SPRINGER BRIEFS IN MOLECULAR SCIENCE

Tatsuhisa Kato
Naoki Haruta
Tohru Sato

Vibronic
Coupling
Density
Understanding
Molecular
Deformation
SpringerBriefs in Molecular Science
SpringerBriefs in Molecular Science present concise summaries of cutting-edge
research and practical applications across a wide spectrum of fields centered around
chemistry. Featuring compact volumes of 50 to 125 pages, the series covers a range
of content from professional to academic. Typical topics might include:
• A timely report of state-of-the-art analytical techniques
• A bridge between new research results, as published in journal articles, and a
contextual literature review
• A snapshot of a hot or emerging topic
• An in-depth case study
• A presentation of core concepts that students must understand in order to make
independent contributions

Briefs allow authors to present their ideas and readers to absorb them with minimal
time investment. Briefs will be published as part of Springer’s eBook collection,
with millions of users worldwide. In addition, Briefs will be available for individual
print and electronic purchase. Briefs are characterized by fast, global electronic
dissemination, standard publishing contracts, easy-to-use manuscript preparation
and formatting guidelines, and expedited production schedules. Both solicited and
unsolicited manuscripts are considered for publication in this series.

More information about this series at http://www.springer.com/series/8898

Tatsuhisa Kato · Naoki Haruta · Tohru Sato

Vibronic Coupling Density

Understanding Molecular Deformation
Tatsuhisa Kato Naoki Haruta
Fukui Institute for Fundamental Chemistry Fukui Institute for Fundamental Chemistry
Kyoto University Kyoto University
Sakyo-ku, Kyoto, Japan Sakyo-ku, Kyoto, Japan

Tohru Sato
Fukui Institute for Fundamental Chemistry
Kyoto University
Sakyo-ku, Kyoto, Japan

ISSN 2191-5407 ISSN 2191-5415 (electronic)

SpringerBriefs in Molecular Science
ISBN 978-981-16-1795-9 ISBN 978-981-16-1796-6 (eBook)
https://doi.org/10.1007/978-981-16-1796-6

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface

Nobel laureate Roald Hoffmann pointed out, “Understanding and Explaining and
their Strong-Tie to Teaching” in the memorial lecture at the Kenichi Fukui 100th
birthday anniversary 2018 in Kyoto.1 He sketched a possible path to coexistence,
way from theory and understanding through simulation and then back again to
understanding. We want to quote his sketches. He gave the process of understanding,
insight, and explanation, and then prediction. Understanding is often tacit, or silent, a
state of mind. It is usually qualitative, though it may have quantitative aspects. Quan-
titative reinforces qualitative. And simulation is around the corner. An explanation is
inherently more pedagogic and storytelling. More useful to science are hypotheses,
alternative narratives. An explanation comes to us through stories, chemical stories.
Prediction is the conceptual passage between understanding and simulation and is
the practical counterpart of contemplative understanding, “I know how” instead of
“I know why.”
A molecule can be regarded as a system of electrons in a framework of nuclei,
that is, a molecular structure. Chemists are interested in molecular motions under
the interactions of other molecules or an electromagnetic field. The motion of a
molecule is decomposed into translational, rotational, and vibrational modes. An
intramolecular motion, i.e., molecular deformation, is expressed as a combination of
vibrational modes.
Exactly speaking, vibrations and motions of electrons cannot be separated. There-
fore, any change of an electronic state gives rise to a molecular deformation.
The couplings among vibrations and electrons are called vibronic coupling (VC).
The magnitude of deformation depends on the strength of VCs, vibronic coupling
constants (VCCs). The direction of deformation depends on the relative values of the
VCCs of vibrational modes. Thus, if we can understand the reason for the relative
strength and magnitude of VCCs, we can explain the molecular deformation under
a certain interaction. Vibronic coupling density (VCD), a density form of a VCC,

1 Roald Hoffmann, “Simulation versus Understanding: A tension, and not just in Quantum Chem-
istry.”, the lecture in the Memorial Symposium of “Kenichi Fukui 100th birthday anniversary,”
Kyoto, 2018.

v
vi Preface

enables us to explain such a reason. In this monograph, we will give the instructive
path to the VCD and the VCC analyses.
Chemistry in alchemy was just aimed at getting gold without understanding.
Modern chemistry emerged with the concept of atoms and molecules. After the
quantum study, the electron behavior circulating nuclei was led to the principal
concept underlying all explanations in chemistry. Many textbooks have given the
plausible explanations to clarify the molecular structure. And the frontier molecular
orbital concepts were proposed to visualize the path of a chemical reaction. The
conventional explanations have provided students with a considerable familiarity
with the molecular structure in terms of the electronic state. However, the more
rational and more convincing ways should be given. Here the VCD and the VCC
analyses are introduced. They are starting from the ab initio molecular Hamilto-
nian, and systematic, rational ways to understand chemical phenomena, and which
can give the quantitative evaluation of the force applied under the chemical defor-
mation process. We offer the guidelines to integrate the traditional “hand-waving”
approach of chemistry with more rational and general VCD and VCC alternative.
Further outlooks for the newly functionalized chemical systems. Thus, through the
visualization by VCD and the evaluation by VCC, the study of chemistry by molec-
ular orbital theory is brought into the domain of substantial science, where qualitative
concepts can be rendered quantitatively and tested rigorously against the quantum
theory.

Kyoto, Japan Tatsuhisa Kato

February 2021 Naoki Haruta
Tohru Sato
Acknowledgements

The authors are greatly indebted to students who made this book possible: Ken Toku-
naga, Motoyuki Uejima, Katsuyuki Shizu, Naoya Iwahara, and Yuichiro Kameoka.
TS acknowledges colleagues for guiding his theoretical studies: Kazuyoshi Tanaka,
Arnout Ceulemans, and Liviu F. Chibotaru, and for discussions on interpretations of
experimental results: Hironori Kaji, Tsunehiro Tanaka, Kentaro Teramura, Saburo
Hosokawa, Kenji Matsuda, and Takashi Hirose.

vii
Contents

1 Qualitative Explanation of Molecular Structures by Various

Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Elongation of Bond on Anionization of H2 and C2 H4 . . . . . . . . . . . . 1
1.2 Bent Deformation of H2 O and NH2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Nonplanar Molecular Structure of NH3 . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Zigzag Conformation of Cycloparaphenylene . . . . . . . . . . . . . . . . . . . 6
1.5 Stable Structure of C3 H3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.6 Frontier Orbital Theory for Diels–Alder Reaction . . . . . . . . . . . . . . . 10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 Molecular Deformations and Vibronic Couplings . . . . . . . . . . . . . . . . . . 13
2.1 Introduction: Molecular Deformations and Its Driving Force . . . . . . 13
2.2 Molecular Geometry and Modes of Motion . . . . . . . . . . . . . . . . . . . . . 14
2.3 Coordinates of Non-linear A3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Coordinates of Non-linear A2 B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.5 Approximate Molecular Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.6 Energy as a Function of Molecular Deformation . . . . . . . . . . . . . . . . 27
2.7 Molecular Deformation Due to Vibronic Couplings . . . . . . . . . . . . . . 28
2.7.1 Force Acting on Molecule: Linear Vibronic Couplings . . . . 29
2.7.2 Spontaneous Symmetry Breaking of Non-linear
Molecule: Jahn–Teller Effect . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.7.3 Spontaneous Symmetry Breaking of Non-linear
Molecule: Pseudo-Jahn–Teller Effect . . . . . . . . . . . . . . . . . . . 32
2.7.4 Spontaneous Symmetry Breaking of Linear Molecule:
Renner–Teller Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3 Vibronic Coupling Density Analyses for Molecular Deformation . . . . 37
3.1 Elongation Force on Anionization of H2 and C2 H4 . . . . . . . . . . . . . . 37
3.2 Molecular Deformation of NH2 and H2 O by Renner–Teller
Effect and Pseudo-Jahn–Teller Effect . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.3 Vibronic Coupling Density of Ammonia Molecule NH3 . . . . . . . . . . 42

ix
x Contents

3.4 Vibronic Coupling Density Explanation for Zigzag

Conformation of 6-Cycloparaphenylene . . . . . . . . . . . . . . . . . . . . . . . 44
3.5 Vibronic Coupling Density and Jahn–Teller Hamiltonian
of C3 H3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.6 Vibronic Coupling Density Picture of Diels–Alder Reaction . . . . . . 51
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4 Design for Functional Molecules by Vibronic Coupling Density . . . . . 57
4.1 Design for Carrier-Transporting Molecules . . . . . . . . . . . . . . . . . . . . . 57
4.2 Design for Light-Emitting Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5 Definitions and Derivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.1 Dirac’s Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.2 Group and Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.2.1 Symmetry Operation and Wave Function:
One-Dimensional Wall as an Example . . . . . . . . . . . . . . . . . . 73
5.2.2 Group and Its Representation: Two-Dimensional Wall
as an Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.3 Point Group and Its Irreducible Representation:
Triatomic Molecules as an Example . . . . . . . . . . . . . . . . . . . . 86
5.3 Molecular Vibrations and Normal Modes . . . . . . . . . . . . . . . . . . . . . . 87
5.4 Vibronic Coupling Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.4.1 Vibronic Coupling Constant and Vibronic Coupling
Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.4.2 Higher-Order Vibronic Coupling Density . . . . . . . . . . . . . . . . 98
5.4.3 One-Electron Property Density . . . . . . . . . . . . . . . . . . . . . . . . . 99
5.4.4 Related Concepts to Vibronic Coupling Density . . . . . . . . . . 102
5.4.5 Vibronic Coupling Density as a Chemical Reactivity
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.4.6 Physical Quantities Related to Vibronic Coupling
Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Acronyms

AO Atomic Orbital
AVCC Atomic Vibronic Coupling Constant
CASSCF Complete Active Space Self-consistent Field
DFT Density Functional Theory
EVCD Effective Vibronic Coupling Density
HOMO Highest Occupied Molecular Orbital
irrep irreducible representation
JT Jahn–Teller
JTE Jahn–Teller Effect
LUMO Lowest Unoccupied Molecular Orbital
MO Molecular Orbital
OVCC Orbital Vibronic Coupling Constant
OVCD Orbital Vibronic Coupling Density
PJTE Pseudo-Jahn–Teller Effect
QVCC Quadratic Vibronic Coupling Constant
QVCD Quadratic Vibronic Coupling Density
SOMO Singly Occupied Molecular Orbital
TDMD Transition Dipole Moment Density
VB Valence Bond
VC Vibronic Coupling
VCC Vibronic Coupling Constant
VCD Vibronic Coupling Density
VSEPR Valence Shell Electron Pair Repulsion

xi
Symbols

x Cartesian coordinate in the three-dimensional space

x, y, z x, y, z-components of x, relative coordinates (Chap. 2.2)
X,Y,Z Coordinates of the center of mass (Chap. 2.2)
r A set of spatial coordinates of all electrons
i, j Labels of electrons, Labels of atom (Chap. 2.2)
li Bond length of atom i and i + 1
ϕi Bond angle of atom i − 1, i, and i + 1
χi Dihedral angle of atom i − 1, i, i + 1, and i + 2
ri Spatial coordinate of electron i
xi , yi , zi x, y, z-components of ri
ω A set of spin coordinates of all electrons
ωi Spin coordinate of electron i
R A set of spatial coordinates of all nuclei
A, B Labels of nuclei
RA Spatial coordinate of nucleus A
XA , YA , ZA x, y, z-components of RA
R0 Reference nuclear configuration
R0A Spatial coordinate of nucleus A in the reference nuclear configu-
ration
XA0 , YA0 , ZA0 x, y, z-components of R0A
R Displacement of all nuclei from R0
RA Displacement of nucleus A from R0A
XA , YA , ZA x, y, z-components of RA
Q A set of all mass-weighted vibrational coordinates
α, β Labels of vibrational modes
Qα Mass-weighted vibrational coordinate of mode α
qα Vibrational coordinate of mode α in the real space
uα Vibrational vector of mode α in the mass-weighted space
vα Vibrational vector of mode α in the real space
μα Reduced mass of mode α
ωα Frequency of mode α
vα (x) Linear potential derivative with respect to vibrational mode α
xiii
xiv Symbols

wαβ (x) Quadratic potential derivative with respect to vibrational modes

α, β
wα1 ···αk (x) k-th-order potential derivative with respect to vibrational modes
α1 · · · αk
m, n Labels of electronic states
m (r, R) m-th total electronic wave function at the nuclear configuration
R
Em (r, R) m-th total electronic energy at the nuclear configuration R
ρm (x) Electron density in the m-th electronic state
ρm (x) Electron density difference between the reference and the m-th
electronic states
ρmn (x) Overlap density between m-th and n-th electronic states
ρab (x) Overlap density between a-th and b-th molecular orbitals
Vmn,α Linear vibronic coupling constant between m-th and n-th elec-
tronic states with respect to vibrational mode α
Wmn,αβ Quadratic vibronic coupling constant between m-th and n-th
electronic states with respect to vibrational modes α, β
Wmn,α1 ···αk k-th-order vibronic coupling constant between m-th and n-th
electronic states with respect to vibrational modes α1 · · · αk
ηmn,α (x) Linear vibronic coupling density between m-th and n-th elec-
tronic states with respect to vibrational mode α
ηmn,αβ (x) Quadratic vibronic coupling density between m-th and n-th
electronic states with respect to vibrational modes α, β
ηmn,α1 ···αk (x) k-th-order vibronic coupling density between m-th and n-th
electronic states with respect to vibrational modes α1 · · · αk
Vab,α Linear orbital vibronic coupling constant between a-th and b-th
molecular orbitals with respect to vibrational mode α
Wab,αβ Quadratic orbital vibronic coupling constant between a-th and
b-th molecular orbitals with respect to vibrational modes α, β
Wab,α1 ···αk k-th-order orbital vibronic coupling constant between a-th and
b-th molecular orbitals with respect to vibrational modes α1 · · · αk
ηab,α (x) Linear orbital vibronic coupling density between a-th and b-th
molecular orbitals with respect to vibrational mode α
ηab,αβ (x) Quadratic orbital vibronic coupling density between a-th and b-th
molecular orbitals with respect to vibrational modes α, β
ηab,α1 ···αk (x) k-th-order orbital vibronic coupling density between a-th and
b-th molecular orbitals with respect to vibrational modes α1 · · · αk
Chapter 1
Qualitative Explanation of Molecular
Structures by Various Approaches

Abstract Valence bond (VB) picture and valence shell electron pair repulsion prin-
ciple (VSEPR) [1–3] are localized bond approaches. VB picture is the historic
landmark for modern chemistry, which is based on the valence bond. It invests each
atom with the valence according to the number of electron, which is the basic concept
for chemical bonds of the molecule; however, it has the fatal fault in which hybridiza-
tion of the bonding orbital is determined after empirically knowing the molecular
structure. VSEPR is more applicable to elucidate various molecular structures but
is still an empirical picture. Both have no proper way to explain molecular struc-
ture in the excited state. On the other hand, molecular orbital (MO) theory is the
approach based on delocalized electrons. It starts from the molecular Hamiltonian,
and is a more systematic and rational way to understand molecular structures. Many
textbooks have given explanations for the molecular deformation [4]. In this chapter,
several well-known issues are discussed, the bond elongation of ethylene under the
anionization, the bent structure of water, the nonplanar structure of NH3 , the tau-
tomerism between benzenoid and quinoid forms of organic molecule, the triangle
structure of C3 H3 , and the chemical process of Diels–Alder reactions.

Keywords Valence bond · Valence shell electron pair repulsion principle ·

Molecular orbital · Anionization · Walsh diagram · Pseudo-Jahn-Teller theory ·
Renner-Teller theory · Highest occupied molecular orbital · Lowest unoccupied
molecular orbital · Tautomerism · Benzenoid · Quinoid · Jahn-Teller effect ·
Frontier orbital theory · Diels-Alder reaction

1.1 Elongation of Bond on Anionization of H2 and C2 H4

Here molecular orbitals of hydrogen are shown in Fig. 1.1. It is well known that
when a neutral hydrogen molecule acquires an electron, and the chemical bond
is elongated, since the additional electron occupies the anti-bonding 1σu lowest
unoccupied molecular orbital (LUMO).

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 1
T. Kato et al., Vibronic Coupling Density,
SpringerBriefs in Molecular Science,
https://doi.org/10.1007/978-981-16-1796-6_1
2 1 Qualitative Explanation of Molecular Structures by Various Approaches

Fig. 1.1 Molecular orbitals

of hydrogen molecule

Anionization of ethylene also gives rise to an elongation of the double bond in the
process of the electron occupation of the anti-bonding π -orbital. The bonding MO is
composed of the in-phase mixing of atomic orbitals, which gives the positive overlap
of the electronic wave function resulting in an attractive force between nuclei. On
the other hand, the anti-bonding MO exhibits a repulsive force between nuclei. In
the case of hydrogen, the effect is partly because an electron occupying the anti-
bonding orbital is excluded from the internuclear region, and hence is distributed
largely outside the bonding region. In effect, whereas an electron occupying the
bonding orbital pulls two nuclei together, an electron occupying the anti-bonding
orbital pulls the nuclei a part [5]. However, in the case of ethylene, no substantial
interpretation of the repulsive force is given, nor the quantitative estimation of the
force so far in both cases. The vibronic coupling density (VCD) analysis will give
the exact physical picture of the repulsive force in Sect. 3.1.

1.2 Bent Deformation of H2 O and NH2

The VSEPR approach [1–3] gives an interpretation for the bent molecular structure
of H2 O as the following. An oxygen atom has six electrons occupying its outermost
shell. Each of the two H atoms donates one electron to the O atom, locating eight
electrons around the O atom. Thus, four pairs of the electrons would be distributed in
a tetrahedral fashion around the central atom if the potential around the O atom was
isotropic. Among them, two pairs form two O–H bonds while the other two pairs
remain unused, which are called lone pairs. Due to repulsion among the bonding
electrons and lone pairs, the bond angle H–O–H is not exactly tetrahedral (109◦ 28 ).
Thus, H2 O becomes a bending molecule with the H–O–H angle of 104◦ . However,
VSEPR approach just gives the interpretation of the actual H–O–H angle of 104◦
less than 109◦ 28 in tetrahedral under the empirical assumption of H2 O molecule’s
electronic structure as shown in Fig. 1.2, not the interpretation why the molecule
becomes bent.
More reasonable interpretation for the bent structure of H2 O has been proposed
by using the Walsh diagram [6] based on the MOs in Fig. 1.2. The Walsh diagram is
the correlation diagram classifying the molecular orbitals of H2 A according to their
1.2 Bent Deformation of H2 O and NH2 3

Fig. 1.2 a Molecular orbitals of H2 A molecule in bent structure of C2v symmetry. b Molecular
orbitals of H2 A molecule in linear structure of D∞h symmetry

symmetries, which leads to the primitive approach to predict the structure for H2 A
molecule, and suggests the useful approach with the orbital symmetry. Supposing the
linear structure of H2 A, the HAH angle is 180◦ , and the point group of the molecule
is D∞h . The MOs in the linear structure are shown in Fig. 1.2b. The fourth and fifth
orbitals are two degenerate πu -orbitals composed of pz and p y atomic orbitals of
the central oxygen. When the H–A–H angle is decreasing from 180◦ to 90◦ , the
former πu ( pz )-orbital interacts with the upper unoccupied 3σg+ -orbital and has an
overlap with the symmetric pair of hydrogen orbitals. As a result, the πu ( pz )-orbital
is stabilized with the decrease of the H–A–H angle as shown in Fig. 1.3. The other
orbital of πu ( p y ) remains at the same energy level. On the other hand, the orbital
energies of lower lying σu and σg increase with the decrease of the overlap of px
with hydrogen orbitals. As a result, the correlation diagram can be drawn according
to their symmetries as shown in Fig. 1.3, which applies to the molecule of type H2 A.
When the diagram is applied to H2 O, and eight valence electrons are configured on
MOs, all four MOs are occupied, and the configuration would be stabilized at the
angle of 104.5◦ , not at 90◦ nor 180◦ . In the case of NH2 , seven electrons occupy
the four orbitals, doubly occupy lower-lying three orbitals and singly occupy the
fourth orbital. Then the electronic state of NH2 is also stabilized at the deformed
bent structure by the same mechanism.
Walsh diagram is intuitive and points out that the symmetry of MO plays a deci-
sive part of determining molecular structures, but is quite qualitative. Going back to
the molecular Hamiltonian to know the origin of molecular deformation, the other
approach of the pseudo-Jahn–Teller theory [7] and the Renner–Teller theory [8] gives
more strict discussion on the symmetry of MO with respect to the deformation coordi-
nate. The mechanism that the ground state of H2 O is stabilized with the deformation
coordinate will be explained by the pseudo-Jahn–Teller effect in Sect. 3.2. In the
case of the linear molecule NH2 , the degenerate states are spontaneously split with
respect to the out of axis deformation by the Renner–Teller effect, also explained in
Sect. 3.2.
4 1 Qualitative Explanation of Molecular Structures by Various Approaches

Fig. 1.3 Walsh diagram: the

correlation diagram of MO
energies according to their
symmetries

1.3 Nonplanar Molecular Structure of NH3

The nonplanar structure of NH3 has been understood in the VSEPR scheme [1–3].
Nitrogen atom has five electrons which occupy its outermost shell. Each of the three
H atoms donate one electron to the N atom, locating eight electrons around the N
atom. Thus, the four pairs of electrons would be distributed in a tetrahedral fashion
around the central atom. Three pairs form the three N–H covalent bonds, while the
fourth pair remains unused as a bond. The fourth pair is called lone pair. Due to
the difference of repulsion between the pair of electrons consisting of covalent bond
and the lone pair, the bond angle H–N–H is not exactly tetrahedral (109◦ 28 ) but
is 107.8◦ . Thus, NH3 is understood as a trigonal pyramidal molecule. However, this
scheme of understanding is again based on the assumption that three N–H bonds are
not aligned in-plane, and just describes the reason why the angle H–N–H is not an
exact tetrahedral angle. There is no explanation of the force to make the structure
bent.
Here, we can explained for the bent structure of NH3 by using Walsh diagram
as shown in Fig. 1.5. Assuming the planar structure of NH3 , the point group of the
molecule is D3h . The MO of the valence electrons is shown in Fig. 1.4. The orbital of
a2 is the HOMO (highest occupied molecular orbital) of NH3 and the orbital 2a1 is
1.3 Nonplanar Molecular Structure of NH3 5

Fig. 1.4 Molecular orbitals

of NH3 with the planar
structure of D3h

Fig. 1.5 Walsh diagram of

NH3
6 1 Qualitative Explanation of Molecular Structures by Various Approaches

the LUMO. Drawing the correlation of MO energies according to their symmetries,

the diagram shown in Fig. 1.5 is given with a planar and pyramidal configuration.
The HOMO orbital is well stabilized with changing from the a2 -orbital in the planar
structure to the a1 in the pyramidal structure because of the resonance interaction of
pz -orbital with the LUMO 2a1 . On the other hand, the energy levels of degenerate
e-orbitals are increasing since the overlap of px - or p y -orbital with the corresponding
combination of the hydrogen decreases. When eight valence electrons occupy the
four MOs, the minimum sum of orbital energies is fulfilled at some position of the
pyramidal structure. The Walsh diagram for NH3 is again entirely qualitative, and
more systematic interpretation will be given by the pseudo-Jahn–Teller theory [7] in
Sect. 3.3.

1.4 Zigzag Conformation of Cycloparaphenylene

The cycloparaphenylene (CPP) is the cyclic molecule that consists of poly-phenylene

connecting in the para positions. We can regard it as the shortest carbon nanotube. For
example, the shortest unit of the (6, 6) armchair carbon nanotube is 6-CPP, as shown
in Fig. 1.6. Then, it is usually supposed to be a belt-like molecule. However, it is
not the case. Experimental data and theoretical calculation show that the most stable
molecular structure is not a belt-like structure but an alternating zigzag conformation
in Fig. 1.7. The X-ray diffraction of 6-CPP reported a non-zero dihedral angle Cortho –
Cipso –Cipso –Cortho [9], which is defined in Fig. 1.7, and the MO calculation predicted
the dihedral angle of 15◦ [10].
The interconversion between belt-like and alternating zigzag conformations is
described in organic-chemical point of view as the tautomerism between benzenoid
and quinoid forms as shown in Fig. 1.8. This is the intuitive way of the explanation

Fig. 1.6 6-CPP as the

shortest unit of the (6, 6)
armchair carbon nanotube
1.4 Zigzag Conformation of Cycloparaphenylene 7

Fig. 1.7 Upper: belt-like

structure of 6-CPP, lower:
alternating zigzag
conformation. Cortho and
Cipso carbons are those at the
designated positions of
molecule, the angle formed
by Cortho , Cipso , Cipso , and
Cortho is defined as a dihedral
angle
Cortho –Cipso –Cipso –Cortho

Fig. 1.8 Tautomerism

between benzenoid and
quinoid forms

of the electron delocalization. The X-ray diffraction of 8-CPP reported that almost
equivalent bond lengths in benzene unit [11], and concluded that the approximately
equal bond lengths indicate that a benzenoid structure is preserved in 8-CPP. As a
result, the bond between neighboring benzene units has a single bond nature and
the dihedral angle is not necessarily zero. On the other hand, the dihedral angle of
the quinoidal form is forced to be zero because of the double bond nature between
neighboring benzene units. However, the description of tautomerism just classifies
the fashion of the electron delocalization. The strain energy with the bend angle at
8 1 Qualitative Explanation of Molecular Structures by Various Approaches

Fig. 1.9 ORTEP figure of

the single crystal of 8-CPP2+
[13]

the Cipso carbon and the steric interactions between neighboring Cortho carbons have
been usually described for the force leading to each form [12].
The 8-CPP radical cation and dication have recently been isolated and character-
ized [13] as shown in Fig. 1.9. The results of single-crystal X-ray analysis showed
a significant decrease in the averaged dihedral angle between the neighboring para
phenylene units of the dication species as compared to that observed in the neutral
species [11]. The analysis also revealed that the Cipso –Cipso and Cortho –Cortho bonds
shorten, while the Cipso –Cortho bonds of the dication elongate as compared to those
in the neutral species. In other words, this signifies a change in the bond-alternation
pattern into the quinoid form. This interconversion from the benzenoid to the quinoid
form due to the ionization would indicate that the zigzag conformation is originated
from the electronic structure of the molecule, not from the steric hindrance between
neighboring Cortho carbons. In Sect. 3.4, the explanation by using the VCD is given
for the origin of the zigzag conformation of 6-CPP.

1.5 Stable Structure of C3 H3

Jahn and Teller found that all the non-linear nuclear configurations are unstable
for an orbitally degenerate electronic state [14]. This effect was named after Jahn
and Teller, called Jahn–Teller effect. The Jahn–Teller effect is usually referred as
the mechanism explaining the ligand deformation of a metal complex in inorganic
chemistry.
For example, a d 9 metal ion of Cu2+ is octahedrally coordinated, the metal dx y -,
d yz -, and dx z -orbitals are equivalent and involve electron density between the axes
containing both the metal ion and the ligands. Both of the dz2 - and dx2 − y 2 -orbitals
direct electron density toward the ligands. The octahedral arrangement of ligands
around the metal ion splits the five d-orbitals into two sets, as shown in Fig. 1.10, one
set (t2g ) being triply degenerate and the other (eg ) being doubly degenerate. The t2g -
orbitals are stabilized, and the eg -orbitals are destabilized relative to their energies
in a spherical field, the energy difference between the t2g - and eg -orbitals being
designated 0 as shown in Fig. 1.10. The driving force of the Jahn–Teller distortion
1.5 Stable Structure of C3 H3 9

Fig. 1.10 The electronic energy levels for Cu2+ in a spherical field (left), an octahedral field
(middle), and an elongated octahedral field (right)

of Cu2+ in octahedral ligand results from the unequal occupancies of the two eg -
orbitals, which are split by the distortion shown in Fig. 1.10. Where the Jahn–Teller
distortion occurs, the singly occupied orbital is destabilized by the same energy as
the doubly occupied orbital is stabilized, resulting in a split by the energy of  J T .
Ligand-field arguments show that single occupancy of the dz2 - and dx2 − y 2 -orbitals
of Cu2+ will result in compressed and elongated octahedral ligand, respectively.
The structure of an organic molecule, for example, C3 H3 can also be predicted
from the Jahn–Teller effect (JTE), which explains the deformation from the equilat-
eral triangle structure of C3 H3 [15]. Let us consider the electronic state of C3 H3 in
the simple Hückel MO theory using valence π -orbitals. In the equilateral triangle
structure of the D3h symmetry, three π -orbitals of eθ , e , and a1 are composed as
shown in Fig. 1.11. The eθ - or e -orbital is partially occupied in the ground electronic
configuration resulting in the doubly degenerated ground electronic state. Among 12
vibrational normal modes, shown in Fig. 3.12, the E  modes give rise to the deforma-
tion causing the energy split for the doubly degenerated ground electronic state. The
E  modes are called Jahn–Teller active modes as referred in Sect. 3.5. Figure 1.11
shows the deformation along the E  (2) mode which stabilizes the orbital energy of
the eθ -orbital, which is symmetric with one plane σv perpendicular to the molecular
plane, and destabilizes that of the e -orbital, anti-symmetric with σv . And vice versa
10 1 Qualitative Explanation of Molecular Structures by Various Approaches

Fig. 1.11 The Jahn–Teller effect of C3 H3

in the opposite direction of the E  (2) vibration. This is the simple interpretation that
the E  (2) deformation causes the energy split for the doubly degenerated ground
electronic state.

1.6 Frontier Orbital Theory for Diels–Alder Reaction

One of the central issues in chemistry is regioselectivity: which region is reactive

in a molecule. The scheme of the HOMO–LUMO interaction between reactants is
stated by the fundamental proposition that a majority of chemical reactions should
take place at the position and in the direction of maximum overlapping of the HOMO
and the LUMO of the reactants [16].
For example, the preferable path of the Diels–Alder reaction between ethylene
and butadiene is explained as the approach between the HOMO of butadiene and the
LUMO of ethylene giving the maximum overlapping shown in Fig. 1.12.
Since molecular deformations play an essential role in chemical reactions, the
concept of VCD can also apply for the regioselectivity problems in chemical reac-
tions.
In the VCD picture of the Diels–Alder reaction, we can apply the similar isolated
reactant approach for larger and complicated reaction systems as of styrene and of
C60 in Sect. 3.6. We will define the effective vibronic coupling density (EVCD) to
clarify the regioselectivity of the Diels–Alder reaction with ethylene to C60 .

Fig. 1.12 Diels–Alder

reaction
References 11

References

1. Gillespie RJ, Nyholm RS (1957) Quart Rev Chem Soc 11:339

2. Gillespie RJ (1963) J Chem Educ 40:295
3. Gillespie RJ (2008) Coord Chem Rev 252:1315
4. Schastnev PV, Schegoleva LN (1995) Molecular distortions in ionic and excited states. CRC
Press Inc, New York
5. Atkins P, de Paula J (2006) Physical chemistry, 8th edn. W. H. Freeman and Company, New
York
6. Walsh AD (1953) J Chem Soc 2260–2331
7. Bersuker I, Gorinchoi N, Polinger V (1984) Theo Chim Acta 66:161
8. Renner R (1934) Z Physik 92:172
9. Xia J, Jasti R (2012) Angew Chem Int Ed 51:2474
10. Wong BM (2009) J Chem Phys C 113:21921
11. Xia J, Bacon JW, Jasti R (2012) Chem Sci 3:3018
12. Bachrach SM, Stück D (2010) J Org Chem 75(19):6595
13. Kayahara E, Kouyama T, Kato T, Takaya H, Yasuda N, Yamago S (2013) Angew Chem Int Ed
52:13722
14. Jahn HA, Teller E (1937) Proc R Soc Lond A 161:220
15. Kayi H, Garcia-Fernandez P, Bersuker IB, Boggs JE (2013) J Phys Chem A 117:8671
16. Fukui K (1971) Acc Chem Res 4(57)
Chapter 2
Molecular Deformations and Vibronic
Couplings

Abstract We define a molecular deformation, and discuss how to specify molecular

deformations in a diatomic and triatomic molecules A3 and A2 B taken as exam-
ples. An approximate molecular Hamiltonian for a polyatomic molecule is intro-
duced. The origins of molecular deformations, vibronic couplings, are discussed for
non-degenerate electronic states and degenerate electronic states (Jahn–Teller effect,
pseudo-Jahn–Teller effect, and Renner–Teller effect).

Keywords Jahn–Teller effect · Pseudo-Jahn–Teller effect · Renner–Teller effect ·

Herzberg–Teller expansion · Crude adiabatic representation · Reduced mass ·
Moment of inertia · Angular momentum · Centrifugal deviation · Translational
mode · Rotational mode · Vibrational mode · Euler angle · Group theory ·
Generator · Representation · Character · Irreducible representation · Cartesian
coordinate · Symmetry-adapted coordinate · Bond angle · Dihedral angle ·
Z-matrix · Normal mode · Vibrational mode · Mass-weighted coordinate ·
Vibronic Hamiltonian · Linear vibronic coupling constant · Quadratic vibronic
coupling constant · Diabatic representation · Adiabatic representation ·
Wigner–Eckart theorem · Clebsch–Gordan coefficient

2.1 Introduction: Molecular Deformations and Its Driving

Force

In the previous chapter, we have glanced at some conventional explanations on the

origins of molecular deformations. This chapter devotes defining molecular defor-
mations and its driving force, vibronic coupling. Furthermore, their driving forces
are analyzed in terms of VCD concepts in the succeeding chapter.
In this book, we define a molecular deformation as a change of the internal coor-
dinates of a molecule from a certain reference geometry. We assume a motion of a
molecule as separated modes: translation, rotation, and vibration. Within the assump-
tion, an internal coordinate can be identified as a vibrational coordinate. The sepa-
ration is illustrated in Sect. 2.2 taking a diatomic molecule as an example.

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 13
T. Kato et al., Vibronic Coupling Density,
SpringerBriefs in Molecular Science,
https://doi.org/10.1007/978-981-16-1796-6_2
14 2 Molecular Deformations and Vibronic Couplings

When we discuss a molecular deformation, we assume some starting structure.

We call the structure a reference geometry. Such a structure may be an equilibrium
structure for an initial electronic state 0 or a hypothetical structure, the symmetry
of which is sometimes higher than the real one. A molecular deformation upon a
change of the electronic state is a real process and could be observable in the former
case. For example, the bond elongations of H2 and C2 H4 discussed in Sect. 1.1 are
a real process. However, the instability of NH3 discussed in Sect. 1.3 is not a real
but hypothetical process just for the explanation. The starting electronic state 0
is called a reference state. Note that the concept of the reference state includes its
structure as reference geometry.
In both cases, the driving forces can attribute to the vibronic coupling. A concept of
the vibronic coupling can be understood starting from the Herzberg–Teller expansion
and crude adiabatic representation. Sections 2.6 and 2.7 devote to the derivation of
the vibronic coupling and discussing its relation to molecular deformations.

2.2 Molecular Geometry and Modes of Motion

In this section, we discuss modes of a molecular motion taking a diatomic molecule,

AB as an example.
Let us consider a molecule consisting of two atoms A and B. The masses are m A
and m B , and the positions are denoted by (X A , Y A , Z A ) and (X B , Y B , Z B ), respec-
tively.
According to classical mechanics, the kinetic energy T is given by, in the space-
fixed Cartesian coordinate,

1   1  
T = m A Ẋ 2A + Ẏ A2 + Ż 2A + m B Ẋ 2B + Ẏ B2 + Ż 2B , (2.1)
2 2
where the dots denote the derivative with respect to time t. Using the coordinates of
the center of mass:
⎧ m A X A +m B X B
⎪ X=
⎨ m A +m B
m A Y A +m B Y B
Y = m A +m B , (2.2)
⎪
⎩ m A Z A +m B Z B
Z= m A +m B

and the relative coordinates:

⎧
⎨ x = XA − XB
y = Y A − YB , (2.3)
⎩
z = ZA − ZB

Equation (2.1) can be written as

2.2 Molecular Geometry and Modes of Motion 15

1  2  1  
T = M Ẋ + Ẏ 2 + Ż 2 + m ẋ 2 + ẏ 2 + ż 2 , (2.4)
2 2
where m Am B
M = m A + mB, m = , (2.5)
mA + mB

M is the total mass, and m is called a reduced mass. The first term of Eq. (2.4)
describes the motion of the center of mass, and the second term the relative motion.
Therefore, the motion of the center of mass is separated from the others.
Using the polar coordinate

x = r sin θ cos φ
y = r sin θ sin φ ,
z = r cos θ
(2.6)

the second term of Eq. (2.4) is written as

1  2  1 1  
T = M Ẋ + Ẏ 2 + Ż 2 + m ṙ 2 + m r 2 θ̇ 2 + r 2 φ̇ 2 sin2 θ . (2.7)
2 2 2
Using angular momentum
L = r 2 θ̇ , (2.8)

the kinetic energy of a diatomic molecule is obtained as

1  2  L2 1
T = M Ẋ + Ẏ 2 + Ż 2 + + m ṙ 2 . (2.9)
2 2mr 2 2
As we will see later, the first term corresponds to a translation, the second term
a rotation, and the third term a vibration, respectively. These are called modes of
motion.
When we regard the molecule as a particle without the internal structure, neglect-
ing (r, θ, φ), the kinetic energy can be written as

1  2  1  2  P2
T = M Ẋ + Ẏ 2 + Ż 2 = PX + PY2 + PZ2 = . (2.10)
2 2M 2M
This is the kinetic energy of a single particle with a mass M which is the total sum of
masses. When we consider the internal structure of the molecule, we can neglect the
time dependence of the interatomic distance among the internal coordinate (r, θ, φ).
Within this approximation, we obtain

P2 L2
T = + , (2.11)
2M 2I
16 2 Molecular Deformations and Vibronic Couplings

where I := mr02 is called moment of inertia. The second term, Eq. (2.11), is the
rotational energy of the rigid rotor. A moment of inertia in the rotational energy is
like the mass in the translational energy as angular momentum in rotation is like
momentum in translation:
P↔L
. (2.12)
M↔I

Equation (2.11) is the sum of the kinetic energies of the single particle and the rigid
rotor.
Furthermore, if we take the derivative of r with respect to time, ṙ , into consider-
ation, the kinetic energy can be written as

P2 L2 1 P2 L2 p2
T = + + m ṙ 2 = + + r , (2.13)
2M 2I 2 2M 2I 2m
where pr = m ṙ. The denominator in the second term of Eq. (2.9) is approximated by
the constant 2I . This approximation enables us to separate the rotation and the vibra-
tion of the kinetic energy which is described by the third term. Exactly speaking, I is
not a constant, and the violation of the approximation can be observed if the molecule
rotates, that is to say, the angular momentum is not equal to zero. The violation of
this approximation is called effect of centrifugal deviation. However, this approxi-
mation is valid in most cases. Therefore, we can discuss the translation (X, Y, Z ),
rotation (θ, φ), and vibration r separately. Hereafter, we assume this approximation
and concentrate ourselves on vibrations, or deformation of a molecule.
It is not easy to define a molecule. However, we can assume that a molecule
has some stationary molecular structure: there exists a certain interatomic distance
r = r0 such that
 
∂ E(r ) ∂ 2 E(r )
=0 and > 0, (2.14)
∂r r0 ∂r 2 r0

where E(r ) is the potential energy. This is because we can expect that interactions
among electrons and nuclei could yield such an attracting potential. The Taylor
expansion of E(r ) around r = r0 is expressed by, using Eq. (2.14),
 
∂ E(r ) 1 ∂ 2 E(r )
E(r ) = E 0 + (r − r0 ) + (r − r0 )2 + · · · (2.15)
∂r r0 2! ∂r 2 r0
1
= E 0 + k(r − r0 )2 + · · · ,
2
where E 0 = E(r0 ) and 
∂ 2 E(r )
k= (2.16)
∂r 2 r0

is called a force constant.

2.2 Molecular Geometry and Modes of Motion 17

Accordingly, we can obtain the following approximate molecular Hamiltonian:

2 2
P̂ L̂ p2 1
Ĥ = + + r + E 0 + k(r − r0 )2 = E 0 + Ĥt + Ĥr + Ĥv , (2.17)
2M 2I 2m 2
where
2
P̂
Ĥt = (2.18)
2M
is the Hamiltonian of the translational modes,
2
L̂
Ĥr = (2.19)
2I
is the Hamiltonian of the rotational modes, and

pr2 1
Ĥv = + k(r − r0 )2 (2.20)
2m 2
is the Hamiltonian of the vibrational mode.
We separate the translation and rotation modes from the molecular motion here-
after since internal molecular motions or deformation can be regarded as a vibration
as shown in Fig. 2.1. In the case of a diatomic molecule AB, the intramolecular struc-
ture is specified by the interatomic distance r . The vibrational Hamiltonian after the
separation is given by

1 2 1 1 2 1 2
Ĥv = P̂r + k(r − r0 )2 = P̂ + kq , (2.21)
2m 2 2m q 2
where q is the displacement or deformation

q = r − r0 . (2.22)

Note that a molecular deformation is described by vibrational coordinates.

Fig. 2.1 Freedom of a

diatomic molecule,
translation, rotation, and
vibration
18 2 Molecular Deformations and Vibronic Couplings

2.3 Coordinates of Non-linear A3

In this section, we discuss the internal coordinates of triatomic molecules, A3 . The

molecular structure can be linear or non-linear. When it comes to a molecular struc-
ture, we implicitly consider the molecule as a rigid body. It is important to distinguish
a linear or non-linear molecular structure since the degree of freedom is different
between them. The rotation of a linear molecule is described by (θ, φ) as the same as
a diatomic molecule, and that of a non-linear molecule is, for example, by Euler angle
(α, β, γ ). Therefore, the rotational degree of freedom is two and three for a linear and
non-linear molecules, respectively. Since the total degree of freedom is 3 × 3 = 9
and the translational degree of freedom is three, the remaining vibrational degree
of freedom is four and three, respectively. A molecular deformation is described by
four or three vibrational coordinates depending on its molecular structure as a rigid
body.
In this section, we concentrate ourselves on a non-linear A3 . The largest point
group of non-linear A3 is D3h . The Cartesian coordinates are denoted by

A1 (X 1 , Y1 , 0)
A2 (X 2 , Y2 , 0) , (2.23)
A3 (X 3 , Y3 , 0)

where the molecule is placed on the x y-plane. We will derive symmetry-adapted

vibrational coordinates of the D3h -A3 (Fig. 2.2).
Since the shape of the molecules is a triangle, we can adopt the three bond lengths
(see Fig. 2.3),

li = (X i+1 − X i )2 + (Yi+1 − Yi )2 , (i = 1, 2, 3) (2.24)

Fig. 2.2 Equilateral triangle

(C3v )

Fig. 2.3 Bond length as an

internal coordinate
2.3 Coordinates of Non-linear A3 19

as internal (vibrational) coordinates. The Cartesian displacements from the equilib-

rium position (X i0 , Yi0 , 0) are written as

xi = X i − X i0 , yi = Yi − Yi0 . (2.25)

A change in bond length li is given by

li = li − li0 . (2.26)

The bond length is, by expanding for xi , xi+1 , yi , yi+1 around the equilibrium posi-
tion,

li = (X i+1
0
+ xi+1 − X i0 − xi )2 + (Yi+1
0
+ yi+1 − Yi0 − yi )2 , (2.27)
1
≈ li0 + (X i+1
0
− X i0 )(xi+1 − xi ) + (Yi+1
0
− Yi0 )(yi+1 − yi ) , (2.28)
li0

where li0 is the equilibrium bond length. Therefore, the change of the bond length is
expressed by

1
li = li − li0 = (X i+1
0
− X i0 )(xi+1 − xi ) + (Yi+1
0
− Yi0 )(yi+1 − yi ) . (2.29)
li0

For D3h -A3 (see Fig. 2.2), all the bond lengths are equal: l10 = l20 = l30 = l0 , and
the Cartesian coordinates can be taken as
 0  0 
l l0 l l0 l0
A1 , − √ , A2 − , − √ , A3 0, √ . (2.30)
2 2 3 2 2 3 3

Accordingly, the changes of the bond lengths are obtained using (2.29) as

l1 = − (x2 − x1 ) (2.31)
√
1 3
l2 = (x3 − x2 ) + (y3 − y2 ) (2.32)
2 √2
1 3
l3 = (x1 − x3 ) − (y1 − y3 ) . (2.33)
2 2
These bond length changes describe bond stretching motions.
We consider the present problem under C3v which is a subgroup of D3h . The
rotation about the threefold axis C3 shown in Fig. 2.4 is written as
    √ 
 x cos 2π − sin 2π x − 1
− 23 x
x = Ĉ3 x = = 3 3 = √2 , (2.34)
y sin 3 cos 2π
2π
3
y 3
− 21 y
2
20 2 Molecular Deformations and Vibronic Couplings

Fig. 2.4 Symmetry

operations of C3v

where 
x
x = (x, y) = T
(2.35)
y

is the Cartesian coordinate of an arbitrary point on the 2D-plane. The reflection

against σ3 is written as
  
x −1 0 x
x  = σ̂3 x = = . (2.36)
y 0 1 y

The atomic position vector Ri = (X i , Yi )T of a deformed molecule with a deforma-

tion vector x i = (xi , yi )T is defined by

Ri = Ri0 + x i , (2.37)

where Ri0 = (X i0 , Yi0 )T is the atomic position vector of a molecule at the reference
geometry R0 . The atomic positions at the reference geometry are transformed as

Ĉ3 : A1 → A2 → A3 , (2.38)

and
σ̂3 : A1 ↔ A2 , A3  → A3 . (2.39)

We will employ group theory hereafter. See Ref. [1] for the definitions and theo-
rems. We can take σ̂3 and Ĉ3 as generators1 of the point group C3v . We can find by
inspection that the actions of these generators on li are as follows:

σ̂3 l1 = l1
σ̂3 l2 = l3 , (2.40)
σ̂3 l3 = l2

1 The generators generate group by multiplying themselves and their products.

2.3 Coordinates of Non-linear A3 21

and
Ĉ3 l1 = l3
Ĉ3 l2 = l1 . (2.41)
Ĉ3 l3 = l2

Therefore, the representations2 of these generators in ( l1 , l2 , l3 ) are obtained

as ⎛ ⎞ ⎛ ⎞
100 001
D (σ̂3 ) = ⎝ 0 0 1 ⎠ , D (Ĉ3 ) = ⎝ 1 0 0 ⎠ . (2.42)
010 010

Using (2.42), we can derive the other representation matrices

⎛⎞
010
D (Ĉ32 ) = D (Ĉ3 )D (Ĉ3 ) = ⎝ 0 0 1 ⎠ , (2.43)
100
⎛
⎞
001
D (σ̂1 ) = D (Ĉ32 σ3 ) = D (Ĉ32 )D (σ3 ) = ⎝ 0 1 0 ⎠ , (2.44)
100

and ⎞⎛
010
D (σ̂2 ) = D (Ĉ3 σ3 ) = D (Ĉ3 )D (σ3 ) = ⎝ 1 0 0 ⎠ . (2.45)
001

Using these matrices, we obtain the characters χ (R)3 of the representation matrices
as tabulated in Table 2.1 as well as the irreps4 of C3v .
By inspection from Table 2.1, we obtain the irreducible decomposition of :

= A1 + E. (2.46)

Using li , we can describe the totally-symmetric A1 deformation which is invariant

under the symmetry operations in D3h as follows:

1 1 1 1 1 1
q A1 = ( l1 + l2 + l3 ) = x1 − √ y1 − x2 − √ y2 + 0x3 + √ y3 .
3 2 2 3 2 2 3 3
(2.47)

2 They represent the symmetry operation as the matrices on the basis of a certain basis, li in the
present case.
3 Character is the trace of a representation matrix, χ (R) = trD (R).
4 irrep is an acronym of irreducible representation. See Ref. [1].
22 2 Molecular Deformations and Vibronic Couplings

Table 2.1 Characters of and the irreducible representations of C3v

Ê 2Ĉ3 3σ̂3
3 0 1
A1 1 1 1
E 2 −1 0

Fig. 2.5 Totally-symmetric deformation of the equilateral triangle described by q A1

Table 2.2 Irreducible representation matrices of C3v

C3v Ê Ĉ3 Ĉ32 σ̂1 σ̂2 σ̂3
A1 1 1 1 1 1 1
A2 1 1 1 −1 −1 −1
   √   √     √   √ 
10 − 21 − 23 − 21 3
1 0 − 21 − 23 − 21 3
E √ √ 2 √ √ 2
01 2
3
− 21 − 2 −2
3 1 0 −1 − 2
3 1
2 2
3 1
2

Under any deformation along q A1 , the molecule keeps its shape a equilateral
triangle as shown in Fig. 2.5. The point group of the deformed molecules is D3h . A
symmetry-adapted coordinate such as q A1 can be obtained using projection operators
[1]:
dim()  
P̂kl = D̄ (R) R̂, (2.48)
|G| R∈G kl

where |G| denotes the number of elements (order) in group G, dim() the dimension
of irrep , and D̄kl (R) the complex conjugate of (k, l) element of the irreducible
representation matrix of symmetry operation R in irrep  (Table 2.2).
For A1 representation, we obtain

1 
P̂ A1 = 1 · Ê + 1 · Ĉ3 + 1 · Ĉ32 + 1 · σ̂1 + 1 · σ̂2 + 1 · σ̂3 . (2.49)
6

When we take l1 and apply P̂ A1 , we can derive the expression of q A1 . There exist
doubly degenerate E coordinates, q Eθ and q E , in D3h -A3 . The projection operators
for q Eθ and q E are
2.3 Coordinates of Non-linear A3 23

Fig. 2.6 Degenerate deformations of the equilateral triangle described by a q E and b q Eθ


2 1 1 1 1
Eθ
P̂11 = 1 · Ê + (− ) · Ĉ3 + (− ) · Ĉ32 + ( ) · σ̂1 + ( ) · σ̂2 + (−1) · σ̂3
6 2 2 2 2
(2.50)
and

2 1 1 1 1
E
P̂22 = 1 · Ê + (− ) · Ĉ3 + (− ) · Ĉ32 + (− ) · σ̂1 + (− ) · σ̂2 + 1 · σ̂3 ,
6 2 2 2 2
(2.51)
Eθ E
respectively. We can obtain q Eθ and q E by applying P̂11 and P̂22 for l2 (note that
q Eθ does not contain l1 ):

1
E
P̂11 l2 = ( l2 − l3 ) =: q E , (2.52)
2
the three bond lengths become different. The point group of the deformed molecules
is C1 as shown in Fig. 2.6a. On the other hand, along

1
E
P̂22 l2 = (−2 l1 + l2 + l3 ) =: q Eθ . (2.53)
6
As shown in Fig. 2.6b, the molecule deforms into an isosceles triangle along q Eθ .
The point group of the deformed molecules is C2v .
24 2 Molecular Deformations and Vibronic Couplings

In summary, the symmetry-adapted vibrational coordinates are expressed by inter-

nal coordinates, l1 , l2 , l3 as

1
q A1 = ( l1 + l2 + l3 ) ,
3
1
q E = ( l2 − l3 ),
2
1
q Eθ = (−2 l1 + l2 + l3 ). (2.54)
6
A molecular deformation of D3h -A3 is described by these coordinates.

2.4 Coordinates of Non-linear A2 B

In the case of A3 , the vibrational coordinates can be obtained by their symmetry

because the representations in terms of ( l1 , l2 , l3 ) do not contain any equivalent
irreducible representation. The largest point group of non-linear A2 B is C2v (Fig. 2.7).
The symmetry operations of C2v are depicted in Fig. 2.8. The characters of the irreps
are tabulated in Table 5.5.
As discussed in Sect. 2.3, we obtain the representation matrices as

Fig. 2.7 Isosceles triangle

(C2v )

Fig. 2.8 Symmetry

operations of C2v
2.4 Coordinates of Non-linear A2 B 25

Table 2.3 Characters of A2 B and the irreducible representations of C2v

Ê Ĉ2 σ̂ (x z) σ̂ (yz)
A2 B 3 1 3 1
A1 1 1 1 1
A2 1 1 −1 −1
B1 1 −1 1 −1
B2 1 −1 −1 1

⎛ ⎞ ⎛ ⎞ ⎞ ⎛
100 100 100
D(Ĉ2 ) = ⎝ 0 0 1 ⎠ , D(σ̂ (yz)) = ⎝ 0 0 1 ⎠ ,
D(σ̂ (x z)) = ⎝ 0 1 0 ⎠ .
010 010 001
(2.55)
The characters in this representation, A2 B , are tabulated in Table 2.3.
By inspection, the representation can be decomposed into

A2 B = 2 A1 + B1 . (2.56)

Therefore, the symmetry-adapted coordinate for B1 can be obtained only from sym-
metry consideration. The projection operators are derived as

1
P̂ A1 = ( Ê + Ĉ2 + σ̂ (x z) + σ̂ (yz)),
4
1
P̂ B1 = ( Ê − Ĉ2 + σ̂ (x z) − σ̂ (yz)). (2.57)
4

Applying P̂ B1 for l2 − l3 , we can obtain

q B1 := P̂ B1 ( l2 − l3 ) = l2 − l3 . (2.58)

The symmetry-adapted coordinate q B1 is one of the normal coordinates of A2 B,

but the normal coordinates for representation A1 cannot be determined only from
symmetry consideration. Applying P̂ A1 for l1 + l2 + l3 and −2 l1 + l2 +
l3 , we can obtain the following two independent symmetry-adapted coordinates:

q A1 a := P̂ A1 ( l2 + l2 + l3 ) = l2 + l2 + l3 ,
q A1 b := P̂ (−2 l1 +
A1
l2 + l3 ) = −2 l1 + l2 + l3 . (2.59)

The normal coordinates for representation A1 can be obtained as a linear combination

of q A1 a and q A1 b from the vibrational analysis.
26 2 Molecular Deformations and Vibronic Couplings

2.5 Approximate Molecular Hamiltonian

For a polyatomic molecule, one way to specify the molecular structure is to use bond
angles and dihedral angles as well as bond lengths as internal coordinates. A bond
angle ϕi (see Fig. 2.9) is defined by
−
→
r a ·−→r b
ϕi = arccos , (2.60)
ra rb

and a dihedral angle χi is defined by the angle between the two planes which are
spanned by ra and rb , and rb and rc (see Fig. 2.10). The dihedral angle between two
planes is given by the angle between the vectors normal to these planes:
−
→
r a · (−
→
r b×− →
r c)
χi = arccos . (2.61)
ra rb rc

A set of bond lengths, bond angles, and dihedral angles to specify a molecular
structure is called Z-matrix. To derive a quantum-mechanical Hamiltonian using
such internal coordinates is difficult since the canonical quantization condition is
only valid for Cartesian coordinates. As the same in a bond length change li , the
changes of a bond angle ϕi and dihedral angle χi can be expressed as a linear
combination of Cartesian coordinates by expanding them around the equilibrium
position. We can obtain an approximate molecular Hamiltonian by expanding poten-
tial energy in terms of Cartesian displacement coordinates up to the second order.
The linear terms are vanishing for an equilibrium structure, and the quadratic terms
can be expressed as a standard form via vibration analysis. In the vibrational analysis
using Cartesian displacement coordinates, normal modes with finite eigenvalues cor-
respond to vibrational modes. Finally, collecting these normal modes, the vibrational
Hamiltonian is expressed by, within the harmonic approximation,

Fig. 2.9 Bond angle as an

internal coordinate

Fig. 2.10 Dihedral angle as

an internal coordinate
2.5 Approximate Molecular Hamiltonian 27

  2 ∂ 2 1

Ĥv = − + K α qα ,
2
(2.62)
α
2μα ∂qα2 2

where qα is called normal coordinate, and μα is the reduced mass of mode α. As dis-
cussed in Sect. 5.3, the Hamiltonian can be simplified by introducing mass-weighted
coordinate
√
Q α = μα qα , (2.63)

  2 ∂ 2 
as
1 2 2
Ĥv = − + ωα Q α , (2.64)
α
2 ∂ Q 2α 2

where
K α = μα ωα2 . (2.65)

We will denote a molecular deformation by a mass-weighted coordinate hereafter.

2.6 Energy as a Function of Molecular Deformation

We have defined in Sect. 2.1 the reference geometry as a stable structure of a certain
electronic state |0 or an assumed geometry such as planar NH3 and linear H2 O
discussed in Chap. 1. We call |0 a reference state. A molecular deformation from
a reference structure R0 = (X i0 , Yi0 , Z i0 ) can be expressed by a set of internal coor-
dinate, li , ϕi , and χi . This set transforms into the set of the normal coordinates,
Q α . We describe a molecular deformation in terms of Q α hereafter. Accordingly,
any deformed structure R = (X i , Yi , Z i ) is written as

R = R0 + R = R0 + R({ li , ϕi , χi }) = R0 + R({Q α }) = R0 + R( Q),

(2.66)
where the normal coordinate is defined for the reference state, |0 , at the reference
geometry, R0 . Therefore, the Hessian

∂ 2 E0
(2.67)
∂ Qα ∂ Qβ R0

is diagonal. Note that the diagonal elements can be positive and negative. As we have
discussed a molecular deformation, taking A3 and A2 B as examples in the previous
sections, a molecular deformation R can be expressed in terms of mass-weighted
normal coordinates as follows (see Sect. 5.3):

R( Q) = Q α uα . (2.68)
α
28 2 Molecular Deformations and Vibronic Couplings

The energy E(R) can be expanded around the reference geometry R0 ,

 
∂E 1 ∂2 E
E(R) = E(R ) + 0
· R+ R T
R + ··· (2.69)
∂ R R0 2 ∂ R∂ R R0
 ∂E 1   ∂2 E
= E(R0 ) + Qα + Q α Q β + · · · (2.70)
α
∂ Q α R0 2 α,β ∂ Q α ∂ Q β R0

The equilibrium point R0 is defined by the following equation:


∂ E ref (R)
= 0, (2.71)
∂ Qα R0

where E ref is the electronic energy for the reference electronic state |ref defined in
the next section.

2.7 Molecular Deformation Due to Vibronic Couplings

We start from the Herzberg–Teller expansion (see Sect. 5.4.1):

   
 ∂ Ĥ 1 ∂ 2 Ĥ
Ĥ (R) = T̂n ( Q) + Ĥe (R ) +
0
Qα + Qα Qβ + · · · .
α
∂ Qα 2 ∂ Qα ∂ Qβ
R0 α,β R0
(2.72)
The second term is called linear vibronic coupling, and the third term quadratic
vibronic coupling.
For vibronic wave function (r, R), we employ the crude adiabatic representa-
tion: 
(r, R) = n (r, R0 )χni (R), (2.73)
n

where r is a set of N -electron coordinates, n (r, R0 ) is an electronic state at a

reference geometry R0 , and χni (R) denotes nuclear wave function with vibrational
quantum number i for electronic state n . The electronic state satisfies

Ĥe (R0 )n (r, R0 ) = E n (R0 )n (r, R0 ). (2.74)

Using the basis set {n (r, R0 )}, Hamiltonian (2.72) is represented as
(Ĥ)mn = m (r, R0 )| Ĥ (R)|n (r, R0 ) (2.75)

 1
= T̂n (Q)δmn + E n (R0 )δmn + Vmn,α Q α + Wmn,αβ Q α Q β + · · · , (2.76)
α
2
α,β
2.7 Molecular Deformation Due to Vibronic Couplings 29

where  
∂ Ĥ
Vmn,α = m (r, R )| 0
|n (r, R0 ) (2.77)
∂ Qα
R0

and  
∂ 2 Ĥ
Wmn,αβ = m (r, R )| 0
|n (r, R0 ) . (2.78)
∂ Qα ∂ Qβ
R0

Hamiltonian matrix (2.76) is called vibronic Hamiltonian. Vmn,α is called a linear

vibronic coupling constant, which describes the electronic part of a linear vibronic
coupling. Wmn,αβ is called a quadratic vibronic coupling constant which describes
the electronic part of a quadratic vibronic coupling.
The first term of Hamiltonian matrix (2.76) is nuclear kinetic energy matrix, and
the remaining terms describe potential energy. When we neglect the nuclear kinetic
energy, we can obtain a static picture of a molecule. A representation that is diagonal
in nuclear kinetic energy is called a diabatic representation, and a representation that
is diagonal in potential energy is called an adiabatic representation. In this respect,
the vibronic Hamiltonian Ĥ is a diabatic representation.

2.7.1 Force Acting on Molecule: Linear Vibronic Couplings

In this subsection, we assume electronic state n is not degenerate. When we neglect

the off-diagonal elements of Vmn,α and Wmn,αβ (m = n), we obtain
 1
(Ĥ)nn = T̂n (Q) + E n (R0 ) + Vnn,α Q α + Wnn,αβ Q α Q β + · · · . (2.79)
α
2 α,β

This approximation is called crude adiabatic approximation. If the normal modes for
n are not so different from those for 0 , the cross terms in the third term of Eq.
(2.79) can be neglected:
 1
(Ĥ)nn = T̂n (Q) + E n (R0 ) + Vnn,α Q α +Wnn,αα Q 2α + · · · ,
α
2 α
  2 ∂ 2 1

= E n (R0 ) + − + Vnn,α Q α + Wnn,αα Q 2α + · · · . (2.80)
α
2 ∂ Qα 2

The second term is the sum of the Hamiltonian for a displaced oscillator with fre-
quency ω:
 2
2 ∂ 2 1 2 2 2 ∂ 2 1 2 |V | V2
− − |V |Q + ω Q = − + ω Q− 2 − . (2.81)
2 ∂ Q2 2 2 ∂ Q2 2 ω 2ω2
30 2 Molecular Deformations and Vibronic Couplings

The appearance of the linear term in Q, −|V |Q, gives rise to the displacement of
the minimum position from the origin by Q = |V |/ω2 , and the minimum energy is
lowered by E = V 2 /2ω2 . Here, since we can choose the direction of a vibrational
vector, we define the direction for the system to be stabilized for a positive Q hereafter.
The choice of the vibrational vectors results in negative VCCs, and thus we can write
Vα = −|Vα |. Equation (2.80) is rewritten as
  
 2 ∂ 2 1 2 |Vα | 2
Vα2
Ĥn = E n (R ) +
0
− + ω Qα − 2 − . (2.82)
α
2 ∂ Q 2α 2 α ωα 2ωα2

It should be noted that the |Vα | is equal to the force f α along Q α . Therefore, the
molecular deformation from R0 by (|Vα |/ωα2 ) occurs because of the linear vibronic
coupling constant Vα . The molecular deformation yields the stabilization energy, or
reorganization energy (Fig. 2.11)

 V2
α
E= 2
. (2.83)
α
2ωα

Fig. 2.11 Force acting on a

molecule: linear vibronic
couplings
2.7 Molecular Deformation Due to Vibronic Couplings 31

2.7.2 Spontaneous Symmetry Breaking of Non-linear

Molecule: Jahn–Teller Effect

For degenerate electronic states of a non-linear molecule A3 with energy E n , for

example,  Eθ and  E , Eq. (2.79) becomes

 2 ∂ 2
(Ĥ)mn = E n (R0 )δmn + − δmn + Vmn,α Q α
α
2 ∂ Q 2α
1
+ Wmn,αβ Q α Q β + · · · (m, n = θ, ). (2.84)
α,β
2

Since the symmetric product is

[E 2 ] = A1 + E, (2.85)

and non-zero linear vibronic couplings arise for E and A1 vibrational modes. Among
these modes, E modes break the molecular symmetry. When we consider the vibra-
tional coordinates (Q θ , Q  ), we obtain the following Jahn–Teller Hamiltonian:

 2 ∂ 2
(Ĥ)mn = E n (R0 )δmn + − δmn + Vmn,α Q α
α=θ,
2 ∂ Q 2α
 1 1
+ Wmn,E Q α Q β E + Wmn,A1 Q α Q β A1 + · · · (m, n = θ, ). (2.86)
α,β=θ,
2 2

We assume C3v symmetry. From the Wigner–Eckart theorem and Clebsch–Gordan

coefficients,  2
2 ∂ ∂2
Ĥ = − + σ 0 + Û(Q θ , Q  ), (2.87)
2 ∂ Qθ 2 ∂ Q 2

where the potential energy matrix Û(Q θ , Q  ) is defined by

1   
Û(Q θ , Q  ) = E n (R0 )σ 0 + VE (Q θ σ z + Q  σ x ) + W E Q 2θ − Q 2 σ x − (2Q θ Q  ) σ y
2
1  
+ W A1 Q 2θ + Q 2 σ 0 , (2.88)
2

and Pauli’s spin matrices are defined by

   
10 01 0 −i 1 0
σ0 = , σx = , σy = , σz = . (2.89)
01 10 i 0 0 −1
32 2 Molecular Deformations and Vibronic Couplings

Polar coordinates are introduced as


Q θ = ρ cos φ
. (2.90)
Q  = ρ sin φ

The diagonalization of the potential energy matrix (2.88) yields

1
E ± (ρ, φ) = E n (R0 ) + W A1 ρ 2 ± VE2 ρ 2 + VE W E ρ 3 cos 3φ + W E2 ρ 4 . (2.91)
2
Figure 2.12 shows the lower energy E − (ρ, φ). There exist three minimum positions
which are equivalent structures.

2.7.3 Spontaneous Symmetry Breaking of Non-linear

Molecule: Pseudo-Jahn–Teller Effect

In the previous subsections, we discuss the cases in which the electronic state(s) with
the energy E n is considered. We discuss the non-degenerate state n considering
other state m with a different energy E m (E m = E n , E n < E m ). For simplicity,
we take two energy levels, n = 1 and m = 2. State m can be degenerate; however,
we assume m also non-degenerate here. Since n and m are non-degenerate, the
irreducible representations n/m are one dimensional. The vibrational modes that
couple with n and m are non-degenerate with = n × m . We take a single
mode Q with here. The vibronic Hamiltonian (2.76) becomes

2 ∂ 2
ĤPJT = − σ 0 + ÛPJT (Q), (2.92)
2 ∂ Q2

where the potential energy matrix is defined by

 
E 1 (R0 ) 0 1 W11 0
ÛPJT (Q) = + V12 Qσ x + Q2. (2.93)
0 E 2 (R0 ) 2 0 W22

When the second term in Eq. (2.93) can be regarded as a perturbation, the lower
eigenenergy of Eq. (2.93) is, up to the second order,

1
E 1 (Q) = E 1 (R0 ) + WPJT Q 2 + · · · , (2.94)
2
where
|V12 |2
WPJT = W11 − . (2.95)
E 2 (R0 ) − E 1 (R0 )
2.7 Molecular Deformation Due to Vibronic Couplings 33

(b) (c)

(d) (e)

Fig. 2.12 Spontaneous symmetry breaking of a non-linear molecule: Jahn–Teller effect. a, b JT

potential energy surface for W E = 0. Note that the energy minima are infinitely degenerate (trough).
c–e JT potential energy surface for W E = 0. There appears three minima and three maxima in the
trough
34 2 Molecular Deformations and Vibronic Couplings

The first term in Eq. (2.95) is positive. If the off-diagonal vibronic coupling constant
V12 is large, or the energy difference between E 1 (R0 ) and E 2 (R0 ) is small, WPJT can
be negative. For such a case, the reference structure R0 is not stable. Therefore, if
2 is different from 1 , the irreducible representation of the vibrational mode is not
totally symmetric, leading to a deformed structure with lower symmetry.

2.7.4 Spontaneous Symmetry Breaking of Linear Molecule:

Renner–Teller Effect

In Sect. 2.7.2, we excluded linear molecules from the discussion. For a linear
molecule with degenerate electronic states, the linear vibronic coupling term is van-
ishing.
Here we assume degenerate electronic states of a molecule with D∞h symmetry.
The ground state of linear NH2 discussed in Sect. 1.2 is such an electronic state.
Since the lowest order of the non-vanishing vibronic couplings is quadratic ones (see
Sect. 5.2.3), instead of Eq. (2.86), vibronic Hamiltonian can be written as

 2 ∂ 2
(Ĥ)mn = E n (R0 )δmn + − δmn +
α=θ,
2 ∂ Q 2α
 1 1
+ Wmn,E {Q α Q β } E + Wmn,A1 {Q α Q β } A1 + · · · , (2.96)
α,β=θ,
2 2

Fig. 2.13 Spontaneous symmetry breaking of a non-linear molecule: pseudo-Jahn–Teller effect

2.7 Molecular Deformation Due to Vibronic Couplings 35

where Q α and Q β denote the coordinates of the E u mode, which are bending modes
shown in Fig. 5.1. The quadratic vibronic coupling terms in (2.96) give rise to lower-
ing the vibrational frequency of the degenerate mode or a bent molecular structure.
This effect is called the Renner–Teller effect (Fig. 2.13).

Reference

1. Ceulemans AJ (2013) Group theory applied to chemistry. Springer, Dordrecht

Chapter 3
Vibronic Coupling Density Analyses
for Molecular Deformation

Abstract The basis of the vibronic coupling density (VCD) analysis has been intro-
duced in Chap. 2. The vibronic coupling constant (VCC) as well as the VCD will
be exactly defined in Sect. 5.4.1. As seen in Sect. 5.4.1, both are starting from the
ab initio molecular Hamiltonian, and systematic, rational ways to understand chem-
ical phenomena, and which can give the quantitative evaluation of the force applied
under the chemical deformation process. We offer the approach of chemistry with
more rational and general way through the visualization by VCD and the evalua-
tion by VCC. Vibronic coupling is the interaction between vibrational and electronic
motions. The local picture of a vibronic coupling can be expressed in terms of elec-
tronic and vibrational structures using VCD. We describe the concepts of VCC and
VCD within the crude adiabatic approximation.

Keywords Vibronic coupling density · Vibronic coupling constant · Elongation ·

Orbital relaxation · Renner–Teller effect · Pseudo-Jahn–Teller effect · Orbital
vibronic coupling density · Orbital overlap · Potential derivative · Imaginary
mode · Quadratic vibronic coupling density · Jahn–Teller effect · Effective
vibronic coupling density · Regioselectivity · Structural relaxation

3.1 Elongation Force on Anionization of H2 and C2 H4

The physical meaning of the diagonal element of vibronic coupling is a force. A

change in the electronic state causes a force to act between nuclei and gives rise to a
geometry change. Therefore, we can understand the structural change by analyzing
the vibronic coupling density (VCD). As defined in Sect. 5.4.1, a linear vibronic
coupling constant (VCC), Vmn,α , is described as an integral of a VCD, ηmn,α (x)
[1–4], 
Vmn,α = ηmn,α (x)d x, (3.1)


ρm (x) × vα (x) (m = n),
ηmn,α (x) := (3.2)
ρmn (x) × vα (x) (m = n).
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 37
T. Kato et al., Vibronic Coupling Density,
SpringerBriefs in Molecular Science,
https://doi.org/10.1007/978-981-16-1796-6_3
38 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.1 Vibronic coupling density analysis for hydrogen molecule anion (ROHF/6-31G with first
derivatives). Top: vibrational mode, lower: vibronic coupling density η, electron-density difference
ρ × potential derivative v. The blue and gray surfaces denote negative and positive densities,
respectively. Reprinted by permission from Springer Nature, The Jahn–Teller Effect: Fundamentals
and Implications for Physics and Chemistry, Springer Series in Chemical Physics, vol. 97, ed. by
H. Köppel, D. R. Yarkony, H. Barentzen (Springer-Verlag, Berlin, 2009)

The concept of VCD enables us to visually and intuitively analyze the origin
of vibronic coupling in terms of ρm (x) or ρmn (x), which is an electron-density
difference or an overlap electron density obtained from electronic structures, and
vα (x), which is an electron potential energy derivation obtained from vibrational
structures. The VCD can be generally understood as the three-dimensional picture,
which describes the interaction between molecular vibration and electron density.
As an example, the vibronic coupling density analysis for a hydrogen molecule
anion is shown in Fig. 3.1. As mentioned in Sect. 1.1, when a neutral hydrogen
molecule acquires an electron, and the chemical bond will be elongated since the
additional electron occupies the anti-bonding LUMO. The vibronic coupling density
analysis reveals this driving force [3].
From Fig. 3.1, it is seen that most of the negative vibronic coupling density, the
blue surface occurs in the bond region, that is, the negative gradient of a potential
energy surface with respect to the vibration of nuclei moving apart. This is the driving
force of the chemical bond being elongated on anionization. The vibronic coupling
density is a product between the electron-density difference ρi and the potential
derivative vα , as defined in Sect. 5.4.1. More specifically, vα is the derivative of the
potential acting on a single electron from all nuclei for the vibrational coordinate of
mode α. In the case of the hydrogen molecule,
   
∂ 1 ∂ 1
vα (x) ∼ − + −
∂(−X HA ) |x − RHA | ∂ X HB |x − RHB |
x − X HA x − X HB
= − , (3.3)
|x − RHA |3 |x − RHB |3

where its molecular axis is taken as the x-axis. Thus, vα (x) has px -type distribution
( px ∼ x/r 3 ) around each nucleus because it is the linear derivative of the totally
symmetric potentials around the nuclei. Negative distribution appears in the positive
direction of the vibrational mode. As with the present case, in general, a potential
derivative always has p-type distribution in the corresponding vibrational direction.
In Fig. 3.1, it is noted that small negative ρi occurs in the bond region. Because the
additional electron distribution represented by the gray surfaces in Fig. 3.1 polarizes
3.1 Elongation Force on Anionization of H2 and C2 H4 39

Fig. 3.2 Top: Vibrational mode, lower: top view and side view of vibronic coupling density η3 ,
electron-density difference ρ × potential derivative v3 . The blue and gray surfaces denote negative
and positive densities, respectively. Reprinted by permission from Springer Nature, The Jahn–Teller
Effect: Fundamentals and Implications for Physics and Chemistry, Springer Series in Chemical
Physics, vol. 97, ed. by H. Köppel, D. R. Yarkony, H. Barentzen (Springer–Verlag, Berlin, 2009)

orbitals occupied by the other electrons. The negative ρi couples with the posi-
tive potential derivative vα .1 The orbital polarization due to anionization or orbital
relaxation plays a crucial role in vibronic coupling.
The importance of orbital relaxation can be observed in π electron systems.
Figure 3.2 shows the vibronic coupling density analysis for the ethylene anion. The
carbon–carbon stretching mode among 12 vibrational modes is focused in terms of
the vibronic coupling via the anionization process. Anionization of ethylene gives
rise to an elongation of the double bond. As shown in Fig. 3.2, even the additional
electron occupies the anti-bonding π LUMO, a negative vibronic coupling density
ηm,α occurs near the carbon atoms in the molecular plane. The negative ηm,α is
originated from the negative electron-density difference ρm due to the electron
polarization of additional π electron.
These simple examples clearly show that orbital relaxation is crucial in vibronic
coupling. Therefore, variationally optimized wave functions should be employed for
vibronic coupling calculations. The frozen orbital approximation is not suitable for
calculation [2].

3.2 Molecular Deformation of NH2 and H2 O by

Renner–Teller Effect and Pseudo-Jahn–Teller Effect

In the initial nuclear configuration R0 of the linear structure, the point group is D∞h ,
the normal modes shown in Fig. 3.3 can be defined. The energy of the molecule
including the vibrational deformation described by normal coordinates Q is given by
the eigenenergies of the molecular Hamiltonian,

1 The positive value of the potential derivative vα between two proton nuclei is due to the increase
of the electronic potential with separating the bonding electrons and two proton nuclei apart.
40 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.3 Vibrational normal

modes of H2 A(D∞h )

   
∂U 1 ∂ 2U
H = H0 + Q+ Q2 + · · · , (3.4)
∂Q 0 2 ∂ Q2 0

where H0 is the Hamiltonian at R0 , U is the interaction potential energy between

nuclei and electrons and among nuclei. Using the perturbation method, the lowest
eigenenergy E 0 is obtained as
 
   2  ∞ |0 | ∂U
∂ Q 0 |k | Q
2 2
∂U 1 ∂ U 
E = E 0 + 0 | |0 Q + 0 | | 0 Q 2
+ ,
∂Q 0 2 ∂ Q2 0 E0 − Ek
k=1
(3.5)
as mentioned in Sects. 2.7.3 and 2.7.4 . The third term causes the degenerate energy
split with
 2 the
 deformation, which is named the Renner–Teller effect [5]. The integral
∂ U
0 | ∂ Q 2 |0  is not zero, and this term is represented as the matrix form for the
0
degenerate 0 .
In the linear structure of NH2 (D∞h ), nine electrons of NH2 configured on lower
lying five MOs depicted in Fig. 3.4, resulting in the doubly degenerate u ground
state, and the third term in Eq. 3.5 becomes the two by two matrix,
⎛     ⎞
  01 | ∂ 2U
|01  01 | ∂ 2U
|02 
∂ U
2
⎜ ∂ Q2 ∂ Q2 ⎟
0 | |0  = ⎝  0  0 ⎠, (3.6)
∂ Q2 0 02 | ∂ U
2
|01  02 | ∂ U
2
|02 
∂ Q2 0 ∂ Q2 0

 
∂ 2U
among four matrix elements, non-zero integral i | ∂ Q2
| j  = 0, (i, j = 01, 02)
0
+ +
is obtained against the vibrational mode of u , because of u × u =
g and g +
 2 
+ ∂ U
u + g as shown in Table 5.9 and that the representation of ∂ Q 2 is identical with
0
that of Q 2 . However, the mode g+ or u+ of the deformation Q does not contribute
to breaking the axial symmetry of the molecule, as shown in Fig. 3.3. On the other
3.2 Molecular Deformation of NH2 and H2 O by Renner–Teller Effect … 41

hand, the u nuclear displacement demolishes  2 the

 axial symmetry.As a result,
∂ U ∂ 2U
the non-zero cross-terms of the matrix 01 | ∂ Q 2 |02  and 02 | ∂ Q 2 |01 
0 0
give rise to the energy split between the doubly degenerate u state concerning the
deformation out of axis u .
In the case of H2 O, the fourth term in Eq. 3.5 causes energy stabilization with
the deformation by the pseudo-Jahn–Teller effect [6]. The denominator E 0 –E k is
minus because
 E k is the upper energy level of the excited state. If the integral
∂U
0 | ∂ Q |k  is not zero. This term always gives a minus energy correspond-
0
ing to the stabilization with deformation. In the linear structure of D∞h , the elec-
tronic configuration of the H2 O ground state is shown in Fig. 3.4. The lowest excited
state corresponds to the excitation πu −→ 3σg . When the molecule is deformed
along with the normal mode of u , which corresponds
  to the bent deformation as
shown in Fig. 3.3, the value of the integral 0 | ∂∂UQ |k  will be non-zero because
0
πu × u × σg includes the totally symmetric representation of g (see Table 5.9).
The VCD analysis can visually understand this. Within the orbital approximation,
0 | ∂∂UQ |k  is given by the integral of the OVCD that is the product of the MO
0
overlap density ρHO LU between the πu (ξ ) HOMO and the σg LUMO with the poten-
tial derivative v u (ξ ) for the u (ξ ) mode (ξ = x or y). Note that the πu HOMOs
are doubly degenerate and distinguished by x and y. The u vibrational modes are
also such a case. Figure 3.5 shows that ρHO LU and v u have the same pattern of
distributions, (−, +, −) on the H–O–H linear structure, and hence their product,
i.e., the OVCD, has the same sign distribution over the molecule. Accordingly, the
integral of the OVCD has a large non-zero value, leading to the pseudo-Jahn–Teller

Fig. 3.4 Molecular orbitals

of water molecule in linear
structure of D∞h symmetry
42 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.5 Upper: imaginary u (x) mode, lower: orbital vibronic coupling density (OVCD) = orbital
overlap density ρHO LU × potential derivative v u (x) . The blue and gray surfaces denote negative
and positive densities, respectively

effect due to the

 mixing
 of the excited state with the ground one via the vibronic
coupling 0 | ∂∂UQ |k Q. Consequently, H2 O molecule is stabilized along the
0
bent deformation.

3.3 Vibronic Coupling Density of Ammonia Molecule NH3

The molecular structure of NH3 can also be well explained by the pseudo-Jahn–Teller
effect [6]. The vibrational modes of NH3 (D3h ) are given, as shown in Fig. 3.6

vib = A1 + A2 + 2E  . (3.7)

From the vibrational analysis by the calculation at the B3LYP/STO-3G level of

theory, we obtained an A2 mode with an imaginary frequency, 902.3i cm−1 .2 The
molecule would be deformed and stabilized along the direction of the imaginary A”2
mode. The pseudo-Jahn–Teller energy curve of the planar D3h NH3 model is given
by  
 2  |ψa2 | ∂∂UQ |ψa1 |2 Q 2
1 ∂ U
E = E 0 + 0 | |0 Q 2 + 0
, (3.8)
2 ∂ Q2 0 E a2 − E a1

as mentioned in Sect. 2.1, where 0 denotes the ground electronic wave function,
ψa2 is the HOMO, ψa1 is the LUMO, E a2 is the energy level of the HOMO, and E a1
is the energy level of the LUMO. The third term on the right-hand side corresponds
 
 ofa2 −→ a1 . When the orbital mixing is
to the lowest unoccupied orbital mixing
∂U
led via the A2 mode, the integral ψa2 | ∂Q
|ψa1  is non-zero because the product
0
of a2 × A2 × a1 includes the totally symmetric representation of A1 as referred


2 The imaginary mode means the vibrational mode having an imaginary frequency, ω κα
α = 1
2π c0 m
[cm−1 ].The harmonic oscillator with an imaginary ωα has a minus force constant κα , which
corresponds to not an oscillation on a concave surface but a motion on a convex one. The movement
on a convex surface is the rolling down along the coordinate of the mode.
3.3 Vibronic Coupling Density of Ammonia Molecule NH3 43

Fig. 3.6 Vibrational normal

modes of NH3 assuming the
D3h structure

in Table 5.10, and as a result, the third term gives the energy stabilization in the
direction of the A2 mode. However, the actual stabilization is determined by the
balance between the second and third terms. Taking the VCD analysis into account,
we can estimate the ratio of the magnitude between the second and third terms as
described below.
If the pseudo-Jahn–Teller effect is discussed by using vibronic coupling density
(VCD) scheme, it is much advantageous that we can see the effect in a visual-
ized manner [4]. The VCD scheme gives that the orbital vibronic coupling density
(OVCD) is localized on the nitrogen atoms, and the instability force originated from
the vibronic coupling of the orbital overlap between the lone pair of the nitrogen
atom via the umbrella bending mode.
The second term is given by the quadratic vibronic coupling density (QVCD), as
defined in Sect. 5.4.2,
ηA1 ,A2 A2 := ρA1 × wA2 A2 , (3.9)

and the third term is given by the OVCD, as described in Sect. 5.4.4.2.

ηHO LU,A2 := ρHO LU × vA2 . (3.10)

The QVCD is shown in Fig. 3.7. The electron density ρ is totally symmetric,
and mainly distributed on the nitrogen atom. However, since the quadratic potential
derivative ωa2 has small distribution on the nitrogen atom, and the QVCD is almost
equally distributed with blue- and gray-colored surfaces on the nitrogen and hydrogen
atoms. The integral of the QVCD on each atom becomes small with the cancellation
between blue (negative) and gray (positive) densities. On the other hand, the OVCD
shown in Fig. 3.8 is localized on the nitrogen atoms with the only blue-colored
surface, which means the pseudo-Jahn–Teller instability originates from the lone
pair of the nitrogen atoms.
44 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.7 Upper: imaginary A2 mode, lower: quadratic vibronic coupling density (ηA1 ,A2 A2 ) =
electron density ρ A1 × quadratic potential derivative w A2 A2 . The blue and gray surfaces denote
negative and positive densities, respectively. Reprinted from J. Phys.: Conf. Ser., vol. 428, T. Sato,
M. Uejima, N. Iwahara, N. Haruta, K. Shizu, and K. Tanaka, Vibronic coupling density and related
concepts, 012010 (2013)

Fig. 3.8 Upper: imaginary A2 mode, lower: orbital vibronic coupling density (ηHO LU,A2 ) = orbital
overlap density ρHO LU × potential derivative v A2 . The blue and gray surfaces denote negative and
positive densities, respectively. Reprinted from J. Phys.: Conf. Ser., vol. 428, T. Sato, M. Uejima,
N. Iwahara, N. Haruta, K. Shizu, and K. Tanaka, Vibronic coupling density and related concepts,
012010 (2013)

3.4 Vibronic Coupling Density Explanation for Zigzag

Conformation of 6-Cycloparaphenylene

Here we introduce the interpretation by the pseudo-Jahn–Teller effect (PJTE) to inter-

pret the tautomerism due to the ionization of 6-CPP molecule. So we visualize the
origin of the vibronic interaction using VCD for the deformation to the zigzag confor-
mation from belt-like one of 6-CPP. The MO calculation using the RB3LYP/STO-3G
level was done for optimizing the geometry of 6-CPP within the D6h symmetry of the
belt-like conformation [7]. The vibrational analysis gives the imaginary frequency of
the B1g normal mode shown in Fig. 3.9. The imaginary frequency indicates that the
D6h structure is unstable for the displacement along this B1g normal mode, and the
symmetry lowers into D3d with non-zero dihedral angles between the neighboring
benzene units.
The electronic ground state is calculated as shown in Fig. 3.10, in which the
HOMO orbital
 is π -orbital with a2g symmetry. As mentioned in Sect. 1.3,
{ψa2 | ∂∂UQ |ψa1 Q}2 /(E a2 − E a1 ) in Eq. 3.8 gives the energy stabilization with
0
the displacement along Q. When the displacement is along the B1g normal mode,
3.4 Vibronic Coupling Density Explanation for Zigzag … 45

Fig. 3.9 The B1g vibrational

normal mode of 6-CPP in
D6h symmetry with the
imaginary frequency.
Reprinted from Chem. Phys.
Lett., vol. 598, Y. Kameoka,
T. Sato, T. Koyama, K.
Tanaka, and T. Kato,
Pseudo-Jahn–Teller origin
of distortion in
6-cycloparaphenylene, 69
(2014), with permission
from Elsevier

Fig. 3.10 Molecular orbitals

obtained by the MO
calculation. The HOMO for
the neutral ground state of
6-CPP is a2g , and the next
HOMOs are e1u with a
degeneracy. The b2g -orbital
is an unoccupied level which
leads to the pseudo-JT effect.
Reprinted from Chem. Phys.
Lett., vol. 598, Y. Kameoka,
T. Sato, T. Koyama,
K. Tanaka, and T. Kato,
Pseudo-Jahn–Teller origin
of distortion in
6-cycloparaphenylene, 69
(2014), with permission
from Elsevier

the term of the pair with ψa2 = a2g (HOMO) and ψa1 = b2g (unoccupied orbital)
yields a maximum coupling because of the selection rule: a2g × B1g × b2g = a1g .
And this pair of orbitals make the largest contribution to the instability of the belt-like
structure, that is, the orbital pair including the HOMO plays an important role in the
structural change to the zigzag conformation.
The OVCD analysis of the pair of the HOMO and the b2g unoccupied orbital is
shown in Fig. 3.11. The symmetries and phase patterns in the pair of orbitals play an
crucial role in OVCC. The b2g unoccupied orbital is a σ -orbital, while the HOMO
is a π -orbital, as shown in Fig. 3.11, respectively. Therefore, the overlap density
ρa2g b2g between them has π -type distribution. The sign of ρa2g b2g is opposite on both
sides of the benzenes of 6-CPP. The function v B1g also has a π -type distribution
46 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.11 VCD of 6-CPP is obtained by the product of the overlap integral with the potential
derivative along the coordinate of the imaginary B1g vibrational normal mode. Reprinted from
Chem. Phys. Lett., vol. 598, Y. Kameoka, T. Sato, T. Koyama, K. Tanaka, and T. Kato, Pseudo-
Jahn–Teller origin of distortion in 6-cycloparaphenylene, 69 (2014), with permission from Elsevier

because of the displacement of the B1g mode. In addition, the signs of v B1g and
ρa2g b2g coincide almost everywhere. Accordingly, OVCD ηa2g b2g ,B1g has a positive
value almost everywhere. The orbital phases and symmetries of the orbital pairs give
rise to a large value of the OVCD with the same sign.
Since the vibronic coupling between the orbital pair, including the HOMO plays
the crucial role in the structural change to the zigzag conformation, the ionization
from the HOMO orbital gives rise to the increasing contribution of the quinoid form
in the cation and dication species of 6-CPP [8].

3.5 Vibronic Coupling Density and Jahn–Teller

Hamiltonian of C3 H3

We refer to the following equation again:

 
   2  ∞ |0 | ∂U
∂ Q 0 |k | Q
2 2
∂U 1 ∂ U 
E = E 0 + 0 | |0 Q + 0 | | 0 Q 2
+ .
∂Q 0 2 ∂ Q2 0 E0 − Ek
k=1
(3.11)

The second term can lead to the linear Jahn–Teller effect causing the insta-
bility. For example, let us look at C3 H3 with the high-symmetry geometry D3h .
We can find its ground electronic state of being doubly degenerate, E  . In such a
case, the second term contributes to the symmetry-lowering distortion in the ground
state via the JT effect. More specifically, the double degeneracy in the ground E 
state is subject to the JT problem E  × (a1 + e ). Because the ground E  state
∂U
gives rise to 0 ∂Q
|0  = 0 for the vibrational modes of A1 and E  , owing
0
to [E 2 ] = A1 + E (see Table 5.10). The vibrational modes of A1 and E  are called


the vibronically active modes among the 12 vibrational modes in Fig. 3.12. In par-
3.5 Vibronic Coupling Density and Jahn–Teller Hamiltonian of C3 H3 47

Fig. 3.12 Vibrational normal modes of C3 H3 (D3h )

ticular, the deformations along the E  modes split the doubly degenerate ground E 
state (JT effect), and hence the E  modes are called the JT active modes. The JT
effect shortens one of the C–C bonds to be ethylenic, resulting in C2v symmetry.
The VCD analysis gives us more detailed picture than the introductory qualitative
discussion. First, we take the non-degenerate C3 H+3 as the reference system with the
high-symmetry geometry of D3h . The geometry optimization and frequency analysis
for the reference system were performed at the RHF/STO-3G level of theory. For
the target system, i.e., C3 H3 , we employed the state-averaged CASSCF method3
including σ -orbitals as well as π -orbitals in the active space, and calculated the
VCCs from the energy gradients [9].
Table 3.1 summarizes the calculated VCCs for the vibronically active modes,
 
2 A1 + 3E , with the reorganization energies in meV. The reorganization energy E A
corresponds to the relaxation energy relative to the reference geometry along with
the A1 modes, as shown in Fig. 3.13, and E JT corresponds to the stabilization energy
due to the JT effect, as shown in Fig. 3.14. The VCCs are represented by Vα ’s in
Table 3.1 in the unit of 10−4 a.u. Among the JT active modes, the E  (2) mode has the
maximum coupling, whereas the A1 (2) mode is the strongest in the totally symmetric
modes. It was confirmed that the calculated VCCs satisfy the relation of the Clebsch–
Gordan coefficients. The negative value of Vα , −9.22, for θ and the positive one

3 CASSCF is the short form of the complete active space self-consistent field. It is one of the
post Hartree–Fock methods, i.e., a specific type of electronic state calculation method including
correlation effects, in which electronic excited configurations are allowed to be mixed into the
ground one by the configuration interaction approach to describe the ground and excited electronic
states at the higher level than the Hartree–Fock one.
48 3 Vibronic Coupling Density Analyses for Molecular Deformation

Table 3.1 Frequencies, ωα (cm−1 ), vibronic coupling constants of C3 H3 , Vα (10−4 a.u.), the
stabilization energy over all the A1 modes (E A ), and that over all the E  modes (E JT ) (meV) [9].
Reprinted from J. Phys.: Conf. Ser., vol. 428, T. Sato, M. Uejima, N. Iwahara, N. Haruta, K. Shizu,
and K. Tanaka, Vibronic coupling density and related concepts, 012010 (2013)
Mode ωα /cm−1 Vα /10−4 a.u.
θ 
A1 (1) 1913 −9.83 −9.83
A1 (2) 3758 −13.63 −13.63
E  (1)θ 1060 −2.83 2.83
E  (2)θ 1480 −9.22 9.22
E  (3)θ 3671 −4.60 4.60
E A /meV 259 259
E JT /meV 311 311

Fig. 3.13 Energy potential

surface of C3 H3 along the
coordinate A1 (2)

of Vα , 9.22, for  means the negative and positive slopes of the potentials at the
reference geometry, as shown in Fig. 3.14. On the other hand, both negative values
of Vα , −13.63, for θ and  represent the negative slopes, as shown in Fig. 3.13.
Figure 3.15 shows the e θ LUMO density ρLUθ of C3 H+ 3 and the electron-density
difference ρ between C3 H+ 3 and C3 H3 . Since the LUMO is a π -orbital, the pos-
itive LUMO density is distributed outside the molecular plane. On the other hand,
ρ is different from ρLUθ , and the negative contribution that originates from the
polarization of doubly occupied σ -orbitals appears in the molecular plane. The VCD
of the strongest mode A1 (2) is shown in Fig. 3.16 and that of the mode E  (2) with
the maximum JT coupling is shown in Fig. 3.17. Since the displacements of all the
vibronically active modes are in the molecular plane, and the σ polarization yields
large contributions to the VCDs. Figure 3.18 shows the AVCCs of carbon atoms and
3.5 Vibronic Coupling Density and Jahn–Teller Hamiltonian of C3 H3 49

Fig. 3.14 Energy potential

surface of C3 H3 along the
coordinate E  (2)θ

Fig. 3.15 The e θ LUMO

density ρLUθ of C3 H+ 3 and
ρ between C3 H+ 3 and
C3 H3 . The gray region is
positive, and the blue region
is negative. The isosurface
value is 0.01 a.u. [9].
Reprinted from J. Phys.:
Conf. Ser., vol. 428, T. Sato,
M. Uejima, N. Iwahara, N.
Haruta, K. Shizu, and K.
Tanaka, Vibronic coupling
density and related concepts,
012010 (2013)

the VCD for the E  (2) mode. The negative σ VCD gives rise to the large AVCC on
one of the carbon atoms. The VCC calculation with σ polarization allowed is indis-
pensable for evaluating the JT deformation from the equilateral triangle geometry.
From the obtained VCCs, we can estimate the mass-weighted displacement Q α
of the structural relaxation along mode α:

|Vα |
Qα = . (3.12)
ωα2

It is of importance to remark that Q α depends on ωα as well as Vα . It can be intuitively

understood as follows: a smaller ωα means a larger amplitude motion, yielding a
50 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.16 Vibronic coupling density η A1 (2) along A1 (2) mode = electron-density difference ρ
× potential derivative v A1 (2) . The isosurface value of η A1 (2) is 0.00005 a.u. [9]. Reprinted from
J. Phys.: Conf. Ser., vol. 428, T. Sato, M. Uejima, N. Iwahara, N. Haruta, K. Shizu, and K. Tanaka,
Vibronic coupling density and related concepts, 012010 (2013)

Fig. 3.17 Vibronic coupling density η E  (2) along E  (2) mode = electron-density difference ρ
× potential derivative v E  (2) . The isosurface value of η E  (2) is 0.00005 a.u. [9]. Reprinted from
J. Phys.: Conf. Ser., vol. 428, T. Sato, M. Uejima, N. Iwahara, N. Haruta, K. Shizu, and K. Tanaka,
Vibronic coupling density and related concepts, 012010 (2013)

Fig. 3.18 Atomic vibronic coupling constants (10−4 a.u.) and vibronic coupling density for the
E  (2) mode [9]. Reprinted from J. Phys.: Conf. Ser., vol. 428, T. Sato, M. Uejima, N. Iwahara,
N. Haruta, K. Shizu, and K. Tanaka, Vibronic coupling density and related concepts, 012010 (2013)
3.5 Vibronic Coupling Density and Jahn–Teller Hamiltonian of C3 H3 51

Table 3.2 Displacements of the structural relaxation of C3 H3 along all the active modes. Here, ωα
means a frequency, μα denotes a reduced mass, Vα stands for a vibronic coupling constant, and qα
is a displacement for mode α
Mode ωα /cm−1 μα /amu Vα /10−4 a.u. qα /Å
A1 (1) 1913 4.37 −9.83 0.077
A1 (2) 3758 1.18 −13.63 0.053
E  (1)θ 1060 1.09 −2.83 0.144
E  (2)θ 1480 5.61 −9.22 0.106
E  (3)θ 3671 1.11 −4.60 0.019

greater structural relaxation. Further, Q α can be converted to the real displacement

qα :
Qα
qα = √ , (3.13)
μα

where μα is the reduced mass of mode α. For all the active modes of C3 H3 , qα ’s
are listed in Table 3.2. The asymmetric deformational modes of the carbon triangle
framework, E  (1) and E  (2) exhibit large displacements in the order of 10−1 Å.

3.6 Vibronic Coupling Density Picture of Diels–Alder

Reaction

Molecular deformations play a crucial role in the course of a chemical reaction. In

this section, we take Diels–Alder reactions as an example.
Diels–Alder reactions can be well interpreted based on the frontier orbital theory
[10, 11], as mentioned for the reaction system of ethylene and butadiene in Sect. 1.6.
The VCD picture of the Diels–Alder reaction of butadiene and styrene is introduced
here (see Fig. 3.19).
Here, we define the effective vibronic coupling density (EVCD), as defined in
Sect. 5.4.4.4, to clarify the regioselectivity of the Diels–Alder reaction. ηs (x) is the
VCD of reaction mode s along which the chemical reaction occurs, and reaction
mode is taken as an effective mode that

Fig. 3.19 Diels–Alder reaction of butadiene and styrene

52 3 Vibronic Coupling Density Analyses for Molecular Deformation

 Vα
ds =  d Qα , (3.14)
α α Vα2

coincides with the steepest descent direction to the minimum of the potential energy
surface in the charge-transfer state. Here the Vα is VCC along the normal mode α,
then ds can be regarded as the normal coordinate deformation weighted with VCC.
ηs is represented as the product of electron-density difference due to charge transfer
and potential derivative in terms of s. The corresponding VCC is obtained from the
gradient of the potential energy surface in the charge-transfer state concerning s. It is
noted that ηs includes the effects of not only electronic but also vibrational structures
on chemical reactivity as discussed in Sect. 5.4.5.
The results calculated at B3LYP/STO-3G level of theory show that the reactive
site of the reaction is the Cα = Cβ double bond. The LUMO of styrene is delocalized
as shown in Fig. 3.20. The VCD analysis for the effective mode of styrene anion is
obtained as shown in Fig. 3.20. The EVCD distribution on Cα and Cβ is fit to that of
ethylene as shown in Fig. 3.20, which causes the reacting system to further stabilized
through a structural relaxation due to vibronic couplings.

Fig. 3.20 LUMO, electron-density difference, effective mode, and EVCD of styrene and ethylene
3.6 Vibronic Coupling Density Picture of Diels–Alder Reaction 53

Fig. 3.21 Averaged density of LUMOs

Fig. 3.22 Vibronic coupling

density analysis for the
reaction mode in C− 60 and
ethylene monoanion

The VCD as a reactivity index is applied to investigate the regioselectivity of the

cycloadditions of fullerene [12–15], metallofullerene [16], and polycyclic aromatic
hydrocarbons [17]. C60 fullerene undergoes nucleophilic cycloadditions at the 6,6-
bonds between two hexagonal rings rather than the 6,5-bonds between the hexagonal
and pentagonal rings [18, 19]. The frontier orbital theory has difficulty in predicting
this regioselectivity of the reactions because the frontier orbitals tend to be delo-
calized in such large systems, as shown in Fig. 3.21. Since the LUMO density is
delocalized in the entire molecule, the reactive sites in C60 cannot be identified by
the frontier orbital theory. In contrast, Fig. 3.22 shows the VCD of C− 60 concerning
the effective mode, which is strongly localized on the 6,6-bonds. Therefore, the VCD
elucidates the reactive sites of C60 . In addition, the VCD of C−
60 is analogous to that
of ethylene mono-anion, which implies that C60 undergoes nucleophilic reactions in
a similar manner to ethylene. The reactive sites for multiple cycloadditions to C60
are obtained in agreement with experiments by successively calculating the VCD of
C60 adducts [14, 15].
The importance of vibronic coupling in chemical reactions can also be cor-
roborated by comparing reaction profiles without and with structural relaxation
54 3 Vibronic Coupling Density Analyses for Molecular Deformation

Fig. 3.23 The Diels–Alder reaction of C60 + butadiene: a its scheme and its energy profiles b with-
out and c with structural relaxation at the B3LYP/6-311G(d,p) level of theory. Structural relaxation
was taken into account by optimizing the total system consisting of the isolated structures of the
reactants at each distance between the reaction centers

(Fig. 3.23). As shown in Fig. 3.23b, if structural relaxation is ignored in the Diels–
Alder reaction of C60 + butadiene, 6,6- and 6,5-additions have no reaction barriers
and make no difference in their potential curves. It means that its regioselectivity
cannot be described only by charge-transfer interaction (orbital interaction), i.e., the
frontier orbital theory. As shown in Fig. 3.23c, however, the consideration of struc-
tural relaxation makes reaction barriers in both the pathways and clearly illustrates
the lower barrier of the 6,6-addition, which is consistent with the experimental fact.
In such a case, vibronic coupling, which is the origin of structural relaxation, is indis-
pensable for describing the regioselectivity. It is the reason for the effectiveness of
the VCD analysis in the prediction of chemical reactions.

References

1. Sato T, Tokunaga K, Tanaka K (2006) J Chem Phys 124

2. Sato T, Tokunaga K, Tanaka K (2008) J Phys Chem A 112:758
3. Sato T, Tokunaga K, Iwahara N, Shizu K, Tanaka K (2009) The Jahn–teller effect: fundamentals
and implications for physics and chemistry. In: Köppel H, Yarkony DR, Barentzen H (eds).
Springer series in chemical physics, vol 97. Springer, Berlin
4. Sato T, Uejima M, Iwahara N, Haruta N, Shizu K, Tanaka K (2013) J Phys Conf Ser 428
5. Renner R (1934) Z Physik 92:172
6. Bersuker I, Gorinchoi N, Polinger V (1984) Theo Chim Acta 66:161
References 55

7. Kameoka Y, Sato T, Koyama T, Tanaka K, Kato T (2014) Chem Phys Lett 598:69
8. Kayahara E, Kouyama T, Kato T, Takaya H, Yasuda N, Yamago S (2013) Angew Chem Int Ed
52:13722
9. Sato T, Uejima M, Iwahara N, Haruta N, Shizu K, Tanaka K (2013) J Phys Conf Ser 428(1)
10. Fukui K (1971) Acc Chem Res 4:57
11. Ota W, Sato T, Tanaka K (2018) J Phys Conf Ser 1148
12. Sato T, Iwahara N, Haruta N, Tanaka K (2012) Chem Phys Lett 531:257
13. Haruta N, Sato T, Tanaka K (2012) J Org Chem 77:9702
14. Haruta N, Sato T, Iwahara N, Tanaka K (2013) J Phys: Conf Ser 428(1)
15. Haruta N, Sato T, Tanaka K (2014) Tetrahedron 70:3510
16. Haruta N, Sato T, Tanaka K (2015) J Org Chem 80:141
17. Haruta N, Sato T, Tanaka K (2015) Tetrahedron Lett 56:590
18. Yurovskaya M, Trushkov I (2002) Russ Chem Bull Int Ed 51:367
19. Thilgen C, Diederich F (2006) Chem Rev 106:5049
Chapter 4
Design for Functional Molecules
by Vibronic Coupling Density

Abstract The vibronic coupling constant (VCC) offers quantitative insight into
the molecular design. The diagonal VCC is the gradient of a potential energy sur-
face concerning the normal coordinate at a fixed nuclear configuration as defined in
Sect. 5.4.1, which can give the quantitative evaluation of the force which occurred
under an electronic transition. The elongation of bond on anionization of hydro-
gen and ethylene is interpreted by the value of vibronic coupling density (VCD),
as shown in Sect. 3.1. The amplitudes of Jahn–Teller deformation for C3 H3 are also
evaluated by the VCCs, which are analyzed based on the VCD concept. As well as
the molecular deformations, the VCD analyses explain carrier-transport efficiency
and the rate of internal conversions for some molecular systems. We give the further
outlooks for novel functional molecules. Through the VCD analyses, highly efficient
molecules for carrier transport and light emission are designed.

Keywords Carrier transporting · Light emission · Franck–Condon state · Internal

conversion · Transition dipole moment density · Organic light-emitting diode

4.1 Design for Carrier-Transporting Molecules

When a voltage is applied to an electrode–molecule–electrode system shown in the

upper picture of Fig. 4.1, carriers can be transported across the molecule interacting
with nuclear vibrations. As shown in the lower picture of Fig. 4.1, the molecule
is once ionized by the excitation with E voltage and neutralized with an electron
through the electrode. The carrier-transporting process across the molecule can be
regarded as the ionizations, or electronic excitation with the applied voltage, followed
by the vibrational relaxation. Since the vibronic coupling causes heat dissipation,
developing a molecule with a small vibronic coupling is important for efficient carrier
transport. Although both intramolecular and intermolecular vibronic coupling can
affect efficiency, the intramolecular vibronic coupling is only considered here.
Carbazole derivatives are known as having the high efficiency of hole-transporting
[1, 2] comparing with the similar structure molecules of biphenyl and fluorene [3].
Their structures are shown in Fig. 4.2, and their VCC and VCD are

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 57
T. Kato et al., Vibronic Coupling Density,
SpringerBriefs in Molecular Science,
https://doi.org/10.1007/978-981-16-1796-6_4
58 4 Design for Functional Molecules by Vibronic Coupling Density

Fig. 4.1 Upper: the hole-transporting process in a electrode–molecule–electrode system with the
voltage E voltage . Lower: the hole-transporting process across a molecule

Fig. 4.2 Molecular structures and the absolute values of vibronic coupling constants of a biphenyl,
b fluorene, and c carbazole cations. Reprinted from J. Phys.: Conf. Ser., vol. 1148, W. Ohta, T.Sato,
and K.Tanaka, Applications of Vibronic Coupling Density, 012004(2018)
4.1 Design for Carrier-Transporting Molecules 59

Fig. 4.3 C–C stretching vibrational modes are depicted on the first row for a biphenyl, b fluo-
rene, and c carbazole. The second row shows their derivatives of the electronic–nuclear potential
derivatives vα , the third row the electron-density differences due to ionization ρm , and the lowest
row are the vibronic coupling densities of the C–C stretching modes ηm,α . The values of ηm,α are
obtained as the product of vα and ρm along each column. White regions denote positive; blue
regions are negative. Reprinted from J. Phys.: Conf. Ser., vol. 1148, W. Ohta, T.Sato, and K.Tanaka,
Applications of Vibronic Coupling Density, 012004(2018)

evaluated to clarify why the carbazole is special. Figure 4.2 shows the diagonal
VCCs of biphenyl, fluorene, and carbazole cations concerning the strongest cou-
pling modes. In these molecules, the vibrational mode that gives the largest VCC is
the C–C stretching mode. The largest VCC values are calculated as −4.192 × 10−4
a.u. for biphenyl, −3.440 × 10−4 a.u. for fluorene, and −2.390 × 10−4 a.u. for car-
bazole.1 The VCC of carbazole is smaller than that of biphenyl and fluorene, which
indicates that intramolecular vibronic coupling is the weakest in carbazole. Figure 4.3
shows the C–C stretching modes and their potential derivatives vα . Biphenyl, flu-
orene, and carbazole exhibit similar distributions of vα with positive and negative

1 The a.u. of mass is m e (electron mass: 0.0005486 amu), a.u. of length a0 (Bohr radius: 0.5292
Å), and a.u. of energy E h (Hartree energy: 27.21 eV). VCC and OVCC given here are the mass-
−1/2 √
weighted constant in E h m e a0−1 , which is converted to the value of 2.195 × 103 × M eV / Å,
where M is the reduced mass of each vibrational mode in amu.
60 4 Design for Functional Molecules by Vibronic Coupling Density

Fig. 4.4 Molecular

structures of TPD and TPA

Fig. 4.5 Molecular structure

of HBCP

phases symmetrically around the C atoms. Figure 4.3 shows the electron-density dif-
ferences ρm due to cationization and the diagonal VCD ηm,α . ρm is extensively
distributed on benzene rings in biphenyl and fluorene, whereas the distribution on
benzene rings is relatively suppressed in carbazole. Also, ρm is localized on the N
atom in carbazole, where vα has small values on the N atom. As shown in Sect. 5.4.1,
the diagonal VCD ηm,α is given as the product of vα and ρm , refer Eq. 5.144. The
diagonal VCD ηm,α decreases for carbazole, resulting in a small VCC. It is the rea-
son why carbazole exhibits high hole-transporting efficiency. The same localization
of ρm on a N atom is also responsible for the weak vibronic coupling in N ,N  -
diphenyl-N ,N  -di(m-tolyl)benzidine (TPD) [4] and triphenylamine (TPA) [5], which
are shown in Fig. 4.4.
Based on the same consideration using the VCD analysis, hexaboracyclophane
(HBCP) [6] and hexaaza[16 ]parabiphenylophane (HAPBP) [7] have been proposed
as highly efficient carrier-transporting molecules with a weak vibronic coupling (see
Figs. 4.5 and 4.6).
4.2 Design for Light-Emitting Molecules 61

Fig. 4.6 Molecular structure

of HAPBP

4.2 Design for Light-Emitting Molecules

Theoretical design of highly efficient light-emitting molecules is of interest for appli-

cation to organic light-emitting diodes. One of the essential factors for efficiency is
the fluorescence quantum yield, determined by the radiative and non-radiative pro-
cesses’ ratio. Therefore, the molecule should be designed to enhance the radiative
process and the non-radiative process should be suppressed. The non-radiative pro-
cess includes vibrational relaxation, internal conversion, and intersystem crossing.
Here we discuss vibrational relaxation processes and the internal conversion and
give their relations with the vibronic coupling. The radiative process is also dis-
cussed because the transition dipole moment density (TDMD) can be considered in
a similar framework as the VCD described in Sect. 5.4.3. Furthermore, the radiative
rate constant depends on a Franck–Condon factor, calculated using diagonal VCCs.
Figure 4.7 describes the photoexcitation process followed by the vibrational relax-
ation and internal conversion. Assuming that nuclei’s motions are slow compared
with that of electrons, a vertical transition occurs from a ground S0 state to a Franck–
Condon state by the absorption of a photon. A molecule at the Franck–Condon state
transforms into an adiabatic state through vibrational relaxation. The molecule is
stabilized due to vibrational relaxation by dissipating the vibrational energy to the
surrounding environment. Within the crude adiabatic approximation, the potential
energy surface at the Sm state is represented as [8]
   
 ωm,α
2
|Vm,α | 2 2
Vm,α
E m (R) = E m (R ) +0
Qα − 2 − , (4.1)
α
2 ωm,α 2
2ωm,α

where ωm,α is the frequency of vibrational mode α. The minima of the energy surface
for each vibrational mode are stabilized by −Vm,α 2
/2ωm,α
2
with the shift of Vm,α /ωm,α
2

in the direction of Q α . The total stabilization energy due to the vibrational relaxation
is obtained by
62 4 Design for Functional Molecules by Vibronic Coupling Density

Fig. 4.7 The process of the

photoexcitation followed by
the vibrational relaxation and
the internal conversion

 Vm,α
2
E mstab = 2
. (4.2)
α
2ωm,α

Thus, vibrational relaxation can be reduced by decreasing the values of diagonal

VCC. Here the normal modes in the excited state are assumed to be the same as that
in the ground state.2
Since vibrational relaxation is a rapid process, radiative transition and internal
conversion mostly occur at the adiabatic state. Assuming that the internal conversion
occurs from |mv  = |χmv | m  to |nv  = |χnv | n , the rate constant of internal
conversion from Sm to Sn states is given by [8]

2π  
IC
km→n (T ) = Vmn,α Vnm,β Pmv (T )ρ(E mv )
 α,β v,v 
×χmv |Q α |χnv χnv |Q β |χmv , (4.3)

where Pmv (T ) is the Boltzmann distribution at |mv  with temperature T , and E mv

and E nv are the eigenvalues corresponding to |mv  and |nv , respectively. ρ(E)
stands for the density of states in the final state |nv . In Eq. (4.3), the term involving

2 The Duschinsky effect is neglected in Eq. (4.2). The stabilization energy due to the Duschinsky
effect is evaluated from the quadratic VCC [8]. Within the Born–Oppenheimer approximation, the
stabilization energy including the Duschinsky effect is calculated from the difference in the energies
of the Franck–Condon Sm state and Sm state at an equilibrium nuclear configuration.
4.2 Design for Light-Emitting Molecules 63

two vibrational modes is neglected. Thus, the internal conversion can be suppressed
IC
by decreasing the values of off-diagonal VCC. km→n also depends on the diagonal
VCC through χnv |Q α |χmv . The initial vibrational state |χmv  is expressed as a
product of the independent initial states |vα  with the number of phonons vα . Sim-
ilarly, the final vibrational state |χnv  is expressed as a product of the independent
final states |vα  with the number of phonons vα . As a result,

χnv |Q α |χmv  = vα |Q α |vα  
vn,β vm,β (4.4)
β=α

with [9]

min[vα ,vα ] v +v  −2l

vα !vα ! − 1 gm,α  α
gm,α α
vα −l l
vα ||vα 
2
= v +v  e
4 (−1) 2 , (4.5)
2 α α
l=0
l!(vα − l)!(vα − l)!

where gm,α = Vm,α / ωm,α 3 is the dimensionless VCC. Equation (4.3) is based on
the crude adiabatic approximation. The rate constant of internal conversion has been
derived in Refs. [10–12] based on the Born–Oppenheimer approximation.
The rate constant of radiative transition from Sm to Sn states is [8]
∞
4ω3   2
r
km→n (T ) = dω 3
Pmv (T )μ2mn χmv χnv ρ(E mv − ω), (4.6)
0 3c mn

where ω and c are the angular frequency and speed of photon, respectively. μmn is the
transition dipole moment between | m  and | n , which affects the enhancement of
radiative transition rate. The VCD is introduced as the integrand of VCC to control
its value based on the local picture. Similarly, the transition dipole moment density
(TDMD) is introduced as the transition dipole moment’s integrand [13]:

μmn = τmn (x) d 3 x. (4.7)

τmn is represented using the overlap density between Sm and Sn states as

τmn (x) = −exρmn (x), (4.8)

where e is the charge of an electron shown in Sect. 5.4.3.

A design principle for fluorescent anthracene derivatives, which are candidates for
highly efficient emitting molecules in an organic light-emitting diode (OLED), has
been proposed from the analyses of VCD and TDMD [14, 15]. Figure 4.8 shows the
molecules of anthracene (A), 9-chloroanthracene (CA), and 9,10-dichloroanthracene
(DCA), whose fluorescence quantum yields were experimentally evaluated as 0.33,
0.11, and 0.53, respectively [16]. First, the contribution of vibrational relaxation to the
64 4 Design for Functional Molecules by Vibronic Coupling Density

Fig. 4.8 The molecular structures of a anthracene (A), b 9-chloroanthracene (CA), and c 9,10-
dichloroanthracene (DCA)

Fig. 4.9 (a1–a3) Vibrational modes susceptible to chlorinations and (b1–b3) derivatives of
electronic–nuclear potentials with respect to their modes vα : (a1, b1) A (ω48 = 1427.55 cm−1 ),
(a2, b2) CA (ω48 = 1420.09 cm−1 ), and (a3, b3) DCA (ω49 = 1409.13 cm−1 ). White regions are
positive; blue regions are negative. Reprinted with permission from Ref. [14]

quantum yield is discussed [14]. Figure 4.9 shows the vibrational modes of A, CA,
and DCA susceptible to chlorinations and the potential derivatives vα concerning
the susceptible modes. The modes and vα exhibit similar behaviors regardless of
chlorinations. Figure 4.10 shows the electron-density differences ρ1 between S1
and S0 states, and the diagonal VCD η1,α . ρ1 are large at the C atoms bonded to
the Cl atoms in CA and DCA. As a result, η1,α , the product of vα and ρ1 , has
large positive values on these C atoms although η1,α localized on the edges of CA
and DCA are negative. Therefore, the diagonal VCCs obtained by the integration
of diagonal VCD are small in CA and DCA compared to A, which explains the
order of the calculated stabilization energies: 0.2332 eV for A, 0.2275 eV for CA,
and 0.2211 eV for DCA. Although the vibrational relaxation is suppressed in CA
and DCA, the quantum yield of CA is lower than that of A. It is attributed to the
difference in the internal conversion rate. The square sums of off-diagonal VCC,
4.2 Design for Light-Emitting Molecules 65

Fig. 4.10 (a1–a3) Electron-density differences between S1 and S0 states ρ1 and (b1–b3) diagonal
vibronic coupling densities in the Franck–Condon S1 state for normal modes susceptible to chlori-
nations η1,α : (a1, b1) A, (a2, b2) CA, and (a3,b3) DCA. Atomic vibronic coupling constants in 10−5
a.u. are shown in (b1–b3). Large AVCC values are shown in red and those reduced significantly by
chlorinations are shown in blue. White regions are positive; blue regions are negative. Reprinted
with permission from Ref.[14]

(a) (b) (c)

Fig. 4.11 Transition dipole moment densities between S1 and S0 states in z-direction τ10,z , for a A,
b CA, and c DCA. White regions are positive; blue regions are negative. Reprinted with permission
from Ref. [14]


α|V10,α |, are calculated as 1.30 × 10−4 a.u. for A, 1.69 × 10−4 a.u. for CA, and
1.24 × 10−4 a.u. for DCA. According to the selection rule, active vibrational modes
that give non-zero VCC are few in a molecule with high symmetry. Therefore, A
and DCA with D2h symmetry have a smaller number of active modes than CA
with C2v symmetry does, which results in the suppression of internal conversion
in A and DCA. Finally, the TDMD is shown in Fig. 4.11 to evaluate the radiative
transition rate. The TDMD is large on Cl atoms because the overlap density ρ10
is localized on the Cl atoms. Therefore, the transition dipole moment is increased
in CA and, even more, in DCA. Indeed, the oscillator strengths, proportional to
the square of the transition dipole moment, are large in the order of 0.0542 for
A, 0.0700 for CA, and 0.0899 for DCA. Since the TDMD is the product of ρ10
66 4 Design for Functional Molecules by Vibronic Coupling Density

Fig. 4.12 The molecular

structures of 5,11-
bis(phenylethynyl)benzo[1,2-
f:4,5-f ]diisoindole-
1,3,7,9(2H ,8H )-tetraone

and distance from the origin, the localization of ρ10 away from the origin leads
to the enhancement of radiative transition. Consequently, the design principle is
summarized as follows: long substituents on which the overlap density is localized
should be introduced, and the D2h symmetry should be maintained. Following these
design principles, an anthracene derivative, 5,11-bis(phenylethynyl)benzo[1,2-f:4,5-
f ]diisoindole-1,3,7,9(2H ,8H )-tetraone, was designed and synthesized [15], whose
structure is shown in Fig. 4.12. The observed fluorescence quantum yield of the
rationally designed molecule was 96%.
In addition to the anthracene derivatives, design principles for triphenylamine
derivatives were presented [17]. Furthermore, a novel electroluminescence mecha-
nism in OLEDs utilizing higher triplet states than the T1 state, called fluorescence
via higher triplets (FvHT) mechanism, has been proposed based on the analyses of
VCD and TDMD [18–20].

References

1. Koene BE, Loy DE, Thompson ME (1998) Chem Mater 10(8):2235

2. Zhang Q, Chen J, Cheng Y, Wang L, Ma D, Jing X, Wang F (2004) J Mater Chem 14(5):895
3. Shizu K, Sato T, Tanaka K, Kaji H (2010) Org Electro 11(7):1277
4. Sato T, Shizu K, Kuga T, Tanaka K, Kaji H (2008) Chem Phys Lett 458(1–3):152
5. Shizu K, Sato T, Tanaka K, Kaji H (2010) Chem Phys Lett 486(4–6):130
6. Shizu K, Sato T, Tanaka K, Kaji H (2010) Appl Phys Lett 97(14)
7. Shizu K, Sato T, Ito A, Tanaka K, Kaji H (2011) J Mater Chem 21(17):6375
8. Uejima M, Sato T, Yokoyama D, Tanaka K, Park JW (2014) Phys Chem Chem Phys
16(27):14244
References 67

9. Hutchisson E (1930) Phys Rev 36(3):410

10. Peng Q, Yi Y, Shuai Z, Shao J (2007) J Chem Phys 126(11)
11. Peng Q, Yi Y, Shuai Z, Shao J (2007) J Am Chem Soc 129(30):9333
12. Niu Y, Peng Q, Deng C, Gao X, Shuai Z (2010) J Phys Chem A 114(30):7817
13. Sato T, Uejima M, Iwahara N, Haruta N, Shizu K, Tanaka K (2013) J Phys Conf Ser 428(1)
14. Uejima M, Sato T, Tanaka K, Kaji H (2014) Chem Phys 430:47
15. Uejima M, Sato T, Detani M, Wakamiya A, Suzuki F, Suzuki H, Fukushima T, Tanaka K,
Murata Y, Adachi C et al (2014) Chem Phys Lett 602:80
16. Ateş S, Yildiz A (1983) J Chem Soc Faraday Trans 179(12):2853
17. Kameoka Y, Uebe M, Ito A, Sato T, Tanaka K (2014) Chem Phys Lett 615:44
18. Sato T, Uejima M, Tanaka K, Kaji H, Adachi C (2015) J Mater Chem C 3(4):870
19. Sato T (2015) J Comput Chem Jpn 14(6):189
20. Sato T, Hayashi R, Haruta N, Pu YJ (2017) Sci Rep 7(1):4820
Chapter 5
Definitions and Derivations

Abstract The vibronic coupling constant (VCC), as well as the vibronic coupling
density (VCD), will be exactly defined in this chapter. Before the precise definition of
VCC and VCD, we will introduce Dirac’s notation and the basic concept of group and
symmetry. The definition of the normal coordinates will be given, and the characters
and the direct products will be tabulated for the treatments of group and symmetry.

Keywords Vibronic coupling constant · Vibronic coupling density · Jahn-Teller

effect · Hellmann-Feynman theorem · Transition density · Transition dipole
moment density · Atomic vibronic coupling constant · Regional vibronic coupling
density · Orbital vibronic coupling constant · Orbital vibronic coupling density ·
Reduced vibronic coupling density · Effective vibronic coupling constant ·
Effective vibronic coupling density · Fukui function · Nuclear Fukui function ·
Electron-phonon coupling · Electron-lattice coupling · Rys-Huang factor

5.1 Dirac’s Notations

Here we briefly summarize Dirac’s notation , bra and ket.

The axioms of quantum mechanics are as follows:
1. (State) A quantum-mechanical state |ψ is a normalized vector in a Hilbert space
on the complex field C.
2. (Observable) An observable, or physical quantity, A is expressed by an Hermi-
tian operator Â.
3. (Measurement) Any observed value of A is one of the eigenvalues {an } of Â.

Â|an  = an |an . (5.1)

4. (Probability interpretation) A state |ψ is expanded by the eigenstates {|an } of

A: 
|ψ = |an cn , cn ∈ C. (5.2)
n

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 69
T. Kato et al., Vibronic Coupling Density,
SpringerBriefs in Molecular Science,
https://doi.org/10.1007/978-981-16-1796-6_5
70 5 Definitions and Derivations

When A is observed, the probability P(an ) obtaining an is

P(an ) = |cn |2 . (5.3)

5. (Schrödinger equation) The time dependence of a state |ψ is described by

∂
i |ψ = Ĥ |ψ, (5.4)
∂t

where Ĥ is Hamiltonian which is the operator corresponding to energy.

The vector |ψ is called a ket.
The inner product between any kets |ψ and |φ in the Hilbert space is written as
ψ|φ. For complex numbers, α, β ∈ C, therefore,
1. φ|ψ = ψ|φ∗
2. ϕ|{α|φ + β|φ} = αϕ|ψ + βϕ|φ
3. φ|φ ≥ 0 (if |φ = 0 (zero vector), then φ|φ = 0)
where ∗ denotes complex conjugate.
Considering an arbitrary |φ for a fixed |ψ, the inner product ψ|φ can be
regarded as a function, ψ|(|φ) acting on a ket |φ giving a complex number ψ|φ.
The function ψ| is called a bra.
For an observable A, we have a set of the eigenvectors {|an } with the eigenvalues
{an } of the corresponding Hermitian operator Â:

Â|an  = an |an  (5.5)

Since  is Hermitian, we can take {|an } as an orthonormal set:

am |an  = δmn . (5.6)

n and m can be a continuous variable, x and y. For such a case,

  
ax a y = δ(x − y), (5.7)

where δ(x) is Dirac’s delta function. Furthermore, we assume that any ket |  can
be expanded in terms of {|an }:

| = |an cn , (5.8)
n

where cn is a complex number. In other words, the basis set {|an } is assumed to be
complete. Applying am | from the left side,
 
am |  = am |an cn = δmn cn = cm . (5.9)
n n
5.1 Dirac’s Notations 71

Substituting (5.9) into Eq. (5.8), the ket |  can be rewritten as


| = |an an | . (5.10)
n

Since |  is arbitrary, we identify


|an an | = 1̂, (5.11)
n

where 1̂ is the unit operator. Equation (5.11) is called the completeness relation. For
a continuous case, (5.11) becomes

d x|ax ax | = 1̂. (5.12)

For B̂, using Eq. (5.11),

   
   
B̂ = 1̂ B̂ 1̂ = |am am | B̂ |an an | = |am am | B̂|an an | = |am Bmn an |,
m n m,n m,n
(5.13)
where
Bmn := am | B̂|an  (5.14)

is called a matrix representation of B̂ in terms of the eigenvectors of Â.

A commutator of  and B̂ is defined by

[ Â, B̂] := Â B̂ − B̂ Â. (5.15)

For operators  and B̂ which are commutable; [ Â, B̂] = 0, there exist simultaneous
eigenvectors such that

Â|an , bn  = an |an , bn , B̂|an , bn  = bn |an , bn . (5.16)

For example, for an electron in a hydrogen atom, Hamiltonian Ĥ , the square of the
2
orbital angular momentum L̂ , the z-component of the orbital angular momentum L̂ z ,
2
the square of the spin angular momentum Ŝ , and the z-component of the spin angular
momentum Ŝz are commutable. Therefore, there exist simultaneous eigenvectors

|n, l, m l , s, m s , (5.17)

where n is the principle quantum number.

72 5 Definitions and Derivations

For position operator r̂, the eigenequation is written as

r̂|r   = r  |r  . (5.18)

The completeness relation is 

dv r |rr| = 1̂, (5.19)

where dv r is the volume element in r-space. For an arbitrary ket | ,

 
|  = 1̂|  = dv r |rr|  = dv r |r (r), (5.20)

where the coefficient (r) is called the wave function.

Operators are quantized by introducing the canonical commutator relation:

[r̂i , p̂ j ] = iδi j , [r̂i , r̂ j ] = 0, [ p̂i , p̂ j ] = 0, (5.21)

where r̂i (i = 1, 2, 3) are Cartesian coordinates; x̂, ŷ, ẑ, describing the position of a
particle, and p̂i (i = 1, 2, 3) are the momenta conjugate to r̂i , p̂x , p̂ y , p̂z . It should
be noted that Eqs. (5.21) are valid only for Cartesian coordinates and their conjugate
momenta. The quantization by Eqs. (5.21) is called canonical quantization.
In the r-representation, the position operator r̂ is

r|r̂|r   = r  δ(r − r  ). (5.22)

Since a function of r̂, f (r̂) is

r| f (r̂)|r   = f (r  )δ(r − r  ), (5.23)

a potential energy operator consisting of the sum of one-electron operators V (r̂) is

r|V (r̂)|r   = V (r  )δ(r − r  ). (5.24)

Using Eq. (5.21), the r-representation of the momentum operator p̂ is expressed by

r| p̂|r   = −i∇ r  δ(r − r  ). (5.25)

Accordingly, the Hamiltonian operator Ĥ in the r-representation is diagonal:

r| Ĥ |r   = Ĥ (r, r  )δ(r − r  ). (5.26)

For the linear vibronic coupling,

5.1 Dirac’s Notations 73
   
∂ Ĥ  ∂ Ĥ
r| |r  = (r, r  )δ(r − r  ). (5.27)
∂ Qα ∂ Qα
R0 R0

In this book, we employ the following notation for a ket which depends on the
coordinates,
 
| n (r, R0 ) := d r |rr| n  = d r |r n (r, R0 ). (5.28)

In other words, we implicitly used r-representation in this book. For an operator Â,
 
 = dv r  dv r" |r  r  | Â|r"r"|. (5.29)

For example,  
∂ Ĥ
Vmn,α =  m (r, R )| 0
| n (r, R0 ) (5.30)
∂ Qα
R0

reads
    
∗  ∂ Ĥ

Vmn,α = dr m (r, R )r|
0
dv r  dv r  |r r |
∂ Qα
R0

|r  r  | d r  |r   n (r

, R0 )
  
∗ ∂ Ĥ 
= d r dv r  dv r  d r  m (r, R )δ(r − r )
0
∂ Qα
R0
(r  , r  )δ(r  − r  )δ(r  − r  ) n (r  , R0 )
  
∗ ∂ Ĥ
= dv r m (r, R ) 0
(r, r) n (r, R0 ). (5.31)
∂ Qα 0
R

5.2 Group and Symmetry

5.2.1 Symmetry Operation and Wave Function:

One-Dimensional Wall as an Example

Representation theory of group is useful in chemistry. Readers should learn the exact
definitions and proofs in Ref. [1]. In this section, we will give some examples to
show the way of applications of the concepts.
74 5 Definitions and Derivations

In quantum mechanics, it is required to evaluate a matrix element of an operator

Ô:  (x)| Ô| (x). This is the calculation of such an integral
  +∞  +∞  +∞
d x f (x) =
3
dx dy dz f (x, y, z), (5.32)
−∞ −∞ −∞

where f (x) = (x) Ô (x). We can say that it is zero without a detailed calculation
applying symmetry consideration.
For simplicity, we discuss a one-dimensional problem. Let us consider a real
function f (x) on R, which is a set of real numbers. If f (−x) = f (x) for any x ∈ R,
f (x) is called an even function. If f (−x) = − f (x) for any x ∈ R, f (x) is called
an odd function. For an odd function f (x),
 ∞
f (x)d x = 0. (5.33)
−∞

This signifies that the symmetry of a function can tell us if the integral is necessarily
zero since the independent variable replacement x → −x is the reflection which
is a symmetry operation. For an even function f (x) which is invariant under the
symmetry operation, the integral can be finite.
When we define a symmetry operator σ̂ acting on position x as

σ̂ : x → x  = −x, (5.34)

the action on an even/odd function f (x) is written as

f (x) (even function)

σ̂ f (x) = (σ̂ f )(x) := f (−x) = (5.35)
− f (x) (odd function)

This signifies that an even/odd function f (x) is the eigenfunction of σ̂ belonging to

the eigenvalue +1 or −1, respectively. For even/odd function f , = S/AS, this
can be rewritten as
(σ̂ f )(x) = D (σ̂ ) f (x), (5.36)

where
+1 ( = S)
D (σ̂ ) = . (5.37)
−1 ( = AS)

D (σ̂ ) is called a representation of σ̂ . For two functions, f 1

and g 2 , the represen-
tation of the product f 1 g 2 is obtained as

+1, ( = S, = S or = AS, 2 = AS)

D 1D 2
= 1 2 1
. (5.38)
−1, ( 1 = S, 2 = AS or 1 = AS, 2 = S)

As an example, for the one-dimensional wall, the Schrödinger equation is

5.2 Group and Symmetry 75

2 d 2
Ĥ ψn (x) = − + V (x) ψn (x) = E n ψn (x), (5.39)
2m d x 2

where
0, (−L/2 ≤ x ≤ L/2)
V (x) = . (5.40)
∞, (x < −L/2, L < x/2)

The wave functions and eigenenergies are

⎧
⎪
⎪ 2 nπ
(n : odd) (−L/2 ≤ x ≤ L/2)
⎨  L sin( L x),
ψn (x) = 2
cos( nπ x), (n : even) (−L/2 ≤ x ≤ L/2) (5.41)
⎪
⎪ L L
⎩
0, (x < −L/2, L < x/2)

and
π 2 2 2
En = n . (5.42)
2m L 2
Therefore, the eigenfunction ψn is either symmetric (S) (n: even) or anti-symmetric
(AS) (n: odd). For an electric dipole transition from state m to n, the transition
probability depends on the following integral:
 ∞
ψm∗ (x)xψn (x) (5.43)
−∞

Note that f (x) = x is transformed as AS under Ŝ. We can easily see, for the product
of three functions, ψm∗ (x) f (x)ψn (x), the integrand of the above integrand is trans-
formed as S if the representations of ψm∗ (x) and ψn (x) are transformed differently:

D m D f D n = +1, (5.44)

where D m , D f , and D n are the representations of ψm∗ (x), f (x), and ψn (x), respec-
tively. Therefore, if m is odd and n is even, or m is even and n is odd, the integrand
is S, we can expect an optical transition between them. This is one of the selection
rules.

5.2.2 Group and Its Representation: Two-Dimensional Wall

as an Example

In the previous section, the representation is a scalar. For more than a two-dimensional
case, a representation can become matrix, D.
Before we proceed with the two-dimensional case, some definitions discussed
in the previous section are extended for a three-dimensional case. For x ∈ R3 , a
76 5 Definitions and Derivations

symmetry operator R̂ such as rotation about n-fold symmetry axis Cn , reflection for
a plane σ , and inversion i,
⎛ ⎞ ⎛ ⎞
x x
R̂ : x = ⎝ y ⎠ → x  = ⎝ y  ⎠ = Rx, (5.45)
z z

where R is the matrix representation of R̂. In general, a symmetry operation R̂ acting

on a function f defined in R3 operates as follows:

R̂ : f (x) → f  (x) = ( R̂ f )(x) = f (R−1 x). (5.46)

We will discuss the two-dimensional wall as an example.

 
2 ∂2 ∂2
Ĥ ψn x ,n y (x, y) = − + 2 + V (x, y) ψn x ,n y (x, y) = E n x ,n y ψn x ,n y (x, y),
2m ∂x 2 ∂y
(5.47)
where

0, (−L x /2 ≤ x ≤ L x /2 and − L y /2 ≤ y ≤ L y /2)

V (x, y) = . (5.48)
∞, (otherwise)

The Hamiltonian is invariant under the following symmetry operations:

   
−1 0 1 0
σx = , σy = . (5.49)
01 0 −1

The wave functions inside the wall and eigenenergies are (Table 5.1)
⎧ 
nyπ
⎪
⎪
⎪
2
sin( nLx xπ x) 2
sin( y) = ψ AS,AS , (n x : odd, n y : odd)
⎪  Lx
⎪  Ly Ly
⎪
⎨ 2 n π
sin( nLx xπ x) 2
cos( Ly y y) = ψ AS,S , (n x : odd, n y : even)
ψn x ,n y (x, y) =  x
L Ly

⎪
⎪ 2
cos( nLx xπ x) 2 n π
sin( Ly y y) = ψ S,AS , (n x : even, n y : odd)
⎪
⎪  Lx
⎪
⎪  Ly
⎩ 2 cos( nLx xπ x) 2 n π
cos( Ly y y) = ψ S,S , (n x : even, n y : even)
Lx Ly
(5.50)
and
π 2 2 2 π 2 2 2
E n x ,n y = n x + n . (5.51)
2m L 2x 2m L 2y y

Further condition, L x = L y = L, induces a higher symmetry in the system. In

addition to the symmetry operations mentioned above, the Hamiltonian is invariant
under the following symmetry operation:
 
0 −1
C4 = . (5.52)
1 0
5.2 Group and Symmetry 77

Table 5.1 Transformation of the wave functions

Ê σ̂x σ̂ y
ψ S,S 1 1 1
ψ S,AS 1 1 –1
ψ AS,S 1 –1 1
ψ AS,AS 1 –1 –1

The solutions are

⎧ nyπ
⎪
⎪
2
sin( n xLπ x) sin( y) (n x : odd, n y : odd)
⎪
⎪
L L
⎪
⎪ nyπ
⎨ 2
L
sin( n xLπ x) cos( L
y) (n x : odd, n y : even)
ψn x ,n y (x, y) = (5.53)
⎪
⎪ 2
cos( n xLπ x) sin(
nyπ
y) (n x : even, n y : odd)
⎪
⎪
⎪
⎪
L L
⎩ nyπ
2
L
cos( n xLπ x) cos( L
y) (n x : even, n y : even)

and
π 2 2  2 
E n x ,n y = n + n 2y . (5.54)
2m L 2 x

For (n x , n y ) = (k, l) and (n x , n y ) = (l, k) with k = l, the energy level E n x ,n y is

doubly degenerate:
π 2 2  2 
E n x ,n y = 2
k + l2 . (5.55)
2m L
    
−1  01 x −y
C4 : x → x = = . (5.56)
−1 0 y x

For n x = 2m + 1, n y = 2l,
   
2 (2m + 1)π 2lπ
ψ2m+1,2l (x, y) = sin x cos y , (5.57)
L L L

and for n x = 2l, n y = 2m + 1

   
2 2lπ (2m + 1)π
ψ2l,2m+1 (x, y) = cos x sin y . (5.58)
L L L

The two functions transformed each other under Ĉ4 :

(Ĉ4 ψ2m+1,2l )(x, y) = ψ2m+1,2l (y, −x) = ψ2l,2m+1 (x, y), (5.59)
(Ĉ4 ψ2l,2m+1 )(x, y) = ψ2l,2m+1 (y, −x) = −ψ2m+1,2l (x, y). (5.60)
78 5 Definitions and Derivations

Therefore, the representation of Ĉ4 , D(Ĉ4 ), is expressed in terms of (ψ2m+1,2l ,

ψ2l,2m+1 ) by the following 2 × 2 matrix:

Ĉ4 (ψ2m+1,2l , ψ2l,2m+1 ) = (ψ2m+1,2l , ψ2l,2m+1 )D(Ĉ4 ), (5.61)

where  
0 −1
D E (Ĉ4 ) = . (5.62)
1 0

The three symmetry operations, Ĉ4 , σ̂x , and σ̂ y , generate the C4v point group. The
representation matrix for (ψ2m+1,2l , ψ2l,2m+1 ) is tabulated in Table 5.2. This repre-
sentation is called E representation.
Since there occurs a product of wave functions in a matrix element  m | Ô| n ,
the transformation property of the product m n is useful to evaluate it. We can
construct other basis set by making products between ψ2m+1,2l and ψ2l,2m+1 :

ψ2m+1,2l (x1 , y1 )ψ2m+1,2l (x2 , y2 ),

ψ2m+1,2l (x1 , y1 )ψ2l,2m+1 (x2 , y2 ),
ψ2l,2m+1 (x1 , y1 )ψ2m+1,2l (x2 , y2 ),
ψ2l,2m+1 (x1 , y1 )ψ2l,2m+1 (x2 , y2 ). (5.63)

This product representation E × E is obtained as follows:

⎛ ⎞
0 0 01
⎜ 0 0 −1 0 ⎟
D E×E (Ĉ4 ) = ⎜ ⎟
⎝ 0 −1 0 0 ⎠ , (5.64)
1 0 00
⎛ ⎞
1 0 00
⎜ 0 −1 0 0 ⎟
D E×E (σ̂x ) = ⎜ ⎟
⎝ 0 0 −1 0 ⎠ , (5.65)
0 0 01

and ⎛ ⎞
1 0 00
⎜ 0 −1 0 0 ⎟
D E×E (σ̂ y ) = ⎜ ⎟
⎝ 0 0 −1 0 ⎠ . (5.66)
0 0 01

Applying the following basis transformation:

 5.2 Group and Symmetry

Table 5.2 Irreducible representation matrices of C4v

R̂ Ê Ĉ4 , Ĉ43 Ĉ2 2σ̂v 2σ̂d basis
               
1 0 0 −1 0 −1 −1 0 −1 0 1 0 0 1 0 −1
D E ( R̂) , , , (ψ2m+1,2l , ψ2l,2m+1 )
0 1 1 0 1 0 0 −1 0 1 0 −1 1 0 −1 0
D A1 ( R̂) (1) (1) (1) (1) (1)
D A2 ( R̂) (1) (1) (1) (−1) (−1)
D B1 ( R̂) (1) (−1) (1) (1) (−1)
D B2 ( R̂) (1) (−1) (1) (−1) (1)
79
80 5 Definitions and Derivations

1
ψ A1 = √ ψ2m+1,2l (x1 , y1 )ψ2m+1,2l (x2 , y2 )
2
1
+ √ ψ2l,2m+1 (x1 , y1 )ψ2l,2m+1 (x2 , y2 ), (5.67)
2
1
ψ B1 = − √ ψ2m+1,2l (x1 , y1 )ψ2m+1,2l (x2 , y2 )
2
1
+ √ ψ2l,2m+1 (x1 , y1 )ψ2l,2m+1 (x2 , y2 ), (5.68)
2
1
ψ B2 = √ ψ2m+1,2l (x1 , y1 )ψ2l,2m+1 (x2 , y2 )
2
1
+ √ ψ2l,2m+1 (x1 , y1 )ψ2m+1,2l (x2 , y2 ), (5.69)
2
1
ψ A2 = √ ψ2m+1,2l (x1 , y1 )ψ2l,2m+1 (x2 , y2 )
2
1
− √ ψ2l,2m+1 (x1 , y1 )ψ2m+1,2l (x2 , y2 ), (5.70)
2

the representation matrices D( R̂) are block diagonalized:

⎛ ⎞ ⎛ ⎞
1 0 0 0 1 0 0 0
⎜0 −1 0 0⎟ ⎜ 0 1 0 0⎟
UD E×E (Ĉ4 )U−1 =⎜
⎝0
⎟ , UD E×E (Ĉ2 )U−1 = ⎜ ⎟ , (5.71)
0 −1 0⎠ ⎝0 0 1 0⎠
0 0 0 1 0 0 0 1
⎛ ⎞ ⎛ ⎞
1 0 0 0 1 0 0 0
⎜0 0 ⎟ ⎜ 0 ⎟
UD E×E (σ̂x )U−1 ⎜
=⎝
1 0 ⎟ , UD E×E (Ĉ4 σ̂ y )U−1 = ⎜ 0 −1 0 ⎟,
0 0 −1 0 ⎠ ⎝0 0 1 0 ⎠
0 0 0 −1 0 0 0 −1
(5.72)
where ⎛ ⎞
1 0 01
1 ⎜ −1 0 0 1⎟
U= √ ⎜ ⎟. (5.73)
2⎝ 0 1 1 0⎠
0 −1 10

These are summarized as follows:

⎛ ⎞
D A1 ( R̂) 0 0 0
⎜ 0 D B1 ( R̂) 0 0 ⎟
UD E×E ( R̂)U−1 =⎜
⎝ 0
⎟. (5.74)
0 D ( R̂)
B2
0 ⎠
0 0 0 D A2 ( R̂)
5.2 Group and Symmetry 81

Table 5.3 Characters of the irreducible representations of C4v

R̂ Ê 2Ĉ4 Ĉ2 2σ̂v 2σ̂d
A1 1 1 1 1 1 z x 2 + y2, z2
A2 1 1 1 −1 −1
B1 1 −1 1 1 −1 x 2 − y2
B1 1 −1 1 −1 1 xy
E 2 0 −2 0 0 (x, y) (x z, yz)

Table 5.4 Characters of D∞h

φ φ
D∞h E 2C∞ ··· ∞σv i 2S∞ ··· ∞C2
g+ (A1g ) 1 1 ··· 1 1 1 ··· 1 x2 +
y2 , z2
g− (A2g ) 1 1 ··· –1 1 1 ··· –1 Rz
g (E g ) 2 2 cos φ ··· 0 2 −2 cos φ · · · 0 (R x , R y ) (xz, yz)
u+ (A1u ) 1 1 ··· 1 –1 –1 ··· –1 z
u− (A2u ) 1 1 ··· –1 –1 –1 ··· 1
u (E u ) 2 2 cos φ ··· 0 –2 2 cos φ ··· 0 (x, y)

This result signifies that the product representation is reduced to the direct sum of
A1 , A2 , B1 , and B2 representations. These elementary representations A1 , A2 , B1 ,
B2 , and E are called irreducible representations (irrep) . In Eq. (5.70), if we make the
following transformation, x1 ↔ x2 , y1 ↔ y2 , ψ A1 , ψ B1 , and ψ B2 are invariant. The
representation in terms of such a representation is called a symmetric product repre-
sentation. On the other hand, since ψ A1 is anti-symmetric, its representation is called
an anti-symmetric product representation. In Tables 5.9, 5.10, and 5.11, symmetric
product representations are denoted by [ ], and anti-symmetric representations are
{ }.
In general, a wave function belongs to one of the irreps of the group under which
the Hamiltonian is invariant. To evaluate a matrix element,  m | Ô| n , we can derive
the condition for a non-zero integral:

m × n contains Ô , (5.75)

where m , n , and Ô are the irreps of m , n , and Ô, respectively. This is because
the product representation between itself, × , always contains a totally symmetric
representation A1 , and the integrand contains a non-vanishing component after the
integration. Accordingly, it is necessary to decompose the product representation
m × n into the direct sum of irreps. This can be easily done by using a character
χ ( ) defined by the trace of a representation matrix. The character table of C4v is
shown in Table 5.3. The characters of the other groups which appeared in this book
are tabulated in Tables 5.4, 5.5, 5.6, 5.7, and 5.8.
82 5 Definitions and Derivations

Table 5.5 Characters of C2v

C2v E C2 σv (x z) σv (yz)
A1 1 1 1 1 z x 2, y2 , z2
A2 1 1 –1 –1 Rz xy
B1 1 –1 1 –1 x, R y xz
B2 1 –1 –1 1 y, Rx yz

Table 5.6 Characters of C3v

C3v E 2C3 3σv
A1 1 1 1 z x 2 + y2, z2
A2 1 1 –1 Rz
E 2 –1 0 (x, y) (Rx , R y ) (x 2 − y 2 ,
2xy)(xz, yz)

Table 5.7 Characters of D3h

D3h E 2C3 3C2 σh 2S3 3σv
A1 1 1 1 1 1 1 x 2 + y2 , z2
A2 1 1 –1 1 1 –1 Rz
E’ 2 –1 0 2 –1 0 (x, y) (x 2 − y 2 , 2x y)
A1 1 1 1 –1 –1 –1
A2 1 1 –1 –1 –1 1 z
E  2 –1 0 –2 1 0 (R x , R y ) (xy, yz)

Table 5.8 Characters of D6h

D6h E 2C6 2C6 2 C2 3C2 3C2 i 2S3 2S6 σh 3σd 3σv
A1g 1 1 1 1 1 1 1 1 1 1 1 1 x 2 + y2 , z2
A1u 1 1 1 1 1 1 –1 –1 –1 –1 –1 –1
A2g 1 1 1 1 –1 –1 1 1 1 1 –1 –1 Rz
A2u 1 1 1 1 –1 –1 –1 –1 –1 –1 1 1 z
B1g 1 –1 1 –1 1 –1 1 –1 1 –1 1 –1
B1u 1 –1 1 –1 1 –1 –1 1 —-1 1 –1 1
B2g 1 –1 1 –1 –1 1 1 –1 1 –1 –1 1
B2u 1 –1 1 –1 –1 1 –1 1 –1 1 1 –1
E 1g 2 1 –1 –2 0 0 2 1 –1 –2 0 0 (Rx , R y ) (x y, yz)
E 1u 2 1 –1 –2 0 0 –2 –1 1 2 0 0 (x, y)
E 2g 2 –1 –1 2 0 0 2 –1 –1 2 0 0 (x 2 −
y 2 , x y)
E 2u 2 –1 –1 2 0 0 –2 1 1 –2 0 0
 5.2 Group and Symmetry

Table 5.9 Direct products of D∞h

D∞h A1g A2g Eg A1u A2u Eu
A1g A1g A2g Eg A1u A2u Eu
A2g A1g Eg A2u A1u Eu
Eg [A1g + E g ] + {A2g } E g Eg A1u + A2 + E u
A1u A1g A2g Eg
A2u A1g Eg
Eu [A1g + E g ] + {A2g }
83
 84

Table 5.10 Direct products of D3h

D3h A1 A2 E A1 A2 E 
A1 A1 A2 E A1 A2 E 
A2 A1 E A2 A1 E 
E [A1 + E  ] + {A2 } E  E A1 + A2 + E 
A1 A1 A2 E
A2 A1 E
E  [A1 + E  ] + {A2 }
5 Definitions and Derivations
Table 5.11 Direct products of D6h
D6h A1g A1u A2g A2u B1g B1u B2g B2u E 1g E 1u E 2g E 2u
A1g A1g A1u A2g A2u B1g B1u B2g B2u E 1g E 1u E 2g E 2u
5.2 Group and Symmetry

A1u A1g A2u A2g B1u B1g B2u B2g E 1u E 1g E 2u E 2g

A2g A1g A1u B2g B2u B1g B1u E 1g E 1u E 2g E 2u
A2u A1g B2u B2g B1u B1g E 1u E 1g E 2u E 2g
B1g A1g A1u A2g A2u E 2g E 2u E 1g E 1u
B1u A1g A2u A2g E 2u E 2g E 1u E 1g
B2g – – A1g A1u E 2g E 2u E 1g E 1u
B2u A1g E 2u E 2g E 1u E 1g
E 1g ([A1g + (A1u + A2u + (B1g + B2g + (B1u + B2u +
E 2g ] + E 2u ) E 1g ) E 1u )
{A2g })
E 1u ([A1g + E 2g ] + (B1u + B2u + (B1g + B2g +
{A2g }) E 1u ) E 1g )
E 2g ([A1g + E 2g ] + (A1u + A2u +
{A2g }) E 2u )
E 2u ([A1g + E 2g ] +
{A2g })
85
86 5 Definitions and Derivations

From Table 5.3, we can derive the characters of the product representation E ×
E and its decomposition into the direct sum of the irreps:

R̂ Ê 2Ĉ4 Ĉ2 2σ̂v 2σ̂d

E×E 2×2=4 0×0=0 (−2) × (−2) = 4 0 × 0 = 0 0×0=0
A1 + A2 +B1 + B4 1+1 + 1 + 1 = 4 1+1 − 1 − 1 = 0 1+1 + 1+1 = 4 1−1 + 1 − 1 = 0 1 − 1 − 1+1 = 0

Accordingly,
E × E = [A1 + B1 + B2 ] + {A2 }. (5.76)

The decompositions of the product representations in the other groups which

appeared in this book are summarized in Tables 5.9, 5.10, and 5.11. Since the z-
component of the electric dipole operator μ̂z = −ez belongs to A1 , the z-component
of the transition dipole moment between the states which belong to E irrep is non-
zero:
SOMO × SOMO = E × E = A1 + B1 + B2  A1 = z . (5.77)

Note that the anti-symmetric product is vanishing because x1 = x2 , y1 = y2 in the

integrand of Eq. (5.32).

5.2.3 Point Group and Its Irreducible Representation:

Triatomic Molecules as an Example

In this section, we discuss the use of the symmetry of triatomic molecules, NH2 , H2 O,
and A3 . Figure 1.2b shows the molecular orbitals of H2 A with the linear structures,
the point group of which is D∞h . The SOMO level of NH2 with the linear geometry
is doubly degenerate (E irrep). As shown in Fig. 1.2b, the SOMO is anti-symmetric
(u irrep) for the inversion. Therefore, from Table 5.4, the SOMOs belong to E u (u )
irrep, and are denoted by the πu -orbitals. The ground electronic state of the linear
NH2 is 2 u or 2 E u .
Figure 5.1 shows the normal modes of a linear H2 A. There are four modes, which
belong to A1g , A1u , and doubly degenerate E u irreps:

vib = A1g + A1u + E u . (5.78)

The selection rule of the linear vibronic coupling between electronic states m and
n is
m × v × n contains A1g , (5.79)

where m and n denote the irreps of the electronic states, and v is the irrep of the
normal mode which couples with the electronic states. From Table 5.9, the product
E u × E u is decomposed as
5.2 Group and Symmetry 87

Fig. 5.1 Vibrational normal

modes of H2 A(D∞h )

E u × E u = [A1g + E g ] + {A2g }, (5.80)

or
[E u2 ] = A1g + E g , {E u2 } = A2g . (5.81)

Thus, the active mode of the linear coupling is only the A1g mode. It should be noted
that, as shown in (5.78), there is no E g mode in the vibrational modes in H2 A, and the
JT distortion, which is due to the linear coupling, is impossible. On the other hand,
a quadratic coupling of mode E u is possible since the product of Q Eu s contains E g :
from Table 5.9,
E u × E u = [A1g + E g ] + {A2g }. (5.82)

Therefore, the quadratic vibronic coupling is not vanishing and gives rise to Renner–
Teller effect.

5.3 Molecular Vibrations and Normal Modes

Let us consider the nuclear motion of an N -atom molecule. Its displacement vector
R from the equilibrium nuclear configuration R0 is written by
⎛ ⎞
..
⎜ . ⎟
R := ⎜ ⎟
⎝ X Aξ ⎠ , (5.83)
..
.

where X Aξ denotes the ξ -component (ξ = x, y, z) of the displacement of an Ath

nucleus in the Cartesian coordinate system. Within the second-order approximation,
88 5 Definitions and Derivations

the nuclear Hamiltonian Ĥ is given by

 
  2 ∂ 2 1
Ĥ := − + RT KR, (5.84)
A ξ =x,y,z
2m A ∂ X 2Aξ 2

where m A stands for the mass of an Ath nucleus,  represents the reduced Planck’s
constant (Dirac constant), and K := (K Aξ,Bξ  ) is a Hessian, i.e., a real symmetric
matrix consisting of the second-order partial derivatives of the potential energy E(R),
 
∂ 2 E(R)
K Aξ,Bξ  := . (5.85)
∂ X Aξ ∂ X Bξ  R0

We subsequently convert the real coordinate X Ar to the mass-weighted coordi-

√
nate X Ar := m A X Aξ :
⎛ ⎞ ⎛ ⎞
.. ..
⎜ .  ⎟ ⎜√ . ⎟
R := ⎜ ⎟ ⎜ ⎟
⎝ X Aξ ⎠ = ⎝ m A X Aξ ⎠ . (5.86)
.. ..
. .

Accordingly, R is expressed as

R = M −1 R, (5.87)

where M is a real diagonal matrix whose element M Aξ,Bξ  is

1
M Aξ,Bξ  = √ δ AB δξ ξ  . (5.88)
mA

Ĥ can therefore be rewritten as

 
  2 ∂ 2 1
Ĥ = − + RT K R , (5.89)
A ξ =x,y,z
2 ∂X Aξ2 2

 
where K := (K Aξ,Bξ  ) is a mass-weighted Hessian whose element K Aξ,Bξ  is

 K Aξ,Bξ 
K Aξ,Bξ  := √ . (5.90)
m Am B

We here discuss the eigenvalue problem of the Hessian K :

K uα = ωα2 uα (α = 1, 2, . . . , 3N ). (5.91)
5.3 Molecular Vibrations and Normal Modes 89

Since K is a real symmetric matrix, ωα2 is a real number, and uα (α = 1, 2, . . . , 3N )

can be chosen such that they are orthonormal with each other. ωα ’s are called fre-
quencies, and uα ’s are normal vectors related to translation, rotation, and vibration
in the mass-weighted space.
Any mass-weighted displacement vector R can be represented as a linear com-
bination of uα (α = 1, 2, . . . , 3N ),

R = Q α uα , (5.92)
α

where Q α ’s are called mass-weighted normal coordinates. Consequently, Ĥ is rewrit-

ten by
 2 ∂ 2 1
Ĥ = − + ωα2 Q 2α , (5.93)
α
2 ∂ Qα2 2

because of the orthonormality, uTα uβ = δαβ .

From Eqs. 5.86 and 5.92, R is expressed as

R = MR = Q α M uα . (5.94)
α

We herein introduce qα and vα as

qα := |Muα |Q α , (5.95)

Muα
vα := , (5.96)
|Muα |

and thereby obtain the following expression:


R = q α vα , (5.97)
α

where vα ’s are called normal vectors in the real space, and qα ’s are called real normal
coordinates. It should be noted that vα ’s are not orthogonal in general. Its exception
is homonuclear molecules.
Finally, we define a reduced mass μα of mode α as

1
μα := . (5.98)
|Muα |2

Therefore, Ĥ can be rewritten as

 2 ∂ 2 1
Ĥ = − + μα ωα2 qα2 , (5.99)
α
2μα ∂qα2 2
90 5 Definitions and Derivations

which means the motion of 3N virtual particles with masses μα ’s. Additionally,
Eqs. 5.95 and 5.98 lead to the relation between qα and Q α :

Qα
qα = √ . (5.100)
μα

5.4 Vibronic Coupling Density

The vibronic coupling density (VCD) and the vibronic coupling constant (VCC) can
be explicitly defined by the ab initio molecular Hamiltonian. And they can give the
quantitative evaluation of the force applied under the chemical deformation process.
We give the definition and evaluation of VCD and VCC in this section.

5.4.1 Vibronic Coupling Constant and Vibronic Coupling

Density

In most quantum chemistry calculations, the Schrödinger equation in which the

kinetic energy operator of nuclei is neglected is solved assuming that the masses
of the nuclei are heavier than that of an electron. However, the motions of nuclei
sometimes play some important roles, as we describe in this monograph. An adiabatic
approximation is usually employed to solve the Schrödinger equation, including
the nuclear kinetic energy operator. As a result of the adiabatic approximation, a
molecular wave function, or a vibronic wave function, is represented as the product
of electronic and vibrational wave functions [2–4]. A vibronic wave function is
expanded using electronic wave functions for a certain nuclear configuration in the
crude adiabatic approximation. Thus, the dependencies of nuclear configurations
are described only by nuclear wave functions. On the other hand, in the Born–
Oppenheimer approximation, a vibronic wave function is expanded so that both
electronic and vibrational wave functions depend on nuclear configurations.
We have investigated vibronic coupling, the interaction between vibrational and
electronic motions, within the crude adiabatic approximation [5, 6]. Vibronic cou-
pling is quantified by a vibronic coupling constant (VCC). Vibronic coupling density
(VCD) is a density form of the VCC. Using the VCD, we can discuss the reason for
the small or large values of VCC based on the electronic and vibrational structures.
Moreover, VCD analysis enables us to control the VCC by a chemical modification.

A well-known phenomenon induced by vibronic coupling is the Jahn–Teller (JT)

effect, in which vibronic coupling lowers the geometrical symmetry of a non-linear
molecule with a degenerate electronic state to stabilize the system by a symmetry-
lowering deformation to lift the degeneracy of the electronic state [7–9]. The VCC
5.4 Vibronic Coupling Density 91

and VCD have been evaluated for C3 H3 in Section 3.5, and the VCC and VCD have
well-characterized the systems such as cyclopentadienyl [10, 11], benzene [12], and
fullerene [13, 14]. In these studies, the JT vibrational modes that strongly couple
to the degenerate electronic states are identified. Various phenomena are related
to vibronic coupling other than the JT effect. The applicability of VCD has been
extended to non-JT molecules of ammonia, cycloparaphenylene, as shown in Sects.
3.3 and 3.4, and naphthalene [15].
A molecular Hamiltonian Ĥ can be written as

Ĥ := T̂n + T̂e + Ûne + Ûee + Ûnn , (5.101)

where T̂n means the sum of nuclear kinetic energy operators, T̂e denotes the sum of
electronic kinetic energy operators, Ûne is the sum of nuclear–electronic attractive
potentials, Ûee stands for the sum of electronic–electronic repulsive potentials, and
Ûnn represents the sum of nuclear–nuclear repulsive potentials. In more detail, these
terms are defined as
 2 ∂2 ∂2 ∂2
T̂n := − + + , (5.102)
A
2m A ∂ X 2A ∂Y A2 ∂ Z 2A

 2 ∂2 ∂2 ∂2
T̂e := − + 2+ 2 , (5.103)
i
2m e ∂ xi
2
∂ yi ∂z i

 z A e2
Ûne := − , (5.104)
A i
4π 0 |r i − R A |

 e2
Ûee := , (5.105)
i=j
4π 0 |r i − r j |

 z A z B e2
Ûnn := , (5.106)
A =B
4π 0 |R A − R B |

where m A denotes the mass of an Ath nucleus;  represents the reduced Planck’s
constant (Dirac constant); R A is the position of an Ath nucleus, whose ξ -component
(ξ = x, y, z) is denoted by X A , Y A , and Z A ; m e stands for the mass of electron; r i
means the position of an ith electron, whose ξ -component (ξ = x, y, z) is denoted by
xi , yi , z i ; 0 is the vacuum permittivity; z A represents the charge of an Ath nucleus;
e is the elementary charge. An electronic Hamiltonian Ĥe including nuclear–nuclear
repulsive potentials is given by

Ĥe := T̂e + Ûne + Ûee + Ûnn , (5.107)

92 5 Definitions and Derivations

and thus Ĥ is rewritten as

Ĥ = T̂n + Ĥe . (5.108)

We here define m (r, R) as an mth eigenfunction of Ĥe and E m (R) as its eigenvalue,

Ĥe m (r, R) = E m (R) m (r, R), (5.109)

where r denotes a set of electronic positions, and R stands for a set of nuclear posi-
tions. Here, spin coordinates of electrons are abbreviated for simplicity. m (r, R) and
E m (R) are called an mth adiabatic electronic wave function and adiabatic potential
energy surface, respectively. A set of m (r, R) (m = 0, 1, . . .) is sometimes called
the Born–Oppenheimer basis.
Vibronic coupling is generally defined using the crude adiabatic basis m (r, R0 )
(m = 0, 1, . . .), whose nuclear configuration is fixed, instead of the Born–
Oppenheimer one m (r, R) (m = 0, 1, . . .). We can expand Ĥ around reference
geometry R0 in the powers of mass-weighted normal coordinates Q α (α = 1, 2, . . .)
as
   
 ∂ Ĥ 1 ∂ 2 Ĥ
Ĥ = Ĥ0 + Qα + Q α Q β + · · · , (5.110)
α
∂ Qα 0 2 α,β ∂ Q α ∂ Q β 0
R R

where Ĥ0 is Ĥ at geometry R0 , which corresponds to Q α = 0 (α = 1, 2, . . .). This

is the so-called Herzberg–Teller expansion. For an eigenfunction (r, R) of Ĥ , we
employ the crude adiabatic representation:

(r, R) = χm (R) m (r, R0 ). (5.111)
m

If we ignore the translations and rotations of a molecule, then (r, R) is a vibronic

wave function. The Hamiltonian matrix Ĥ of Ĥ represented by the crude adiabatic
basis m (r, R0 ) (m = 0, 1, . . .) is given by
⎛ ⎞
 0 (r, R0 )| Ĥ | 0 (r, R0 )  0 (r, R0 )| Ĥ | 1 (r, R0 ) · · ·
⎜ 1 (r, R )| Ĥ |
0
0 (r, R ) 
0
1 (r, R )| Ĥ |
0 ⎟
1 (r, R ) · · · ⎠ .
0
Ĥ = ⎝ (5.112)
.. .. ..
. . .

The matrix element (Ĥ)mn is written as

 1
(Ĥ)mn = T̂n δmn + E m (R0 )δmn + Vmn,α Q α + Wmn,αβ Q α Q β + · · · ,
α
2 α,β
(5.113)
where a linear vibronic coupling constant (VCC), Vmn,α , and quadratic VCC (QVCC),
Wmn,αβ , between mth and nth states are introduced as
5.4 Vibronic Coupling Density 93
 
∂ Ĥ
Vmn,α :=  m (r, R )| 0
| n (r, R0 ), (5.114)
∂ Qα
R0
 
∂ 2 Ĥ
Wmn,αβ :=  m (r, R )|0
| n (r, R0 ). (5.115)
∂ Qα ∂ Qβ 0
R

Such a Hamiltonian matrix is called vibronic Hamiltonian. For simplicity, we here-

after represent diagonal elements of VCCs, Vmm,α and Wmm,αβ , as Vm,α and Wm,αβ ,
respectively.
We can obtain the adiabatic potential energy surface E m (R) from the diagonal-
ization of the potential part after the second term in Eqs. 5.112 and 5.113. With the
help of the perturbation theory, E m (R) is given by
⎡ ⎤
 1 ⎣ m  Vmn,α Vnm,β
E m (R) = E m (R0 ) + Vm,α Q α + Wαβ − 2 ⎦ Qα Qβ + · · · .
α
2
α,β
E (R0 ) − E m (R0 )
n =m n
(5.116)
In particular, if only the single vibrational mode α is taken into account, then E m (R)
is written as
⎡ ⎤
1  2
Vmn,α
E m (R) = E m (R0 ) + Vm,α Q α + ⎣Wm,αα − 2 ⎦ Q 2α + · · · .
2 E
n =m n
(R 0 ) − E m (R 0)

(5.117)
Within the orbital approximation, the orbital mixing (pseudo-Jahn–Teller) term is
given by
 Vmn,α Vmn,β   Vab,α Vab,β
− Qα Qβ = − Qα Qβ , (5.118)
n =m
E n (R ) − E m (R )
0 0
a∈occ b∈vir
b − a

where a denotes the ath-orbital energy at R0 and Vab,α stands for the orbital vibronic
coupling constant between ath and bth molecular orbitals for mode α defined later.
Similar to E m (R), the adiabatic electronic wave function m (r, R) is given by
 Vnm,α
m (r, R) = m (r, R0 ) − n (r, R0 ) Qα + · · · .
α n =m
E n (R ) − E m (R0 ) 0

(5.119)
As described above, an electronic wave function and a potential energy surface
change as a function of vibrational coordinates, Q α (α = 1, 2, . . .), owing to the
mixing of electronic states. This is the concept of vibronic coupling, which is the inter-
action between the motion of electrons and that of nuclei. The strength of vibronic
coupling is estimated by Vmn,α , Wmn,αβ , and higher-order ones.
94 5 Definitions and Derivations

Vm,α , a diagonal element of vibronic coupling, is the origin of structural/vibrational

relaxation [5, 6]. The following relation can be obtained based on the Hellmann–
Feynman theorem [16, 17]:
 
∂ E m (R)
Vm,α = . (5.120)
∂ Qα R0

Equation 5.120 shows that Vm,α is a force acting on nuclei along mode α in state
m (r, R ). If the Hessian is diagonalized with the vibrational analysis as
0

 Vmn,α Vnm,β ωm,α

2
(α = β),
Wm,αβ − 2 = (5.121)
E n (R0 ) − E m (R0 ) 0 (α = β),
n =m

then E m (R) is given by

 1 2
E m (R) ≈ E m (R0 ) + Vm,α Q α +
ωm,α Q 2α
α
2 α
  2 
 1 |Vm,α | V 2
m,α
= E m (R0 ) + ω2 Qα − 2 − , (5.122)
α
2 m,α ωm,α 2
2ωm,α

where we choose the direction of a vibrational vector such that the system is sta-
bilized for a positive Q α . The choice of the vibrational vectors results in negative
VCCs, i.e., Vα = −|Vα |. If the value of Vm,α is not zero, molecular deformation
spontaneously occurs with the displacement of Q α = |Vm,α |/ωm,α 2
to yield the sta-
bilization/reorganization energy E m,α [5, 6],
stab

2
Vm,α
E m,α
stab
:= 2
. (5.123)
2ωm,α

The summation of E m,α

stab
over all the normal modes gives the total stabilization
energy E m , defined as follows:
stab


E mstab := E m,α
stab
. (5.124)
α

As described above, diagonal vibronic couplings give rise to structural relaxation in

an electronic state. Figure 5.2 schematically illustrates the structural relaxation.
In contrast to diagonal vibronic couplings, off-diagonal ones give rise to internal
conversion, i.e., non-radiative transition, between different electronic states [6, 18].
Within the crude adiabatic approximation, an eigenfunction mv of the molecular
Hamiltonian Ĥ is written as follows:

mv (r, R) = χmv (R) m (r, R0 ), (5.125)

5.4 Vibronic Coupling Density 95

Fig. 5.2 A conceptual

presentation of structural
relaxation in an mth
electronic state

where R0 is the equilibrium geometry in the state m (r, R). If we ignore the transla-
tions and rotations, then χmv (R) is the vth vibrational wave function, and mv (r, R)
is a vibronic wave function. According to Fermi’s golden rule, which describes a
transition rate between different quantum states in a general manner, when perturba-
tion Ĥ  is applied, the transition rate kmv,nv from mv (r, R) to nv (r, R) is given
by
 2
2π  
kmv,nv = ρ(E)  mv (r, R)| Ĥ  | nv (r, R) , (5.126)

where E is E mv − ω for the radiative transition, in which E mv denotes the eigenvalue
of Ĥ corresponding to mv (r, R) and ω is the frequency of the light, and E is E mv
for the internal conversion. ρ(E) stands for the density of states in the final state
nv (r, R). Therefore, the internal-conversion rate km→n (T ) from an mth electronic
IC

state to an nth electronic state at the absolute temperature T is given by


IC
km→n (T ) = Pmv (T )kmv,nv
v,v
2π 
= Pmv (T )ρ(E mv )
 v,v
 2
 
×  mv (r, R)| Ĥ  | nv (r, R) , (5.127)

where Pmv (T ) denotes a statistical weight of the initial state mv (r, R) owing to
Boltzmann’s distribution and Ĥ  stands for a perturbation caused by molecular vibra-
tions:  
 ∂ Ĥ

Ĥ = Qα + · · · . (5.128)
α
∂ Qα 0
R

Within the first-order approximation, a matrix element of this perturbation is given

by
96 5 Definitions and Derivations
 
 ∂ Ĥ

 mv (r, R)| Ĥ | nv (r, R) =  m (r, R )| 0
| n (r, R0 )
α
∂ Qα
R0
×χmv (R)|Q α |χnv (R)

= Vmn,α χmv (R)|Q α |χnv (R). (5.129)
α

The substitution of Eq. 5.129 into Eq. 5.127 leads to [17]

2π  
IC
km→n (T ) = Vmn,α Vnm,β Pmv (T )ρ(E mv )
 α,β v,v
×χmv (R)|Q α |χnv (R)χnv (R)|Q β |χmv (R). (5.130)

In particular, if only the single vibrational mode α is taken into consideration,

IC
km→n (T ) is given by

2π 2 
IC
km→n (T ) = V Pmv (T )ρ(E mv )χmv (R)|Q α |χnv (R)2 . (5.131)
 mn,α v,v

Equations 5.130 and 5.131 show that off-diagonal vibronic couplings, Vmn,α (α =
1, 2, . . .), govern the strength of internal conversion. Figure 5.3 schematically illus-
trates internal conversion.
We can introduce a density form of a VCC, vibronic coupling density (VCD) [5,
6, 10, 15]. It is a useful concept to analyze the origin of vibronic coupling. VCD is
formulated as follows.
First, we employ the equilibrium geometry in a reference electronic state
ref (r, R ) as reference geometry R . The reference geometry is optimized in the
0 0

state ref (r, R):

   
∂ E ref (R) ∂ Ĥ
=  ref (r, R )|
0
| ref (r, R0 ) = 0. (5.132)
∂ Qα R 0 ∂ Qα 0
R

Fig. 5.3 A conceptual

presentation of internal
conversion from an mth
electronic state to an nth
electronic state
5.4 Vibronic Coupling Density 97
! "
The electronic part of a linear vibronic coupling operator, ∂ Ĥ /∂ Q α 0 , can be
R
written as follows:
     
∂ Ĥ ∂ Ûne ∂ Ûnn
= + . (5.133)
∂ Qα 0 ∂ Qα 0
∂ Qα 0
R R R

On the other hand, Ûne can be represented as the sum of one-electron potentials u(r i )
(i = 1, 2, . . .), 
Ûne = u(r i ), (5.134)
i

where u(r i ) is defined as

 Z A e2
u(r i ) = − . (5.135)
A
4π 0 |r i − R A |

Accordingly, using a one-electron potential derivative vα (r i ),

 
∂u(r i )
vα (r i ) := , (5.136)
∂ Qα R0
! "
∂ Ĥ /∂ Q α can be rewritten by
R0
   
∂ Ĥ  ∂ Ûnn
= vα (r i ) + . (5.137)
∂ Qα ∂ Qα
R0 i R0

The substitution of Eq. 5.137 into Eq. 5.114 leads to

 
Vmn,α = ∗ 0 0
m (r, R ) n (r, R )d r 1 dω1 · · · d r i−1 dωi−1 dωi d r i+1 dωi+1 · · · d r N dω N
i
 
∂ Ûnn
vα (r i )d r i + δmn
∂ Qα
R0
 
 1 ∂ Ûnn
= ρmn (r i )vα (r i )d r i + δmn
N ∂ Qα
i R0
  
∂ Ûnn
= ρmn (x)vα (x)d x + δmn , (5.138)
∂ Qα 0 R

where N is the number of electrons, and ρmn (r i ) stands for an overlap density between
the mth and nth states,
98 5 Definitions and Derivations

∗
ρmn (r i ) := N m (r, R0 ) n (r, R0 )d r 1 dω1 · · · d r i−1 dωi−1 dωi d r i+1 dωi+1 · · · d r N dω N ,
(5.139)
in which r i and ωi denote the spatial and spin coordinates of an ith electron, respec-
tively. A diagonal element of an overlap density, ρmm (x), equals the total electron
density in the state m (r, R0 ). We hereafter abbreviate a diagonal element ρmm (x)
as ρm (x). An off-diagonal element of ρmn (x) (m = n) is equivalent to a transition
density within the orbital approximation, or the independent-particle approximation.
Equations 5.132 and 5.138 are combined to yield
  
∂ Ûnn
0= ρref (x)vα (x)d x + , (5.140)
∂ Qα 0
R

where ρref (x) is the total electron density in the state ref (r, R0 ). Then, subtracting
Eq. 5.140 from Eq. 5.138 leads to the simplification of a diagonal linear VCC,

Vm,α = ρm (x)vα (x)d x, (5.141)

where an electron-density difference ρm (x) is defined as

ρm (x) := ρm (x) − ρref (x). (5.142)

As shown
! in Eq. 5.141,
" the introduction of ρm (x) apparently vanishes the second
term, ∂ Ûnn /∂ Q α 0 .
R
The obtained relations are summarized as follows. A linear VCC, Vmn,α , is
described as an integral of a VCD, ηmn,α (x) [5, 6, 10, 15],

Vmn,α = ηmn,α (x)d x, (5.143)

ρm (x) × vα (x) (m = n),

ηmn,α (x) := (5.144)
ρmn (x) × vα (x) (m = n).

The concept of VCD enables us to visually and intuitively analyze the origin of
vibronic coupling in terms of ρm (x)/ρmn (x), which is obtained from electronic
structures, and vα (x), which is obtained from vibrational structures.

5.4.2 Higher-Order Vibronic Coupling Density

Similar to a linear VCC, QVCC and higher-order VCCs can also be represented in
density forms [6]. A kth-order VCC Wmn,α1 ···αk is defined by
5.4 Vibronic Coupling Density 99
 
∂ k Ĥ
Wmn,α1 ···αk :=  m (r, R )| 0
| n (r, R0 ). (5.145)
∂ Q α1 · · · ∂ Q αk
R0
! "
In particular, Wmn,α is Vmn,α . ∂ k Ĥ /∂ Q α1 · · · ∂ Q αk can be rewritten by
R0
   
∂ k Ĥ  ∂ k Ûnn
= wα1 ···αk (r i ) + , (5.146)
∂ Q α1 · · · ∂ Q αk ∂ Q α1 · · · ∂ Q αk
R0 i R0

where wα1 ···αk (r i ) is a kth one-electron potential derivative defined by

 
∂ k u(r i )
wα1 ···αk (r i ) := . (5.147)
∂ Q α1 · · · ∂ Q αk R0

The substitution of Eq. 5.146 into Eq. 5.145 leads to

 
∗
Wmn,α1 ···αk = m (r, R0 ) n (r, R0 )d r 1 dω1 · · · d r i−1 dωi−1 dωi d r i+1 dωi+1 · · · d r N dω N
i
 
∂ k Ûnn
×wα1 ···αk (r i )d r i + δmn
∂ Q α1 · · · ∂ Q αk
R0
 
 1 ∂ k Ûnn
= ρmn (r i )wα1 ···αk (r i )d r i + δmn
N ∂ Q α1 · · · ∂ Q αk
i R0
  
∂ k Ûnn
= ρmn (x)wα1 ···αk (x)d x + δmn . (5.148)
∂ Q α1 · · · ∂ Q αk 0 R

Therefore, Wmn,α1 ···αk is an integral of a kth-order VCD ηmn,α1 ···αk (x):

  
∂ k Ûnn
Wmn,α1 ···αk = ηmn,α1 ···αk (x)d x + δmn , (5.149)
∂ Q α1 · · · ∂ Q αk
R0

where
ηmn,α1 ···αk (x) := ρmn (x) × wα1 ···αk (x). (5.150)

5.4.3 One-Electron Property Density

We herein review vibronic coupling density from a general standpoint, leading to the
concept of one-electron property density. For an N -electron state m (r 1 ω1 , r 2 ω2 ,
. . . , r N ω N ), the anti-symmetric principle for Fermions states that

m (r 1 ωσ (1) , r 2 ωσ (2) , . . . , xωσ (i)=1 , . . . , r N ωσ (N ) )

= sgn(σ ) m (xω1 , r 2 ω2 , . . . , r N ω N ), (5.151)
100 5 Definitions and Derivations

where σ ∈ S N . The product of two electronic wave functions satisfies

∗
m (r σ (1) , . . . , xω1 , . . . , r N ωσ (N ) ) n (r σ (1) . . . , xω1 , . . . , r N ωσ (N ) )
= {sgn(σ )}2 m∗ (xω1 , r 2 ω2 , . . . , r N ω N ) n (xω1 , r 2 ω2 , . . . , r N ω N )
∗
= m (xω1 , r 2 ω2 , . . . , r N ω N ) n (xω1 , r 2 ω2 , . . . , r N ω N ). (5.152)

We can therefore obtain the following relation:


∗
m (x ω1 , r 2 ω2 , . . . , r N ω N ) n (x ω1 , r 2 ω2 , . . . , r N ω N )dω1 d r 2 · · · d r N

∗
= m (r 1 ω1 , x ω2 , . . . , r N ω N ) n (r 1 ω1 , x ω2 , . . . , r N ω N )
d r 1 dω1 dω2 d r 3 dω3 · · · d r N dω N
= ·· ·
∗
= m (r 1 ω1 , r 2 ω2 , . . . , r N −1 ω N −1 , x ω N ) n (r 1 ω1 , r 2 ω2 , . . . , r N −1 ω N −1 , x ω N )
d r 1 dω1 · · · d r N −1 dω N −1 dω N
1
= ρmn (x). (5.153)
N

Suppose that an operator Ô consists of one-electron operators ô without any

differential operators,
 N
Ô = ô(r i ), (5.154)
i=1

where r i denotes the spatial coordinate of electron i. A matrix element of Ô is given

by [19]
 
∗
Omn = ··· m Ô nd r1 · · · drN
N 
  
∗
= ··· m nd r1 · · · d r i−1 dωi d r i+1 · · · d r N ô(r i )d r i
i=1
N 
 1
= ρmn (r i ) ô(r i )d r i
i=1
N

= mn (x)d x, (5.155)

where r i = (r i , si ) with spatial coordinate r i and spin coordinate si for electron i.

Additionally, a one-electron property density mn (x) is defined as

mn (x) := ρmn (x) × ô(x). (5.156)

5.4 Vibronic Coupling Density 101

We can therefore discuss a one-electron property in the density form.

The vibronic coupling constant Vmn,α is
   
∂ Ûne
∗
Vmn,α = ··· nd r1 · · · d r N
∂ Qα
m
R0
   
∗ ∂ Û nn
+ ··· nd r1 · · · d r N
m
∂ Qα
R0
     
∗ ∂ Ûne ∂ Ûnn
= ··· nd r1 · · · d r N + δmn , (5.157)
m
∂ Qα ∂ Qα
R0 R0

where Une denotes the sum of nuclear–electronic potentials, Unn stands for the sum of
nuclear–nuclear potentials, and Q α is a mass-weighted normal coordinate of mode α.
Since (∂Une /∂ Q α ) R0 can be written as the sum of one-electron potential derivatives:

  
N
∂Une
= vα (r i ), (5.158)
∂ Qα R0 i=1

Vmn,α is given by the integral of the!vibronic coupling

" density ηmn,α (x) = ρmn (x) ×
vα (x) [18, 19], except the term of ∂ Ûnn /Q α 0 δmn .
R
Let us consider an electric dipole moment operator

µ̂ = er i , (5.159)
i

which is another example of one-electron operators: ô = e r i . We define an electric

dipole moment density as
τmn (x) = ρmn (x) ex. (5.160)

For an electronic state 0, the diagonal element of µ̂ yields a permanent electric

dipole moment:  
µ= ρ0 (x) exd x = τ00 (x)d x. (5.161)

The integrand is called a permanent dipole moment density.

The off-diagonal element of µ̂ between m and n is a transition dipole moment.
 
µmn = ρmn (x) exd x = τmn (x)d x, (5.162)

where τmn (x) is called the transition dipole moment density (TDMD). The intro-
duction of TDMD enables us to analyze and control optical processes via electronic
102 5 Definitions and Derivations

structures of molecules because absorptions, emissions, and scattering are governed

by transition dipole moments.

5.4.4 Related Concepts to Vibronic Coupling Density

5.4.4.1 Atomic Vibronic Coupling Constant

The decomposition of Vmn,α into its atomic components would be useful for grabbing
a local picture of vibronic coupling based on the evaluation. The partial derivative
∂/∂ Q α is given by

∂  ∂XA ∂ ∂Y A ∂ ∂ZA ∂
= + +
∂ Qα A
∂ Q α ∂ X A ∂ Q α ∂Y A ∂ Qα ∂ Z A
 u α(Ax) ∂ u α(Ay) ∂ u α(Az) ∂
= √ + √ + √ , (5.163)
A
m A ∂ X A m A ∂Y A mA ∂ZA

because of the following relation:

 u α(Ax)
X A = X 0A + √ Qα ,
α
mA
 u α(Ax)
Y A = Y A0 + √ Qα ,
α
mA
 u α(Ax)
Z A = Z 0A + √ Qα , (5.164)
α
mA

where X A , Y A , and Z A are the ξ -components (ξ = x, y, z) of the position of an Ath

nucleus; X 0A is X A at R0 , and u α(Aξ ) is the component of uα for an Ath nucleus and
ξ (= x, y, z). Accordingly, Vmn,α is rewritten by
  
 ∂ Ĥ u α(Ax)
Vmn,α =  m (r, R )| 0
| n (r, R0 ) √
∂ X A mA
A R0
 
∂ Ĥ u α(Ay)
+ m (r, R0 )| | n (r, R0 ) √
∂Y A 0 mA
 R 
∂ Ĥ u α(Az)
+ m (r, R0 )| | n (r, R0 ) √ . (5.165)
∂ZA 0 mA
R

Thus, Vmn,α can be regarded as the sum of Vmn,α(A) over all the nuclei:
5.4 Vibronic Coupling Density 103

∂ Ĥ u α(Ax)
Vmn,α(A) :=  m (r, R )| 0
| n (r, R0 ) √
∂XA 0 mA
 
R

∂ Ĥ u α(Ay)
+ m (r, R0 )| | n (r, R0 ) √
∂Y A 0 mA
 R
∂ Ĥ u α(Az)
+ m (r, R0 )| | n (r, R0 ) √ , (5.166)
∂ZA 0 mA
R

where Vmn,α(A) is called an atomic VCC (AVCC) .

Instead of AVCC, we can also use regional VCD for decomposing Vmn,α into
its atomic components . It is defined as the integration of VCD over the Wigner–
Seitz cell of each nucleus, which is the smallest polyhedron formed by planes that
perpendicularly bisect the segments between the nucleus and the neighboring nuclei.

5.4.4.2 Orbital Vibronic Coupling Density

For a Slater determinant 0 constructed by orthonormal spin orbitals, {χ (r i , ωi ) =

ψ(r i )σ (ωi )}, a diagonal element of the one-electron property can be decomposed as
[20] 
 0 | Ô| 0  = n a ψa (r i )|ô(r i )|ψa (r i ),
a∈occ

where n a is the occupation number of ψa . For an off-diagonal element between 0

and ab with orbital ψa in 0 replaced by ψb ,

 0 | Ô| a
b
= ψa (x)|ô(x)|ψb (x)

. We define orbital one-electron property density ωab (x) as

ωab (x) := ψa∗ (x)ψb (x)ô(x) = ρab (x) × ô(x),

where ρab (x) denotes orbital overlap density (transition density). We can decompose
orbital one-electron property in the density form:

ψa (x)|ô(x)|ψb (x) = ωab (x)d x. (5.167)

A kth-order orbital vibronic coupling constant (OVCC) between orbitals a and b

can be expressed in a density form:

Vab,α1 ···αk := ψa (x)|wα1 ···αk (x)|ψb (x) = ηab,α1 ···αk (x)d x, (5.168)
104 5 Definitions and Derivations

where orbital vibronic coupling density (OVCD) is defined as

ηab,α1 ···αk (x) := ρab (x) × wα1 ···αk (x). (5.169)

5.4.4.3 Reduced Vibronic Coupling Density

Within the linear combination of atomic orbitals (LCAO) approximation, an overlap

density, ρmn , is given by a linear combination of the products of two basis functions
based on atomic orbitals (AOs),

ρmn (x) = AμBν χ Aμ (x)χ Bν (x),
D mn (5.170)
A,B μ,ν

where χ Aμ denotes the μth basis function belonging to nucleus A. An electron-

density difference, ρm , is similar to the overlap density. Accordingly, the kth-order
VCD, ηmn,α1 ···αk , can be written by

ηmn,α1 ···αk (x) = AμBν χ Aμ (x)χ Bν (x)wα1 ···αk (x),
D mn (5.171)
A,B μ,ν

where wα1 ···αk (x) is the kth-order potential derivative. The kth-order VCC,
Wmn,α1 ···αk , can be obtained by its spatial integration,
 
Wmn,α1 ···αk = D mn
AμBν χ Aμ (x)χ Bν (x)wα1 ···αk (x)d x. (5.172)
A,B μ,ν

In general, wα1 ···αk (x) has symmetric distribution near atoms. In particular, the linear
one, vα (x), has p-type distribution around each atom, which is locally transformed
as ξ = x, y, z:

 z A e2 x − X A u α(Ax) y − Y A u α(Ay) z − Z A u α(Az)

vα (x) = √ + √ + √ ,
A
4π 0 |x − R A | 3 mA |x − R A | 3 mA |x − R A |3 m A
(5.173)
where z A denotes the charge of an Ath nucleus; X A , Y A , and Z A are the ξ -components
(ξ = x, y, z) of the position of an Ath nucleus; and u α(Aξ ) is the component of the
normal vector uα for an Ath nucleus and ξ (= x, y, z). This gives rise to cancel-
lation due to the symmetrical reason. For example, if both χ Aμ (x) and χ Bν (x) are
s-type, then the integrand, χ Aμ (x)χ Bν (x)vα (x), is p-type (= s × s × p) whose inte-
gral is exactly zero. We can eliminate such a zero contribution ηmn,α1 ···αk  (x) from
ηmn,α1 ···αk (x) to define the reduced vibronic coupling density (RVCD), η̄mn,α1 ···αk (x),
which describes net contributions to the VCC [21]:

η̄mn,α1 ···αk (x) = ηmn,α1 ···αk (x) − ηmn,α1 ···αk  (x). (5.174)
5.4 Vibronic Coupling Density 105

5.4.4.4 Effective Vibronic Coupling Density

According to the Hellmann–Feynman theorem [16, 17], the total differential of the
total electronic energy E is written by

 ∂E  
dE = d Qα = Vα d Q α , (5.175)
α
∂ Q α R0 α

where Q α is a mass-weighted normal coordinate of mode α. We here introduce

gradient and displacement vectors, V and d Q, as
 
V := Vα uα , d Q := uα d Q α , (5.176)
α α

where uα is a vibrational vector of mode α, which is orthonormal with each other.

Thereby, d E can be rewritten as

d E = V · d Q. (5.177)

The Cauchy–Schwarz inequality states that |d E| has the maximum value of |V ||d Q|
when V and d Q are parallel. Such a displacement vector is given by

V  Vα
d Q = us ds, us := = # uα . (5.178)
|V | α V2 α α

This is the steepest descent direction of E. The direction vector us with the corre-
sponding coordinate s is called the effective mode. Since each element of d Q is
given by
Vα
d Q α = # ds, (5.179)
2
α  Vα 

ds is expressed as
 Vα
ds = # d Qα . (5.180)
α α Vα2

The VCD for the effective mode, which is called the effective VCD (EVCD), is also
introduced as
 
∂u(x)
ηs (x) := ρ(x) × vs (x), vs (x) := . (5.181)
∂s R0

The spatial integration of ηs (x) yields the VCC for the effective mode,
106 5 Definitions and Derivations
 
∂ Ĥ
Vs :=  (r, R )| 0
| (r, R0 ). (5.182)
∂s
R0

It is called the effective VCC (EVCC) .

5.4.5 Vibronic Coupling Density as a Chemical Reactivity

Index

A VCD can be formulated as a reactivity index based on Parr and Yang’s theory [22,
23]. Before describing the VCD theory for chemical reactions, we briefly review
the conventional Parr and Yang’s theory to make this document self-contained. They
derived the frontier orbital theory in conceptual density functional theory (DFT) [24]
without the orbital approximation.
The total electronic energy of a reactant, E, is a functional of the number of total
electrons, N , and a one-electron nuclear–electronic potential, u(x),

E = E[N , u(x)]. (5.183)

Its differential, d E, is given by

    
∂E δE
dE = dN + du(x)d x
∂N δu
 u N

= μd N + ρ(x)du(x)d x, (5.184)

where an electronic chemical potential, μ, is defined as

 
∂E
μ := , (5.185)
∂N u

and ρ(x) denotes the total electron density, which is equal to a functional derivative
of E for u as per the first-order perturbation theory,
 
δE
ρ(x) = . (5.186)
δu N

As the same as E, μ is a functional of N and u(x),

μ = μ[N , u(x)]. (5.187)

Its differential dμ is given by

5.4 Vibronic Coupling Density 107
    
∂μ δμ
dμ = dN + du(x)d x
∂N δu N
u

= 2ηd N + f (x)du(x)d x, (5.188)

where absolute hardness η is defined as

 
1 ∂2 E
η := , (5.189)
2 ∂N2 u

and a Fukui function f (x) is defined by

 
δμ
f (x) := , (5.190)
δu N

which is equal to a derivative of ρ(x) with respect to N ,

 
∂ρ
f (x) = , (5.191)
∂N u

because of the Maxwell relation,

       
δμ δ ∂E ∂ δE ∂ρ
= = = . (5.192)
δu N δu ∂ N u N ∂N δu N u ∂N u

Differentials d E and dμ have the physical meanings of the linear responses of

an isolated reactant to the perturbations, d N and δu(x), which describe a reaction
process.
Parr and Yang proposed the following principle in analogy to thermodynamics: a
chemical reaction occurs such that |dμ| is maximum in the early stage of a reaction.
The first term on the right-hand side of Eq. 5.188 is a global quantity uniquely deter-
mined by the reactant species. Accordingly, the first term is independent of regios-
electivity. On the other hand, the second term on the right-hand side of Eq. 5.188 is
local. According to the principle, a chemical reaction occurs such that the second
term is maximum. That is, f (x) must have a large distribution around a reactive
region.
Within the finite difference approximation, f (x) is equal to the electron-density
difference ρ(x) caused by a charge transfer,

f (x) ≈ ρ(x). (5.193)

In addition, f (x) is equal to a frontier orbital density within the frozen orbital approx-
imation, namely, HOMO density |ϕHO |2 for electrophilic attacks or LUMO density
|ϕLU |2 for nucleophilic attacks. In this way, Parr and Yang’s principle yields the fron-
108 5 Definitions and Derivations

Fig. 5.4 Three types of reaction profiles. a E ES+PL+EX , where only the sum of electrostatic,
polarization, and exchange interactions is taken into consideration without any structural relaxation;
b charge-transfer interaction, E CT , is considered along with E ES+PL+EX without any structural
relaxation; c structural relaxation, E VC , is included in addition to all the above interactions

tier orbital theory, according to which a chemical reaction usually occurs around a
region with a large HOMO/LUMO density.
We herein discuss the effect of molecular deformation based on Parr and Yang’s
theory (Fig. 5.4). We assume that an effective mode s coincides with the direction of
the reaction path.
 Vα
us := # uα , (5.194)
2
α 
α αV 

where uα denotes the vibrational vector of an αth normal mode. The effective mode is
the steepest descent direction of the adiabatic potential surface in the charge-transfer
state. Accordingly, it can be regarded as an intramolecular reaction mode. Then,
du(x) defined in the conceptual DFT can be expressed using a reaction coordinate
s,  
∂u(x)
du(x) = ds + · · · = vs (x)ds + · · · , (5.195)
∂s R0

where vs (x) has the same definition as vα (x),

 
∂u(x)
vs (x) := . (5.196)
∂s R0

Within the finite difference approximation, Eq. 5.188 along with Eq. 5.195 can be
rewritten by 
dμ = 2ηd N + ρ(x)vs (x)d xds. (5.197)
5.4 Vibronic Coupling Density 109

The product of ρ(x) and vs (x) is equal to a diagonal linear VCD ηs (x) with respect
to s,
ηs (x) := ρ(x) × vs (x). (5.198)

Thus, the following relation can be obtained:


dμ = 2ηd N + ηs (x)d xds. (5.199)

To predict a reactive region that gives a large |dμ|, it is necessary but not sufficient
for Fukui function f (x) to take a large value in the reactive region. The potential
derivative vs (x) also plays an important role in |dμ|. Consequently, the vibronic
coupling density for the reactive mode s can be a chemical reactivity index that can
predict a large |dμ| region.
For example, Fig. 5.5 shows the vibronic coupling density of the naphthalene
cation for the reaction mode s, defined by the steepest direction.
We can find that the vibronic coupling density has a large value near the α-carbons.
This means that the motion of the α-carbon couples with the hole. This is consistent
with the prediction of the frontier orbital theory.
Using the nuclear Fukui function [25, 26] defined by
     
∂U ∂U ∂U
φX A =− , φY A =− , φZ A =− , (5.200)
∂XA N ∂Y A N ∂ZA N

the total differential of the chemical potential is written as


dμ = 2ηd N − (φ X A d X A + φY A dY A + φ Z A d Z A ) . (5.201)
A

Fig. 5.5 Vibronic coupling

density of the naphthalene
cation. Blue and gray
surfaces denote negative and
positive densities,
respectively. Reprinted by
permission from Springer
Nature, The Jahn–Teller
Effect: Fundamentals and
Implications for Physics and
Chemistry, Springer Series in
Chemical Physics, vol. 97,
ed. by H. Köppel, D.R.
Yarkony, H. Barentzen
(Springer-Verlag, Berlin,
2009)
110 5 Definitions and Derivations

The mass-weighted normal coordinate is expressed in terms of the nuclear coordinate

R as 
Qα = Aα X A X A + AαY A Y A + Aα Z A Z A . (5.202)
A

Therefore,
  
dμ = 2ηd N + Vα Aα X A d X A + AαY A dY A + Aα Z A d Z A , (5.203)
α A

and
  
φX A = − Vα Aα X A , φY A = − Vα AαY A , φ Z A = − Vα Aα Z A . (5.204)
α α α

The relation between the Jahn–Teller system and the Fukui function has been dis-
cussed by Balawender et al. [26].

5.4.6 Physical Quantities Related to Vibronic Coupling

Constant

In solid-state physics, the vibronic coupling is sometimes called electron–phonon

or electron–lattice coupling, mainly used for crystal vibrations. Let us look at the
other form of the VCC familiar with physicists. The electron–phonon coupling is
defined in the second quantized form of the vibronic Hamiltonian,

 2 ∂ 2 ω2
Ĥ = E 0 + − + α Q 2α + Vα Q α
α
2 ∂ Qα 2 2
  
1
= E 0 ĉ† ĉ + ωα b̂α† b̂α + + λα ĉ† ĉ(b̂α† + b̂α ) , (5.205)
α
2

where ĉ† and ĉ are the creation and annihilation operators of the one-electron state,
respectively. b̂α† and b̂α are the creation and annihilation operators of the vibrational
state with vibrational energy ωα :
$  
ωα i  ∂
b̂α† := Qα − √ , (5.206)
2 2ωα i ∂ Qα
$  
ωα i  ∂
b̂α := Qα + √ . (5.207)
2 2ωα i ∂ Qα
5.4 Vibronic Coupling Density 111

λα is the electron–phonon coupling constant, the dimension of which is that of energy.

Since Q α is rewritten as %
 ! † "
Qα = b̂α + b̂α , (5.208)
2ωα

λα is given by %

λα := Vα . (5.209)
2ωα

The dimensionless electron–phonon coupling constant, i.e., the dimensionless VCC,

is also defined by
λα Vα
gα := =& , (5.210)
ωα 2ωα3

and the Hamiltonian becomes

 
1
Ĥ = E 0 ĉ† ĉ + ωα b̂α† b̂α + + gα ĉ† ĉ(b̂α† + b̂α ) . (5.211)
2

The Rys–Huang factor is another measure of the coupling strength, which is defined
by
E αstab Vα2 1
Sα := = = gα2 , (5.212)
ωα 2ωα 3 2

where E α is the reorganization energy of mode α, which is equal to the Jahn–Teller

stabilization energy in the case of a degenerate system. Sα is the stabilization energy
measured by vibrational energy.
We also mention the relation between the VCC and mechanical force. Table 5.12
lists the fundamental physical quantities and their derivatives. Based on these quan-
tities, physical dimensions and atomic units of VCCs are given in Table 5.13. The
VCC is the derivative of energy for a mass-weighted coordinate. On the other hand,
the force has the dimension of the derivative of energy for length. The kth-order
partial derivatives satisfy
∂k k/2 ∂
k
= μ α , (5.213)
∂qαk ∂ Q kα

where qα and Q α stand for vibrational coordinates of mode α in the real and mass-
weighted spaces, and μα denotes the reduced mass of the mode. Accordingly, a linear
√
VCC, Vα , can be converted to the force, Fα , by multiplying μα :
√
Fα = μα Vα . (5.214)

As with the linear VCC, higher-order VCCs can be converted to those without
the dimension of mass. For example, if μα = κμ amu and Vα = κV a.u. (a.u. =
112 5 Definitions and Derivations

Table 5.12 Atomic units of fundamental physical quantities and their derivatives. Physical dimen-
sions and their symbols follow the International System of Units
Quantity Dimension Atomic unit of quantity
Definition Name Expression
Mass M Electron mass me
Charge IT Elementary e
charge
Action ML2 T−1 Reduced Planck’s constant (Dirac constant)
Length L Bohr radius Bohr a0
2
energy ML2 T−2 m e a02
Hartree Eh

Time T Eh

Table 5.13 Atomic units related to vibronic coupling constants and their related quantities. Physical
dimensions and their symbols follow the International System of Units
Quantity Expression Dimension Atomic unit
Wavenumber ν̃α L−1 a0−1
! "−1
Speed of light c LT−1 a0 Eh
! "−1

Frequency να = ν̃α c T−1 Eh
! "−1

Angular frequency ωα = 2π να T−1 Eh
−2
Force constant κα = ωα2 T−2 E h m −1
e a0
−1/2 −1
Linear vibronic Vmn,α M1/2 L T−2 Eh m e a0
coupling constant
−2
Quadratic vibronic Wmn,αβ T−2 E h m −1
e a0
coupling constant
−k/2 −k
kth-order vibronic Wmn,α1 ···αk M1−k/2 L2−k T−2 Eh m e a0
coupling constant

−1/2 √
E h m e a0−1 ), then Fα = κμ κV × 3.52 × 103 nN. In the usual case, the reduced
mass is about ∼ 1 to ∼ 10 amu. The linear VCC of ∼ 10−4 a.u. means that the force
2

in the order of nN is applied to nuclei in a molecule.

References

1. Ceulemans AJ (2013) Group theory applied to chemistry. Springer, Dordrecht

2. Fischer G (1984) Vibronic coupling: the interaction between the electronic and nuclear motions.
Academic Press, London
3. Bersuker I, Polinger V (1989) Vibronic interactions in molecules and crystals. Springer, Berlin
4. Azumi T, Matsuzaki K (1977) Photochem Photobiol 25(3):315
References 113

5. Sato T, Tokunaga K, Iwahara N, Shizu K, Tanaka K (2009) In: Köppel H, Yarkony DR,
Barentzen H (eds) The Jahn–Teller effect: fundamentals and implications for physics and
chemistry, Springer series in chemical physics, vol 97. Springer, Berlin
6. Sato T, Uejima M, Iwahara N, Haruta N, Shizu K, Tanaka K (2013) J Phys: Conf Ser 428
7. Bersuker I, Gorinchoi N, Polinger V (1984) Theo Chim Acta 66:161
8. Jahn HA, Teller E (1937) Proc R Soc London A 161:220
9. Bersuker IB (2001) Chem Rev 101(4):1067
10. Sato T, Tokunaga K, Tanaka K (2006) J Chem Phys 124
11. Tokunaga K, Sato T, Tanaka K (2007) J Mol Struct 838(1–3):116
12. Tokunaga K, Sato T, Tanaka K (2006) J Chem Phys 124(15)
13. Iwahara N, Sato T, Tanaka K, Chibotaru LF (2010) Phys Rev B 82(24)
14. Iwahara N, Sato T, Tanaka K (2012) J Chem Phys 136(17)
15. Sato T, Tokunaga K, Tanaka K (2008) J Phys Chem A 112:758
16. Hellmann H (1937) Einführung in die Quantenchemie. Deuticke and Company, Leipzig
17. Feynman RP (1939) Phys Rev 56(4):340
18. Uejima M, Sato T, Yokoyama D, Tanaka K, Park JW (2014) Phys Chem Chem Phys
16(27):14244
19. Sato T, Uejima M, Iwahara N, Haruta N, Shizu K, Tanaka K (2013) J Phys: Conf Ser 428(1)
20. Szabo A, Ostlund N (1982) Modern quantum chemistry: introduction to advanced electronic
structure theory. Macmillan, New York
21. Shizu K, Sato T, Tanaka K (2010) Chem Phys Lett 491:65
22. Parr RG, Yang W (1984) J Am Chem Soc 106:4049
23. Lee C, Yang W, Parr RG (1988) J Mol Struct: THEOCHEM 163:305
24. Geerlings P, Proft FD, Langenaeker W (2003) Chem Rev 103:1793
25. Cohen M (1994) J Chem Phys 101:8988
26. Balawender R, Proft F, Geerings P (2001) J Chem Phys 114:4441