293 900 —

1960.4 1560 1080.4 730.7 684 990.9 2315.6 948 912 894 505 929.8 685 505 2155.7 1691 2481 587 582 2544 1415 1290 1261 794 985 1540, 1532 1522.9 293 2588 862 381 446 310 580,603 716.3 605.0 1256 1057.8 1035 1025 1033 997 635 438 260 862 858 472 471

167,9 595 351.3 246.5 224 875.0 425.3 239 163 115 288 430.8 220 163 181.8 528 208 149 217 480 473 267 173 534 696 535 524.6 242 629 571 601 669 645 392 549.7 (354) 594 257.8 165 120 — 106 338 295 230 183 120 318 121

320.6 960 847.8 785.2 755 — 997.8 640 520 424 940 2209.3 492 360 708.6 883 634 500 657 1090 990 581 488 1125 947 1063 814.6 493 1200 1302 1300 1123 1120 968,974 1317.9 974 870 806 504 400 430 448 955 954 950 470 355 846 342

609.3 422 242.8 290.2 284 962.2 843.6 282 226 194 338 996.4 — — 454 558 280 172 235 989 485 337 267 — 442 508 432.3 140 509 619 553 541 530 409 589.8 (376) 623 204.3 129 83 — 225 370 365 350 183 120 318 121

— 576 395.3 188.0 188 730.0 306.2 167 110 71 200 632.4 225 178 145.0 400 653 120 159 380 345 193 118 382 405 387 346.2 255 1123 424 312 335 — 301 403.9 (296) 457 308 249 212 — 110 261 209 (175) 263 188 239 —

987 988,989 990 990 991 992 993 994 994 994 994 995 996 996 997 998 999 1000 1000 1001,1001a 1002 1002 764,1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015,1016 1017,1018 1019 1020 1020 1020 1021 1021a 1022 1023 1024 1023 1025 1026 985

(continued)

206

APPLICATIONS IN INORGANIC CHEMISTRY

TABLE 2.6b. 2.6g. (Continued (Continued ) C3v ZXY3 2

[SeMoS3] [OWS3]2 [OWSe3]2 [SWSe3]2 FMnO3 ClMnO3 FTcO3 ClTcO3 FReO3 ClReO3 BrReO3 [NReO3]2 [SReO3] [NOsO3] [FOsO3]þ

~ n1 (A1) n (XY3)

~n2 (A1) n (XZ)

~n3 (A1) d (XY3)

~n4 (E) n (XY3)

~n5 (E) d (XY3)

~n6 (E) rr (XY3)

Ref.

349 474 292 468 903.6 889.9 962 948 1013.2 994 997 878 948 892.5 1002

458 878 878 281 715.5 458.9 696 451 701 436,427 350 1022 528 1026.2 745

— 182 (120) 108 339.0 305 317 300 305.5 291 195 315 322 310.0 333

473 451 312 311 950.6 951.6 951 932 978.3 963 963 830 906 872.0 992

150 182 (120) 150 380.0 365 347 342 346.9 337 332 273 322 299.7 372

183 264 194 108 264.0 200 231 197 234.2 192 168 380 (240) 371.5 238

1027 1023 1025 1027 1028,1029 1026,1029 1030 1031 1032 1032 1022 1033 1022 1034 1035

The first example of the trifluorosulfite anioin (OSF3) was obtained as the Me4Nþ salt, and its IR and Raman spectra assignments based on a theoretically predicted pseudo-trigonai–bipyramidal structure with two long SF bonds in the axial direction and one F, one O atom and a sterically active lone-pair electrons at the triangular corners of the equatorial plane [1036]. The Raman spectrum of the [NVO3]4 ion has been assigned as a pseudo-tetrahedral anion of Cs symmetry [1037]. Vibrational spectra have been reported for a number of mixed halogeno complexes. Some references are as follows: [AlClnBr4n] [1038], SiFnCl4n [1039], SiClnBr4n [1040], and [FeClnBr4n] [1041]. It is interesting to note that the SiFClBrI molecule exhibits the SiF, SiCl, SiBr, and SiI stretching bands at 910, 587, 486, and 333 cm1, respectively [1042]. The vibrational spectrum of OClF3 suggests a trigonal–bipyramidal structure similar to that of the [OXeF3]þ ion shown above [1043]. Table 2.6h lists the vibrational frequencies of tetrahedral Z2XY2 molecules. The vibrational spectrum of O2XeF2 can be interpreted on the basis of a trigonal– bipyramidal structure in which two F atoms are axial, and two O atoms and a pair of electrons are equatorial [1059]. The structures of [O2ClF2] [1057] and [O2BrF2] [1058] are similar to that of O2XeF2, but that of [O2ClF2]þ [1067] is pseudo-tetrahedral. The gas-phase Raman spectrum of Cl2TeBr2 at 310 C is indicative of the C1 symmetry shown below [1068]. Table 2.6i lists the vibrational frequencies of tetrahedral ZWXY2 molecules. Other references are as follows: SFPC12 [1075], FOPCl2 [1076], FOPB2 [1077], ClOPBr2 and BrOPCl2 [1003], FBrSO2 [1078], and FCIClO2 [1079].

207

þ

[F2NH2] [O2PH2] F2CI2 H2SiCl2 H2SiBr2 F2SiBr2 H2GeF2 H2GeCl2 H2GeBr2 H2GeI2 [Cl2PBr2]þ [O2PF2]þ O2SF2 O2SCl2 F2SCl2 [O2SeF2] [O2ClF2] [O2BrF2] O2XeF2 [O2VF2] [O2VCl2] [O2NbS2]3

Z2XY2

2637 2365 605 942 925 414 860 840 848 821 326 910 847.4 405 527 396 363 374 490 631 431 464

~n1 (A1) n (XY2)

1060 1046 1064 2221 2206 891 2155 2132 2122 2090 584 1179 1265.5 1182 592 859 1070 885 845 969 968 897

~n2 (A1) n (XZ2) 1543 1160 272 514 393 270 720 404 298 220 191 269 388.7 218 — 241 198 201 198 199 128 246

~n3 (A1) d (XY2) 528 470 112 188 122 115 (270) 163 104 96 132 567 552.0 560 — 445 559 429 333 365 323 356

~n4 (A1) d (XZ2)

TABLE 2.6h. Vibrational Frequencies of Z2XY2 Molecules (cm1)

— 930 251 710 688 187 (664) 648 105 628 150 — 390.8 388 — — 480 405 — — 200 —

~n5 (A2) rt (XY2) 2790 2308 740 868 828 540 814 772 757 706 518 962 884.1 362 533 — 510 448 578 664 436 514

~n6 (B1) n (XY2) 1176 820 200 566 456 241 720 420 324 294 173 492 545.0 380 — — 337 339 313 285 277 (297)

~n7 (B1) rw (XY2) 1036 1180 1110 2221 2232 974 2174 2155 2138 2110 616 1269 1497.0 1414 770 833 1221 912 902 960 958 872

~n8 (B2) n (XZ2) 1487 1093 278 602 — 257 596 533 492 451 201 528 538.2 282 — 304 337 314 313 265 194 (271)

~n9 (B2) rr (XY2)

(continued)

1044 1045 1046 1047 1048 1049 1050 1087 1050 1050 1022 1051 1052,1053 1054 1055 1056 1057 1058 1059 1060 1060 1061

Ref.

208

~n1 (A1) n (XY2)

727 475 399.0 262 473 283 454 282

Z2XY2

O2CrF2 O2CrCl2 O2CrBr2 O2MoBr2 2 ½O92 2 MoS2 [O2MoSe2]2 [O2WS2]2 [O2WSe2]2

TABLE 2.6b. 2.6h. (Continued (Continued )

1006 995 982.6 995 819 864 886 888

~n2 (A1) n (XZ2) 208 140 — 147 200 114 196 116

~n3 (A1) d (XY2) 364 356 (305) 373 307 339 310 319

~n4 (A1) d (XZ2) (259) (224) — — 267 251 280 235

~n5 (A2) rt (XY2) 789 500 403.5 338 506 353 442 329

~n6 (B1) n (XY2) 304 257 — 161 267 251 280 235

~n7 (B1) rw (XY2) 1016 1002 995.4 970 801 834 848 845

~n8 (B2) n (XZ2)

274 215 — 184 246 — 235 156

~n9 (B2) rr (XY2)

1062 1063 1064 1065 1066 1061 1061 1061

Ref.

209

b

X denotes the central atom. These assignments may be interchanged.

623 549 561 477 907 915 888 887 432 372 552 500 355 320 461 459

900 949 884 938 546 479 472 377 545 493 391 333 478b 473 360b 317b

OClPF2 SClPF2 OBrPF2 SBrPF2 OFPCl2 SFPCl2 OFPBr2 SFPBr2 OBrPCl2 SBrPCl2 OClPBr2 SClOBr2 [OSeMoS2]2 [OSeWS2]2 [OSMoSe2]2 [OSWSe2]2

a

~n2 (A0 ) n (XW)

~n1 (A0 ) ns (XY2)

ZWXY2a 1384 735 1380 719 1358 750 1337 713 1285 743 1275 729 869 879 865 882

~n3 (A0 ) n (XZ) (419) 394 (413) 389 205 193 133 129 242 (230) 130 121 190 190 — —

~n4 (A0 ) ds (XY2)

TABLE 2.6i. Vibrational Frequencies of ZWXY2 Molecules (cm1)

(274) 363 (240) 288 330 328 273 274 172 150 209 196 — — — —

~n5 (A0 ) ds (WXY) 412 209 316 175 382 268 304 218 285 206 291 190 273 265b — —

~n6 (A0 ) ds (ZXY2) 960 925 947 911 626 574 536 470 580 536 492 436 467b 458 320b 312b

~n7 (A00 ) na (XY2) 274 252 240 231 253 193 220 162 161 150 157 136 — — — —

~n8 (A00 ) d (ZXW) 419 613 413 297 374 317 290 254 327 230 271 205 273 255b — —

~n9 (A00 ) da (XY2)

639,1069 639,1069 639,1069 639,1069 639,1069 639,1069 639,1069 639,1069 1070 1070 1070 1070 1071,1072 1072,1073 1074 1074

Ref.

210

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.21. Normal modes of vibration of square–planar XY4 molecules.

2.6.3. Square–Planar XY4 Molecules (D4h) Figure 2.21 shows the seven normal modes of vibration of square–planar XY4 molecules. Vibrations n3, n6, and n7 are infrared-active, whereas n1, n2, and n4 are Raman-active. However, n5 is inactive both in infrared and Raman spectra. In the case of the [PdCl4]2 ion, the n5 was observed at 136 cm1 by inelastic neutron scattering (INS) spectroscopy [1088]. Table 2.6j lists the vibrational frequencies of some ions belonging to this group. Chen et al. [1088a] also reported the IR and Raman spectra of K2[MX4] (M ¼ Pt or Pd and X ¼ Cl or Br) and band assignments on the basis of normal coordinate analysis. XeF4 (Sec. 1.11) is an unusual example of a neutral molecule that takes a square– planar structure. This structure is predicted by the valence shell electron pair repulsion (VSEPR) theory [1089] which states that two lone-pairs in the valence shell occupy the axial positions because they exert greater mutual repulsion than

211

b

a

505 523 288 554.3 347 212 148 303 188 330 208 155

~n1 (A1g) ns (XY)

288 246 128 218 171 102 75 164 102 171 106 85

~n2 (B1g) d (XY2) 425 317 — 291 — — — 150 114 147 105 105

~n3 (A2u) p 417 449 261 524 324 196 110 275 172 312 194 142

~n4 (B2g) na (XY) 680–500 580–410 266 586 350 252b 192 321 243 313 227 180

~n6 (Eu) nd (XY) — (194) — (161) 179 110b 113 161 104 165 112 127

~n7 (Eu) dd (XY2)

For these molecules n5 is inactive. The designations B1g and B2g may be interchanged, depending on the definition of symmetry axes involved. From Ref. 1087.

[ClF4] [BrF4] [ICl4] XeF4 [AuCl4] [AuBr4] [AuI4] [PdCl4]2 [PdBr4]2 [PtCl4]2 [PtBr4]2 [Ptl4]2

XY4

TABLE 2.6j. Vibrational Frequencies of Square-Planar XY4 Molecules (cm1)a

1080 1080 1081 1082,1083 1084,1085 1084,1085 1084 1085,1086 1084,1086 1085,1086 1085,1086 1084,1086

Ref.

212

APPLICATIONS IN INORGANIC CHEMISTRY

single-bond pairs:

The structures of fluorides, oxides, and oxyfluorides of xenon discussed in other sections can also be rationalized on the basis of VSEPR theory. Figure 2.22

Fig. 2.22. (a) IR [1090] and (b) Raman [1082] spectra of XeF4 vapor.

TRIGONAL–BIPYRAMIDAL AND TETRAGONAL–PYRAMIDAL XY5 AND RELATED MOLECULES

213

shows the IR and Raman spectra of gaseous XeF4 obtained by Claassen et al. [1082,1090]. Bosworth and Clark [1091] measured the relative intensities of Raman-active fundamentals of some of these ions, and calculated their bond polarizability derivatives and bond anisotropies. They [1087] also measured the resonance Raman spectra of several [AuBr4] salts in the solid state, and observed progressions such as nn1 (n ¼ 1–9) and n2þ nn1 (n ¼ 1–5). For normal coordinate analyses of square–planar XY4 molecules, see Refs. [1081], [1092], and [1093].An empirical relationship between the antisymmetric MCl stretching frequency [na(MCl), cm1] and the MCl distance (RMCl, A) has been derived using IR and structural data for a large number of planar Pt(II) and Pd(II) complexes of MCI2L2, MC12LL0 , and MC1L3 types (L: unidentate ligand) ½na ðM ClÞ2 ¼

P ðRMCl 1:6Þ3

where P is 38988 for Pt(II) and 41440 for Pd(II) complexes [1094]. The IR and Raman spectra of the [AuCl4nBrn] ions (n ¼ 1–3) [1095] and Raman spectra of the square–planar tetrahydrido[PtH4]2 and related ions [[1095]a] are reported.

2.7. TRIGONAL–BIPYRAMIDAL AND TETRAGONAL–PYRAMIDAL XY5 AND RELATED MOLECULES An XY5 molecule may be a trigonal bipyramid (D3h) or a tetragonal pyramid (C4v). 00 If it is trigonal–bipyramidal, only two stretching vibrations (A2 and E0 ) are infraredactive. If it is tetragonal–pyramidal, three stretching vibrations (two A1 and E) are infrared-active. As discussed in Sec. 1.11, however, it is not always possible to make clear-cut distinctions of these structures based on the selection rules since practical difficulties arise in counting the number of fundamental vibrations in infrared and Raman spectra. 2.7.1. Trigonal–Bipyramidal XY5 Molecules (D3h) Figure 2.23 shows the eight normal vibrations of a trigonal–bipyramidal XY5 0 00 molecule. Six of these eight (A1 , E0 E00 ) are Raman-active, and five (A2 and E0 ) are infrared-active. Three stretching vibrations (n1, n2, and n5) are allowed in the Raman, whereas two (n3 and n5) are allowed in the infrared. Table 2.7a lists the observed frequencies and band assignments of trigonal-bipyramidal XY5 molecules. The IR spectrum of the ion-paired complex, Csþ[GeF5], in inert gas matrices has been reported [1116]. It takes a trigonal–bipyramidal structure perturbed by the presence of the Csþ ion in its vicinity.

214

APPLICATIONS IN INORGANIC CHEMISTRY

Y′ Y X Y

Y

Y′ ν1(A′1) νs(XY3)

ν5(E′1) νa(XY3)

ν2(A′1) νs(XY′2)

ν3(A′′2) νa(XY′2)

ν4(A′′2) π (XY3)

ν6(E′′) δ(XY3)

ν7(E′) δ(XY ′2)

ν8(E′′) ρr(XY′2)

Fig. 2.23. Normal modes of vibration of trigonal–bipyramidal XY5 molecules.

The majority of trigonal–bipyramidal compounds show the following frequency trend: v5 Equatorial stretch

>

v3 > Axial stretch

v1 Equatorial stretch

>

v2 Axial stretch

Although the IR spectrum of MoCl5 in the gaseous phase has been assigned based on a trigonal–bipyramidal structure [1107] the corresponding spectrum in N2 matrices suggests a tetragonal–pyramidal structure of C4v symmetry [1117]. This is supported by the observation that it shows only two prominent IR banos at 473 (A1) and 408 cm1 (E) in the v(MoCl) region. Normal coordinate analyses on trigonal–bipyramidal XY5 molecules have been carried out by several investigators [1118–1121]. These calculations show that equatorial bonds are stronger than axial bonds. Most neutral XY5 molecules are dimerized or polymerized in the condensed phases. The molecules MoCl5, [1122], NbCl5 [1113,1123], TaCl5 [1124], and WCl5 [1125] are dimeric in the liquid and solid states (D2h), whereas NbF5 and TaF5 are known to be tetrameric in the crystalline state [1126]. Some of these molecules are polymerized even in the gaseous phase. For example, SbF5 is monomeric (D3h) at 350 C but polymeric at 140 C in the gaseous phased [1127], and NbF5 and TaF5 are polymeric in the gaseous phase if the temperature is below 350 C. [1128]. Although PC15 exists as a D3h molecule in the gaseous and liquid states, it has an ionic structure consisting of [PC14]þ[PC16] units in the crystalline state, as

215

Gas Solid Solid Solid Sol’na Solid Gas Matrix Gas Gas Gas Matrix Gas

SbCl5 [SeO5]4 [CuCl5]3 [CdCl5]3 [TiCl5] [TiCl5] VF5 NbCl5 NbBr5 TaCl5 TaBr5 MoF5 MoCl5

b

Nonaqueous solution. May be assigned to n~7. c May be assigned to n~8.

a

Sol’n Sol’na Solid Solid Solid Gas Sol’na Gas Sol’na Gas

a

Phase

[SiF5] [SiCl5] [GeCl5] [SnCl5] [SnBr5] PF5 PCl5 AsF5 AsCl5 SbF5

Molecule

355 756 260 251 348 — 719 (349) 234 406 240 — 390

708 372 348 340 — 817 392 733 369 667

~n1

309 736 — — 302 — 608 (349) 178 324 182 — 313

519 — 236 — — 640 281 642 295 264

~n2

— 775 268 236 355 346 784 396 288 — — 683 —

785 395 310 314 208 944 443 784 385 —

~n3

— 674 — — 178 170 331 126 (93) — — — —

481 271 200 160 106 575 300 400 184 —

~n4 874 550 395 350 256 1026 580 809 437 716 710 400 344 170b 157b 411 385 810 444 315 — — 713 418

~n5

TABLE 2.7a. Vibrational Frequencies of Trigonal-Bipyramidal XY5 Molecules (cm1)

173 806 95c 98c 190 212 282 159 119 181 110 261 200

449 250 200 150 111 532 272 366 220 498

~n6

58 — — — 66 (83) (200) 99 67 54 70 112 100

— — — 66 — 300 102 123 83 90

~n7

120 355 — — 166 — 350 (139) 101 127 93 — 175

— — — 169 — 514 261 388 213 228

~n8

1107 1108 1109 1109 1110 1100 1111,1112 1113 1107,1114 1107 1107 1115 1107

1096 1097 1098 1099 1100 1101 1102,1103 1104 1105 1106

Ref.

216

APPLICATIONS IN INORGANIC CHEMISTRY

proved by Raman spectroscopy [1129]. The importance of the v7 vibration in the intramolecular conversion of pentacoordinate molecules has been discussed by Holmes [1130].

Vibrational spectra of mixed halogeno compounds such as PFnCl5n, [1131,1132] PF3X2(X ¼ Cl, Br), [1133], PHnF5n, [1134,1135], PH3F2, [1136,1137] AsCl3F2 [1138] and AsCl4F [1139] have been assigned. The structure of OSF4 is pseudo-trigonal–bipyramidal with two fluorines and one oxygen occupying the three equatorial positions. Thus, the IR and Raman spectra were assigned on the basis of C2v symmetry [1140]. Since the OPF4 ion is isoelectronic with OSF4, the structure is similar to that of OSF4, and assigment of its Raman spectrum is based on the same symmetry [1141]. The FOsO4 ion assumes a distorted trigonal–bipyramidal structure with three oxygen atoms occupying the equatorial positions, and the IR and Raman spectra were assigned under Cs symmetry [1142]:

Although the Raman spectrum of Csþ [trans-XeO2F3] was reported previously [1143], a later study showed that its structure is Csþ [XeO2F3nXeF2], in which the two oxygen atoms are cis to each other [1144]. In the [F(cis-OsO2F3)2]þ ion, a fluorine bridge connects two trigonal bipyramidal cations in which two oxygen atoms and one fluorine atom occupy the equatorial plane and two fluorine atoms are in axial positions. The Raman spectra of its AsF6 and Sb2F11 salts have been assigned on the basis of

TRIGONAL–BIPYRAMIDAL AND TETRAGONAL–PYRAMIDAL XY5 AND RELATED MOLECULES

217

such a dimeric structure [1145]. MO2F3SbF5 (M ¼ Te, Re) consists of infinite chains of alternating MO2F4 and SbF6 units in which the bridging fluorine atoms on the antimony are trans to each other. The Raman spectra of these and related compounds have been reported [1146]. 2.7.2. Tetragonal–Pyramidal XY5 and ZXY4 Molecules (C4v) Figure 2.24 shows the nine normal modes of vibration of a tetragonal–pyramidal ZXY4 molecule. Only A1 and E vibrations are infrared-active, whereas all nine vibrations belonging to the A1, B1, B2, and E species are Raman–active. Table 2.7b lists the vibrational frequencies of tetragonal–pyramidal XY5 and ZXY4 molecules. In the majority of XY5 molecules, the axial stretching frequency (n1) is higher than the equatorial stretching frequencies (n2, n4, and n7). This is opposite to the trend found for trigonal–bipyramidal XY5 molecules, discussed in the preceding section. For normal coordinate analyses on these compounds, see Refs. [1150], [1153], [1170], and [1173]. An adduct, XeF6:BiF5, is formulated as [XeF5]þ[BiF6] in the solid state; its [XeF5]þ cation is tetragonal–pyramidal [1157]. This should be contrasted to the [XeF5] anion which is pentagonal planar, as shown in the following section. In the OCrF4KrF2 adduct, KrF2 coordinates to the vacant axial position of OCrF4 through a

Z Y

Y X

Y

ν1(A1) ν (XZ)

Y

ν4(B1) νa(XY4)

ν7(E) νd(XY4)

ν2(A1) νs(XY4)

ν3(A1) π(ZXY4)

ν5(B1) δa(XY4)

ν6(B2) δs(XY4)

ν8(E) πd (XY4)

ν9(E) δd(XY4)

Fig. 2.24. Normal modes of vibration of tetragonal–pyramidal ZXY4 molecules.

218

294 557 445 796 666 616 363 723 682 698 740 660

837 1216

932

[OTeF4]2 [OClF4]

[OBrF4]

A1 ~n1

[InCl5] [SbF5]2 [SbCl5] [SF5] [SeF5] [TeF5] [TeCl5] ClF5 BrF5 IF5 NbF5 [XeF5]þ

2

XY5 or ZXY4

525

461 462

283 427 285 522 515 483 254 545 570 593 686 598

A1 ~n2

312

— 339

140 278 180 469 332 280 136 487 365 315 513 355

A1 ~n3

459

287 388 420 (435) 460 505 270 498 535 575 — 635 628 390 350

B1 ~n4

236

— —

193 — — 269 236 231 98 317 (281) (257) — 242

B1 ~n5

248

190 283

165 220 117 342 282 183 82 385 312 273 — 291

B2 ~n6

335 600,550 506 483

274 375,347 300 590 480 452 246 732 644 640 729 650

E ~n7

TABLE 2.7b. Vibrational Frequencies of Trigonal–Pyramidal XY5 and ZXY4 Molecules (cm1)

108 307 255 (435) 399 162 90 500 414 374 261 415 395 — 415,394 ) 421 407 398

E n~8

194 164

129 213

143 142 90 241 202 342 169 301 237 189 103 218

E n~9

1160,1161

1158 1159

1147 1148 1149 1150 1150 1148,1151 1151 1152–1154 1153,1155 1148,1153 1156 1157

Ref.

219

[OIF4] OXeF4 OCrF4 OMoF4 OMoCl4 [OMoCl4] OWF4 OReF4 [OReI4] [NRuCl4] [NRuBr4] [NOsCl4] [NOsBr4] ORuF4 OOsF4

XY5 or ZXY4

888 920 1028 1050 1015 1008 1055 1082 1014 1092 1088 1123 1119 1060 1080

A1 ~n1

TABLE 2.7b. (Continued )

533 567 686 714 450 354 733 685 169 346 224 358 162 685 695

A1 ~n2

273 285 277 267 143 184 248 — 85 197 156 184 122 — —

A1 ~n3 475 527 — — 400 327 631 — 183 304 187 352 156 — —

B1 ~n4 — (230) — — 148 158 328 — — 154 103 149 110 — —

B1 ~n5 214 233 — — 220 167 291 — — 172 128 174 120 — —

B2 ~n6 485 608 744 708 396 364 698 710 240 378 304 365 220 713/710 685

E ~n7 365 365 320 306 256 240 298 302 181 267 211 271 273 310 319

E n~8 124 161 271 238 172 114 236 245 — 163 98 132 98 — —

E n~9 1162 1153 1163,1164 1165–1167 1168,1169 1170 1166,1167 1171 1172 1170 1170 1170 1170 1171 1171

Ref.

220

APPLICATIONS IN INORGANIC CHEMISTRY

predominantly covalent bond [1174]:

As a result, the KrF2 vibrations are markedly shifted in frequency relative to those of free KrF2 [305]: n1

n2

n3

Free KrF2

449 233

596; 580

cm1

Bound KrF2

487 176

550; 542

cm1

2.7.3. Pentagonal–Planar XY5 Molecules (D5h) The first example of a pentagonal–planar XY5 type molecule was the XeF5 anion [1175]. This anion was obtained as the 1:1 adduct of XeF4 with N(CH3)4F, CsF or other alkali fluorides, and its structure was confirmed by X-ray analysis, IR/Raman, and NMR spectroscopy. Figure 2.25 shows the structure of the XeF5 anion, which can be derived by replacing the two axial fluorine ligands of the pentagonal–bipyramidal IF7 molecule with two sterically active lone-pair electrons. The difference in structure between this anion and the XeF5 cation discussed in the preceding section may be understood based on the VSERP theory [1089].

Fig. 2.25. Structures of the XeF5 and XeF3þ ions.

OCTAHEDRAL MOLECULES

221

The XeF5 anion has 12 normal vibrations, which are classified into 00 0 0 00 A1 ðRÞþA2 ðIRÞþ2E1 ðIRÞþ2E2 ðRÞþE2 (inactive) under D5h symmetry. Thus, only 00 0 0 0 three vibrations (A2 and 2E1 ) are IR-active and only three vibrations ðA 1 þ2E 2 Þ are Raman-active. In agreement with this prediction, the observed Raman spectrum of 0 (CH3)4N[XeF5] exhibits three bands at 502 (symmetric stretch, A1 ), 423 (asymmetric 0 0 1 stretch, E2 ), and 377 cm (in-plane bending, E2 ). As discussed in Sec. 1.11, the trigonal–bipyramidal and tetragonal-pyramidal structures can be ruled out since many more IR- and Raman-active bands are expected for these structures. The second example of a pentagonal planar XY5-type molecule is the IF52 ion, which is isoelectronic with the XeF5 ion. It was obtained as the N(CH3)4þ salt from the reaction of N(CH3)4IF4 and N(CH3)4F in CH3CN solution, and the IR and Raman spectra were assigned under D5h symmetry [1176]. 0

2.8. OCTAHEDRAL MOLECULES 2.8.1. Octahedral XY6 Molecules (Oh) Figure 2.26 illustrates the six normal modes of vibration of an octahedral XY6 molecule. Vibrations n1, n2, and n5 are Raman-active, whereas only n3 and n4 are infrared-active. Since n6 is inactive in both, its frequency is estimated by several

Y

Y

X

Y

Y Y

Y ν1(A1g) ν (XY)

ν4(F1u) δ (YXY)

ν2(Eg) ν (XY)

ν3(F1u) ν (XY)

ν5(F2g)

ν6(F2u)

δ (YXY)

δ (YXY)

Fig. 2.26. Normal modes of vibration of octahedral XY6 molecules.

222

APPLICATIONS IN INORGANIC CHEMISTRY

methods including the analysis of nonfundamental frequencies and rotational fine structures of vibration spectra. The G and F matrix elements of the octahedral XY6 molecule are listed in Appendix VII. Table 2.8a lists the vibrational frequencies of a number of hexahalogeno compounds. In general, the order of the stretching frequencies is n1 > n3 ( n2 or n1 < n3 ( n2, depending on the compound. The order of the bending frequencies is n4 > n5 > n6 in most cases. Parker et al. [1263] measured the IR, Raman, and INS (inelastic neutron scattering) spectra of the [FeH6]4 ion and its deuterated analog listed at the end of Table 2.8a. These hexahydrido complexes are unsual in that the v1(A1g) and v2(Eg) frequencies are very close, although they are well separated in other XY6 molecules. This result indicates that the stretching–stretching interaction force constant (frr in Appendix VII.6 is very small in the case of Fe–H(D) bonds. The v6 frequencies of these compounds listed in Table 2.8a were obtained by INS spectroscopy. Several trends are noted in the octahedral XY6 molecules listed in Table 2.8a, as described in Secs. 2.8.1.1–2.8.1.12. 2.8.1.1. Mass of Central Atom Within the same family of the periodic table, the stretching frequencies decrease as the mass of the central atom increases, for example [AlF6]3 v~1 ðcm1 Þ v~3 ðcm1 Þ

541 568

[GaF6]3 > >

[InF6]3 > >

535 481

497 447

[TiF6]3 > >

478 412

The trend in n1 directly reflects the trend in the stretching force constant (and bond strength) since the central atom is not moving in this mode. In n3, however, both X and Yatoms are moving, and the mass effect of the X atom cannot be ignored completely. 2.8.1.2. Oxidation State of Central Atom Across the periodic table, the stretching frequencies increase as the oxidation state of the central atom becomes higher. Thus we have [AlF6]3 1

v~1 ðcm Þ v~3 ðcm1 Þ

541 568

[PF6]

[SiF6]2 <

584 .578

[VF6]3 > >

533 511

As in many other cases, the higher the oxidation state, the higher the frequency. The bending frequencies do not exhibit clear-cut trends.

223

[AlF6] [GaF6]3 [InF6]3 [InCl6]3 [TlF6]3 [TlCl6]3 [TlBr6]3 [SiF6]2 [GeF6]2 [GeCl6]2 [SnF6]2 [SnCl6]2 [SnBr6]2 [SnI6]2 [PbCl6]2 [PF6] [PCl6] [AsF6] [AsCl6] [SbF6] [SbCl6] [SbCl6]3 [SbBr6] [SbBr6]3 [SbI6]3 [BiF6] [BiCl6]3 [BiBr6]3 [BiI6]3

3

Molecule

541 535 497 277 478 280 161 663 671 318 592 311 190 122 281 756 360 689 337 668 330 327 192 180 107 590 259 156 114

~n1

(400) (398) (395) 193 387 262 153 477 478 213 477 229 144 93 209 585,570 283 573 289 558 282 274 169 153 96 547 215 130 103

~n2 568 481 447 250 412 294 190,195 741 624 310 559 303 224 161 262 865,835 444 700 333 669 353 — 239,224 180 108 585 172 128 96

~n3

TABLE 2.8a. Vibrational Frequencies of Octahedral XY6 Molecules (cm1)

387 298 226 157 202 222,246 134,156 483 365,355 213 300 166 118 84 142 559,530 285 385 220 350 180 — 119 107 82 — 130 75 (59)

~n4 322 281 229 (149) 209 155 95 408 333/325 191 252 158 109 78 139 480 468 238 375 202 294 175 137 103,78 73 54 231/247 115 62 54

~n5 (228) (198) (162) — (148) (136) (80) — — — — — — — — — — (252) — — — — — — — — — — —

~n6a

(continued)

1177,1178 1178,1179 1177,1178 1180 1177,1178 1181 1181 1182 1183 1184 1182 1185 1186 1187 1185 1188 1184,1189 1182 1184 1182,1190 1191 1192 1193 1194,1195 1194,1195 1190,1196 1194,1195 1194,1195 1194,1195

Ref.

224

~n2

643 658 255 155 672 250 192 153 384 630 660 454 530 375 67 382 334 380 370 265 (440) 271 278 — (416) 237 144 (389) 257 142 538

~n1

775 708 299 176 698 298 264 174 525 679 658 568 595 495 119 476 443 473 491 295 618 320 322 192 589 327 194 572 325 197 676

Molecule

SF6 SeF6 [SeCl6]2 [SeBr6]2 TeF6 [TeCl6]2 [TeCl6]3 [TeBr6]2 [ClF6] [ClF6]þ [BrF6]þ [BrF6] [AuF6] [ScF6]3 [ScI6]3 [XF6]3 [LaF6]3 [GaF6]3 [YbF6]3 [CeCl6]2 [TiF6]2 [TiCl6]2 [TiCl6]3 [TiBr6]2 [ZrF6]2 [ZrCl6]2 [ZrBr6]2 [HfF6]2 [HfCl6]2 [HfBr6]2 [VF6]

TABLE 2.8a. (Continued )

939 780 280 221 752 250 230 198 — 890 — 400 — 458 — 160 130 140 156 268 615,660 316 304,290 244 537,522 290 223 448,490 275 189 646

~n3 614 437 160–140 129 325 136 146 — — 582 — 204,184 — 257 — 74 63 72 70 117 315,281 183 — 119 241,192 150 106 217,184 145 102 300

~n4 524 403 165 138 312 140 (135) 75 289 513 405 250 225 235 80 194 171 185 196 120 308,300 173 175 115 258,244 153 99 259,247 156 101 330

~n5 (347) (264) — — (197) — — — — — — (138) — — — — — — — (86) — — — — — — — — (80) — —

~n6a

1197–1199 1197,1198 1180 1200 1197,1198 1201 1170 1201,1202 1203 1204 1205 1206 1207 1208 898 1209 1209 1209 1209 1210 1211 1212 1213 1212 1214 1215 1216 1211 1185 1216 1217

Ref.

225

584 533 — 683 368 — — — — 692 378 — — — — (720) 286 741.8 685 356 329

305

770 437

382

341

[MoCl6]3

WF6 WCl6

[WCl6]

[WCl6]2

~n1

[VF6] [VF6]3 [VCl6]2 [NbF6] [NbCl6] [NbCl6]2 [NbBr6] [NbBr6] [NbI6] [TaF6] [TaCl6] [TaCl6]2 [TaBr6] [TaBr6]2 [Tal6] CrF6 [CrCl6]3 92 MoF6 [MoF6]2 [MoCl6] [MoCl6]2

2

Molecule

TABLE 2.8b. (Continued )

—

—

676 331

—

— — — 562 288 — — — — 581 298 — — — — (650) 237 652.0 598 — —

~n2

293

578 511 355,305 602 333 314 240 216 236 180 560 330 297 234 223 217 160 790 315 749.5 653 327 308 ) 268 286 302 711 373 332 312

~n3

166 150

157

258 160

273 292 — 244 162 165 — 112 70,66 240 158 160 — 109 80 (266) 199 265.7 250 162 168 167 187

~n4

—

168

321 182

150

— — — 280 183 — — — — 272 180 — — — — (309) 162 317 274 — 154

~n5

—

—

(127) —

—

— — — — — — — — — (192) — — — — — (110) 182 117 — — —

~n6a

(continued)

1230

1230

1197,1198 1185

1230

1217 1217 1218 1219,1220 1185 1220,1221 1222 1220,1221 1220,1223 1224 1185 1185 1222 1221 1223 1225,1226 1227 1228 1229 1230 1230

Ref.

226

[MnF6] TcF6 [ReF6]þ ReF6 [ReF6]2 [ReCl6] [ReCl6]2 [ReBr6]2 [FeF6]3 [NiF6]2 [PdF6]2 [PdCl6]2 [PdBr6]2 PtF6 [PtF6]2 [PtCl6]2 [PtBr6]2 [PtI6]2 [FeF6]3 RuF6 [RuCl6]2 OsF6 [OsCl6] [OsCl6]2 [OsBr6]2 [OsI6]2 RhF6 [RhCL6]2 IrF6 [IrCl6]2

2

Molecule

592 713 797 754 611 — 346 213 507 555 558 318 198 656 611 348 213 — 538 (675) — 731 375 345.3 210.6 152 (634) — 701.1 341.3

~n1

TABLE 2.8a. (Continued )

508 (639) 734 (671) 530 — (275) (174) — 512 276 289 176 (601) 576 318 190 — 374 (624) — (668) 302 245.2 169.2 121 (595) 320 647.3 289.4

~n2 620 748 783 715 535 318 313 217 456 648 615 346 253 705 571 342 243 186 — 735 346 720 325 326 227 170 724 335 715.6 —

~n3 335 275 353 257 249 161 172 118 — 332 — 200 130 273 281 183 146 46 — 275 188 268 168 176 122 91 283 — 267,5 —

~n4 308 (297) 359 (295) 221 — 159 104 273 307/298 212 178 100 (242) 210 171 137 — 253 (283) — (276) 183 160 100 80 (269) 253 265.2 159.6

~n5 — (145) — (147) (181) — — — — — — — — (211) (143) (88) — — — (186) — (205) — — — — (192) — 208.3c —

~n6a

1231 1198 1232 1198 1233 1234 1185,1235 1236 1179 1237 1238 1239 1240 1198 1241 1186 1239 1242 1243 1198,1244 1244 1198 1245 1246,1247 1246,1248 1249 1198,1244 1250 1251,1252 1253

Ref.

227

1342

[FeD6]4

1363

— 175.1 133 255 530 — 273 237 — 525 574 — 519 — 1878

~n2

1260

296 — 175 259 619 525 310 262 214 618 604 265 612 — 1746

~n3

a

pﬃﬃ The value of v6 can also be estimated by the relation, v6 ¼ v5 = 2 [1264,1265]. b Ground state. c Excited state. d INS frequency.

— 209.6 149 294 666 — 343 299 — 646 643 310 625 615 1873

~n1

[IrCl6] [IrBr6]2 [IrI6]3 [ThCl6]2 UF6 [UF6] [UCl6] [UCl6]2 [Ubr6] NpF6b NpF6c [NpCl6]2 PuF6b PuF6c [FeH6]4

3

Molecule

TABLE 2.8b. (Continued )

655d

200 — 87 — 184 173 122 114 87 198 191 117 200 202 899d

~n4 — 103.2 88 114 200 — 136 121 — 208 — 128 209 198 1019 1057 730

~n5

549d

— — — — — — — (80) — 169 138/149 — 177 172 836d

~n6a

1263

1242 1253 1254 1255 1256 1257 1257,1258 1255 1257 1259 1259 1260 1261 1262 1263

Ref.

228

APPLICATIONS IN INORGANIC CHEMISTRY

2.8.1.3. Halogen Series of series, for example

The effect of changing the halogen is seen in a number

[SnF6]2 v~1 ðcm1 Þ v~3 ðcm1 Þ

592 559

[SnCl]2 > >

[SnBr6]2 > >

311 303

190 224

[SnI6]2 > >

122 161

The stretching force constants also follow the same order. The ratios n(MBr)/n (MCl) and n(MI)/n(MCl) are about 0.61 and 0.42, respectively, for n1, and about 0.76 and 0.62, respectively, for n3. 2.8.1.4. Electronic Structure follows: v~3 ðcm1 Þ v~3 ðcm1 Þ

In the [MCl6]3 series, n3 and n4 change as

Cr3þ(d3)

Mn3þ (d 4)

Fe3þ(d 5)

In3þ

315 200

342 183

248 184

248 161

3 2 All these metal ions are in the high-spin state. For Fe3þ (t2g eg ), occupation of the antibonding orbitals lowers n3 drastically in comparison to the Cr3þ complex; its n3 is comparable to that of the In3þ complex, whose n3 is lowered because of the increased mass of the metal. On the other hand, the v3 of [MnCl6]3 is higher than that of [CrCl6]3 because the static Jahn–Teller effect of the Mn3þ ion causes a tetragonal distortion [1210].

2.8.1.5. Electronic Excited States Table 2.8a includes PuF6 and NpF6, for which vibrational frequencies have been determined both for the ground and excited states. The excited frequencies were obtained from the analysis of vibrational fine structure of electronic spectra. 2.8.1.6. Raman Intensities The Raman intensity of an XY6 molecule normally follows the order I(n1) > I(n2) > I(n5). Adams and Downs [1266] noted that I (n2)/I(n1) is 0.5 to 1 for [TeCl6]2 and [TeBr6]2, although it normally ranges from 0.05 to 0.1. Furthermore, they observed n3, which is not allowed in Raman spectra. From these and other items of evidence, they proposed that the Oh selection rule breaks down in [TeX6]2 because less symmetric electronic excited states perturb the Oh ground state. They also noted the distortion of [SbX6]3 ions to C3v symmetry from their Raman spectra in solution. Woodward and Creighton [1267] noted that I(n2) > I(n1) holds in the aqueous Raman spectra of Na2PtX6 (X ¼ Cl, Br) and Na2PdCl6, and attributed this unusual trend to the presence of six nonbonding d electrons in the valence shell. 2.8.1.7. Symmetry Lowering Table 2.8a contains XY6-type ions that exhibit splitting of degenerate vibrations due to lowering of symmetry in the crystalline state. For example, the n3 vibration of the ReF62 ion in K2 [ReF6] splits into two bands (565 and 543 cm1) due to trigonal distortion in the crystalline state. Similar splitting is observed for the v4 vibration (222 and 257 cm1) [1268,1269].

OCTAHEDRAL MOLECULES

229

More examples are found in the [ReCl6] (C4v) [1270] and [CuCl6]4(D2h) [1271] ions in the crystalline state. 2.8.1.8. Adduct Formation Hunt et al. [1272] obtained two types of 1:1 adducts by codepositing UF6 with HF in excess Ar at 12 K; UF6–HF exhibits a strong band at 3848 cm1, while the anti-hydrogen-bonded complex, UF6FH, shows a weak, broad band at 3903 cm1 in IR spectra. Similar experiments show, however, WF6HF exhibits a weak band at 3911 cm1, while WF6HF shows a strong band at 3884 cm1. These results indicate that, in hydrogen-bonded complexes, the order of proton affinity is WF6 < UF6, with HF serving as a proton donor and that, in antihydrogen-bonded complexes, the order of acidity is UF6 < WF6, with HF serving as a Lewis base. Relative yields of these two types are determined by the acidity/basicity of these fluorides. 2.8.1.9. Anomalous Polarization Preresonance Raman spectra of 17 XY6type metal halides have been measured by Bosworth and Clark [1192]. Hamaguchi et al. [1273] observed anomalous polarization (Sec. 1.23) for all the Raman bands of the IrCl62 ion measured under resonance condition. Figure 2.27 shows the polarized Raman spectra (488.0 nm excitation) of this ion in aqueous solution together with band assignments and depolarization ratios for each band. Since the Raman tensor for the n1, n2, and v5 vibrations cannot have an antisymmetric part, group theoretical arguments such as those used for heme proteins (Sec. 1.23) are not applicable to octahedral XY6 ions. A more detailed study by Hamaguchi [1274] shows that electronic degeneracy in the IrCl62 (5d5) ion induces an antisymmetric part of vibrational Raman tensors. Anomalous polarization was also observed for the [IrBr6]2 ion [1275]. 2.8.1.10. Jahn–Teller Effect Weinstock et al. [1276] noted that the combination bands (n1þ n3) and (n2þ n3) appear with similar frequencies, intensities, and shapes in the infrared Spectra of MoF6(d 0) and RhF6(d3). As shown in Fig. 2.28 however, (n2þ n3) was very broad and weak in TcF6(d1), ReF6(d1), and RuF6(d2), and OsF6(d2). This anomaly was attributed to a dynamic Jahn-Teller effect. The static Jahn–Teller effect does not seem to operate in these compounds since no splittings of the triply degenerate fundamentals were observed. Perhaps the most fascinating XY6-type molecule is XeF6. In their earlier work, Claassen et al. [1277] suggested the distortion of XeF6 from Oh symmetry since they observed two stretching bands in infrared and three stretching bands in Raman spectra. It was not possible, however to determine the precise structure of XeF6 until they [1278] carried out a detailed IR, Raman, and electronic spectral study of XeF6 vapor as a function of temperature. They were then able to show that XeF6 consists of the three electronic isomers shown in Fig. 2.29 and to explain subtle differences in spectra at different temperatures as a shift of equilibrium among these three isomers. 2.8.1.11. 235U / 238U Isotope Shift UF6 in natural abundance consists of 99.3% 238 UF6 and 0.7 % 235 UF6 . Since the isotope shifts between them is extremely small, it

230

APPLICATIONS IN INORGANIC CHEMISTRY

Fig. 2.27. Polarized Raman spectra of [(n-C4H9)4N]2 [IrCl6] in acetonitrile (488 nm excitation) [1273].

Fig. 2.28. Band profiles for (n1þ n3) and (n2þ n3) for the 4d transition series hexafluorides [1276].

OCTAHEDRAL MOLECULES

231

Fig. 2.29. Structures of three isomers of XeF6 and their IR-active stretching frequencies (cm1). The Xe atom at the center is not shown.

is difficult to observe the vibrational spectrum of the latter without serious overlapping of the former. Takami et al. [1279] have overcome this difficulty by measuring the spectrum of a UF6 (natural abundance)/Ar mixture as a supercooled vapor (below 100 K) using the cold-jet technique [1280]. Figure 2.30 shows the high-resolution IR spectrum of UF6 in the n3 region thus obtained [1279]. It exhibits sharp rotational fine structures [1281] without hot band contributions, and the rotational components of 238 UF6 and 235 UF6 are well separated; the former shows the Q-branch as well as R (0)–R(6) branches whereas the latter exhibits the Q-branch between R(5) and R(6) of the former. Although the isotope shift is estimated to be only 0.6040 cm1, Takeuchi et al. [1282] have found that UF6 gas in the supercooled state (

Our partners will collect data and use cookies for ad personalization and measurement. Learn how we and our ad partner Google, collect and use data. Agree & close