Upload
fredrick-mutunga
View
46
Download
0
Tags:
Embed Size (px)
DESCRIPTION
CH3I
Citation preview
5/24/2018 High Resolution Infrared
1/173
OULU 1999
HIGH RESOLUTION INFRARED
SPECTROSCOPY ON THE
FUNDAMENTAL BANDS OF13
CH3I
SEPPO
ALANKO
Department of Physical Sciences
5/24/2018 High Resolution Infrared
2/173
OULUN YLIOPISTO, OULU 1999
HIGH RESOLUTION INFRARED
SPECTROSCOPY ON THE
FUNDAMENTAL BANDS OF13
CH3I
SEPPO ALANKO
Academic Dissertation to be presented with the assent
of the Faculty of Science, University of Oulu, for public
discussion in Raahensali (Auditorium L 10), on April 9th,
1999, at 12 noon.
5/24/2018 High Resolution Infrared
3/173
Copyright 1999
Oulu University Library, 1999
OULU UNIVERSITY LIBRARY
OULU 1999
ALSO AVAILABLE IN PRINTED FORMAT
Manuscript received 26.2.1999Accepted 18.3.1999
Communicated by
Professor Hans Brger
Doctor Esko Kyr
ISBN 951-42-5185-7
ISBN 951-42-5184-9
ISSN 0355-3191 (URL: http://herkules.oulu.fi/issn03553191/)
5/24/2018 High Resolution Infrared
4/173
Abstract
This thesis deals with the rotation-vibration theory and high resolution infrared spectroscopy of
semirigid C3nmolecules. Semirigid molecules form a class of molecules which are strongly bound
with one well defined structure, and without low frequency internal motions. The theory, as well as
the experimental studies of semirigid molecules are of special importance in the field of rotation-
vibration spectroscopy. They provide a good starting point for interpreting and analyzing the spectra
of practically all types of molecules.
In this work, the theory is reviewed fromthe standpoint of one particular molecule, 13CH3I, which
is a prolate symmetric top with C3nsymmetry. The origin and the properties of the rotation-vibration
Hamiltonian are discussed in detail. Molecular symmetry plays an important role in these studies.
The expansion of the Hamiltonian for nuclear motion in powers of the vibrational operatorsconverges rapidly as numerical examples thoughout the treatment indicate. The molecule is thus a
good subject for the perturbation calculations, also reviewed here in detail.13CH3I can be considered
as a model example of semirigid molecules.
From the spectroscopic point of view, this thesis is a study of the six fundamental bands of 13CH3I.
The rotational analysis of the vibrational ground state is first given. Special attention is paid to
obtaining the axial rotational constants which are problematic for symmetric top molecules. The
relatively high energy level density of13CH3I leads to several resonances. The fundamental bands,
especially the higher ones, must therefore be treated as parts of band systems. Care is paid to properly
take into account the effects of th e near-lying vibrational levels on the constants of the fundamentals.
Certain ambiguities in the rotation-vibration Hamiltonian of 13CH3I are also discussed.
Keywords:molecular constants, rotation-vibration theory, semirigid molecules, symmetric
top.
5/24/2018 High Resolution Infrared
5/173
To My Mother and Father
5/24/2018 High Resolution Infrared
6/173
5/24/2018 High Resolution Infrared
7/173
5/24/2018 High Resolution Infrared
8/173
Acknowledgments
This work has been carried out at the department of Physical Sciences of the University ofOulu. I would like to thank professor Rauno Anttila, the head of the department during the
course of this work, for placing the facilities at my disposal. I also would like to express
my sincere gratitude to him, as my supervisor, for suggesting me the topic, giving me free
hands in performing it, and for encouraging me during the work.
I am grateful to all the other members of the IR research group in Oulu for their co-operation and valuable discussions over the course of this work. I wish to express mygratitude to Dr. Veli-Matti Horneman, Dr. Matti Koivusaari, Dr. Risto Paso, and Dr. KariPekkala for their guidance and support. I also wish to thank Mr. Jyrki Schroderus, Mr.Tarmo Ahonen, Mr. Jari Pietil, Mr. Hannu Sarkkinen, Mr. Jarmo Lohilahti, and Mr. Jarmo
Lehtomaa with whom it has been a pleasure to work. Special thanks go to Ms. Anne-MaariaAhonen who has accompanied me exploring theeld of IR spectroscopy. I am indebted toall these people, as well as to the rest of the staff of the Department for their friendship andtheir contributions in creating a pleasant and inspiring working atmosphere.
I would like to thank Mr. Gordon Roberts for revising the language of the manuscript.Many thanks are also due to Ms. Anne-Marie Elo who gave several valuable commentsregarding the manuscript.
This thesis is dedicated to my parents who have supported and encouraged me in manyways over the years. I also thank all the other members of my family for showing interest inthis work. Special thanks go to my niece Ms. Annamari Alanko for checking some angularmomentum commutator calculations presented here.
Oulu, February 1999 Seppo Alanko
5/24/2018 High Resolution Infrared
9/173
5/24/2018 High Resolution Infrared
10/173
List of original papers
The present thesis consists of a theoretical part and the analysis of the fundamental bandsof CH
I. The analysis is based on the following six papers, which are referred to in the
text by their Roman numerals:
I Alanko S (1991) The
Band of CH
I. J Mol Spectrosc 147: 406-413.
II Anttila R, Alanko S, Horneman V-M & Koivusaari M (1990) Ground state constants
and
of CH
I from perturbation-allowed transitions. Mol Phys 71: 1433-1436.
III Pietil J, Koivusaari M, Alanko S & Anttila R (1996) Ground state constants
,
and
of CH
I from normally allowed infrared absorption bands. Mol Phys 87:523-527.
IV Alanko S & Karhu J (1995) The Lowest Fundamental Band
of CH
I : Simulta-neous Analysis of the Levels
,
, and
. J Mol Spectrosc 174:215-222.
V Alanko S (1996) A Detailed Analysis of the
Band System of CH
I
and
CH I. J Mol Spectrosc 177: 263-279.
VI Alanko S (1998) Infrared Spectrum of CH
I in the cm Region: Ro-tational Analysis of the Fundamentals
and
Together with
and
.J Mol Spectrosc 188: 43-62.
5/24/2018 High Resolution Infrared
11/173
Contents
Abstract
Acknowledgments
List of original papers
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2. Symmetry considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.1. General symmetry properties of the molecular Hamiltonian . . . . . . . . . . . . . . . . 16
2.2. Molecular symmetry group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3. Equivalent rotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4. Point group. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5. Group theory and representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3. Hamiltonian for a vibrating and rotating molecule . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.1. Born-Oppenheimer approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293.2. Vibrational and rotational coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.1. Euler angles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323.2.2. Normal coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3. Classical Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.4. Harmonic forceeld calculations for CH
I . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423.4.1. Vibrational problemdirect approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.4.2. Vibrational problem using internal coordinates . . . . . . . . . . . . . . . . . . . . . 45
3.4.3. Normal coordinates for
CH I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493.4.4. Vibration-rotation interaction parameters . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.5. Quantum mechanical Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4. Expansion of the Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 594.1. Rigid rotor and harmonic oscillator approximation. . . . . . . . . . . . . . . . . . . . . . . . 60
4.1.1. Rigid rotor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 624.1.2. One-dimensional harmonic oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684.1.3. Two-dimensional harmonic oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714.1.4. Combined rigid rotor and harmonic oscillator . . . . . . . . . . . . . . . . . . . . . . 75
4.2. Expansion of the Hamiltonian and orders of magnitude . . . . . . . . . . . . . . . . . . . . 78
4.3. Expanded Hamiltonian for molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5/24/2018 High Resolution Infrared
12/173
5. Transformations on the Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.1. Vibrational contact transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.2. Transformed Hamiltonian for
molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
5.2.1. Vibrational Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 995.2.2. Rotational Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.2.3. Vibration-rotation interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1105.3. Resonances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.3.1. A method of calculating non-diagonal matrix elements . . . . . . . . . . . . . . 1165.3.2. Essential resonances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1205.3.3. Accidental resonances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.4. Nuclear quadrupole hyperne structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6. Intensities of infrared transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1296.1. Selection rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1316.2. Intensities in harmonic oscillator and rigid rotor approximation . . . . . . . . . . . . 1326.3. Spin statistical weights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1366.4. Perturbation calculations on the dipole moment operator . . . . . . . . . . . . . . . . . . 139
7. Rotational analysis of the fundamental bands of CH
I . . . . . . . . . . . . . . . . . . . . . . 1467.1. Experimental methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1477.2. Computational principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1507.3. Analysis of the ground state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.3.1. - dependent constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1537.3.2. - dependent constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
7.4. Band system
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1557.5. Band system
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1577.6. Band system
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1587.7. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1627.8. Future prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
9. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
5/24/2018 High Resolution Infrared
13/173
1. Introduction
This thesis deals with molecular spectroscopy, the art of measuring and interpreting re -sponses of molecules to stimuli induced by electromagnetic radiation. Molecules, consist-ing of electrically charged nuclei and electrons, may interact with the oscillating electricand magneticelds of light absorbing its energy. Not all the energy of the radiationeldis absorbed, however. Each molecule, obeying the laws of quantum mechanics, can absorbonly distinct, characteristic wavelengths corresponding to the differences of its discrete en-ergy levels. The experimental data obtained are the particular wavelengths and the amountsof radiation at each wavelength being absorbed by the molecule. In other words, an absorp-tion spectrum is measured.
The different categories of molecular spectroscopy are classied according to the type
of molecular energy being probed. The photons in the X-Ray-Ultraviolet-Visible regionof the electromagnetic spectrum (100 eV - 1 eV) can cause core and valence electron re-arrangements. They are therefore the tools for electronic spectroscopy. The infrared quanta(10 000 cm - 10 cm ) are capable of exciting vibrational motions and they are used ininfrared spectroscopy (IR). Changes in rotational energy can be stimulated by microwavefrequencies (1000 GHz-1 GHz), and this branch of spectroscopy is accordingly namedmicrowave spectroscopy (MW). The regions of the electromagnetic spectrum mentionedhere are only suggestive, possessing no sharp boundaries. They are also given in unitsthat are customary for each branch of spectroscopy. The energyin electron volts (eV),the wavenumberin waves per centimeter (cm ), and the frequencyin Hertz (Hz) arerelated (in vacuum) by
(1)
where is Plancks constant, eVs, and is the velocity of light, cm/s.
The transitions studied by electronic, infrared, and microwave spectroscopy are all in-duced by the oscillating electriceld of the radiation. The magneticeld component, onthe other hand, is utilized in the spectroscopies of the spin, which are nuclear magneticresonance (NMR) and electron spin resonance (ESR) spectroscopy. In these techniques,the molecule is placed in an external magneticeld. Those molecular particles having a
nonzero spin, and thus a magnetic moment, tend then to orientate themselves in certaindirections with respect to the direction of the externaleld. The different orientations cor-
5/24/2018 High Resolution Infrared
14/173
14
respond to different energies, and transitions between them can be stimulated by radiation
of suitable frequency. In NMR spectroscopy radio frequency photons can ip the nuclear
spins in magnetic elds of a few tesla. In ESR spectroscopy, microwave frequencies are
used.The techniques to study molecules are numerous as too are the types of molecules. Some
are small, some large, some are tightly bound and some may have loose structures allowing
various types of internal motions. Some molecules have a well dened geometrical struc-ture whereas some can pass from one conformation to another. The present thesis appliesinfrared spectroscopy to one limited type, the semirigid molecule. A semirigid moleculeis strongly bound, with a well dened geometrical structure. It has no low potential bar-riers, which would allow internal motions with large amplitudes. The phenomena understudy here are related to the vibrational motion of the molecule. For a semirigid molecule,the expansion of the Hamiltonian for nuclear motion in powers of the vibrational opera-tors converges rapidly. This forms a good basis for perturbation treatments. We deal with
the vibration-rotation spectroscopy of a relatively small molecule
CH I, which can beconsidered as a model example of semirigid molecules.
The present thesis can be divided logically into two parts. After this brief introductorychapter, the next ve chapters (2-6) are devoted to the general theory of vibrational androtational motions of semirigid molecules. In these chapters, the CH
I molecule is usedin illustrations and examples. Chapter 7 then, together with six separate papers, gives thedetailed analysis of the rotational structures of the vibrational ground state, as well as thesix fundamental vibrational levels of the CH
I molecule.In chapter 2, we begin with a discussion of molecular symmetry and the tool which
makes use of it, namely the group theory. Symmetry, both geometrical and that of the
Hamiltonian operator, plays an important role in molecular studies. The degeneracy of theenergy levels, for example, is just one aspect of the symmetry of the system. The moresymmetrical the molecule is, the more degenerate its energy levels are. On the other hand,the advantages of making use of the symmetry are numerous. The existence of variousterms in the Hamiltonian, as well as certain involved integrals, may be determined purelyby group theoretical arguments. The quantum mechanics of a molecular system may thusbe simplied, sometimes considerably. The symmetry studies also enable us to classify thewavefunctions in a useful way.
In chapter 3, the quantum mechanical Hamiltonian operator for a semirigid vibratingand rotating molecule is formed. We start with a very important simplication, the Born-Oppenheimer approximation. As a result, the nuclear motion is separated from the elec-tronic motion. Henceforth, a vibrating and rotating molecule can be treated as a collectionof atomic point masses connected by forces keeping the individual atoms at or near theirequilibrium positions. In the frame of this model, we then introduce new coordinates, theEuler angles and the normal coordinates, which facilitate the separation of the molecularmotions into translational, vibrational, and rotational parts. The translational motion is thenignored, and the classical Hamiltonian is formed. The molecular vibrations can be treatedclassically and therefore the harmonic forceeld calculations for CH
I are presented alsoat this point. Finally, the classical Hamiltonian is transformed to the quantum mechanicalHamiltonian operator.
Chapter 4 is devoted to the expansion of the Hamiltonian. To explicitly solve the
Schrdinger equation, approximate methods involving numerical approaches must be uti-
5/24/2018 High Resolution Infrared
15/173
15
lized. For semirigid molecules the perturbation method has proven to be efcient. The
Hamiltonian is rst expanded in terms of normal coordinates, their conjugate momenta,
and the rotational momenta. Then, to a good approximation, the vibrational-rotational en-
ergies and wavefunctions are obtained using the harmonic oscillator and rigid rotor termsfrom the expansion. The higher order terms are treated as small perturbations and the cor-rections of energies and wavefunctions are obtained using the perturbation theory. Theimportant solutions of the harmonic oscillator and the rigid rotor Schrdinger equations arerst considered in detail. The general expansion of the Hamiltonian is then given togetherwith some numerical examples to illustrate its rapid convergence. Finally, the terms thatcan appear in the expansion of
molecules are discussed.The next chapter, Chapter 5, then presents the perturbation calculations. The basic prin-
ciples of vibrational contact transformation, the form of perturbation theory used, arerstreviewed. Then, the most important terms appearing in the transformed Hamiltonian for
molecules are discussed. The unperturbed vibrational and rotational Hamiltonians are
derived and the effects of the vibration-rotation interaction are considered. Distinct formu-lae to evaluate the unperturbed vibration-rotation energy of a C
molecule are given. Nextthe resonances, the anomalies in unperturbed Hamiltonians caused by failures in the per-turbation theory, are discussed. The resonances appear as coupling of otherwise separatevibrational levels. In this chapter, a general symmetry-based method is given to calculatethe form of any coupling matrix element. The main purpose of Chapter 5 is to provide thetools to form the rotation-vibration Hamiltonian, and thus the energies of any given blockof interacting vibrational levels of C
molecules. In addition to this, the effect of anotheroperator giving a measurable contribution, the nuclear quadrupole hyperne coupling op-erator, is briey reviewed.
Chapter 6 deals with transitions between vibration-rotation energy levels. Rigoroussymmetry based selection rules for electric dipole transitions arerst obtained. Then the in-tensities in harmonic oscillator and rigid rotor approximation for symmetric top moleculesare discussed. Spin statistical weights are also considered. In the study of intensities, weshall not involve ourselves in the art of extracting intensity information from the experi-mental lines. Our purpose is simply to gain understanding about why the experimental IRspectra appear as they do. For that purpose, we use more or less qualitative simulations.In the spectra, there are usually a number of transitions that cannot be understood by theharmonic oscillator and rigid rotor approximation. To study the origin of these so-calledperturbation allowed or forbidden transitions, the perturbation calculations on the electricdipole moment operator are also discussed.
In Chapter 7, the infrared analysis of CH
I is thennally carried out. Only the mainfeatures of the analysis are discussed and the most important results are summarized. Thedetailed step by step analyses are presented in the individual Papers, I-VI.
5/24/2018 High Resolution Infrared
16/173
2. Symmetry considerations
Most molecules possess some symmetry. A molecule may, for example, have such a struc-ture that interchanging the positions of some of its atoms results in the molecule lookingexactly the same as before. Procedures of this type are called symmetry operations. Acollection of symmetry operations forms a group in mathematical sense and we can usethe formal group theory to systematically utilize the symmetry in molecules. What is sur-prising is that so much detailed, fundamental, and important understanding can result fromsuch a simple concept as symmetry. The quantum mechanics is simplied considerablyand spectroscopic information, like the nature of normal vibrations, the allowed terms inHamiltonian, the nuclear spin statistical weights, the selection rules for transitions, the al-lowed rotation-vibration interactions, for example, can be obtained.
In the next sections, we begin by considering the general molecular symmetry prop-erties, of which the above mentioned geometrical symmetry is but one example. Thesymmetry that we shall especially exploit is the symmetry of the Hamiltonian operatorsince, in general, the complete molecular symmetry group consists of operations leavingthe molecular Hamiltonian unchanged. Secondly, we shall study the most important sym-metry properties of the CH
I molecule. The concepts of molecular symmetry group (MSgroup) and point group will be introduced. Finally, we shall briey overview the abstractmathematical group theoretical machinery considering only the most important conceptsfrom the standpoint of a spectroscopist.
2.1. General symmetry properties of the molecular Hamiltonian
In theeld of molecular spectroscopy, a molecule is regarded as a collection of electronsand nuclei (nuclear structure neglected) held together by electromagnetic forces and obey-ing the laws of quantum mechanics. The allowed stationary molecular energiesand thewavefunctions are provided by the time-independent Schrdinger equation ,where is the molecular Hamiltonian operator. For an isolated molecule in the absenceof external electric or magneticelds, the nonrelativistic Hamiltonian in a space-xed co-ordinate system with its origin at the center of mass can be written in the form, see e.g.
5/24/2018 High Resolution Infrared
17/173
17
[13]
(2)
where the Greek subscripts () refer to the nuclei of the molecule, while the Romansubscripts () refer to its electrons. Therst two terms include the massesand themomentum operators of the particles, thus representing kinetic energies of thenuclei and electrons, respectively. The next three potential energy terms are the nuclear-nuclear repulsion energy, the electron-nuclear attraction energy, and the electron-electronrepulsion energy, respectively. These terms involve the interparticle distances and theelectrostatic charges of the particles. The term
arises from the interaction of eachof the electron spin magnetic moments with the other magnetic moments in the molecule.The term
is the nuclear hyperne structure term resulting from the interactions of theelectric and magnetic moments of the nuclei with the other electric and magnetic momentsin the molecule. These two terms include the-factors and the spin angular momentumvectors
and
for electrons and nuclei, respectively. The terms are of the type [2]
and
, for example, where thevectors are the position vectors of the particles.
It is now relatively easy to verify that the Hamiltonian is invariant in any of the followingoperations (see e.g. [2,48])
(a) any overall translation of the molecule in space(b) any rotation of the molecule about any space-xed axis through the center of mass(c) any permutation of the space and spin coordinates of the electrons(d) any permutation of the space and spin coordinates of identical nuclei(e) inversion of the coordinates of all the particles (nuclei and electrons) in the center of
mass of the molecule(f) time reversal(g) Hermitian conjugation
The invariance to translation (a) and rotation (b) results from the fact that free space is
uniform and isotropic, meaning that the molecule experiences the space in the same wayeverywhere and in every orientation. A permutation of electrons (c) merely involves in-terchanging the electron subscripts () in. The invariance to this operation followsimmediately from the fundamental fact that all the electrons are identical and indistin-guishable. The same applies to permutations of identical nuclei (d), too. The effect of theinversion operation (e) is to change all position vectors to and momentum vectorsto . The spin angular momentum vectorsand transform as any angular momentumvector, and behave therefore as . Sinceand are reversed, the vector product isinvariant and it is readily seen that is invariant to inversion. The time reversal operation(f) consists of changing the sign of the time variableaccompanied with complex conju-
gation. This operation reverses all momenta,as well as angular momenta and , butnot the coordinates. The Hamiltonian is not altered as , ,
5/24/2018 High Resolution Infrared
18/173
18
and . Finally, the Hamiltonian is invariant to Hermitian conjugation (g) because itseigenvalues are real.
2.2. Molecular symmetry group
The full symmetry group of a molecule consists of all the operations that leave the complete
molecular Hamiltonian unchanged, i.e. (a)-(f) in the previous section. There is an innitenumber of such operations, but in practical molecular spectroscopy not all the elementsof the complete group need be considered. The most important subgroup is the molecularsymmetry group, or MS group, introduced by Longuet-Higgins in 1963 [4,9,10]. It consistsof the feasible operations of the complete permutation - inversion group (PI group) of themolecule. The PI group includes the following operations:
(i) All permutationsof the positions and spins of identical atomic nuclei(ii) All permutation-inversions , where is the inversion of the
positions of electrons and atomic nuclei in the center of mass.
For most molecules, not all of these operations are feasible because there may be very highpotential barriers opposing some of the transformations. Attention can be restricted to thefeasible transformations, to those achievable without passing over insuperable barriers.
For example, the potential energy surface of the methyl iodide molecule in the groundelectronic state has two deep minima centered at the clockwise (C) and anticlockwise (A)forms of the molecule (see Fig. 1).
3
5
4
I
C
H13
I
C
5
4
(C) (A)
H2
H3H1
H2
H3
Fig. 1. The two symmetrically equivalent nuclear equilibirium structures of a methyl iodide
molecule. The form (C) is called the clockwise form because the protons are numbered clock-wise when viewed from the positive end of the C-I axis, and (A) is called the anticlockwise form,respectively.
The complete PI group of CH
I has, as illustrated in Fig. 2, the following 12 elements:
(3)
Here, is the identity operation, denotes the cyclic transposition of the hydrogennuclei in Fig. 2, , and for example, is the permutation of the hydrogennuclei 1 and 2. The tunnelling through the potential barrier from conguration (C) to (A)
would appear as splitting of energy levels in experimental spectra. A classic example ofsuch inversional splitting is the spectrum of ammonia, NH
[11]. No experimental tech-
5/24/2018 High Resolution Infrared
19/173
19
313
2
5
4
E
321
3
5
4
(123)
332
1
5
4
(132)
323
1
5
4
(12)
331
2
5
4
(13)
31
2
3
5
4
(23)
3
3
2
5
4
E*1 3
1
3
5
4
2(123)*
3
2
1
5
4
3(132)*
3
3
1
5
4
2(12)*
3
1
2
5
4
3(13)*
3
2
3
5
4
1(23)*
Fig. 2. The effect of the complete PI group operations on the equilibrium conguration ofmethyl iodide, CH
I.
nique, however, has so far been able to resolve such splittings in CH
I, indicating that the
molecule does not pass over from (C) to (A) on the time scale of the experiments being per-formed. This means that all the symmetry properties of CH
I can be derived by studyingthe molecule in one minimum, say (C), only. Half of the symmetry operations in group (3)become now useless i.e. unfeasible. Careful inspection of Fig. 2 reveals that the set of op-erations that does not interconvert the clockwise and anticlockwise forms is the subgroupconsisting of the following six operations
(4)This can now be considered as the molecular symmetry group (MS group) for methyl io-dide in its electronic ground state. By geometrical symmetry, the methyl iodide moleculebelongs to the
point group (see point groups later) and the name of the MS group istaken to be that of the point group, followed by , e.g. the MS group of CH
I is called
.Since the rotational-vibrational-electronic (rovibronic) Hamiltonian is invariant to any
operation of the MS group of the molecule, the rovibronic wave functions of the moleculecan be classied in the MS group.
2.3. Equivalent rotations
Each operation in the MS group of a molecule may be associated with an equivalent rota-tion. Following Bunker [2], werst dene two rotation operators,
and
. Operation
rotates the molecule-xed axis systemby an angleabout theaxis. The an-gleis measured in a right handed sense, i.e. anticlockwise about theaxis when viewedfrom its positive end. The operator
rotates the axes by angle about an axis which is
in the plane making an angle with theaxis. Alsois measured in a right handedsense about theaxis.
5/24/2018 High Resolution Infrared
20/173
20
H1
H2
H3
x
y
(+z)
(23)*
H3
H2
H1x
y
(-z)
(a) (b)
Fig. 3. (a) Theaxis denition for a methyl iodide molecule. The origin is at the centerof mass and the axis points up out of the page. The three protons are below the plane of thepage. In (b) the effect of the operation on the molecule is shown. The axis points nowinside the paper and the protons are above the plane of the page. The operation has resulted inthe axes being rotated through radians about the axis.
In Fig. 3a a methyl iodide molecule is shown in its equilibrium conguration with the
molecule-xed () axes attached. The axes coincide with the three principal axes ofinertia and the following conventions are adopted. The origin lies at the center of mass andthe iodine and carbon atoms on the axis so that iodine lies on the positive side. Atom H
isin the plane with a negative coordinate and the axis is located so that the axis systemis right handed. Fig. 3b shows now the effect of the operation on the molecule. Theconvention of attaching the axes on the molecule is not changed by this operation and theaxes have thus rotated with the molecule. It can be seen that the same result obtained by operation is achieved by
. Similar examination of the other MS group operationsshows that the rotations equivalent to the operations (4) are, respectively
(5)
where
denotes no rotation, i.e. the identity operation of the group.Equivalent rotations are especially suitable in determining how the Euler angles trans-
form under the MS group operations. Euler angles give the orientation of the molecule-xed rotating () axis system relative to the space-xed ( ) axis system (seepage 32). Knowing the transformation properties of the Euler angles (see Table 5 on page35), the properties of rovibronic angular momenta are easily deduced and then the proper-ties of the rotational wavefunctions are obtained (see Eqs.(209) and (210) on page 67).
2.4. Point group
The point group makes use of the geometrical symmetry of the molecule, while the MSgroup utilizes the invariance of the Hamiltonian. The point group is dened as the set ofsymmetry operations that transform the rigid nuclear reference conguration on to itself.In the point group operations, one point in a molecule, the center of mass, always remainsunchanged, and also all of the symmetry elements pass through this point. For the CH
I
molecule, the equilibrium conguration in the electronic ground state forms a structurewith
point group symmetry. The symmetry elements are now
5/24/2018 High Resolution Infrared
21/173
21
H1 x
y
(+z)
(23)*
H 1x
y
(+z)
1vE
R/2
(a) (c)(b)
x
y
(-z)
+
+ -
H1 H1
H2 H2 H2
H3 H3 H3
Fig. 4. The effect of the operations
, , and
on an arbitrarily vibrating and rotatingmethyl iodide molecule. Only H atoms are considered. Black dots represent the instantaneouspositions of H atoms in space, and symbols and represent the position of an electron above
and below the plane, respectively.
(1) 3-fold rotation axis (C-I bond)
. There are two symmetry operations that make useof this elementthe rotation through 2/3 radians designated by
and the rotationthrough 4/3 radians designated by
.(2) three symmetry planes (H-C-I planes)
,
, and
. The symmetry operationsare reections in the planes and are designated by
,
, and
. Atom H
lies inthe
plane.
The point group associated with the methyl iodide molecule has thus six operations:
v v
v (6)
The point group symmetry is established by studying the molecule in its equilibriumconguration. The application of the point group to molecular wavefunctions is howeverdifferent. By denition [12], the elements of the point group are interpreted as rotating andreecting the electronic coordinates and the vibrational displacements of the nuclei fromtheir equilibrium positions, the moleculexed axes remaining unchanged.
The operations of the point group do not appear in the list of symmetry operations onpage 17 because these operations do not commute with the complete Hamiltonian. It mustbe emphasized that the point group is not a symmetry group of the molecular Hamilto-
nian. The point group is however a symmetry group of the vibrational-electronic (vibronic)Hamiltonian since all interparticle distances are maintained in its operations. The Eulerangles are also unchanged by these operations.
To understand how the elements of the point group are related to the elements of themolecular symmetry group, in Fig. 4 the effects of
and on a vibrating and ro-tating methyl iodide molecule are compared [13]. For simplicity, only the H atoms areconsidered. The three black dots represent instantaneous positions of the H atoms in space,and the open circles denote the equilibrium conguration. The small arrows representatomic motions relative to the molecule-xedaxes. The molecule-xed axes fora deformed molecule are oriented at all times so that the Eckart conditions (see page 41)
are instantaneously satised. The orientation of the axes smoothly coincides with the prin-cipal axes of the equilibrium as the distortion is removed. The MS group operation
5/24/2018 High Resolution Infrared
22/173
22
converts the molecular conguration from the one shown in Fig. 4a to that shown in Fig.
4c. In Fig. 4b is shown the result of the point group operation
which rotates the elec-tronic and vibrational coordinates but does not affect on the rotational coordinates. It is
seen that a rotation operation
introduced in the previous section converts theconguration of Fig. 4b to that of Fig. 4c, and we can therefore write
or
. The corresponding equations can be set up to relate in a one-to-one fashion the six MS group elements and the point group elements for the methyl iodidemolecule. The relations are the following:
(7)
The rotation operationsdo not affect the vibrational and electronic variables, and there-fore the point group element and its partner in the MS group have the same effect on thevibronic variables. Thus, either group can be used to classify the vibronic wave functionsand the same results will be obtained.
In Fig. 4a the overall rotation of the molecule is given by the dotted arrow and the vi-bronic angular momentum is represented by having H
move in a circle. The point groupoperations do not affect the overall rotational angular momentum, but change the vibronicangular momentum. If we neglect those terms in the Hamiltonian that couple the vibroniccoordinates to the rotational coordinates (Coriolis coupling and centrifugal distortion), weobtain an approximate Hamiltonian that commutes with the elements of the molecular pointgroup. The neglected terms are small, and in principle the point group could be used forlabelling rovibronic states. However, the point group is generally used only for labellingvibrational and electronic states.
2.5. Group theory and representations
The mathematical formalism used to exploit symmetries in molecules is the group theory,which had its form as a modern abstract mathematical theory by the end of the nineteenthcentury [14]. It entered physics in crystallography and has since been used extensively in
theoretical physics and chemistry [7, 12,1518] for example.An abstract mathematical group is a set of elements with a multipli-
cation operation obeying the following four rules: (i) The set is closed, meaning that theproduct of any two group elements must also be a group element. (ii) The elements obeythe law of association, e.g., for every . As long as theelements are not interchanged the order of multiplication is immaterial. (iii) There is a unitelementsuch thatfor every . (iv) Every element has aninverse element such that . We shall restrict our attentiontonite groups containing anite numberof group elements.is said to be the order ofthe group.
The multiplication of two symmetry operations is dened as their sequential perfor-mance. For example, the product means operation followed by.
5/24/2018 High Resolution Infrared
23/173
23
Table 1. The multiplication table of the MS group
.
The operations on top are performed rst.
The result of carrying out these two operations is the same as performing the single opera-
tion
, as can be veried from Fig. 2. A multiplication table, Table 1, for cannow be constructed, and it is readily seen that elements fulll the requirements for a groupin a mathematical sense.
The group elements can be partitioned into subsets in many different ways. For example,from Table 1 we observe that the elements form a group by themselves.Other subgroups in
are , , and . The most impor-tant partitioning of the elements is however obtained through the concept of classes. Twomembers of a group, and , are said to belong to the same class if another group membercan be found such that
(8)and are then said to be conjugate to each other. By use of (8) the
group canbe partitioned into three classes
and (9)
The importance of classes is in following: when the group elements are represented bymatrices the traces of all elements in a class are equal. This is because the operation ofconjugation (8) becomes now a similarity transformation, leaving the trace invariant.
Two groups are said to be isomorphic if their multiplication tables are identical. This
means that there is a one-to-one correspondence between the elements of the two groups.It is readily clear that our three groups introduced so far, the
group, the group ofequivalent rotations, and the point group
are all isomorphic with respect to each other.The concept of isomorphism is important because it can be used to construct physical
representations for the abstract groups. A true representation of a group is in general anygroup being composed of concrete physical entities and being isomorphic to the originalgroup. Among the representations, the matrix representations are of special importancefor physical applications. The group elements are square matrices of linear operators con-structed with respect to a given base vectors. The group multiplication is matrix multi-plication, the unit element is the unit matrix, and the inverse element of each matrix is
its inverse matrix. The dimension of the representation is dened as the dimension of therepresentation matrices.
5/24/2018 High Resolution Infrared
24/173
24
Table 2. The character table of the
group
Equiv.rot.
1 1 1
1 1 -1
2 -1 0
Depending on the basis chosen, the dimension of the representation can be very large.However, a matrix representation is in general reducible. This means that it is often possibletond a similarity transformation that reduces the matrices representing the elements of thegroup into block diagonal form with all the matrices having the same block structure. Ifthis reduction cannot be done, the representation is said to be an irreducible representation,meaning that it cannot be expressed in terms of representations with lower dimensions.The structure of reducible representation is conveniently indicated by giving the irreduciblerepresentations which form the blocks after reduction to block form. We can write
red
(10)
wherered stands for the reducible representation and
gives the number of irreduciblerepresentation occurring inred. The symbol is now a direct sum of the representa-tions rather than ordinary addition. It can be shown [7] that for anite group, the numberof inequivalent irreducible representations is the same as the number of classes.
An important property of the matrices representing a group is their character. The char-acterof a matrix is its trace, the sum of its diagonal elements. The character of a matrixis independent of the choice of the basis and remains unchanged in similarity transforma-tions. It is thus a specic property for a matrix belonging to the representation, and can beused to identify an irreducible representation. The characters of the irreducible represen-tations for a given symmetry group can be obtained withoutnding the actual irreducible
matrices (see e.g. any of the references given at the beginning of this section). In Table2, the characters of the three possible different irreducible representations for the
group are listed. The table is square since the number of classes equals the number of ir-reducible representations. On the top title row, one element from each class is given andthe number of elements in the class is indicated underneath the element. Further, becauseof the isomorphism, the character table for the
point group is the same as that of theMS group and also the appropriate element of each class of the point group are given. Theequivalent rotations corresponding the MS group elements are given, too. The rows of thetable are labelled by the irreducible representations using the Mulliken notation [19]: labeldenotes one-dimensional representation that is symmetric (
with respect
to rotation about the symmetry axis of the molecule. The numeric subscripts are added todistinguish the two such irreducible representations. Labelis for two-dimensional irre-
5/24/2018 High Resolution Infrared
25/173
25
Table 3. The direct products of the irreducible representations in
group.
ducible representation. In character tables, also the symmetry properties of the translations
and rotations
are usually given. These are discussed in detail in the context ofnormal coordinates, starting on page 34.
In practical calculations, the matrix representation is rst generated by considering the
effect of symmetry operations on the chosen basis. For example, in deriving the symmetry
of the vibrational modes of the CH
I molecule, the 15 Cartesian displacement coordi-nates can be used as the basis (see page 36). A matrix representation with threematrices (one for each class) is generated by studying how the displacement coordinatesare transformed under the operations,
, and
. The characters, red,red
,andred
in our example now form the reducible representationred in Eq. (10). Ourobjective is to determine how many times each irreducible representation (
) ap-pears in this representation, i.e. tond out the integral coefcients
in Eq. (10). Inmany cases, this can be done by direct inspection using the character table, but from grouptheory there is also a systematic method, a formula, to calculate how many times a given
irreducible representation appears in any reducible representation. This is given by
red
(11)
whereruns over the classes,is the order of the group,
is the number of members ineach class,red
is the character of the reducible representation in class, and
is the character of the irreducible representation taken from the character table for the class. For
, and
,
, and
An important concept related to the representations is the direct product of two repre-sentations. Symbolically, we write
(12)
where the representation is the direct product of the representations and Ana-logically to the direct sum, where the characters of the resulting representation are obtainedby summing the characters of the component representations, the characters of the directproduct representation are obtained by multiplication, i.e.
(13)
5/24/2018 High Resolution Infrared
26/173
26
In general, the representation will not be irreducible, even if the component representa-tions are. However, it can be decomposed using (11). The direct products of the irreduciblerepresentations of
group are given in Table 3.
The multiplication table can now be used to determine the symmetries of various prod-ucts of functions whose symmetries are known. For example, if a non-degenerate func-tion
belongs to a one-dimensional irreducible representation, say
and a pair offunctions
to a two-dimensional irreducible representationthe two functionsformed by multiplication are
This pair now belongs to since
When we multiply two pairs of degenerate functions, the situation is more com-plex. The four functions obtained when
and say
are multiplied are
and
From the multiplication table, we see that these fourfunctions together form the representation
but the question re-mains as to what are the combinations forming the irreducible representations. It can beshown [2, 8], by using projection operators, that the combinations of the product functions
transforming irreducibly are
(14)
where the functions generating representation
are called symmetric product func-tions, and the functions generating
antisymmetric product functions. We write []symm
and { }antisymm
for the symmetric and antisymmetric prod-ucts, respectively. In general, the characters in the symmetric and antisymmetric productsare given by [8,12]
(15)
(16)
The two pairs of functions may be the same, that is
and
. In thiscase
and there are only three independent functions in the product. From(14) we see that the antisymmetric product vanishes and the three independent functionstransform as
. Symbolically we say that in this case, the product of the degeneraterepresentationwith itself is the second power ofand we write
symm (17)In applications, see Eqs.(478) on page 105 for example, the higher order products are alsoimportant. In general, the characters in theth power are obtained from the characters inthe th power by using [2,12]
(18)
and for example
(19)
(20)
5/24/2018 High Resolution Infrared
27/173
3. Hamiltonian for a vibrating and rotating molecule
A general procedure (see e.g. [20]) for obtaining the quantum mechanical Hamiltonianoperator of a system of particles starts with setting up the classical energy in terms of
Cartesian generalized coordinates
and velocities
as
(21)
where the coefcients
in the kinetic energymay be functions of the coordinates
. The potential energyis a function of coordinates, but here we assume that it is notdependent on velocities or time. The next step is to express (21) in terms of momenta
conjugate to coordinates
, i.e.
. The energy is then
obtained in the Hamiltonian form
(22)
where the matrix of is the inverse of the matrix of
. The quantum mechanical Hamil-tonian operator is now obtained by replacing
by the operator
.In this procedure, problems can arise due to the commutative nature of the ordinary clas-
sical algebra. The order of the various factors in the classical Hamiltonian is unimportant,and one Hamiltonian with a different order of factors transfers into quantum mechanicswith different results. A solution to this problem is given by Podosky [21], who showedthat reordering (22) into the form
(23)
where is the determinant of the matrix , the correct quantum mechanical form of theHamiltonian operator is obtained with the replacement of
by its quantum mechanicalcorrespondence.
Another difculty is that generally the coordinates used are not Cartesian. Often it is alsomore convenient to use momenta, such as the components of the total angular momentum
that are not conjugate to any of the coordinates. This problem has been discussed by Wilsonand Howard [22] in cases where the momenta used, denoted as
, can be written as linear
5/24/2018 High Resolution Infrared
28/173
28
combinations of the real conjugate momenta
as
(24)
with the inverse
(25)
They then showed that under certain conditions the correct form of the quantum mechanical
operator is now obtained, analogously to (23), from the Hamiltonian
(26)
Here, is the determinant of the matrix , which analogously to (22), includes thecoefcients of the kinetic energy in terms of the new momenta:
(27)
The conditions which would justify the use of (26) are:
(28)
and
(29)
where is the determinant of matrix
and is the weight function in the volumeelement written in terms of the non-Cartesian coordinates used. For example, in polar coordinates where or in Euler angle space where .
The following sections deal with the determination of quantum mechanical Hamiltonianfor a vibrating and rotating molecule. We begin by considering the Born-Oppenheimerapproximation which allows the problems of nuclear and electronic motions largely to beseparated. As a result, a vibrating and rotating molecule can be treated as a collectionof atomic point masses connected by forces keeping the individual atoms at or near theirequilibrium positions. In the frame of this model, the classical Hamiltonian is then set up incoordinates (Euler angles and normal coordinates) that facilitate the maximal separation ofvibration and rotation. The molecule CH
I is considered as an example, and in Section3.4, starting on page 42, we perform some numerical calculations for this molecule. Finally,
in Section 3.5 the quantum mechanical Hamiltonian for a vibrating and rotating moleculeis obtained.
5/24/2018 High Resolution Infrared
29/173
29
3.1. Born-Oppenheimer approximation
The principal approximation in molecular spectroscopy is the Born-Oppenheimer approx-
imation based on the great energy difference of the electronic and nuclear motions. As aresult the rotation-vibration-electronic Schrdinger equation is separated into two parts, theelectronic equation and the rotation-vibration equation.
An important factor in determining the general energy distribution of a molecule is theBorn-Oppenheimer expansion parameter [23]
, where
is the electronmass and
typical nuclear mass, depending upon the specic case. Following the dis-cussion of Bethe and Jackiw [24], the approximate orders of magnitude of the energies ofelectronic (
), vibrational (
), and rotational (
) motions are
and
(30)
The electronic energy
is for a single electron in the molecule resulting from Heisen-bergs uncertainty principle, with the assumption that the electron is conned within adistance of a typical molecular dimension, e.g. intermolecular distance. The result for
is obtained from the following deduction. The vibrational energy of the nuclei for a lowmode of vibration is
, where is the appropriate force constant and is the vibrational quantum number. The force constant may be estimated by assuming thatthe displacements of the nuclei must be to the order of if the molecule is to dissociate.The energy of such a vibration is then about and being now the dissociation energy, itmust be to the order of
. Therefore the energy difference between two neighboring vibra-
tional levels is about
resulting in
The actual nucleardisplacementscan now be approximated by writing
and
.From these results. Thus is to the order of the ratio of a vibrational displacementto an internuclear distance. The moments of inertia of a molecule are about
.The elementary quantum theory of rotators shows that the low rotational energies of thenuclei are calculated from the moment of inertia roughly as
and from this weget
.Let us now consider some numeric values. The Born-Oppenheimer parameteris usu-
ally taken to be to the order of corresponding to typical nuclear masses of roughly 10. Taking the typical molecular dimension to be to the order of ngstrms, say 1 (10
m), the equations (30) give
eV per electron,
eV cm per
vibrational mode, and eV cm
for low rotational states. The corre-sponding values from physical experience for methyl iodide are
eV determinedfrom ionization potentials [25],
cm calculated as mean value from the nine(!) fundamental frequencies (see papers IV - VI), and
cm for low rotationalstates calculated from the rotational constants cm and cm (see Pa-pers I - III). The experimental values agree surprisingly well with the orders of magnitudeand therefore methyl iodide can be considered as a model example of semirigid molecules.
The order of magnitude scheme gives the typical electronic/nuclear velocity ratio of . The electrons, therefore, move much faster than the nuclei and, as a con-sequence, during the time scale of their motion, the electrons see the nuclei practically at
xed conguration. The nuclei, on the other hand, experience the electrons as an averagecharge distribution within the molecule. In making the Born-Oppenheimer approxima-
5/24/2018 High Resolution Infrared
30/173
30
tion, we separate the molecular energy into two parts, one for the electronic motions in
a particular nuclear conguration, and the other for the nuclei vibrating under a potential
eld provided by the electrons in a particular electronic state. The molecular rotations are
considered together with the vibrations.The neglected terms and the way the average potential for the nuclear motion is formed
from the electrons were shown by Born and Oppenheimer in their classic 1927 paper [23].
In reviewing the results of their calculations, the Hamiltonian is taken as
(31)
where runs over the nuclei and over the electrons, and the potential energy terms depend on the electronic coordinates (
) and nuclear coordinates (
). As compared to
the Hamiltonian (2), the small interaction terms
and
have been neglected. If
the nuclei are now frozen in positions as if they were innitely massive, the Hamiltonianbecomes
(32)
and we assume that the electronic eigenvalue problem [
]
hasbeen solved at this particular nuclear conguration. It should be noted that the electronic
Hamiltonian (32) depends in a parametric way (not included in the differential operators)
on the nuclear coordinates. This results in a situation where the energies
, as well asthe wavefunctions
depend on the nuclear coordinates, too. This is because of the
term . In addition, as long as the nuclei are frozen in place, the term isan additive constant in the Hamiltonian having no effect on
but adds directly tothe eigenvalue
The total Hamiltonian becomes now
(33)
What Born and Oppenheimer did was to show that, because
is so small,
can be treated as a small perturbation. They carried out a systematic expansion of the
eigenfunctions and eigenvalues extending the analysis through fourth order to include rota-tional as well as vibrational and electronic parts properly. The suitable expansion parameterturned out to be
To conrm that the results are reasonable, we now write an approximate total eigenfunc-
tion of as a product
, and try to satisfy the equation
(34)
By using the identity and the fact that
isindependent of
but
depends on
, the operations
and
may be expanded. Eq. (34) now yields, after some rearrangements
(35)
5/24/2018 High Resolution Infrared
31/173
31
where
(36)
Ifis now neglected, the electronic wavefunction
is no longer operated on, and it can
be factored out. We are left with
(37)
which is the Schrdinger equation for
showing that the effective potentialeld providedby the electrons is the electronic energy (
included) as a function of the internucleardistances.
The validity of the neglect of the two terms in (36) may be examined by comparing theneglected terms with the nuclear kinetic energy term, which is the smallest term left in the
Hamiltonian. The ratios of these terms are
and
which are small, because
depends much less on
than does
. This results from thefact that
approaches zero as
is displaced by the order of vibrational displacement,whereas
changes signicantly only when
is displaced by the order of internucleardistance. On page 29 the order of the ratio oftois shown to be.
3.2. Vibrational and rotational coordinates
At this point, we adopt the Born-Oppenheimer approximation and do not consider theelectrons in a molecule as specic particles. Our molecule, from now on, is a collectionof nuclei (with atomic masses) moving under a potential eld provided by the averagedmotion of the electrons. We consider isolated, non-degenerate, singlet electronic states andignore the possible electron spin effects. We also assume that the potential surface has arelatively deep minimum with respect to the nuclear displacements holding the nuclei in aconguration with approximatelyxed bond lengths and valence angles. In other words,
we deal with a semirigid molecule with a well dened equilibrium structure.A molecule is free to propagate and rotate in space and simultaneously all its nuclei arein continuous movement relative to each other. The Schrdinger equation of such a com-plex system is virtually impossible to solve accurately. In this section, we shall study in de-tail the coordinates that facilitate the separation of the molecular motion into translational,vibrational, and rotational parts so that as little approximation as possible is necessary.
Two coordinate systems are rst dened. A Cartesian space-xed system is needed to locate the molecule in space and a Cartesian molecule-xed systemis used to describe the instantaneous displacements of the nuclei from their equilibriumpositions in the molecule. The following vectors are then dened (see also Fig. 5):
(1)
(with components
and
) points to the origin of the rotating molecule-xed axis system The origin is by denition at the center of mass of the
5/24/2018 High Resolution Infrared
32/173
32
X
Y
Z
z
x
y
R0
Ri
di
rie
ri
Fig. 5. The position of the-th nucleus with respect to the space-xed axis systemand the moving, molecule-xed axis system. The empty andlled dots represent theequilibrium and the instantaneus positions of the nucleus, respectively.
molecule.
(2)
(
and
) specify the instantaneous positions of the nuclei of the molecule, with respect to space-xedaxes.
(3)
(
, and
) are the equilibrium positions of thenuclei with respect to themolecule-xed axis system
(4)
(
, and
) specify the instantaneous positions of the nuclei in the molecule-xedaxis systemThe displacement vectors for the nuclei are then
The mathematical treatment of the overall motion of a molecule starts with the calcula-tion of the classical kinetic energy in these coordinates but before going into that, the Eulerangles and the normal coordinates are reviewed.
3.2.1. Euler angles
To relate the coordinates of a point in the space-xed () axis system to those inthe molecule-xed rotating () axis system the Euler angles are used. The Euler an-
gles anddene the orientation of the () axis system relative to the ()axis system, both systems having the same origin. The angles are dened, following theconvention of Wilson, Decius, and Cross [12], in Fig. 6a.
The angles andare the polar coordinates of the axis in the () axis systemandis the angle in theplane measuring the rotation about the axis. The angles arerestricted to the ranges , , and are all measuredanticlockwise, as shown in Fig. 6a. The axes () can be obtained from the referenceframe () with three successive rotations represented by the following orthogonal
5/24/2018 High Resolution Infrared
33/173
33
X
Y
Z
x
N
z
x
x
yX
Z
z
x
y
..
.
(a) (b)
O
N
XeX
YeY
ZeZ
eZzez
ez
e
N
yey
xexY
Fig. 6. (a) The denition of the Euler anglesthat relate the orientation of the molecule-xed () axes to the space-xed () axes. The axis is the line of nodes, i.e. theintersection of the and planes. (b) Components of the total angular velocity vector.
rotation operators
=
-
-
-
(38)
The rst matrix (the right-most) rotates () by an angle about the axis to anew orientation ( ), see Fig. 6a. The second (the middle) matrix then preforms therotation by angleabout the newaxis giving the orientation ( ), and the third(left-most) matrix represents the rotation by angle about the newaxis leadingnally to(). As a result, for any particle in space its coordinates in the ( ) and ()axis systems can be related in the following way
=
(39)Here
, etc., are the elements of the direction cosine matrixobtained from the matrixmultiplication of the three matrices in (38). The elements of matrix are given in Table4. It should be noted that the rotation is an orthogonal transformation and therefore theinverse of the direction cosine matrix is equal to its transpose,
An important concept related to the Euler angles is the total angular velocity vectorof the rotating axis system. In Fig. 6b, the components of this vector in the two coordinatesystems are shown. The angular velocity satises the relationship [12]
(40)
5/24/2018 High Resolution Infrared
34/173
34
Table 4. The direction cosines in terms of the Euler angles between the space xed
and the rotating Cartesian axes.
- + - - - - +
where , , and Its components in the rotating axes are nowgiven by the dot products like
, etc, (41)
and from Fig. 6b, with the help of Table 4, we have for example
,
, and
. A short calculation gives the results
(42)
with the inverse equations
(43)
It is also important to determine the transformations that the Euler angles undergo whenthe molecule-xed () axes are rotated about various axes. The rotation
(see page19) of the () axes byradians about an axis in theplane making an anglewiththe axis change the Euler angles fromto . This result isobtained from a careful study of Fig. 6a. Similarly, the rotation
byangles about the
axis changes the angles to . These results, together with the equivalent rotationsdened on page 20, may now be used to determine the changes on the Euler angles causedby the MS group operations. The results for
group are summarized in Table 5.
3.2.2. Normal coordinates
The normal coordinates describe the vibrations in molecules. They are dened as linearcombinations of the mass-weighted Cartesian displacement coordinates
, where
runs over the atomic nuclei,
are the corresponding nuclear masses, and ,,are the Cartesian components of the displacement vectors of the nuclei from their
5/24/2018 High Resolution Infrared
35/173
35
Table 5. The transformation properties of the Euler angles in the
group.
equilibrium positions. These mass-weighted Cartesian coordinates generate, in additionto the genuine normal coordinates
, six nongenuine or zero frequency normalcoordinates
and
( ), which describe the overall translational and rotationalmotions of the molecule. The normal coordinates are written as
(44)
(45)
(46)
(47)
(48)
(49)
(50)
where is the total mass of all the nuclei in the molecule and
is the moment of inertia about the principal axis coinciding with the axis.
and
aregiven by cyclic permutations of,, andin
.The normal coordinates dened above are not unique but must have denite properties.
They must be normal, orthogonal, and they must transform properly under the MS groupoperations of the particular molecule under study. The orthonormality among
and
may easily be veried from Eqs. (45) - (50). The orthonormality among
, Eq. (44), issatised if the transformation coefcients
fulll the relations
(51)
5/24/2018 High Resolution Infrared
36/173
36
From the orthonormality between the coordinates
and
we have
(52)
and the coordinates
and
give
(53)
(54)
(55)
In the study of molecular vibrations, the displacements of the nuclei from their equilib-rium positions should not produce translation or rotation of the molecule. For this reason,the translational and rotational coordinates
and
are chosen to be such combinationsof
that they vanish when the six Eckart conditions [26,27] are used. The Eckart condi-tions ensure that the moleculexed coordinate system is properly locked with the molecule(see Eqs.(66) and (67) later) and they, in this case, may be written as
(56)
Because the transformation (44) - (50) is orthonormal, the inverse transformation for thegenuine normal coordinates may now be written as
(57)
The calculation of the transformation coefcients
requires knowledge of inter-atomic forces (force constants) and the conguration (bond lengths and angles) of the par-ticular molecule. The numeric calculations of frequencies and forms of the normal modesfor CH
I are presented later, starting on page 42.
The normal coordinates form a basis for an irreducible representation of the molecularsymmetry group and they can be symmetry classied accordingly. The determination of thesymmetry species starts with the construction of a reducible representation,red in Eq.(10),using themass-weighted Cartesian displacement coordinates as a basis. For CH
I,we begin by determining the transformation properties of the Cartesian displacement coor-dinates under the symmetry operations of the point group
. The point group, instead ofthe MS group, may now be used since, as discussed in Chapter 2, it is the symmetry groupof the purely vibrational part of the Hamiltonian.
In Fig. 7a, the displacement coordinates for CH
I are shown, and Fig. 7b shows as anexample how the coordinates for the three protons transform under rotation operation
.
The transformation properties under all
operations can now be determined using visualinspection ofgures like Fig. 4b or Fig. 7b, or more easily by directly applying the rotation
5/24/2018 High Resolution Infrared
37/173
37
3z3x1
z1
x3
y2z2
x5
y5z5
(a) (b)
y2
x2
y3
x3
x1
y1
H1
C
I
C3
x2
H3
y2x2
y3x3
x1y1
1
2 2
1
3 3
Fig. 7. (a) The Cartesian displacement coordinates for CH
I, perspective view. (b) Top viewto show the effect of C
on the displacements of the three protons.
and reection operators [12,18,28]
and
, (58)where the mirror plane makes an angle with the axis. For planes
and
theangletakes values,, and , respectively. The results are summarized in Table6.
The symmetry operations mix only symmetrically equivalent nuclei, and Table 6 can
now be used to generate the representation in the mass-weighted basis, too. To determinethe characters inred we only need to consider the diagonal elements of the representationmatrices, and we get
red 15 0 3
This can now be reduced using Eq.(11) and the
character table (Table 2) into
This result is summarized as
red
(59)
5/24/2018 High Resolution Infrared
38/173
38
Table 6. The effect of the
point group symmetry operations on the displacement coor-dinates of the CH
I molecule.
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
+
-
-
-
-
-
-
+
-
+
-
-
-
-
+
-
+
-
-
-
-
+
-
+
-
-
-
-
+
-
+
-
-
-
-
+
-
+
The representation (59) describes the symmetry of all molecular motions, including thethree translations and the three rotations of the molecule as a whole. To obtain the sym-metries of the genuine vibrations it is therefore necessary torst determine the symmetriesof the coordinates
and
. The transformation properties of these coordinates can bedetermined using Eqs.(45) - (50) and the transformation properties of the
coordinatesgiven in Table 6. They are presented in Table 7 where the results for
coordinates areobtained straightforwardly, but the properties for
require some calculation. They in-clude the equilibrium structure of the molecule and in our case we can use relations like
and
, for example, to simplify the calculation. FromTable 7 we now get
rot,tra 6 0 0
which, using Eq.(11), reduces to
rot,tra
(60)
The symmetry of translation
is
and that of rotation
is
(see Table 7). The twopairs
and
belong to the doubly degenerate symmetry species. The
5/24/2018 High Resolution Infrared
39/173
39
Table 7. The effect of the
point group symmetry operations on the translational and
rotational normal coordinates of the CH
I molecule.
-
+
-
-
-
+
-
-
-
-
-
-
+
-
+
-
+
-
-
-
-
+
-
-
-
-
-
-
-
-
-
symmetry properties of the translations and rotations are usually included in the charactertables, see Table 2.
Substraction of rotational and translational motions, Eq. (60), from the total representa-tion Eq. (59) yields the normal coordinate representation for genuine vibrations
vib
(61)
Thus, for CH
I, there are three one dimensional normal vibrations of the type
andthree doubly degenerate vibrations of type . Naturally, the same applies for all symmetrictop isotopomers of methyl iodide.
3.3. Classical Hamiltonian
To obtain the quantum mechanical Hamiltonian operator it isrst necessary to write theclassical Hamiltonian, Eq. (23). The classical kinetic energyin space-xed coordinatestakes a very simple but useless form (we writefor convenience)
(62)
where runs over all the nuclei,
are the nuclear (atomic) masses, and
,where is time. The separation of the motions is achieved by the coordinate transformationfrom space-xed coordinates to molecule-xed coordinates by the equations
, (63)
where
,
,
, and
are the components of the position vectors of the nuclei, described
on page 31 (and in Fig. 5) and
are the direction cosines relating the rotating and non-rotating coordinate axes as functions of Euler angles, as described in Table 4.
5/24/2018 High Resolution Infrared
40/173
40
Substitution of (63) into (62), with some elementary vector calculus with non-inertialcoordinate systems involved, leads to (see e.g. [12])
(64)
where is the total angular velocity vector of the rotating axis system. Its componentswith respect to the rotating axes are given in Eqs. (42).
The rst term on the right hand side of Eq. (64) represents the kinetic energy of thetranslation of the molecule with total mass concentrated in the center of mass. The secondterm is the kinetic energy of the overall rotation of the molecule and the third term is thekinetic energy related to the nuclear vibrations with respect to the moving axes.
The three last terms represent the interaction energies of translation-rotation, translation-vibration, and rotation-vibration, respectively.By denition, the origin of the moving axis system is in the center of mass of the
molecule so that
This results in
, and thus the interactionterms between translation and the other motions vanish. The kinetic energy of the transla-tion is therefore exactly separated. In the absence of externalelds, it has no spectroscopicinterest and will henceforth be ignored. The remaining interaction term connects rotationand vibration and it is called the Coriolis energy. Introducing now the displacement vectors
and using the relation for the scalar triple product, the kinetic energy for avibrating and rotating molecule becomes
(65)
Sofarwehavedened that the moving axis system follows the translation of thewhole molecule. In addition, we have assumed that the axes have somehow been attachedto the molecule following also its rotation. At this point of the calculation, it becomesnecessary to establish exactly how this attachment has been realized.
The means tox the moving axes on to the molecule comes from the redundancies inthe coordinates
and the Euler anglesand . A molecule withatoms hasdegrees of freedom, and so independent coordinates are needed to describe itstranslational, rotational, and vibrational kinetic energies. In the coordinate transformation(63) there are exactly coordinates
on the left-handside of the equation. On the right-hand side, however, the number of coordinates is since there are three coordinates for the center of mass
, three Euler anglesandCartesian coordinates
. So, there are sixredundant coordinates (not independent from the others) and it is necessary to introduce sixconstraints for the
coordinates. These constraints dene how the moving axis systemis attached to the molecule and they can, in principle, be chosen in an innite number ofdifferent ways.
For a molecule in its equilibrium conguration, the molecule-xed axes are locatedalong the principal axes of inertia. For semirigid molecules the nuclear displacements are
5/24/2018 High Resolution Infrared
41/173
41
restricted to small amplitudes, and we could dene theaxes as the instantaneousprinciple axes. However, for a deformed molecule it is more useful to choose the axes to be
the Eckart axes rather than the principal axes. The Eckart axes coincide with the principle
axes when the molecule is in equilibrium, and during the vibrations they are located so thatthe following Eckart conditions are instantaneously met [26, 27]:
(66)
(67)
Therst condition (66) is called the center of mass condition, which we already have
applied earlier. The origin of the axis system is the nuclear (atomic) center ofmass. This condition implies also that and meaning thatduring the vibrations the origin is maintained at the center of mass. The second condition(67) means that during the molecular vibration there must be no zero -order vibrationalangular momentum. The exact vibrational angular momentum is given by
and in demanding that it vanishes, there would be no Coriolis coupling terms to spoil theseparation of rotation from vibration. Unfortunately, with only three conditions left at ourdisposal we cannot accomplish this. To a good approximation, however, we can use
instead of
and condition (67) follows. From (67) we get
and theCoriolis coupling (the second row in Eq. (65)) is minimized.
Introducing now the normal coordinates from Eq.(57), and substituting the componentsof from Eqs. (42) into Eq. (65) gives, after a rather involved but straightforward calcula-tion, (see e.g. [12]) the Hamiltonian form of the kinetic energy:
(68)
Here, indicesandrun over all three components of the rotating axes,
isthe component of the total angular momentum about the center of mass along the rotatingdirection,
is (almost) the vibrational angular momentum along the rotating direction,
is the-component of the inverse of the instantaneous inertia tensor (againalmost), and
is the momentum conjugate to normal coordinate
. We can write
(69)
(70)
(71)
(72)
In these equations,
is called the component of the vibrational angular momentum, al-
though strictly speaking the component of the classical vibrational angular momentumis dened as
This differs from
because
does not exactly equal
5/24/2018 High Resolution Infrared
42/173
42
but contains an additional rotational contribution, Eq. (72). Also
is not exactly the
instantaneous inertial tensor
but contains an additional rotation-vibration contributionas
(73)
The coefcients
are the Coriolis coupling constants depending on the
coefcientsaccording to
(74)
and cyclically for theand coefcients. It must also be pointed out here that the coef-
cients
are not constants, but are functions of the normal coordinates.To write the classical Hamiltonian, we need an expression for the potential energytoo. In general the potential energy is a complicated function of the Cartesian coordinatesof the nuclei, and we do not usually have an analytical form for it. For a semirigid moleculethe nuclei vibrate only in the neighborhood of their equilibrium positions so that the nor-mal coordinates are restricted to small amplitudes. The potential functioncan then beexpanded in a power series of the normal coordinates
in the point of equilibrium
(75)
where
are successive potential energy derivatives. In (75) the poten-tial energy is chosen to be zero at equilibrium, and by denition all the linear terms vanish.The transformation coefcients
are chosen to reduce the quadratic terms in (75) todiagonal form.
3.4. Harmonic forceeld calculations for CH
I
The CH
I molecule has ve atoms and hence nine fundamental vibrations, three
vibrations, and three doubly degenerate typevibrations, as was shown on page 39. Ourpurpose is now to calculate for CH
I numerical values for the transformation coefcients
transforming the mass-weighted Cartesian displacement coordinates to the normalcoordinates (Eq. (44)) and thus determine the actual forms of the normal vibrations. Inaddition to this, the calculation gives numerical values for the vibrational frequencies, aswell as for some vibration-rotation interaction parameters.
The normal coordinate analysis of molecular vibrations has become a routine approachin assigning individual bands in the spectra to the characteristic vibrations of certain atomicgroups in the molecules. The mathematical principles of the calculations may be found innumerous references, for example Refs. [12,15,2831]. The normal coordinates are closely
related to the quadratic terms in the expansion of the potential function and we shall there-fore adopt the harmonic oscillator approximation in the treatment of the vibrations. Werst
5/24/2018 High Resolution Infrared
43/173
43
consider the direct approach to solve the vibrational problem in mass-weighted Cartesiandisplacement coordinates. Then, two sets of internal coordinates, the valence coordinatesand the symmetry coordinates, are introduced. Next, we adopt from literature values for
the molecular structure and for the harmonic forceeld expressed in symmetry coordi-nates, and carry out the numerical calculations for CH
I. Finally, the Coriolis couplingconstants and the centrifugal distortion parameters describing the vibrational-rotational in-teraction will be calculated.
3.4.1. Vibrational problemdirect approach
To obtain the equation of motion for our molecule, werst write the vibrational kinetic en-ergy and the potential energy in terms of mass-weighted Cartesian displacement coor-dinates introduced in Section 3.2.2. The coordinates are
for atom 1,
for atom 2, ,
for atom , or more generally
The energies are (we writeagainandfor convenience)
and
(76)
where and
. The coefcients
are the quadraticforce constants characterizing the restoring forces acting on the distorted molecule. Usingnow Newtons equations of motion in the Lagrangian form
(77)
we obtain for each
,
(78)
This is a set ofsimultaneous second-order differential equations having the general
solution
(79)
This is the equation of a simple harmonic oscillator with frequency (in cm ) ,amplitude
and phase constant.By substituting (79) back to (78) we obtain a set oflinear homogenous equations
(80)
where
is the Kronecker delta. This set of equations has nontrivial (not zero) solutionsfor the amplitudes
only for certain specic values of. These can be found by setting
5/24/2018 High Resolution Infrared
44/173
44
the determinant of the coefcients
equal to zero. The equation then takes the form
......
. . ....
(81)which is called the secular determinant. The secular determinant provides us withcharacteristic values for
from which the harmonic frequencies can be calculated. By
substituting each
back to equations (80), the corresponding amplitudes
are obtained.
Equations (80) do not determine the amplitudes uniquely, but only their ratios are obtained.
This is, however, sufcient to characterize the vibrations, and the coefcients
are
obtained from normalized amplitudes as
.
The results describe the typical behavior of nucle