NoN Covalente Forces

Challenges and Advances in Computational
Chemistry and Physics
Volume 19
Series Editor
Jerzy Leszczynski
Department of Chemistry and Biochemistry
Jackson State University Chemistry, Jackson, Mississippi, USA
This book series provides reviews on the most recent developments in computational
chemistry and physics. It covers both the method developments and their applica-
tions. Each volume consists of chapters devoted to the one research area. The series
highlights the most notable advances in applications of the computational methods.
The volumes include nanotechnology, material sciences, molecular biology, struc-
tures and bonding in molecular complexes, and atmospheric chemistry. The authors
are recruited from among the most prominent researchers in their research areas. As
computational chemistry and physics is one of the most rapidly advancing scientific
areas such timely overviews are desired by chemists, physicists, molecular biologists
and material scientists. The books are intended for graduate students and researchers.
More information about this series at http://www.springer.com/series/6918

Steve Scheiner
Editor
Noncovalent Forces
2123
Editor
Steve Scheiner
Department of Chemistry & Biochemistry
Utah State University
Logan
Utah
USA
Challenges and Advances in Computational Chemistry and Physics

ISBN 978-3-319-14162-6 ISBN 978-3-319-14163-3 (eBook)
DOI 10.1007/978-3-319-14163-3
Library of Congress Control Number: 2015931919
Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Contents
1 Introduction to the Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Steve Scheiner
Part I Hydrogen Bonds
2 Hydrogen Bonds Involving Sulfur: New Insights from ab Initio

Calculations and Gas Phase Laser Spectroscopy . . . . . . . . . . . . . . . . . . 15
Himansu S. Biswal
3 CH· · · π Interaction in Organic Molecules . . . . . . . . . . . . . . . . . . . . . . . 47

Osamu Takahashi
4 The CH··O H-Bond as a Determining Factor

in Molecular Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Steve Scheiner
5 Hydrogen Bonds Involving Radical Species . . . . . . . . . . . . . . . . . . . . . . 107

Qing-Zhong Li and Hai-Bei Li
6 Agostic and Hydrogen-Bonding X–H· · · M Interactions Involving a

d8 Metal Center: Recent Advances Towards Their Understanding 129
Jiří Kozelka
7 What is Common for Dihydrogen Bond and H· · ·σ

Interaction—Theoretical Analysis and Experimental Evidences . . . . 159
Sławomir J. Grabowski
Part II Other Bridging Atoms
8 The Pnicogen Bond in Review: Structures, Binding Energies,

Bonding Properties, and Spin-Spin Coupling Constants of Complexes
Stabilized by Pnicogen Bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Janet E. Del Bene, Ibon Alkorta and José Elguero
v
vi Contents
9 Chalcogen Bonds in Protein Architecture . . . . . . . . . . . . . . . . . . . . . . . . 265

Michio Iwaoka
10 A Unified View of Halogen Bonding, Hydrogen Bonding and Other

σ-Hole Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Peter Politzer and Jane S. Murray
11 The X-C· · ·Y Carbon Bond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

Devendra Mani and E. Arunan
12 Interplay of Hydrogen, Halogen, Lithium and Beryllium Bonds

in Complexes of Thiirane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Sean A. C. McDowell and Jerelle A. Joseph
13 Understanding Lone Pair-π Interactions

from Electrostatic Viewpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Shridhar R. Gadre and Anmol Kumar
Part III Aromatic Systems
14 Unraveling the Origin of Substituents Effects in π-Stacking

Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
Steven E. Wheeler
15 Noncovalent Interactions of Organic Ions with Polar Molecules

in the Gas Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
M. Samy El-Shall, Isaac K. Attah and Sean P. Platt
16 Anion-π Interactions in Supramolecular Chemistry and Catalysis 471

Antonio Bauzá, Pere M. Deyà and Antonio Frontera
17 A Survey of DNA–Protein π–Interactions: A Comparison of Natural

Occurrences and Structures, and Computationally Predicted
Structures and Strengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
Katie A. Wilson and Stacey D. Wetmore
Contributors
Ibon Alkorta Instituto de Química Médica (IQM-CSIC), Madrid, Spain

E. Arunan Department of Inorganic and Physical Chemistry, Indian Institute of
Science, Bangalore, India
Isaac K. Attah Department of Chemistry, Virginia Commonwealth University,
Richmond, VA, USA
Antonio Bauzá Departament de Química, Universitat de les Illes Balears, Palma
de Mallorca, Baleares, Spain
Himansu S. Biswal School of Chemical Sciences, National Institute of Science
Education and Research, Institute of Physics Campus, Bhubaneswar, Odisha, India
Janet E. Del Bene Department of Chemistry, Youngstown State University,
Youngstown, OH, USA
Pere M. Deyà Departament de Química, Universitat de les Illes Balears, Palma de
Mallorca, Baleares, Spain
José Elguero Instituto de Química Médica (IQM-CSIC), Madrid, Spain
M. Samy El-Shall Department of Chemistry, Virginia Commonwealth University,
Richmond, VA, USA
Department of Chemistry, Faculty of Science, King Abdulaziz University, Jeddah,
Saudi Arabia
Antonio Frontera Departament de Química, Universitat de les Illes Balears, Palma
de Mallorca, Baleares, Spain
Shridhar R. Gadre Department of Chemistry, Indian Institute of Technology
Kanpur, Kanpur, India
Sławomir J. Grabowski Kimika Fakultatea, Euskal Herriko Unibertsitatea
UPV/EHU, Donostia International Physics Center (DIPC), Donostia, Euskadi, Spain
IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
vii
viii Contributors
Michio Iwaoka Department of Chemistry, School of Science, Tokai University,

Hiratsuka-shi, Kanagawa, Japan
Jerelle A. Joseph Department of Biological and Chemical Sciences, The University
of the West Indies, Wanstead, Barbados
Jiří Kozelka Laboratoire de Chimie et Biochimie Pharmacologiques et Toxi-
cologiques, UMR 8601 CNRS, Université Paris Descartes, Paris, France
Department of Condensed Matter Physics, Faculty of Science, Masaryk University,
Brno, Czech Republic
Anmol Kumar Department of Chemistry, Indian Institute of Technology Kanpur,
Kanpur, India
Hai-Bei Li School of Ocean, Shandong University, Weihai, People’s Republic of
China
Qing-Zhong Li The Laboratory of Theoretical and Computational Chemistry,
School of Chemistry and Chemical Engineering, Yantai University, Yantai, People’s
Republic of China
Devendra Mani Department of Inorganic and Physical Chemistry, Indian Institute
of Science, Bangalore, India
Sean A. C. McDowell Department of Biological and Chemical Sciences, The
University of the West Indies, Wanstead, Barbados
Jane S. Murray Department of Chemistry, University of New Orleans, New
Orleans, LA, USA
Sean P. Platt Department of Chemistry, Virginia Commonwealth University,
Richmond, VA, USA
Peter Politzer Department of Chemistry, University of New Orleans, New Orleans,
LA, USA
Steve Scheiner Department of Chemistry & Biochemistry, Utah State University,
Logan, UT, USA
Osamu Takahashi Institute for sustainable sciences and development, Hiroshima
University, Higahshi-Hiroshima, Japan
Stacey D. Wetmore Department of Chemistry and Biochemistry, University of
Lethbridge, Lethbridge, Alberta, Canada
Steven E. Wheeler Department of Chemistry, Texas A&M University, College
Station, TX, USA
Katie A. Wilson Department of Chemistry and Biochemistry, University of
Lethbridge, Lethbridge, Alberta, Canada
Chapter 1
Introduction to the Volume
Steve Scheiner
Abstract The reader is introduced to the definition and diversity of noncovalent

forces. The division of these interactions into various subtopics is explained by way
of introducing each of the chapters. A brief exposition is provided of the typical
means by which computational chemists study these forces, and the language that is
commonly used.
Perhaps the first question a reader may have concerns the title of this volume: what
is a noncovalent force? By its very name, it is clear that these forces exclude the
strong bonds that hold atoms together within a molecule. What is usually meant is
the collection of phenomena that attract molecules or ions toward one another. And
this collection is indeed diverse. One can think first as one extreme of the strong
Coulombic attractions between ions of opposite charge in a crystal of NaCl. On
the other end of the spectrum lie the very weak forces between noble gas atoms as
for example in an Ar matrix. These two extremes bracket a wide array of forces of
intermediate strength. And it is this diverse set of phenomena which this volume is
intended to address. Some will be quite familiar, as for example the H-bond, which
has been part of the chemistry lexicon for over a century. But this familiar concept
has undergone an expansion over the years, and now covers a range of different
proton donor and acceptor groups not anticipated earlier. Other forces may be less
familiar, as in the case of attractions between electronegative atoms as occurs in
halogen, chalcogen, and even pnicogen bonds.
This volume is intended to provide the reader with an exposition of current thinking
about the forces that bind molecules together. It is aimed toward a diverse audience.
Computational chemists will be familiar with some of the concepts and tools, but
may not be fully up to date in terms of the rapidly evolving field of noncovalent
forces, and its various subfields. Experimentalists may not be fully cognizant of the
ways that modern computational chemists think about these forces, the tools that
are used to analyze them, and the power of computations to understand the nature
of these forces. The various chapters are thus targeted toward both audiences. The
reader will perhaps be impressed by the diversity of the types of noncovalent bonds
S. Scheiner ()
Department of Chemistry & Biochemistry,
Utah State University, 84322-0300 Logan, UT, USA
e-mail: steve.scheiner@usu.edu
© Springer International Publishing Switzerland 2015 1
S. Scheiner (ed.), Noncovalent Forces, Challenges and Advances in
Computational Chemistry and Physics 19, DOI 10.1007/978-3-319-14163-3_1
2 S. Scheiner
that attract molecules toward one another. But at the same time, there may be surprise
at some of the common themes that emerge in this seemingly disparate set of forces.
The various chapters comprise a tour of this diverse array of noncovalent bonds.
Some of the interactions will be well known, such as hydrogen bonds (HBs). But
our understanding of these venerable bonds has evolved over the past years, and
encompasses a range of new types which provide the focus here. The involvement of
sulfur as both proton donor and acceptor atom has been developed in recent years,
as described in Chap. 2. The next two chapters shift attention to the nominally weak
proton donor atom C, dividing into π system acceptors in Chap. 3, and lone pairs
of electronegative atoms in Chap. 4. The concept of HBs is expanded to open shell,
radical systems in Chap. 5, and the ability of metal centers to serve as proton acceptor
is explored in the following chapter. The normal expectation of a partially positively
charged bridging H atom is reversed in Chap. 7, which describes not only dihydrogen
bonds, but others where the bridge is better thought of as a hydride.
The next section expands further on the idea of HBs, in that there are a range
of electronegative atoms which can replace H as the bridge. Atoms from Group V
offer one such example, and the idea of the correspondingly named pnicogen bonds
is developed in Chap. 8. Chalcogen atoms (O, S, etc) represent the topic of Chap. 9,
and the underpinnings of halogen bonds are described in Chap. 10. Even the C atom
gets into the act, as is explored in Chap. 11. Just like the old school HBs, these sorts
of pnicogen, halogen, etc bonds can occur in aggregates of molecules, and strengthen
one another through the phenomenon of cooperativity, the subject of Chap. 12. The
next chapter brings to our consciousness the idea that lone pairs of one molecule can
interact attractively with the π systems of another, and the mechanism for how this
might be.
The extended π-clouds of aromatic systems represent an entire subject of their
own. Chap. 14 explores the way in which substituents on one π-system affect its
interaction with another. The presence of a charge on the system can be expected to
have a major effect on interactions involving π-systems, and these perturbations are
described in Chap. 15. The systems are enlarged in Chap. 16, with the focus remaining
on ion-π interactions. Finally Chap. 17 brings us closer to the macroscopic world of
biology, with its analysis of the interactions between DNA and proteins.
There are a number of computational tools and procedures that are particularly
common in the computational study of molecular interactions. These are briefly
introduced below, so that they might be familiar to the reader when encountered in
the ensuing chapters.
1.1 Energetics
Probably the most often asked question about any particular interaction has to do
with its “strength”. But that is a multifaceted issue, that can have several different
answers. Most commonly, the questioner is implicitly wondering how much energy
is required to break the noncovalent bond in question. Taking the hydrogen bond
1 Introduction to the Volume 3
between proton donor AH and B as an example, reaction 1 describes the association

or binding, so is typically exothermic, with a negative E, which is computed as the
difference between the energy of the HB complex and the sum of the energies of the
two monomers AH and B, Eq. (1.2)
AH + B → AH · · · B (1.1)
E = E(AH · · · B) − {E(AH) + E(B)} (1.2)

But there are some devils in the details of the computation of E. In the first place,
it is assumed that the geometry of AH · · B will be fully optimized, as will those of
the isolated monomers. But an alternate prescription would leave AH and B in the
geometries which they adopt within the AH · · B complex. There is a generally, but
by no means universally accepted, convention that the former procedure is called
the binding energy. The latter is commonly termed the interaction energy because
it more directly accesses the interaction of the two monomers within the complex.
In the majority of cases, the geometrical rearrangement of AH and B within the
complex is minor, and has only a small effect on their energies, so that the binding
and interaction energies are usually fairly similar. But the reader should be aware
that this is not always the case, so some attention to computational detail may be
called for. It is also important to recognize that different authors may use different
nomenclatures; for example E is the most common symbol but sometimes refers
to binding and sometimes to interaction energy. As written, E will be negative for
reaction 1, so will typically appear as such in the literature. But this is not always the
case, particularly within the written text where it is not unusual for authors to say,
for example, that the binding energy of the water dimer is 5 kcal/mol, but a value of
− 5 kcal/mol may appear for E in the pertinent table.
A second complication arises in that most calculations utilize a finite basis set
when computing E. Specifically, a given set of orbitals will be used to describe the
electronic structure of AH, and another set for B. The AH · · · B complex, however,
will utilize the union of the two latter sets of orbitals. The larger basis set of the
complex gives it an unfair advantage in the sense that the variation principle tells
us that the energy of a system becomes more negative as the basis is enlarged. A
more negative energy for AH · · B will leads to an overly negative E for reaction 1.
This spurious inflation of the binding energy, called basis set superposition error, is
usually corrected by a counterpoise procedure [1] that dates back to 1970, wherein
the energy of monomer AH is computed in a basis set that includes not only its own
orbitals, but also those of B, the so-called “ghost orbitals” since neither the nuclei
nor the electrons of B are present (and vice versa for monomer B).
The interaction energy up to this point refers only to the electronic contribution
to the thermodynamic E. The full E includes also vibrational, rotational, and
translational terms. Of these, the term that introduces the largest contribution is the
zero-point vibrational energy (ZPE), which takes account of the various vibrational
modes of reactants and product. Following this correction, it is common to denote
the new quantity as E + ZPE. The quantum calculation of the various entities also
permits the ready computation of H, S, and G. So even in energetic terms alone,
4 S. Scheiner
the strength of the bond can be measured by any of the above quantities, depending
upon one’s inclination.
The reader may have noticed that AH and B in reaction 1 are separate molecules.
Suppose one is interested in the strength of an intra molecular interaction. Reaction
1 is no longer relevant as there is no AH and B; that is the two constituents of the
noncovalent bond cannot be fully separated from one another. How then can the
system with this bond broken be defined? Unfortunately, there is no simple answer
to this thorny question. This situation has motivated a number of different definitions
over the years, but all provide somewhat different answers, so the reader should be
cautious in such cases.
1.2 Energy Dissection
There is often some curiosity expressed as to what exactly is the nature of the bind-
ing. What is meant by this can be vague, and there is a long list of different terms in
common usage: electrostatic, covalent, van der Waals, London forces, charge trans-
fer, donor/acceptor, and on and on. Some of these terms have clearer definition than
others and it is not unusual to see these terms applied differently by different authors.
There have been numerous attempts to dissect the total interaction energy into its
constituent parts. One of the earliest was due to Kitaura and Morokuma [2–4] who de-
fined the electrostatic component as the Coulombic attraction (or repulsion) between
the charge distributions of the two monomers in their pristine state, i.e. before the two
molecules are allowed to perturb one another’s charge clouds. This definition, which
seems sensible and along the lines of what most would suggest, implicitly contains
within it dipole-dipole, dipole-quadrupole, and higher terms in the multipole expan-
sion. Using the same frozen wave functions, these authors defined an exchange, or
steric, repulsion that results when the charge clouds penetrate one another.
Additional attractive forces are connected with the modification of each molecular
charge distribution, and it is here that the different energy decomposition schemes
differ most. Kitaura and Morokuma differentiated between charge that crossed a
boundary from one molecule to another, which they termed charge transfer, and
charge redistributions that remained on a single molecule, referred to as polarization.
It must be understood however, that such a distinction is artificial, dependent upon
where the boundary is drawn, as well as other factors. This fact motivated others to
provide a more rigorous means to separate the two phenomena [5–10]. Other schemes
avoid this distinction, leaving the charge redistribution as a single term which goes
by several names, including induction, orbital interaction, or simply polarization.
The last major contributor to the attractive force between molecules originates
in the instantaneous fluctuations of charge of one molecule, and its effect upon its
partner. This phenomenon is commonly dubbed dispersion, but also goes by London
or even van der Waals force. With respect to calculations, dispersion does not appear
at the SCF level, but only when electron correlation is added. For that reason, some
refer to any attraction found at the correlated level, over and above SCF attraction,
as dispersion, but this is not correct. Correlation causes other sorts of perturbations
as well, as for example, modifications of multipole moments within each monomer
which in turn change the electrostatic energy. So while dispersion is indeed contained
within the correlation energy, it represents only part of the latter. While dispersion
energy is by definition attractive, there is no such restriction on the full correlation
contribution.
Due in part to its limitation to the SCF level, and therefore inability to address
dispersion, the KM scheme faded in usage over the years. It was replaced by other
schemes, perhaps the most widespread of which is symmetry-adapted perturba-
tion theory [11–14] or SAPT. This formalism provides electrostatic (ES), induction
(IND), and dispersion (DISP) energies, but the effects of exchange reside not only
in a first-order exchange energy, but also its effects upon other terms in the form of
exchange-induction and exchange-dispersion. SAPT is particularly flexible in that it
can be applied at progressively higher levels of perturbation theory, leading to ad-
ditional terms, and to wave functions computed at either the Hartree-Fock or higher
levels. Although SAPT does not normally attempt to divide induction into charge
transfer and polarization, there have been some attempts [15–17] to do so, although
not frequently applied in the literature.
Among other energy decomposition schemes in common usage, there is LMO-
EDA which is based [18] on localized orbitals. In addition to electrostatic, polar-
ization, and dispersion terms, LMO-EDA provides separate attractive exchange and
repulsion terms. Another method that has found wide application is based on Natural
Bond Orbital (NBO) treatment, and is termed NEDA [19–22]. While the preceding
may be the methods in most widespread use, it is by no means an all-encompassing
list; there are also other procedures that have been proposed [23–31].
It is stressed that the reader should exercise caution in comparing the components
derived from one method to those from another, as different schemes can result in
discrepant data. One may wonder which procedure is best, which most accurately
reflects reality. It must be understood that just as a frog can be dissected in various
ways, there is no single correct way to partition the total interaction energy; each
method contains a certain degree of arbitrariness. Nonetheless, the use of these
methods has provided key insights into the physical nature of the interactions, as is
described in the following chapters.
1.3 Cooperativity
Most chemical systems contain more than just two molecules. For example, liquid
water is a rapidly fluctuating system of H-bonded molecules, where most are engaged
in H-bonds with more than one other molecule. It has long been recognized that
the whole is greater than the sum of its parts. That is, the total binding energy of
a AH · · · AH · · · AH trimer is greater than twice the binding energy of the simple
AH · · · AH dimer. This amplification is frequently denoted as cooperativity. To be
more exact, this enhancement through multiple noncovalent bonds can be termed
6 S. Scheiner
positive cooperativity, or sometimes as synergistic. But the formation of a trimer

does not always result in a stronger binding. If the central molecule, for example,
acts as double proton donor, as in H2 O · · · HOH · · · OH2 , the total binding energy is
less than that of a pair of simple dimers. This negative cooperativity is also sometimes
called antagonistic or diminutive.
Another way in which computational chemists address the specifics of coopera-
tivity is via multibody terms [32, 33]. Taking the H2 O · · · HF · · · HCl system as an
example, the trimer is first fully optimized. Holding the geometry fixed, the total in-
teraction energy E is computed as the difference between the energy of the trimer,
and the sum of the energies of the separate monomers. Pairwise energies are also
computed for each pair. E2 (H2 O · · · HF) is defined as the binding energy of the
indicated pair, at the geometry of the trimer, but with the third molecule excluded.
Likewise, E2 (HF · · · HCl) and E2 (H2 O · · · · · · HCl) make up the other two pair-
wise energies. (Note that the last term will be small due to the distance between
H2 O and HCl.) If there were no cooperativity at all, then E would be equal to the
sum of the three E2 quantities. But there is usually some degree of cooperativity,

and three-body interaction E3 is defined as the difference between E and E2 .
This same logic can be used to define four and higher body terms in complexes larger
than a trimer.
1.4 Electrostatic Potentials
As stated earlier, the electrostatic energy is usually defined as the Coulombic interac-
tion between the charge distributions of the two monomers. There are various means
to visualize this interaction, the most common of which presents the electrostatic
potential around each molecule in a color format. Red usually indicates the most
negative regions, and blue the most positive (but this pattern is not always followed,
so the reader should carefully read the figure caption). But in addition to the color
scheme, there are other matters that may differ from one work in the literature to
the next. For example, the potential is sometimes presented on a surface of con-
stant electron density, which is frequently 0.001 au, but can just as easily be another
isocontour. Alternately, the surface may be that which corresponds to the van der
Waals surface surrounding each atom, or in other cases a longer distance from the
nuclei may be deemed more useful, so can be twice the vdW radius, for example. On
any given surface, the potential will have one or more minima and maxima, whose
values can provide a quantitative means of comparing one molecule with another. It
is common for these extrema to be denoted Vs,min and Vs,max . Still another measure
dispenses with the idea of a fixed density or atomic radius and instead presents an
isopotential surface. In other words, the figure may illustrate a surface on which the
potential is equal to a preselected constant, for example 0.01 au.
Other measures of the electrostatic potential do not require a visual presentation.
For example, a charge can be assigned to each atom. But of course a table of atomic
charges provides a much cruder picture of the full potential. This approach can be
improved by adding atomic dipoles, or even quadrupoles. There remains, however,

the question of how to make these assignments. There are numerous prescriptions
for this purpose, some of the most common of which are Mulliken, NBO, and AIM
schemes. Comparison shows that different schemes often provide very different
charges, so some caution is necessary in the interpretation of this data.
1.5 Atoms in Molecules (AIM)
The electron density contains a great deal of information about the character of
the bonding within any system. One means of extracting the character of chemical
bonding rests on the topology of the density ρ, and most particularly its Laplacian
∇ 2 ρ. This “atoms-in-molecules” (AIM) analysis [34–37] leads to the concept of
basins that surround each atom and separate it from the others, and the total density
within this basin can then lead to the assignment of AIM atomic charges. Also of
importance, there are zero-gradient curves, termed bond paths, between atoms that
are viewed as noncovalent bonds in the context of intermolecular interactions. Along
each path there is a bond critical point, more or less midway between the two atoms.
Bond paths are not restricted to pairs of nuclei, but can also connect other areas,
such as π bonds [38–40]. The numerical value of the density ρ and Laplacian ∇ 2 ρ at
the bond critical point, has been shown in many cases to correlate nicely with other
measures of the strength of the bond, e.g. distance or energetics. However, the reader
should be aware that the identification of a bond between two atoms via AIM is not
necessarily entirely consistent with other means of analysis, particularly in the case
of weak noncovalent bonds. There are examples in the literature where AIM data are
at odds with other bond indicators [41–53].
1.6 NBO
The natural bond orbital (NBO) method transforms the fully delocalized wave func-
tion into one more in line with chemists’conventional ideas about individual chemical
bonds and lone pairs. In terms of noncovalent forces, NBO analysis of a dimeric
complex provides a series of second-order perturbation energies E(2) that corre-
spond to the energetic consequence of the transfer of charge from one orbital of the
first molecule, to another orbital of the second. In the case of an H-bonded complex
AH · · · B, for example, there is typically a sizable E(2) that corresponds to the transfer
of charge from the lone pair of B to the σ* antibonding orbital of AH: nB → σ*(AH).
It is this buildup of density in the latter antibond which has been implicated as the
source of the weakening and stretching of the A–H covalent bond in the H-bonded
complex. Other orbitals can be involved as well, for example π and π*, in other
systems, some of which are described in the ensuing chapters. The NBO scheme
offers a very useful picture by which to understand the nature of binding in various
sorts of noncovalent interactions.
8 S. Scheiner
The intermolecular charge transfers of the NBO prescription would best fall under
the rubric of induction energy. But it must be understood that induction is by definition
a broader quantity, including all motions of electron density, in all orbitals, and
encompasses both inter and intramolecular transfers. As such, it would be a mistake
to equate a single value of E(2) with the full measure of induction or charge transfer
energy. This scheme also offers a means of evaluating atomic charges, which are
often used in the literature.
1.7 Electron Density Redistributions
The mutual perturbations of each molecule upon its partner can be explicitly vi-
sualized as a density shift map. One starts with the total electron density of the
complex and then subtracts from it the sum of densities of the monomers, prior to
the interaction occurring, analogous to Eq. 1.1 above.
ρ = ρ(AH · · · B) − {ρ(AH) + ρ(B)} (1.3)
The resulting map offers a three-dimensional perspective on shifts within each

monomer, as well as density transfers from one molecule to the other. Maps such as
these have clearly displayed, for example, the loss of density that occurs around the
bridging proton in HBs, and the increase within the region occupied by the lone pair
of the electron donor atom.
Quantitative encapsulations of the charge shifts are often provided by comparisons
of atomic charges before and after the interaction occurs. But of course a single num-
ber assigned to an atom can miss important details that are clearly visible in the full
map. Another measure of density shifts is associated with the dipole moment. Paral-
lel with Eq. (1.3), the enhancement of the dipole moment is assessed by comparison
with the vector sum of the dipoles of the unperturbed monomers.
1.8 Perturbations of the Monomers
When two molecules engage in an interaction with one another, even relatively weak
ones, they exert a force which perturbs the internal properties of one another. One
manifestation is the change in internal geometries. Perhaps the most famous example
of this effect is the stretch undergone by the A–H covalent bond when it participates
in a AH · · B H-bond. But other geometrical changes can occur as well, including
stretches and contractions of other bonds, as well as modifications of internal bond
angles.
The rearrangement of electron density that accompanies formation of a nonco-
valent bond has repercussions on the spectral features of each monomer as well.
Taking HBs as an example once again the reduction of electron density in the vicin-
ity of the bridging proton leads to a diminution of the NMR chemical shielding σ
and consequently to a downfield shift of the signal of this proton. Another example
from the H-bonding interaction is the red shift of the A–H stretching frequency in
the vibrational spectrum of AH · · · B, along with an intensification of this band. This
reduced frequency is connected with a weakening of the covalent A–H bond, which
has been attributed to the increased population of the σ*(AH) antibonding orbital.
The latter is consistent with the idea that charge transfer into this antibonding orbital
takes place from the lone pair of the electron donor atom B. The reader is forewarned,
however, that most calculations of vibrational frequencies employ a fully harmonic
approximation, so should make direct comparison to experimental quantities with
caution. The magnitudes of the downfield NMR shift as well as the red shift of
ν(AH) have been shown in numerous studies to correlate with energetic and geo-
metric measures of the strength of the HB. These spectroscopic perturbations offer
a particularly useful bridge with experimental observations, to assess the accuracy
of the calculations.
The reader is now hopefully prepared for the chapters that follow with some
understanding of the lexicon and computational tools that are being applied by the
various contributing authors. You will probably note that there is some disagreement
from one set of authors to the next as to the precise mechanisms underlying the
molecular attractions. This disagreement is a healthy sign that the topic of this volume
represents a vibrant field that is still encompassing new ideas. It is hoped that you
will leave your reading desk with a well-rounded comprehension of modern views
of the forces that hold molecules together: May the Noncovalent Force be with you.
References
1. Boys SF, Bernardi F (1970) The calculation of small molecular interactions by the differences
of separate total energies. Some procedures with reduced errors. Mol Phys 19:553–566
2. Morokuma K (1971) Molecular orbital studies of hydrogen bonds. III. C=O... H–O hydrogen
bond in H2 CO... H2 O and H2 CO... 2H2 O. J Chem Phys 55:1236–1244
3. Kitaura K, Morokuma K (1976)A new energy decomposition scheme for molecular interactions
within the Hartree-Fock approximation. Int J Quantum Chem 10:325–340
4. Morokuma K (1977) Why do molecules interact? The origin of electron donor-acceptor
complexes, hydrogen bonding, and proton affinity. Acc Chem Res 10:294–300
5. Stevens WJ, Fink WH (1987) Frozen fragment reduced variational space analysis of hydrogen
bonding interactions. Application to the water dimer. Chem Phys Lett 139:15–22
6. Chen W, Gordon MS (1996) Energy decomposition analyses for many-body interaction and
applications to water complexes. J Phys Chem 100:14316–14328
7. van der Vaar A, Merz KM (1999) Divide and conquer interaction energy decomposition. J Phys
Chem A 103:3321–3329
8. Salvador P, Duran M, Mayer I (2001) One- and two-center energy components in the atoms in
molecules theory. J Chem Phys 115:1153–1157
9. Fedorov DG, Kitaura K (2007) Pair interaction energy decomposition analysis. J Comput Chem
28:222–237
10. Khaliullin RZ, Bell AT, Head-Gordon M (2008) Analysis of charge transfer effects in molecular
complexes based on absolutely localized molecular orbitals. J Chem Phys 128:184112
11. Szalewicz K, Jeziorski B, Rybak S (1991) Perturbation theory calculations of intermolecular
interaction energies. Int J Quantum Chem QBS 18:23–36
10 S. Scheiner
12. Jeziorski B, Moszynski R, Szalewicz K (1994) Perturbation theory approach to intermolecular

potential energy surfaces of van der waals complexes. Chem Rev 94:1887–1930
13. Szalewicz K, Jeziorski B (1997) In: S. Scheiner (ed), Molecular Interactions. From Van der
Waals to strongly bound complexes. Wiley, New York, p 3–43
14. Parrish RM, Sherrill CD (2014) Spatial assignment of symmetry adapted perturbation theory
interaction energy components: the atomic SAPT partition. J Chem Phys 141:044115
15. Stone AJ (1993) Computation of charge-transfer energies by perturbation theory. Chem Phys
Lett 211:101–109
16. Stone AJ, Misquitta AJ (2009) Charge-transfer in Symmetry-Adapted Perturbation Theory.
Chem Phys Lett 473:201–205
17. Misquitta AJ (2013) Charge Transfer from Regularized Symmetry-Adapted Perturbation
Theory. J Chem Theory Comput 9:5313–5326
18. Su P, Li H (2009) Energy decomposition analysis of covalent bonds and intermolecular
interactions. J Chem Phys 131:014102
19. Glendening ED, Streitwieser A (1994) Natural energy decomposition analysis: An energy
partitioning procedure for molecular interactions with application to weak hydrogen bonding,
strong ionic, and moderate donor–acceptor interactions. J Chem Phys 100:2900–2909
20. Glendening ED (1996) Natural energy decomposition analysis: explicit evaluation of electro-
static and Polarization effects with application to aqueous clusters of alkali metal cations and
neutrals. J Am Chem Soc 118:2473–2482
21. Schenter GK, Glendening ED (1996) Natural energy decomposition analysis: the linear
response electrical self energy. J Phys Chem 100:17152–17156
22. Glendening ED (2005) Natural energy decomposition analysis: extension to density functional
methods and analysis of cooperative effects in water clusters. J Phys Chem A 109:11936–11940
23. Ziegler T, Rauk A (1977) On the calculation of bonding energies by the Hartree Fock Slater
method. Theor Chim Acta 46:1–10
24. Bagus PS, Hermann K, Bauschlicher CWJ (1984) A new analysis of charge transfer and po-
larization for ligand–metal bonding: Model studies of Al4 CO and Al4 NH3 . J Chem Phys
80:4378–4386
25. van der Vaart A, Merz KM (1999) Divide and conquer interaction energy decomposition. J Phys
Chem A 103:3321–3329
26. Mo Y, Gao J, Peyerimhoff SD (2000) Energy decomposition analysis of intermolecular
interactions using a block-localized wave function approach. J Chem Phys 112:5530–5538
27. Khaliullin RZ, Head-Gordon M, Bell AT (2006) An efficient self-consistent field method for
large systems of weakly interacting components. J Chem Phys 124:204105
28. Mitoraj M, Michalak A (2007) Natural orbitals for chemical valence as descriptors of chemical
bonding in transition metal complexes. J Mol Model 13:347–355
29. Reinhardt P, Piquemal J-P, Savin A (2008) Fragment-Localized Kohn–Sham orbitals via a sin-
gles configuration-interaction procedure and application to local properties and intermolecular
energy decomposition analysis†. J Chem Theory Comput 4:2020–2029
30. Wu Q, Ayers PW, Zhang Y (2009) Density-based energy decomposition analysis for inter-
molecular interactions with variationally determined intermediate state energies. J Chem Phys
131:164112
31. Horn PR, Sundstrom EJ, Baker TA, Head-Gordon M (2013) Unrestricted absolutely lo-
calized molecular orbitals for energy decomposition analysis: Theory and applications to
intermolecular interactions involving radicals. J Chem Phys 138:134119
32. Xantheas SS, Dunning THJ (1993) Ab initio studies of cyclic water clusters (H2 O)n , n = 1–6.
I. Optimal structures and vibrational spectra. J Chem Phys 99:8774–8792
33. Xantheas SS (1994) Ab initio studies of cyclic water clusters (H2 O)n , n = 1–6. II. Analysis of
many-body interactions. J Chem Phys 100:7523–7534
34. Bader RFW (1990) Atoms in molecules, a quantum theory. Clarendon Press, Oxford
35. Bader RFW, Cheeseman JR, Laidig KE, Wiberg KB, Breneman C (1990) Origin of rotation
and inversion barriers. J Am Chem Soc 112:6530–6536
36. Popelier PLA (2000) Atoms in molecules. An introduction. Prentice Hall, Harlow
37. Popelier PLA, Bader RFW (1992) The existence of an intramolecular C–H–O hydrogen bond
in creatine and carbamoyl sarcosine. Chem Phys Lett 189:542–548
38. Domagala M, Grabowski SJ (2009) XH... π and X–H... N hydrogen bonds—acetylene and
hydrogen cyanide as proton acceptors. Chem Phys 363:42–48
39. Grabowski SJ, Ugalde JM (2010) Bond Paths Show Preferable Interactions: Ab Initio and
QTAIM Studies on the X–H... π Hydrogen Bond. J Phys Chem A 114:7223–7229
40. Grabowski SJ (2013) Dihydrogen bond and X–H... σ interaction as sub-classes of hydrogen
bond. J Phys Org Chem 26:452–459
41. Lane JR, Contreras-García J, Piquemal J-P, Miller BJ, Kjaergaard HG (2013) Are bond critical
points really critical for hydrogen bonding? J Chem Theory Comput 9:3263–3266
42. Jablonski M (2012) Energetic and geometrical evidence of nonbonding character of some
intramolecular halogen · · · oxygen and other Y · · ·Y interactions. J Phys Chem A 116:3753–
3764
43. Jablonski M, Palusiak M (2013) The halogen... oxygen interaction in 3-halogenopropenal
revisited—the dimer model vs. QTAIM indications. Chem Phys 415:207–213
44. Varadwaj PR, Varadwaj A, Jin B-Y (2014) Significant evidence of CO and CC long-range
contacts in several heterodimeric complexes of CO with CH3 –X, should one refer to them as
carbon and dicarbon bonds! Phys Chem Chem Phys 16:17238–17252
45. Cormanich RA, Moreira MA, Freitas MP, Ramalho TC, Anconi CPA, Rittner R, Contreras RH,
Tormena CF (2011) 1h JFH coupling in 2-fluorophenol revisited: Is intramolecular hydrogen
bond responsible for this long-range coupling? Magn Reson Chem 49:763–767
46. Johnson ER, Keinan S, Mori-Sanchez P, Contreras-Garcia J, Cohen AJ, Yang W (2010)
Revealing noncovalent interactions. J Am Chem Soc 132:6498–6506
47. Azofra LM, Scheiner S (2014) Complexation of n SO2 Molecules (n = 1,2,3) with formaldehyde
and thioformaldehyde. J Chem Phys 140:034302
48. Azofra LM, Scheiner S (2014) Complexes containing CO2 and SO2 . mixed dimers, trimers
and tetramers. Phys Chem Chem Phys 16:5142–5149
49. Weinhold F, Schleyer PvR, McKee WC (2014) Bay-type H · · · H “bonding” in cis-2-butene
and related species: QTAIM versus NBO description. J Comput Chem 35:1499–1508
50. Alkorta I, Sanchez-Sanz G, Elguero J (2014) Pnicogen bonds between X = PH3 (X = O, S,
NH, CH2 ) and phosphorus and nitrogen bases. J Phys Chem A 118:1527–1537
51. Foroutan-Nejad C, Shahbazian S, Marek R (2014) Toward a consistent interpretation of the
QTAIM: tortuous link between chemical bonds, interactions, and bond/line paths. Chem Eur J
20:10140–10152
52. Ma F, Li A (2014) A computational study of pnicogen−hydride interaction in complexes
XH2 P... HBeY. Comput Theor Chem 1045:78–85
53. Cormanich RA, Rittner R, O’Hagan D, Bühl M (2014) Analysis of CF · · · FC interactions on
cyclohexane and naphthalene frameworks. J Phys Chem A 118:7901–7910
Part I
Hydrogen Bonds
Chapter 2
Hydrogen Bonds Involving Sulfur: New Insights
from ab Initio Calculations and Gas Phase
Laser Spectroscopy
Himansu S. Biswal
Abstract The hydrogen bonds involving sulfur (sulfur center hydrogen bonds;
SCHBs) are generally regarded as weak H-bonds in comparison with the conven-
tional N–H· · · O, O–H· · · O, N–H· · · N and O–H· · · N H-bonds. One of the reasons
being considered for this is the smaller electronegativity of S than O or N. However,
recent high resolution laser spectroscopy in combination with quantum chemical
calculations reveals that SCHBs can be as strong as conventional H-bonds. Surpris-
ingly, in the case of methionine containing dipeptides the amide-N–H· · · S H-bonds
are even stronger than amide-N–H· · · O = C H-bonds. Sulfur is not only a potential
H-bond acceptor, but the S–H group is also a very good H-bond donor and capable
of forming a variety of H-bonds. For example, the S–H· · · π H-bond between H2 S
and indole/benzene is found to be the strongest H-bond among O–H· · · π, O–H· · · π,
and C–H· · · π H-bonds. In general the SCHBs are dispersive in nature. This chapter
details about few SCHB systems, many more systems need to be studied extensively
and carefully to unravel many facts and facets about SCHBs. The major challenge for
the experimentalists is to accurately determine the intra- and intermolecular H-bond
energies and for the theoreticians to propose a universal H-bond descriptor.
2.1 Introduction
The IUPAC’s new definition of hydrogen bonding (H-bonding) [1] is stated as “The
hydrogen bond is an attractive interaction between a hydrogen atom from a molecule
or a molecular fragment X–H in which X is more electronegative than H, and an atom
or a group of atoms in the same or a different molecule, in which there is evidence
of bond formation.” As is mentioned in the definition electro negativities of X and
Y in X–H· · ·Y regulates the H-bond strength, i.e. the H-bond strength increases
with increasing the electronegativity values of donor and acceptor atoms [2, 3]. This
is exactly followed by the H-bonds involving second row elements in general and
H. S. Biswal ()
School of Chemical Sciences, National Institute of Science Education and Research, Institute of
Physics Campus, PO: Sainik School, 751 005 Bhubaneswar, Odisha, India
e-mail: himansu@niser.ac.in

16 H. S. Biswal et al.
oxygen and nitrogen in specific. In fact, N–H· · · O, O–H· · · O, N–H· · · N and O–

H· · · N H-bonds are very essential in supramolecular chemistry, crystal engineering,
designing new materials and structure and function of biomolecules [4–10]. Apart
from nitrogen and oxygen, halogens (F, Cl, Br and I) are found to be potential H-bond
acceptors and form weaker H-bonds with NH, OH and CH protons [11–16]. Similarly
the possibility of H-bond formation by the higher group members of chalcogens
(sulfur and selenium) have also been explored and documented in the literature [17–
19]. As the atomic number in the chalcogen family increases, the electronegativity
of the group members decreases and their metallic character increases. The electro
negativities of sulfur (2.58) and selenium (2.55) in the Pauling scale are much smaller
than that of oxygen atom (3.44) and comparable with the electronegativity of carbon
(2.55). The CH group is regarded as a weak hydrogen bond donor [20–25] and very
recently it is reported that sp3 -C can be involved in non-covalent interactions very
similar to halogen and hydrogen bond termed as carbon bond [26–28]. Comparable
electronegativity of sulfur with that of carbon perhaps leads to the general consensus
that the hydrogen bonds involving sulfur (Sulfur Center Hydrogen Bond, SCHB)
are very weak. Surprisingly nature has chosen two amino acids such as cysteine and
methionine bearing sulfur atom in their side chains. It is in fact observed in the protein
structure data bank that sulfur can form many non-covalent interactions including
H-bonds that influence structure and function of proteins [17, 29–39]. However the
strength (weak or strong), nature (electrostatic or dispersive), directionality (linear or
non-linear) of SCHBs are still debatable and need to be investigated at the molecular
level.
This chapter summarizes recent progress in the assessment of hydrogen bonds
involving sulfur. Hydrogen sulfide dimer (H2 S–H2 S) is the simplest model system
for intermolecular SCHBs. In this case S–H· · · S hydrogen bond is formed between
the two monomer units. The computed S–H· · · S H-bond energy at semi-empirical
level is 0.71 kcal/mol [40], which is much smaller than the conventional hydro-
gen bond energies. In this example sulfur acts as a H-bond acceptor as well as a
donor. There are few matrix isolation IR spectroscopy studies on the SCHBs [41–
44]. The experimental results demonstrate the sulfur atom of Methanethiol (MeSH)
and dimethyl sulfide (DMS) can be a potential hydrogen bond acceptor, although
a weaker acceptor than oxygen [41]. However, the work by M. Wierzejewska [42]
demonstrated that the sulfur atom of DMS is better hydrogen bond acceptor than the
sulfur atom of dimethyldisulfide (DMDS) and hydrogen sulfide (H2 S) and compa-
rable to the oxygen atom of dimethylether (DME) as observed in the DMS–HNO3
and DME–HNO3 binary complexes [43]. Apart from the intermolecular S–H· · · S
and O–H· · · S hydrogen bonds, intramolecular SCHBs are frequently observed in
the crystals [45–48]. The spectroscopic evidence for intramolecular SCHB is the mi-
crowave study of the conformers of jet-cooled thiodiglycol (TDG) [49]. It is observed
that intramolecular O–H· · · S hydrogen bond provides a highly compact and folded
structure to the most stable conformer of TDG. Sulfur is also capable of forming
S–H· · · π π-type H-bonds along with S–H· · · S, N–H· · · S and O–H· · · S σ-type H-
bonds as cited above. For instance, H2 S forms strong S–H· · · π H-bonded complex
with benzene π-electrons. The computed H-bond energy in H2 S and benzene dimer
2 Hydrogen Bonds Involving Sulfur 17
at CCSD(T)/CBS level is − 2.85 kcal/mol. This binding energy is comparable to

that of O–H· · · π and N–H· · · π H-bond energy and almost two times of C–H· · · π
H-bond energy [50].
Many theoretical approaches have been adopted to shed light on the experimental
observations on SCHBs. The main focus is to understand the directionality, nature
and hydrogen bond energies of SCHBs [18, 44, 51, 52, 53–56]. One of the detailed
theoretical study by Platts et al. [18] on SCHBs suggest that sulfur atoms of H2 S and
H2 CS are weaker H-bond acceptor than oxygen atoms of H2 O and H2 CO. Sulfur
prefers to form perpendicular H-bonds where as oxygen forms linear hydrogen bonds,
the reason being (X) S· · · H–F H-bond is dominated by charge (H)-quadrupole (S) in-
teraction while charge-charge attraction is the main contributor for the (X)O· · · H–F
H-bonded complexes. In other words SCHBs are dispersive in nature. One of the im-
portant finding by them is that the Laplacian and charge density derived from atoms
in molecules (AIM) theory is unable to predict the directionality of SCHBs in these
complexes. The H-bond energy in (X)S· · · H–F is almost half of that in (X)O· · · H–F
complexes. On contrary in the case of methanol–dimethylsulfide (OH· · · S H-bond)
and methanol–dimethylether (OH· · · O H-bond), the H-bond energies are very sim-
ilar. The estimated H-bond energy of OH· · · S H-bond at the coupled cluster level is
− 5.46 kcal/mol which is slightly smaller than the H-bond energy of − 5.97 kcal/mol
for OH· · · O H-bond. Similarly in a recent work on the nature of the SCHBs, Singh
and co-workers [56] from the symmetry adopted perturbation theory analysis (SAPT)
of variety of intermolecular H-bond complexes with DMS suggest that electrostatic
component of the stabilization energy is the major contributor in the SCHBs rather
than the dispersive energy component as mentioned by Barcke and co-workers [18].
One thing emerges out from these few examples is that the hydrogen bond acceptor
strength of sulfur, directionality and the electrostatic/dispersive nature of SCHB are
not very conclusive and still debatable.
Apart from spectroscopic and computational evidences, the existences SCHBs
have been realized in many organic crystals, peptides and proteins [48, 57–73,
74–76]. The intrachain S–H· · · O = C H-bonds involving sulfur atom of cysteine
side chain and backbone carbonyl oxygen have been observed for globular proteins
[17, 77]. The strength and occurrence of SCHBs in proteins have recently been in-
vestigated by Zhou et al. [78]. By analyzing the geometrical parameters of 500 high
resolution protein structures, they concluded that “(i) SCHBs regulate secondary
structure of peptides, (ii) SCHBs have longer hydrogen bond length (d) and smaller
hydrogen bond angle (θ), (iii) sulfur atoms in the side chain of cysteine, half-cysteine
and methionine forms weaker hydrogen bond, (iv) the disulfide bonds are capable
of forming SCHBs, (v) the SH group of cysteine acts as a hydrogen bond donor and
forms weak S–H· · · π hydrogen bond, and (vi) methionine sulfur is a poor hydrogen
bond acceptor.” Some of these findings are very pertinent and useful while some need
to be studied at the molecular level. For example, looking at the electro negativities
of acceptor atoms and distance between X and Y in X–H· · ·Y H-bond systems, it is
very difficult to say whether Y (in this case sulfur) will be weak or strong H-bond
acceptor. A comprehensive search of crystal structure data base by Allen et al. [19]
and Steiner [79] emphasizes that except for few thioethers sulfur is a poor hydrogen
bond acceptor than oxygen and the role of SCHBs in bimolecular structure function
are minimal. On the contrary, Francois et al. [61] found very strong N–H· · · S hydro-
gen bond in the crystals of (Triazocyclononane)2 Fe2 S6 . They also observed that the
strength of N–H· · · S and N–H· · · O H-bonds highly depend on the hydrogen bond
angle (θ). For the mentioned crystal structure the N–H· · · S hydrogen bond strength
is optimal for θ ∼ 800 , where as N–H· · · O hydrogen bond is favorable for θ > 1150 .
Hence, one can conclude that the strength of SCHB is dependent on the systems and
geometry of hydrogen bond. Similarly the role of SCHB in biomolecular structure
and function cannot be neglected. For example, in a recent report [80] methionine is
found to be very essential for the catalytic role of phosphite dehydrogenase (PTDH),
that catalyses the oxidation of phosphite to phosphate. In fact, the N–H· · · S hydro-
gen bond interaction between His292 and Met53 stabilizes the transition state of the
reaction, thereby lowering the activation energy barrier and increasing the catalytic
activity of PTDH.
The above cited few examples from the literature manifest many facts and facets
of SCHBs, that are still unexplored and need to be addressed very accurately. It is
very clear that sulfur can form hydrogen bond and SCHBs are ubiquitous in proteins
and organic crystals. For the last few years we have been trying to unravel SCHBs at
molecular level using the arsenal of high resolution gas phase laser spectroscopy and
computational methods. Hope our results will mark a point with respect to vexata
quaestio of the acceptor strength of sulfur, directionality and nature of SCHBs.
The aim is to mimic the SCHBs observed in the organic crystals, peptides, and
proteins by choosing relevant simpler model compounds. The model compounds
and their clusters are prepared in the supersonic jet cooled condition by exploiting
isentropic phenomenon. The UV and IR spectroscopy of these compounds and their
clusters are recorded in the jet cooled condition. Supersonic jet spectroscopy has
many advantages over the conventional spectroscopy and used by many research
groups to study molecular complexes that ranges from weakly bound van der Waal’s
complexes to very strongly bound ionic-hydrogen bonded complexes [81–87]. One
of the major benefits of this technique is that it reduces thermal broadening of the
spectra and helps the molecular complexes to form at a lower temperature that cannot
be possible in the ambient temperature and pressure. Combining this with TOF-Mass
spectrometer allows us to detect the complexes of our interest and studying them
in isolated conditions. Many double resonance spectroscopic methods are used to
record the high resolution conformer specific UV and IR spectra. The computational
methods are very handy in assigning the spectra, estimating the H-bond energy and
predicting the nature of the SCHBs. The experimental data are also very useful
in benchmarking the computational methods and provide plenty opportunities to
develop new theoretical methods to understand and correctly predict non-covalent
interactions such as SCHBs. In the next few sections we will discuss how the marriage
between the computational methods and high resolution laser spectroscopic methods
help us to shed light on the H-bonds sulfur atoms; sulfur being a H-bond acceptor or
S–H as a H-bond donor.
Ser- 361
Met- 350
Met- 389
Thr- 368
Met- 36
Tyr- 38
Pyrrolidone carboxyl peptidase Glycogen phosphorylase b Aspartate aminotransferase

( PDB: 1a2z) ( PDB: 1a8i) ( PDB: 1ajs)
Cys- 145
Tyr- 447 Cys- 283
Ser- 285
Cys- 460
Thr- 148
Catalase, human erythrocyte 5'-deoxy-5'-methylthioadenosine Creatine kinase

( PDB: 1dgf) phosphorylase ( PDB: 1cb0) ( PDB: 1qh4)
Fig. 2.1 O–H· · · S Hydrogen bonds involving sulfur atoms of methionine and cysteine in proteins
2.2 “Sulfur” as a H-Bond Acceptor
Sulfur can be a potential H-bond acceptor. The mostly studied H-bonds with sulfur
as a H-bond acceptor are O–H· · · S and N–H· · · S H-bonds. More detail of these
H-bonds will be discussed in the subsequent sections.
2.2.1 O–H· · · S Hydrogen Bonding
It is evident from the protein structure data bank that sulfur atoms of methionine and
cysteine are capable of forming O–H· · · S H-bonds with the side chain OH groups
of tyrosine, threonine and serine. Few selected examples are shown in the Fig. 2.1.
Several computational and spectroscopic studies have been devoted to mimic
and understand such interactions at the molecular level [18, 41–43, 88–99]. The
matrix isolated FTIR spectroscopy in combination with ab initio calculation provides
valuable information to categories and characterize SCHBs. Maes and coworkers
used the extent of the red shift of H−Cl frequency in 1:1 complexes of HCl and YR2
(Y = O, S, Se and R = CH3 , C2 H5 ) as the measure of the strength of H-bond, more is
the red shift stronger is the H-bond [90]. The red shift in H-Cl stretching frequency is
higher for S than O acceptors and it correlates with the proton affinity of the acceptor.
Similar studies have been performed for intermolecular and intramolecular H-bonds
involving S as an acceptor and various proton donors such as H2 O, HF, HCl, HNO2 ,
HNO3 , and CF3 CCH [42, 43]. The O–H· · · S H-bond between the OH group of the
HNO3 and sulfur atoms of dimethylsulfide (DMS), dimethyldisulfide (DMDS) and
H2 S was observed. In these binary complexes O–H· · · S H-bond strength increases
in the order as H2 S–HNO3 , DMDS–HNO3 and DMS–HNO3 , suggesting that the
sulfur atom of DMS is the strongest H-bond acceptor among them. The conclusion is
purely based on the frequency shift in the O–H stretching frequencies of HNO3 . One
of the inherent problems in the matrix isolation studies is the effect of matrix on the
O–H frequencies, which prohibits precise determination of vibrational frequencies.
Computational efforts have also been put forward to determine the H-bond strength
of O and S acceptors. Wennmohs et al. [52] computed the interaction energies of the
DMS-MeOH, DME-MeOH and dimethylthiocarbonyl–MeOH at the CCSD (T)/aug-
cc-pVQZ level of theory. The interaction energies are − 5.46, − 5.97 and − 5.33
kcal/mol respectively. It suggests that the O–H· · · S H-bond strength is very similar
to that of O–H· · · O H-bond strength in these complexes. The authors also found that
in case of O–H· · · S H-bonding the dispersion energy contribution is about ∼ 70 %
to the total interaction energy. However, the electrostatic component controls the
H-bond geometry. The theoretical predication of S as potential H-bond acceptor as
O in DMS-MeOH and DME-MeOH complexes was confirmed experimentally by
Daryl L. Howard and Henrik G. Kjaergaard [99]. From the vapor phase infrared
spectroscopic study of the complexes of MeOH it was claimed that S is nearly
equivalent to O as H-bond acceptor. The problem with the vapor phase IR spectra
is that they are very broad and no control over mass selectivity. Hence it is very
difficult to obtain precise information about the shift in O−H vibrational frequency
in DMS-MeOH and DME-MeOH complexes.
The importance of SCHBs and some debatable conceptions about SCHBs
prompted us to study them systematically. A combined experimental and compu-
tational approach has been deployed to extract molecular level information about
SCHBs. The computational methods and experimental techniques are complemen-
tary to each other. For example computed vibrational frequencies and structure of
molecules are very helpful in assigning the experimental spectral features while the
experimental data can be used for bench marking and developing computational
methods. To mimic the O–H· · · S interaction between the side chain of tyrosine
and S of methionine, para-cresol (p-CR) and thioethers are taken as the respective
model compounds. As shown in Fig. 2.2, the alkyl chain length of the thioethers is
systematically varied to see its effect on the nature and strength of the SCHBs.
Laser spectroscopy of jet-cooled molecules and molecular clusters is very handy
to extract precise experimental data on non-covalent bonded clusters, in this case
the H-bonded clusters. In most of the cases double resonance spectroscopy such as
fluorescence dips infrared (FDIR) spectroscopy and resonant ion dip infrared spec-
troscopy (RIDIR) coupled with mass spectrometry are used to get the X–H stretching
frequencies (here, O–H, N–H, S–H stretching frequencies) of the monomers and their
H-bonded complexes. The relative changes in the X-H stretch in the complexes are
Fig. 2.2 Model compounds used to study O–H· · · S H-bond between OH group of tyrosine and
sulfur atoms of methionine and cysteine
Fig. 2.3 FDIR spectra of 1:1 complexes of p-CR and YHR (Y = O, S and R = H, CH3 , and C2 H5 ).
H2 O, MeOH and EtOH are used as “O” H-bond acceptors and H2 S and EtSH are used as “S”
H-bond acceptors
generally taken as a measure of the H-bond strength. It is a general practice for the
conventional H-bond that more red shift in the X–H stretch, stronger is the X–H· · ·Y
H-bond. The X–H· · ·Y H-bonds becomes stronger if Y becomes more basic.
Figure 2.3 presents the FDIR spectra of 1:1 complexes of p-CR and YHR (Y = O,
S and R = H, CH3 , and C2 H5 ). Here the OH group of p-CR acts as a H-bond donor
and alcohols and thiols as H-bond acceptor.
The OH stretching frequencies of p-CR are red shifted in the complexes irre-
spective of the acceptor atoms, suggesting formation of O–H· · · O(S) H-bonds.
The formations of O–H· · · O(S) H-bonds are also confirmed by DFT calculations
[100, 101]. If one goes by the red shift of OH stretch, it is very clear from the
figure that O of H2 O and alcohols are better H-bond acceptors than the S of H2 S
and corresponding thiols. There are multiple conformers observed experimentally.
In such situations computed vibrational frequencies are very helpful in assigning

the spectra. The computed H-bond energy and vibrational frequencies along with
the experimental frequencies are tabulated in Table 2.1. As one can see the H-bond
energy of O–H· · · S H-bonds are smaller compared to that of O–H· · · O H-bond.
However, with increasing alkyl chain length in the acceptor moiety the O–H· · · S
H-bond becomes stronger. One more thing can be noticed from the table; the gas
phase basicities of alcohols are smaller than their corresponding thiols; however the
red shifts of O–H stretching frequencies for thiols as H-bond acceptors are less com-
pared to their oxygen counterpart, suggesting that SCHB does not follow the acid
base formalism.
The situation is little different in case of ethers and thioethers as H-bond acceptor.
The FDIR spectra of 1:1 complexes of p-CR with ethers and thioethers are shown in
Fig. 2.4.
The red shifts of O− H stretching frequency of p-CR•DES is larger compared
to that of p-CR•DEE, suggesting the acyclic thioethers can be better H-bond ac-
ceptors than the ethers [102, 103]. Comparing the H-bond acceptor quality of cyclic
thioether and cyclic ether, the cyclic ether is found to be better acceptor than its sulfur
counterpart. Thioethers are better H-bond acceptors than the thiols [104]. Multiple
conformers are observed for the ethers and thioether complexes [102–104]. In those
cases DFT calculations are very useful to assign the experimental spectral features.
Two conformers in the complex of cyclic ether, tetrahydrothiophene (THT), with
p-CR are found experimentally. The two O–H stretching frequencies are observed
for p-CR•THT complex while probing different electronic transitions. The FDIR
spectra of p-CR•THT are shown in Fig. 2.5 with C2 and Cs symmetry in THT
These two conformers have different electronic transitions, but same IR transitions.
In this case it is difficult to assign the IR spectra without the help of quantum chem-
ical calculations. The non-covalent interaction (NCI) plots [105, 106] and atoms
in molecules (AIM)[107–111] molecular graph assist the assignment. As shown in
the figure conformer A of p-CR•THT does not show any vibronic coupled band in
the FDIR spectra, while conformer B shows a strong vibronic coupled band. The
vibronic couple band arises because of the coupling between the O− H oscillator
of p-CR and C–H oscillator of THT. This is in fact happens for the p-CRTHT (Cs)
conformer as predicted by NCI as well as AIM topology.
The binding and the red shift in O−H in the H-bonded complexes are the net
result of several fundamental interactions, such as charge-dipole, dipole-dipole,
dipole-induced dipole and the higher order multipole interactions. The magnitudes
of these interactions are very difficult to determine experimentally. However theo-
retical insights are useful to account for the individual energy contribution to the
total interaction energy. Various energy decomposition analyses such as. Natural
energy decomposition analysis (NEDA) [112–114], Kitaura and Morokuma (KM)
[115, 116], and reduced variational space self-consistent field (RVS) [117] decom-
position analyses are used to solve the purpose. It is noticed that dispersion energy
component is very important in SCHBs [100–104]. Table 2.2 summarizes dispersion
energy contribution to the total interaction energy in the O–H· · · O and O–H· · · S
H-bonded intermolecular complexes.
2
Table 2.1 Experimental red shifts of O–H stretching frequencies, red shifts of band origin (BO) positions, gas phase basicity of acceptors (GB), computed O–H
shifts, changes in the O–H bond lengths (r) and bond dissociation energies (D0 ) of p-CR•YRR’ (Y = O and S, R(R’) = H, CH3 and C2 H5 ) complexes
Species O–H Shift BO shift GB in kJ/mol O–H shift (B3LYP/aDZ) r (B3LYP/aDZ) BEa (MP2/aDZ)
(Expt) in cm−1 (Expt) in cm−1 in cm−1 in Å in kcal/mol
“O” Acceptors p-CR•H2 O − 127 − 357.0 660.0 − 192 0.0095 − 4.29
p-CR•MeOH − 191 − 425.0 724.5 − 246 0.0120 − 5.76
Hydrogen Bonds Involving Sulfur
p-CR•EtOH − 217 − 422.4 746.0 − 256 0.0125 −6.33

p-CR•DEE − 280 − 451.3 801.0 − 292 0.0140 −7.67
p-CR•THF − 315 − 482.0 794.7 − 335 0.0160 −7.42
“S” Acceptors p-CR•H2 S − 102 − 239.0 673.0 − 138 0.0064 −2.49
p-CR•MeSH 742.0 − 210 0.0096 −4.08
p-CR•EtSH − 174 − 316.3 758.4 − 219 0.0102 −4.70
p-CR•DES − 292 − 384.5 827.6 − 299 0.0146 −6.23
p-CR•THT − 285 − 373.0 819.3 − 296 0.0142 −6.15
a
Bond dissociation energies (D0 ) are computed at MP2/aug-cc-pVDZ//B3LYP-aug-cc-pVDZ level of theory, the basis set superposition error (BSSE) and zero
point energy (ZPE) corrections are also applied. aDZ: aug-cc-pVDZ, DME dimethylether, DEE diethylether, THF tetrahydrofuran, DMS dimethylsulfide, DES
diethylsulfide, and THT tetrahydrothiophene
23
Fig. 2.4 FDIR spectra of 1:1 complexes of p-CR and YR (Y = O, S and R = H, CH3 , and C2 H5 ).
H2 O, DEE and THT are used as “O” H-bond acceptors and H2 S DES, and THT are used as “S”
H-bond acceptors
Fig. 2.5 Left panel: FDIR spectra of p-CR•THT conformers probed at their respective band ori-
gins. The stick diagrams in the spectra are the computed OH stretching frequencies obtained at
B3LYP/aug-cc-pVDZ level of theory. Middle panel: Non-covalent interaction (NCI) plot of two
conformers of p-CR•THT, showing O–H· · · S interactions in both the conformers and C–H· · · O
interaction only in p-CR•THT (Cs) conformer. Right panel: Molecular graph of generated using
atoms in molecules (AIM) electron density topology. Bond critical points are located for O–H· · · S
and C–H· · · O interactions
In general the O–H· · · S HBs are dispersive in nature while O–H· · · O HBs are
electrostatic. However, the dispersion energy contribution in the O–H· · · O HB case
increases with the increase of alkyl chain length of the hydrogen bond acceptors. In
the case of the O–H· · · S HB, the dispersion energy contribution decreases from H2 S
2
Table 2.2 Total interaction energy without the addition of ZPE (kcal/mol) at the MP2 level (EMP2 int ), according to the NEDA, KM, and RVS energy
decomposition analyses (Eint ), and dispersion energy (Edisp ) for the X = O vs S complexes. Numbers in parenthesis denote the percentage of contribution
of dispersion interaction to the total interaction energy. In all the cases single point energy calculation was done at the MP2/aug-cc-pvDZ for the respective
B3LYP/aug-cc-pvDZ optimized structures
Species EMP2
int Eint Edisp
KM RVS NEDA KM RVS NEDA
“O” Acceptors p-CR•H2 O − 6.19 − 4.77 − 4.87 − 4.77 − 1.42 (23 %) − 1.32 (21 %) − 1.42 (23 %)s
Hydrogen Bonds Involving Sulfur
p-CR•MeOH − 7.25 − 5.20 − 5.34 − 5.20 − 2.05 (28 %) − 1.91 (26 %) − 2.05 (28 %)
p-CR•EtOH (anti) − 7.79 − 5.27 − 5.44 − 5.27 − 2.52 (32 %) − 2.35 (30 %) − 2.52 (32 %)
p-CR•DME − 7.91 − 5.21 − 5.21 − 2.70 (34 %) − 2.70 (34 %)
p-CR•DEE (TT) − 9.26 − 4.99 − 5.28 − 4.99 − 4.27 (46 %) − 3.98 (43 %) − 4.27 (46 %)
p-CR•THF − 9.02 − 6.65 − 6.27 − 6.04 − 2.37 (26 %) − 2.75 (30 %) − 2.98 (33 %)
“S” Acceptors p-CR•H2 S − 3.76 − 1.19 − 1.23 − 1.19 − 2.57 (68 %) − 2.53 (67 %) − 2.57 (68 %)
p-CR•MeSH − 5.16 − 2.40 − 2.61 − 2.40 − 2.76 (53 %) − 2.55 (49 %) − 2.76 (53 %)
p-CR•EtSH (anti) − 5.74 − 2.64 − 2.84 − 2.64 − 3.10 (54 %) − 2.90 (51 %) − 3.10 (54 %)
p-CR•DMS − 6.40 − 3.05 − 3.05 − 3.35 (52 %) − 3.35 (52 %)
p-CR•DES (TT) − 7.30 − 3.30 − 3.57 − 3.30 − 4.00 (55 %) − 3.73 (51 %) − 4.00 (55 %)
p-CR•THT − 7.20 − 3.98 − 3.60 − 3.37 − 3.22 (45 %) − 3.60 (50 %) − 3.83 (53 %)
25
Fig. 2.6 The correlation plots of red shift in O−H frequency (νO−H ) vs proton affinity (PA) for the
O−H· · · O and O−H· · · S bound complexes. (Reprinted with permission from ref 121. Copyright
2013 American Chemical Society)
to MeSH (from 68 to 53 %), but remains unchanged with further increase of alkyl
chain length of the hydrogen bond acceptors.
Another aspect about SCHBs, is to look for H-bond descriptor. The H-bond
descriptors are the computational and experimental molecular properties that are used
to predict the H-bond strength. One such parameter is the gas phase proton affinity
(PA) and/or gas phase basicity of the H-bond donor. It is expected that for a particular
H-bond donor, H-bond acceptors with higher PA will form stronger H-bonds. PA has
been a well established H-bond descriptor for the conventional O–H· · · O H-bonded
systems [118–120] where electrostatic interaction is dominant. For example a good
correlation between the red shift in O–H frequency of phenol with the PA of acceptor
bases has been observed [118, 119]. Very recently Bhattacharyya et al. [121] explored
the possibility of extending the acid-bas formalism to the O–H· · · S H-bonds. para-
Fluorophenol (FP) was considered as H-bond donor for several S containing solvents
of varying PA such as, H2 S, Me2 S, MeSH, and THT. The correlation plots between
the experimentally determined red shifts of O–H stretching frequencies and PA of O
and S acceptors are shown in the Fig. 2.6.
Excellent linear correlations are observed between red shift of O−H stretching
frequency (νO−H ) and PA for the individual O and S acceptor groups. However,
Fig. 2.7 Correlation plot of (Ect + Eex ) (kcal/mol) vs red shift in O−H frequency (νO−H ) for the
O−H· · · O and O−H· · · S bound complexes. (Reprinted with permission from ref 121. Copyright
there is no linear correlation exist between the red shift of O−H stretching frequency
(νO−H ) and PA if one combines both the acceptor group members. Nevertheless
these correlations are useful in predicting νO−H from the PA of the H-bond acceptor,
e. g. νO−H of FP-MeSH can be predicted to be 202 cm−1 from the PA of MeSH
184.8 kcal mol−1 . They also observed very good correlation between the dissociation
energies (D0 ) and νO−H for the individual O and S acceptor groups. The energy
decomposition analysis showed that irrespective of O and S acceptor groups there
is a unified correlation exist between the sum of the charge transfer (Ect ) and the
exchange (Eex ) energy component of the total binding energy and red shift of O−H
stretching frequency. The correlation graph is depicted in Fig. 2.7.
The major challenge is to obtain the interaction energy of O–H· · · S H-bond exper-
imentally. Mass analyzed threshold ionization (MATI) spectroscopy [122–129] and
zero kinetic energy photoelectron spectroscopy (ZEKE-PES) [130–134] are the two
popular laser spectroscopy techniques used to obtain the bond dissociation energies
(D0 ) for H-bonded complexes, but the success rates are limited. Few attempts have
been made by our group to get the experimental D0 values of O–H· · · S H-bonded
complexes. These experimental D0 will be used for benchmarking the quantum
chemical methods.
Met- 1
Met- 233
Met- 149 His- 43
Trp- 223
Tyr-617
Bacteriochlorophyll Human thioredoxins Galactose oxidase

( PDB: 1bcl) ( PDB: 1erv) (PDB: 1gof)
Fig. 2.8 N–H· · · S H-bonds involving sulfur atoms of methionine and N–H groups in proteins
2.2.2 N–H· · · S Hydrogen Bonding
The sulfur atom of methionine also forms H-bonds with the N–H donor groups.
Figure 2.8 displays some of the cases where methionine sulfur is involved in H-
bonding with side chain N–H of tryptophan and histidine and backbone amide N–H.
One of the exhaustive analysis of N–H· · · S H-bond was reported by Wategaonkar
and co workers [135]. This is one of the very first reports where both the computa-
tional and experimental techniques are used to explain the nature of N–H· · · S H-bond
between NH of indole and S of DMS. In fact, indole is used as the model compound
of tryptophan and to probe local environment and dynamics of proteins in solution
[136–138]. The IR spectra of H-bonded complexes of indole (IND) with various
H-bond acceptors are shown in Fig. 2.9. The highest red shift of N–H stretching fre-
quency is observed for IND-DMS complex. The red shift of N–H for this complex
is the highest among other complexes studied. The red shift for N–H· · · S H-bond is
very similar to N–H· · · O H-bond in IND-DME complex, but almost two times larger
than that for N–H· · · O H-bond in IND-H2 O complex and ∼ 3.5 times larger than
N–H· · · π H-bond in IND-Benzene complex [139]. If one goes by the red shift of
N–H stretching frequency, then N–H· · · S H-bond should be considered the strongest
H-bond among all other H-bonds shown in the figure.
The bond dissociation energies (D0 ) of indole-DMS and 3-methylindole-DMS
complexes are determined at the MP2/CBS and compared with the experimental
determined D0 of IND-H2 O and IND-Bz [135]. The D0 of IND-DMS at MP2/CBS
is − 5.59 kcal/mol which is little higher than that of IND-H2 O (− 4.67 kcal/mol) [122]
and IND-Bz (− 5.21 kcal/mol) [140]. Similar D0 values and very different N–H red
shift in N–H· · · S, N–H· · · O and N–H· · · π H-bond complexes urge us to investigate
the nature of these three different type H-bonds. It has already been established
that X–H· · · π H-bonds are dispersive in nature [141–145]. Energy decomposition
analysises of IND-DMS and IND-H2 O complexes suggest that without dispersion
interaction N–H· · · S H-bonded complexes are not stable complexes. The dispersion
energy is about 115 % of the total interaction energy [135]. It can be inferred that the
nature of N–H· · · S is very different from N–H· · · O H-bond and somewhat similar to
Fig. 2.9 FDIR spectra of H-bonded complexes of indole with various H-bond acceptors. IND
Indole, DMS Dimethylsulfide, DME Dimethylether, Bz Benzene
N–H· · · π H-bond. The correlation plots of band origin shift (δ) versus polarizabilty
(α) and proton affinity (PA) for different types of N–H· · ·Y H-bonded and van der
Waals complexes of indole as shown in Fig. 2.10 also confirms that IND-DMS can
be categorized with the IND-Bz and van der Waals complexes.
Going by the correlation presented above, one can put N–H· · · S H-bond and N–
H· · · π H-bond in the same class, but the 3.5 times larger red shift of N–H stretching
frequency in IND-DMS than in IND-Bz is completely unexplained by this corre-
lations. One can safely say that N–H· · · S H-bond (in IND-DMS) is very different
from the conventional N–H· · · O (in IND-H2 O), N–H· · · N (in IND-NH3 ) H-bonds
and unconventional N–H· · · π (in IND-Bz) H-bond, but how N–H· · · S H-bond is
different from others is a challenging topic of further explorations both by theories
and experiments.
The intra and intermolecular amide N–H· · · S H-bonds are frequently observed
in the crystals of proteins [57, 146, 147] and simple organic molecules [58, 65, 148,
149]. The H-bond length of N–H· · · S H-bond varies from system to system and it
depends on the spatial arrangement of the crystals and the orientation of amide N–H
towards sulfur atom. By considering H-bond distances in most of these cases the
Fig. 2.10 The correlations plot band origin shift (δ) versus polarizabilty (α) and proton affinity (PA)
for different types of N−H· · ·Y hydrogen-bonded and van der Waals complexes of indole. a The
band origin shift (δ) versus polarizabilty (α). b the band origin shift (δ) versus proton affinity (PA)
plots for all of the complexes. c the band origin shift (δ) versus polarizabilty (α) for the van der Waals
complexes and M·Me2 S, M·H2 S, and M·benzene. d the band origin shift (δ) versus proton affinity
(PA) for the N−H· · · O and N−H· · · N type HB complexes and M·Me2 S, M·H2 S, and M·benzene.
(Reprinted with permission from ref 135. Copyright 2009 American Chemical Society)
amide N–H· · · S H-bond is regarded as a weak H-bond like N–H· · · π H-bond, even
if the statistical analysis of methionine and cysteine containing proteins supports
N–H· · · S H-bond as a weak H-bond [78]. There are hardly any comprehensive
study of amide N–H· · · S H-bond using laser spectroscopy and high level ab intio
calculations. Recently Mons and co-workers studied amide N−H· · · S H-bonds in
a systematic manner using methionine containing model peptides [150]. The model
peptides are capped dipeptides of phenylalanine and methionine residues. They are
N-acetyl-Lphenylalaninyl-L-methionine-amide (FM) and N-acetyl-L-methioninyl-
L-phenylalanine-amide (MF) as shown in Fig. 2.10. In these peptides a variety of
H-bonds can be possible with the amide N−H. The amide N−H forms N–H· · · O = C
H-bond with the back bone carbonyls and it provides different folding patterns and
structures to the peptides. The N–H· · · S and N–H· · · π H-bonds are also formed
with the side chain sulfur atom of methionine and π electron cloud of phenylalanine,
respectively. Hence, these peptides are ideal peptides to study the N–H· · · S H-bonds
CH3
S
tr
O H O O H O
N H N H
H3C N N H3C N N
H H
H O H O
S
H3C
a Ac-Phe-Met-NH2 (FM) Ac-Met-Phe-NH2 (MF)

Backbone H-Bond Paern Side chain H-Bond Paern
Fully Extended Form
C5 CH3
O R H O N-H•••O H-bond S C10
N H C6 C7
H3C N N O H O
Two C5 (β-strand like structure)
H O R H N H
C5 H3C N N
H
H O
Parally Folded Form
O R H O N-H•••S H -bond
N H CH3
H3C N N S
One C5 + One γ-turn
H O R H
C5 C7 O H O
N H
Fully Folded Form H3C N N
H
C7 H O
O R H O O R H O
N H N H
H3C N N H3C N N
H O R H H O R H N-H•••π H-bond
C7 C10
One β-turn (C10)
b Two γ-turns (C7)
Fig. 2.11 a Model peptides, N-acetyl-Lphenylalaninyl-L-methionine-amide (FM) and N-acetyl-

L-methioninyl-L-phenylalanine-amide (MF). b Left Panel: Selected sterically allowed backbone
H-bond patterns. Right panel: Local H-bonding patterns of the side-chain of methionine (N–H· · · S
H-bond), or phenylalanine (N–H· · · π H-bond). The H-bond patterns are labelled Cn , according to
the “n” number of atoms present in the ring formed by the H-bond
and compare them with other type H-bonds existing in the same system. As shown
in Fig. 2.11b, the sulfur atom of methionine can either form an intra residue five
membered (C5 ) N–H· · · S H-bond with its own residue or inter six membered (C6 ) and
ten membered (C10 ) N–H· · · S H-bonds with “i” and “i + 1” residues, respectively.
The IR-UV double resonance spectroscopy was employed to get the IR spectral
signature of these peptides in the N–H stretch region. Since, these peptides are
studied in isolated gas phase conditions, the effect of solvent environment and crystal
packing pattern on N–H· · · S, N–H· · · O = C, and N–H· · · π H-bonds are avoided.
In most of the proteins the methionine residues reside in the hydrophobic core of
Ac-Phe-Met-NH2
(β –turn Type I )
NH---S C10 COAGULATION FACTOR VIII PRECURSOR

π NH2
a d Phe-2200 PDB: 1D7P
Ac-Met-Phe-NH2 (A)
(β –turn Type I )
Met-2199
NH---S HYPOXANTHINE-GUANINE
C10 π NH2 PHOSPHORIBOSYLTRANSFERASE
b e PDB: 1QK5
Ac-Met-Phe-NH2 (B) Met-210
(β –strand + γ-turn)
Phe-211
NH---S
C5
c C7 NH2 f
3350 3400 3450 3500
Wavenumber /cm-1
Fig. 2.12 Left Panel: Experimental infrared spectra versus B97-D/TZVPP computed vibrational
frequencies of a the main conformer of FM and b, c, the stick diagrams represent the computed N–H
frequencies. Middle panel: The H-bond pattern and structures of conformers of FM and MF whose
computed N–H stretching frequencies best fit the experimental ones. Right panel: The H-bond
parameters and examples of locally folded methionine residues in proteins as obtained from X-ray
diffraction crystal. (Reprinted with permission from ref 150. Copyright 2012 American Chemical
Society)
the protein. Hence, studying them by gas phase spectroscopy without any solvent
(very low dielectric medium) will allow us to mimic the environment of methionine
in proteins.
The experimental IR spectra and the computed N–H stretching frequencies of
assigned conformers of FM and MF at B97-D/TZVPP level are shown in Fig. 2.12.
The most red shifted N–H stretch is observed for N–H· · · S H-bond. The shift is
even stronger than that observed for N–H· · · O = C H-bonds. It is really surprising
that sulfur can form very strong H-bonds as opposed to earlier observation that it is
as weak as N–H· · · π H-bonds. The B97-D methods are very helpful in computing
the vibrational frequencies of a large number of conformers of the model peptides
and those frequencies are used to assign the correct conformer to experimental IR
spectra. The folding pattern as observed in these small model peptides are also found
in proteins. Two of the examples are shown in the Fig. 2.11. The major challenge
is to determine the intramolecular N–H· · · S, N–H· · · O = C, and N–H· · · π H-bond
energies. The donor-acceptor overlap energies and electron density at the bond crit-
ical point as obtained from NBO and AIM calculations do not correlate with the
experimental observed red shifts of N–H stretching frequencies. The donor-acceptor
overlap energies and electron density at the bond critical point for N–H· · · O = C
Cys- 263
Cys- 49
Ser-436
His- 460
Human methionine aminopeptidase-2 Bence-Jones protein

( PDB: 1b6a) ( PDB: 1bfd)
S-H•••N H-bond S-H•••O H-bond
Cys- 182
Cys- 302
Cys-13
Trp- 171
Formate dehydrogenase. Human kinesin motor domain

( PDB: 2nac) ( PDB: 1bg2)
S-H•••π H-bond S-H•••S H-bond
Fig. 2.13 S–H· · · N, S–H· · · O, S–H· · · S, and S–H· · · π H-bonds involving S–H group of cysteine
H-bonds are always higher than for N–H· · · S H-bonds. Molecular tailoring [151]
and fragmented molecular orbital [152] methods may be useful in getting the exact
intramolecular N–H· · · S H-bond energies.
2.3 S–H as a H-Bond Donor
Sulfur is not only a potential H-bond acceptor forming N–H· · · S and O–H· · · S
H-bonds as strong as N–H· · · O and O–H· · · O H-bonds, S–H also participates as
H-bond donor with many different types of acceptors. From a detailed survey of
the protein structure data bank it is observed that S–H group of cysteine can make
S–H· · · N, S–H· · · O, S–H· · · S, and S–H· · · π H-bonds. Few examples of SCHBs
involving cysteine S–H group are shown in Fig. 2.13.
The simplest example where S–H acts as an H-bond acceptor is the H2 S dimer.
The structural similarity between H2 O and H2 S has attracted many researchers to
compare their molecular properties. One of them is the H-bond acceptor and donor
potentialities of H2 O and H2 S [53, 153–160]. There are many reports on the ma-
trix isolated IR spectra of H2 S dimer. The S–H stretching frequencies varies with
the matrix; e.g. the symmetric H-bonded S–H stretching frequencies (ν1 ) are ob-
served at 2580.3, 2567, 2569.5, and 2566.4 cm−1 for N2 [161], CO, Ar, and Kr [162]
matrices respectively. The matrix effect can be avoided by studying those dimers
in supersonic-jet condition with mass selection. Very recently, the H2 S dimer was
studied by Wategaonkar’s group using VUV ionization-detected IR predissociation
spectroscopy (VUV-ID-IRPDS) [159]. The symmetric H-bonded S–H stretching
frequencies(ν1 ) of H2 S dimer was observed at 2590 cm−1 , which is about 24 cm−1
red shifted compared to that of the monomer [163]. On the other hand red shift for
H2 O dimer is 56 cm−1 which is ∼ 2.5 times of that of H2 S dimer. The binding energy
of the H2 S dimer was estimated at the MP2/CBS limit level and using the Helgaker
two-point extrapolation formula [164]. It is found to be − 0.97 kcal mol−1 . The bind-
ing energy of H2 S dimer is one third of the binding energy of H2 O dimer (− 3.16 kcal
mol−1 ) [165]. The red shift of X-H stretching frequencies observed experimentally
correlate well with computed binding energies. An energy decomposition analysis of
these two complexes suggests that H2 S dimer is predominantly stabilized by the dis-
persion interaction where as electrostatic energy component is the major contributor
for the H2 O dimer binding energy. The authors [159] extended their study further to
S–H· · · O H-bonded complex between H2 S and MeOH using VUV-ID-IRPDS and
computational methods. In this case H2 S prefers to act as a H-bond donor rather than
H-bond acceptor. This is evidenced as the S–H· · · O H-bonded complex between
H2 S and MeOH is not observed experimentally. The symmetric H-bonded S–H (ν1 )
stretching frequencies of H2 S-MeOH complex is observed at 2577 cm−1 which is
37 cm−1 red shifted compared to the monomer stretching frequencies. The red shift
is larger compared to the H2 S–H2 S case. The binding energy of H2 S-MeOH (S–
H· · · O H-bonded dimer) complex at MP2/CBS level is − 2.65 kcal mol−1 which is
0.21 kcal mol−1 higher than that of the O–H· · · S H-bonded dimer. The difference
in the stability comes from the electrostatic component. The electrostatic interaction
in S–H· · · O H-bonded dimer is more than in the O–H· · · O H-bonded dimer. The
computed binding energy of S–H· · · O = C H-bond (∼-4 kcal mol−1 ) [166] is com-
parable with that of N–H· · · S H-bonds and stronger than C–H· · · O and C–H· · · S
H-bonds.
The S–H· · · N H-bond between H2 S and different amines has been characterized
by Scheiner and coworkers [167] using ab intio methods. The bond dissociation
energies (D0 ) of H2 S-NH3 and H2 S-N (CH3 )3 are − 1.76 and − 3.55 kcal mol−1 ,
respectively. These values are quite larger compared to the bond dissociation energy
of H2 S dimer. The complexes are linear in structure with H-bond angle (θ) very
close to 180◦ . The S–H bond length of H2 S increases in all the complexes and red
shifts of the S–H stretching frequencies are observed. In another report Scheiner
and coworkers mentioned blue-shifted S–H· · · N and S–H· · · P H-bonds [168]. Here
the S–H of SHN is the H-bond donor and amines and phosphines are the H-bond
acceptors. The maximum blue shift (110 cm−1 ) was observed NSH· · · NH3 complex.
Such a large blue shift is really surprising and needs further rather experimental
evidences.
The S–H· · · π H-bonds involving the S–H group of cysteine and π electron den-
sity of aromatic amino acids are very important as they are frequently observed
in proteins [34, 76, 138, 166, 169, 170]. Most commonly observed X–H· · · π H-
bonds in nature are O–H· · · π, N–H· · · π, and C–H· · · π interactions and they are
also vital in predicting and dictating structural and conformational preferences of
biomolecules [138, 141, 142, 144, 171–173]. Many efforts have been made to char-
acterize S–H· · · π H-bond and compare it with other X–H· · · π H-bonds. In most of
the cases benzene and H2 S have been considered as the model compounds to study
such H-bonds [50, 145, 155, 157, 174, 175].
We reported first ever IR spectroscopic studies of the S–H· · · π H-bonds in the
gas phase in complexes of H2 S with indole (IND) and 3-methylindole (3-MI) [176].
The red shift in the band origin of the IND–H2 S (44 cm−1 ) and 3-MI–H2 S (71 cm−1 )
complexes were found to be much smaller than those of their respective H2 O com-
plexes (132 and 239 cm−1 ) which are N–H· · · O H-bonded complexes. On the other
hand band origin shift of IND–H2 S is very similar to that of the IND–CH4 complex
(76 cm−1 ) which is a C–H· · · π H-bonded complex [177]. This provides an indirect
evidence of H2 S forming S–H· · · π H-bond with indole and 3-MI. The direct evi-
dences come from the IR spectra of the complexes (Fig. 2.14). The N–H stretching
frequency is almost unchanged upon the complex formation, thereby indicating that
the N–H stretch remains free in the complex. At the same time both complexes fur-
nished an H-bonded symmetric S–H stretching frequency at 2593 and 2589 cm−1 ,
respectively. The H-bonded S–H stretch for IND–H2 S was red shifted by 21 cm−1 ,
while that for the 3-MI–H2 S was red shifted by 25 cm−1 . This gives a confirmatory
inference that S–H group of H2 S prefers to act as an H-bond donor to the π electron
density of IND and 3-MI.
The bond dissociation energies of IND-H2 S and 3-MI-H2 S (π-type H-bonded
complexes) were calculated at the MP2/CBS and compared with those of IND-H2 O
and 3-MI-H2 O complexes which form σ-type H-bonds. Table 2.3 summarizes the
bond dissociation energies, band origin shift and S(O)-H stretching frequency shifts
for the above mentioned complexes.
The S–H· · · π H-bond energies of both the complexes are larger than the N–H· · · O
H-bond energies of IND.H2 O and 3-MI-H2 O complexes. Moreover, comparison of
the computed binding energies of the π-type H-bonded complexes of IND and ben-
zene with H2 O, H2 S, NH3 , and CH4 showed that the X–H· · · π H-bond energies
lie in the order as S–H· · · π > O–H· · · π > N–H· · · π > C–H· · · π. H-bonds [176].
Energy decomposition analysis of the H-bonded complexes with IND revealed the
net electrostatic components to lie in the order H2 O ≈ H2 S > NH3 > CH4 whereas
the dispersion contribution was found to follow the order H2 S > CH4 > NH3 > H2 O.
Very recently attempts have been made by Boxer and coworkers to quantify the elec-
trostatic contribution in X–H· · · π H-bonds using vibrational Stark spectroscopy
in combination with DFT calculations [178, 179]. It is observed that the red
shifts of the H-bonded O H stretching frequencies in O–H(phenol)· · · π H-bonds
correlate linearly with the strength of the applied electric fields indicating that
they are primarily governed by electrostatic interactions. However in the case of
N–H(indole)· · · π and S–H(thiophenol)· · · π H-bonds deviations from the linear
IND
3525
IND-H2S
3523
2593
3MI
3526
3MI-H2S
3523
2589
2550 2600 3350 3400 3450 3500 3550 3600

Wavenumber / cm-1
Fig. 2.14 FDIR spectra of indole (IND) and 3-methylindole (3-MI) and their H-bonded com-
plexes with H2 S probing at their respective band origins. (Reprinted with permission from ref 176.
Copyright 2009 American Chemical Society)
Table 2.3 Experimentally observed parameters such as XH stretching frequency shift, the band
origin shift, The S1 state lifetime, and binding energy of indole·L and 3-MI·L (L = H2 O and H2 S)
σ- and π-type hydrogen-bonded complexes. (Reprinted with permission from ref 176. Copyright
Complex HB type νNH νXH EBO τ E0
(cm−1 ) (cm−1 )X = O,S (cm−1 ) (ns) (kcal/mol)
IND·H2 O N−H· · · O σ HB − 89 − 5a − 132 22.8 − 4.67a
IND·H2 S S–H· · · π π HB −2 − 21 − 44 11.6 − 4.89b
3-MI·H2 O N−H· · · O σ HB − 84 − 5a − 239 14.8 − 4.49a
3-MI·H2 S S–H· · · π π HB −3 − 25 − 71 7.9 − 5.17b
a
Experimental value from ref 124
b
Binding energy computed at the CBS limit
correlations were observed. These observations were found to be consistent with the
atomic polarizabilities of the associated X–H groups and hence can be attributed to
the dispersion dominance in N–H (indole)· · · π and S–H(thiophenol)· · · π H-bonds.
Acknowledgment I am very much thankful to Prof. Sanjay Wategaonkar and Dr. Michel Mons,
who introduced me to the exciting field of laser spectroscopy. My special thanks to Dr. Rudresh
Acharya for stimulating discussions and for his help in rendering figures for SCHBs in protein
structures. I am also grateful to my coworkers, students and authors of the cited references who
have contributed to this work in many ways. This work is financially supported by DST-Inspire
faculty fellowship, Department of Science and Technology (DST, Govt. of India) and National
Institute of Science Education and Research, (Department of Atomic Energy, DAE, Govt. of India).
References
1. Arunan E, Desiraju GR, Klein RA, Sadlej J, Scheiner S, Alkorta I, Clary DC, Crabtree RH,
Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ (2011) Defining
the hydrogen bond: an account (IUPAC Technical Report). Pure Appl Chem 83(8):1619–1636
2. Pauling L (1960) The nature of the chemical bond, 3rd edn. Cornell University Press, Ithaca
3. Pimentel GC, McClellan AL (1960) The hydrogen bond. W. H. Freeman and Company, New
York
4. Jeffrey GA, Saenger W (1991) Hydrogen bonding in biological structures. Springer-Verlag,
Berlin
5. Jeffrey G A (1997) An introduction to hydrogen bonding. Oxford University Press, New York
6. Scheiner S (1997) Hydrogen bonding: a theoretical perspective. Oxford University Press,
New York
7. Desiraju GR, Steiner T (1999) The weak hydrogen bond in structural chemistry and biology.
Oxford University Press, New York
8. Grabowski SJ (2006) Hydrogen bonding—new insights challenges and advances in compu-
tational chemistry and physics, vol 3, 1st edn. Springer, New York
9. Gilli G, Gilli P (2009) The nature of the hydrogen bond: outline of a comprehensive hydrogen
bond theory. Oxford University Press, New York
10. Hobza P, Müller-Dethlefs K (2010) Non-covalent interactions: theory and experiment. Royal
Society of Chemistry, Cambridge
11. Dunitz JD, Taylor R (1997) Organic fluorine hardly ever accepts hydrogen bonds. Chem Eur
J 3:89–98
12. Dunitz JD (2004) Organic fluorine: odd man out. ChemBioChem 5:614–621
13. Frey JA, Leist R, Leutwyler S (2006) Hydrogen Bonding of the Nucleobase Mimic 2-Pyridone
to Fluorobenzenes: an ab Initio Investigation. J Phys Chem A 110:4188–4195
14. Toth G, Bowers SG, Truong AP, Probst G (2007) The role and significance of unconventional
hydrogen bonds in small molecule recognition by biological receptors of pharmaceutical
relevance. Curr Pharm Des 13:3476–3493
15. Zhu Y-Y, Yi H-P, Li C, Jiang X-K, Li Z-T (2008) The N–H· · · X (X = Cl, Br, and I) hydrogen-
bonding pattern in aromatic amides: a crystallographic and 1H NMR study. Cryst Growth
Des 8:1294–1300
16. Koller AN, Bozilovic J, Engels JW, Gohlke H (2010) Aromatic N versus aromatic F: bioisos-
terism discovered in RNA base pairing interactions leads to a novel class of universal base
analogs. Nucleic Acids Res 38:3133–3146
17. Gregoret LM, Rader SD, Fletterick RJ, Cohen FE (1991) Hydrogen bonds involving sulfur
atoms in proteins. Proteins 9(2):99–107
18. Platts JA, Howard ST, Bracke BRF (1996) Directionality of hydrogen bonds to sulfur and
oxygen. J Am Chem Soc 118(11):2726–2733
19. Allen FH, Bird CM, Rowland RS, Raithby PR (1997) Hydrogen-bond acceptor and donor
properties of divalent sulfur (Y–S–Z and R–S–H). Acta Crystallogr, Sect B: Struct Sci 53:696–
701
20. Sutor DJ (1963) Evidence for existence of C–H· · · O hydrogen bonds in crystals. J Chem Soc
(FEB) 1105–1110
21. Taylor R, Kennard O (1982) Crystallographic evidence for the existence of CH· · · O, CH· · · N
and CH· · · Cl hydrogen bonds. J Am Chem Soc 104(19):5063–5070
22. Brandl M, Weiss MS, JabsA, Suhnel J, Hilgenfeld R (2001) C–H· · · π -interactions in proteins.
J Mol Biol 307(1):357–377
23. Scheiner S, Kar T, Pattanayak J (2002) Comparison of various types of hydrogen bonds
involving aromatic amino acids. J Am Chem Soc 124(44):13257–13264
24. Manikandan K, Ramakumar S (2004) The occurrence of C–H· · · O hydrogen bonds in alpha-
helices and helix termini in globular proteins. Proteins 56(4):768–781
25. Sarkhel S, Desiraju GR (2004) N–H· · · O, O–H· · · O, and C–H· · · O hydrogen bonds in
protein-ligand complexes: strong and weak interactions in molecular recognition. Proteins
54(2):247–259
26. Mani D, Arunan E (2013) The X-C· · ·Y (X = O/F, Y = O/S/F/Cl/Br/N/P) ‘carbon bond’ and
hydrophobic interactions. Phys Chem Chem Phys 15:14377–14383
27. Bauza A, Mooibroek TJ, Frontera A (2014) Influence of ring size on the strength of car-
bon bonding complexes between anions and perfluorocycloalkanes. Phys Chem Chem Phys.
Ahead of Print. doi:10.1039/c4cp01983k
28. Thomas SP, Pavan MS, Guru Row TN (2014) Experimental evidence for ‘carbon bonding’ in
the solid state from charge density analysis. Chem Commun 50:49–51
29. Adman E, Watenpaugh KD, Jensen LH (1975) NH· · · S Hydrogen bonds in Peptococcus
aerogenes ferredoxin, clostridium pasteurianum rubredoxin, and chromatium high potential
iron protein. Proc Nat Acad Sci U S A 72(12):4854–4858
30. Morgan RS, Tatsch CE, Gushard RH, McAdon JM, Warme PK (1978) Chains of alternating
sulfur and p-bonded atoms in 8 small proteins. Int J Pept Protein Res 11(3):209–217
31. Reid KSC, Lindley PF, Thornton JM (1985) Sulfur-aromatic interactions in proteins. FEBS
Lett 190(2):209–213
32. Li HM, Thomas GJ (1991) Cysteine conformation and sulfhydryl interactions in proteins and
viruses.1. Correlation of the Raman S–H B and with hydrogen-bonding and intramolecular
geometry in model compounds. J Am Chem Soc 113(2):456–462
33. Renault L, Nassar N, Vetter I, Becker J, Klebe C, Roth M, Wittinghofer A (1998) The 1.7
angstrom crystal structure of the regulator of chromosome condensation (RCC1) reveals a
seven-bladed propeller. Nature 392(6671):97–101
34. Iwaoka M, Takemoto S, Okada M, Tomoda S (2002) Weak nonbonded S–X (X = O, N, and
S) interactions in proteins. Statistical and theoretical studies. Bull Chem Soc Jpn 75(7):1611–
1625
35. Pantoja-Uceda D, Shewry PR, Bruix M, Tatham AS, Santoro J, Rico M (2004) Solution
structure of a methionine-rich 2S albumin from sunflower seeds: Relationship to its allergenic
and emulsifying properties. Biochem 43(22):6976–6986
36. Brosnan JT, Brosnan ME (2006) The sulfur-containing amino acids: An overview. J Nutr
136(6):1636S–1640S
37. Nicoletti FP, Comandini A, Bonamore A, Boechi L, Boubeta FM, Feis A, Smulevich G, Boffi
A (2010) Sulfide binding properties of truncated hemoglobins. Biochem 49(10):2269–2278
38. Iwaoka M, Isozumi N (2012) Hypervalent nonbonded interactions of a divalent sulfur atom.
Implications in protein architecture and the functions. Molecules 17(6):7266–7283
39. Kolluru GK, Shen XG, Bir SC, Kevil CG (2013) Hydrogen sulfide chemical biology:
pathophysiological roles and detection. Nitric Oxide-Biol Ch 35:5–20
40. Sabin JR (1971) Hydrogen bonds involving sulfur. I. Hydrogen sulfide dimer. J Am Chem
Soc 93(15):3613–3620
41. Li S, Li YS (1991) FTIR spectra of matrix isolated complexes between sulfur-compounds.

Spectrochim Acta Part A 47(2):201–209
42. Wierzejewska M (2000) FTIR matrix isolation studies of complexes of dimethylsulfide,
dimethyldisulfide and hydrogen sulfide with nitric acid. Vib Spectrosc 23(2):253–262
43. Wierzejewska M (2000) Infrared matrix isolation studies of complexes formed between
dimethylsulfide, dimethyldisulfide and nitrous acid. J Mol Struct 520:199–214
44. Wierzejewska M, Saldyka M (2004) Are hydrogen bonds to sulfur and oxygen different?
Theoretical study of dimethylsulfide and dimethylether complexes with nitric acid. Chem
Phys Lett 391(1–3):143–147
45. Galardon E, Roger T, Deschamps P, Roussel P, Tomas A, Artaud I (2012) Synthesis of a
(FeSH)-S-II complex stabilized by an intramolecular N–H· · · S hydrogen bond, which acts
as a H2 S donor. Inorg Chem 51(19):10068–10070
46. Okamura T-a, Kunisue K, Omi Y, Onitsuka K (2013) Strong NH· · · S hydrogen bonds in
molybdoenzyme models containing anilide moieties. Dalton Trans 42:7569–7578
47. Okamura T-a, Ushijima Y, Omi Y, Onitsuka K (2013) Systematic investigation of relationship
between strength of NH· · · S hydrogen bond and reactivity of molybdoenzyme models. Inorg
Chem 52:381–394
48. ZhouYJ, Zhang MM, Wang XS (2013) The N–H· · · X hydrogen bonds in the crystal structures
of (Thio)Isochromene Derivatives. J Chem Crystallogr 43(1):26–30
49. Xu LH, Liu Q, Suenram RD, Lovas FJ, WalkerARH, Jensen JO, SamuelsAC (2004) Rotational
spectra, conformational structures, and dipole moments of thiodiglycol by jet-cooled FTMW
and ab initio calculations. J Mol Spectrosc 228(2):243–250
50. Crittenden DL (2009) A systematic CCSD(T) study of long-range and noncovalent interac-
tions between benzene and a series of first- and second-row hydrides and rare gas atoms. J
Phys Chem A 113(8):1663–1669
51. Cabaleiro-Lago EM, Rodriguez-Otero J (2002) Methanethiol dimer and trimer. An ab initio
and DFT study of the interaction. J Phys Chem A 106(32):7440–7447
52. Wennmohs F, Staemmler V, Schindler M (2003) Theoretical investigation of weak hydrogen
bonds to sulfur. J Chem Phys 119(6):3208–3218
53. Vila A, Mosquera RA (2006) Are the hydrogen bonds involving sulfur bases inverse or
anomalous? Int J Quantum Chem 106(4):928–934
54. Jezierska A, Panek JJ, Mazzarello R (2009) Structural and electronic structure differences
due to the O–H· · · O and O–H· · · bond formation in selected benzamide derivatives: a first-
principles molecular dynamics study. Theor Chem Acc 124(5–6):319–330
55. Steudel R, Steudel Y (2009) Microsolvation of thiosulfuric acid and its tautomeric anions
[HSSO3 ]− and [SSO2 (OH)]− Studied by B3LYP-PCM and G3X(MP2) calculations. J Phys
Chem A 113(36):9920–9933
56. Kaur D, Aulakh D, Khanna S, Singh H (2014) Theoretical study on the nature of S· · · H and
O· · · H hydrogen bonds. J Sulfur Chem 35(3):290–303
57. Ueyama N, Nishikawa N, Yamada Y, Okamura T, Oka S, Sakurai H, Nakamura A (1998)
Synthesis and properties of octaethylporphinato(arenethiolato)iron(III) complexes with in-
tramolecular NH–S hydrogen bond: chemical function of the hydrogen bond. Inorg Chem
37(10):2415–2421
58. Nangia A, Desiraju GR (1999) Axial and equatorial conformations of penicillins, their
sulphoxides and sulphones: the role of N–H· · · S and C–H· · · O hydrogen bonds. J Mol Struct
65–79
59. Word JM, Lovell SC, LaBean TH, Taylor HC, Zalis ME, Presley BK, Richardson JS, Richard-
son DC (1999) Visualizing and quantifying molecular goodness-of-fit: small-probe contact
dots with explicit hydrogen atoms. J Mol Biol 285(4):1711–1733
60. Word JM, Lovell SC, Richardson JS, Richardson DC (1999) Asparagine and glutamine:
using hydrogen atom contacts in the choice of side-chain amide orientation. J Mol Biol
285(4):1735–1747
61. Francois S, Rohmer MM, Benard M, Moreland AC, Rauchfuss TB (2000) The N–H–S hy-
drogen bond in (TACN)2 Fe2 S6 (TACN = triazacyclononane) and in model systems involving
the persulfido moiety: an ab initio and DFT study. J Am Chem Soc 122(51):12743–12750
62. Steiner T (2000) S–H· · · S hydrogen-bond chain in thiosalicylic acid. Acta Crystallogr Sect
C-Cryst Struct Commun 56:876–877
63. Krepps MK, Parkin S, Atwood DA (2001) Hydrogen bonding with sulfur. Cryst Growth Des
1(4):291–297
64. Hambley TW, Hibbs DE, Turner P, Howard ST, Hursthouse MB (2002) Insights into bonding
and hydrogen bond directionality in thioacetamide from the experimental charge distribution.
J Chem Soc-Perkin Trans 2(2):235–239
65. Lynch DE, McClenaghan I, Light ME, Coles SJ (2002) The solid-state packing of sulfur
substituted 2-aminopyrimidines and the occurrence of N–H-S hydrogen-bonding associations.
Cryst Eng 5(1):79–94
66. UsmanA, Fun HK, Ganguly NC, Datta M, Ghosh K (2003) 2-(2-hydroxyphenyl)-1,3-dithiane.
Acta Crystallogr Sect E-Struct Rep Online 59:O773–O775
67. Sundaresan CN, Dixit S, Venugopalan P (2004) A supramolecular assembly dominated by
N–H· · · S hydrogen bonds: structure of 2-thioureidobenzoxazole by single crystal X-ray
diffraction. J Mol Struct 693(1–3):205–209
68. Valdes-Martinez J, Hernandez-Ortega S, Rubio M, Li DT, Swearingen JK, Kaminsky W,
Kelman DR, West DX (2004) Study of the sulfur atom as hydrogen bond acceptor in N(2)-
pyridylmethyl-N-arylthioureas. J Chem Crystallogr 34(8):533–540
69. Moroder L (2005) Isosteric replacement of sulfur with other chalcogens in peptides and
proteins. J Pept Sci 11(4):187–214
70. Tutughamiarso M, Egert E (2011) Cocrystals of 6-propyl-2-thiouracil: N–H· · · O versus N–
H· · · S hydrogen bonds. Acta Crystallogr, Sect C: Cryst Struct Commun 67:O439–O445
71. Naktode K, Kottalanka RK, Panda TK (2012) N-(2,6-Dimethylphenyl)diphenyl-
phosphinamine chalcogenides (S, Se) and a zirconium complex possessing phosphanylamide
in the coordination sphere. New J Chem 36(11):2280–2285
72. Minkov VS, Boldyreva EV (2013) Weak hydrogen bonds formed by thiol groups in N-Acetyl-
L-Cysteine and their response to the crystal structure distortion on increasing pressure. J Phys
Chem B 117(46):14247–14260
73. Olivella M, Caltabiano G, Cordomi A (2013) The role of Cysteine 6.47 in class A GPCRs.
BMC Struct Biol 13:3
74. AlDamen MA, Sinnokrot M (2014) Crystallographic and theoretical studies of 1-(1-naphthyl)-
2-thiourea with intermolecular N–H· · · S heteroatom interaction and N–H· · · π interaction.
J Struct Chem 55(1):53–60
75. Beck BW, Xie Q, Ichiye T (2001) Sequence determination of reduction potentials by cysteinyl
hydrogen bonds and peptide dipoles in 4Fe-4S ferredoxins. Biophys J 81(2):601–613
76. Duan GL, Smith VH, Weaver DF (2001) Characterization of aromatic-thiol π-type hydrogen
bonding and phenylalanine-cysteine side chain interactions through ab initio calculations and
protein database analyses. Mol Phys 99(19):1689–1699
77. Pal D, Chakrabarti P (1998) Different types of interactions involving cysteine sulfhydryl group
in proteins. J Biomol Struct Dyn 15(6):1059–1072
78. Zhou P, Tian FF, Lv FL, Shang ZC (2009) Geometric characteristics of hydrogen bonds
involving sulfur atoms in proteins. Proteins Struct Funct Bioinf 76(1):151–163
79. Steiner T (2002) The hydrogen bond in the solid state. Angew Chem Int Ed 41(1):48–76
80. Ranaghan KE, Hung JE, Bartlett GJ, Mooibroek TJ, Harvey JN, Woolfson DN, van der Donk
WA, Mulholland AJ (2014) A catalytic role for methionine revealed by a combination of
computation and experiments on phosphite dehydrogenase. Chem Sci 5(6):2191–2199
81. Duncan MA (2003) Infrared spectroscopy to probe structure and dynamics in metal ion-
molecule complexes. Int Rev Phys Chem 22(2):407–435
82. Ebata T, Fujii A, Mikami N (1998) Vibrational spectroscopy of small-sized hydrogen-bonded
clusters and their ions. Int Rev Phys Chem 17(3):331–361
83. Fujii A, Mizuse K (2013) Infrared spectroscopic studies on hydrogen-bonded water networks
in gas phase clusters. Int Rev Phys Chem 32(2):266–307
84. Fujii M, Dopfer O (2012) Ionisation-induced site switching dynamics in solvated aromatic
clusters: phenol-(rare gas)(n) clusters as prototypical example. Int Rev Phys Chem 31(1):131–
173
85. Kang C, Pratt DW (2005) Structures, charge distributions, and dynamical properties of weakly
bound complexes of aromatic molecules in their ground and electronically excited states. Int
Rev Phys Chem 24(1):1–36
86. Simons JP, Jockusch RA, Carcabal P, Hung I, Kroemer RT, Macleod NA, Snoek LC (2005)
Sugars in the gas phase. Spectroscopy, conformation, hydration, co-operativity and selectivity.
Int Rev Phys Chem 24(3–4):489–531
87. Wild DA, Bieske EJ (2003) Infrared investigations of negatively charged complexes and
clusters. Int Rev Phys Chem 22(1):129–151
88. Buckingham AD, Fowler PW (1985) A model for the geometries of van der Waals complexes.
Can J Chem 63:2018–2025
89. Graindourze M, Maes G (1985) Matrix isolation vibrational spectra of alkyl chalcogenides
complexed with hydrogen chloride. Matrix isolation IR spectra as a guide for the analysis of
solution IR spectra of alkyl sulfide and alkyl selenide complexes with hydrogen chloride. J
Mol Spectrosc 114:97–104
90. Maes G, Graindourze M (1985) Matrix isolation vibrational spectra of alkyl chalcogenides
complexed with HCl: structure of alkyl sulfide and alkyl selenide complexes with hydrochloric
acid in ar matrices from infrared spectra. J Mol Spectrosc 113(2):410–425
91. Maes G, Graindourze M (1985) Conformational isomerism of ethyl chalcogenides in inert
matrixes: influence on the νs band structure of hydrogen-bonded complexes with hydrogen
chloride. J Mol Spectrosc 114:280–288
92. Andrews L, Arlinghaus RT, Hunt RD (1986) FTIR spectra of dialkyl sulfide and alkanethiol
complexes with hydrogen fluoride in solid argon. Inorg Chem 25(18):3205–3209
93. Barnes AJ, Wright MP (1986) Molecular complexes of hydrogen halides with ethers and
sulfides studied by matrix isolation vibrational spectroscopy. J Mol Struct: Theochem 28:21–
30
94. Jeng MLH, Ault BS (1990) Infrared matrix isolation study of hydrogen bonds involving
carbon–hydrogen bonds: alkynes with bases containing second- and third-row donor atoms.
J Phys Chem 94(4):1323–1327
95. Jeng MLH, Ault BS (1990) Infrared matrix isolation studies of hydrogen bonds involving
carbon–hydrogen bonds: alkenes with selected bases. J Phys Chem 94:4851–4855
96. Li S, Li YS (1991) FT-IR spectra of matrix isolated complexes between some alkanethiols
and sulfur dioxide. J Mol Struct 248:79–88
97. Kim KS, Tarakeshwar P, Lee JY (2000) Molecular clusters of π-systems: theoretical studies
of structures, spectra, and origin of interaction energies. Chem Rev 100(11):4145–4185
98. Li S, Kurtz H, Korambath P, Li YS (2000) Infrared spectra, photochemistry, and ab
initio calculations of matrix isolated methanethiol/sulfur dioxide complex. J Mol Struct
550–551:235–244
99. Howard DL, Kjaergaard HG (2008) Hydrogen bonding to divalent sulfur. Phys Chem Chem
Phys 10(28):4113–4118
100. Biswal HS, Shirhatti PR, Wategaonkar S (2009) O–H· · · O versus O–H· · · S hydrogen bonding
I: experimental and computational studies on the p-Cresol·H2 O and p-Cresol·H2 S complexes.
J Phys Chem A 113(19):5633–5643
101. Biswal HS, Shirhatti PR, Wategaonkar S (2010) O–H· · · O versus O–H· · · S hydrogen bonding
2: alcohols and thiols as hydrogen bond acceptors. J Phys Chem A 114:6944–6955
102. Biswal HS, Chakraborty S, Wategaonkar S (2008) Experimental evidence of O–H· · · S
hydrogen bonding in supersonic jet. J Chem Phys 129(18):184311–184317
103. Biswal HS, Wategaonkar S (2010) O–H· · · O versus O–H· · · S Hydrogen Bonding. 3. IR–UV
double resonance study of hydrogen bonded complexes of p-cresol with diethyl ether and its
sulfur analog. J Phys Chem A 114(19):5947–5957
104. Biswal HS, Wategaonkar S (2011) OH· · · X (X = O, S) hydrogen bonding in tetrahydrofurane
and tetrahydrothiophene. J Chem Phys 135(13):134306
105. Contreras-García J, Johnson ER, Keinan S, Chaudret R, Piquemal J-P, Beratan DN, Yang W
(2011) NCIPLOT: a program for plotting noncovalent interaction regions. J Chem Theory
Comput 7(3):625–632
106. Otero-de-la-Roza A, Johnson ER, Contreras-Garcia J (2012) Revealing non-covalent interac-

tions in solids: NCI plots revisited. Phys Chem Chem Phys 14(35):12165–12172
107. Bader RFW, Nguyen-Dang TT (1981) Quantum theory of atoms in molecules—Dalton
revisited. Adv Quantum Chem 14:63–124
108. Bader RFW (1990) Atoms in molecules: a quantum theory. Clarendon Press, Oxford
109. Bader RFW (1991) A quantum theory of molecular structure and its applications. Chem Rev
91(5):893–928
110. Koch U, Popelier PLA (1995) Characterization of C–H–O hydrogen bonds on the basis of
the charge density. J Phys Chem 99(24):9747–9754
111. Popelier P (2000) Atoms in molecules, an introduction. Prentice Hall, Englewood Cliffs
112. Glendening ED, Streitwieser A (1994) Natural energy decomposition analysis: an energy
partitioning procedure for molecular-Interactions with application to weak hydrogen bonding,
strong ionic, and moderate donor-dcce tor interactions. J Chem Phys 100(4):2900–2909
113. Glendening ED (1996) Natural energy decomposition analysis: explicit evaluation of electro-
static and polarization effects with application to aqueous clusters of alkali metal cations and
neutrals. J Am Chem Soc 118(10):2473–2482
114. Schenter GK, Glendening ED (1996) Natural energy decomposition analysis: the linear
response electrical self energy. J Phys Chem 100(43):17152–17156
115. Kitaura K, Morokuma K (1976) A new energy decomposition scheme for molecular
interactions within the Hartree-Fock approximation. Int J Quantum Chem 10(2):325–340
116. Umeyama H, Morokuma K (1977) The origin of hydrogen bonding. An energy decomposition
study. J Am Chem Soc 99(5):1316–1332
117. Stevens WJ, Fink WH (1987) Frozen fragment reduced variational space analysis of hydrogen
bond in interactions-application to the water dimer. Chem Phys Lett 139(1):15–22
118. Jouvet C, Lardeux-Dedonder C, Richard-Viard M, Solgadi D, Tramer A (1990) Reactivity of
molecular clusters in the gas phase: proton-transfer reaction in neutral phenol-(ammonia)n
and phenol-(ethanamine)n. J Phys Chem 94:5041–5048
119. Iwasaki A, Fujii A, Watanabe T, Ebata T, Mikami N (1996) Infrared spectroscopy of hydrogen-
bonded phenol-amine clusters in supersonic jets. J Phys Chem 100(40):16053–16057
120. Biswas N, Wategaonkar S, Watanabe T, Ebata T, Mikami N (2004) Fluorescence, REMPI,
hole-burning, and FDIR spectroscopy of para-cyanophenol-water1 complex. Chem Phys Lett
394:61–67
121. Bhattacharyya S, Bhattacherjee A, Shirhatti PR, Wategaonkar S (2013) O–H· · · S hydrogen
bonds conform to the acid-base formalism. J Phys Chem A 117(34):8238–8250
122. Braun JE, Mehnert T, Neusser HJ (2000) Binding energy of van der Waals- and hydrogen-
bonded clusters by threshold ionization techniques. Int J Mass Spectrom 203(1–3):1–18
123. Braun J, Neusser HJ, Hobza P (2003) N–H· · · p interactions in Indole-Benzene-h6 ,d6
and Indole-Benzene-h6 ,d6 radical cation complexes. Mass Analyzed threshold ionization
experiments and correlated ab initio quantum chemical calculations. J Phys Chem A
107(19):3918–3924
124. Georgiev S, Neusser HJ (2004) Investigation of hydrogen bonding in 3-methylindole–H2 O
cluster by mass analyzed threshold ionization. Chem Phys Lett 389(1–3):24–28
125. Zhang B, Li C, Su H, Lin JL, Tzeng WB (2004) Mass analyzed threshold ionization spec-
troscopy of p-fluorophenol cation and the p-fluoro substitution effect. Chem Phys Lett
390:65–70
126. Georgiev S, Neusser HJ (2005) Mass analyzed threshold ionization of hydrogen bonded clus-
ters of biological molecules: the 3-methylindole center.C6 H6 complex. J Electron Spectrosc
Relat Phenom 142(3):207–213
127. Shibasaki K, Fujii A, Mikami N, Tsuzuki S (2006) Magnitude of the CH/π interaction in the
gas phase: Experimental and theoretical determination of the accurate interaction energy in
benzene-methane. J Phys Chem A 110(13):4397–4404
128. Huang JH, Huang K, Liu SQ, Luo Q, Tzeng WB (2007) Molecular structures and vibrations
of cis and trans m-cresol in the electronically excited S1 and cationic D0 states. J Photochem
Photobiol A 188(2–3):252–259
129. Shirhatti PR, Wategaonkar S (2012) Mass analyzed threshold ionization (MATI) spectroscopy
of p-cresol. Indian J Phys 86(3):159–164
130. Cordes E, Dopfer O, Wright TG, Mullerdethlefs K (1993) Vibrational spectroscopy of the
phenol-ethanol cation. J Phys Chem 97(29):7471–7479
131. Vondrak T, Sato S, Kimura K (1997) Cation vibrational spectra of indole and indole-argon van
der Waals complex. A zero kinetic energy photoelectron study. J Phys Chem A 101(13):2384–
2389
132. Dessent CEH, Muller-Dethlefs K (2000) Hydrogen-bonding and van der Waals complexes
studied by ZEKE and REMPI spectroscopy. Chem Rev 100(11):3999–4021
133. Zierhut M, Dummler S, Roth W, Fischer I (2003) Multiphoton ionization and zero kinetic
energy photoelectron spectroscopy of the 1-naphthol(H2 O) cluster. Chem Phys Lett 381:346–
353
134. Cockett MCR (2005) Photoelectron spectroscopy without photoelectrons: twenty years of
ZEKE spectroscopy. Chem Soc Rev 34(11):935–948
135. Biswal HS, Wategaonkar S (2009) Nature of the N–H· · · S hydrogen bond. J Phys Chem A
113(46):12763–12773
136. Ballew RM, Sabelko J, Gruebele M (1996) Direct observation of fast protein folding: the
initial collapse of apomyoglobin. Proc Nat Acad Sci U S A 93(12):5759–5764
137. Desfrancois C, Carles S, Schermann JP (2000) Weakly bound clusters of biological interest.
Chem Rev 100(11):3943–3962
138. Meyer EA, Castellano RK, Diederich F (2003) Interactions with aromatic rings in chemical
and biological recognition. Angew Chem Int Ed 42(11):1210–1250
139. Biswal HS, Gloaguen E, Mons M, Bhattacharyya S, Shirhatti PR, Wategaonkar S (2011)
Structure of the indole-benzene Dimer Revisited. J Phys Chem A 115(34):9485–9492
140. Braun JE, Grebner TL, Neusser HJ (1998) Van der Waals versus hydrogen-bonding in com-
plexes of indole with argon, water, and benzene by mass-analyzed pulsed field threshold
ionization. J Phys Chem A 102(19):3273–3278
141. Tsuzuki S, Honda K, Uchimaru T, Mikami M, Tanabe K (2000) Origin of the attraction and
directionality of the NH/π interaction: Comparison with OH/π and CH/π interactions. J Am
Chem Soc 122(46):11450–11458
142. Tsuzuki S, Honda K, Uchimaru T, Mikami M, Tanabe K (2000) The magnitude of the CH/p
interaction between benzene and some model hydrocarbons. J Am Chem Soc 122(15):3746–
3753
143. Morita S-i, Fujii A, Mikami N, Tsuzuki S (2006) Origin of the attraction in aliphatic C–H/π
interactions: Infrared spectroscopic and theoretical characterization of gas-phase clusters of
aromatics with methane. J Phys Chem A 110(36):10583–10590
144. Tsuzuki S, Honda K, Uchimaru T, Mikami M, Fujii A (2006) Magnitude and directionality of
the interaction energy of the aliphatic CH/π interaction: Significant difference from hydrogen
bond. J Phys Chem A 110(33):10163–10168
145. Alberti M, Aguilar A, Huarte-Larranaga F, Lucas JM, Pirani F (2014) Benzene-hydrogen bond
(C6 H6 -HX) Interactions: the influence of the X nature on their strength and anisotropy. J Phys
Chem A 118(9):1651–1662
146. Ueyama N, Nishikawa N, Yamada Y, Okamura T, Nakamura A (1996) Cytochrome P-450
model (porphinato)(thiolato)iron(III) complexes with single and double NH· · · S hydrogen
bonds at the thiolate site. J Am Chem Soc 118(50):12826–12827
147. Suzuki N, Higuchi T, UranoY, Kikuchi K, Uekusa H, OhashiY, Uchida T, Kitagawa T, Nagano
T (1999) Novel iron porphyrin-alkanethiolate complex with intramolecular NH· · · S hydrogen
bond: Synthesis, spectroscopy, and reactivity. J Am Chem Soc 121(49):11571–11572
148. Liu S-G, Li Y-Z, Zuo J-L, You X-Z (2004) N, N’-Bis[2-(methylsulfanyl)phenyl]pyridine-2,6-
dicarboxamide. Acta Cryst E 60(9):o1527–o1529
149. Du P, Jiang XK, Li Z-T (2009) Five- and six-membered N–H· · · S hydrogen bonding in
aromatic amides. Tetrahedron Lett 50(3):320–324
150. Biswal HS, Gloaguen E, Loquais Y, Tardivel B, Mons M (2012) Strength of (NHS)-S-
· · · hydrogen bonds in methionine residues revealed by gas-phase ir/uv spectroscopy. J Phys
Chem Lett 3(6):755–759
151. Khedkar JK, Deshmukh MM, Gadre SR, Gejji SP (2012) Hydrogen bond energies and
cooperativity in substituted calix[n]arenes (n = 4, 5). J Phys Chem A 116(14):3739–3744
152. Kitaura K, Ikeo E,Asada T, Nakano T, Uebayasi M (1999) Fragment molecular orbital method:
an approximate computational method for large molecules. Chem Phys Lett 313(3–4):701–
706
153. Amos RD (1986) Structures, harmonic frequencies and infrared intensities of the dimers of
H2 O and H2 S. Chem Phys 104(1):145–151
154. Gutowsky HS, Emilsson T, Arunan E (1997) Rotational spectra, structure, and internal
dynamics of Ar–H2 S isotopomers. J Chem Phys 106(13):5309–5315
155. Arunan E, Emilsson T, Gutowsky HS, Fraser GT, de Oliveira G, Dykstra CE (2002) Rotational
spectrum of the weakly bonded C6 H6 –H2 S dimer and comparisons to C6 H6 –H2 O dimer. J
Chem Phys 117(21):9766–9776
156. Goswami M, Mandal PK, Ramdass DJ, Arunan E (2004) Rotational spectra and structure of
the floppy C2 H4 –H2 S complex: bridging hydrogen bonding and van der Waals interactions.
Chem Phys Lett 393(1–3):22–27
157. Hermida-Ramon JM, Cabaleiro-Lago EM, Rodriguez-Otero J (2005) Theoretical character-
ization of structures and energies of benzene–(H2 S)n and (H2 S)n (n = 1–4) clusters. J Chem
Phys 122(20):204315
158. Goswami M, Arunan E (2011) Microwave spectrum and structure of C6 H5 CCH· · · H2 S
complex. J Mol Spectrosc 268(1–2):147–156
159. Bhattacherjee A, Matsuda Y, Fujii A, Wategaonkar S (2013) The intermolecular S–H· · ·Y
(Y = S, O) hydrogen bond in the H2 S Dimer and the H2 S-MeOH complex. Chemphyschem
14(5):905–914
160. Goswami M, Neill JL, Muckle M, Pate BH, Arunan E (2013) Microwave, infrared-
microwave double resonance, and theoretical studies of C2 H4 · · · H2 S complex. J Chem Phys
139(10):104303
161. Tursi AJ, Nixon ER (1970) Infrared spectra of matrix-isolated hydrogen sulfide in solid
nitrogen. J Chem Phys 53(2):518–521
162. Barnes AJ, Howells JDR (1972) Infra-red cryogenic studies. Part 7 hydrogen sulphide in
matrices. J Chem Soc Faraday Trans 68(4):729–736
163. Lechugafossat L, Flaud JM, Camypeyret C, Johns JWC (1984) The spectrum of natural
hydrogen sulfide between 2150 and 2950 cm−1 . Can J Phys 62(12):1889–1923
164. Helgaker T, Klopper W, Koch H, Noga J (1997) Basis-set convergence of correlated
calculations on water. J Chem Phys 106:9639–9646
165. Rocher-Casterline BE, Ch’ng LC, Mollner AK, Reisler H (2011) Communication: determi-
nation of the bond dissociation energy (D0 ) of the water dimer, (H2 O)2 , by velocity map
imaging. J Chem Phys 134:211101/211101–211101/211104
166. Adhikari U, Scheiner S (2012) Contributions of various noncovalent bonds to the interaction
between an amide and S-containing molecules. Chemphyschem 13:3535–3541
167. Solimannejad M, Scheiner S (2011) Unconventional H-bonds: SH· · · N interaction. Int J
Quantum Chem 111:3196–3200
168. Solimannejad M, Gharabaghi M, Scheiner S (2011) SH· · · N and SH· · · π blue-shifting H-
bonds and N· · · P interactions in complexes pairing HSN with amines and phosphines. J Chem
Phys 134:024312/024311–024312/024316
169. Steiner T, Koellner G (2001) Hydrogen bonds with π-acceptors in proteins: frequencies and
role in stabilizing local 3D structures. J Mol Bio 305(3):535–557
170. Ringer AL, Senenko A, Sherrill CD (2007) Models of S/π interactions in protein structures:
comparison of the H2 S-benzene complex with PDB data. Protein Sci 16(10):2216–2223
171. Tsuzuki S, Uchimaru T (2006) Magnitude and physical origin of intermolecular interactions of
aromatic molecules: recent progress of computational studies. Curr Org Chem 10(7):745–762
172. Vaupel S, Brutschy B, Tarakeshwar P, Kim KS (2006) Characterization of weak NH-π in-
termolecular interactions of ammonia with various substituted π-systems. J Am Chem Soc
128(16):5416–5426
173. Salonen LM, Ellermann M, Diederich F (2011) Aromatic rings in chemical and biological
recognition: energetics and structures. Angew Chem Int Ed 50(21):4808–4842
174. Tauer TP, Derrick ME, Sherrill CD (2005) Estimates of the ab initio limit for sulfur-π
interactions: the H2 S-benzene dimer. J Phys Chem A 109(1):191–196
175. Wang YX, Paulus B (2007) A comparative electron correlation treatment in H2 S-benzene
dimer with DFT and wavefunction-based ab initio methods. Chem Phys Lett 441(4–6):187–
193
176. Biswal HS, Wategaonkar S (2009) Sulfur, not too far behind O, N, and C: S–H· · · π hydrogen
bond. J Phys Chem A 113(46):12774–12782
177. Hager J, Wallace SC (1983) Laser spectroscopy and photodynamics of indole and indole-
vanderwaals molecules in a supersonic beam. J Phys Chem 87(12):2121–2127
178. Saggu M, Levinson NM, Boxer SG (2011) Direct measurements of electric fields in weak
OH· · · π hydrogen bonds. J Am Chem Soc 133(43):17414–17419
179. Saggu M, Levinson NM, Boxer SG (2012) Experimental quantification of electrostatics in
X-H· · · π hydrogen bonds. J Am Chem Soc 134(46):18986–18997
Chapter 3
CH· · · π Interaction in Organic Molecules
Osamu Takahashi
Abstract CH· · · π interaction, which is one of weak non-covalent hydrogen bonds

(H-bonds), plays an important role in physics, chemistry, and biology. First, defini-
tion of the H-bond is introduced, and a position of the CH· · · π H-bond is confirmed.
Next, several experimental evidence of the CH· · · π H-bond are described. Then,
consequence of the CH· · · π H-bond in organic chemistry such as conformation,
reaction selectivity, and molecular recognition is discussed especially from recent
results.
3.1 Introduction
A molecule consists of atoms linked with various kinds of chemical bonds. A covalent
bond, the strongest chemical bond, connects atoms in the molecule with shared two
or more electrons, while non-covalent bonds are weaker but play a dominant role
in determining molecular shape. The non-covalent bond also plays important roles
in various chemical phenomena such as structure of supramolecules, chiroptical
property, enantiomeric selectivity, etc. Importance of weak hydrogen bonds has
been dealt with in monographs written by Scheiner [1] and Nishio et al. [2].
Importance of the non-covalent bond depends not only on its strength but also the
directionality. A hydrogen bond (H-bond) is one of the non-covalent bonds, which
plays important roles in physics, chemistry, and biology.
In the present chapter, definition of the H-bond is given first and the position of
the CH· · · π interaction or CH· · · π H-bond is confirmed. Next, several experimental
evidences for the CH· · · π H-bond in organic chemistry and spectroscopy with theo-
retical calculations are summarized. As evidences in organic chemistry, some topics
related to conformation, reaction selectivity, molecular recognition, and chiroptical
properties are discussed, especially from recent reports.
O. Takahashi ()
Institute for sustainable sciences and development, Hiroshima University,
739-8526 Higahshi-Hiroshima, Japan
e-mail: shu@hiroshima-u.ac.jp

48 O. Takahashi
3.2 What is the CH· · · π Interaction?
Definition of the H-bond in early days has been changed and extended [3, 4]. Accord-
ing to Pauling, the H-bond originates from the Coulombic or dipole-dipole interaction
between two polar atoms (X, Y): XH· · ·Y. The XH (hydrogen donor) is a hydrogen
donor and positively polarized, whileY (hydrogen acceptor) bears lone-pair electrons
and negatively polarized.
3.2.1 Definition of the CH· · · π Interaction
A modern and more comprehensive definition of the H-bond has been presented,
in 1960, by Pimentel and McClellan [5]. According to this definition, there is no
restriction to the nature of X andY, energy, and directionality of the H-bond. Recently,
the IUPAC commission has published a technical report to define the H-bond: This is
an attractive interaction between a hydrogen atom from a molecule or a molecular
fragment X–H in which X is more electronegative than H, and an atom or a group
of atoms in the same or a different molecule, in which there is evidence of bond
formation [6]. In their report, six criteria and characteristics of H-bonds are listed.
A target system in the present chapter, CH· · · π H-bond, is also included in this
definition, as the weakest limit of H-bonds.
There are several types of H-bonds. The OH· · · O bond is a typical one and
relatively strong. Not only the OH and NH groups but also CH groups are hydrogen
donors for XH groups in the second row elements of periodic table of the elements.
Further, not only a lone pair provided by O and N atoms but also π electrons are good
candidates as hydrogen acceptor. For example, CH· · · O, OH· · · π, and NH· · · π H-
bonds are weaker than the conventional H-bond; the interaction energy is from ca. 2
to 4 kcal mol−1 . The CH· · · π interaction is the weakest H-bond and its interaction
energy ranges from ca. 0.5 to 2 kcal mol−1 ; this is comparable with the thermal energy,
however, plays a very important role in organic chemistry and biology. A number of
review articles were written and its importance is widely recognized [4, 7–11].
3.2.2 Energy Decomposition Scheme
The interaction energy between groups in H-bonds can be decomposed into several
physical components. Several energy decomposition schemes have been reported.
The first one is the so-called Kitaura-Morokuma partitioning [12], which was de-
veloped within the framework of Hartree-Fock approximation. Following their
pioneering work, some alternative schemes were proposed. Reed et al. [13] presented
another energy decomposition scheme based on natural bond orbitals (NBOs). Meth-
ods based on electron density in a molecule have also been developed. For instance,
Bader’s “atoms in molecules” (AIM) analysis is applied to partition the electron
3 CH· · · π Interaction in Organic Molecules 49
Table 3.1 Energy decomposition for various types of hydrogen bonds using SAPT method (in kcal
mol−1 ).
Type of Example Eatotal Eaes Eaer Eaind Eadisp Eaother Ees Edisp
H-bond /Etotal /Etotal
OH· · · Ob H2 O· · · H2 O − 4.60 − 8.22 8.30 − 1.38 − 2.36 − 0.94 1.79 0.51
OH· · · N b
H2 O· · · NH3 − 6.10 − 11.15 11.76 − 2.10 − 3.05 − 1.55 1.83 0.5
OH· · · π b
H2 O· · · C2 H4 − 2.57 − 3.82 4.96 − 0.78 − 2.24 − 0.68 1.49 0.87
NH· · · Ob NH3 · · · H2 O − 2.15 − 3.81 4.14 − 0.50 − 1.61 − 0.37 1.77 0.75
NH· · · N b
NH3 · · · NH3 − 2.75 − 5.14 5.84 − 0.79 − 2.05 − 0.60 1.87 0.75
NH· · · πb NH3 · · · C2 H4 − 1.34 − 1.79 2.55 − 0.23 − 1.59 − 0.28 1.33 1.19
CH· · · O b
CH4 · · · H2 O − 0.45 − 0.89 1.61 − 0.10 − 0.98 − 0.09 2.00 2.18
CH· · · N b
CH4 · · · NH3 − 0.65 − 1.32 2.32 − 0.30 − 1.17 − 0.18 2.03 1.8
CH· · · πb CH4 · · · C2 H4 − 0.81 − 0.57 1.45 − 0.03 − 1.58 − 0.07 0.71 1.95
OH· · · π c
H2 O· · · C6 H6 − 2.86 − 2.94 3.78 − 1.05 − 3.18 1.03 1.11
NH· · · πc NH3 · · · C6 H6 − 2.08 − 2.05 3.40 − 0.60 − 3.32 0.98 1.60
CH· · · π c
CH4 · · · C6 H6 − 1.27 − 0.86 2.23 − 0.10 − 2.4 − 0.16 0.68 1.89
CH· · · π d
C2 H2 · · · C6 H6 − 2.54 − 2.89 5.46 − 0.55 − 3.75 − 0.72 1.14 1.48
a
The total interaction (Etotal ), electrostatic (Ees ), exchange repulsion (Eer ), induction (Eind ),
dispersion (Edisp ), and other effects (Eother ) energies, respectively
b
This work
c
Reference [18]
d
Reference [19]
density of a many-electron system into basins uniquely [14]. In these schemes, sym-
metry adapted perturbation theory (SAPT) [15–17] is widely applied to study the
intermolecular interaction within the framework of density functional theory (DFT).
Some typical H-bonds and their energy components are listed in Table 3.1. The elec-
trostatic energy (Ees ) is dominant for relatively strong H-bonds such as OH· · · O and
the dispersion energy (Edisp ) is minor. On the other hand, Edisp is dominant for weak
H-bonds such as CH· · · π.
3.2.3 AIM Analysis
The AIM analysis is a valuable tool in hands of both experimental and theoreti-
cal chemists. Using this analysis, topological properties of electron density in the
system can be evaluated [14, 20]. This is an elegant and sophisticated theory of chem-
ical bonding based on the topological analysis of the electron density. The electron
density can be measured by either X-ray and neutron diffraction crystallography or
computation with the aid of high-level ab initio and DFT calculation. It is possible to
decompose and analyze quantitatively the intra- and inter-molecular interactions that
50 O. Takahashi
Table 3.2 Topological parameters of some types of H-bondsa

Type of H-bonds Example ρ XH ∇ 2 ρ XH ρH...Y ∇ 2 ρH . . . Y
OH· · · O H2 O· · · H2 O 0.356 − 2.512 0.023 0.091
OH· · · N H2 O· · · NH3 0.348 − 2.450 0.028 0.085
OH· · · π H2 O· · · C2 H2 0.283 − 1.033 0.014 0.052
CH· · · π C2 H2 · · · C2 H2 0.284 − 1.029 0.007 0.019
a
Reference [22]
characterize any molecular systems. The aspect of AIM theory is that it redefines the
concept of the chemical bond in terms of the topological properties of ρ(r), namely
its gradient field, ∇ρ(r), and its curvature or Laplacian, ∇ 2 ρ(r) [3]. A bond critical
point (BCP) is a point along the trajectory of the gradient path connecting two local
electron density maxima with ∇ρ(r) = 0 (nuclei) and lying at the borderline of the
two atomic basins involved. The theory shows that the properties of ∇ 2 ρ(r) = 0 at
this BCP can discriminate among the different types of chemical bonds. Koch and
Popelier proposed empirical criteria for XH· · ·Y interactions to establish true H-
bonds, based on the AIM analysis [21]. In Table 3.2, typical topological parameters
for H-bonds are compiled. All cases are satisfied with the Popelier’s criteria.
3.2.4 Cambridge Structural Database (CSD) Analysis
In studying the weak H-bond, a definite evidence is provided by database analysis

using the Cambridge Structural Database (CSD) [23]. Various weak H-bonds were
studied such as CH· · · O [24], OH· · · π, NH· · · π [25], and CH· · · π H-bonds [26, 27].
Umezawa et al. [27] searched for short intramolecular and intermolecular CH· · · π
H-bonds contacts. They found that ca. 29 % of compounds in their entries are contact-
ing with short CH· · · π (aromatic) distance. The corresponding values for OH· · · π
and NH· · · π interactions are 1.4 and 2.7 %, respectively. Note that these values were
estimated with CSD version 515, 181309 entries. If CSD data of other version were
used, the results would be changed. However the number of the compound entries is
so large, the final conclusions would not be affected even if the recent CSD was used.
The same trends were obtained for intermolecular XH· · · π H-bonds. It is obvious
that the CH· · · π H-bond is quite important to describe the intra- and intermolecular
structure in crystal.
As described in the previous section, the H-bond shows directionality, in which
the three atoms XH· · ·Y usually tend toward linearity. For the CH· · · π H-bond, the
same trends have been found, computationally and spectroscopically. A minimum
of potential energy surface as a function of bending angle of XH· · ·Y showed linear
from ab initio calculations although its depth was shallow as already discussed in
the section of energy decomposition. However, according to an analysis using the
CSD, the directionality and the CH· · · π-plane distance (Dpln ) have been found to
Table 3.3 Structural database analysis for CH· · · π interactiona

Contact type Dpln /Åb α/degreec α(corrected)/degreed
Cl3 CH· · · π 2.53 ± 0.17 157 ± 12 169 ± 11
Cl2 CH· · · π e
2.62 ± 0.15 151 ± 13 159 ± 14
sp-CH· · · π 2.62 ± 0.13 152 ± 13 159 ± 13
sp2 -CH· · · πf 2.73 ± 0.11 148 ± 11 154 ± 13
sp -CH· · · π
2 g
2.70 ± 0.11 146 ± 9 149 ± 11
CCH3 · · · π 2.75 ± 0.10 148 ± 13 157 ± 15
a
Reference [29]
b
The mean values and the standard deviations of CH· · · π distance
c
The mean values and the standard deviations of CH· · · π access angle
d
The mean values and the standard deviations of CH· · · π access angle α corrected by a factor of
1/sinα
e
Duplicate hits by two (Cl2 CH2 ) or three (CCH3 ) hydrogen atoms to a single aromatic rung were
not counted
f
Organic crystals with no disorder and R ≤ 5 %
g
Neutron data including organometallic compounds
correlate, depending on the strength (or acidity) of the proton donor [27, 28]. In
Table 3.3, results of a CSD analysis for CH· · · π interactions are shown. Histograms
in Fig. 3.1 show the distribution of CH hydrogens as a function of C–H· · · π access
angle (α) [29]. Correlation diagrams between Dpln and α are given in Fig. 3.2. In
crystals, the directionality may not be linear by various limitations. Although energy
components of CH· · · π interaction is mostly dispersive, (no directionality), it is
obvious that the CH· · · π H-bond has directionality.
3.3 Evidence of CH· · · π H-bond
A number of experimental evidences of CH· · · π H-bond have accumulated by var-

ious spectroscopic techniques. In the present section, especially recent results of
spectroscopic evidence are reviewed.
3.3.1 Infra-red (IR) Spectroscopy
IR spectra of hydrogen-bonded systems have been studied extensively. It is popular

to examine a peak shift of XH stretching mode and to determine the thermodynamic
quantities of CH· · · π interaction [30]. Recently, direct observation of intermolecular
CH· · · π H-bonds of benzene with acetylene [31], benzene with methane [32], and
benzene with ethane [33] was reported by Fujii et al., using infrared-ultraviolet
double resonance spectroscopic technique in the gas phase. In Fig. 3.3, observed
52 O. Takahashi
Fig. 3.1 Histograms showing the distribution of CH hydrogen atoms (Dpln < 2.9 Å) against the
distance from the center of the ring (Dpx1 ). a Cl3 CH, b Cl2 CH2 , c sp-CH, d sp2 -CH (aromatic CH),
e sp2 -CH (aromatic CH, neutron data), f CCH3 . The data of only organic crystals with no disorder
and R ≤ 5 % were used for (d) and (f). Open bars: observed. Shaded bars: corrected by a factor
of 1/Dpx1 . Dpx1 : Horizontal distance of H from the center of the ring. (Reprinted (adapted) with
permission from the Chemical Society of Japan. Copyright 2001 the Chemical Society of Japan)
Fig. 3.2 Histograms showing the distribution of CH hydrogen atoms (Dpx1 < 1.4 Å, Dpln < 2.9 Å)
against the C–H· · · π access angle (α). a Cl3 CH, b Cl2 CH2 , c sp-CH, d sp2 -CH (aromatic CH), e
sp2 -CH (aromatic CH, neutron data), f CCH3 . The data of only organic crystals with no disorder
and R ≤ 5 % were used for (d) and (f). Open bars: observed. Shaded bars: corrected by a factor of
1/sinα. (Reprinted (adapted) with permission from the Chemical Society of Japan. Copyright 2001
the Chemical Society of Japan)
54 O. Takahashi
Fig. 3.3 IR spectrum of benzene-acetylene complex. Unperturbed vibrational frequencies of CH

stretching vibrations in bare acetylene are indicated by the dashed lines. (Reprinted (adapted) with
permission from American Chemical Society. Copyright 2004 American Chemical Society)
IR spectrum of benzene-acetylene complex is shown. The observed large peak of

3266.7 cm−1 was assigned to the ν3 band, which is the anti-symmetric CH stretching
vibration mode, and this band shows a remarkable low-frequency shift of 22 cm−1
from the unperturbed frequency of the bare molecule. This special feature of the
complex clearly demonstrates that the intermolecular bond between acetylene and
benzene is of a π H-bond type.
3.3.2 Nuclear Magnetic Resonance (NMR) Spectroscopy
NMR spectroscopy is one of the most powerful tools in studying chemistry; struc-
ture, thermodynamic properties, and reactivity. Various techniques were used to
demonstrate CH· · · π H-bond [30]. Plevin et al. [34] reported that weak H-bonds
such as OH, NH, and CH· · · π are important in three-dimensional structures of pro-
teins. Direct observation of CH· · · π H-bond in proteins was performed by NMR
spectroscopy, together with DFT calculations.
Zhao et al. [35] designed a system to investigate the difference of strength between
CH· · · π and CD (deuterium)· · · π H-bonds by 1 H NMR spectra and theoretical
calculations. They concluded that the D/H isotope effect is either very small or
nonexistent in the CH· · · π H-bond (Fig. 3.4).
3.3.3 Microwave Spectroscopy
Microwave spectroscopy is used to examine the rotational motion of a molecule, and

accurate molecular structures can be determined by analyzing spectral data. Direct
microwave spectra of complexes including CH· · · π H-bonds have been recorded.
Fig. 3.4 Scheme showing the unfolded, folded conformational equilibrium of the molecular bal-
ances, which can be used to measure changes in the strength of the intramolecular CH· · · π H-bond
in the folded conformer. (Reprinted (adapted) with permission from American Chemical Society.
Table 3.4 Structual r0 (Rcm ) rs ab initio

parameters for C6 H6 . . . C2 Ha2
Rcm / Åb 4.1546 4.1430 4.0387
c
RCH...π /Å 2.4921 2.4717 2.3694
a
Reference [38]
b
The center of mass distance
c
Perpendicular distance from H atom to the benzene plane
Tubergen et al. [36] observed rotational spectra of 1-phenyl-2-propanol, metham-

phetamine, and 1-phenyl-2-propanone. The lowest energy conformations of these
species were found to be stabilized by weak OH· · · π, NH· · · π, and CH· · · π in-
tramolecular H-bonds. The folded structure for 1-phenyl-2-propanol was predicted
by ab initio calculations previously [37]. Very recently, Ulrich et al. [38] studied
the rotational spectra of benzene-acetylene complex; their results are consistent
with intermolecular CH· · · π H-bonds observed by other spectroscopies and the-
oretical calculations. In Table 3.4, experimentally determined structural parameters
are compiled together with theoretical ones.
3.3.4 X-Ray Electron Spectroscopy
Recent progress in synchrotron radiation facility enabled to detect weak H-bonds in

molecule using a soft X-ray. Energy of soft X-ray is in the range of 100 eV to a few
keV, and 1s electron for most of the second and third row elements can be excited
directly. Binding energy of 1s electron calls core-electron binding energy (CEBE),
and its value is atom-specific, i.e., 290 eV for carbon and 410 eV for nitrogen, etc.
And even if the same atomic spices are included in a molecule, the CEBE is varied
with chemical environment of an excited atom. So such properties can be useful for
various chemical analysis, and are widely used as electron spectroscopy for chemical
analysis (ESCA) [39].
56 O. Takahashi
Fig. 3.5 The experimental

C1s photoelectron spectrum
of 1-pentyne (circles) is
shown together with a model
that takes into account both
anti and gauche confomers.
(Reprinted (adapted) with
permission from Elsevier B.V.
Copyright 2009 Elsevier B.V.)
Holme et al. [40] studied the conformations of 1-pentyne using X-ray photoelec-
tron spectroscopy (XPS) at the C K-edge. They decomposed the observed spectra
for anti and gauche conformations together with the theoretical calculations, and
their analysis yields 29 ± 3 % anti and 71 ± 3 % gauche conformers. In Fig. 3.5, the
experimental XPS spectrum of 1-pentyne and decomposed ones are depicted.
3.4 Consequent Phenomena Under CH· · · π H-bond
As described in the previous sections, although the CH· · · π H-bond is so weak and
is comparable to the thermal energy, it plays a critical effect in various phenomena
in organic chemistry.
3.4.1 Conformation
Stabilities of conformation are one of key factors to determine the molecular structure.
A lot of reviews have appeared relating to this issue [10, 11].
According to textbooks of organic chemistry, discussions regarding to the stability
of conformation start from anti-gauche form of n-butane, as shown in Fig. 3.6. The
C–C bond axis in the center of the molecule can rotate freely, and the anti conformer,
in which two methyl groups are located at the opposite sites is known to be more
stable than the gauche. This is interpreted generally as the steric repulsion between
two methyl groups, i.e., electrostatic repulsion between two bulkier moieties than
hydrogen atoms. However, for molecules by substituting one methyl group to another,
the anti is not necessarily more stable than the gauche due to the internal H-bond or
Fig. 3.6 Conformation of

n-butane. a anti and b gauche
forms
Table 3.5 Difference in the Y=0 Y = CHb2

Gibbs energy between the
anti- and gauche conformers X MP2 G3 MP2 G3
of CH3 -Y-CH2 -X (X; OH, OH 2.35 2.20 0.04 0.06
OCH3 , F, Cl, CN, CCH, and
C6 H5 : Y = O, CH2 ) at OCH3 2.86 2.45 1.30 0.11
MP2/6-311++ G(3df,3pd) F 4.02 3.76 0.12 0.12
with thermal corrections and
Cl 4.00 3.71 0.17 0.10
G3 theorya
CN 1.68 1.26 0.24 0.14
CCH 0.98 0.94 0.09 0.05
C6 H5 1.09c 0.07c
a
Reference [43]
b
This work
c
MP2/6-311G(d, p) level of approximation
the gauche effect [41]. It is well-known that the gauche for ethylenediamine is more
stable than the anti [42]. Takahashi et al. studied the stability of the gauche form
for CH3 -Y-CH2 -X system (X = OH, OCH3 , F, Cl, CN, CCH, and C6 H5 ; Y = O)
using high-level ab initio calculation [43]. Together with the result for Y = CH2 ,
difference in the Gibbs energy between the anti- and gauche conformers at MP2/6-
311++ G(3df,3pd) and G3 level of theory are listed in Table 3.5. Note that these
combinations include weak H-bonds such as CH· · · O, CH· · · N, and also CH· · · π.
The energy difference between the stabilization energy for X = O is larger than
that for X = CH2 because the distances between functional groups become shorter.
For these molecules, it is interpreted that an attractive force is operating between
functional groups.
Next intramolecular interactions related to the CH· · · π H-bond will be discussed.
Iitaka et al. reported the X-ray structure of 1-(p-bromophenyl)ethyl t-butyl sulphox-
ide; there, its t-butyl group was found to orient itself gauche to the phenyl group
[44]. Later, their group studied the conformation of related compounds, and found
that folded conformers are the most favored [[7], Fig. 3.7].
58 O. Takahashi
Fig. 3.7 Gauche form of

1-(p-bromophenyl)ethyl
t-butyl sulphoxide
They also carried out computational studies using molecular mechanics (MM)
techniques for 1-substituted 2-phenylpropanols to reproduce the order of the sta-
ble rotamers from Lanthanide-induced shift technique [7]. However, MM results
sometimes failed to reproduce the order of the rotamers because of the simplicity of
computational model. In the 2000s, Takahashi et al. tried high-level ab initio calcula-
tions at MP2 level of approximation for a series of compounds such as Ph-CH2 -X-R
and Ph-CHMe-X-R (X = CO, CH2 , O, S, SO, SO2 ; R = H, Me, Et, i-Pr, t-Bu) to find
stable rotamers and reproduce the order of them [45–49]. The rotamers with R· · · Ph
in gauche relationship are generally more stable than the R· · · Ph anti rotamers. Con-
trary to the MM calculations, the energy order of rotamers can be predicted more
accurately by ab initio calculations with electron correlation. The attractive CH· · · π
H-bond has been suggested to be a dominant factor in determining the conformation
of this series of compounds. Success of the use of ab initio calculations suggests that
the importance of the dispersion force is important, which can be described with the
above computational level (Fig. 3.8).
The CH· · · π H-bond operate not only in acyclic molecules but also in cyclic
molecules [50–53]. It is well-known that an isopropyl group in unsaturated
cyclohexane derivatives prefers to be axial for several terpenic compounds including
isomenthone 1 and isocarvomenthone 2. This unusual preference can be understood
by CH· · · π H-bonds operating between CHs in this groups and C = O π-system.
Takahashi et al. [50] calculated the conformational energy of these compounds.
They found that the most stable conformer has been found to have the axial isopropyl
group (Fig. 3.9).
Furthermore, the genesis of stabilization of the axial conformers in 2- and 3-
alkyl cyclohexanones, as compared to the structurally corresponding cyclohexane
derivatives, was sought in the context of the attractive CH. . . π(C = O) H-bond.
Fig. 3.8 Stable rotamers of Ph-CHMe-X-R (X = CO, CH2 , O, S, SO, SO2 ; R = H, Me, Et, i-Pr,
t-Bu)
Fig. 3.9 Conformation of isomenthone 1 and isocarvomenthone 2
Occasionally the folded conformations are observed in steroidal compounds.

Burgstahler et al. [54] reported that levopimaric acid 3 exists in the folded confor-
mation. The reason may be explained by intramolecular CH. . . π H-bond between
CH in the 10β angular methyl group and sp2 carbons of the conjugated diene ring.
Takahashi et al. [53] performed ab initio calculations using model compounds of 3
and concluded that the folded conformers were more stable than the extended ones
(Fig. 3.10).
60 O. Takahashi
Fig. 3.10 Folded and extended conformations of levopimaric acid 3
Fig. 3.11 Diels-Alder reactions of vinyl ketones and dienes
3.4.2 Reaction Selectivity
Folded conformer by weak intramolecular H-bond sometimes plays a critical role

in chemical reaction, i.e., stereoselective reactivity and chiral recognition. A great
number of examples have reported already and some of them were compiled in review
articles [9]. In the present, recent results are mainly introduced.
The Diels-Alder reaction is one of the most versatile synthetic transformations
for the construction of the cyclohexane framework. Carmona et al. [55] pre-
pared the aqua complexes [(η5 -C5 Me5 )M(PROPHOS)-(H2 O)][SbF6 ]2 (M = Rh, Ir;
PROPHOS = diphenylphosphane), and found that these complexes efficiently cat-
alyze the asymmetric Diels-Alder reaction between ketones and dienes. The structure
of the intermediate complexes of these reactions indicates that the CH· · · π H-bond
is operating between a phenyl group of the PROPHOS ligand and the α-vinyl proton
of the ketones (Fig. 3.11).
Allen et al. [56] examined NHC-catalyzed [4+2] cycloaddition between an eno-
late derived from α,β-unsaturated aldehydes or α-functionalized aldehydes and an
Fig. 3.12 Diels-Alder reactions of α,β-unsaturated aldehydes and an enone as the diene catalized
by N-heterocyclic carbenes
Fig. 3.13 Diastereo-differentiating reactions with LiAlH4
enone as the diene, and these reactions displayed remarkable diastereo- and enan-
tioselectivity, producing γ,δ-unsaturated δ-lactones in up to 99 % ee and greater than
20:1 de, as well as near-quantitative yields. They also examined mechanism using ab
initio calculations and concluded that an oxyanion-steering and a CH· · · π H-bond
play important roles to determine the high stereoselectivity (Fig. 3.12).
As have been done by Allen et al., computational chemistry can now explain this
kind of reaction mechanisms. To investigate the Diels-Alder reaction theoretically,
multireference procedure such as CASSCF may be used to reproduce a correct be-
havior of the reaction, because more than two electrons are contributed to the reaction
mechanism [57, 58]. However, target systems including the CH· · · π H-bond are still
too large for recent computer resources. Thus most of recent theoretical studies for
Diels-Alder reaction often used single reference methodology such as DFT.
As another reaction mechanism, a diastereo-differentiating reactions was inves-
tigated. Takahashi et al. [59] calculated the transition states of a model reaction of
alkyl 1-phenylethyl ketones with LiAlH4 . The transition-state geometries leading
to the predominant product are similar to those of the ground-state conformation.
In geometries leading to the minor product, the relevant torsion angles are twisted
to avoid unfavorable steric interaction. Short CH· · · π and CH· · · O distances sug-
gest that these weak hydrogen bonds are operating in stabilizing the structure of the
transition state (Fig. 3.13).
3.4.3 Molecular Recognition
During more than 40 years, a lot of studies related to molecular recognition operating
by CH· · · π interaction in liquid and solid phases have been accumulated, and the
knowledge contributed to the development of host-guest chemistry. Several inclusion
host compounds such as cyclodextrin, calixarene, cyclophane, pseudorotaxane, and
62 O. Takahashi
Fig. 3.14 X-ray structures of

4·MeOH. Dotted lines
represents intermolecular
H-bond. (Reprinted (adapted)
with permission from
American Chemical Society.
Copyright 2012 American
Chemical Society)
catenane have been synthesized. To interpret the driving force of the complex forma-
tion and the directionality, weak interactions such as π-π, CH· · · π, CH· · · O H-bond
are important [8].
In the solid phase, molecular orientation can be easily determined by X-ray and
neutron diffraction analysis. The molecular orientation is discussed with a help of
computational chemistry.
Morohashi et al. [60] synthesized powdery crystals of p-tert-butylthiacalix[4]
arene 4 which selectively include EtOH from 1:1 mixtures of MeOH-EtOH and
Et-PrOH, and EtCOOH from HCOOH-EtCOOH. Through CH· · · π and other H-
bonds, EtOH entered into two neighboring host molecules in a head-to-tail manner
to construct an infinite columnar structure along the c-axis (Fig. 3.14).
Goel et al. [61] reported CH· · · π H-bond-driven self-assembly in π-conjugated
skeletons based on oligophenylenevinylenes (OPVs) and trace the origin of interac-
tions at the molecular level by using single-crystal structures. OPVs were designed
with appropriate pendants in the aromatic core and varied by hydrocarbon or flu-
orocarbon tails along the molecular axis. Single-crystal structures of hydrocarbon
OPVs provided direct evidence for the existence of CH· · · π interaction (Fig. 3.15).
3.4.4 Chiral Recognition and Chiroptical Properties
Molecules bearing an asymmetric atom are optically active and rotate the plane of
polarized light. As a consequence, the absorption or emission intensity by left- or
right-handed light becomes different. Now circular dichroism (CD) which observe
Fig. 3.15 CH· · · π H-bond in

an OPV single crystal.
(Reprinted (adapted) with
permission from American
Chemical Society. Copyright
2004 American Chemical
Society)
difference of intensity between left- or right-handed light is popular. There has been
a long history in the investigation for the relationship between molecular structure
and chiroptical property. The theoretical background of chiroptical spectroscopy is
described in Ref. [62–64]. Various spectroscopy has been developed, such as elec-
tronic CD (ECD) and vibrational CD (VCD), respectively, for the visible/ultraviolet
and infrared region [65, 66].
CD spectral properties are useful in determining the absolute configuration of
optically active molecules. Several empirical rules were proposed. Moffitt et al. [67]
presented an empirical rule that the sign of the rotatory intensity of a n→ π* tran-
sition of a carbonyl chromophore in saturated chiral molecules: the so-called octant
rule. Later, Scott et al. [68] proposed another octant rule for the π → π* transition
of olefins. Recently, theoretical calculations by ab initio calculation have been de-
veloped, and calculations for relatively large molecules can be done to determine the
absolute configuration.
Intramolecular interaction between groups affects optical properties of molecule;
this is reflected on the CD spectra. In 1976, Burgstahler et al. [69, 70] presented
evidence that the effect of an axial alkyl group allylic to a C = C double-bond was
more important than that of the skewness of the diene chromophore (so-called diene
helicity rule). Introduction of an axial methyl group to a bridgehead carbon next to the
π-group has shown to enhance the CD amplitude of an exomethylene steroids, such
as 4- and 6-methylene-5a-estrane at ca. 200 nm. Takahashi et al. [71] investigated the
effect of a methyl group on the rotational strength of these steroids by time-dependent
density functional theory at the M06-2X/6-311 + + G(d, p)//MP2/6-31G(d, p) level
of approximation. Replacement of the bridgehead hydrogen atom of these steroids by
a methyl group influenced the CD amplitude at the π → π* transition. They [72] also
demonstrated that the theoretical CD amplitude of 1,3-cyclohexadiene compounds
increased by introduction of a methyl group at the bridgehead carbon. In Fig. 3.16,
calculated CD amplitudes of three model compounds of tricyclic 1,3-dienes are
shown. The result suggests that introduction of a methyl group increased the CD
amplitude at around 270 nm; this is fully consistent with the experimental data.
64 O. Takahashi
Fig. 3.16 Calculated CD

spectra of tricyclic 1,3-diene
model steroids. Reproduced
from Ref. [72] with
permission from the Centre
National de la Recherche
Scientifique (CNRS) and The
Royal Society of Chemistry
3.5 Summary and Perspective
More than 40 years have been passed since the weak H-bond such as CH· · · π in-
teraction was proposed for the first time. Now, this interaction is widely recognized
in physics, chemistry, and biology. Weak H-bonds play important roles in various
fields of organic chemistry. In the present chapter, several experimental and com-
putational results of the CH· · · π H-bond are introduced. The theoretical results
obtained by high-level ab initio and DFT calculations can support the experimental
data. Joint studies experiments and MO calculations will increase more and more to
realize reaction mechanism and biomolecules, which did not described in the present
chapter.
Acknowledgment The author thanks Dr. Motohiro Nishio for his continuous encouragements
and reading the manuscript. This work was supported by the Cooperative Research Program of the
Network Joint Research Center for Materials and Devices of MEXT, the Asahi Glass Foundation,
and a Grant-in-Aid for Scientific Research from JSPS. The author thanks the Information Media
Center at Hiroshima University for the use of a grid of high-performance PCs and the Research
Center for Computational Science in Okazaki, Japan.
References
1. Scheiner S (1997) Hydrogen bonding. Oxford Universty Press, New York

2. Nishio M, Hirota M, Umezawa Y (1998) The CH/π Interaction. Wiley-VCH, New York
3. Gilli G, Gilli P (2009) The nature of the hydrogen bond: outline of a comprehensive hydrogen
bond theory. Oxford University Press, Oxford
4. Nishio M, Umezawa Y, Fantini J, Weiss MS, Chakrabarti P (2014) CH-π hydro-
gen bonds in biological macromolecules. Phys Chem Chem Phys 16(25):12648–12683.
doi:10.1039/c4cp00099d
5. Pimentel GC, McClellan AL (1960) The Hydrogen Bond. Freeman, San Francisco
Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ (2011) Defin-
ing the hydrogen bond: An account (IUPAC Technical Report). Pure Appl Chem 83:1619.
doi:10.1351/pac-rep-10-01-01
7. Nishio M, Hirota M (1989) CH/π interaction: implications in organic chemistry. Tetrahedron
45(23):7201–7245. doi:10.1016/S0040-4020(01)89185-7
8. Nishio M (2004) CH/π hydrogen bonds in crystals. CrystEngComm 6(27):130–158.
doi:10.1039/b313104a
9. Nishio M (2005) CH/π hydrogen bonds in organic reactions. Tetrahedron 61(29):6923–6950.
doi:10.1016/j.tet.2005.04.041
10. Takahashi O, KohnoY, Nishio M (2010) Relevance of weak hydrogen bonds in the conformation
of organic compounds and bioconjugates: evidence from recent experimental data and high-
level ab Initio MO calculations. Chem Rev 110(10):6049–6076. doi:10.1021/cr100072x
11. Nishio M (2011) The CH/π hydrogen bond in chemistry. Conformation, supramolecules, opti-
cal resolution and interactions involving carbohydrates. Phys Chem Chem Phys 13(31):13873–
13900. doi:10.1039/C1CP20404A
12. Kitaura K, Morokuma K (1976) A new energy decomposition scheme for molecular in-
teractions within the Hartree-Fock approximation. Int J Quantum Chem 10(2):325–340.
doi:10.1002/qua.560100211
13. Reed AE, Curtiss LA, Weinhold F (1988) Intermolecular interactions from a natural bond
orbital, donor-acceptor viewpoint. Chem Rev 88(6):899–926. doi:10.1021/cr00088a005
14. Bader RFW (1990) Atoms in molecules: A quantum theory. Oxford University Press, Oxford
15. Jeziorski B, Moszynski R, Szalewicz K (1994) Perturbation theory approach to intermolec-
ular potential energy surfaces of van der waals complexes. Chem Rev 94(7):1887–1930.
doi:10.1021/cr00031a008
16. Chen W, Gordon MS (1996) Energy decomposition analyses for many-body interaction and
applications to water complexes. J Phys Chem 100(34):14316–14328. doi:10.1021/jp960694r
17. Williams HL, Chabalowski CF (2000) Using Kohn–Sham orbitals in symmetry-adapted per-
turbation theory to investigate intermolecular interactions. J Phys Chem A 105(3):646–659.
doi:10.1021/jp003883p
18. Bloom JWG, Raju RK, Wheeler SE (2012) Physical nature of substituent effects in XH/π
interactions. J Chem Theory Comput 8(9):3167–3174. doi:10.1021/ct300520n
19. Mishra BK, Karthikeyan S, Ramanathan V (2012) Tuning the C–H· · · π interaction
by different substitutions in benzene-acetylene complexes. J Chem Theory Comput.
doi:10.1021/ct300100h
20. Popelier P (2000) Atoms in molecules: an introduction. Prentice Hall
21. Koch U, Popelier PLA (1995) Characterization of C–H–O hydrogen bonds on the basis of the
charge density. J Phys Chem 99(24):9747–9754. doi:10.1021/j100024a016
22. Grabowski SJ (2001) A new measure of hydrogen bonding strength—ab initio and atoms in
molecules studies. Chem Phys Lett 338(4–6):361–366. doi:10.1016/S0009-2614(01)00265-2
23. Allen F (2002) The Cambridge structural database: A quarter of a million crystal structures and
rising. Acta Crystallographica Sect B 58(3 Part 1):380–388. doi:10.1107/S0108768102003890
24. Taylor R, Kennard O (1982) Crystallographic evidence for the existence of CH· · · O, CH· · · N
and CH· · · Cl hydrogen bonds. J Am Chem Soc 104(19):5063–5070. doi:10.1021/ja00383a012
25. Malone JF, Murray CM, Charlton MH, Docherty R, Lavery AJ (1997) X-H· · · π (phenyl)
interactions theoretical and crystallographic observations. J Chem Soc, Faraday Trans
93(19):3429–3436. doi:10.1039/a700669a
26. Umezawa Y, Tsuboyama S, Takahashi H, Uzawa J, Motohiro (1999) CH/π interaction
in the conformation of peptides. A database study. Biorg Med Chem 7(9):2021–2026.
doi:10.1016/S0968-0896(99)00123-6
27. Umezawa Y, Tsuboyama S, Takahashi H, Uzawa J, Nishio M (1999) CH/π interaction in the
conformation of organic compounds. A database study. Tetrahedron 55(33):10047–10056.
doi:10.1016/S0040-4020(99)00539-6
66 O. Takahashi
28. Takahashi H, Tsuboyama S, Umezawa Y, Honda K, Nishio M (2000) CH/π Interactions as

demonstrated in the crystal structure of host/guest compounds. A database study. Tetrahedron
56(34):6185–6191. doi:10.1016/S0040-4020(00)00575-5
29. Takahashi O, Kohno Y, Iwasaki S, Saito K, Iwaoka M, Tomoda S, Umezawa Y, Tsuboyama S,
Nishio M (2001) Hydrogen-bond-like nature of the CH/π interaction as evidenced by crystal-
lographic database analyses and ab initio molecular orbital calculations. Bull Chem Soc Jpn
74(12):2421–2430. doi:10.1246/bcsj.74.2421
30. Nishio M, Hirota M, Umezawa Y (1998) The CH/π interaction. Evidence, nature, and
consequences. Wiley-VCH, New York
31. Fujii A, Morita S, Miyazaki M, Ebata T, Mikami N (2004) A molecular cluster study on
activated ch/π interactions: infrared spectroscopy of aromatic molecule–acetylene clusters. J
Phys Chem A 108(14):2652–2658. doi:10.1021/jp049946b
32. Morita S, Fujii A, Mikami N, Tsuzuki S (2006) Origin of the attraction in aliphatic C-H/π
interactions: infrared spectroscopic and theoretical characterization of gas-phase clusters of
aromatics with methane. J Phys Chem A 110(36):10583–10590. doi:10.1021/jp064297k
33. Fujii A, Hayashi H, Tsuzuki S (2012) Preference of the monodentate contact in the CH/π
interaction between an alkyl group and a single phenyl ring: stable structures of benzene–ethane
clusters. Chem Phys Lett 537(0):11–15. doi:10.1016/j.cplett.2012.04.005
34. Plevin MJ, Bryce DL, Boisbouvier J (2010) Direct detection of CH/π interactions in proteins.
Nat Chem 2(6):466–471. doi:10.1038/nchem.650
35. Zhao C, Parrish RM, Smith MD, Pellechia PJ, Sherrill CD, Shimizu KD (2012) Do
deuteriums form stronger CH–π interactions? J Am Chem Soc 134(35):14306–14309.
doi:10.1021/ja305788p
36. Tubergen MJ, Lavrich RJ, Plusquellic DF, Suenram RD (2006) Rotational spectra and confor-
mational structures of 1-Phenyl-2-propanol, methamphetamine, and 1-phenyl-2-propanone. J
Phys Chem A 110(49):13188–13194. doi:10.1021/jp064810u
37. Takahashi O, Saito K, Kohno Y, Suezawa H, Ishihara S, Nishio M (2004) The conformation of
alkyl benzyl alcohols studied by ab initio MO calculations—A comparison with IR and NMR
spectroscopic data. Eur J Org Chem (11):2398–2403. doi:10.1002/ejoc.200300801
38. Ulrich NW, Seifert NA, Dorris RE, Peebles RA, Pate BH, Peebles SA (2014) Benzeneacetylene:
a structural investigation of the prototypical CH. . . π interaction. Phys Chem Chem Phys
16(19):8886–8894. doi:10.1039/c4cp00845f
39. Siegbahn K, Nordling C, Fahlman A, Nordberg R, Hamrin K, Hedman J, Johansson G,
Bergmark T, Karlsonn S, Lindgren I, Lindberg B (1966) ESCA-atomic, molecular and solid
state structure studied by means of electron spectroscopy. North-Holland Publishing Company
40. Holme A, Sæthre LJ, Børve KJ, Thomas TD (2009) Carbon 1s photoelectron spectroscopy of
1-pentyne conformers. J Mol Struct 920(1–3):387–392. doi:10.1016/j.molstruc.2008.11.035
41. Wolfe S (1972) Gauche effect. Stereochemical consequences of adjacent electron pairs and
polar bonds. Acc Chem Res 5(3):102–111. doi:10.1021/ar50051a003
42. Kazerouni MR, Hedberg L, Hedberg K (1994) Conformational analysis. 16. Ethylenediamine.
An electron-diffraction and ab initio investigation of the molecular structure, conforma-
tional composition, anti-gauche energy and entropy differences, and implications for internal
hydrogen bonding. J Am Chem Soc 116(12):5279–5284. doi:10.1021/ja00091a036
43. Takahashi O, Yamasaki K, Kohno Y, Ueda K, Suezawa H, Nishio M (2009) The origin of the
generalized anomeric effect: possibility of CH/n and CH/π hydrogen bonds. Carbohydr Res
344(10):1225–1229. doi:10.1016/j.carres.2009.04.011
44. Iitaka Y, Kodama Y, Nishihata K, Nishio M (1974) X-Ray structure determination of 1-(p-
bromophenyl)ethyl t-butyl sulphoxide. Configuration and conformation. J Chem Soc Chem
Commun (10):389–390. doi:10.1039/c39740000389
45. Takahashi O, Yasunaga K, Gondoh Y, Kohno Y, Saito K, Nishio M (2002) The conformation
of 2-phenylpropionaldehyde and alkyl 1-phenylethyl ketones as evidenced by ab initio calcu-
lations. Relevance of the CH/π and CH/O interactions in stereochemistry. Bull Chem Soc Jpn
75(8):1777–1783. doi:10.1246/bcsj.75.1777
46. Takahashi O, Gondoh Y, Saito K, Kohno Y, Suezawa H, Yoshida T, Ishihara S, Nishio M

(2003) The alkyl/phenyl-folded conformation of alkyl 1-phenylethyl sulfides and sulfones as
evidenced by ab initio MO calculations. Implication for the 1,2-asymmetric induction. New J
Chem 27(11):1639–1643. doi:10.1039/B303268J
47. Takahashi O, Kohno Y, Gondoh Y, Saito K, Nishio M (2003) General preference for
alkyl/phenyl folded conformations. Relevance of the CH/π and CH/O interactions to stere-
ochemistry as evidenced by ab initio MO calculations. Bull Chem Soc Jpn 76(2):369–374.
doi:10.1246/bcsj.76.369
48. Takahashi O, Kohno Y, Saito K, Nishio M (2003) Prevalence of the alkyl/phenyl-folded con-
formation in benzylic compounds C6 H5 CH2 -X-R (X = O, CH2 , CO, S, SO, SO2 ): significance
of the CH/π interaction as evidenced by high-level ab initio MO calculations. Chem Eur J
9(3):756–762. doi:10.1002/chem.200390084
49. Takahashi O, Saito K, Kohno Y, Suezawa H, Ishihara S, Nishio M (2003) The conformation
of 1-alkyl-2-phenylpropan-1-ols studied by ab initio MO calculations. Relevance of the CH/π
and OH/π hydrogen bonds. Bull Chem Soc Jpn 76(11):2167–2173. doi:10.1246/bcsj.76.2167
50. Takahashi O, Yamasaki K, Kohno Y, Kurihara Y, Ueda K, Umezawa Y, Suezawa H, Nishio M
(2008) The conformation of alkyl cyclohexanones and terpenic ketones. Interpretation for the
‘alkylketone effect’ based on the CH/π(CO) hydrogen bond. Tetrahedron 64(10):2433–2440.
doi:10.1016/j.tet.2007.12.052
51. Takahashi O, Yamasaki K, Kohno Y, Ueda K, Suezawa H, Nishio M (2008) Origin of the
axial-alkyl preference of (R)α-phellandrene and related compounds investigated by high-level
ab initio MO calculations. Importance of the CH/π hydrogen bond. Tetrahedron 64(24):5773–
5778. doi:10.1016/j.tet.2008.04.007
52. Takahashi O, Yamasaki K, Kohno Y, Ueda K, Suezawa H, Nishio M (2009) The origin
of the relative stability of axial conformers of cyclohexane and cyclohexanone derivatives:
importance of the CH/n and CH/π hydrogen bonds. Bull Chem Soc Jpn 82(2):272–276.
doi:10.1246/bcsj.82.272
53. Takahashi O, Yamasaki K, Kohno Y, Ueda K, Suezawa H, Nishio M (2009) The conformation
of levopimaric acid investigated by high-level ab initio MO calculations. Possibility of the
CH/π hydrogen bond. Tetrahedron 65(17):3525–3528. doi: 10.1016/j.tet.2009.02.014
54. Burgstahler AW, Ziffer H, Weiss U (1961) The configurations of levopimaric acid and α-
phellandrene; interpretation of their rotatory dispersions. J Am Chem Soc 83(22):4660–4661.
doi:10.1021/ja01483a039
55. Carmona D, Viguri F, Asenjo A, Lahoz FJ, García-Ordua P, Oro LA (2012) Enantioselective
catalytic Diels–Alder reactions with enones as dienophiles. Organometallics 31(12):4551–
4557. doi:10.1021/om300346s
56. Allen SE, Mahatthananchai J, Bode JW, Kozlowski MC (2012) Oxyanion steering and CH–π
interactions as key elements in an n-heterocyclic carbene-catalyzed [4 + 2] cycloaddition. J
Am Chem Soc 134(29):12098–12103. doi:10.1021/ja302761d
57. Li Y, Houk KN (1993) Diels-Alder dimerization of 1,3-butadiene: An ab initio CASSCF study
of the concerted and stepwise mechanisms and butadiene-ethylene revisited. J Am Chem Soc
115(16):7478–7485. doi:10.1021/ja00069a055
58. Sakai S (2000) Theoretical analysis of concerted and stepwise mechanisms of diels–alder reac-
tion between butadiene and ethylene. J Phys Chem A 104(5):922–927. doi:10.1021/jp9926894
59. Takahashi O, Saito K, Kohno Y, Suezawa H, Ishihara S, Nishio M (2004) Origin of the di-
astereofacial selectivity in the nucleophilic addition to chiral acyclic ketones. An ab initio MO
study. New J Chem 28(3):355–360. doi:10.1039/B310173H
60. Morohashi N, Noji S, Nakayama H, Kudo Y, Tanaka S, Kabuto C, Hattori T (2011) Unique
Inclusion Properties of Crystalline Powder p-tert-Butylthiacalix[4]arene toward Alcohols and
Carboxylic Acids. Org Lett 13(13):3292–3295. doi:10.1021/ol200506p
61. Goel M, Jayakannan M (2012) C–H/π-interaction-guided self-assembly in π-conjugated
oligomers. Chem Eur J 18(10):2867–2874. doi:10.1002/chem.201102670
62. Schellman JA (1968) Symmetry rules for optical rotation. Acc Chem Res 1(5):144–151.
doi:10.1021/ar50005a003
68 O. Takahashi
63. Buckingham AD, Stiles PJ (1974) Theory of natural optical activity. Acc Chem Res 7(8):258–
264. doi:10.1021/ar50080a004
64. Hansen AE, Bouman TD (1980) Natural chiroptical spectroscopy: theory and computations.
In: Adv Chem Phys 44 Wiley, pp 545–644. doi:10.1002/9780470142639.ch5
65. Nicu VP, Autschbach J, Baerends EJ (2009) Enhancement of IR and VCD intensities due to
charge transfer. Phys Chem Chem Phys 11(10):1526–1538. doi:10.1039/B816151H
66. Nicu VP, Heshmat M, Baerends EJ (2011) Signatures of counter-ion association and hydrogen
bonding in vibrational circular dichroism spectra. Phys Chem Chem Phys 13(19):8811–8825.
doi:10.1039/C0CP02701D
67. Moffitt W, Woodward RB, Moscowitz A, Klyne W, Djerassi C (1961) Structure and
the optical rotatory dispersion of saturated ketones. J Am Chem Soc 83(19):4013–4018.
doi:10.1021/ja01480a015
68. ScottAI, WrixonAD (1970)A symmetry rule for chiral olefins. Tetrahedron 26(15):3695–3715.
doi:10.1016/S0040-4020(01)92948-5
69. Burgstahler AW, Weigel LO, Gawronski JK (1976) Cotton effect behavior of skewed 1,3-
cyclohexadienes. Evidence for dominance of homoannular allylic chirality contributions. J
Am Chem Soc 98(10):3015–3016. doi:10.1021/ja00426a057
70. Burgstahler AW, Jahansouz H, Véliz EA, Takasugawa F (2002) Diene chirality and cotton
effects of nonhomoannular cisoid conjugated dienes: circular dichroism and crystal structure
of a steroidal 19-nor-1(10),9(11)-diene derived from westphalen’s diol diacetate. Chirality
14(2–3):180–186. doi:10.1002/chir.10067
71. Takahashi O, Yamasaki K, Kohno Y, Ueda K, Nishio M (2012) Importance of the CH/π hy-
drogen bond in the enhancement of CD amplitude of exomethylene steroids. RSC Advances
2(29):10891–10898. doi:10.1039/C2RA21446F
72. Takahashi O, Kohno Y, Ueda K, Nishio M (2013) Influence of the CH/π hydrogen bond on the
enhancement of circular dichroism (CD) amplitude of 1,3-cyclohexadiene compounds. New J
Chem 37(3):843–849. doi:10.1039/C2NJ40955K
Chapter 4
The CH··O H-Bond as a Determining Factor in
Molecular Structure
Steve Scheiner
Abstract The factors that affect the strength of CH··O hydrogen bonds (HBs) are
enumerated, along with the source of their stability, and the structural and spectro-
scopic features that signal their presence. Their influence upon a number of chemical
and biological processes is discussed. Within the context of proteins, the Cα H group
of protein residues can engage in CH··O HBs, as can the CH groups of a number of
amino acid side chains including aromatic residues. CH··O HBs have the potential
to make a major contribution to protein folding, particularly in an aqueous environ-
ment. These unconventional HBs may play as important a role in the formation of
protein β-sheet structures as do NH··O HBs. The strength of CH··O HBs is magnified
several-fold by the presence of positive charge on the proton donor. The implications
of these principles are discussed for a number of specific biological and chemical
problems, which include the catalytic mechanisms of methyltransferases and serine
proteases, and the structural properties of fluoroamides.
4.1 Introduction
Although it is widely thought that the first ideas of H-bonds (HBs) were limited to O,
N, and F as proton donor or acceptor atoms, the idea that the much less electronegative
C atom could also serve as donor appeared in the literature very soon after the concept
of a HB was originally proposed. Perhaps the first suggestion [1] was derived from
boiling point measurements. A related work [2] was published in 1937, arising from
studies of the liquid phase for systems like CHCl3 and ethers, followed 2 years
later by a study [3] that used this concept to explain the high dissociation constant
of o-toluic and n-butyric acids. Heat of mixing data for a range of molecules by
Marvel et al led this group to suggest [4] CH··X bonds in 1940, and to propose that
neighboring groups such as Cl, Br, and phenyl could activate the CH donor. The
idea was extended [5] to biological systems such as proteins like collagen shortly
thereafter. The 1950s yielded multiple confirmatory reports [6–12] that included not
S. Scheiner ()
Department of Chemistry & Biochemistry,
Utah State University, 84322-0300 Logan, UT, USA
e-mail: steve.scheiner@usu.edu
70 S. Scheiner
only additional heat of mixing results, but also NMR, IR, and X-ray data. More
accurate data of various sorts [13–22] continued to buttress the idea in the 1960s,
and greatly extended the range of systems in which they were observed to include
collagen and polyglycine among others.
The data emanating from the experimental work was highly suggestive of the for-
mation of CH··O HBs. X-ray diffraction data, for example, had documented the fairly
close approach of the C and O atoms. Unexpectedly high heats of mixing pointed to
stronger intermolecular attractions than might have been anticipated without invok-
ing CH··O HBs. Shifts of the bridging proton’s NMR chemical shift were consistent
with what was seen in conventional HBs, as were the changes in the CH stretching
frequency. But this evidence was inconclusive in that alternate explanations could be
offered. The proximity of the C and O atoms could be the result of crystal packing
forces, and not the product of a genuine attractive force, particularly as the position
of the bridging H was left unresolved by the refinement of the diffraction data. The
change in the CH stretching frequency might be dictated instead by intermolecular
geometry [11] or perhaps steric crowding [23].
Quantum chemical methods are well suited to fill this gap, and to definitively
answer the question as to whether a CH··O interaction is an attractive one, and to
quantitatively assess its strength. Prototype systems can be designed in which a pair
of molecules is held together solely by the interaction of a CH donor and O acceptor,
eliminating any ambiguities associated with larger systems with multiple points of
contact. The interaction energy can be dissected into its constituent components,
which can then be compared to those of conventional HBs of the OH··O variety. The
effects of the interaction upon the NMR chemical shielding or the C–H stretching
frequency can be analyzed so as to determine whether they are due to a HB or to some
other phenomenon. Even small changes in molecular geometry that are signatures of
HBs can be determined by these calculations, in many cases too small to be observed
experimentally. And the wave function derived from these calculations is amenable
to analysis in a number of ways, which can directly answer the question as to whether
or not a HB is present.
4.2 Early Calculations
Due to the generally weak nature of CH··O HBs, it was not until the 1970s that
quantum chemical techniques had matured to the point where it was reasonable to
take up this question. Even so, many of the earliest calculations [24–30] were limited
by the constraints at the time to semiempirical methods, small basis sets, or failure to
include electron correlation. Despite these limitations, the calculations supported the
idea of CH··O as an attractive force, even if the computational level was inadequate
for quantitative purposes. The coming of age of some of the newer quantum chemistry
codes, along with improved computers, catalyzed a blossoming of work in this area
in the 1980s. Relevant publications since that time are far too numerous to describe
in any detail here, so only a representative sample will be summarized.
4 The CH··O H-Bond as a Determining Factor in Molecular Structure 71
As a strong acid, it was no surprise that HCN could serve as a potent proton donor
[31–36], as could the triply bonded C in HC≡CH [37–40]. The sp2 -hybridized
C of the vinyl [41] or phenyl group [42–45] could also donate a proton, as could
other aromatic systems such as imidazole [46], pyrimidine [47] or nucleic acid
bases [48–50]. Other work concluded that the very simple CH4 could engage in a
HB [51–61], albeit a weak one. An early study of chlorosubstituted methanes [62]
indicated that the CH··O HB with water would be progressively strengthened by
each exchange of H with Cl. Other electron-withdrawing substituents like F, NO2 ,
etc have a similar strengthening effect [63–70].
4.3 Recent Work
It was at this point in time that this group became involved in this issue. Some of our
earlier work up through 2006 has been summarized already [71–73] so will only be
briefly outlined here, with emphasis on major findings. Beginning with the simplest
system which might contain a CH··O HB, the binding energy between CH4 and
OH2 , is very weak [74], on the order of 0.5 kcal/mol or less. But there is a strong
substituent effect in that electron withdrawing agents on the proton donor molecule
polarize the CH bond, making it a more potent donor. In numerical terms with each
replacement of a H atom on CH4 by F, the HB energy climbs by about 1 kcal/mol,
to the point where the F3 CH··OH2 HB is very nearly as strong as that in the water
dimer. As typical of HBs, the progressive strengthening is also accompanied by a
shortening; in this case R(H··O) contracts by about 0.14 Å with each F substitution.
Comparison of a number of computational protocols showed [74] that the trends are
largely independent of the level of theory, and even the quantitative data show an
insensitivity to basis set. These findings were later confirmed by others [75, 76] for
very similar systems. This idea of inductive effects strengthening a CH··O HB is not
limited to F, but has been demonstrated for a range of other electron-withdrawing
substituents like Cl [70, 77–79].
Unlike conventional HBs such as OH··N where the OH bond is stretched upon
formation of the complex, and its stretching frequency shifted to the red, there were
some indications that CH··O and CH··N HBs commonly behave in a contrary fashion
[80–86]. The work [74, 87] addressed the matter of C–H bond contraction and blue
shift, specifically to determine whether such behavior might exclude this interaction
from its HB classification. It was determined first that in all other respects, the CH··O
interaction fits into this rubric. Electron density shifts are characteristic of typical
HBs, and the NMR signal of the bridging proton shifts downfield [88] by an amount
proportional to the HB strength. Also in line with HB expectations are the components
of the interaction energy arising from an energy decomposition. In fact, this sort of
dissection provided some insight into the nature of this contrary behavior. All HBs,
whether conventional or CH··O, exhibit a large electrostatic attraction, which is
complemented by smaller polarization, charge transfer, and dispersion terms, all
balanced by a repulsive exchange energy. And in all cases, a small stretch of the
72 S. Scheiner
CH/OH covalent bond magnifies all of these factors. The difference resides only
in the amount of this magnification. The combined magnification of the attractive
components outweighs the greater repulsion energy when the OH bond is stretched,
albeit by only a small amount, so this bond is stretched when the OH··O bond is
formed. The opposite is true for CH··O where the enhanced repulsion overwhelms
the magnified attractive term, forcing the CH bond to contract. This central idea,
that the direction of shift of the CH stretching frequency resides in a fine balance
between two sets of opposing forces, has been confirmed by other research groups,
each using a different definition of these forces [78, 89–114].
As one might expect, O is not the only atom that can accept a CH proton. Cal-
culations [80, 87, 115–122] have shown that N serves an equivalent purpose. In
fact, due its greater basicity, CH··N HBs tend to be stronger than CH··O [123]. Like
their CH··O counterparts, the NMR signal of the bridging proton is also deshielded in
CH··N HB [41]. Despite the generic weakness of CH as a proton donor, computations
have shown that it is strong enough to engage F as an acceptor [79, 103, 124–131].
Other proton-accepting atoms are S [100, 132–136], even as strong as 7 kcal/mol
[137, 138], Cl [139–142] or I [143]. Even an atom as weakly basic as P [144, 145]
can serve this function as can carbenes [146]. And there is an extensive literature
concerning CH··π HBs, where the electron donor is an aromatic π system [147–152].
Such CH··π HBs can serve to guide structural changes as in N-heterocyclic carbene
palladium complexes [153].
Another major factor which influences the direction of C–H stretching frequency
shift is the hybridization of the C atom. Calculations [87] show that whereas the sp3
CH of alkanes commonly (but not universally) leads to a blue shift, the opposite
result of a red shift occurs for the sp-hybridized alkynes, which generally form
stronger CH··O HBs. Given their intermediate sp2 hybridization, and HB strengths
intermediate between alkanes and alkynes, it was not surprising to learn that alkenes
manifest only very small shifts, sometimes to the red and other times to the blue.
But again, whether alkane, alkene, or alkyne, all properties of these CH··O HBs, e.g.
NMR chemical shifts or electron density shifts, mimic those of conventional HBs
such as OH··O. Comparable trends were noted later with other systems [111, 132,
154], some of which contained the analogous CH··N HBs [155].
HBs are well known for their ability to reinforce one another. That is, a string of
n H-bonded molecules is commonly bound together by a force which exceeds that
occurring within n − 1 simple dimers, in what is frequently referred to as coopera-
tivity. A detailed examination [156] revealed that this same phenomenon is common
to CH··O HBs as well. This cooperativity is observed not only in the energetics,
but also in the HB lengths and in the electron density shifts resulting from the HB
formation. On the other hand, there is little evidence of cooperativity within the
context of CH covalent bond contraction or CH stretching frequency shifts. Whether
conventional or CH··O, the cooperativity in either sort of HB is reduced when the
system is immersed in a model solvent. The cooperative aspects of CH··O HBs have
been the subject of some inquiry by others as well [157–162]. The overriding conclu-
sion is that these HBs act much like any others in the sense that synergistic, positive
cooperativity will occur if a central molecule acts simultaneously as both electron
Fig. 4.1 Interaction energies

of each indicated proton
donor molecule with H2 O as
proton acceptor. Horizontal
axis Fo = (ε − 1)/(ε + 2)
measures polarizability of
surrounding medium, where ε
is dielectric constant.
B3LYP/6-31 + G** and
MP2/aug-cc-pVTZ results
indicated by solid and broken
lines, respectively. (Reprinted
with permission from
Scheiner and Kar [163].
Chemical Society)
donor and acceptor, while double donor or acceptor activity will result in an overall
weakening or negative cooperativity.
4.4 Biological Systems
Turning once more to the simple fluoromethanes as model proton donor, consid-
eration of how its CH··O HBs are affected by solvent provides some insight into
the situation within a large biomolecule such as a protein. Figure 4.1 shows how
the binding energies of any of the Fn CH4−n molecules with water are progressively
weakened as the polarizability of the surrounding medium is increased [163], and
the same is true of the conventional OH··O HB in the water dimer.
What are the implications of this observation for protein folding? The contribution
of any HB to folding may be thought of as the process that begins with the two subunits
disengaged in an unfolded protein, and thus both exposed to aqueous solvent. They
then approach one another and form a connecting HB, but now within the confines
of a protein, i.e. in a less polarizable medium than within water. In other words,
the impact of the HB is a combination of two conceptually separate processes. As
illustrated schematically in Fig. 4.2, the two subunits (1) form a HB within water,
and then (2) the H-bonded pair is removed from water and placed instead in a protein
interior. The exothermicity of the former process (1) is countered in part by the
endothermicity of the latter (2), sometimes referred to as a desolvation penalty.
How does this scenario play out for CH··O vs OH··O HBs? Whether in vacuo or
in solution, the CH··O HB is weaker [163] than is the conventional OH··O, so ΔE
74 S. Scheiner
Fig. 4.2 Representation of 1

formation of AH··B HB
within an aqueous A-H---B
environment, followed by 2
displacement of H-bonded
pair from water into the
interior of a protein.
(Reprinted with permission
from Scheiner and Kar [163]. (2)
Chemical Society) A-H B
(1)
A-H--- B
for process (1) for CH··O is less negative. In contrast, however, ΔE of step (2) is less
positive for CH··O. In other words, it takes less energy to move a CH··O bond from
water to a less polarizable medium, like a protein interior, than it does to desolvate a
OH··O HB. The net result is that it is more favorable for a CH··O HB to participate
in the protein folding process than for OH··O.
From a quantitative perspective, the binding energy of F3 CH + OH2 is 3.7
kcal/mol in vacuum, and diminished to 2.1 kcal in water; an intermediate value
of 2.9 kcal/mol occurs in a protein interior, modeled with dielectric constant of 4.
The corresponding values for the OH··O HB in the water dimer are all larger: 5.5,
3.1, and 4.1 kcal/mol. The desolvation penalty of the CH··O HB, viz. the energy
required to take this complex from water to protein is 3.3 kcal/mol, but this quantity
is higher for the OH··O HB, 5.4 kcal/mol. When these pieces are all assembled, the
contribution of the CH··O HB is 1.0 kcal/mol greater than that for OH··O.
Of course, a continuum homogeneous polarizable medium is only a rough ap-
proximation of the interior of a protein. In order to bring the model one step closer
to the real situation, a number of discrete water molecules were placed [163] around
the H-bonded systems, forming a first solvation sphere. HB energies in this primitive
solvated system were rather close to the same quantities computed with the dielectric
continuum model. A second issue with the latter solvation model is the choice as
to what value of dielectric constant most correctly models a protein interior. It is
not uncommon in the literature for a value of ε = 4 to be considered to simulate a
generic protein interior but of course each protein is different, and even within a
single protein, some domains will be more polarizable than others. But in any case,
the conclusion that a CH··O HB can be as much of a contributor to protein folding
as OH··O has a certain degree of experimental support [164].
4.4.1 Amino Acids
F3 CH + OH2 is only the roughest of models of a CH··O HB within a protein, even if it

does provide some fundamental insights. A first effort to study a more realistic model
donor started [165] with a set of amino acids, NH2 Cα HRCOOH, all of which contain
a Cα H group. Gly, Ala, Val, Ser, and Cys were all paired with a water molecule as
they engaged in a Cα H··OH2 HB. There was very little sensitivity to the nature of
the sidechain R, with binding energies all in the 1.9–2.5 kcal/mol range. Also rather
constant was the R(Cα ··O) distance, 3.31–3.35 Å, and the NMR downfield chemical
shift of the Cα H proton which varied between − 1.35 and − 1.71 ppm. There was a
bit more sensitivity in the contraction of the C–H bond, from 0.3 to 3.1 mÅ; all of
the C–H stretching frequencies were to the blue, in the 14–56 cm−1 range, so this
pattern fits the idea of blue stretches accompanying sp3 hybridization
In order to expand the scope to charged amino acids, Lys+ was modeled [165]
by the R sidechain of (CH2 )4 NH3 + and Asp− by CH2 COO− . As one might expect,
the presence of a positive charge, even one that is removed from the Cα H by a
hydrocarbon chain, enhanced the binding energy with OH2 up to 4.9 kcal/mol. But
perhaps surprisingly, this stronger bond did not result in much change in any of the
other parameters: R(Cα ··O), Δr(Cα H), and ΔσH were in line with the values obtained
for the neutral amino acids. Likewise for the anionic Asp− , even though its binding
to OH2 is much weaker.
4.4.2 Dipeptides
Within the context of the amino acid model, the Cα H is surrounded by a –NH2 on
one side and –COOH on the other. An expansion of each to a full peptide group leads
to a glycyl dipeptide CHONHCα H2 CONH2 that better represents the setting of this
central group within a protein. The ability of this Cα H group to participate in a HB
was tested [166], this time using the carbonyl O of formamide H2 NCHO as a more
representative proton acceptor within a protein.
The dipeptide model introduces a good deal of flexibility into the donor molecule.
This flexibility is represented primarily by the dihedral angles ϕ and ψ, that are
commonly used to denote the rotation of the two peptide groups around the C–Cα and
Cα –N bonds. This work [166] centered around the two conformations of dipeptides
that represent minima on their potential energy surface. The C7 minimum derives its
name from the presence of a seven-membered ring that contains an intramolecular
NH··O HB, while a five-membered ring occurs in the C5 structure. Note that neither
internal HB directly involves the Cα H2 group which is available to participate in a
CH··O HB with the neighboring formamide.
The interaction energy of a conventional NH··O HB with the formamide carbonyl
O atom of the C7 dipeptide is 7.5 kcal/mol, considerably larger than the 2.3 kcal/mol
of the Cα H··O HB. In the case of the C5 structure, however, the NH··O HB is cut
by a factor of three to only 2.5 kcal/mol, whereas the Cα H··O HB is slightly larger,
76 S. Scheiner
at 3.8 kcal/mol. In other words, the CH··O HB is stronger than the NH··O HB for
the C5 geometry of the dipeptide. This dipeptide structure with (ϕ,ψ) = (180◦ ,180◦ )
represents an extended structure of a polypeptide backbone, so is certainly an impor-
tant segment of the Ramachandran (ϕ,ψ) region, not far from the β-sheet geometry
(more about the β-sheet below).
What can account for this surprising and remarkable sensitivity of the NH··O HB
to the conformation of the dipeptide? The threefold reduction of this HB strength
is particularly puzzling as the geometry of this bond, including R(NH··O), is nearly
the same in the C5 and C7 conformers. A detailed inquiry [167] expanded the (ϕ,ψ)
conformational space of the dipeptide to cover the entire Ramachandran map, not
just the C5 and C7 areas, as shown in Fig. 4.3. It was found that the NH··O HB
energy is nearly uniform over the majority of the (ϕ,ψ) map, as indicated by the
red and orange sections of Fig. 4.3, but becomes progressively weaker as one moves
toward the fully extended (−180◦ ,180◦ ) structure in the upper left corner, at which
point the HB energy nearly vanishes entirely. The region of weakened NH··O HB
extends over a fairly wide area, which may be categorized as −180◦ < ϕ < − 100◦ ,
and 100◦ < ψ < 180◦ .
Careful scrutiny of the data [167] pointed to one particular feature as the prime
culprit. In the extended structure of the dipeptide, Fig. 4.4a shows that the proton
donor NH is close to the carbonyl O of the neighboring peptide unit. (It is in fact this
proximity which leads to the common characterization of this structure as C5.) A neg-
ative region of electrostatic potential emanates from this carbonyl O atom, which acts
as a shield of sorts, pushing an approaching proton acceptor away from the NH group
as indicated by the red double arrow in Fig. 4.4a, and thereby weakening the incipient
NH··O HB. As the dipeptide curls away from the (−180◦ ,− 180◦ ) extended structure,
the neighboring carbonyl O moves away from the NH, leaving this group exposed to
the approaching proton acceptor group, and allowing the NH··O HB to achieve its
normal potential, as for example in Fig. 4.4b where (ϕ,ψ) = (− 80◦ , + 80◦ ).
An important conclusion arising from this work is that it is incorrect to consider
the strength of any particular HB, NH··O or otherwise, as a given or constant. The
actual binding energy can be heavily influenced by the conformation adopted by the
protein, particularly if the structure places the proton donor group in the vicinity of
another group with a strong electrostatic potential.
Of course, the peptide units that interact with one another within proteins are
not necessarily adjacent along the polypeptide backbone. The folding of the protein
brings peptide groups from quite different segments of the backbone into close coinci-
dence, so it is important to consider a fuller range of interpeptide geometries than the
restricted low-energy sections of the (ϕ,ψ) space. A full search of the potential energy
surface of a pair of CH3 NHCOCH3 (NMA) molecules was thus undertaken [168] so
as to identify any minimum-energy structures, free of the restrictions imparted by a
connecting unit.
The primary minimum pictured in Fig. 4.5a unsurprisingly contained a strong
NH··O HB [168]. This bond was able to achieve full strength since there was no
peptide adjacent to the NH group whose carbonyl O could impede the approach of the
carbonyl from the other NMA molecule. The less expected result concerned a second
Fig. 4.3 Binding energy of

water to
CH3 CONHCH2 CONH2 , as a
function of internal dihedral
angles ϕ and ψ of glycyl
dipeptide. Numerical labels
indicate binding energy
(kcal/mol); Letters indicate
standard locations of a
α-helix, b 310 -helix, c
π-helix, d parallel β-sheet, e
antiparallel β-sheet, f 2.27
ribbon, g collagen triple
helix, h PPII, and i type II
β-bend. (Reprinted with
permission from Scheiner
[167]. Copyright 2007
American Chemical Society)
pair of minima, almost as stable as the first. In the first of these structures, Fig. 4.5b,
the planes of the two NMA molecules lie parallel to one another. Quite similar
in energy is Fig. 4.5c which contains an anti-parallel, but still stacked, geometry.
Analysis of the binding forces led to two primary components. First, there is a
π → π* transfer of charge from the π(CO) orbital of one molecule to the π*(CO)
antibonding orbital of the other amide, and vice versa. This attraction is augmented
by CH··O HBs between the methyl H atoms of one molecule and the carbonyl O
of the other. It is noted that this attractive force would be strengthened in a protein
where the –CH3 group would be replaced by a –CH2 -methylene, surrounded on both
sides by electron-withdrawing amide units. And further, that the directions of the
two NMA molecules are immaterial: both parallel and antiparallel geometries are
equally stable.
78 S. Scheiner
Fig. 4.4 Geometries optimized for dipeptide-water system, with dipeptide in a (ϕ,ψ))
(− 180◦ ,− 180◦ ) and b (− 80◦ ,80◦ ) conformations. (Reprinted with permission from Scheiner [167].
Fig. 4.5 Optimized geometries of N-methylacetamide dimers, with binding energies (kcal/mol)
shown in blue. Distances and angles in Å and degs. (Reprinted with permission from Adhikari and
Scheiner [168]. Copyright 2013 American Chemical Society)
4.4.3 Longer Chains
Given the variability of NH··O HB strengths in polypeptides, and its weakness in par-
ticular in extended conformations, one is naturally led to consider β-sheets, wherein
each strand adopts a fairly extended geometry. The conventional wisdom holds that
the strands are held together by interstrand NH··O HBs, but even an idealized visu-
alization of the structure of these sheets shows that CH groups of one strand lie in
close proximity to the carbonyl O atoms of the next strand. Could not the ensuing
CH··O HBs contribute to the stability of the interchain linkages, just as the NH··O
HBs do?
This question was specifically addressed [169] in both parallel and anti-parallel
configurations of β-sheets. Beginning with the geometry of a full double-strand extent
of anti-parallel polyglycine, a piece was excised such that each strand contains both
CH and NH donor groups, lying opposite the carbonyl O of the other strand. The
HCONHCH2 CONH2 dimer was then optimized to yield both NH··O and CH··O HBs.
The former were shorter than the latter, with R(NH··O) = 1.97 Å and R(CH··O) =
2.57 Å, as displayed in Fig. 4.6. But as shown above, the length can be deceiving, as
a short HB is not necessarily a strong one. It was necessary to extract the energetic
contribution of each sort of HB, separate from the others. The question of how to
disentangle individual interactions, when both are present simultaneously, has been
a thorny problem for some time.
The issue was addressed [169] by removing one proton donor group at a time,
and then computing the interaction energy of the remaining dimer. As indicated in
Fig. 4.6a, the NH··O HB was deleted by replacing the terminal NH2 groups on each
strand by a H atom, leaving a CH stump that is both a weak proton donor, and too far
away from the carbonyl O to engage in a HB in any case. The interaction between
the two HCONHCα H2 COH molecules consists only of Cα H··O HBs. The CH··O
HBs were deleted by simply removing the central CH2 group of each strand, which
leaves only a set of HCONH2 molecules, bound only by NH··O HBs, as indicated in
Fig. 4.6b. It should be stressed that the geometry of the dimer was left unchanged by
each deletion, to avoid contamination of the results by changes in HB geometry. In
this manner, it was determined that the pair of NH··O HBs contribute 14 kcal/mol,
as compared to 10 kcal/mol for the CH··O HBs.
The location where the strands are cut is an arbitrary one. If this location is
shifted down a bit, the energetic contributions of the NH··O HBs drop from 14 to
only 10 kcal/mol, placing it precisely on a par with the CH··O HBs. And it should be
stressed that this result is not an artifact of the use of a pair of dipeptides; extension to
tripeptides has no substantive effect on the results. The situation is somewhat different
in the case of a parallel β-sheet. An equivalent partitioning into separate NH··O and
CH··O HB energetic contributions shows that the latter makes a larger contribution
than does the former. Specifically, the 8.3 kcal/mol contribution from the pair of
NH··O HBs is superseded by a 9.5 kcal/mol contribution from the CH··O HBs.
In summary, whichever model is adopted, whether parallel or anti-parallel,
whether dipeptide or tripeptide, and wherever the cut is made to excise the sys-
tem for study, it can hardly be said that the interstrand binding is solely due to NH··O
HBs. The CH··O HBs clearly make contributions which are comparable to, and
perhaps even larger than, those of NH··O HBs.
Later calculations verified some of the primary conclusions. For example, the
importance of the Cα H···O = C HB to the interaction energy in a β-sheet was sup-
ported by examinations [170, 171] of β-sheet models of (Gly)n and (Ala)n which
found CH··O HB to be almost as strong as NH··O in the anti-parallel structure, but
stronger in parallel. Additional confirmation came from Guo et al. in 2009 [172]. In
that same year, the Cα H··O HB was computed to be stronger than NH··O in dipeptide
and tripeptide models of the parallel β-sheet geometry of Ala [173]. In fact, CH··O
HBs have shown some propensity to stabilize the α-helix as well [173–175]. There
80 S. Scheiner
Fig. 4.6 Atom excisions made in HCONHCH2 CONH2 dimer. a Replacement of terminal NH2
groups by H, to leave only CH··O HBs. b Removal of central CH2 groups, leaving behind only
NH··O. (Reprinted with permission from Scheiner [169]. Copyright 2006 American Chemical
Society)
is also experimental support for the importance of CH··O HBs is β-sheets [176]; the
β-pleated sheet structure for β-sulfidocarbonyls [177] contains not only CH··O, but
also CH··S HBs.
With the realization that CH··O HBs have the potential to make significant contri-
butions to the stability of fully formed β-sheets, the next question might relate to their
ability to actually influence the formation of such sheets. This subject was addressed
[178] by the design of a novel molecule in Fig. 4.7a which contains first a dipep-
tide segment –NHCOCH2 NHCOCH3 which has the full (ϕ,ψ) Ramachandran space
Fig. 4.7 System which

contains both upper dipeptide
and lower peptide unit. a
Schematic diagram showing
atomic labeling. b Global
minimum with interatomic
distances in Å, c secondary
minimum with ψ = 150◦ .
from Adhikari and Scheiner
[178]. Copyright 2013
available to it. Another segment (the lower one in Fig. 4.7a) of the same molecule
contains a –CH2 NHCOCH3 peptide unit with which the first dipeptide might en-
gage in NH··O and/or CH··O H-bonding. These two segments are both attached to a
phenyl/ether connector unit which holds them together, but whose nature and rigidity
prevents its participation in any HBs of its own.
82 S. Scheiner
The upper dipeptide segment has its natural disposition in terms of preferred
conformations, in the absence of the rest of the molecule. As a dipeptide, the two
minima on its potential energy surface are the C5 and C7 structures, with the latter
lying 1 kcal/mol higher in energy than the former [178]. Within the framework of
the full molecule, this native conformational surface undergoes some important and
fundamental changes. These perturbations can be understood on the basis of the
interactions that occur between the upper dipeptide and the lower peptide segment
in the full molecule.
While C7 remains the global minimum, its stability can no longer be understood
on the basis of a simple intramolecular NH··O HB. Instead, HBs with the lower
peptide segment replace this internal C7 HB as the most important component. As
indicated in Fig. 4.7b, in addition to the conventional interpeptide NH··O HB, with
R(H··O) = 2.33 Å, there are a pair of CH··O HBs, with R(H··O) distances of 2.42
and 2.48 Å, which contribute heavily to its stability. The C5 configuration loses its
status as a minimum in the isolated dipeptide, and in fact becomes a maximum on
the surface of the full molecule. A new minimum in Fig. 4.7c appears in its place,
with ψ = 150◦ , which resembles a C5 structure, but like the global minimum, is also
dependent upon an interpeptide CH··O HB. Of course, this system is too small to
adopt a full β-sheet structure, but the results clearly indicate the importance of CH··O
HBs, not only to the stability of the fully formed sheet, but also to the process of
“zipping up” this sheet from separate strands.
4.4.4 Amino Acid Side Chains
Of course, the Cα H groups of the polypeptide skeleton are not the only ones capable
of engaging in a CH··O HB. The amino acid sidechains also contain CH protons,
some of which are situated near electron withdrawing groups that ought to impart to
them an added potency. Those situated near the positively charged termini of the Lys
and Arg residues come immediately to mind, as do the Cβ H hydrogens of Ser that
are adjacent to a hydroxyl group. A particularly interesting class are the aromatic
CH atoms of Phe, His, Tyr, and Trp whose sp2 hybridization should enhance their
potency, particularly if they lie near an electron-withdrawing atom.
The ability of these aromatic CH groups to engage in a CH··O HB was assessed
[179], and placed in the context of other HBs with which these side chains might
participate. In particular, the CH··O HBs were compared to conventional NH··O,
OH··N HBs, as well as OH··π interactions where the π system of the aromatic group
serves as electron donor; water was used as the partner molecule.
As anticipated the conventional HBs were the strongest, with binding energies
varying between 4 and 7 kcal/mol. OH··π HBs were weaker, between 2 and 4
kcal/mol, and CH··O slightly weaker still lying in the 1–2 kcal/mol range. The
HB lengths correlated with these binding energies, with R(H··O) varying between
2.9 and 3.0 Å for conventional HBs, up to 3.3–3.4 Å for CH··O. Consistent with
the sp2 hybridization of the proton donor C atom, both contractions and stretches
were observed for the CH··O HBs, and stretching frequency changes were small,
and both red and blue. As in the case of other HBs, the isotropic NMR signal
of the bridging CH··O proton shifted downfield. The OH··π proton’s NMR shift
went in the opposite direction, due to the usual magnetic field currents of the
aromatic ring above which it lies. The CH protons lying in closer proximity to
electron-withdrawing atoms, such as the N atoms of His, displayed a somewhat
greater propensity to engage in CH··O interactions.
Of particular interest was the effect of placing a charge on the aromatic proton
donor. The protonated imidazole model of His formed a very strong HB, amounting
to 10–11 kcal/mol. Consonant with this greater strength was a HB length that was
reduced by 0.3 Å, and a strong blue shift of its CH stretching frequency, of 78–
118 cm−1 . The NMR shift of this CH proton was also enhanced, by a factor of 2.
Other work has supported these ideas. A CH+ ··O HB of protonated imidazole was
in part responsible [180] for the self-assembly of a triple helical structure.
An early calculation of the interaction involving a methylpyridinium cation with
dimethyl ether [181] also found a strong CH+ ··O HB with a binding energy of as
much as 13 kcal/mol. Regarding other amino acid side chains, the Cδ H group of
proline engages in CH··O HBs, within the context of real protein geometries [182].
4.4.5 Effect of Charge
Indeed placing a charge on either subunit has been known for some time [183, 184]
to amplify the binding in HBs, so it is natural to expect that the same ought to
be true for CH··O HBs as well. And in fact, there was some evidence this might
be true in a few cases [185, 186]. This idea was probed systematically [187] in a
series of systems that all employed the carbonyl O of N-methylacetamide (NMA)
as the proton acceptor. Beginning with neutral proton donors S(CH3 )2 and N(CH3 )3
as a point of reference, both formed optimized complexes with NMA in which a
CH of each methyl group present was engaged in a CH··O HB. The total binding
energies amounted to 4.9 and 2.1 kcal/mol, respectively for the S and N systems.
The situation changed dramatically when an extra methyl group was added to each,
so that the new donors were the S(CH3 )+ +
3 and N(CH3 )4 cations. The various CH··O
HBs all contracted by 0.2–0.4 Å, and the binding energies rose to 20.6 and 18.8
kcal/mol, for the S and N systems, respectively, an amplification by a factor of 4–9.
(The optimized geometries for the amines are illustrated in Fig. 4.8a, 4.8b) These
HBs are very strong, stronger than any neutral conventional HBs, and in fact on a
par with ionic HBs of the OH+ ··O or NH+ ··O sort.
It is understood that displacing any HB from a gas-phase situation to a solvated
environment will weaken it. And such was found [187] to be the case with all HBs
studied, ionic as well as neutral. Indeed, the binding strengths of the ionic systems
were more dramatically reduced with increasing dielectric constant ε of the surround-
ing polarizable continuum model of solvation. But nevertheless, the ionic systems
remained more tightly bound than the neutral complexes, even for high ε of 78 that
simulates water.
84 S. Scheiner
Fig. 4.8 Optimized geometries of indicated amines with N-methylacetamide (NMA). Distances
in Å. (Reprinted with permission from Adhikari and Scheiner [187]. Copyright 2013 American
Chemical Society)
As the methyl groups on these cations are lengthened to ethyl, propyl, etc, the
alkyl H atoms occur at various distances from the heteroatomic center of charge.
For example, the methylene CH2 hydrogens lie adjacent to N/S, while the terminal
methyl protons are removed by one additional C–C linkage. This distinction was
found [187] to make an important difference. If the CH··O HBs involved the terminal
methyl groups, then each chain lengthening, i.e. methyl → ethyl → propyl, very
substantially reduced the binding energy with the NMA acceptor. Taking the N
series as an example, the binding energy of 20.3 kcal/mol of N(CH3 )+ 4 was reduced
to 14.1 kcal/mol for N(Et)+ +
4 , and then to 10.5 kcal/mol for MeN(Pr)3 . As displayed in
Fig. 4.8, this reduced binding energy is accompanied by elongations of the relevant
R(CH+ ··O) HBs. This decrease is much more gradual if instead of terminal methyl
groups, the HBs are rather formed with the CH2 protons lying adjacent to S/N.
The elongation to ethyl and then to propyl results in binding energies of 18.2 and
17.5 kcal/mol, greatly diminished reductions. One might anticipate that the positive
charge on CH protons drops as one moves along the alkyl chain further from the
heteroatomic center of charge, which would help to explain this distinction. And
indeed, careful scrutiny of the electrostatic potential confirmed this suspicion.
Since these are charged systems, it is tempting to presume that the two entities are
bound together primarily by Coulombic attraction. And indeed a SAPT decompo-
sition of the interaction energy reveals a strong electrostatic component, exceeding
20 kcal/mol. However, one cannot ignore other attractive forces: induction and dis-
persion together contribute between 13 and 16 kcal/mol. The induction signals its
presence in a number of ways, that also confirm that there are indeed bona fide HBs
present in these ionic complexes. First of all, there are large NBO values of E(2)
for the charge transfer from the proton-accepting O atom to the σ*(CH) antibond-
ing orbitals of the donors, as much as 18 kcal/mol, characteristic of HBs. Second,
maps of electron density redistribution that accompany complexation contain the
characteristic trademarks of HBs: losses of density around the bridging proton and
gains in the area of the lone pair of the proton acceptor atom. The identity of these
interactions as true HBs is further confirmed by downfield shifts of the NMR signal
of the bridging protons. These shifts are as large as 2 ppm for the ionic systems, more
than twice the magnitude of the corresponding quantities in the neutral counterparts.
As indicated above, the optimal arrangements of these ionic systems contain a
trifurcated HB wherein one H atom from each of three different methyl groups inter-
acts directly with the proton acceptor O. The binding is weakened if this trifurcation
involves three protons from the same methyl group, by as much as 35 %. This re-
duction likely has a geometric cause in that there is a good deal of deviation of each
θ(CH··O) angle from linearity when all H atoms come from the same methyl group.
This sort of nonlinearity can be avoided if there is but a single proton involved in the
CH··O bond. While this single linear HB is slightly superior to a trifurcated interac-
tion with a single methyl group, it remains weaker than the optimal arrangement of
three CH··O HBs arising from three separate methyls.
4.5 Specific Examples of Biological Implications
4.5.1 Methyltransferases
The strength of ionic CH··O HBs has some direct applications to biochemistry. As
one example, consider methylation of proteins which is essential to the metabolism of
amino acids, cofactors, hormones, lipids, nucleic acids, and polysaccharides as well
as covalent modification of DNA and proteins. Many of the enzymes that carry out
this function are dependent upon S-adenosylmethionine (AdoMet) which transfers
its methyl group to the acceptor substrate via an SN 2 reaction, wherein the C of the
transferring methyl adopts a planar transition state. There has been some evidence
that the process is aided by CH··O HBs in the active site of SET domain class of
protein lysine methyltransferases. These HBs help to enhance the binding of AdoMet,
to hold the transferring methyl group in its proper orientation, and to stabilize the
transition states partial positive charge.
This issue was addressed [188] by combining quantum calculations with ex-
perimental measurements. Human SET7/9 lysine methyltransferase, was chosen as
86 S. Scheiner
representative of this class of proteins. This system displays two highly conserved
AdoMet methyl CH··O HBs that are conserved in the SET domain methyltransferase
class. One of the proton acceptors is the backbone carbonyl oxygen atom of His-293,
whereas the second acceptor is the hydroxyl group of Tyr-335, the invariant tyro-
sine in the active site of SET domain enzymes. The tyrosine OH accepts a number
of CH··O HBs from the AdoMet methyl group, methylene groups, and adenine C8
atom, in addition to donating a conventional HB to a carbonyl group of Ala-295 in
the enzyme active site.
The calculations were carried out by modeling the AdoMet donor by S(Et)2 Me+ ;
a N-methylacetamide (NMA) molecule was used in place of the His-293 peptide
linkage that engages in a HB with the AdoMet, and phenol replaced the full Tyr-335
residue. In order to first derive a sense of how strongly the AdoMet might bind to each
proton acceptor, each was allowed to separately interact directly with S(Et)2 Me+ .
An optimization of the AdoMet··NMA pair yielded a geometry much like that in the
earlier work [187] with the very similar S(Et)3+ , involving three CH··O HBs to the
NMA oxygen, with R(H··O) of 2.2 Å, and a binding energy of 20.5 kcal/mol. The
O of phenol is a weaker proton acceptor, which binds to the AdoMet model by 7.4
kcal/mol. The three HB lengths vary between 2.4 and 2.9 Å.
Previous studies of SET domain methyltransferases had demonstrated that the
invariant tyrosine is critical to enzyme function but its roles in catalysis and AdoMet
recognition remained unresolved. For example, mutation of this residue to a pheny-
lalanine (Y335F) severely impaired AdoMet binding affinity, demonstrating the
importance of the CH··O hydrogen bonds to substrate recognition. In order to exam-
ine this question, the Tyr was mutated to Phe in the experiments; this mutation was
modeled in the calculations by replacing the phenol model by benzene, and also by
the isosteric aniline. Experimental analysis demonstrated that these mutations retain
the structure and protein substrate binding properties of the wild-type enzyme.
The effects of charge were first modeled [188] by comparing the binding of the
charged AdoMet model S(Et)2 Me+ with its neutral S(Et)2 analog. The active site
model also contained phenol in place of Tyr335, NMA modeled Ala-295, and the
adenine group of the cofactors, as displayed in Fig. 4.9a, b (adenine has been deleted
for purposes of clarity). Consistent with experimental data, the binding energy of
the ionic sulfonium was stronger than that of its neutral analogue by 11 kcal/mol.
Breakdown of the total into pairwise interaction energies allowed this difference to
be traced to the attraction of the sulfonium to the NMA (Ala) and phenol groups.
The H-bond distances in Fig. 4.9a, b support this idea: there are two CH··O distances
in Fig. 4.9a that are shorter than 2.4 Å, both marked by broken red lines. This
selectivity, i.e. the difference in binding energy between sulfonium and thioether,
was considerably reduced when phenol was replaced by benzene. As indicated in
Fig. 4.9c, d, the HB lengths for the charged and uncharged AdoMet model differ
very little, consistent with the much smaller selectivity. The primary cause of this
change is the loss of HBs involving the OH group of phenol. This loss is particularly
striking in the cases of the CH··O HBs involving the sulfonium.
The replacement of the phenol by aniline substitutes –OH with –NH2 , which
in principle can also serve as both proton donor and acceptor. So it was initially
puzzling that the aniline mutant, like benzene, also lost selectivity for the sulfonium.
Fig. 4.9 Optimized geometries of model active site of SET7/9, using phenol model of Tyr in (a and
b), and benzene model of Phe in (c and d). S(Et)2 Me+ models AdoMet in (a and c); neutral S(Et)2
in (b and d). HB lengths in Å. (Reprinted with permission from Horowitz et al. [188]. Copyright
And indeed, the N forms a strong CH··N HB with the sulfonium, with R(CH··N)
only 2.30 Å in length. A more thorough analysis revealed, however, that this HB
is not without cost. In order to accommodate this HB, the amine group rotates 58◦
about the plane of the aniline ring (compared to the optimized monomer) into a less
energetically favorable position. Accordingly, there is an energetic penalty incurred
by HB formation in the active site, thus sacrificing the lowest-energy aniline amine
conformation in order to better accommodate CH··N hydrogen bonding to AdoMet.
88 S. Scheiner
There was a second factor observed as well. The limited angular range of the amine
group compared with the OH of phenol eliminates the ability for the aniline amine
group to simultaneously serve as an effective CH··N HB acceptor to the adenine
group, and negates the CH··O HB which occurs for phenol.
The calculations [188] yielded another unanticipated observation about the
SET7/9 active site: the electrostatic repulsion between the AdoMet sulfonium cation
and the C8 of adenine, the most acidic carbon atom in the purine ring system. Remov-
ing the charge from the sulfonium cation alleviated this repulsion. This conformation
is dissimilar to that found in solution and in other methyltransferase classes. This
observation raises the intriguing possibility that the different AdoMet binding con-
formations in various methyltransferase classes may serve to tune the substrate’s
methyl transfer reactivity and merits further investigation.
4.5.2 Serine Proteases
Another application arises in connection with the catalytic mechanism of the serine
protease class of enzymes. One common feature of these enzymes is the presence of
what has come to be called a charge relay system. An Asp-His-Ser triad of residues
are situated adjacent to one another in the active site. The Asp is thought [189, 190]
to initially be in its anionic –COO− state, which acts through the intermediacy of
the His imidazole ring to facilitate the ability of the Ser Oγ nucleophile to attack a
C atom of the substrate. The His residue is thought to act in part by picking up a
proton in one stage of the catalytic cycle and delivering it to another site later. The
orientation of this His ring is thus very important for this process.
There had been some strong indications that a CH··O HB between the Cε1 H of His
and the O of a neighboring Ser residue could be an important functional component
of the mechanism [191–193]. This idea led to the proposal [194, 195] of a “ring-flip”
mechanism involving a 180◦ rotation of the His. A rotation of this sort is disfavored
in the initial state of the enzyme, as it would delete one of the four HBs in which the
His is thought to participate, two of which are of the CH··O variety. Upon formation
of the tetrahedral intermediate, on the other hand, the His becomes protonated which
now permits retention of all 4 of these HBs, thus facilitating the rotation of the
imidazole ring of His, and the entire enzymatic process, by better positioning it for
the next step in this reaction.
These ideas were tested in a set of calculations that modeled each of the relevant
residues by a smaller, and computationally tractable, molecule. The business end of
His-57 was modeled by methylimidazole, Ser-195 by ethanol, the aspartate residue
by CH3 COO− , and the peptide group of Ser-214 by formamide [196]. In order to
permit the necessary geometry optimizations that allow the groups a range of motion,
while maintaining the positions of each of these groups within the active site structure,
a couple of atoms of each residue were frozen in their X-ray coordinate positions.
The initial optimized structure is illustrated in Fig. 4.10a, which does indeed
suggest a Cε1 H··O HB to be present, along with the expected conventional HBs. A
flip of the ring, to the rotated configuration of His-57 in Fig. 4.10b, does indeed
result in a minimum, but this process is rather endothermic, on the order of 15
kcal/mol. While the Cε1 H does not have a proton acceptor partner in 10b, the same
was true of Cδ2 H in 10a. More importantly, the potent Nδ1 H donor has lost its partner
in b. Consistent with the hypothesis, the ring flip is more favorable energetically
after the tetrahedral intermediate has been formed, going from Fig. 4.10c–4.10d. In
connection with Ser-214 and its purported CH··O HB with His-57, although this
HB appears to be quite weak in the initial structure, it does appear to help stabilize
the tetrahedral intermediate. Further, the CH··O HB acts to hold the His ring in a
position appropriate for the reaction to proceed. Arguing against the validity of the
ring-flip are the small populations of the relevant configurations, based on Boltzmann
factors. All told, while providing some structural data, the calculations were unable
to establish unequivocally the validity of the ring-flip mechanism.
One unanticipated finding of the calculations relates to the second O atom of the
Asp-102, the one that is not itself involved in a HB to the Nδ1 of His-57. This atom
serves as proton acceptor in a CH··O HB with the Cβ H of His-57, with a length of
2.47 Å in Fig. 4.10a. This HB persists throughout the catalytic cycle, whether or not
the His-57 ring has flipped around by 180◦ . This HB is fairly strong with R(H··O) in
the 2.4–2.6 Å range.
In another biomolecular context, CH··O HBs appear to be quite prominent in
oligosaccharides and carbohydrates [197] where they can amount to some 40 %
of the total interaction energy, and can lock particular configurations [198]. The
interactions between RNA bases and phosphate include CH··O HBs [199] as do 2-
deoxyribonucleosides [200, 201]. Studies of a model protein backbone combined
with nucleic acid bases showed evidence of CH··O HBs as well [202].
4.5.3 Other Systems
The importance of CH··O to biomolecular structure and function has become more
clear with each passing year. A statistical analysis [203] of protein-ligand complexes
provided geometrical evidence of their widespread occurrence. There is evidence
of their role also in interhelical packing forces [204]. The Cα H· · ·O = C HB has
been shown to be a major driving factor in interhelical interactions [205], where Gly
residues are beneficial due to the lesser steric repulsions arising from the lack of an R
group. This finding reinforces the earlier idea [206] that fibrils of poly-Glu are stabi-
lized by bifurcated H-bonds and the frequency [207] with which Gly participates in
CH··O HBs at protein-ligand interfaces. Quadrupole couplings and 2 H solution NMR
data suggest Cα H· · ·O = C HBs in ubiquitin [208]. They have been proposed [209]
to act as the principal driving force for folding of β,γ-hybrid model peptide and of the
structure of amicyanin [210]. Ultrahigh resolution neutron diffraction data of proteins
[211] have noted CH··O HBs with an average HB length of 2.0 Å. These bonds are
essential ingredients of oligosaccharides and carbohydrates where they have been es-
timated [198] to account for as much as 40 % of the total interaction energy. In another
related case, there appear to be as many as 24 CH··O HBs in cellotetraose [212].
90 S. Scheiner
Fig. 4.10 Optimized structures relevant to ring-flip hypothesis of serine proteases. Residue numbers
are displayed in (a) for each model unit. a and b represent the initial binding, before and after His-
57 flip, respectively. c and d following formation of tetrahedral intermediate. Distances in Å.
(Reprinted with permission from Scheiner [196]. Copyright 2008 American Chemical Society)
4.6 Implications for Organic Chemistry
4.6.1 Fluoroamides
There is a strong tendency of the F atom in an α-fluoroamide to adopt a position

anti to the carbonyl O. This trend is likely due at least in part to an electrostatic
repulsion between the F and O atoms, both of which carry a partial negative charge.
Another contributing factor may be a weak HB between the NH of the amide and
the neighboring F atom [213, 214]. But as with any other general tendency, one can
envision over-riding factors that might push the system in another direction. Sup-
pose, for instance, that the molecule containing the α-fluoroamide, had appended to
it another group, one which contained a proton-acceptor atom. Consider the confor-
mation wherein this proton acceptor sits opposite the carbonyl O of the fluoroamide.
In this position, it would of course be unable to interact in an attractive way with the
F atom. But if the CH2 F group of the fluoroamide were to rotate around its C–C bond
such that one of the H atoms might approach the acceptor, one has the possibility
of a CH··O HB. The strength of this HB might be sufficient to hold the F atom up
near the carbonyl O, in violation of the general tendency for an anti configuration.
Further, if a second electron-withdrawing F atom were added to the terminal group,
the resulting CHF2 group would likely be an even stronger proton donor, enhancing
the intergroup HB, and magnifying the push away from the usual anti structure.
These ideas were tested in a combined quantum chemical and experimental ap-
proach [215]. The first order of business was to assess the native trend toward the
anti configuration in the absence of any intergroup HB. The potential energy surface
for a rotation around the C–C bond in CH3 NHCOCH2 F was calculated and the anti
structure was indeed found to be its minimum, as displayed in Fig. 4.11a. Internal ro-
tation to place one of the CH2 F protons opposite the carbonyl O costs 6 kcal/mol. An
analogous calculation of the difluorosubstituted CH3 NHCOCHF2 shows this system
(Fig. 4.11b) also prefers to keep a F atom anti to O; the cost to rotate a H atom into
this anti position is 4 kcal/mol.
Can an intergroup CH··O be strong enough to counter this native pull? In order to
answer this question, the fluoroamide was placed onto a larger molecule, to which was
also added a carbamate group whose carbonyl O could serve as proton acceptor, as
displayed in Fig. 4.12a. Focusing first on the monofluoro CH2 F group, full geometry
optimization led to a global minimum with the F anti to the carbonyl O, as would
be expected were there no proton accepting group. However, in contrast to the small
CH3 NHCOCH2 F model, there is a second minimum as well, one which contains a
CH··O HB to the carbamate, with R(CH··O) = 2.49 Å, as shown in Fig. 4.12a. And
whereas such a structure is 6 kcal/mol higher in energy than the anti configuration
in the small model, it lies only 0.9 kcal/mol higher in this larger system. It might
be concluded then that the presence of the CH··O HB in the secondary minimum
stabilizes this structure by the difference, some 5 kcal/mol.
As indicated above, the addition of a second F atom ought to strengthen any
intergroup CH··O HB. The same system was therefore again studied, but with the
CH2 F replaced by CHF2 . The CH··O HB has now been empowered to reverse the
native tendency for an anti conformation, and the structure shown in Fig. 4.12b with
the intergroup CH··O HB now represents the global minimum. The anti structure is
now relegated to secondary minimum status, even if only higher in energy by 0.1
kcal/mol.
To be sure that it is the intergroup HB that causes this reversal, and not some
artifact of the large connecting group, the carbamate was deleted [215] and the
geometry optimized. Confirming the influence of the CH··O HB, in its absence the
geometry reverts to the anti conformation displayed in Fig. 4.12c, with no hint of a
minimum when the H atom lies opposite the carbonyl O. Another verification of the
92 S. Scheiner
Fig. 4.11 Rotational profiles

computed for a monofluoro
and b difluoroamides.
from Jones et al. [215].
Chemical Society)
influence of the CH··O HB arises when the terminal CHF2 is replaced by a simple
methyl group, as in Fig. 4.12d. The absence of an electron-withdrawing substituent
would severely weaken any CH··O HB. And indeed, following this substitution, the
methyl group can rotate freely, with a nearly flat rotational profile.
The energy difference between the conformations with the CH syn and anti to the
carbonyl O offers only one estimate of the CH··O HB energy. There are of course
other factors that influence this energy difference, e.g. differing interaction between
the neighboring NH and either the CH or CF, depending upon the conformation
chosen, or interactions between the CF2 H and the terminal methyl group on the
Fig. 4.12 Optimized geometries of molecules containing fluoroamide units; insets focus on the
rotation around the C–C bond of the fluoroamide. (Reprinted with permission from Jones et al.
[215]. Copyright 2012 American Chemical Society)
carbamate. A second opinion regarding this quantity of interest, as it were, was

obtained by focusing only on the groups participating in the CH··O HB, and deleting
the remaining groups and their complicating effects.
More specifically, employing a scheme that was also used to analyze protein
β-sheet energetics [169], the upper CH donor was reduced to only a CHOCFH2
molecule and the lower carbamate to HCOOH. Importantly, both of these groups
were forced to retain the precise alignment which they had in the full molecular
system. So for example, the R(CH··O) distance of 2.49 Å in the fully optimized
molecule was held constant in the smaller CHOCFH2 /HCOOH dimer, as illustrated
in Fig. 4.13a. The same is true of other geometrical aspects, e.g. the θ(CH··O) angle
and even internal bond lengths. This procedure not only eliminates spurious and non-
relevant interactions, but also provides a complex containing two separate molecules,
whose interaction energy can be computed unambiguously, something which is not
possible when the two groups are parts of the same molecule.
The binding energies of these complexes, which are directly attributable solely to
the CH··O HB, were found in this manner to be 3.0 kcal/mol for the monofluorosub-
stituted molecule, and 3.5 kcal/mol for the analogous disubstituted system. It was
gratifying to note that the latter value is quite close to the 4.0 kcal/mol obtained in the
less direct way in the full molecule, by comparing syn-anti energy differences in the
94 S. Scheiner
Fig. 4.13 Means employed to estimate CH··O HB energy, freezing geometry of upper and lower
segments as they are in the full molecule on the left. Monofluoro and difluoro derivatives in (a
and b), respectively. (Reprinted with permission from Jones et al. [215]. Copyright 2012 American
Chemical Society)
full molecule vs the same difference in the small model. While the calculated results
might be questioned, experimental data can be more convincing. The conclusions
arrived at by the computations were fully supported by the X-ray geometries [215] of
a host of relevant mono and difluoroamides, some with and some without the proton-
accepting carbamate group. This combined computational and experimental study
thus provided strong evidence that a CH··O HB can be strong enough to influence
the conformation adopted by an organic system.
Additional evidence derives from other work, for example, that showed that CH··O
HBs are instrumental in making the equatorial isomer of tropinone, a tropane alka-
loid, dominant in its complex with water [216]. There is a CH··O/S intramolecular
HB that influences the conformation of Meldrum’s acid derivatives, a bond that
persists in the gas phase [137]. CH··O HBs can control the stereochemistry during
palladium-catalyzed arylation and vinylation of lactones [217, 218]. They are part of
the mechanism by which aminophosphonate diesters adopt their equilibrium struc-
tures [219]. Intramolecular bonds of this type can be utilized to induce aromatic
1,2,3-triazole oligomers to form [220] folded and helical secondary structures, con-
taining a 18 Å diameter cavity. In terms of catalytic function and design, CH··O
HBs are intimately involved in binding a substrate in thiourea organocatalysis [221].
Bonds of this sort also lower the barrier to H transfer of keto-enol tautomerization
in β-cyclohexanedione [222] and differentially dictate the conformation in solid and
solution of substituted cyclohexanes, driving them toward the boat structure [223],
or [224] enhance chiral recognition.
4.7 Perspective
It is clear that the CH group can be a potent donor within the context of HBs. Moti-
vated by early experimental indications of the presence of CH··O HBs, computations
over recent years have provided solid support for their presence, as well as important
details. These interactions behave in most respects like other more conventional HBs:
electron-withdrawing substituents strengthen their proton-donating power, as does
the change in hybridization from sp3 to sp2 and then to sp. Their strength derives
primarily from a combination of Coulombic attraction and charge transfer from the
O lone pairs into the σ*(CH) antibonding orbital. Also like standard HBs, CH··O
interactions display cooperativity when aligned properly.
Just like any HB, CH··O bonds can have profound influence on molecular struc-
ture. Calculations suggest that they actively participate in the folding of proteins,
particularly in the assembly of β-sheets. In addition to the ubiquitous Cα H group
on the protein backbone, the sidechains of numerous protein residues also engage
in CH··O HBs. Regardless of the HB considered, its strength is very substantially
magnified when the proton donor carries a positive charge. This effect is attenuated
as the CH donor occurs further from the center of charge, but is still present even
when several atoms removed.
CH··O HBs are directly involved in the catalytic mechanism of several enzymes,
including methyltransferases and serine proteases as specific examples. This inter-
action is also capable of influencing the conformation of organic molecules, with
mono and di-fluoroamides taken as simple examples.
It seems clear that the study of CH··O HBs will continue to accelerate, from both
experimental and computational perspectives. As in the past, numerous cases will
emerge where these interactions have been present all along but unrecognized and
unappreciated. Future work will likely identify new aspects of these noncovalent
bonds that will guide the synthesis of molecular systems with novel and useful
properties.
Acknowledgments I am indebted to my coworkers whose names are listed in the references of

the individual papers, and who made very substantial contributions. Collaborators Professors Ray
Trievel and Martin Smith inspired the work on the methyltransferases and fluoroamides. Some of
this work was supported financially by the NIH GM57936 and NSF CHE-1026826.
References
1. Kumler WD (1935) The effect of the hydrogen bond on the dielectric constants and boiling
points of organic liquids. J Am Chem Soc 57:600–605
2. Glasstone S (1937) The structure of some molecular complexes in the liquid phase. Trans
Faraday Soc 33:200–214
3. Dippy JFJ (1939) The dissociation constants of monocarboxylic acids; their measurement
and their significance in theoretical organic chemistry. Chem Rev 25:151–211
4. Marvel CS, Copley MJ, Ginsberg E (1940) Hydrogen bonds involving the C–H link. XI. Effect
of structure on bonding of donor and acceptor molecules. J Am Chem Soc 62:3109–3112
5. Huggins ML (1943) The structure of fibrous proteins. Chem Rev 32:195–218
6. Kosolapoff GM, McCullough JF (1951) Comparison of hydrogen bonding abilities of some
organic compounds of phosphorus. J Am Chem Soc 73:5392–5393
7. Dougill MW, Jeffrey GA (1953) The structure of dimethyl oxalate. Acta Cryst 6:831–837
8. Parry GS (1954) The crystal structure of uracil. Acta Cryst 7:313–320
9. Huggins CM, Pimentel GC, Shoolery JN (1955) Proton magnetic resonance studies of
chloroform in solution: evidence for hydrogen bonding. J Chem Phys 23:1244–1247
10. Pinchas S (1955) Infrared absorption of the aldehydic C–H group. Anal Chem 27:2–6
11. Schneider WG, Bernstein HJ (1956) Molecular association and infra-red spectrum of solid
formaldehyde and acetaldehyde. Trans Faraday Soc 52:13–18
96 S. Scheiner
12. Pinchas S (1957) Infrared absorption of aldehydic C–H group. Ortho-substituted benzalde-
hydes. Anal Chem 29:334–339
13. Brown I, Smith F (1960) Liquid-vapour equilibria. X. The systems acetone + nitromethane
and acetone + acetonitrile at 45 ◦ C. Austr. J Chem 13:30–37
14. Sutor DJ (1962) The C–H... O hydrogen bond in crystals. Nature 195:68–69
15. Sutor DJ (1963) Evidence for the existence of C–H... O hydrogen bonds in crystals. J Chem
Soc 1105–1110
16. Allerhand A, Schleyer PvR (1963) A survey of C–H groups as proton donors in hydrogen
bonding. J Am Chem Soc 85:1715–1723
17. Pinchas S (1963) Intramolecular hydrogen bonding in o-nitrobenzaldehyde and related
compounds. J Phys Chem 67:1862–1865
18. Ferguson G, Tyrrell J (1965) C–H... O hydrogen bonding. Chem Commun 195–196
19. Bednowitz AL, Post B (1966) Direct determination of the crystal structure of β-fumaric acid.
Acta Cryst 21:566–571
20. Krimm S (1967) Hydrogen bonding of C–H... O = C in proteins. Science 158:530–531
21. Krimm S, Kuroiwa K (1968) Low temperature infrared spectra of polyglycines and
C–H. . . O = C hydrogen bonding in polyglycine II. Biopolymers 6:401–407
22. Ramachandran GN, Chandrasekharan R (1968) Interchain hydrogen bonds via bound water
molecules in the collagen triple helix. Biopolymers 6:1649–1658
23. Forbes WF (1962) The study of hydrogen bonding and related phenomena by spectro-
scopic methods. Part VII. Intramolecular hydrogen bonding and steric interactions in
o-nitrobenzaldehyde and related compounds. Can J Chem 40:1891–1898
24. Morokuma K (1971) Molecular orbital studies of hydrogen bonds. III. C = O... H–O hydrogen
bond in H2 CO... H2 O and H2 CO... 2H2 O. J Chem Phys 55:1236–1244
25. van Duijneveldt-van de Rijdt JGCM van Duijneveldt FB (1971) Perturbation calculations on
the hydrogen bonds between some first-row atoms. J Am Chem Soc 93:5644–5653
26. Johansson A, Kollman P, Rothenberg S (1972) Hydrogen bonding and the structure of the
HF-HCN dimer. Chem Phys Lett 16:123–127
27. Johansson A, Kollman P, Rothenberg S (1972) The electronic structure of the HCN dimer and
trimer. Theor Chim Acta 26:97–100
28. Bonchev D, Cremaschi P (1974) C–H group as proton donor by formation of a weak hydrogen
bond. Theor Chim Acta 35:69–80
29. Kollman P, McKelvey J, Johansson A, Rothenberg S (1975) Theoretical studies of hydrogen-
bonded dimers. Complexes involving HF, H2 O, NH3 , HCl, H2 S, PH3 , HCN, HNC, HCP,
CH2 NH, H2 CS, H2 CO, CH4 , CF3 H, C2 H2 , C4 H4 , C6 H6 , F− , and H3 O+ . J Am Chem Soc
97:955–965
30. Vishveshwara S (1978) Ab-initio molecular orbital studies on CH. . . X hydrogen bonded
systems. Chem Phys Lett 59:26–29
31. Frisch MJ, Pople JA, Del Bene JE (1983) Hydrogen bonds between first-row hydrides and
acetylene. J Chem Phys 78:4063–4065
32. Somasundram K, Amos RD, Handy NC (1986) Ab initio calculation for properties of hydro-
gen bonded complexes H3 N... HCN, HCN... HCN, HCN... HF, H2 O... HF. Theor Chim Acta
69:491–503
33. Hinchliffe A (1986) Ab initio study of the hydrogen-bonded complexes NH... 3 HCN,
PH... ... ... ... ...
3 HCN, AsH3 HCN, H2 O HCN, H2 S HCN, H2 Se HCN. J Mol Struct (Theochem)
136:193–199
34. Chattopadhyay S, Plummer PLM (1990) Ab initio studies on the dimer of sulfur dioxide and
hydrogen cyanide. J Chem Phys 93:4187–4191
35. Peralta JE, Ruiz de Azua MC, Contreras RH (1999) Natural bond orbitals analysis of C–H... O
interactions in NCH/H2 O and NCH/OCH2 , and their effect on nuclear magnetic shielding
constants. J Mol Struct (Theochem) 491:23–31
36. Heikkilä A, Pettersson M, Lundell J, Khriachtchev L, Räsänen M (1999) Matix isolation and
ab initio studies of 1:1 hydrogen-bonded complexes HCN–H2 O and HNC–H2 O produced by
photolysis of formaldoxime. J Phys Chem A 103:2945–2951
37. Turi L, Dannenberg JJ (1993) Molecular orbital studies of C–H... O H-bonded complexes. J
Phys Chem 97:7899–7909
38. Schulz B, Botschwina P (1996) Ab initio calculations for propyne and the hydrogen-bonded
complex NH... 3 H–CaC–CH3 . Mol Phys 89:1553–1565
39. Simon S, Duran M, Dannenberg JJ (1996) How does basis set superposition error change the
potential surfaces for hydrogen-bonded dimers? J Chem Phys 105:11024–11031
40. Philp D, Robinson JMA (1998) A computational investigation of cooperativity in weakly
hydrogen-bonded assemblies. J Chem Soc Perkin Trans II 1643–1650
41. Afonin AV, Vashchenko AV, Takagi T, Kimura A, Fujiwara H (1999) Specific intramolecular
interactions C–H... N in heteroaryl vinyl ethers and heteroaryl vinyl sulfides studied by1 H,13 C,
and15 N NMR spectroscopies and by ab initio calculations on molecular structures as well as
on nuclear shieldings. Can J Chem 77:416–424
42. Cense JM, Agafonov V, Ceolin R, Ladure P, Rodier N (1994) Crystal and molecular structure
analysis of flutamide. Bifurcated helicoidal C–H... O hydrogen bonds. Struct Chem 5:79–84
43. Kim K, Friesner RA (1997) Hydrogen bonding between amino acid backbone and side chain
analogues: a high-level ab initio study. J Am Chem Soc 119:12952–12961
44. Hobza P, Spirko V, Selzle HL, Schlag EW (1998) Anti-hydrogen bond in the benzene dimer
and other carbon proton donor complexes. J Phys Chem A 102:2501–2504
45. Samanta U, Chakrabarti P, Chandrasekhar J (1998) Ab initio study of energetics of X-H... π (X
= N, O, and C) interactions involving a heteroaromatic ring. J Phys Chem A 102:8964–8969
46. Ornstein RL, Zheng Y (1997) Ab initio quantum mechanics analysis of imidazole C–H. . . O
water hydrogen bonding and a molecular mechanics forcefield correction. J Biomol Struct
Dyn 14:657–665
47. McCarthy W, Smets J, Adamowicz L, Plokhotnichenko AM, Radchenko ED, Sheina GG,
Stepanian SG (1997) Competition between H-bonded and stacked dimers of pyrimidine: IR
and theoretical ab-initio study. Mol Phys 91:513–525
48. Kratochvil M, Engkvist O, Sponer J, Jungwirth P, Hobza P (1998) Uracil dimer: potential en-
ergy and free energy surfaces. Ab initio beyond Hartree-Fock and empirical potential surfaces.
J Phys Chem A 102:6921–6926
49. Shishkin OV, Sponer J, Hobza P (1999) Intramolecular flexibility of DNA bases in adenine-
thymine and guanine-cytosine Watson-Crick base pairs. J Mol Struct 477:15–21
50. Hobza P, Sponer J (1999) Structure, energetics, and dynamics of nucleic acid base pairs:
nonempirical ab initio calculations. Chem Rev 99:3247–3276
51. Woon DE, Zeng P, Beck DR (1990) Ab initio potentials and pressure second virial coefficients
for CH4 –OH2 and CH4 –SH2 . J Chem Phys 93:7808–7812
52. Novoa JJ, Whangbo M-H, Williams JM (1990) Ab initio computational study of the C–
H... donor and C–H... anion contact interactions in organic donor salts. Mol Cryst Liquid
Cryst 181:25–42
53. Legon AC, Roberts BP, Wallwork AL (1990) Rotational spectra and geometries of the gas-
phase dimers (CH4 ,HF) and (CH4 ,HCl). Chem Phys Lett 173:107–114
54. Novoa JJ, Tarron B, Whangbo M-H, Williams JM (1991) Interaction energies associated with
short intermolecular contacts of C–H bonds. Ab initio computational study of the C–H..O
contact interaction in CH4 ..OH2 . J Chem Phys 95:5179–5186
55. Wiberg KB, Waldron RF, Schulte G, Saunders M (1991) Lactones. 1. X-ray crystallographic
studies of nonanolactone and tridecanoloactone: nature of CH... O nonbonded interactions. J
Am Chem Soc 113:971–977
56. Novoa JJ, Whangbo M-H, Williams JM (1991) Interaction energies associated with short
intermolecular contacts of C–H bonds. Structure and energetics of the interaction between
CH4 and CN− . Chem Phys Lett 177:483–490
57. Sennikov PG, Sharibdjanov RI, Khoudoinazarov K (1992) The molecular orbital study of the
structures and energies of bimolecular complexes of CH4 and SiH4 with water. J Mol Struct
270:87–97
58. Szczesniak MM, Chalasinski G, Cybulski SM, Cieplak P (1993)Ab initio study of the potential
energy surface of CH4 –H2 O. J Chem Phys 98:3078–3089
98 S. Scheiner
59. Vizioli C, de Azúa MCR, Giribet CG, Contreras RH, Turi L, Dannenberg JJ, Rae ID, Weigold
JA, Malagoli M, Zanasi R, Lazzeretti P (1994) Proximity effects on nuclear spin-spin coupling
constants. 1.1 J(CH) couplings in the vicinity of an atom bearing lone pairs. J Phys Chem
98:8858–8861
60. Novoa JJ, Planas M, Rovira MC (1996) On the usefulness of the counterpoise method on
hydrogen-bonded complexes: a numerical test using near complete basis sets on H2 O. . . HF,
(H2 O)2 , (HF)2 , and CH4 . . . H2 O. Chem Phys Lett 251:33–46
61. Rovira C, Novoa JJ (1997) Strength and directionality of the C(sp3 )–H... S(sp3 ) interaction.
An ab initio study using the H2 S... CH4 model complex. Chem Phys Lett 279:140–150
62. Hobza P, Sandorfy C (1984) Quantum chemical and statistical thermodynamic investigations
of anesthetic activity. 3. The interaction between CH4 , CH3 Cl, CH2 Cl2 , CHCl3 , CCl4 , and
an O–H. . . O hydrogen bond. Can J Chem 62:606–609
63. Kumpf RA, Damewood JR (1988) Do nitromethane and malonitrile form C–H... O hydrogen
bonds? Implications for molecular recognition by crown ethers. J Chem Soc Chem Commun
621–622
64. Neuheuser T, Hess BA, Reutel C, Weber E (1994) Ab initio calculations of supramolecular
recognition modes. Cyclic versus noncyclic hydrogen bonding in the formic acid/formamide
system. J Phys Chem 98:6459–6467
65. Erickson JA, McLoughlin JI (1995) Hydrogen bond donor properties of the difluoromethyl
group. J Org Chem 60:1626–1631
66. Turi L, Dannenberg JJ (1995) Molecular orbital studies of the nitromethane-ammonia
complex. An unusually strong C–H... N hydrogen bond. J Phys Chem 99:639–641
67. Alkorta I, Maluendes S (1995) Theoretical study of CH... O hydrogen bonds in H2 O–CH3 F,
H2 O–CH2 F2 , and H2 O–CHF3 . J Phys Chem 99:6457–6460
68. Novoa JJ, Mota F (1997) Substituent effects in intermolecular C(sp3 )–H... O(sp3 ) contacts:
how strong can a C(sp3 )–H... O(sp3 ) hydrogen bond be? Chem Phys Lett 266:23–30
69. Boldeskul IE, Tsymbal IF, Ryltsev EV, Latajka Z, Barnes AJ (1997) Reversal of the usual
ν(C–H/D) spectral shift of haloforms in some hydrogen-bonded complexes. J Mol Struct
436:167–171
70. Cubero E, Orozco M, Luque FJ (1999) Electron density topological analysis of the C–H... O
anti-hydrogen bond in the fluoroform-oxirane complex. Chem Phys Lett 310:445–450
71. Scheiner S (2000) CH· · · O hydrogen bonding. In: Hargittai M, Hargittai I (eds) Advances in
molecular structure research, JAI Press, Stamford, pp 159–207
72. Scheiner S (2005) The CH--O hydrogen bond. A historical account. In: Dykstra C, Frenking
G, Kim K, Scuseria G (eds) Theory and applications of computational chemistry: the First 40
years, Elsevier, Amsterdam, p. 831–857
73. Scheiner S (2006) Contribution of CH··X hydrogen bonds to biomolecular structure. In:
Grabowski S. J. (ed) Hydrogen bonding—new insights. Springer, pp 263–292
74. Gu Y, Kar T, Scheiner S (1999) Fundamental properties of the CH... O interaction: is it a true
hydrogen bond? J Am Chem Soc 121:9411–9422
75. Alkorta I, Rozas I, Elguero J (2000) Effects of fluorine substitution on hydrogen bond
interactions. J Fluor Chem 101:233–238
76. Martins JBL, Politi JRS, Braga AD, Gargano R (2006) Complexes of water with the
fluoromethanes. Chem Phys Lett 431:51–55
77. Kryachko ES, Zeegers-Huyskens T (2001) Theoretical study of the CH... O interaction in
fluromethanes. H2 O and chloromethanes. H2 O complexes. J Phys Chem A 105:7118–7125
78. Delanoye SN, Herrebout WA, van der Veken BJ (2002) Improper or classical hydrogen bond-
ing? A comparative cryosolutions infrared study of the complexes of HCClF2 , HCCl2 F, and
HCCl3 with dimethyl ether. J Am Chem Soc 124:7490–7498
79. van der Veken BJ Delanoye SN Michielsen B Herrebout WA (2010) A cryospectroscopic
study of the blue-shifting C–H... O bonded complexes of pentafluoroethane with dimethyl
ether-d6 , acetone-d6 and oxirane-d4 . J Mol Struct 976:97–104
80. Bedell BL, Goldfarb L, Mysak ER, Samet C, Maynard A (1999) Matrix isolation infrared and
ab initio study of the 1:1 complexes of bromocyclopropane with NH3 and (CH3 )3 N: evidence
for a novel C–H... N hydrogen bond. J Phys Chem A 103:4572–4579
81. Karger N, Amorim da Costa AM, Ribeiro-Claro JA (1999) C–H... O bonded dimers in liquid 4-
methoxybenzaldehyde: a study by NMR, vibrational spectroscopy, and ab initio calculations.
J Phys Chem A 103:8672–8677
82. Hobza P, Havlas Z (2000) Blue-shifting hydrogen bonds. Chem Rev 100:4253–4264
83. Marques MPM, da Costa AMA, Ribeiro-Claro PJA (2001) Evidence of C–H... O hydrogen
bonds in liquid 4-ethoxybenzaldehyde by NMR and vibrational spectroscopies. J Phys Chem
A 105:5292–5297
84. Reimann B, Buchhold K, Vaupel S, Brutschy B, Havlas Z, Spirko V, Hobza P (2001) Improper,
blue-shifting hydrogen bond betwen fluorobenzene and fluoroform. J Phys ChemA 105:5560–
5566
85. Hobza P, Havlas Z (2002) Improper, blue-shifting hydrogen bond. Theor Chem Acc 108:325–
334
86. Gopi R, Ramanathan N, Sundararajan K (2014) Experimental evidence for blue-shifted hy-
drogen bonding in the fluoroform–hydrogen chloride complex: a matrix-isolation infrared
and ab initio study. J Phys Chem A 118:5529–5539
87. Gu Y, Kar T, Scheiner S (2000) Comparison of the CH... N and CH... O interactions involving
substituted alkanes. J Mol Struct 552:17–31
88. Gu Y, Kar T, Scheiner S (2000) Evaluation of the H-bonding properties of CH... O interactions
based upon NMR spectra. J Mol Struct (Theochem) 500:441–452
89. Masunov A, Dannenberg JJ, Contreras RH (2001) C–H bond-shortening upon hydrogen bond
formation: influence of an electric field. J Phys Chem A 105:4737–4740
90. Pejov L, Hermansson K (2003) On the nature of blueshifting hydrogen bonds: ab initio and
density functional studies of several fluoroform complexes. J Chem Phys 119:313–324
91. Qian W, Krimm S (2002) Vibrational spectroscopy of hydrogen bonding: origin of the different
behavior of the C–H... O hydrogen bond. J Phys Chem A 106:6628–6636
92. Qian W, Krimm S (2002) C–H... O and O–H... O hydrogen bonding in formic acid dimer
structures: a QM/MM study confirms the common origin of their different spectroscopic
behavior. J Phys Chem A 106:11663–11671
93. Hermansson K (2002) Blue-shifting hydrogen bonds. J Phys Chem A 106:4695–4702
94. Delanoye SN, Herrebout WA, van der Veken BJ (2002) Blue shifting hydrogen bonding in
the complexes of chlorofluoro haloforms with acetone-d6 and oxirane-d6 . J Am Chem Soc
124:11854–11855
95. Li X, Liu L, Schlegel HB (2002) On the physical origin of blue-shifted hydrogen bonds. J
Am Chem Soc 124:9639–9647
96. Alabugin IV, Manoharan M, Peabody S, Weinhold F (2003) Electronic basis of improper
hydrogen bonding: a subtle balance of hyperconjugation and rehybridization. J Am Chem
Soc 125:5973–5987
97. Alabugin IV, Manoharan M, Weinhold FA (2004) Blue-shifted and red-shifted hydrogen bonds
in hypervalent rare-gas FRg–H···Y. J Phys Chem A 108:4720–4730
98. Rutkowski KS, Rodziewicz P, Melikova SM, Herrebout WA, van der Veken BJ, Koll
A (2005) Blue shifted F3 CH... FCD3 and Cl3 CH... FCD3 weakly H-bound complexes.
Cryospectroscopic and ab initio study. Chem Phys 313:225–243
99. Yang Y, Zhang W-J, Gao X-M (2006) Theoretical study on N–H···O blue-shifted H-bond for
HNO···H2 O2 complex. Chin. J Chem 24:887–893
100. Jablonski M (2007) Blue-shifting intramolecular C–H... O(S) contacts in sterically strained
systems. J Mol Struct (Theochem) 820:118–127
101. Li AY (2007) Chemical origin of blue- and redshifted hydrogen bonds: intramolecular
hyperconjugation and its coupling with intermolecular hyperconjugation. J Chem Phys
126:154102
102. Joseph J, Jemmis ED (2007) Red-, blue-, or no-shift hydrogen bonds: a unified explanation.
J Am Chem Soc 129:4620–4632
103. Michielsen B, Herrebout WA, van der Veken BJ (2008) C–H bonds with a positive dipole gra-
dient can form blue-shifting hydrogen bonds: the complex of halothane with methyl fluoride.
ChemPhysChem 9:1693–1701
100 S. Scheiner
104. Li AY (2008) Theoretical study of linear and bifurcated H-bonds in the systems Y... H2 CZn
(n = 1,2; Z = O, S, Se, F, Cl, Br; Y = Cl− ,Br− ). J Mol Struct (Theochem) 862:21–27
105. Karpfen A, Kryachko ES (2009) On the intramolecular origin of the blue shift of A–H
stretching frequencies: triatomic hydrides HAX. J Phys Chem A 113:5217–5223
106. Grabowski SJ (2011) Red- and blue-shifted hydrogen bonds: the Bent rule from quantum
theory of atoms in molecules perspective. J Phys Chem A 115:12789–12799
107. Karpfen A, Kryachko ES (2005) Blue-shifted hydrogen-bonded complexes. II.
CH3 F... (HF)1≤n≤3 and CH2 F...2 (HF)1≤n≤3 . Chem Phys 310:77–84
108. Kryachko ES, Karpfen A (2006) Theoretical force-field model for blue-shifted hydrogen
bonds with fluoromethanes. Chem Phys 329:313–328
109. Karpfen A (2011) Blue-shifted A–H stretching frequencies in complexes with methanol: the
decisive role of intramolecular coupling. Phys Chem Chem Phys 13:14194–14201
110. Chandra AK, Zeegers-Huyskens T (2013) Theoretical study of the cooperativity in substi-
tuted dimethyl ethers complexed with two water molecules. Red or blue shifts of the ν(CH)
vibrations? Chem Phys 410:66–70
111. Donoso-Tauda O, Jaque P, Santos JC (2011) Theoretical analysis based on X–H bond-
ing strength and electronic properties in red- and blue-shifting hydrogen-bonded X–H... π
complexes. Phys Chem Chem Phys 13:1552–1559
112. Mo Y, Wang C, Guan L, Braı̈da B, Hiberty PC, Wu W (2014) On the nature of blueshifting
hydrogen bonds. Chemistry. Eur J 20:8444–8452
113. Jabłoński M (2014) Red and blue shifted hydridic bonds. J Comput Chem 35:1739–1747
114. Struble MD, Kelly C, Siegler MA, Lectka T (2014) Search for a strong, virtually “No-Shift”
hydrogen bond: a cage molecule with an exceptional OH–F interaction. Angew Chem Int Ed
53:8924–8928
115. Gou Q, Feng G, Evangelisti L, Caminati W (2014) Interaction between Freons andAmines: the
C–H···N weak hydrogen bond in quinuclidine–trifluoromethane. J Phys Chem A 118:737–740
116. Dey A, Mondal SI, Patwari GN (2013) Binary complexes of ammonia with phenylacetylenes:
a combined experimental and computational approach to explore multiple minima on
intermolecular potentials. Chem Phys Chem 14:746–753
117. Kharat B, Deshmukh V, Chaudhari A (2012) Hydrogen-bonding interactions in acetonitrile
oligomers using density functional theory method. Struct Chem 23:637–644
118. Maity S, Dey A, Patwar GN, Karthikeyan S, Kim KS (2010) A combined spectroscopic and
abinitio investigation of phenylacetylene-methylamine complex. Observation of σ and π type
hydrogen-bonded configurations and fluorescence quenching by weak C–H···N hydrogen
bonding. J Phys Chem A 114:11347–11352
119. Howard AA, Tschumper GS, Hammer NI (2010) Effects of hydrogen bonding on vibrational
normal modes of pyrimidine. J Phys Chem A 114:6803–6810
120. Jablonski M, Palusiak M (2010) Basis set and method dependence in atoms in molecules
calculations. J Phys Chem A 114:2240–2244
121. Liu G, Wang H, Li W (2006) Solvent effect on the type (red-shifted or blue-shifted) of hydrogen
bond. J Mol Struct (Theochem) 772:103–108
122. Mons M, Dimicoli I, Tardivel B, Piuzzi F, Brenner V, Millié P (2002) Energetics of a model
NH-π interaction: the gas phase benzene-NH3 complex. Phys Chem Chem Phys 4:571–576
123. Shirhatti PR, Maity DK, Wategaonkar S (2013) C–H···Y hydrogen bonds in the complexes of
p-cresol and p-cyanophenol with fluoroform and chloroform. J Phys Chem A 117:2307–2316
124. Zhang G, He W, Chen D (2014) On difference of properties between organic fluorine hydrogen
bond C–H···F–C and conventional hydrogen bond. Mol Phys 112:1736–1744
125. Christenholz CL, Obenchain DA, Peebles RA, Peebles SA (2014) Rotational spectroscopic
studies of C–H···F interactions in the vinyl fluoride···difluoromethane complex. J Phys Chem
A 118:1610–1616
126. Lu N, Ley RM, Cotton CE, Chung W-C, Francisco JS, Negishi E-I (2013) Molecular tuning
of the closed shell C–H···F–C hydrogen bond. J Phys Chem A 117:8256–8262
127. Biller MJ, Mecozzi S (2012) A high level computational study of the CH4 /CF4 dimer: how
does it compare with the CH4 /CH4 and CF4 /CF4 dimers? Mol Phys 110:377–387
128. Vincent MA, Hillier IH (2011) The structure and interaction energies of the weak complexes
of CHClF2 and CHF3 with HCCH: a test of density functional theory methods. Phys Chem
Chem Phys 13:4388–4392
129. Saar BG, O’Donoghue GP, SteevesAH, John W., Thoman J (2006) Evidence for a blue-shifting
intramolecular hydrogen bond in the vibrational overtone spectrum of 1H-nonafluorobutane.
130. Favero LB, Giuliano BM, Melandri S, Maris A, Ottaviani P, Velino B, Caminati W (2005)
CH. . . O and CHF links form the cage structure of dioxane-trifluoromethane. J Phys Chem A
109:7402–7404
131. Lu P, Liu G-Q, Li J-C (2005) Existing problems in theoretical determination of red-shifted or
blue-shifted hydrogen bond. J. Mol. Struct (Theochem) 723:95–100
132. Domagala M, Grabowski SJ (2010) Hydrocarbons as proton donors in C–H... N and C–H... S
hydrogen bonds. Chem Phys 367:1–6
133. Wolstenholme DJ, Weigand JJ, Cameron EM, Cameron TS (2008) The progression of strong
and weak hydrogen bonds in a series of ethylenediammonium dithiocyanate derivatives—A
new bonding protocol for macromolecules? Phys Chem Chem Phys 10:3569–3577
134. Domagala M, Grabowski SJ (2005) C–H... N and C–H... S hydrogen bonds—influence of
hybridization on their strength. J Phys Chem A 109:5683–5688
135. Afonin AV, Toryashinova DD, Schmidt EY (2004) Investigation of C–H... X (X = N, O, S)
intramolecular hydrogen bond in 1-vinyl-2-(2’-heteroaryl)pyrroles by ab initio calculations.
J Mol Struct (Theochem) 680:127–135
136. McDowell SAC (2006) Blue and red shifts in F3 C–H... B (B = FH, ClH, OH2 , SH2 and Cl− )
complexes predicted by a perturbative model. Chem Phys Lett 424:239–242
137. Hopkins WS, Hasan M, Burt M, Marta RA, Fillion E, McMahon TB (2014) Persistent in-
tramolecular C–H···X (X = O or S) hydrogen-bonding in benzyl meldrum’s acid derivatives.
J Phys Chem A 118:3795–3803
138. Tamasi G, Botta F, Cini R (2006) DFT-molecular modeling analysis of C–H... N and C–H... S
hydrogen bond type interactions in selected platinum–purine/pyrimidine complexes. J Mol
Struct (Theochem) 766:61–72
139. Feng G, Gou Q, Evangelisti L, Vallejo-López M, Lesarri A, Cocinero EJ, Caminati W (2014)
Competition between weak hydrogen bonds: C–HCl is preferred to C–HF in CH2 ClF–H2 CO,
as revealed by rotational spectroscopy. Phys Chem Chem Phys 16:12261–12265
140. Ito F (2012) Observation of CHF2 Cl trimer in Xe matrix and calculations of stable isomers
by density functional theory. J Mol Struct 1012:43–49
141. Castro M, Nicolas-Vazquez I, Zavala JI, Sanchez-Viesca F, Berros M (2007) Theoretical
study of intramolecular CH–X (X = N, O, Cl) hydrogen bonds in thiazole derivatives. J Chem
Theory Comput 3:681–688
142. Balamurugan V, Mukherjee J, Hundal MS, Mukherjee R (2007) Supramolecular architectures
with ladder and lamellar topologies using metal-ligand coordination units via C–H···Cl and
O–H···Cl hydrogen bonding. Struct Chem 18:133–144
143. Pazout R, Houskova J, Dusek M (2011) Platinum precursor of anticancer drug: a structure fixed
by long intermolecular N–H... I and C–H... I hydrogen bonds. Struct Chem 22:1325–1330
144. Michielsen B, Verlackt C, van der Veken BJ, Herrebout WA (2012) C–H... X (X = S, P)
hydrogen bonding: the complexes of halothane with dimethyl sulfide and trimethylphosphine.
J Mol Struct 1023:90–95
145. Ramasami P, Ford TA (2012) Ab initio studies of some hydrogen-bonded complexes of
fluoroform—evidence for blue-shifted behaviour. J Mol Struct 1023:163–169
146. Howard ST, Abernethy CD (2004) Intramolecular CH. . . Ccarbene hydrogen bonds and
competing interactions in monoprotonated tripodal carbenes. J Comput Chem 25:649–659
147. Yu F (2013) Intermolecular interactions of formic acid with benzene: energy decomposition
analyses with ab initio MP2 and double-hybrid density functional computations. Int J Quantum
Chem 113:2355–2360
148. Karthikeyan S, Ramanathan V, Mishra BK (2013) Influence of the substituents on the CH... π
interaction: benzene–methane complex. J Phys Chem A 117:6687–6694
102 S. Scheiner
149. Majumder M, Mishra BK, Sathyamurthy N (2013) CH... π and π... π interaction in benzene-
acetylene clusters. Chem Phys 557:59–65
150. Michielsen B, Dom JJJ, van der Veken BJ, Hesse S, Xue Z, Suhm MA, Herrebout WA
(2010) The complexes of halothane with benzene: the temperature dependent direction of the
complexation shift of the aliphatic C–H stretching. Phys Chem Chem Phys 12:14034–14044
151. Shirhatti PR, Wategaonkar S (2010) Blue shifted hydrogen bond in 3-methylindole·CHX3
complexes (X = Cl, F). Phys Chem Chem Phys 12:6650–6659
152. Takahashi O, Kohno Y, Gondoh Y, Saito K, Nishio M (2003) General preference for
alkyl/phenyl folded conformations. Relevance of the CH/π and CH/O interactions to
stereochemistry as evidenced by ab initio MO calculations. Bull Chem Soc Jpn 76:369–374
153. Xu X, Pooi B, Hirao H, Hong SH (2014) CH–π and CF–π interactions lead to structural
changes of N-heterocyclic carbene palladium complexes. Angew Chem Int Ed 53:1283–1287
154. Wang Y, Balbuena PB (2001) Associations of alkyl carbonates: intermolecular C–H... O
interactions. J Phys Chem A 105:9972–9982
155. Wetmore SD, Schofield R, Smith DM, Radom L (2001) A theoretical investigation of the
effects of electronegative substitution on the strength of C–H... N hydrogen bonds. J Phys
Chem A 105:8718–8726
156. Kar T, Scheiner S (2004) Comparison of cooperativity in CH··O and OH··O hydrogen bonds.
J Phys Chem A 108:9161–9168
157. Malenov DP, Janjic GV, Veljkovic DŽ, Zaric SD (2013) Mutual influence of parallel, CH/O,
OH/π and lone pair/π interactions in water/benzene/water system. Comput. Theor. Chem
1018:59–65
158. Gao X, Liu Y, Li H, Bian J, Zhao Y, Cao Y, Mao Y, Li X, Xu Y, Ozaki Y, Wu J (2013) A
cooperative hydrogen bonding system with a C–H... O hydrogen bond in ofloxacin. J Mol
Struct 1040:122–128
159. Lee K-M, Chen JCC, Chen H-Y, Lin IJB (2012) A triple helical structure supported solely by
C–H... O hydrogen bonding. Chem Commun 48:1242–1244
160. Solimannejad M, Malekani M, Alkorta I (2011) Cooperativity between the hydrogen bonding
and halogen bonding in F3 CX ··· NCH(CNH) ··· NCH(CNH) complexes (X = Cl, Br). Mol
Phys 109:1641–1648
161. Samanta AK, Pandey P, Bandyopadhyay B, Chakraborty T (2010) Cooperative strengthening
of an intramolecular O–H···O hydrogen bond by a weak C–H···O counterpart: matrix-isolation
infrared spectroscopy and quantum chemical studies on 3-methyl-1,2-cyclohexanedione. J
Phys Chem A 114:1650–1656
162. Li Q, An X, Gong B, Cheng J (2007) Cooperativity between OH... O and CH... O hydrogen
bonds involving dimethyl sulfoxide-H2 O–H2 O complex. J Phys Chem A 111:10166–10169
163. Scheiner S, Kar T (2005) Effect of solvent upon CH··O hydrogen bonds with implications for
protein folding. J Phys Chem B 109:3681–3689
164. Shi Z, Olson CA, Bell AJ, Kallenbach NR (2002) Non-classical helix-stabilizing interactions:
C–H... O H-bonding between Phe and Glu side chains in α-helical peptides. Biophys Chem
101–102:267–279
165. Scheiner S, Kar T, GuY (2001) Strength of the Cα H.. O hydrogen bond of amino acid residues.
J Biol Chem 276:9832–9837
166. Scheiner S (2005) Relative strengths of NH··O and CH··O hydrogen bonds between
polypeptide chain segments. J Phys Chem B 109:16132–16141
167. Scheiner S (2007) The strength with which a peptide group can form a hydrogen bond varies
with the internal conformation of the polypeptide chain. J Phys Chem B 111:11312–11317
168. Adhikari U, Scheiner S (2013) Preferred configurations of peptide-peptide interactions. J Phys
Chem A 117:489–496
169. Scheiner S (2006) Contributions of NH··O and CH··O H-bonds to the stability of β-sheets in
proteins. J Phys Chem B 110:18670–18679
170. Vener MV, Egorova AN, Fomin DP, Tsirelson VG (2007) QTAIM study of the closed-shell
interactions in peptide secondary structures: a cluster treatment of oligo- and polyalanines.
171. Parthasarathi R, Raman SS, Subramanian V, Ramasami T (2007) Bader’s electron density
analysis of hydrogen bonding in secondary structural elements of proteins. J Phys Chem A
111:7141–7148
172. Guo H, Gorin A, Guo H (2009) A peptide-linkage deletion procedure for estimate of energetic
contributions of individual peptide groups in a complex environment: application to parallel
β-sheets. Interdiscip Sci Comput Life Sci 1:12–20
173. Vener MV, Egorova AN, Fomin DP, Tsirel’son VG (2009) A quantum-topological analysis of
noncovalent interactions in secondary polyalanine structures. Russ. J Phys Chem B 3:541–547
174. LaPointe SM, Farrag S, Bohórquez HJ, Boyd RJ (2009) QTAIM study of an α-helix hydrogen
bond network. J Phys Chem B 113:10957–10964
175. Park H, Yoon J, Seok C (2008) Strength of Cα –H... O = C hydrogen bonds in transmembrane
protein. J Phys Chem B 112:1041–1048
176. Pohl G, Plumley JA, Dannenberg JJ (2013) The interactions of phenylalanines in β-sheet-like
structures from molecular orbital calculations using density functional theory (DFT), MP2,
and CCSD(T) methods. J Chem Phys 138:245102
177. Hussain S, Das G, Chaudhuri MK (2007) Intermolecular hydrogen bonded and self-assembled
β-pleated sheet structures of β-sulfidocarbonyls. J Mol Struct 837:190–196
178. Adhikari U, Scheiner S (2013) First Steps in growth of a polypeptide toward β-sheet structure.
J Phys Chem B 117:11575–11583
179. Scheiner S, Kar T, Pattanayak J (2002) Comparison of various types of hydrogen bonds
involving aromatic amino acids. J Am Chem Soc 124:13257–13264
180. Chang H-C, Lee KM, Jiang J-C, Lin M-S, Chen J-S, Lin IJB, Lin SH (2002) Charge-enhanced
C–H... O interactions of a self-assembled triple helical spine probed by high-pressure. J Chem
Phys 117:1723–1728
181. Raymo FM, Bartberger MD, Houk KN, Stoddart JF (2001) The magnitude of [C–H... O]
hydrogen bonding in molecular and supramolecular assemblies. J Am Chem Soc 123:9264–
9267
182. Guo H, Beahm RF, Guo H (2004) Stabilization and destabilization of the Cδ –H... O = C
hydrogen bonds involving proline residues in helices. J Phys Chem B 108:18065–18072
183. Deakyne CA, Meot-Ner M (1999) Ionic hydrogen bonds in bioenergetics. 4. Interaction
energies of acetylcholine with aromatic and polar molecules. J Am Chem Soc 121:1546–1557
184. Cannizzaro CE, Houk KN (2002) Magnitude and chemical consequences of R3 N+ –C–H... O
= C hydrogen bonding. J Am Chem Soc 124:7163–7169
185. Liu T, Gu J, Tan X-J, Zhu W-L, Luo X-M, Jiang H-L, Ji R-Y, Chen K-X, Silman I, Sussman
JL (2002) The relationship between binding models of TMA with furan and imidazole and
the molecular electrostatic potentials: DFT and MP2 computational studies. J Phys Chem A
106:157–164
186. Shishkin OV, Palamarchuk GV, Gorb L, Leszczynski J (2008) Opposite charges assisted extra
strong C–H... O hydrogen bond in protonated 2’-deoxyadenosine monophosphate. Chem Phys
Lett 452:198–205
187. Adhikari U, Scheiner S (2013) The magnitude and mechanism of charge enhancement of
CH··O H-bonds. J Phys Chem A 117:10551–10562
188. Horowitz S, Dirk LMA, Yesselman JD, Nimtz JS, Adhikari U, Mehl RA, Scheiner S,
Houtz RL, Al-Hashimi HM, Trievel RC (2013) Conservation and functional importance of
carbon-oxygen hydrogen bonding in AdoMet-dependent methyltransferases. J Am Chem Soc
135:15536–15548
189. Kraut J (1977) Serine proteases: structure and mechanism of catalysis. Ann Rev Biochem
46:331–358
190. Wang JH (1970) Directional character of proton transfer in enzyme catalysis. Proc Natl Acad
Sci U S A 66:874–881
191. Topf M, Várnai P, Richards WG (2002) Ab initio QM/MM dynamics simulation of the tetrahe-
dral intermediate of serine proteases: insights into the active site hydrogen-bonding network.
J Am Chem Soc 124:14780–14788
104 S. Scheiner
192. Derewenda ZS, Derewenda U, Kobos PM (1994) (His)Cε –H... O = C < hydrogen bond in the
active sites of serine hydrolases. J Mol Biol 241:83–93
193. Ishida T, Kato S (2003) Theoretical perspectives on the reaction mechanism of serine
proteases: the reaction free energy profiles of the acylation process. J Am Chem Soc
125:12035–12048
194. Ash EL, Sudmeier JL, Day RM, Vincent M, Torchilin EV, Haddad KC, Bradshaw EM, Sanford
DG, Bachovchin WW (2000) Unusual 1 H NMR chemical shifts support (His) Cε1 –H... O = C
H-bond: proposal for reaction-driven ring flip mechanism in serine protease catalysis. Proc
Natl Acad Sci U S A 97:10371–10376
195. Bachovchin WW (2001) Contributions of NMR spectroscopy to the study of hydrogen bonds
in serine protease active sites. Magn Reson Chem 39:39S199–S213
196. Scheiner S (2008) Analysis of catalytic mechanism of serine proteases. Viability of ring-flip
hypothesis. J Phys Chem B 112:6837–6846
197. Chen F, Selvam L, Wang F (2010) Blue shifted intramolecular C–H···O improper hydrogen
bonds in conformers of zidovudine. Chem Phys Lett 493:358–363
198. Zierke M, Smieško M, Rabbani S, Aeschbacher T, Cutting B, Allain FH-T, Schubert M, Ernst
B (2013) Stabilization of branched oligosaccharides: Lewis benefits from a nonconventional
C–H···O hydrogen bond. J Am Chem Soc 135:13464–13472
199. Zgarbová M, Jurecka P, Banás P, Otyepka M, Sponer JE, Leontis NB, Zirbel CL, Sponer J
(2011) Noncanonical hydrogen bonding in nucleic acids. benchmark evaluation of key base-
phosphate interactions in folded RNA molecules using quantum-chemical calculations and
molecular dynamics simulations. J Phys Chem A 115:11277–11292
200. Shishkin OV, Palamarchuk GV, Gorb L, Leszczynski J (2006) Intramolecular hydrogen bonds
in canonical 2’-deoxyribonucleotides: an atoms in molecules study. J Phys Chem B 110:4413–
4422
201. Shishkin OV, Gorb L, Zhikol OA, Leszczynski J (2004) Conformational analysis of canonical
2-deoxyribonucleotides. 1. Pyrimidine nucleotides. J Biomol Struct Dyn 21:537–554
202. Alkorta I, Elguero J (2003) Interaction of protein backbone with nucleic acid bases. J Phys
Chem B 107:5306–5310
203. Liu Z, Wang G, Li Z, Wang R (2008) Geometrical preferences of the hydrogen bonds on
protein-ligand binding interface derived from statistical surveys and quantum mechanics. J
Chem Theory Comput 4:1959–1973
204. Hildebrand PW, Günther S, Goede A, Forrest L, Frömmel C, Preissner R (2008) Hydrogen-
bonding and packing features of membrane proteins: functional implications. Biophys J
94:1945–1953
205. Mueller BK, Subramanian S, Senes A (2014) A frequent, GxxxG-mediated, transmembrane
association motif is optimized for the formation of interhelical Cα–H hydrogen bonds. Proc
Natl Acad Sci U S A 111:E888–E895
206. Fulara A, Dzwolak W (2010) Bifurcated hydrogen bonds stabilize fibrils of poyl(L-glutamic)
acid. J Phys Chem B 114:8278–8283
207. Panigrahi SK, Desiraju GR (2007) Strong and weak hydrogen bonds in the protein-ligand
interface. Proteins 67:128–141
208. Sheppard D, Li D-W, Godoy-Ruiz R, Brschweiler R, Tugarinov V (2010) Variation in
quadrupole couplings of a deuterons in ubiquitin suggests the presence of Cα –Hα ···O = C
hydrogen bonds. J Am Chem Soc 132:7709–7719
209. Venugopalan P, Kishore R (2013) Unusual folding propensity of an unsubstituted β, γ-hybrid
model peptide: importance of the C–H. . . O intramolecular hydrogen bond. Chem Eur J
19:9908–9915
210. Sukumar N, Mathews FS, Langan P, Davidson VL (2010) A joint x-ray and neutron study on
amicyanin reveals the role of protein dynamics in electron transfer. Proc Natl Acad Sci U S
A 107:6817–6822
211. Chen JC-H, Hanson BL, Fisher SZ, Langan P, Kovalevsky AY (2012) Direct observa-
tion of hydrogen atom dynamics and interactions by ultrahigh resolution neutron protein
crystallography. Proc Natl Acad Sci U S A 109:15301–15306
212. Yoneda Y, Mereiter K, Jaeger C, Brecker L, Kosma P, Rosenau T, French A (2008) van der
Waals versus hydrogen-bonding forces in a crystalline analog of cellotetraose: cyclohexyl
4’-O-cyclohexyl β-D-cellobioside cylohexane solvate. J Am Chem Soc 130:16678–16690
213. Banks JW, Batsanov AS, Howard JAK, O’Hagan D, Rzepa HS, Martin-Santamaria S (1999)
The preferred conformation of α-fluoroamides. J Chem Soc Perkin Trans II 2409–2411
214. Tormena CF, Amadeu NS, Rittner R, Abraham RJ (2002) Conformational analysis in N-
methylfluoroamides. A theoretical, NMR and IR investigation. J Chem Soc Perkin Trans II
773–778
215. Jones CR, Baruah PK, ThompsonAL, Scheiner S, Smith MD (2012) Can a C–H···O interaction
be a determinant of conformation. J Am Chem Soc 134:12064–12071
216. Écija P, Vallejo-López M, Evangelisti L, Fernández JA, Lesarri A, Caminati W, Cocinero EJ
(2014) O–H... N and C–H... O hydrogen bonds control hydration of pivotal tropane alkaloids:
Tropinone... H2 O complex. Chem Phys Chem 15:918–923
217. Huang Z, Chen Z, Lim LH, Quang GCP, Hirao H, Zhou JS (2013) Weak arene C–H... O
hydrogen bonding in palladium-catalyzed arylation and vinylation of lactones. Angew Chem
Int Ed 52:5807–5812
218. Huang Z, Lim LH, Chen Z, Li Y, Zhou F, Su H, Zhou JS (2013) Arene CH–O hydrogen
bonding: a stereocontrolling tool in palladium-catalyzed arylation and vinylation of ketones.
Angew Chem Int Ed 52:4906–4911
219. Juribasic M, Bellotto L, Tusek-Bozic L (2012) N–H···O=P hydrogen-bonded dimers as the
main structural motif of aminophosphonate diesters. Struct Chem 23:257–266
220. You L-Y, Chen S-G, Zhao X, Liu Y, Lan W-X, Zhang Y, Lu H-J, Cao C-Y, Li Z-T (2012)
C–H... O hydrogen bonding induced triazole foldamers: efficient halogen bonding receptors
for organohalogens. Angew Chem Int Ed 51:1657–1661
221. Lippert KM, Hof K, Gerbig D, Ley D, Hausmann H, Guenther S, Schreiner PR (2012)
Hydrogen-bonding thiourea organocatalysts: The privileged 3,5-bis(trifluoromethyl)phenyl
group. Eur J Org Chem 2012:5919–5927
222. Bandyopadhyay B, Pandey P, Banerjee P, Samanta AK, Chakraborty T (2012) CH···O inter-
action lowers hydrogen transfer barrier to keto-enol tautomerization of β-cyclohexanedione:
combined infrared spectroscopic and electronic structure calculation study. J Phys Chem A
116:3836–3845
223. Vibhute AM, Gonnade RG, Swathi RS, Sureshan KM (2012) Strength from weakness: op-
portunistic CH.. O hydrogen bonds differentially dictate the conformational fate in solid and
solution states. Chem Commun 48:717–719
224. Solà J, Riera A, Verdaguer X, Maestro MA (2005) Phosphine-substrate recognition through
the C–H... O hydrogen bond: application to the asymmetric Pauson-Khand reaction. J Am
Chem Soc 127:13629–13633
Chapter 5
Hydrogen Bonds Involving Radical Species
Qing-Zhong Li and Hai-Bei Li
Abstract In this chapter, we focused on the structures, patterns, energies, and nature
of hydrogen bonds involving radicals, such as H3 C, OH, BH2 , and BeH, based on
the fact that hydrogen-bonded complexes involving radicals may be formed in the
related reactions and processes, and are useful for understanding their mechanisms.
Theses radicals as the proton donor and acceptor may participate in the formation of
different types of hydrogen bonds, including single-electron hydrogen bonds with
the single electron of radicals as the proton acceptor, dihydrogen bonds with the
hydridic hydrogen of radicals as the proton acceptor, conventional hydrogen bonds
with the lone-pair electron of radicals as the proton acceptor or with the proton
of radicals as the proton donor. In addition, a covalent interaction is also formed
between radicals and the other molecule. The formation of these interactions was
understood from the view of HOMO and LUMO of radicals, and their nature was
analyzed by the energy decomposition scheme, showing similar nature in most cases
with conventional hydrogen bonds. We paid a particular attention to the cooperative
effect of single-electron hydrogen bond with other types of interactions as well as
the competition among different types of interactions involving radical species.
5.1 Introduction
The radicals, such as hydroxyl radical (HO), hydroperoxyl radical (HOO), organic
peroxy radical (ROO), alkyl radical (R), particularly methyl radical (H3 C), have been
considered to be extremely important in many chemical reactions and the related pro-
cesses, including combustion chemistry and celestial chemistry. For instance, ROO
radicals, usually from the atmospheric oxidation of hydrocarbons, are responsible for
Q.-Z. Li ()
The Laboratory of Theoretical and Computational Chemistry,
School of Chemistry and Chemical Engineering, Yantai University,
264005 Yantai, People’s Republic of China
e-mail: liqingzhong1990@sina.com
H.-B. Li ()
School of Ocean, Shandong University, 264209 Weihai, People’s Republic of China
e-mail: lihaibei@sdu.edu.cn
108 Q.-Z. Li and H.-B. Li
the formation of tropospheric ozone, and also the generation of HOx (HO and HOO)
radicals [1]. Reactive free radical, HOO, plays significant roles in the stratospheric
chemistry and oxidation processes in the troposphere, such as the generation of sul-
phuric acid through the oxidation of SO2 [2]. In these reactions or processes, radicals
probably form the complexes through hydrogen bonds or other types of interactions.
Hydrogen bond is one of the most interesting topics of intermolecular interactions
because it plays an important role in molecular recognition, crystal engineering, and
chemical reactions [3]. Based on the electronic character of the monomers, two
types of hydrogen bond interactions are defined, that is, closed-shelled and open-
shelled ones, in which the latter has attracted a growing attention in recent years.
The open-shelled hydrogen bonds are able to regulate the electron transfer processes
in many enzymatic systems [4] and affect the chemical properties of many radical
species, such as bond dissociation energies and reduction/oxidation potentials [5]. It
has been demonstrated that radical species plays the roles of proton acceptor/donor
in hydrogen-bonded complexes. A single-electron hydrogen bond could form if the
single electron of a radical acts as the proton acceptor [6].
A great number of theoretical and experimental investigations of the complexes
involving radicals [7] have been performed to understand the mechanisms of the
related reactions and processes, such as the proton transfer between neutral molecules
and radical species in organic reactions. In the following sections, we introduced
different types of hydrogen bonds involving radicals, and paid more attention to the
effects of competition and cooperativity between them and other types of interactions.
In order to compare the complexes at the same level, we recalculated all the complexes
using Gaussian 09 program [8] at the UMP2/aug-cc-pVTZ(PP) level, even some of
them have been studied before. The binding energy of these complexes is obtained
using supermolecular method, that is, the difference between the sum of the energy
of the monomers and the total energy of the complex. In general, the energy of
the monomer is obtained from the optimized geometries of the isolated molecules
with the exception if the significant variations in geometry occur between the isolated
monomers and the ones in the complex, the geometry of the monomer in the complex
will be used.
5.2 Single-Electron Hydrogen Bonds
5.2.1 Alkyl Radicals as Proton Acceptors
Methyl radical (H3 C) is a simple prototype for a wide class of organic radicals. It usu-
ally acts as the proton acceptor in single-electron hydrogen bonds and plays important
roles as an intermediate in the field of chemistry and biochemistry involving methyl
radical. There are two types of the highest occupied molecular orbital (HOMO) and
the lowest unoccupied molecular orbital (LUMO) for H3 C radical (Fig. 5.1): alpha
and beta. The shape of alpha-LUMO is similar to that of beta-LUMO, and both or-
bitals are near degenerate due to their slight difference in energy. However, alpha-
and beta-HOMOs of H3 C radical are different. The alpha-HOMO is perpendicular to
5 Hydrogen Bonds Involving Radical Species 109
Fig. 5.1 HOMOs and LUMOs of CH3 (a), H2 B (b), and HBe (c) radicals with the orbital energy (E)
the plane of methyl radical, describing the distribution of the single free electron on
the C atom, while the beta-HOMO is in the plane of methyl radical, which describes
the C–H σ-bonds. In combination with the higher energy of the former than that of
the latter, it is predictable that the alpha-HOMO will provide the unpaired electron
when H3 C radical forms a single-electron hydrogen bond with the proton donors.
H3 C· · · HF complex was detected first in the reaction of F atoms with methane by
matrix isolation infrared spectroscopy [9], characterized by a red shift of –198 cm−1
for the H–F stretching vibration [10], and later was further confirmed by the hy-
perfine coupling constant using electron paramagnetic resonance [11]. Theoretical
calculations showed that this complex forms a single-electron hydrogen bond with
C3v symmetry (Fig. 5.2a). The bind energy of the complex is around 8 ∼ 13 kJ/mol,
depending on the computational levels. [6, 11, 12]. The complexes of H3 C radical
with other hydrogen halides were also studied, and different types of interactions
Fig. 5.2 Geometrical structures of a single-electron hydrogen-bonded complexes, b single-electron

halogen-bonded complexes, reaction products of H3 C radical and hydrogen halide, c methane and
halogen radical, and d halomethane and H radical with the binding energy (BE)
were obtained (Fig. 5.2a, 5.2b). For instance, H3 C radical and HBr molecule could
form single electron hydrogen-bonded (Fig. 5.2a) and also halogen-bonded com-
plexes (Fig. 5.2b) between the single electron of the former monomer and the proton
and halogen of the latter, respectively. At the MP2/aug-cc-pVTZ(PP) level, the
binding energies corrected for basis set superposition error (BSSE) were calcu-
lated to be 12.00, 9.24, 8.65, 3.26, and 4.85 kJ/mol for the complexes H3 C· · · HF,
H3 C· · · HCl, H3 C· · · HBr, H3 C· · · BrH, and H3 C· · · IH, respectively. Thus, the
strength of the single electron hydrogen bond is in order of HF > HCl > HBr, and
it is in order of HBr < HI for the single electron halogen bond, where the former is
related with the electronegativity of the halogen and the latter is with the magnitude
of σ-hole, a region of positive electrostatic potential on the outer of the covalently
bonded halogen’s surface [13]. In nature, this character is consistent with that of
the conventional hydrogen bonds and halogen bonds. It has been suggested that the
complexes H3 C· · · HX and H3 C· · · XH (X = F, Cl, Br, and I) are the intermediates
of the reactions: CH3 + HX → CH4 + X (Fig. 5.2c) and CH3 + HX → CH3 X + H
(Fig. 5.2d) [14]. With the development of the high-resolution detective technique,
these preliminary precursors of the reactions are expected to be observed in the near
future. Wang et al. performed a detailed theoretical study on the single electron
hydrogen-bonded complexes of H3 C· · · HF and H3 C· · · HCCH, and suggested that
the single-electron hydrogen bond has many similar features with the conventional
one, such as the lengthening of the X–H (X = F and C) bond of the proton donor
and the red shift of the stretching vibrational mode [6]. Raghavendra and Arunan
[15] compared the single-electron hydrogen, lithium, and chlorine bonds with H3 C
as the electron donor and HF, LiF, and ClF as the electron acceptor, respectively,
that is, H3 C· · · HF, H3 C· · · LiF, and H3 C· · · ClF. Interestingly, they found the similar
features of the interactions in all these complexes using Atoms in Molecules (AIM)
theoretical analysis.
Besides the complexes H3 C· · · HX (X = F, Cl, Br, and I), many other single-
electron hydrogen-bonded complexes involving H3 C radical have been studied, such
as H3 C· · · HCN, H3 C· · · HNC, H3 C· · · HCCH, and H3 C· · · H2 O. Figures 5.3, and 5.4
depict the optimized structures of complexes H3 C· · ·Y (Y = HCN, HNC, H2 O, H2 S,
and NH3 ), where H3 C radical also plays the role of the proton acceptor in the com-
plexes. In most cases, H3 C radical forms the weak single-electron hydrogen bonds
with the other neutral proton donors with the binding energies less than 12 kJ/mol. On
the other hand, the strength of this weak hydrogen bonding interaction is able to be
regulated by the substitution effects. Generally, both electron-donating substituents in
H3 C radical and electron-withdrawing ones in the proton donor could strengthen the
single-electron hydrogen bonds. Furthermore, the single-electron hydrogen bonds
become much strong if the proton donor is protonated. For example, the binding
energy is about 62 kJ/mol in the complex H3 C· · · H3 O+ at the MP2/aug-cc-pVTZ
level.
Complex H3 C· · · HCN (Fig. 5.3a) was suggested to be a post-reactive interme-
diate of the reaction CN + CH4 → HCN + CH3 [16], and it has been studied using
many experimental and theoretical methods. This complex was characterized by a
red shift of − 45.51 cm−1 for the C–H stretching vibration in helium nanodroplets
using infrared laser spectroscopy [16]. Theoretically, the red shift of C–H bond was
estimated to be − 48, − 38, and − 23 cm−1 at the UCCSD(T)/6-311++G(2d,2p)
[17], UMP2/6-311++G(2d,2p) [17], and UMP2/6-311++G(d, p) levels [16], re-
spectively. It is obvious that this red shift is closely correlated with the theoretical
methods applied and is well reproduced at the UCCSD(T)/6-311++G(2d,2p) level,
but underestimated at the UMP2/6-311++G(d, p) level. This indicates that it is
necessary to apply the high level methods considering the electron correlation in
combination with the large basis sets to quantitatively reproduce the properties of
open-shelled systems [18]. On the other hand, complex H3 C· · · HNC (Fig. 5.3b)
forms a stronger single-electron hydrogen bond than that of H3 C· · · HCN complex,
due to the higher acidity of HNC monomer where the proton connects to atom N
with the higher electronegativity than that of C in HCN. This is consistent with the
variation trend of the strength of hydrogen bond in the complexes H3 C· · · HX (X = F,
Cl, Br, and I). The binding energies were estimated to be 6.40 and 11.41 kJ/mol for
H3 C· · · HCN and H3 C· · · HNC complexes, respectively.
The reaction of OH radical with methane is important in combustion and the
atmosphere, which produces one H3 C radical and one water molecule via direct
abstraction of one H atom from methane by OH [19], and its channel complex
H3 C· · · H2 O (Fig. 5.4a) has been evidenced in helium nanodroplets using infrared
laser spectroscopy [20]. The H3 C· · · H2 O complex was characterized with a red shift
of the H–O–H symmetric stretching vibration compared with that of H2 O monomer.
Further ab initio calculations [21] predicted that it is a weak single electron hydrogen-
bonded complex with Cs symmetry, which is in good agreement with the experiment
results. Complexes H3 C· · · H2 S (Fig. 5.4d) and H3 C· · · H3 N (Fig. 5.4g) were also
studied in order to compare with H3 C· · · H2 O, and the results show that the stability
Fig. 5.3 Complexes with a single-electron hydrogen bond (a and b), a hydrogen bond with H3 C
as a proton donor (c and d), and a covalent interaction (e–h) formed between the H3 C radical and
HCN or HNC with binding energy (BE)
of these complexes is in order of H3 C· · · H3 N < H3 C· · · H2 S < H3 C· · · H2 O, which

is consistent with the acidity of molecules NH3 , H2 S and H2 O.
The effects of substitution, hybridization, and solvent on the properties of
C· · · H–O single-electron hydrogen bond in the H3 C· · · H2 O complex have been in-
vestigated by means of quantum chemical calculations [22]. This hydrogen bond
is able to be strengthened by the halogenations of the proton donor, such as
H3 C· · · HOCl. The methyl group in the proton donor and acceptor plays different
roles in the formation of the C· · · H–O hydrogen bond, where the former plays the
role of electron-withdrawing and the latter is electron-donating. Both cases make
positive contributions to the enhancement of the hydrogen bonding interaction, but
the latter has larger one than the former. When the proton acceptor varies from methyl
radical to vinyl one, the binding energy increases by 1 kJ/mol. On the other hand,
the binding energy also increases due to the enhancement of the acidity of the proton
Fig. 5.4 Complexes of a single-electron hydrogen bond (a, d, g), a hydrogen bond with H3 C as a
proton donor (b, e, h), and a covalent interaction (c, f, i) formed between the H3 C radical and H2 O,
H2 S, or NH3 with binding energy (BE)
donor, such as the proton donor changing from water to vinyl alcohol [22]. Further-
more, it has been confirmed that the solvents also have an enhancing effect on the
strength of the single-electron hydrogen bond, as evidenced by the increase of the
binding energy and the shortening of the binding distance [22].
On the basis of the measurement of the binding distance and binding energy,
the nonadditivity effect of methyl group has been studied in the single electron
hydrogen-bonded complexes with H3 C, (H3 C)H2 C, (CH3 )2 HC, and (CH3 )3 C as the
proton acceptors and H2 O as the proton donor at the UMP2/6-311++G(2df,2p)
level [23]. It is found that the increase of the number of the methyl substituents in
the proton acceptor results in a shorter binding distance and a larger binding energy,
indicating that there is nonadditivity effect of the methyl groups in the single electron
hydrogen-bonded complexes. Furthermore, this nonadditivity is negative, that is, the
contribution of each methyl group decreases with the increase of the number of the
methyl group in the proton donor. This is different from the positive nonadditivity
of methyl groups in H3 C· · · BrH complex [24].
As mentioned above, alkyl radicals form the weak single-electron hydrogen bonds
with the neutral proton donors, such as HF, HCl, HBr, H2 O, NH3 , H2 S, HCN,
HNC, and HCCH. When ionic proton donors are present, however, alkyl radicals are
able to form moderate even strong hydrogen bonds with the corresponding donors
[25], which have been evidenced by the proton transfer from water and alcohols
to alkyl radicals in the presence of Lewis acids [25, 26]. For instance, the methyl
radical forms a single-electron hydrogen bond with H3 O+ , the binding energy is
51.41 kJ/mol at the CCSD(T)/6-311G++(3df,2pd)//MP2/6-31G++(d, p) level [28],
which is much larger than that in H3 C· · · H2 O complex. These interactions present
the same characteristics with those from closed-shell species; however, the charge
transfer contributions stand out more clearly in these interactions since the electro-
static contribution diminishes when the acceptor molecule is nonpolar and nonbasic
[25]. For the complexes with alkyl radicals as proton acceptor and methanol as pro-
ton donor, the strengths of single-electron hydrogen bonds are in order of methyl
radical < ethyl radical < n-propyl radical < iso-propyl radical < sec-butyl radical
< tert-butyl radical [25], and the tert-butyl radical is a much better proton acceptor
than formaldehyde, which is characterized by a larger elongation of O–H bond and
a bigger red shift of O–H stretch vibration of the proton donor methanol [25].
5.2.2 Other Types of Radicals as Proton Acceptors
Besides alkyl radicals, the proton acceptors in single-electron hydrogen bonds could
also be other types of radical species, such as H, Li, HBe, HMg, H2 B, H2Al, and
H2 Ga. Similar with those of H3 C radical, the alpha- and beta-HOMOs of HBe and
H2 B are different in energy and shape for both radicals (Fig. 5.1). The alpha-HOMOs
of H2 B and HBe also direct to another hybridization orbital occupied by the single
free electron, while the beta-HOMOs describe theY–H σ-bonds (Y = B and Be). Both
the higher energy and the shape of alpha-HOMO indicate that H2 B and HBe radicals
are favorable to act as the proton acceptor with the single electron participating
in the formation of single-electron hydrogen bonds. On the basis of the similarity
between H2 B (HBe) and H3 C radicals, it is predictable that the complexes with H2 B
(HBe) and H3 C as the proton acceptor might have similar features. Figures 5.5a–5.5c
show the optimized structures of the single electron hydrogen-bonded complexes of
H2 B· · · HCN, H2 B· · · HNC, and H2 B· · · HF, where the proton is introduced to the
single electron along the direction of the alpha-HOMO of H2 B radical. Solimannejad
andAlkorta [29] performed ab initio study of the single-electron hydrogen bonds with
AH2 radicals (A = B, Al, and Ga) as the proton acceptors and HX (X = F, Cl, Br, CN,
and CCH) as the proton donors to investigate their binding energies, frequency shifts,
and geometrical properties. The stability of complexes is in order of H2 B· · · HX >
H2Al· · · HX > H2 Ga· · · HX, and with the same proton acceptor, the binding energy is
H2A· · · HF > H2A· · · HCl > H2A· · · HBr > H2A· · · HCN > H2A· · · HCCH [29]. For
instance, the complex H2 B· · · HNC, with the binding energy of 13.71 kJ/mol and the
binding distance of 2.390 Å, is more stable than H2 B· · · HCN, which has the binding
energy and distance of 7.19 kJ/mol and 2.684 Å, respectively [30]. This is similar to
the complexes of H3 C radical with HCN and HNC.
Fig. 5.5 Complexes of a single-electron hydrogen bond (a–c), a dihydrogen bond (d–f), and a
covalent interaction (g–k) formed between BH2 and HCN, HNC, and HF with binding energy (BE)
Figure 5.6 presents the optimized structures of the single electron hydrogen-
bonded complexes of HBe· · · HCN, HBe· · · HNC, and HBe· · · HF with the binding
energies of 5.81, 11.12, and 13.46 kJ/mol, respectively. Thus, the stability of the com-
plexes is in the same order as that of H2 B as the proton acceptor. In order to compare
the capability of accepting protons of HBe, H2 B, and H3 C in the single electron
hydrogen-bonded complexes with HF as the proton donor, we found that the binding
energy increases in order of H3 C < HBe < H2 B [31]. Additionally, theses complexes
were studied at the levels of QCISD/aug-cc-pVTZ and CCSD(T)/aug-cc-pVTZ, and
the results showed that the properties of open-shelled systems are significantly af-
fected by the computational levels, which is consistent with the conclusion proposed
by Qi et al. [18].
Fig. 5.6 Complexes with a single-electron hydrogen bond (a–c). a dihydrogen bond (d–f), and a
covalent interaction (g–k) formed of BeH with HCN, HNC, and HF with binding energy (BE)
5.3 Other Types of Hydrogen Bonds with Radicals as Proton

Acceptors
For a radical, it is probable to have one lone pair electron besides a single electron.
In this case, the lone pair electron will be much more favorable to interact with the
proton of the other monomer compared with the single electron. Thus, an interesting
point is raised about the capability of accepting protons for such radicals and the
neutral molecules.
The complex OH· · · H2 O has attracted much attention [32] due to its significance
in chemical processes in the Earth’s atmosphere [33], crystalline and amorphous
ices [34], and aqueous environments including biological systems [35]. Both OH
radical and H2 O molecule are able to be taken as the proton donor and acceptor in
the formation of hydrogen bond, therefore, they could form two different types of
hydrogen-bonded complexes: OH radical as a proton donor and H2 O molecule as
the proton acceptor giving rise to the complex with Cs symmetry (Fig. 5.7a), and on
the contrary, the complex (Fig. 5.7c) with OH as a proton acceptor and H2 O as the
proton donor. The former is much more stable than the latter (the binding energy
22.61 vs. 13.38 kJ/mol), in which the former stable complex has been identified in the
gas phase via microwave spectroscopy [32]. Thus, molecule H2 O is a better proton
acceptor than OH in the complexes composed of H2 O and OH. Lai and Chou [36]
performed a detailed study of the above two complexes composed of OH and H2 O
at the UCCSD/6-31++G(d, p) level, and figured out the nature of OH and H2 O as
proton acceptor in the complexes using NBO analysis. The LPO → BD*O−H orbital
Fig. 5.7 Complexes formed between OH radical and H2 O with binding energy (BE)
interaction plays an important role in stabilizing the complexes (LP and BD* denote
the lone pair and anti-bonding orbital, respectively), where OH radical possesses a
lower LPO and BD*O−H in energy than H2 O. They also performed the calculations
to compare the complexes between H2 O· · · HO and H2 S· · · HO [36]. The binding
energies were estimated to be 14.96 and 4.01 kJ/mol for the proton acceptor H2 O
and H2 S, respectively, which are well correlated with the basicity of H2 O and H2 S
as a proton acceptor.
In order to a get deeper understanding of such interactions, we studied the com-
plexes composed of radicals OH/SH and molecules HCN/HNC [36], which are
important species in interstellar space. The results further confirmed the conclu-
sion that neutral molecules are a better proton acceptor than the radicals, and the
center atom O of H2 O has greater ability of donating electron than the S atom of
H2 S. However, in the complexes, O(S)· · · HCN(HNC), S atom becomes a stronger
proton acceptor than O [36]. The binding energy was estimated to be 4.93 and 6.35
kJ/mol for the complexes O· · · HCN and S· · · HCN, respectively, at the QCISD/6-
311++G(2df,2p) level. Comparing them with that of the complexes HO· · · HCN
(11.79 kJ/mol) and HS· · · HCN (7.65 kJ/mol), we come to the conclusion that the H
atom in OH and SH plays an electron-donating role in the proton acceptor, which is
favorable for the formation of hydrogen bond. This is the first report about the role
of H in the monomer as the proton acceptor.
The interaction mode between OH and H2 O in Fig. 5.7b is the same as that in
Fig. 5.7a, but the structure in the former has C2v symmetry with one imaginary
frequency of 102 cm−1 at the MP2/aug-cc-pVTZ level. The structure in Fig. 5.7d is
stabilized by a two-center three-electron bonding (hemibonding) between the singly
occupied local-π orbital on OH and the doubly occupied local-π lone pair on H2 O
[36]. The hemibonding interaction between OH and H2 O was considered to have
some contribution to the major ultraviolet absorption band of hydroxyl radical in
water [40], even there is some controversy over its existence and the hemibonded
water structure was considered to be an artifact of the exchange-correlation functional
applied [41].
Besides the OH radical, naked atoms could also form the hydrogen-bonded com-
plexes with the proton donor. The NH3 + Cl → NH2 + HCl reaction is relevant to the
chemistry of stratosphere, thus its rate constant and reaction mechanism have been
paid much attention [42, 43]. The reactants NH3 and Cl could form two intermedi-
ates. One hydrogen-bonded complex NH3 · · · Cl with C3v symmetry is formed when
the Cl atom approaches the basin of the three H atoms of NH3 , and a more stable
intermediate complex Cl· · · NH3 with the binding energy of 29.68 kJ/mol is formed
when the Cl atom approaches the N atom of NH3 . The former can eliminate the
HCl molecule via a 32.60 kJ/mol barrier to the final products, NH2 + HCl, whereas
the further decomposition of the latter to Cl–NH2 + H needs to pass a much higher
potential barrier of 203.57 kJ/mol [43].
It has been well established that halogen oxide radicals (ClO) play an important
role in the ozone depletion events during the artic polar springtime [44], thus much
attention has been paid to the hydration of ClO. Experimental investigations have
confirmed the formation of ClO· · · H2 O complex [45], and its structures have also
been unveiled by the theoretical calculations. Two isomers were found: a halogen-
bonded structure with the O atom of water as the electron donor and with the Cl atom
as the electron acceptor, and a hydrogen-bonded structure with the O atom of ClO
as the proton acceptor and water as the proton donor [46]. It has been demonstrated
that the halogen-bonded complex is more stable than the hydrogen-bonded one. The
hydrated structures of ClO(H2 O)n with n = 2–6, were also studied and found that ClO
radical is bonded to water clusters by means of the formation of the OCl· · · OH and
ClO· · · HO network structures and give rise to the global minimum cyclic complex.
For HBe, HMg, H2 B, H2Al, and H2 Ga radicals, they are able to form hydrogen
bonds via not only the single electron on the center atom but also the hydridic hy-
drogen atom. These radicals could form dihydrogen bonds, which is an attractive
interaction between the protonic hydrogen and the hydridic hydrogen [47]. The struc-
tures of dihydrogen-bonded complexes of H2 B and HBe with HCN, HNC, and HF
are presented in Figs. 5.5 and 5.6, respectively. The BeH· · · HCN and BeH· · · HNC
complexes are linear, and the BeH· · · HF complex is bent with the largest binding
energy 16.30 kJ/mol. For H2 B radical, instead of the formation of the conventional
dihydrogen bonds, the bifurcate dihydrogen bonds are formed with two hydrogen
atoms participating in the H· · · H contacts (Fig. 5.5) [29]. The bifurcate dihydrogen
bond is symmetric in BH2 · · · HCN and BH2 · · · HNC complexes but is asymmetric
in BH2 · · · HF.
5.4 Hydrogen Bonds with Radicals as Proton Donors
For radical species, such as CH3 and OH, they possess not only the single electron
but also the acidic protons. Such species, generally, could also provide the protons to
form hydrogen bonds. For instance, the methyl radical, it is able to form the single-
electron hydrogen bonds as a proton acceptor (Fig. 5.4b, 5.4e, 5.4h), and to supply
the proton forming the hydrogen bonds as a proton donor.
We compared the ability of donating protons for different types of radicals and
the corresponding neutral molecules in the complexes Hm X–H· · · NH3 and Hm−1
X–H· · · NH3 (X = O, S, Se, m = 1; X = N, m = 2; and X = C, m = 3) using the MP2,
QCISD, and CCSD(T) methods [48]. The following conclusions were obtained: (1)
the methyl radical could act as the proton donor besides the proton acceptor in the
formation of hydrogen bonds, (2) the methyl radical is more facile to provide the
proton than methane because the binding energy is much higher for H2 CH· · · NH3
complex (4.39 kJ/mol) than that for H3 CH· · · NH3 complex (2.34 kJ/mol) at the
QCISD/aug-cc-pVTZ level, (3) Other radicals are also a better proton donor than
the corresponding neutral molecules, (4) the ability of the radicals to provide the
proton is in order of C–H < N–H < O–H and Se–H < S–H < O–H, which is similar
with that of the neutral molecules, (5) the difference of donating protons between
the radicals and the corresponding neutral molecules is related to both the acidity of
proton donor and the basicity of proton acceptor. The binding energy is larger for the
stronger acidic species, and on the other hand, the stronger proton acceptor leads to
a bigger difference in the interaction strength between the complexes composed of
Hm−1 X–H and Hm X–H, respectively.
As discussed above, the CH3 radical is able to act as the proton acceptor and
donor in the formation of hydrogen bonds, thus, there is a competition between both
types of interaction modes. CH3 radical is facile to act as the proton acceptor when
it interacts with water because the binding energy is 6.44 kJ/mol for H3 C· · · HOH
complex and 4.18 kJ/mol for H2 CH· · · OH2 (Fig. 5.4b). The similar trend is also
found in the systems of H3 C with HCN, HNC, and H2 S. On the contrary, CH3
radical is more favorable to act as the proton donor when it interacts with NH3 with
the binding energy of 3.76 kJ/mol for H3 C· · · HNH2 complex and 4.77 kJ/mol in
H2 CH· · · NH3 (Fig. 5.4h). Thus the ability of donating the proton and electron for
CH3 radical is also related to the acidity and basicity of the other monomer in the
complexes.
The spectroscopic properties of the SH radical and its hydrated complex are cru-
cial in understanding sulfur transformation in the atmosphere. Du and Francisco [49]
performed the high level quantum chemical calculations on the equilibrium geome-
tries, binding energies, and spectroscopic properties of the complex SH· · · H2 O. Two
stable isomers were found with the SH radical as the proton donor forming the global
minimum complex and as the proton acceptor forming the local minimum one. The
binding energies of both complexes are 12.16 and 10.95 kJ/mol, respectively. The
global minimum of SH· · · H2 O complex is less stable than that of OH· · · H2 O due
to less overlap of the lone pair orbital in water LPO with the BD*S−H antibonding
orbital.
Solimannejad and Ghafari [50] analyzed the intermolecular interactions in ternary
radical–molecule complexes between molecules HCN(HNC) and radicals HO(HS)
in gas phase and in water media at the MP2/cc-pVTZ level. There are three structures
for each ternary system with the radical as the proton donor, the proton acceptor,
and a dual role of both donor and acceptor, respectively. The geometries of the first
two are chained with a favorable cooperative effect, whereas the last one is cyclic
with a diminutive effect. Many-body interaction analyses indicate that hydrogen
bonding between two HCN (HNC) molecules gives more stability to the triads than
hydrogen bonding between HCN (HNC) and OH (SH) species. The water media has
an enhancing effect on the stabilities of the complexes. Solimanejad and Scheiner
[51] compared the ability of donating proton of SH and H2 S in the SH· · · N hydrogen
bond with a series of small molecules containing nitrogen atom as the proton acceptor.
They found that the SH· · · N interaction involving with SH radical is slightly stronger
than the H2 S· · · N hydrogen bond in which the closed-shell H2 S serves as the proton
donor.
5.5 Competition Between Single-Electron Hydrogen Bonds and

Other Interactions
It is possible for some radicals that there are several sites of Lewis acid/base, thus,
when they form complexes with other monomers, several types of interactions proba-
bly coexist. Unavoidably, there will be competition among the formation of different
types of interactions. We selected some examples to address this phenomenon.
Besides the formation of the single electron hydrogen-bonded complex, CH3
radical is able to form a covalent-bonded complex with HCN and HNC (Fig. 5.3e,
5.3f). The binding energies are 62.99 and 167.37 kJ/mol, respectively, which are
much higher than that of single-electron hydrogen bonds [17]. H3 C radical essentially
plays different roles in both interactions. It acts as a Lewis acid in the covalent-bonded
complex (Fig. 5.3e, 5.3f), and a Lewis base in the single electron hydrogen-bonded
one (Fig. 5.3a, 5.3b). In the covalent interaction, the CH3 radical interacts with the
Lewis base through the alpha-LUMO. The stronger covalent interaction leads to a
more remarkable change of H3 C geometry, varying from the plane structure in the
isolated monomer to the umbrella-like configuration in the complex. Simultaneously,
the geometries of HCN and HNC also have a big change, varying from a linear
structure in the monomer to a bent one in the complex. Another covalent-bonded
structure is also formed as shown in Fig. 5.3g, 5.3h, where the middle C atom of
HCN and the middle N atom of HNC are introduced into the alpha-LUMO of the
CH3 radical. Similarly, the geometries of this type covalent-bonded complex also
changed significantly. Compared the binding energies of four complexes (Fig. 5.3e,
5.3f, 5.3g, 5.3h), the strength of the covalent-bonded C–N bond is much weaker than
that of the C–C bond, which is in contrast with the bonding energy of the conventional
C–C and C–N bonds.
In a similar way, H2 B radical also forms a covalent-bonded complex (Fig. 5.5g,
5.5h) with HCN and HNC, besides the single electron hydrogen-bonded complex
[30]. The binding energy in the covalent-bonded complex is 241.69 kJ/mol and
291.97 kJ/mol at the MP2/aug-cc-pVTZ level for H2 B· · · HCN and H2 B· · · HNC,
respectively, which are much higher than that in the single electron hydrogen-bonded
complexes (Fig. 5.5a, 5.5b). The covalent-bonded complex of H2 B· · · HNC is more
stable than that of H3 C· · · HNC with the binding energy of 167.28 kJ/mol at the
same level. As a consequence, the H2 B radical is not only a stronger Lewis base in
the hydrogen bond but also a stronger Lewis acid in the covalent interaction than
the H3 C radical. Similar with H3 C, H2 B radical also interacts with the middle C
atom of HCN and the middle N atom of HNC giving rise to the other patterns of
the covalent-bonded structures presented in Fig. 5.5j, 5.5k. Both types of covalent-
bonded complexes, resulting in a change of geometry of HCN and HNC, are more
stable for the HNC complex than for the HCN complex. During the formation of
both types of covalent-bonded complexes, H2 B provides the beta-LUMO to interact
with HCN and HNC, even the energy of the beta-LUMO is higher than that of the
alpha-LUMO. This is different from that in the H3 C counterpart. However, the F
atom of HF attacks the alpha-LUMO of H2 B, giving rise to a complex (Fig. 5.5i)
with the binding energy of 12.92 kJ/mol, which can compete with the single electron
hydrogen-bonded complex with the binding energy of 15.80 kJ/mol.
HBe also forms two types of covalent-bonded complexes with HCN and HNC
by its beta-LUMO, and its alpha-LUMO is used as the Lewis acid to interact with
HF, as shown in Fig. 5.6g–5.6k. However, different from the complexes composed
of H2 B and CH3 radicals, instead of singly covalent-bonded to middle atoms of
HCN and HNC, HBe radical could be bicovalent-bonded to both middle and side
N/C atoms and forms tricyclic structures (Fig. 5.6j, 5.6k). Both covalent interactions
between HBe and HCN are stronger than those in the HNC counterpart, which are
in contrast with the complexes of H2 B with HCN(HNC). On the other hand, the
conformation of the HBe covalent-bonded complex with HF is different from that
with HCN and HNC. The covalent interaction between HBe and HF is roughly equal
to that of the single-electron hydrogen bond and the dihydrogen bond (discussed in
the following paragraph), indicating that there is a competition during the formation
of the three-type complexes between HBe radical and HF molecule.
Besides the formation of the single-electron hydrogen bond, for radicals HBe,
HMg, H2 B, H2Al, and H2 Ga, they could also form dihydrogen bonds with the basic
H atom participating in the hydrogen bond formation. Radicals H2 B, H2Al, and
H2 Ga have two types of dihydrogen bond structures: one has C2v symmetry with the
two basic H atoms interacting with the proton of the other monomer, and the other
is of Cs symmetry with one basic H atom to interact with the proton. The stability
of the dihydrogen-bonded complex is in order of H2 B < H2 Ga < H2Al. For the BH2
radical, the single-electron hydrogen bond is stronger than the dihydrogen bond,
while the opposite phenomenon is found for the AlH2 and GaH2 systems [29].
In H3 C· · · H2 O complex, a local minimum (Fig. 5.4b) is also found with one
of the hydrogen atoms of H3 C pointing towards the oxygen atom [20], which is
lower in binding energy than the most stable H3 C· · · H2 O complex with a single-
electron hydrogen bond. Similar with the covalent bonded complexes of H3 C radical
with molecules HCN and HNC, the covalent-bonded complex of H3 C and H2 O is
also obtained (Fig. 5.4c). This covalent-bonded complex is weaker than the H2 S
counterpart (Fig. 5.4f), which is reverse to the electron-donating ability of O and
S atoms. In the covalent-bonded complex of H3 C and H2 S, the strong covalent
interaction leads to a big change of H3 C and H2 S geometries, and the H–S–H angle
has a large increase from the monomer (92.2◦ ) to the complex (150.6◦ ). Similarly,
the methyl radical also forms a covalent bonded complex with NH3 (Fig. 5.44i) with
the C· · · N distance of 1.497 Å, which is much smaller than the sum of van der Waals
of the C and N atoms (3.2 Å). However, its binding energy is negative, indicating
that it is not favorable to form this complex.
When the C atom in the CH3 radical is replaced with other atom in the Group IV,
the corresponding radicals H3A (A = Si, Ge, Sn, and Pb) could also form the single-
electron hydrogen bond with H3 O+ although it is weaker than that of H3 C. For the
heavier Group IV atom, the hydrogen atoms of H3A have a partial negative charge,
thus they are able to form a dihydrogen bond with the proton in H3 O+ [28]. For the
complexes composed of radicals H3 Si and H3 Ge, the single-electron hydrogen bond
is about 8 ∼ 17 kJ/mol more stable than the dihydrogen bond, whereas the dihydrogen
bond is 33.44 kJ/mol more stable than the single electron hydrogen bond for the H3 Sn
complex. The corresponding dihydrogen bond involving H3 Pb was not obtained
because it changed to be other products such as H2 in the optimization process, due
to the strong dihydrogen bond. These results suggested that oxidation can induce
formation of molecular hydrogen through the formation of unusual hydrogen-bonded
systems [28].
5.6 Nature of Hydrogen Bonds Involving Radicals
Many studies have been performed for hydrogen bonds involving radicals; however,
their nature has not been investigated systematically. Here, we applied the localized
molecular orbital energy decomposition analysis (LMOEDA) [52] implemented in
the GAMESS program [53] to decompose the binding energy of hydrogen bonds
involving radical species into five physical components: electrostatic (E ele ), exchange
(E ex ), repulsion (E rep ), polarization (E pol ), and dispersion (E disp ) energies, and the
results are listed in Table 5.1. It is obvious from Table 5.1 that for each type of
interaction the E ex contribution is the largest among the four attractive terms. The
exchange interaction is correlated with the overlap of the molecular orbitals, thus,
the larger E ex means the larger orbital interactions between two monomers. On the
other hand, the large overlap of orbitals indicates that two interacting monomers are
close in spatial distance, which results in the high repulsive energy E rep . This has
been confirmed by the positive correlation between E ex and E rep in their absolute
value (Table 5.1), where the absolute value of the latter is nearly two times higher
than that of the former. Considering the dependency of E rep on E ex and the fact that
both terms are often not discussed separately in most energy analyses, we paid our
main attention to comparing the contribution of E ele , E pol , and E disp attractive terms
to the stability of the complexes.
For the single electron hydrogen-bonded complexes H3 C· · · HX (X = F, Cl, and
Br), the E ele term is the most negative, indicating a dominant electrostatic contribution
in stabilizing these complexes. With the increase of X atomic number, the E disp term
becomes more negative, and the most negative E ele and E pol occur in H3 C· · · HBr,
which is inconsistent with the smallest binding energy in this complex. This abnor-
mal result is different from most conventional hydrogen bonds, and is rationalized by
Table 5.1 Electrostatic (E ele ), exchange (E ex ), repulsion (E rep ), polarization (E pol ), and dispersion
(E disp ) energies in the complexes at the MP2/aug-cc-pVTZ level. All are in kJ/mol
Eele Eex Erep Epol Edisp
Fig. 5.2a-F − 16.93 − 25.67 47.23 − 11.12 − 16.93
Fig. 5.2a-Cl − 16.47 − 37.16 65.12 − 11.04 − 16.47
Fig. 5.2a-Br − 20.65 − 53.75 94.72 − 15.63 − 20.65
Fig. 5.2b-Br − 6.69 − 16.72 28.51 − 1.80 − 6.69
Fig. 5.2b-I − 12.50 − 31.48 53.71 − 4.60 − 12.50
Fig. 5.3a − 9.45 − 13.21 23.78 − 4.14 − 9.45
Fig. 5.3b − 14.88 − 22.82 42.22 − 8.86 − 14.88
Fig. 5.3c − 4.31 − 7.48 12.37 − 1.21 − 4.31
Fig. 5.3d − 4.35 − 8.61 14.09 − 1.17 − 4.35
Fig. 5.4a − 10.07 − 17.56 30.68 − 4.22 − 10.07
Fig. 5.4b − 6.02 − 9.45 15.80 − 1.63 − 6.02
Fig. 5.4d − 7.44 − 19.40 32.52 − 3.34 − 7.44
Fig. 5.4e − 3.64 − 8.90 14.46 − 0.92 − 3.64
Fig. 5.4g − 5.35 − 10.37 17.39 − 1.30 − 5.35
Fig. 5.4h − 7.69 − 13.50 22.11 − 2.30 − 7.69
Fig. 5.5a − 12.50 − 13.75 24.91 − 4.10 − 12.50
Fig. 5.5b − 21.65 − 28.26 52.50 − 10.07 − 21.65
Fig. 5.5c − 27.42 − 37.12 −68.30 − 14.50 − 27.42
Fig. 5.5d − 2.68 − 7.06 12.54 − 2.38 − 2.68
Fig. 5.5e − 4.77 − 12.16 22.07 − 4.81 − 4.77
Fig. 5.5f − 5.85 − 13.75 24.83 − 6.14 − 5.85
Fig. 5.6a − 9.78 − 11.08 20.15 − 3.34 − 9.78
Fig. 5.6b − 17.18 − 23.07 42.76 − 8.07 − 17.18
Fig. 5.6c − 22.78 − 32.23 59.23 −12.46 − 22.78
Fig. 5.6d − 12.54 − 13.79 24.66 − 4.72 − 12.54
Fig. 5.6e − 17.85 − 23.24 42.80 − 10.12 − 17.85
Fig. 5.6f − 24.08 − 34.65 63.66 − 15.76 − 24.08
the largest E rep in H3 C· · · HBr. However, the single electron halogen-bonded com-
plexes H3 C· · · XH (X = Br and I) are jointly stabilized by electrostatic and dispersion
energies, which respectively accord with the positive electrostatic potential on the
halogen atom [54] and the atomic radius.
For the single electron hydrogen-bonded complexes H3 C· · · HCN and
H3 C· · · HNC, the values of E pol and E disp are almost equal, but both of them are
smaller than E ele , indicative of the electrostatic nature of the single-electron hydro-
gen bond. For the C–H· · · N/C hydrogen bonds in H2 CH· · · NCH and H2 CH· · · CNH
complexes, the contribution of E disp is comparable to that of E ele , and the relatively
large E disp is consistent with the weak interaction.
The dominant role of electrostatic interaction is the same case for the single
electron hydrogen-bonded complexes of H2 B and BeH with HCN, HNC, and HF,
and this conclusion is similar to that in hydrogen bond of water dimer [52]. With
the increase of binding energy, the ratio of E pol to E ele is also increased from 0.33
in H2 B· · · HCN to 0.53 in H2 B· · · HF, and from 0.34 in HBe· · · HCN to 0.55 in
HBe· · · HF. The larger E pol means that the orbitals undergo more observable changes
in their shapes. The E disp contribution is the smallest in these complexes, although it
is greater for the stronger interaction.
For the weak single-electron hydrogen bonds in the complexes of H3 C· · · HOH,
H3 C· · · HSH, and H3 C· · · HNH2 , the contribution of E disp is close to that of E ele .
A similar result is also found for the weak C–H· · ·Y (Y = O, S, and N) hydrogen
bonds.
The dihydrogen bonds in the complexes of BeH with HCN, HNC, and HF are
comparable in strength to the corresponding single-electron hydrogen bonds, thus
the three attractive terms (E ele , E pol , and E disp ) show similar roles in both types of
hydrogen bonds. However, for the weak dihydrogen bonds in the complexes of H2 B
with HCN, HNC, and HF, the three attractive terms make almost equal contribution
to stabilize these complexes.
5.7 Cooperative Effects
Cooperative effect is an important property of hydrogen bonds, which plays an

important role for the applications of hydrogen bonds in molecular recognition,
crystal engineering, and chemical reactions. Thus much attention has been paid
to the cooperativity among hydrogen bonds or with other types of non-covalent
interactions.
An ab initio calculation at the MP2/aug-cc-pVTZ level has been performed
for the complexes H3 C· · · HCN· · · HCN and H3 C· · · HNC· · · HNC [55], where a
single-electron hydrogen bond coexists with another type of hydrogen bond. The
equilibrium structures, frequency shifts, NMR chemical shifts, binding energies,
atom charges, charge transfers, and electron densities in both trimers were studied
in order to investigate the cooperative effect between the single-electron hydrogen
bond and the N· · · H–C and C· · · H–N hydrogen bonds in H3 C· · · HCN· · · HCN and
H3 C· · · HNC· · · HNC, respectively. In H3 C· · · HCN· · · HCN trimer, the binding en-
ergy of single-electron hydrogen bond increases by 22 % compared to that of dimer
H3 C· · · HCN, whereas that of N· · · H–C hydrogen bond increases by 7 %. These re-
sults show significant cooperativity between both types of hydrogen bonds, as well
as the changes in the binding distances and bond lengths. A similar result is also
found for the H3 C· · · HNC· · · HNC trimer.
In H2 B· · · HCN· · · HCN and H2 B· · · HNC· · · HNC trimers, the single-electron

hydrogen bond exhibits a similar cooperative effect with another type of hydrogen
bond [30]. The binding energies are 7.19 and 19.40 kJ/mol in H2 B· · · HCN and
HCN· · · HCN dimers, respectively, that is, the single-electron hydrogen bond in
H2 B· · · HCN dimer is weaker than the hydrogen bond N· · · H–C in HCN· · · HCN.
The formation of the trimer enhances the binding energy of single-electron hydrogen
bond by 170 % in H2 B· · · HCN· · · HCN trimer and 110 % in H2 B· · · HNC· · · HNC,
and by 34 and 37 % for the hydrogen bond in H2 B· · · HCN· · · HCN and
H2 B· · · HNC· · · HNC trimers, respectively. These results support the conclusion
that the stronger interaction has a greater effect on the weaker one [56]. Compared
the binding energy of different system, the enhancement of single-electron hydrogen
bond in H2 B· · · HCN· · · HCN and H2 B· · · HNC· · · HNC trimers is more prominent
than that in H3 C· · · HCN· · · HCN and H3 C· · · HNC· · · HNC ones.
In H2 O· · · H3 O+ · · · C(CH3 )3 complex, where H3 O+ as a double proton donor in-
teracts with H2 O and C(CH3 )3 , respectively, both O· · · H–O and C· · · H–O hydrogen
bonds in the ternary system are less strong than those in the corresponding binary
system, giving rise to smaller bond length changes and red-shifts, thus a negative
cooperative effect is present [25].
5.8 Conclusions
The hydrogen bonds involving radical species have been discussed in this chapter.
In many chemical reactions and processes, it is possible that the radicals form the
complexes with other molecules via different types of interactions, especially the
hydrogen bonds. Great progress has been achieved in the understanding of the hy-
drogen bonds involving radical species, both experimentally and theoretically. Many
radical species as the proton acceptor are able to form the single electron hydrogen-
bonded complexes with the neutral molecules. The free single electron in radical
species participates in the formation of the open-shelled hydrogen bond with its
direction towards the proton of the other molecule. The binding energies of the com-
plexes are generally smaller than 20 kJ/mol. The magnitude of the binding energy
is proportional to the acidity of the proton donor and will be enhanced significantly
with the protonation of the proton donor or the substitutions of the radical species
and neutral molecules with electron-withdrawing and electron-denoting groups, re-
spectively. Theoretically, the high level computational method applied is important
to reproduce the properties of the open-shelled hydrogen-bonded complexes, for
instance, UCCSD(T)/6-311++G(2d,2p) method, including the large basis set and
electron strong correlation effect, could give rise to a good result [18]. On the other
hand, radicals could also form other types of hydrogen bonds including dihydrogen
bonds, with binding energy usually smaller than that of the single-electron hydrogen
bonds. Therefore, there will be competitions to form different types of interactions
between radical species and other molecules. Essentially, the nature of the hydrogen
bonds involving radical species is similar with the conventional ones where the at-
tractive contribution depends on the properties of both the radical itself and the other
monomer. Furthermore, there also exists the cooperativity in the trimers or larger
clusters involving radicals, with the larger effect on the weaker interaction, which is
coincident with that of the conventional hydrogen bonding interactions.
Acknowledgements This work was supported by the Outstanding Youth Natural Science Foun-
dation of Shandong Province (JQ201006) and the Program for New Century Excellent Talents in
University (NCET-2010-0923).
References
1. Wallington TJ, Dagaut P, Kurylo MJ (1992) Chem Rev 92:667–710

2. Calvert JG, Lazrus A, Kok GL, Heikes BG, Walega JG, Lind J, Cantrell CA (1985) Nature
317:27–35
3. Jeffrey GA (1997) An introduction to hydrogen bonding; Oxford University Press: New York
4. de Visser SP Shaik S (2003) J Am Chem Soc 125:7413–7424
5. Lucarini M, Mugnaini V, Pedulli GF, Guerra M (2003) J Am Chem Soc 125:8318–8329
6. Wang BQ, Li ZR, Wu D, Hao XY, Li RJ, Sun CC (2003) Chem Phys Lett 375:91–95
7. Sander W, Roy S, Polyak I, Ramirez-Anguita JM, Sanchez-Garcia E (2012) J Am Chem Soc
134:8222–8230
8. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Montgomery
JA Jr, Vreven T, Kudin KN, Burant JC, Millam JM, Iyengar SS, Tomasi J, Barone V, Mennucci
B, Scalmani G, Cossi M, Rega N, Petersson GA, Nakatsuji H, Hada M, Ehara M, Toyota K,
Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, KleneMLX, Knox
JE, Hratchian HP, Cross JB, Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O,
Austin AJ, Cammi R, Pomelli C, Ochterski JW, Ayala PY, Morokuma K, Voth GA, Salvador
P, Dannenberg JJ, Zakrzewski VG, Dapprich S, Daniels AD, Strain MC, Farkas O, Malick
DK, Rabuck AD, Raghavachari K, Foresman JB, Ortiz JV, Cui Q, Baboul AG, Clifford S,
Cioslowski J, Stefanov BB, Liu G, Liashenko A, Piskorz P, Komaromi I, Martin RL, Fox DJ,
Keith T, Al-Laham MA, Peng CY, Nanayakkara A, Challacombe M, Gill PMW, Johnson B,
Chen W, Gonzalez C, Wong MW, Pittsburgh PA, Pople JA (2009) Gaussian 09, revision A02.
Gaussian Inc, Wallingford
9. Jacox ME (1979) Chem Phys 42:133–148
10. Johnson GL, Andrews L (1980) J Am Chem Soc 102:5736–5741
11. Misochko EY, Benderskii VA, GoldschlegerAU, Akimov AV, Shestakov AF (1995) J Am Chem
Soc 117:11997–11998
12. Tachikawa H (1998) J Phys Chem A 102:7065–7069
13. Clark T, Hennemann M, Murray JS, Politzer P (2007) J Mol Model 13:291–296
14. Chen Y, Tschuikow-Roux E, Rauk A (1991) J Phys Chem 95:9832–9836
15. Raghavendra B, Arunan E (2007) J Phys Chem A 111:9699–9706
16. Rudić S, Merritt JM, Miller RE (2006) J Chem Phys 124:104305
17. Solimannejad M, Alikhani ME (2005) Chem Phys Lett 406:351–354
18. Qi XJ, Liu L, Fu Y, Guo QX (2005) Struct Chem 16:347–353
19. Zhang B, Shiu W, Liu K (2005) J Phys Chem A 109:8983–8988
20. Rudic S, Merritt JM, Miller RE (2009) Phys Chem Chem Phys 11:5345–5352
21. Hashimoto T, Iwata S (2002) J Phys Chem A 106:2652–2658
22. An XL, Liu HP, Li QZ, Gong BA, Cheng JB (2008) J Phys Chem A 112:5258–5263
23. Li QZ, Zhu HJ, An XL, Gong BA, Cheng JB (2009) Int J Quantum Chem 109:605–611
24. Li QZ, An XL, Gong BA, Cheng JB (2008) J Mol Struct Theochem 866:11–14
25. Hammerum S (2009) J Am Chem Soc 131:8627–8635

26. Spiegel DA, Wiberg KB, Schacherer LN, Medeiros MR, Wood JL (2005) J Am Chem Soc
127:12513–12515
27. Pozzi D, Scanlan EM, Renaud P (2005) J Am Chem Soc 127:14204–14205
28. Gil A, Sodupe M, Bertran (2004) Chem Phys Lett 395:27–32
29. Solimannejad M, Alkorta I (2006) J Phys Chem A 110:10817–10821
30. Wang WJ, Li QZ (2011) Comput Theor Chem 966:128–133
31. Li QZ, Li R, Yi SC, Li WZ, Cheng JB (2012) Struct Chem 23:411–416
32. Ohshima Y, Sato K, Sumiyoshi Y, Endo Y (2005) J Am Chem Soc 127:1108–1109
33. Aloisio S, Francisco JS (2000) Acc Chem Res 33:825–830
34. Langford VS, McKinley AJ, Quickenden TI (2000) Acc Chem Res 33:665–671
35. Hamad S, Lago S, Mejias JA (2002) J Phys Chem A 106:9104–9113
36. Lai CH, Chou PT (2007) J Comput Chem 28:1357–1363
37. Jing B, Li QZ, Gong BA, Cheng JB, Li WZ, Liu ZB (2010) Mol Phys 108:1655–1664
38. Jing B, Li QZ, Gong BA, Liu ZB, Li WZ, Cheng JB, Sun JZ (2011) Mol Phys 109:831–838
39. Yamaguchi M (2011) J Phys Chem A 115:14620–14628
40. Chipman DM (2011) J Phys Chem A 115:1161–1171
41. Vande J, Sprik M (2005) Phys Chem Chem Phys 7:1363–1367
42. Gao Y, Alecu IM, Hsieh PC, Morgan BP, Marshall P, Krasnoperov LN (2006) J Phys Chem A
110:6844–6850
43. Xu ZF, Lin MC (2007) J Phys Chem A 111:584–590
44. Sander SP, Friedl RR, Yung YL (1989) Science 245:1095–1098
45. McKeachie JR, Appel MF, Kirchner U, Schindler RN, Benter T (2004) J Phys Chem B
108:16786–16797
46. Galvez O, Gomez PC (2007) Chem Phys Lett 448:16–23
47. Custelcean R, Jackson JE (2001) Chem Rev 101:1963–1980
48. Li QZ, Kou H, Li R, Li WZ, Cheng JB (2011) Comput Theor Chem 976:83–87
49. Du S, Francisco JS (2009) J Chem Phys 130:124304
50. Solimannejad M, Ghafari S (2013) Struct Chem 24:1493–1498
51. Solimanejad M, Scheiner S (2011) Int J Quantum Chem 111:3196–3200
52. Su PF, Li H (2009) J Chem Phys 131:014102
53. Schmidt MW, Baldridge KK, Boalz JA, Elbert ST, Gorden MS, Jensen JH, Koseki S, Matsunaga
N, Nguyen KA, Su SJ, Windus TL, Dupuis M, Montgomery JA Jr (1993) J Comput Chem
14:1347–1363
54. Politzer P, Murray JS, Clark T (2010) Phys Chem Chem Phys 12:7748–7757
55. Li QZ, An XL, Luan F, Li WZ, Gong BA, Cheng JB, Sun JZ (2008) J Chem Phys 128:154102
56. Li QZ, Lin QQ, Li WZ, Cheng JB, Gong BA, Sun JZ (2008) ChemPhysChem 9:2265–2269
Chapter 6
Agostic and Hydrogen-Bonding X–H· · · M
Interactions Involving a d8 Metal Center: Recent
Advances Towards Their Understanding
Jiří Kozelka
Abstract The binding of d8 transition metal ions to X–H bonds (X = non-metal) has
been subject of intense research in the last two decades. Two different types of orbital
interactions can stabilize X–H· · · M bonds: (1) charge transfer from a filled orbital
of the metal into the empty σ*-antibonding orbital of the X–H bond; (2) charge
transfer from the filled σ-bonding orbital of the X–H bond into an empty orbital of
the metal. The first type corresponds to a hydrogen bond, whereas the second is
commonly designated as an agostic bond. The present article analyses experimental
and theoretical approaches to the characterization of these two interaction types in
d8 transition metal complexes, points out some assignment errors that occurred in
the past, and summarizes recent advances towards the understanding of the structure,
dynamics, and physical origin of these weak interactions.
6.1 Introduction
Metal centers can interact with a hydrogen atom which is bound to another atom,
thus forming a weak M· · · H–X bond. Such bonds can have two different origins. In
the first case, they result from the donation of electron density of the X–H σ bond to
an empty orbital of the metal. Such bonds, named agostic bonds, usually fill a vacant
coordination site, i.e., complete the coordination sphere of the metal ion. In the
second type, the σ*-antibonding X–H orbital interacts with a lone-pair of the metal,
thus forming a (non-conventional) hydrogen bond. Here, a metal complex extends its
usual coordination sphere by the hydrogen-bonding interaction. Whereas the agostic
bonds are typical of electron-deficient metal centers, the non-conventional hydrogen
bonds occur in electron-rich centers. According to the number of electrons involved
J. Kozelka ()
Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, UMR 8601 CNRS,
Université Paris Descartes, 45, rue des Saints-Pères, 75270 Paris, France
e-mail: kozelka.jiri@gmail.com
Department of Condensed Matter Physics, Faculty of Science, Masaryk University, Kotlářská 2,
611 37 Brno, Czech Republic

130 J. Kozelka
in the bonding of the system, agostic bonds are formally 3-center-2-electron (3c–2e)
bonds, whereas hydrogen bonds correspond to 3-center-4-electron (3c–4e) bonds.
A particularly interesting class is constituted by square-planar d8 complexes.
These possess sterically accessible filled ndxz , ndyz , and ndz2 orbitals (z axis as-
sumed to be perpendicular to the coordination plane) which can serve as hydrogen
bond acceptors, but they can also form agostic bonds with their vacant ndx2 − y2 and
(n + 1)s orbitals, when no ligands are available to form more stable bonds. Fig-
ure 6.1 shows representative examples of C–H· · · Pt interactions, allowing the two
interaction types to be distinguished at first glance: whereas the hydrogen-bonding
interaction (1) involves a coordinatively saturated square-planar metal ion and a C–H
bond contacting Pt(II) along the z-axis, in the agostic interactions (2–4), the C–H
bonds assume the function of the fourth ligand in an otherwise coordinatively un-
saturated Pt(II) center. Both interaction types involve charge transfer (CT), but in
opposite directions: in the hydrogen-bonding case, from a filled orbital of platinum
to the σ*-antibonding orbital of the C–H bond, and in the agostic interactions, from
the C–H σ-bond to the ndx2 − y2 orbital.
Several review articles on agostic [1–4] and hydrogen-bonding [5–7] X–H· · · M
interactions have appeared. In the present review, we focus on contributions to the
understanding of the physical basis of X–H· · · M interactions involving a d8 metal
center that were published during the last two decades.
6.2 Structural Studies Revealing Agostic or Hydrogen-Bonding

X–H· · · M(d8 ) Interactions
Although it is apparent from Fig. 6.1 that a straightforward distinction between

agostic and hydrogen bonding X–H· · · M(d8 ) interactions can be made simply by
considering whether the interaction is part of the square-planar coordination sphere
(yes: agostic, no: hydrogen-bonding), some confusion occurred in the past. Al-
ready the seminal paper by Brookhart and Green in which the designation “agostic
bond” was introduced [12] contains two erroneous assignments of C–H· · · M(d8 )
hydrogen bonds observed in the structures of the palladium(II) complexes trans-
[PdBr{C4 (CO2 Me)4 H}(PPh3 )2 ] (Fig. 6.2, 5) [13] and trans-[PdI2 (PMe2 Ph)2 ] [14]
to agostic bonds. It is interesting to note that Maitlis et al., who reported the struc-
ture of 5, already suggested that the C–H· · · Pd interaction may be considered as a
hydrogen bond.
In the following decade, Albinati, Pregosin et al. characterized a number of Pt(II),
Pd(II), and Rh(I) complexes featuring a saturated square-planar coordination sphere
and showing an extra intramolecular axial contact with a C–H or an N–H bond
[15–20]. The authors observed that these interactions show spectroscopic features
different from those of typical agostic interactions but did not identify them as hydro-
gen bonds, rather, they designated them as “weak” or “pregostic”. Nevertheless, they
noted that the interaction is 3c–4e in nature [20], and thus similar to hydrogen bond-
ing. Similar intramolecular C–H· · · M contacts were identified in 8-methylquinoline
CCl3
H
S N
6 Agostic and Hydrogen-Bonding X–H· · · M . . .
Pt
S N
1 2 3 4
hydrogen-bonding β-agostic γ-agostic bifurcated δ-agostic
Fig. 6.1 Crystal structures and structural formulas of [Pt(C6 H5 )2 (2,2’-Bi-5,6-dihydro-4H-1,3-thiazine-N, N’)].CHCl3 (1), [Pt(norbornyl) (P, P’–But 2 PCH2 –
CH2 PBut 2 )]+ (2), [Pt(Me)(Pi Pr3 )2 ]+ (3) and [Pt{P(2,6-Me2 C6 H3 )Cy2 }{κ2 -P, C–P(2-Me-6-CH2 –C6 H3 )Cy2 }]+ (4), exemplifying hydrogen-bonding (1) and
agostic (2–4) C–H· · · Pt interactions, respectively. The ORTEP figures were adapted with permission from refs. [8] (http://journals.iucr.org/), [9–11], respectively
131
132 J. Kozelka
Fig. 6.2 Crystal structure

details of trans-
[PdBr{C4 (CO2 Me)4 H}(PPh3 )2 ]
(5) showing the
hydrogen-bonding interaction
which was later assigned
erroneously as agostic [12].
Phenyl substituents on the
phosphorus atoms and ester
groups on the butadienyl
carbons were omitted for
clarity. Reproduced with
permission from [13]
complexes of Rh(I) and Ir(I) by Neve et al. [21, 22]; these authors called the inter-
actions “weak” or “remote agostic”, and incorrectly interpreted them as interactions
with “the unsaturated rhodium (or iridium) center”, although these d8 metal centers
were coordinatively saturated. Obviously, the principal distinction between the situa-
tion where a C–H bond donates electron density into a vacant orbital of an unsaturated
metal center (such as in 2–4), and the opposite case where a lone-pair of a saturated
metal center donates electrons to a vacant, antibonding orbital of a C–H bond (such
as in 1), was not yet made that time. Adding to the confusion, the designations
“pregostic” and “remote agostic” have misled many subsequent authors [2,23–27] to
assume that there exists a class of C–H· · · M(d8 ) interactions intermediate between
agostic and hydrogen-bonding, or constituting a “prestage” to agostic bonding [28].
The assumption of such an intermediate class gained support from a study by Crab-
tree et al. in which the authors examined the structural characteristics of 50 selected
X–H· · · M(d8 ) interactions (X = N, C) [29]. In their selection of compounds, most
of the N–H· · · M angles were over 160◦ , whereas the C–H· · · M angles were gener-
ally lower. The authors concluded that N–H· · · M(d8 ) systems were “best regarded as
hydrogen bonds”, but the C–H· · · M(d8 ) systems were “intermediate in geometry be-
tween the hydrogen bonded N–H· · · M systems and true agostic bonds”. The problem
of the structural analysis by Crabtree et al. [29] was that the compounds examined in-
cluded almost exclusively intramolecular X–H· · · M contacts. Thus, the X–H· · · M
angles reflected primarily geometrical constraints of the ligands and only in the sec-
ond place the electronic requirements of the M· · · H bond. In this context, it is worth
mentioning that intramolecular X–H· · · M contacts are not necessarily attractive but
can also be repulsive. An example of such a repulsive contact is given by the com-
plex cis-[PtCl(NH3 )2 (N9-9-aminoacridine)]+ (6, Fig. 6.3), where it is manifest by the
widened Pt–N9-C9 bond angle of 136◦ [30]. Although in 6, the N–H· · · M contact
has an attractive hydrogen-bonding component and the compound shows NMR-
spectroscopic characteristics of hydrogen bonds, the geometrical constraints of the
6 Agostic and Hydrogen-Bonding X–H· · · M . . . 133
Fig. 6.3 Crystal structure of

cis–[PtCl(NH3 )2 (N9-9-
aminoacridine)]+ (6) showing
a repulsive H1· · · Pt contact
(dashed line). That this
contact is repulsive in spite of
a hydrogen-bonding
component can be inferred
from the widened C9-N9-Pt
angle (136◦ ). Adapted with
permission from [30]
ligand obviously force the hydrogen atom to a repulsive, (i.e., too short) distance,
and the repulsion has to be relieved by widening the Pt–N–C bond angle. Simi-
lar distortions of bond angles, apparently caused by repulsive C–H· · · Pt contacts,
have been seen in [Pd(benzo[h]quinoline)(H2 O){2-(dimethylaminomethy1)phenyl-
N}]+ [31], as well as in complexes reported by Albinati, Pregosin et al. [16–18], as it
has been noted by Hambley [32]. It appears clear from these examples that extract-
ing geometrical preferences of X–H· · · M bonding contacts from crystal structures
of intramolecular H· · · M associations is problematic.
More pertinent was therefore a structural study by Desiraju et al. [33] in which
the authors examined intermolecular bonding X–H· · · M (X = C, N, O) interactions,
searching the Cambridge Structural Database (CSD) for H· · · M contacts shorter than
3.2 Å. In this case, the examined sample was statistically relevant and the scattering
of X–H· · · M bond angles did reflect intrinsic structural preferences of the individ-
ual (X = C, N, or O) groups. Since agostic interactions observed in crystals are
almost exclusively intramolecular (with the exception of the so-called “intermolec-
ular pseudo-agostic bonds”1 [33]), the search yielded predominantly non-agostic
interactions. The conclusion of Desiraju et al. [33] was exactly opposite to that of
Crabtree et al. [29]: C–H· · · M interactions follow the same trend as N–H· · · M and
O–H· · · M interactions. Since it is of interest to deplete from each other the structural
preferences of early and late transition metals (which the study of Desiraju et al. [33]
did not separate), and since in the present review, we are interested principally in d8
metal complexes, we have carried out our own search of the 2014 version of the CSD
database for intermolecular X–H· · · M (M = Pd(II), Pt(II), Rh(I), Ir(I); X = C, N, O)
1
In contrast to “pregostic” or “remote agostic” interactions, pseudoagostic interactions share all
characteristics with agostic interactions; their only particularity is being intermolecular.
134 J. Kozelka
Fig. 6.4 Analysis of CSD

entries for tetracoordinate d8
complexes of Pd(II), Pt(II),
Rh(I), and Ir(I), featuring a
short ( ≤ 2.8 Å)
intermolecular contact
X–H· · · M; X = C (red
symbols) or N (blue
symbols). Plot of the
X–H· · · M angle against the
H· · · M distance. Structures in
which the X–H· · · M
interaction was affected by
disorder were discarded. The
full lines are regression lines
through the two ensembles of
points. The data used for the
plot are available as
supporting material
contacts (M· · · H ≤ 2.8 Å) within square-planar, coordinatively saturated complexes

(such as 1). This search yielded 317 C–H· · · M, 15 N–H· · · M, and 9 O–H· · · M con-
tacts for Pd(II), 132 C–H· · · M, 52 N–H· · · M, and 5 O–H· · · M contacts for Pt(II),
45 C–H· · · M, 4 N–H· · · M, and 0 O–H· · · M contacts for Rh(I), and 8 C–H· · · M,
2 N–H· · · M, and 0 O–H· · · M contacts for Ir(I). The positions of hydrogen atoms
in the O–H· · · M interactions proved generally ill-defined, therefore, we analyzed
only the C–H· · · M and N–H· · · M interactions. The CSD codes of the selected struc-
tures and the corresponding H· · · M distances and X–H· · · M angles are listed in the
Supporting Material. Figure 6.4 shows the plot of the X–H· · · M angle as a function
of the M· · · H distance. It can be seen that both subclasses have a preference for a
linear X–H· · · M angle, and this preference becomes more stringent with decreasing
M· · · H length. The scattering of the values for C–H· · · M and N–H· · · M interac-
tions is similar and the regression lines through the two subsets of points virtually
coincide. This result confirms that the conclusion of Crabtree et al. [29], according
to which C–H· · · M(d8 ) interactions involving a coordinatively saturated d8 metal
atom are structurally different from N–H· · · M(d8 ) interactions and tend to be bent,
is invalid. Nevertheless, the myth of intrinsically bent non-agostic C–H· · · M(d8 )
bonds has propagated until the present time.
Although the suggestion that the C–H· · · M(d8 ) interaction seen in the crystal
structure of compound 5 might be considered as a hydrogen bond appeared as early
as in 1972 [13], the insight that coordinatively saturated M(d8 ) centers can function as
acceptors of hydrogen bonding met growing recognition only after the publication
of the neutron diffraction structure of [NPrn4 ]2 {[PtC14 1 · cis−[PtC12 (NH2 Me)2 ]}
by Brammer et al. [34] which showed a textbook example of a nearly linear N–
H· · · Pt hydrogen bond between the Pt-coordinated methylamine and the Pt center
of [PtC14 12− . This intermolecular hydrogen bond is, of course, strongly aided
by electrostatic attraction, since both the (formally cationic) methylamine and the
Fig. 6.5 X-ray structure of cis–[Pt(C6 F5 )2 (I, N-2-iodoaniline)] (7) and neutron diffraction struc-
ture of trans–[PtCl2 (NH3 )(N–glycine)]·H2 O (8), showing intermolecular N–H· · · Pt and O–H· · · Pt
hydrogen bonds, respectively, between electrically neutral moieties. Reproduced with permission
from [35] and [36], respectively
tetrachloroplatinate dianion bear net charges. However, subsequent crystal structures

revealed cases of intermolecular hydrogen bonds between electrically neutral part-
ners: 1 [8], 7 [35] and 8 [36] (Fig. 6.5) are such examples showing intermolecular
C–H· · · Pt, N–H· · · Pt, and O–H· · · Pt hydrogen bonds, respectively. These structures
are particularly convincing demonstrations that d8 centers can be hydrogen-bonding
acceptors from O–H, N–H, and C–H bonds.
6.3 Spectral Characteristics of Agostic Versus

Hydrogen-Bonding X–H· · · M(d8 ) Interactions
In both agostic and hydrogen-bonding X–H· · · M interactions the X–H bond is weak-
ened which can be detected by measuring a reduced JXH coupling constant and/or a
red-shifted X–H stretching frequency [12, 17, 20]. For both types JMH -coupling can
be frequently detected (if the M nuclide has a nonzero nuclear spin), reflecting the
metal-hydrogen orbital interaction. X–H· · · M interactions also affect the chemical
shift of the metal-bound proton, as outlined in detail below.
The strong anisotropy of square-planar complexes of d8 metal ions, and in particu-
lar, of their electronic d8 system, becomes manifest in the magnetic field surrounding
the complex [37, 38]. Thus, nuclei, and most sensitively those of protons, experience
a deshielding effect if they occupy the region above or below the coordination plane,
close to the z-axis (if the z-axis is defined as the normal to the coordination plane
through the metal atom). In contrast, nuclei located close to the coordination plane
experience a shielding effect.
The shielding effect on in-plane nuclei has been found to be particularly strong
in platinum(II) hydride complexes whose hydridic protons show chemical shifts
136 J. Kozelka
JCH = 149 Hz + δH = +1.1 ppm

+
H
H H
JCH = 64 Hz JCH = 68 Hz H
δH = -1.05 ppm H H H
δH = -5.8 ppm H
JPtH = 136 Hz
Pt Ni
Me3C CMe3 Me3C CMe3
P P P P
Me3C CMe3 Me3C CMe3
9 2
10 4
Fig. 6.6 NMR parameters of agostic complexes of d8 metal ions. Note that in the crystal structures
of 4 and in 9, the methyl protons were localized from the difference Fourrier maps. In the NMR
spectra of 9, the agostic proton could be resolved from the non-agostic ones at 175 K. For 4, the
individual H-resonances of the interacting methyl group could not be resolved
between − 5 and − 24 ppm [39–41]. Buckingham and Stevens [37] used ligand field
theory calculations to show that the observed chemical shifts originate mainly from
paramagnetic shielding from the incompletely filled d-shell of platinum, and not
from high electron density at the hydrogen nucleus as suggested by Chatt et al. [39].
Concurring with this result, d0 and d10 metal hydrides do not show such shielding and
the hydridic protons exhibit positive δ-values [42–44]. In a brilliant paper, Ziegler
et al. reported state-of-the-art DFT-GIAO calculations for a number of octahedral
d6 hydride complexes to show that their negative hydride chemical shifts originated
exclusively from paramagnetic shielding by the metal d6 system [45]. The calcu-
lations which perfectly reproduced the experimental chemical shifts showed that if
only diamagnetic shielding operated, the chemical shifts would be positive (with
respect to TMS) in all cases.
Hydrogen atoms in typical agostic d8 metal complexes, such as those displayed in
Fig. 6.6, show negative chemical shifts. Referring to the negative chemical shifts of
metal hydrides, these negative shifts were sometimes called “hydridic shifts” [46]. As
in the case of the hydride complexes, these negative shifts arise from paramagnetic
shielding, and not from enhanced electron density at the hydrogen nucleus. This has
been shown explicitly by Scherer et al. [46] for several agostic diphosphine-stabilized
Ni(II) alkyl complexes of the type [Ni(alkyl)(P, P’–But 2 PCH2 –CH2 PBut 2 )]+ (alkyl
= ethyl (9) [47], norbornyl, dicyclopentadienyl). For these compounds, DFT/AIM-
derived atomic charges carried by the agostic hydrogen were close to zero, and
DFT-calculated shielding constants, which reproduced well the experimental values,
demonstrated that the experimentally found negative shifts arise from paramagnetic
shielding, as in the case of the hydride shifts. For the dicyclopentadienyl derivative,
these results could be also confirmed by an experimental charge-density analysis
[46]. Another experimental proof that the negative chemical shifts observed for β-
agostic complexes (e.g., 1, 9, 10) arise from the d8 anisotropy and not from enhanced
electron density is provided by the δ-agostic complex 4, whose NMR parameters are
included in Fig. 6.6. We can see that in the three β-agostic complexes, the agostic
proton, which lies close to the coordination plane, is shielded (δ < 0). The difference
between the chemical shifts of the agostic and non-agostic hydrogen atoms attached
to the same carbon is − 6.9 ppm in 9 and − 5.0 ppm in 10. In contrast, in the δ-agostic
complex 4 where the agostic interaction is shared between two methyl protons both
of which lie outside the coordination plane (as shown by the X-ray structure where
the methyl protons were localized from the difference Fourrier maps [11], Fig. 6.1),
the methyl group containing the agostic protons shows a positive chemical shift (2.70
ppm), comparable to that of the other, non-agostic ortho methyl group of the same
dimethylphenyl ligand (2.54 ppm). The fact that only those agostic protons which
lie in the coordination plane show a negative chemical shift supports the above
conclusion that negative shifts of agostic protons arise from paramagnetic shielding
and not from increased electron density on the proton.
Figure 6.6 displays also salient spin-spin coupling constants. The agostic interac-
tion is manifest in a decrease of the1 JCH coupling constant from ∼ 150 to 40–70 Hz,
consistent with a weakening of the agostic C–H bond. In the platinum complexes,
we observe J-coupling of the agostic hydrogen with the195 Pt nucleus (136 Hz in 2,
77 Hz in 10, 64 Hz in 4 (average value for the three methyl Hs)), clearly indicating
orbital interaction.
Hydrogen bonds to d8 metal centers occur exclusively at the apical site of the metal
ion, and the hydrogen atoms experience therefore always paramagnetic deshielding
from the metal atom [37, 38]. This deshielding could, in principle, contain a contri-
bution from a decrease of electron density at the hydrogen nucleus, as observed for
classical hydrogen bonds; however, it has not yet been shown whether this contribu-
tion is significant in X–H· · · M(d8 ) hydrogen bonds. So far, NMR manifestations of
X–H· · · M(d8 ) hydrogen bonding have been reported only for intramolecular bonds,
where these hydrogen bonds have sufficient lifetimes to be observed by NMR. The
first report of low-field shifts of hydrogen atoms interacting with a d8 center by hydro-
gen bonding was that by Miller et al. [38] who reported that nickel(II) complexes of
styrene derivatives such as 11 (Fig. 6.7) show deshielding of protons that are likely to
approach the nickel center from the apical site. Miller et al. brought convincing argu-
ments, referring to the calculations of Buckingham and Stevens [37], to show that the
138 J. Kozelka
11 12
13 14
Fig. 6.7 Salient NMR parameters of d8 metal complexes displaying C–H· · · Ni (11) [38], C–H· · · Pt
(12) [17], and N–H· · · Pt (13, 14) [48] hydrogen bonding. Note that for 11 and 12, the chemical
shift difference between the complexed and uncomplexed ligand is given, whereas for 13 and
14, the chemical shift in the complex is indicated (the chemical shift in the free ligand has not
been reported). The high δ values in 13 and 14 show nevertheless clearly that the N–H hydrogen
experiences a deshielding effect
apical proton experiences paramagnetic deshielding from the metal. However, the au-
thors did not recognize the interaction as hydrogen-bonding. Albinati, Pregosin et al.
subsequently examined d8 complexes such as 12, where C-H bonds undergo similar
interactions with the central atom, and identified the direct Pt· · · H interaction by
measuring the1 JPtH spin-spin coupling constants. Although X–H· · · M(d8 ) hydrogen
bonding is expected to weaken the X–H bond, the Albinati-Pregosin complexes with
C–H· · · M(d8 ) hydrogen bonds showed1 JCH coupling constants which were virtually
unaffected (within experimental error) by the coordination to the metal. On the other
hand, in the complexes 13 and 14 (Fig. 6.7) featuring N–H· · · Pt hydrogen-bonding
interactions, a reduction of the1 JNH coupling constants by ∼20 % with respect to
the free ligands was observed, and the JPtH coupling constants were larger, reaching
the order of magnitude of JPtH observed in agostic complexes (Fig. 6.6). It has to be
emphasized that in 13 and 14, the interaction is aided by the negative charge on the
Pt center and the formally positive charge of the ammonium hydrogen-bond donor,
which adds a large stabilizing electrostatic component to the interaction.
Fig. 6.8 Four d8 complexes containing H· · · M interactions studied by DFT and AIM theory by
Oldfield et al. Reproduced with permission from [25]
6.4 Characterization of Agostic and Hydrogen-Bonding

X–H· · · M(d8 ) Interactions Using AB Initio and DFT
Methods
6.4.1 AIM Analyses of X–H· · · M(d8 ) Interactions
The assumption that “weak” or “pregostic” C–H· · · M(d8 ) interactions constitute

a separate interaction type between agostic bonds and hydrogen bonds (Sect. 6.2)
has prompted several groups to carry out computational studies of this presumed
intermediate class. For instance, Oldfield et al. [25] studied three d8 complexes sus-
pected to show such intermediate C–H· · · M(d8 ) interactions using DFT calculations
and the AIM theory (Fig. 6.8): complex 15, that was reported to show a “weak” C–
H· · · Rh(I) interaction ascribed tentatively (and incorrectly) to a donation of C–H σ
electrons to the (coordinatively saturated) Rh(I) center, [49], complexes 16 [17] and
17 [24] designated as “pregostic”, and complex 18 which was initially assigned as
agostic [50] but later identified as hydrogen-bonding by Brammer et al. [34]. Oldfield
et al. showed that all four complexes had similar electron densities at the bond critical
points (BCP) of the H· · · M bonds; the Laplacians of the electron density were also
similar (Table 6.1), and conform to expected ranges for hydrogen bonds (see the
discussion below). Calculations of the chemical shifts indicated deshielding of the
protons contacting M, in agreement with reported NMR data. Calculation of the total
energy density at the BCP yielded slightly negative values for 15 and 18, for which
partial covalent character of the H· · · M bond was concluded, but slightly positive
values for 16 and 17, for which a purely electrostatic interaction was concluded. How-
ever, this conclusion was in contradiction with a significant overlap found between
the metal dxz/yz orbitals and the C–H σ* orbital in these complexes. In fact, as pointed
out in the recent review by Weinhold and Klein [51], a hydrogen bond always has a
covalent n→ σ* component, thus, considering a hydrogen bond a purely electrostatic
interaction seems problematic anyway. Interestingly, the orbitals overlapping with
the C–H σ* orbital in 16 and 17 were dxz/yz , and not dz2 . This might be related to the
fact that in all four complexes studied the X–H· · · M interaction was intramolecular,
140
Table 6.1 Electron density ρ(rc ) and the Laplacian of electron density ∇ 2 (ρ(rc )) at the H· · · M bond critical point, determined from AIM analyses for selected
hydrogen-bonding and agostic X–H· · · M(d8 ) complexes. For hydrogen-bonding complexes, the stabilization energies SE calculated by the supermolecule
method (where available), and their estimates from ρ(rc ) (Fig. 6.9) are also given.
Complex interaction el. density Laplacian of el. SEa SE(estimate)b Level of theory Ref.
[e/a30 ] density [e/a50 ] [kcal.mol− 1 ] [kcal.mol− 1 ]
a) Hydrogen bonding complexes
c
15 C–H· · · Rh 0.024 0.067 4.8 [25]
c
16 C–H· · · Pt 0.023 0.072 4.6 [25]
c
17 C–H· · · Rh 0.012 0.031 2.4 [25]
c
18 N–H· · · Pt 0.025 0.059 5.0 [25]
d
[Ni(bhac)(phim)] O–H· · · Ni 0.011 0.039 2.2 B3LYP/6-311G(d,p) [26]
[Ni(ahac)(phim)]d O–H· · · Ni 0.012 0.041 2.4 B3LYP/6-311G(d,p) [26]
7 O–H· · · Pt 0.009 0.025 1.8 B3LYP/LANL2DZ/6- [52]
311++G (d,p)e
7 N–H· · · Pt 0.012 0.034 2.4 B3LYP/LANL2DZ/6- [52]
311++G (d,p)e
cis-[PtCl2 (NH3 )2 ].H2 O O–H· · · Pt 0.024 0.061 4.7 4.8 MP2/LANL2DZ(f)/6- [53]
311++G (2d,f,p)
cis-[PtCl2 (NH3 )2 ].H2 O O–H· · · Pt 0.020 0.047 6.0 4.0 MP2/DP/6-311+G (2d,2p)f [54]
f
trans-[PtCl2 (NH3 )2 ].H2 O O–H· · · Pt 0.020 0.047 4.1 4.0 MP2/DP/6-311+G (2d,2p) [54]
f
cis-[Pt(OH)2 (NH3 )2 ].H2 O O–H· · · Pt 0.020 0.045 7.6 4.0 MP2/DP/6-311+G (2d,2p) [54]
trans-[Pt(OH)2 (NH3 )2 ].H2 O O–H· · · Pt 0.020 0.045 5.5 4.0 MP2/DP/6-311+G (2d,2p)f [54]
−
[Pt(C6 F6 )3 (8-OH-quinaldine)] O–H· · · Pt 0.021 0/089 4.2 B3LYP/ LANL2DZ/6- [55]
311++G(d,p)g
J. Kozelka
Table 6.1 (continued)
[Pt(C6 F6 )(bzq)(8-OH-qn)]h C–H· · · Pt 0.034 0.070 6.8 M06/SDD/6-31G*/6- [56]
31G**
[Pt(C6 F6 )(bzq)(2-Me-8-OH-qn)]i C–H· · · Pt 0.036 0.066 7.2 M06/SDD/6-31G*/6- [56]
31G**
[Pt(C6 F6 )3 (8-OH-qn)]−h C–H· · · Pt 0.035 0.070 7.0 M06/SDD/6-31G*/6- [56]
31G**
[Pt(C6 F6 )3 (bzq)(2-Me-8-OH-qn)]−i C–H· · · Pt 0.039 0.079 7.8 M06/SDD/6-31G*/6- [56]
31G**
[Pt(C6 F6 )3 (8-Me-quinoline)]− C–H· · · Pt 0.018 0.064 3.6 B3LYP/LANL2DZ/6- [57]
31G(d,p)
6 Agostic and Hydrogen-Bonding X–H· · · M . . .
1 C–H· · · Pt 0.018 0.048 3.6 B3LYP/LANL2DZ/6- [57]

31G(d,p)
b) Agostic complexes
2 C–H· · · Pt 0.059 0.183 B3LYP/LANL2DZ/ [57]
6-31G(d,p)
3 C–H· · · Pt 0.013 0.049 B3LYP/LANL2DZ/6- [57]
31G(d,p)
4 C–H· · · Pt 0.041 0.134 B3LYP/LANL2DZ/6- [57]
31G(d,p)
[Pt(Et)P,P’-But 2 P(CH2 )3 PBut 2 )]+ C–H· · · Pt 0.053 0.176 B3LYP/LANL2DZ/6- [57]
31G(d,p)
[PdBr(Ph)(1-adamantylPt Bu2 )] C–H· · · Pd 0.023 0.075 BP86/ECP1/6-31G* [58]
141
142
[PdBr(Ph)(2-adamantylPt Bu2 )] C–H· · · Pd 0.029 0.083 BP86/ECP1/6-31G* [58]
[PdBr(Ph)(Pt Bu3 )] C–H· · · Pd 0.018 0.058 BP86/ECP1/6-31G* [58]
t
[PdI(Ph)(P Bu3 )] C–H· · · Pd 0.017 0.053 BP86/ECP1/6-31G* [58]
19 C–H· · · Rh 0.026 − 0.021 B3LYP/DZVP/6-31G* [59]
C–H· · · Rh 0.021 − 0.016
a
Calculated using the supramolecule approach, BSSE-corrected
b
According to the relationship SE = 200*ρ(rc ) determined from Fig. 6.10
c
mPW1PW91/SDD for M, 6-311++G(2d,2p) for the interacting H and its neighboring atoms, 6-31G* and 3-21G* for the other non-metal atoms
d
H2 bhac = acetylacetone-benzoylhydrazone; H2 ahac = acetylacetone-acetylhydrazone; phim = 2-phenylimidazole
e
One unit cell from the experimental neutron-diffraction crystal structure [36] was used without geometry optimization
f
DP = Dolg-Pélissier pseudopotential/pseudoorbital basis including two polarization f orbitals [60, 61]
g
Experimental X-ray coordinates, only the position of the interacting H-atom was optimized
h
bzq = 7,8-benzoquinolinate, 8-OH–qn = 8-hydroxyquinoline
i
bzq = 7,8-benzoquinolinate, 2-Me-8-OH–qn = 2-methyl-8-hydroxyquinoline
J. Kozelka
and the sterical constraints did not allow the X–H bond to get oriented exactly along
the z axis. Overall, Oldfield’s study has confirmed that C–H· · · M and N–H· · · M
contacts involving a coordinatively saturated d 8 center are of the same type, i.e.,
they correspond to hydrogen-bonding interactions.
Another study inspired by the presumed existence of an intermediate class of
square-planar d8 complexes showing an apical C–H· · · M interaction whose nature
was “not very clear” was that of Mukhopadhyay and Pal [26]. The authors synthe-
sized two square-planar Ni(II) complexes bearing a 2-phenylimidazole ligand which
contacts the Ni(II) center with one of its ortho hydrogen atoms. A natural population
analysis based on wavefunctions obtained from DFT calculations showed that the
ortho hydrogen atom involved in the interaction with the Ni(II) atom bears a higher
positive charge than the other ortho hydrogen atom of the same phenyl ring, in line
with a hydrogen bond-like interaction. The topological parameters of the C–H· · · Ni
interaction calculated using the AIM theory (see Table 6.1) were also in agreement
with hydrogen bonding.
Thakur and Desiraju [57] examined twenty crystal structures of metal complexes
from the Cambridge Structural Database (CSD) presenting C–H· · · M interactions,
and investigated the ability of the AIM and NBO theories to distinguish between
agostic and hydrogen-bonding interactions. In particular, using AIM analyses, they
checked whether the criteria of Popelier et al. [62, 63] for the electron density at
the bond critical point and for its Laplacian were applicable. Popelier et al. had
reported ranges 0.002 e/a30 ≤ ρ(rc ) ≤ 0.035 e/a30 and 0.024 e/a50 ≤ ∇ 2 ρ(rc ) ≤ 0.139
e/a50 for hydrogen bonds, whereas for agostic interactions the values were reported to
lie outside these ranges, namely 0.04 e/a30 ≤ ρ(rc ) and 0.15 e/a50 ≤ ∇ 2 ρ(rc ) ≤ 0.25 e/a50
[63]. Among the twenty complexes examined by Thakur and Desiraju figured also
six X–H· · · M(d8 ) interactions (M = Pt(II)), two hydrogen-bonding and four agostic.
The data for these interactions constitute the two bottom hydrogen-bonding entries
and the four top agostic entries of Table 6.1. It can be seen that compound 3, a textbook
example of a γ-agostic interaction, shows ρ(rc ) and ∇ 2 ρ(rc ) values rather typical of
hydrogen bonds. The agostic interaction in 3 is particularly weak, as apparent from
the rapid exchange between the methyl groups of all isopropyl substituents observed
in NMR solution spectra even at 190 K [10]; however, the crystal structure (Fig. 6.1)
leaves no doubt that it is an agostic, and not a hydrogen-bonding interaction. Thus, it
is apparent that the BCP analysis only allows for the distinction between weaker and
stronger X–H· · · M interactions, but is incapable of indicating whether the metal is
electron donor or acceptor. The Popelier criteria simply exploit the fact that hydrogen
bonding is usually weaker than agostic interactions, however, they cannot be used
to distinguish between a hydrogen bond and a particularly weak agostic interaction.
Also, Thakur and Desiraju concluded [57] that “the BCP analysis, based upon the
Popelier criteria, showed poor prediction of the interaction type, when compared
to the experimental reports”. They suggested that the bond ellipticity at the H· · · M
BCP might provide a supplementary parameter, being higher for agostic than for
hydrogen-bonding interactions. Nevertheless, even this is not a general rule, since,
for instance, for the agostic complex 4, the ellipticity at BCP is 0.273 a.u., a value
more typical for hydrogen bonds than for agostic interactions (Table 9 of [57]).
144 J. Kozelka
Fig. 6.9 Crystal structure of

[RhCl(CO)(CAAC)] (CAAC
= 2-(2,6-di-isopropylphenyl)-
6-isopropyl-3,3,9-trimethyl-
2-azaspiro(4.5)decan-1-
ylidene) (19) displaying the
two agostic H· · · Rh bonds as
dashed lines. Reproduced
with permission from [64]
The last two entries in Table 6.1 show BCP data for the rhodium complex 19
[59] featuring two particularly strong agostic C–H· · · Rh interactions that fill jointly
the fourth coordination site (Fig. 6.9) [64]. Here the Laplacians of the BCP electron
densities are negative, indicating rather strong covalent interactions. It is interesting
to recall that very strong hydrogen bonds also show negative ∇ 2 ρ(rc ) values [51].
However, those strong hydrogen bonds show also high ρ(rc ) values which is not the
case here.
Zhang et al. performed DFT (B3LYP/6-311++G(d,p)) calculations and an AIM
analysis on trans–[PtCl2 (NH3 )(N-glycine)]·H2 O (8), and determined ρ(rc ) and
∇ 2 ρ(rc ) values for the two hydrogen-bonding X–H· · · Pt (X = O, N) interactions
occurring in the crystal lattice (Fig. 6.5) [52]. The electron densities and their Lapla-
cians (Table 6.1) are rather low, compared to other hydrogen-bonding interactions
shown in Table 6.1. However, this may be partly caused by the fact that the calcu-
lations were performed on the neutron diffraction coordinates without optimization.
The same team carried out a similar study on the [Pt(C6 F5 )3 (8-hydroxyquinaldine)]−
complex showing an intramolecular O–H· · · Pt hydrogen bond. For this complex,
ρ(rc ) and ∇ 2 ρ(rc ) values were comparable to those of other X–H· · · Pt hydrogen
bonds (Table 6.1). For an intramolecular hydrogen bond, a calculation of the interac-
tion energy using the supermolecule approach is problematic; therefore, the authors
estimated the interaction energy using the observation that the hydrogen-bond energy
is proportional to the electron density at the BCP, as reported by different authors.
In fact, Parthasarathi et al. [65] and Weinhold and Klein [51] have determined, for
various hydrogen-bonded pairs ranging from very weak to very strong, the stabiliza-
tion energy (SE) using the supramolecule approach, and found that it is proportional
to the electron density at the hydrogen bond BCP. Although the level of theory was
not the same in the two studies (MP2/aug-cc-pVDZ versus B3LYP/aug-cc-pVTZ),
the correlation diagrams are remarkably similar, as shown in Fig. 6.10. In particu-
lar, for weak hydrogen bonds (SE < 10 kcal.mol− 1 ), the regression lines virtually
coincide, with a proportionality constant of ∼200 kcal.mol− 1 /ea30 . As the investi-
gated hydrogen-bonded pairs were constituted exclusively from main-group atoms,
we decided to test whether the proportionality would hold also for X–H· · · M(d8 )
Fig. 6.10 Hydrogen-bonding

energy (determined by the
supermolecule approach) as a
function of the electron
density at the hydrogen bond
BCP, for various
hydrogen-bonded pairs. Blue
symbols: Data from
MP2/aug-cc-pVDZ
calculations by Parthasarathi
et al. [65]. Red symbols: Data
from B3LYP/aug-cc-pVTZ
calculations by Weinhold and
Klein [51]
hydrogen bonds, where the lone-pairs at the acceptor atom are very different, and
have included, in Table 6.1, the estimates for hydrogen-bonding energy from ρ(rc )
with the above proportionality constant. It can be seen that the agreement with the cal-
culated interaction energy (where available) is relatively reasonable, the two values
never diferring more than by a factor of 2.
6.4.2 Quantification of Dispersion and Charge-Transfer

Components of X–H· · · M(d8 ) Interactions
Our own interest in hydrogen bonding involving d8 metal ions as acceptors resulted
from our Hartree-Fock and MP2 calculations on platinum(II) complexes interacting
with an axial water molecule [66], which indicated that an uncharged Pt(II) com-
plex such as trans–[Pt(OH)2 (NH3 )2 ] undergoes associations with water molecules
preferentially with the H-atom oriented towards the filled 5dz2 orbital of platinum,
thus forming an O–H· · · Pt hydrogen bond. The energy well of ∼4 kcal.mol− 1
observed in the MP2 binding curve was only very shallow in the HF curve, in-
dicating a major dispersion contribution to the interaction. A similar result was
obtained for trans–[PtCl2 (NH3 )(N-glycine)] interacting with one water molecule
(8) (Figs. 6.5, 6.11); in this case, the hydrogen-bonding orientation of the water
molecule was confirmed by a neutron diffraction analysis [36]. A hydrogen-bonding
orientation with a major dispersion component resulted also as the most favor-
able geometry from high level ab initio calculations on axial water associations of
146 J. Kozelka
Fig. 6.11 Interaction energy

between trans–
[PtCl2 (NH3 )(N-glycine)] and
H2 O (8) calculated as a
function of Pt· · · O distance.
MP2 ( ) and HF ( )
calculations were
BSSE-corrected. Reproduced
with permission from [36]
cis–[PtCl2 (NH3 )2 ] by de Almeida et al. [53, 67], and from MP2 calculations on axial
water interactions with other electrically neutral Pt(II) complexes [54]. Comparing
the axial water association of cis–[PtCl2 (NH3 )2 ] with that of the aquated derivatives
cis–[PtCl(NH3 )2 (H2 O)]+ and cis–[Pt(NH3 )2 (H2 O2 )]2+ , de Almeida et al. found that
whereas for cis–[PtCl2 (NH3 )2 ], the hydrogen-bonding (H-ahead) orientation of the
axial water molecule is favored both by electrostatic and dispersion forces, for the
aquated species, the O-ahead orientation is strongly favored due to electrostatics
[67]. This result is in agreement with our calculations on water interaction with
trans–[Pt(OH)2 (NH3 )2 ] and [Pt(NH3 )4 )]2+ where electrostatic forces favor the H-
ahead orientation in the former case but the O-ahead orientation in the latter, and
significant dispersion interaction is found only in the former case [66].
De Almeida et al. also used perturbation methods to calculate, for the axial
hydrogen-bonding water interaction with cis–[PtCl2 (NH3 )2 ], the electrostatic, in-
duction, and dispersion components of the interaction energy [53]. They showed
that at the MP2 energy minimum, the electrostatic and induction components (the
latter including the charge-transfer energy) are not negligible. Our conclusion that the
polarization and charge-transfer components are not important in axial water associ-
ations of trans-[Pt(OH)2 (NH3 )2 ] [66] and trans–[PtCl2 (NH3 )(N-glycine)] [36] was
therefore probably premature. That at the MP2 energy minimum the HF energy is
close to zero for these interactions (see Fig. 6.11) could be due to the fact that the sum
of electrostatic, polarization, and charge-transfer components is counterbalanced by
exchange-repulsion. Supporting this interpretation, our laterAIM analysis of these in-
teractions indicated significant charge transfer for the H-ahead water orientation [54].
Zhang et al. performed DFT (B3LYP/6-311++G(d, p)) calculations and an NBO
analysis of trans–[PtCl2 (NH3 )(N-glycine)]·H2 O (8) to quantify the charge transfer
Table 6.2 NBO analysis of donor-acceptor interactions for hydrogen-bonding and agostic
C–H· · · Pt(II) complexesa
(2) (2)
Complex Interaction EM→σ ∗ (C−H ) Eσ (C−H )→M CSD Refcode
typeb −1
[kcal.mol ] [kcal.mol− 1 ]
[Pt(C6 F6)3 (8-mehq)]−c HB 4.38 2.75 TOXJEV
1 HB 8.57 1.52 KEKZAB
2 agostic 16.66 39.06 GELKAJ
3 agostic 1.43 3.96 BIMHAH
4 agostic 8.81 20.62 WURJAU
[Pt(Et)P, P’– agostic 3.78 28.28 KILKOF10
But 2 P(CH2 )3 PBut 2 )]+
a
From [57]
b (2) (2) (2)
Considering that EM→σ ∗ (C−H ) > Eσ (C−H )→M for hydrogen bonds, EM→σ ∗ (C−H ) <
(2)
Eσ (C−H )→M for agostic interactions
c
8-mehq = 8-methylquinoline
(CT) from platinum lone-pairs to the σ* antibond of the interacting water molecule,
and to the σ* antibond of the NH+ 3 group of a second complex molecule interacting
from the other apical site (Fig. 6.5). The CT to the O–H and N–H σ* antibonds was
0.0051 and 0.020 e, respectively, and the corresponding second-order perturbative
(2)
energy En→σ ∗ , summed over the three platinum lone-pairs (presumably 5dz2 , 5dxz
and 5dyz ) was calculated as 0.97 and 1.42 kcal.mol− 1 , respectively [52]. In a subse-
quent NBO analysis of the [Pt(C6 F5 )3 (8-hydroxyquinaldine)]− complex showing an
O–H· · · Pt hydrogen bond from the 8-hydroxy group, the same authors determined
the CT from the platinum lone pairs to the O–H antibond of 0.0054 e, and the sum
(2) −1
of the En→σ ∗ values over the platinum lone-pairs of 2.06 kcal.mol [55]. These
values are similar to those calculated for the water-to-platinum hydrogen bond in
8, which may appear surprising, as in the hydroxyquinaldine complex the hydrogen
bond is aided by the net − 1 charge on the platinum center, and thus, considerably
stronger charge transfer would be expected.
Thakur and Desiraju [57] used B3LYP/6-31G(d, p) calculations and an NBO
analysis to examine the power of the NBO method to distinguish between agostic
and hydrogen-bonded C–H· · · M systems. They found that of all examined methods,
NBO analysis “was in best agreement with experimental characterizations by NMR
spectroscopy and X-ray crystallography”. As shown in Table 6.2 for the six investi-
gated C–H· · · Pt(II) interactions, the agostic bonds are distinguished from hydrogen
bonds by the relative CT energies corresponding to C–H to metal CT versus metal to
C–H CT: for agostic complexes, the former is larger whereas for hydrogen-bonded
complexes, the latter is larger. This criterion allows 19 of the 20 complexes inves-
tigated by Thakur and Desiraju to be correctly assigned. The assignment fails for
the agostic complex cis, trans-[RuH2 (1,3-di-tert-butylimidazol-2-ylidene)(PPh3 )2 ]
where the coordination geometry is a distorted octahedron and the calculated σ →M
and M→ σ* CT energies (20 and 29 kcal.mol− 1 ) are similar.
148 J. Kozelka
6.4.2.1 Relationship Between Transferred Charge and Charge-Transfer

Energy
In the 2013 paper, Zhang et al. made the attractive suggestion that one could make
use of the proportionality between transferred charge and charge-transfer energy,
and calculate one value from the other on the basis of a known proportionality con-
stant [55]. This interesting but somewhat problematic issue deserves a comment
here. In the zero-overlap approximation, the charge transfer q can be expressed
(2) (2)
as q = En→σ ∗ /(εσ ∗ − εn ), where En→σ ∗ is the second-order perturbative en-
ergy and εσ ∗ and εn are the unperturbed NBO orbital energies [68]. If variations in
(εσ ∗ − εn )−1 can be neglected with respect to En→σ (2)
∗ , then a proportionality between
(2)
q and En→σ ∗ should be observed. In fact, B3LYP/aug-cc-pVTZ calculations and
an NBO analysis of a series of hydrogen-bonding systems ranging from very weak
(2) −1 (2) −1
( En→σ ∗ ≈ 0.4 kcal.mol ) to very strong ( En→σ ∗ ≈ 180 kcal.mol ) by Wein-
hold and Klein has shown a very good linear correlation (R = 0.994) passing through
0, with a proportionality constant of 417 kcal.mol− 1 /electron transferred, when the
(2) −1
very strong hydrogen bonds ( En→σ ∗ > 100 kcal.mol ) were excluded (Table 6.2
of [51]). An earlier, similar study done at the HF/6–31G* level of theory also yielded
a good proportionality dependence (R = 0.987 with a proportionality constant of 683
kcal.mol− 1 /electron transferred [69]). This indicates that, except for very strong hy-
drogen bonds, for which the zero overlap approximation obviously does not hold,
(2) −1
En→σ ∗ is in fact proportional to q (i.e. the variations in (εσ ∗ − εn ) are not strong
enough to perturb the proportionality); however, the proportionality constant depends
substantially on the level of theory. A still different proportionality constant was de-
rived by Cappelletti et al. [70]. These authors used high-resolution molecular-beam
scattering experiments to determine experimentally the intermolecular potential for
various gas-phase binary complexes, including water complexes with rare gases, H2 ,
N2 and O2 . They evaluated the CT energy to be proportional to the CT (calculated
either using NBO theory or the “charge displacement analysis” [71]) with a rough
proportionality constant of 2–3 eV (46–69 kcal.mol− 1 ) per electron transferred. The
authors also performed NBO analyses of the hydrogen-bonded systems at a very
(2)
high level of theory (CCSD/aug-cc-pVQZ), but did not report the En→σ ∗ CT ener-
gies; therefore, it is not clear whether the discrepancy between their proportionality
constant and that obtained by Weinhold and Klein [51] is due to a discrepancy be-
tween the experimental and theoretical (NBO) CT energies, or from the fact that
the relationship is different for the extremely weak hydrogen bonds (ΔECT ≤ 0.1
kcal.mol− 1 ) for which the experimental CT energies were determined [70].
Cappelletti et al. have also compared the CT in water-noble gas complexes evalu-
ated by the “charge displacement analysis” [71], NBO and AIM methods, and found
that the NBO charges agreed with those from the charge displacement analysis rea-
sonably well [70, 72], whereas the AIM charges were considerably larger [70]. This
finding goes in line with considerably larger CT from trans-[PtCl2 (NH3 )(N-glycine)]
to water determined using AIM calculations (0.04 e) [54] as compared to that calcu-
lated using an NBO analysis (0.005 e) [52], although the latter value was obtained on
neutron diffraction coordinates [36] without optimization, which could reduce the
calculated CT on its own. Clearly, care has to be taken when comparing CT values
obtained with different methods and from different geometries.
6.4.3 Ab-Initio Molecular Dynamics (AIMD) Simulations

of X–H· · · M(d8 ) Interactions
6.4.3.1 Water Solvation of Pt(II) Complexes Probed by AIMD Simulations
Vidossich et al. reported an ab initio molecular dynamics (AIMD) study of the

complex trans–[PtCl2 (NH3 )(N-glycine)] (8) with ∼ 100 explicit water molecules
under periodic boundary conditions [73], in order to test whether the axial hydrogen-
bonding interaction with a water molecule seen in the neutron diffraction structure
[36] would hold in aqueous solution. As can be seen from the radial distribution
functions (rdf) of the axial region (Fig. 6.12, top), the AIMD calculations predict
water molecules to assume preferentially the H-ahead (i.e., hydrogen-bonding) ori-
entation with respect to the platinum atom. This result is in agreement with static in
vacuo calculations on 8 [36] and on other uncharged Pt(II) complexes [53, 54, 66],
and suggests that O–H· · · Pt hydrogen bonds between water and electrically neutral
platinum complexes may hold also in solution.
Similar rdf diagrams indicating hydrogen-bonding orientation of solvating water
molecules towards platinum were derived by Autschbach et al. from AIMD simula-
tions of cis–[PtCl2 (NH3 )2 ] in a water bath consisting of 64 water molecules [74] and
of solvated dianionic complexes [PtX4 ]2− (X = Cl, Br, CN) [75].
Lau and Ensing [76] carried out AIMD simulations on cis–[PtCl2 (NH3 )2 ] and its
two hydrolyzed derivatives, cis–[PtCl(NH3 )2 (H2 O)]+ and cis–[Pt(NH3 )2 (H2 O)2 ]2+ ,
each solvated with ∼ 50 explicit water molecules. For the electroneutral complex cis–
[PtCl2 (NH3 )2 ], they observed a somewhat smaller population of hydrogen-bonding
axial water molecules than the previous authors. For cis–[PtCl(NH3 )2 (H2 O)]+ , the
corresponding rdf peak near Pt–H distance of 2.5 Å had an even smaller amplitude, as
expected for the increased positive complex charge, and for cis–[Pt(NH3 )2 (H2 O)2 ]2+ ,
it was completely absent (Fig. 6.12, middle). These results are in perfect agreement
with the in vacuo calculations of de Almeida et al. [67], and also with the observation
that rate constants for the hydrolysis of chlorido Pt(II) complexes depend only very
slightly on the complex charge ([77], see also discussion in [36]).
In contradiction to the above results are AIMD simulations of [Pt(H2 O)4 ]2+
solvated in a periodic box of 70 water molecules by Marcos et al. [78, 79]. In these
simulations of a dicationic complex, axial water molecules significantly populated
the hydrogen-bonding orientation (Fig. 6.12, bottom).
There is a neutron diffraction structure of a dicationic Pt(II) complex having axial
interactions available: that of [Pt(py)4 ]Cl2 .3H2 O [80]. In this structure, Pt(II) in-
teracts axially with one Cl− counter-ion from one side and with a water molecule
150 J. Kozelka
Fig. 6.12 Radial distribution functions (rdf) in the axial regions of different Pt(II) complexes
solvated in water, obtained from AIMD simulations. Pt-O distance distributions are shown as
solid line, Pt-H distance distributions as dashed or pointed line. Top: trans–[PtCl2 (NH3 )(N-
glycine)] from AIMD simulations based on BLYP (left) and BLYP–D3 (right) functionals.
Reproduced with permission from [73]. Middle: cis–[PtCl2 (NH3 )2 ], cis-[PtCl(NH3 )2 (H2 O)]+ , and
cis–[Pt(NH3 )2 (H2 O)2 ]2+ from AIMD simulations based on the B3LYP functional. Reproduced with
permission from [76]. Bottom: [Pt(H2 O)4 ]2+ from AIMD simulations based on the PBE functional.
Reproduced with permission from [79]
oriented towards platinum with the oxygen lone-pairs from the other side. Obvi-
ously, in a dicationic complex, the axial sites above and below the Pt atom are sites
of highly positive electrostatic potential and favor a lone-pair-ahead orientation of
axial water molecules, as also indicated by in vacuo calculations of interaction en-
ergy [66, 67]. Although Marcos et al. [78, 79] argued that the bulk solvent favors
the hydrogen-bonding orientation, it is difficult to imagine that it would overcom-
pensate the effect of the electrostatic potential. In any case, the AIMD simulation of
Lau and Ensing of cis–[Pt(NH3 )2 (H2 O)2 ]2+ does not indicate any tendency of this
similar dicationic complex to assume the hydrogen-bonding orientation. The differ-
ence observed by these authors between the behavior of cis–[PtCl2 (NH3 )2 ] versus
cis–[Pt(NH3 )2 (H2 O)2 ]2+ in solution perfectly matches the switch from the hydrogen-
bonding orientation seen in the neutron diffraction structure of trans–[PtCl2 (NH3 )(N-
glycine)].H2 O (Fig. 6.5, [36]) to the lone-pair-to-platinum orientation seen for
[Pt(py)4 ]Cl2 .3H2 O [80].
6.4.3.2 Probing the Solution Dynamics of Agostic Interactions in Pt(II)

Complexes with AIMD Simulations
Tricoordinated complexes of d8 metal ions are intermediates in many organometallic

transformations ([58], [81], and references therein). Agostic interactions are fre-
quently involved in the stabilization of these unsaturated complexes. In solution,
and even in crystal structures, these interactions can be fluxional, as exemplified by
the crystal structure of [Pd(4-OMe–Ph){N(3,5-(CF3 )2 C6 H3 )2 }(Pt Bu3 )] where two
molecules are present in the asymmetric unit, one without and one with an agostic
interaction involving a t-butyl substituent of Pt Bu3 [82]. The solution dynamics of
agostic interactions is related to their chemical transformations. Occupation of the
vacancy in unsaturated complexes is an important issue since it can be directly re-
lated to their reactivity. Vidossich, Lledós et al. have therefore performed QM/MM
molecular dynamics simulations of three representative Pt(II) complexes displaying
β-, δ-, and remote, ξ-agostic interactions, respectively, in explicit dichloromethane
solvent [83]. The metal complexes were treated quantum mechanically (PBE/DZVP),
whereas ∼ 1000 dichloromethane solvent molecules and the counterions were de-
scribed using molecular mechanics. The simulations of 15 ps showed a stable agostic
bond for the β-agostic interaction, and exchange between C–H bonds from different
methyl groups of a t-butyl group for the δ-agostic interaction. For the ξ-agostic in-
teraction, two types of movement were observed, exchange between the three C–H
bonds of the agostic methyl group, and movements of this methyl group in and out of
the coordination sphere. Although more studies and longer simulations are needed
to make more specific statements and to correlate the simulations with NMR exper-
iments, this simulation certainly represents a promising initial step towards detailed
dynamic description of agostic bonds at an atomistic level.
152 J. Kozelka
Fig. 6.13 a Alternate mechanistic pathway for MetH-bound Co(II)/Co(I) reduction. b BP86-
optimized Co+ Cbi with an axial water molecule. Reproduced with permission from [90] and [87],
respectively
6.5 Co(I) In Cobalamin-Dependent Transferases: A Biologically

Relevant d8 Acceptor of Hydrogen-Bonding?
Cobalamin-dependent methyltransferases and adenosyltransferases share a com-

mon mechanistic feature: reduction of cob(II)alamin (Co2+ Cbi) to cob(I)alamin
(Co+ Cbi). For example, in the reactivation cycle of E. coli cobalamin-dependent
methionine synthase (MetH), Co2+ Cbi has to be reduced to Co+ Cbi when Co+ Cbi
is accidentally oxidized to Co2+ Cbi [84]. Co+ Cbi is an extremely powerful reduc-
ing agent (the midpoint potential for the Co2+ /Co+ couple at pH 7 is − 490 mV
against SHE for the wild-type E. Coli MetH [85]) and it is not clear how the
Co2+ Cbi→Co+ Cbi reduction is achieved with common biological reducing agents
whose midpoint potentials are higher [86]. A generally accepted hypothesis, based
on spectroscopic and structural studies, invokes the stabilization of the Co+ state by
creating a tetracoordinate environment favoring the d8 configuration ([84] and ref-
erences therein). Although this hypothesis plausibly fits the known principles of the
ligand field theory, Kozlowski et al. challenged it asking why the enzyme should put
up with the energetic disadvantage of creating an empty space on both sides of the
Co-corrin plane and leave the Co+ ion vulnerable to unwanted nucleophilic attack,
when it has the possibility to stabilize the d8 state by an axial O-H· · · Co(I) hydro-
gen bond, either from an inversely coordinated water ligand (Fig. 6.13a) [87–89],
or from a nearby tyrosine ligand [90]. As a test, the authors have used various DFT
functionals to optimize Co2+ Cbi and Co+ Cbi with an axial water molecule. The
calculations have shown that whereas Co2+ Cbi prefers classical coordination with
oxygen lone-pairs, Co+ Cbi attracts water with one of its hydrogen atoms, forming an
O–H· · · Co(I) hydrogen bond (Fig. 6.13b) [87]. The hydrogen bond was found to be
thermodynamically stable and to consist of charge-transfer, correlation, dispersion,
and electrostatic components.
6.6 Conclusion
A considerable amount of experimental and theoretical data has accumulated on

d8 metal complexes presenting X–H· · · M agostic or hydrogen-bonding interac-
tions. Some initially unclear aspects could be clarified, and general characteristics
distinguishing these two different interaction types can be now delineated:
i) In agostic interactions, one (or two in some cases) C-H bond interacts with
a tricoordinate d8 center, completing thus the square-planar tetracoordination typi-
cal of the d8 complexes. Only intramolecular agostic complexes of d8 metals have
been characterized to date. In contrast, hydrogen bonding contacts occur exclusively
between a coordinatively saturated d8 center and an intra- or intermolecular X–H
hydrogen bond donor (X = C, N, O) along the z-axis of the coordination square. The
two cases are exemplified by the structures shown in Fig. 6.1.
ii) Both interaction types involve a covalent component. In agostic bonds, the
latter arises from the charge transfer from the filled C–H σ-orbital to antibonding
orbitals of the metal. In hydrogen bonds, the transfer goes from the filled dz2 , dxz ,
and/or dyz metal orbitals to the X–H σ∗-antibonding orbital. Counting the valence
electrons involved, agostic bonding corresponds to a 3-center-2-electron bond and
hydrogen-bonding to a 3-center-4-electron bond. The donor-acceptor orbital overlap
is the principal determinant of the directionality of X–H· · · M(d8 ) bonding: hydrogen
bonds tend to be linear, whereas agostic bonds tend to be bent. There is no interme-
diate class of C–H· · · M(d8 ) bonds involving a coordinatively saturated d8 center; all
interactions previously interpreted as pregostic, preagostic etc., were shown to be
of the hydrogen-bonding type, and as all hydrogen bonds, tend to linearity. In in-
tramolecular hydrogen bonds, however, a non-linear C–H· · · M angle can be imposed
by the structural constraints of the ligand.
iii) The magnetic anisotropy of the d8 electronic system causes shielding of nuclei
in the coordination plane and deshielding along the z-axis. H-atoms involved in
hydrogen bonding with d8 centers are therefore always deshielded. The effect can be
enhanced by the induction effect typical of all hydrogen bonds; however, its relative
contribution to the overall deshielding in X–H· · · M(d8 ) hydrogen bonds has not been
evaluated so far. Agostic H-atoms are usually located in the coordination plane and
are accordingly shielded. However, in particular cases such as that of complex 4, the
agostic hydrogens lie outside the coordination plane and are not deshielded.
iv) The Bond Critical Point (BCP) analysis using theAIM theory does not allow for
an unambiguous distinction between agostic bonding and hydrogen bonding. Agos-
tic bonds do usually show ρ(rc ) and ∇ 2 ρ(rc ) values outside the range observed for
hydrogen bonds, however, weak agostic bonds can show values in the hydrogen bond-
ing range. Nevertheless, ρ(rc ) has been shown to correlate with hydrogen-bonding
energy, and can be therefore used as an estimate of strength of hydrogen bonding.
v) Natural Bond Orbital (NBO) analysis has been instrumental in evaluating the
σ →M and M→ σ* charge transfer, and the concomitant charge-transfer energies.
Generally, the former is larger for agostic bonds, and the latter for hydrogen bonds.
154 J. Kozelka
Recent computational advances towards the understanding of agostic & hydrogen

bonding X–H· · · M(d8 ) interactions include ab initio molecular dynamics simula-
tions, which allow insights into the dynamic behavior of these systems in solution. A
particularly interesting result from computational studies is the prediction of models
for the Co+ “supernucleophile” occurring in the active site of cobalamin-dependent
methyl and adenosyl transferases, where the d8 state is stabilized by tetracoordina-
tion complemented with axial hydrogen bonding [87, 90]. It will be a challenging
task to test these models experimentally against currently accepted models.
Supporting Material Table of CSD Refcodes and geometrical parameters of X–
H· · · M contacts used in Fig. 6.4
Acknowledgment I would like to thank all students and colleagues whom I was lucky to work with.
Their names figure in the correponding references. I acknowledge support with platinum chemicals
from W. C. Heraeus GmbH and financial support from the Hubert-Curien program “Galileo” and
from COST (Action D39/0004/06).
References
1. Clot E, Eisenstein O (2004) Agostic interactions from a computational perspective: one name,
many interpretations. Struct Bond 113:1–36
2. Brookhart M, Green MLH, Parkin G (2007) Agostic interactions in transition metal compounds.
Proc Nat Acad Sci USA 104:6908–6914
3. Lein M (2009) Characterization of agostic interactions in theory and computation. Coord Chem
Rev 253:625–634
4. Scherer W, Herz V, Hauf C (2012) On the nature of β-agostic interactions: a comparison
between the molecular orbital and charge density picture. Struct Bond 146:159–208
5. Martín A (1999) Hydrogen bonds involving transition metal centers acting as proton acceptors.
J Chem Ed 76:578–583
6. Brammer L (2003) Metals and hydrogen bonds. Dalton Trans 3145–3157
7. Calhorda MJ (2006) In: Grabowski SJ (ed) Hydrogen bonding- new insights, vol 3. Springer,
Dordrecht, pp 245–262
8. Bruno G, Lanza S, Nicolò F (1990) Structure of [Pt(C6 H5 )2 (btz-N, N’)].CHCl3 , btz=2,2’-Bi-
5,6-dihydro-4H-1,3-thiazine. Acta Cryst C 46:765–767
9. Carr N, Dunne BJ, Orpen AG, Spencer JL (1988) Co-ordinatively unsaturated diphosphine
platinum(II) alkyl cations: a new class of β-agostic complexes. J Chem Soc Chem Commun
926–928
10. Ingleson MJ, Mahon MF, Weller AS (2004) [PtMe(i Pr3 P)2 ]+ : a Pt(II) complex with an agostic
interaction that undergoes C–H activation. Chem Commun 2398–2399
11. Baratta W, Stoccoro S, Doppiu A, Herdtweck E, Zucca A, Rigo P (2003) Novel T-shaped
14-electron platinum(II) complexes stabilized by one agostic interaction. Angew Chem Int Ed
42:105–108
12. Brookhart M, Green MLH (1983) Carbon-hydrogen-transition metal bonds. J Organomet Chem
250:395–408
13. Roe DM, Bailey PM, Moseley K, Maitlis PM (1972) Structure of bromobis(triphenyl-
phosphine)-(1,2,3,4-tetrakisrnethoxycarbonylbuta-1,3-dienyl)palladium and evidence for a
C–H· · ·Pd interaction. J Chem Soc Chem Commun 1273–1274
14. Bailey NA, Jenkins JM, Mason R, Shaw BL (1965) Unusual co-ordination of the palladous
ion in the structure of trans-di-iodobis(dimethylphenylphosphine)palladium(II). J Chem Soc
Chem Commun 237
15. Albinati A, Anklin CG, Pregosin PS (1984) Platinum induced C–H activation in aromatic
aldehydes. Unusual J(Pt, H) coupling constants and structure of trans-dichloroquinoline-8-
carboxaldehyde triethylphosphine platinum(II). Inorg Chim Acta 90:L37–L38
16. Albinati A, Arz C, Pregosin PS (1987) n J(Pt, H) and Pt–H–C interactions in Schiff base com-
plexes of 2-(benzylideneamino)-3-methylpyridine. Molecular structures of dichloro(2-((2,4,6-
trimethylbenzylidene)amino)-3-methylpyridine)(triethylarsine)platinum(II) and dichloro-(2-
amino-3-methylpyridine)(triethylphosphine)palladium(II). Inorg Chem 26:508–513
17. Albinati A, Anklin CG, Ganazzoli F, Rüegg H, Pregosin PS (1987) Preparative and 1 H NMR
spectroscopic studies on palladium(II) and platinum(II) quinoline-8-carbaldehyde (1) com-
plexes. X-ray Structures of the cyclometalated acyl complex PdCl(C(O)C9 H6 N)(PPh3 ).PPh3
and trans-PtCl2 (1)(PEt3 ). Inorg Chem 26:503–508
18. Albinati A, Pregosin PS, Wombacher F (1990) Weak Pt–H–C interactions. Extensions to 8-
methylquinoline, benzoquinoline, and a tetralone schiff base. X-ray crystal structure of trans-
PtCl2 (benzoquinoline)(PEt3 ). Inorg Chem 29:1812–1817
19. Bortolin B, Bucher U, Ruegger H, Venanzi LM, Albinati A, Lianza F (1992) Synthesis and
2D NMR studies of {bis(pyrazolyl)borato}rhodium(I) complexes with weak Rh· · · H–C inter-
actions and the x-ray crystal structure of {(cyclooctane-1,5-diyl)bis(pyrazol-1-yl)borato}(1,5-
cyclooctadiene)rhodium. Organometallics 11:2513–2521
20. Albinati A, Lianza F, Pregosin PS, Müller B (1994) New N–H· · · Pt Interactions. The nature
of the bond. Inorg Chem 33:2522–2526
21. Neve F, Ghedini M, Crispini A (1992) Weak Rh–H–C interactions. Molecular structure of
[trans–Rh(CO)(8-methylquinoline)(PPh3 ]BF4. Organometallics 11:3324–3327
22. Neve F, Ghedini M, De Munno G, Crispini A (1991) Aromatic and benzylic C–H activation.
Synthesis and structural characterization of iridium 2-phenylpyridine and 8-methylquinoline
complexes. Organometallics 11:1143–1148
23. Calhorda MJ (2000) Weak hydrogen bonds: theoretical studies. Chem Comun 801–809
24. Lewis JC, Wu J, Bergman RG, Ellman JA (2005) Preagostic Rh–H interactions and C–H
bond functionalization: a combined experimental and theoretical investigation of rhodium(I)
phosphinite complexes. Organometallics 24:5737–5746
25. Zhang Y, Lewis JC, Bergman RG, Ellman JA, Oldfield E (2006) NMR shifts, orbitals, and
M· · · H–X bonding in d8 square planar metal complexes. Organometallics 25:3515–3519
26. Mukhopadhyay A, Pal S (2006) Intramolecular apical C–H· · · M interactions in square-planar
nickel(II) complexes with dianionic tridentate ligands and 2-phenylimidazole. Eur J Inorg
Chem 4879–4887
27. Taubmann C, Öfele K, Herdweck E, Hermann WA (2008) Complexation of (5H)-dibenzo[a,
d]cyclohepten-5-ylidene to palladium(II) via the diazo route and evidence of C–H· · · Pd
Interactions. Organometallics 28:4254–4257
28. Schöler S, Wahl MH, Wurster NIC, Puls A, Hättig C, Dyker G (2014) Bidentate cycloimidate
palladium complexes with aliphatic and aromatic anagostic bonds. Chem Comun 50:5909–
5911
29. Yao W, Eisenstein O, Crabtree RH (1997) Interactions between C–H and N–H bonds and d8
square planar metal complexes: hydrogen-bonded or agostic? Inorg Chim Acta 254:105–111
30. Sundquist WI, Bancroft DP, Lippard SJ (1990) Synthesis, characterization, and biological
activity of cis-diammineplatinum(II) complexes of the DNA intercalators 9-aminoacridine and
chloroquine. J Am Chem Soc 112:1590–1596
31. Deeming AJ, Rothwell IP, Hursthouse MB, New L (1970) Comparison of 8-methylquinoline
and benzo[h]quinoline complexes of palladium(II) with those of related ligands. Crys-
tal and molecular structure of aqua(benzo[h]quinoline)[2-(dimethylaminomethyl)phenyl-
N]palladium(II) perchlorate. J Chem Soc Dalton Trans 1489–1496
32. Hambley TW (1998) Van der Waals radii of Pt(II) and Pd(II) in molecular mechanics models
and an analysis of their relevance to the description of axial M· · · H(–C), M· · · H(–N), M· · · S
and M· · · M (M=Pd(II) or Pt(II)) interactions. Inorg Chem 37:3767–3774
33. Braga D, Grepioni F, Tedesco E, Biradha K, Desiraju GR (1997) Hydrogen bonding in
organometallic crystals. 6. X--H---M hydrogen bonds and M---(H--X) pseudo-agostic bonds.
Organometallics 16:1846–1856
156 J. Kozelka
34. Brammer L, Charnock JM, Goggin PL, Goodfellow RJ, Koetzle TF, OrpenAG (1987) Hydrogen
bonding by cisplatin derivatives: evidence for the formation of N–H· · · Cl and N–H· · · Pt bonds
in [NPrn 4 ]{[PtCl4 ]· cis–[PtCl2 (NH2 Me)2 ]}. J Chem Soc Chem Commun 443–445
35. Casas JM, Falvello LR, Forniés J, Martín A (1996) Syntheses and structures of the complexes
cis–[M(C6 F5 )2 (N–X)] (M=Pd, Pt; N–X=2-iodoaniline, 2-benzoylpyridine) containing N–X
acting as a didentate chelating ligand and displaying I–M or O–M interactions. Inorg Chem
35:56–62
36. Rizzato S, Bergès J, Mason SA, Albinati A, Kozelka J (2010) Dispersion-driven hydrogen bond-
ing: predicted hydrogen bond between water and platinum(II) identified by neutron diffraction.
37. Buckingham AD, Stephens JP (1964) Proton chemical shifts in the nuclear magnetic resonance
spectra of transition-metal hydrides: square-planar platinum(II) complexes. J Chem Soc 4583–
4587
38. Miller RG, Stauffer RD, Fahey DR, Parnell DR (1970) Alkenaryl compounds of nickel(II) and
palladium(II). Influence of the transition metal on ligand proton chemical shifts. J Am Chem
Soc 92:1511–1521
39. Chatt J, Duncanson LA, Shaw BL (1957) A volatile chlorohydride of platinum. Proc Chem
Soc 343
40. Chatt J, Shaw BL (1962) Hydrido-complexes of platinum(II). J Chem Soc 5075–5084
41. Church MJ, Mays MJ (1968) Spectroscopic studies on some new cationic complexes of
platinum(II). J Chem Soc (A):3074–3078
42. Bercaw JE, Marvich RH, Bell LG, Brintzinger HH (1972) Titanocene as an intermediate in
reactions involving molecular hydrogen and nitrogen. J Am Chem Soc 94:1219–1238
43. Manriquez JM, McAlister DR, Sanner RD, Bercaw JE (1978) Reduction of carbon monoxide
promoted by alkyl and hydride derivatives of permethylzirconocene. JAm Chem Soc 100:2716–
2724
44. Caulton KG, Goeden GV (1981) Soluble copper hydrides: solution behavior and reactions
related to CO hydrogenation. J Am Chem Soc 103:7354–7355
45. Ruiz-Morales Y, Schreckenbach G, Ziegler T (1996) Origin of the hydridic 1 H NMR chemical
shift in low-valent transition-metal hydrides. Organometallics 15:3920–3923
46. Scherer W, Herz V, Brück A, Hauf C, Reiner F, Altmannshofer S, Leusser D, Stalke D (2011)
The nature of β-agostic bonding in late-transition-metal alkyl complexes. Angew Chem Int Ed
50:2845–2849
47. Conroy-Lewis FM, Mole L, Redhouse AD, Lister SA, Spencer JL (1991) Synthesis of co-
ordinatively unsaturated diphosphine nickel(II) and palladium(II) β-agostic ethyl cations:
X-ray crystal structure of [Ni{But2 P(CH2 )2 PBut2 }(C2 H5 )][BF4 ]. J Chem Soc Chem Commun
1601–1603
48. Pregosin PS, Rüegger H, Wombacher F, van Koten G, Grove DM, Wehman-Ooyevaar
ICM (1992) New Pt· · · H–N bonds characterized by15 N-filtered and 2D NOESY 1 H NMR
spectroscopy. Magn Reson Chem 30:548–551
49. Yoshida T, Tani K, Yamagata T, Tatsuno Y, Saito T (1990) Preparation and struc-
ture of [Rh{(h5 –C5 H4 (2-C5 H4 N))(h5 –C5 H4 PPh2 )}(cod)]PF6 and [Ir(H){Fe[h5 –C5 H3 (2-
C5 H4 N)](h5 –C5 H4 PPh2 )}(cod)]PF6 ; a RhI complex having a C–H· · · RhI interaction and a
hydrido IrIII complex (where cod = cyclo-octa-1,5-diene). Chem Commun 292–294
50. Hedden D, Roundhill DM, Fultz WC, Rheingold AL (1986) Reaction chemistry of some new
hybrid phosphine amide complexes of platinum(II) and palladium(II). Isolation and X-ray
structure determination of an ortho-metalated platinum(II) complex derived from a chelated
phosphine amide complex of platinum(II). Organometallics 5:336–343
51. Weinhold F, Klein RA (2012) What is a hydrogen bond? Mutually consistent theoretical and
experimental criteria for characterizing H-bonding interactions. Mol Phys 110:565–579
52. Li Y, Zhang G, Chen D (2012) Theoretical investigation of hydrogen bonding between water
and platinum(II): an atom in molecule (AIM) study. Mol Phys 110:179–184
53. Fedoce Lopes J Da Silva JCS Rocha WR De Almeida WB Dos Santos HF (2011) Quantum
chemical study of cisplatin-water complexes: an investigation of electron correlation effects.
J Chem Theory Comput 10:371–391
54. Bergès J, Fourré I, Pilmé J, Kozelka J (2013) A quantum chemical topology study of the
water-platinum(II) interaction. Inorg Chem 52:1217–1227
55. Zhang G, Li X, Li Y, Chen D (2013) Electron density characteristics and charge transfer effect
of hydrogen bond O–H· · · Pt(II): atoms in molecules study and natural bond orbital analysis.
Mol Phys 111:3276–3282
56. Baya M, Belio U, Martin A (2014) Synthesis, characterization, and computational study of
complexes containing Pt· · · H hydrogen bonding interactions. Inorg Chem 53:189–200
57. Thakur TS, Desiraju GR (2007) Theoretical investigation of C–H· · · M interactions in
organometallic complexes: a natural bond orbital (NBO) study. J Mol Struct: THEOCHEM
810:143–154
58. Stambuli JP, Incarvito CD, Bühl M, Hartwig JF (2004) Synthesis, structure, theoretical studies,
and ligand exchange reactions of monomeric, T-shaped arylpalladium(II) halide complexes
with an aditional, weak agostic interaction. J Am Chem Soc 126:1184–1194
59. Sassmanshausen J (2011) Agostic or not? Detailed density functional theory studies of the
compounds [LRh(CO)Cl], [LRh(COD)Cl] and [LRhCl] (L = cyclic (alkyl)(amino)carbene,
COD = cyclooctadiene). Dalton Trans 40:136–141
60. Andrae D, Häussermann U, Dolg M, Stoll H, Preuss H (1990) Energy-adjusted ab initio
pseudopotentials for the second and third row transition elements. Theor ChimActa 77:123–141
61. Kozelka J, Bergès J (1998) Ab Initio calculations on cis–[PtCl2 (PMe3 )2 ]: search for a model
chemistry for platinum(II) complexes. J Chim Phys 95:2226–2240
62. Koch U, Popelier PLA (1995) Characterization of C–H–O hydrogen bonds on the basis of the
charge density. J Phys Chem 99:9747–9754
63. Popelier PLA, Logotheties G (1998) Characterization of an agostic bond on the basis of the
electron density. J Organomet Chem 555:101–111
64. Lavallo V, Canac Y, DeHope A, Donnadieu B, Bertrand G (2005) A rigid cyclic
(alkyl)(amino)carbene ligand leads to isolation of low-coordinate transition-metal complexes.
65. Parthasarathi R, Subramanian V, Sathyamurthy N (2006) Hydrogen bonding without borders:
an atoms-in-molecules perspective. J Phys Chem A 110:3349–3351
66. Kozelka J, Bergès J, Attias R, Fraitag J (2000) O–H· · · Pt(II): hydrogen bond with a strong
dispersion component. Angew Chem Int Ed Engl 39:198–201
67. Fedoce Lopes J, Rocha WR, Dos Santos HF, De Almeida WB (2008) Theoretical study of the
potential energy surface for the interaction of cisplatin and their aquated species with water. J
Chem Phys 128:165103
orbital, donor-acceptor viewpoint. Chem Rev 88:899–926
69. Reed AE, Weinhold F, Curtiss LA, Pochatko DJ (1986) Natural bond orbital analysis of molec-
ular interactions: theoretical studies of binary complexes of HF, H2 O, NH3 , N2 , O2 , F2 , CO,
and CO2 with HF, H2 O and NH3. J Chem Phys 84:5687–5705
70. Cappelletti D, Ronca E, Belpassi L, Tarantelli F, Pirani F (2012) Revealing charge-transfer
effects in gas-phase water chemistry. Acc Chem Res 45:1571–1580
71. Belpassi L, Infante I, Tarantelli F, Visscher L (2008) The chemical bond between Au(I) and the
noble gases. Comparative study of NgAuF and NgAu+ (Ng = Ar, Kr, Xe) by density functional
and coupled cluster methods. J Am Chem Soc 130:1048–1060
72. Belpassi L, Tarantelli F, Pirani F, Candori P, Cappelletti D (2009) Experimental and theoretical
evidence of charge transfer in weakly bound complexes of water. Phys Chem Chem Phys
11:9970–9975
73. Vidossich P, Ortuno MA, Ujaque G, Lledos A (2011) Do metal· · · water hydrogen bonds hold in
solution? Insight from ab initio molecular dynamics simulations. Chem Phys Chem 12:1666–
1668
74. Truflandier LA, Sutter K, Autschbach J (2011) Solvent effects and dynamic averaging of 195 Pt
NMR shielding in cisplatin derivatives. Inorg Chem 50:1723–1732
75. Truflandier LA, Autschbach J (2010) Probing the solvent shell with 195 Pt chemical shifts:
density functional theory molecular dynamics study of PtII and PtIV anionic complexes in
aqueous solution J Am Chem Soc 132:3472–3483
158 J. Kozelka
76. Lau JK-C, Ensing B (2010) Hydrolysis of cisplatin—a first-principles metadynamics study.
Phys Chem Chem Phys 12:10348–10355
77. Martin DS Jr (1967) Anomalies in ligand exchange reactions for platinum(II) complexes. Inorg
Chim Acta Reviews 87–97
78. Beret EC, Pappalardo RR, Doltsinis NL, Marx D, Sánchez Marcos E (2008) Aqueous PdII and
PtII : anionic hydration revealed by Car-Parrinello simulations. ChemPhysChem 9:237–240
79. Beret EC, Martínez JM, Pappalardo RR, Sánchez Marcos E, Doltsinis NL, Marx D (2008)
Explaining asymmetric solvation of Pt(II) versus Pd(II) in aqueous solution revealed by ab
initio molecular dynamics simulations. J Chem Theory Comput 4:2108–2121
80. Wei CH, Hingerty BE, Busing WR (1989) Structure of tetrakis(pyridine)platinum(II) chloride
trihydrate: unconstrained anisotropic least-squares refinement of hydrogen and non-hydrogen
atoms from combined X-ray-neutron diffraction data. Acta Cryst C 45:26–30
81. Urtel H, Meier C, Eisenträger F, Rominger F, Joschek JP, Hofmann P (2001) A neutral three-
coordinate alkylrhodium(I) complex: Stabilization of a 14-electron species by γ–C–H agostic
interactions with a saturated hydrocarbon group. Angew Chem Int Ed 40:781–784
82. Yamashita M, Hartwig JF (2004) Synthesis, structure, and reductive elimination chemistry of
three-coordinate arylpalladium amido complexes. J Am Chem Soc 126:5344
83. Ortuno MA, Vidossich P, Ujaque G, Conejero S, Lledos A (2013) Solution dynamics of agos-
tic interactions in T-shaped Pt(II) complexes from ab initio molecular dynamics simulations.
Dalton Trans 42:12165–12172
84. Koutmos M, Datta, Pattridge KA, Smith JL, Matthews RG (2009) Insights into the reactivation
of cobalamin-dependent methionine synthase. Proc Nat Acad Sci USA 106:18527–18532
85. Jarrett JT, Choi CY, Matthews RG (1997) Changes in protonation associated with substrate bind-
ing and cob(I)alamin formation in cobalamin-dependent methionine synthase. Biochemistry
36:15739–15748
86. Jarrett JT, Hoover DM, Ludwig ML, Matthews RG (1998) The mechanism of
adenosylmethionine-dependent activation of methionine synthase: a rapid kinetic analysis of in-
termediates in reductive methylation of cob(II)alamin enzyme. Biochemistry 37:12649–12658
87. Kumar M, Kozlowski PM (2011) A biologically relevant Co1+ · · · H bond: possible implications
in the protein-induced redox tuning of Co2+ /Co1+ reduction. Angew Chem Int Ed 123:8861–
8864
88. Kumar M, Kumar N, Hirao H, Kozlowski PM (2012) Co2+ /Co+ redox tuning in methyltrans-
ferases induced by a conformational change at the axial ligand. Inorg Chem 51:5533–5538
89. Kumar M, Hirao H, Kozlowski PM (2012) Co+ –H interaction inspired alternate coordina-
tion geometries of biologically important cob(I)alamin: possible structural and mechanistic
consequences for methyltransferases. J Biol Inorg Chem 17:1107–1121
90. Kumar M, Kozlowski PM (2013) Can the local enzyme scaffold act as an H-donor for a Co(I)–
H bond formation? The curious case of methionine synthase-bound cob(I)alamin. J Inorg
Biochem 126:26–34
Chapter 7
What is Common for Dihydrogen Bond
and H· · ·σ Interaction—Theoretical Analysis
and Experimental Evidences
Sławomir J. Grabowski
Abstract Two types of the hydrogen bond are described and compared here; the A–
H· · ·H–B dihydrogen bond and the A–H· · ·σ interaction. In a case of the dihydrogen
bond the H· · ·H contact between the hydrogen atoms characterized by the opposite
charges is observed; i.e. between the protonic (A)H and hydridic (B)H hydrogens.
For the A–H· · ·σ hydrogen bond the A–H proton donating bond interacts with the
σ-electrons of the molecular hydrogen. These interactions are topologically different
since for DHB the bond path linking the attractors of H-atoms with the corresponding
bond critical point is observed. For the A–H· · ·σ interaction the bond path between
the (A)H-atom attractor and the bond critical point of the H–H bond of the molecular
hydrogen is observed. Both types of the hydrogen bond are characterized by the
significant σ → σ* orbital-orbital interaction, σBH → σAH * in a case of DHB and
σHH → σAH * for A–H· · ·σ. There are also evidences that A–H· · ·H–B and A–H· · ·σ
may be classified as the hydrogen bonds. The examples of complexes characterized
by the mentioned above types of the hydrogen bond are analyzed in this chapter, the
theoretical as well as experimental examples are presented.
7.1 Introduction
It was 20 years ago when R. H. Crabtree and coworkers noted that ¨ Typical hydrogen
bond acceptors possess an O or N lone pair, but more recently, the σ-electron pair of
a transition metal hydride has been shown to give intramolecular N–H· · ·H–M and
O–H· · ·H–M hydrogen bonds of an unconventional type, for which we suggest the
term dihydrogen bonds. These interactions have bond strengths of 4–6 kcal/mol and
H· · ·H distances of ca. 1.7–1.9 Å and seem to play a role in proton transfer, fluxional
processes and other reactions [1].¨ The authors also wrote in the footnote of the
article that ä referee points out that these unconventional hydrogen bonds require a
S. J. Grabowski ()
Kimika Fakultatea, Euskal Herriko Unibertsitatea UPV/EHU, Donostia International Physics
Center (DIPC), P.K. 1072, 20080 Donostia, Euskadi, Spain
e-mail: s.grabowski@ikerbasque.org
IKERBASQUE, Basque Foundation for Science, Maria Diaz de Haro 3, 48013 Bilbao, Spain

160 S. J. Grabowski
Scheme 7.1 The hydrogen

bond and dihydrogen bond hydrogen bond A-H +δ…-δ B
interactions
dihydrogen bond A-H +δ…-δ H-B
new name, for which we suggest the term dihydrogen bond [1]. ¨ The authors refer
to the intramolecular N–H· · ·H–M and O–H· · ·H–M interactions but later articles
concern also intermolecular dihydrogen bonds. Early studies on dihydrogen bond
are briefly described in the next section of this chapter. Additionally the dihydrogen
bonds are classified as ¨ hydrogen bonds of an unconventional type ¨ with typical
proton donating bond, O–H or N–H, and an unconventional proton acceptor, another
hydrogen atom of the hydride bond. Even more, the authors claim that the σ-electron
pair of the hydride bond plays the role of the Lewis base center. The H· · ·H distances
are in the range of 1.7–1.9 Å, less than the sum of the van der Waals radii of H-
atoms (∼ 2.0–2.4 Å). The range of the interaction energy for dihydrogen bonds, 4–6
kcal/mol [2], is situated around the hydrogen bond energy for the water dimer, ca. 5
kcal/mol [3], where typical, conventional O–H· · ·O hydrogen bond exists.
In the article mentioned above [1] and other early studies of Crabtree and co-
workers [4, 5] the dihydrogen bond (DHB) term was reserved for O–H· · ·H–M
and N–H· · ·H–M interactions (where M designates the transition metal or boron)
but the authors have claimed that possible elements connected with the hydridic
hydrogen are ¨currently known as boron and the transition metals; other cases will
no doubt be discovered in the future [6]. ¨ And really further studies have shown
that numerous other interactions can be classified as DHBs [7], even C–H· · ·H–C
contacts were considered as the interactions of this type [8, 9]. It was also found that
for numerous H· · ·H links the energy of interaction exceeds 10 or even 20 kcal/mol
[10–16]. Scheme 7.1 presents the main difference between hydrogen and dihydrogen
bond. For the first interaction the Lewis base center (B) is usually electronegative
atom possessing at least one lone electron pair, for the dihydrogen bond the hydridic
negatively charged H-atom plays the role of the Lewis base center. The same types
of A–H proton donating bonds with the excess of the positive charge on H-atom are
observed for both interactions (Scheme 7.1).
One should note that the dihydrogen bond plays a substantial role in numerous
processes; controlling reactivity and selectivity in solution, σ-bond metathesis, hy-
dride reduction or the ligand attachment to transition metal clusters [7, 17]. Its role
in the hydrogen storage design should be also mentioned [18–21]. The dihydro-
gen bond may be also treated as the preliminary stage of the reaction leading to
the separation of the gaseous hydrogen [17, 22–25]. Reversely, the activation of
the molecular hydrogen may lead to further processes through the dihydrogen bond
formation [26–28]. The dihydrogen bond seems to be important phenomenon since
there are numerous studies concerning this kind of interaction, one can mention re-
view articles [7, 29–32], book chapters [33–35] and even the book monograph [36].
This chapter refers to older concepts and to recent studies to present the current state
of knowledge on the dihydrogen bond interaction.
7 What is Common for Dihydrogen Bond and H· · ·σ Interaction 161
It is interesting to note that Crabtree and coworkers claimed in their early studies
that the σ-bond is the electron donor for dihydrogen bonded systems [37]. They
have presented the concept of the A–H· · ·σ hydrogen bond where A–H is the proton
donating bond (Lewis acid) and σ-electrons (Lewis base) are the proton acceptor.
This is why it is also discussed in this chapter if the A–H· · ·σ and A–H· · ·H–B refer to
the same kind of interactions, i.e. if they may be classified as the dihydrogen bonds,
or if these are different interactions.
7.2 Early Studies and First Concepts on Dihydrogen Bond

Interaction
In 1934 Zachariasen and Mooney reported results on the crystal structure of am-
monium hypophosphite and they have written that “the hydrogen atoms of the
hypophosphite group behave toward ammonium as if they were H− ions” [38]. This
is probably the first study where the existence of the interaction between hydrogen
atoms characterized by the opposite charges was detected. In the later study Burg
analyzed the IR spectrum of the liquid (CH3 )2 NH–BH3 and suggested the existence
of N–H· · ·H3 B interactions possessing the characteristics similar to hydrogen bonds
[39]. Titov and coworkers analyzed the reaction of aminoboranes which leads to the
loss of molecular hydrogen. They have claimed that such a reaction is an effect of
“close spatial arrangement of the oppositely charged hydrogen atoms” [40]. This is
a very important study since it was pointed out there that the interaction between
the opposite charged H-atoms, named in later studies as the dihydrogen bond, is
the preliminary stage of the reaction of the separation of the molecular hydrogen.
This was analyzed in the later studies that the dihydrogen bond may be considered
as the initiation of the separation of the gaseous hydrogen [17, 22–25]. Similarly
the hydrogen bond interaction is often treated as the preliminary stage of the proton
transfer process [41, 42] or the σ-hole bond as the initiation of the SN 2 reaction [43].
In one of early studies Brown and Heseltine have reported the infrared evidence
for a moderately strong and specific interaction between the coordinated BH3 group
in amine-boranes and a protonic H atom [44]. The authors have concluded that the
appearance of a strong, broad, lower frequency stretching absorption in the infrared
spectrum is the proof of the existence of the hydrogen bond. They have also noted
that neither the BH3 group nor the amine-borane molecule possesses a lone pair of
electrons or any electron rich system thus this is the untypical kind of the hydrogen
bond. In another study, Brown and coworkers [45] have reported the appearance of
low-frequency O–H stretching absorptions upon addition of Me3 N and BH3 , to dilute
solutions of methanol and phenol in carbon tetrachloride. The authors concluded that
these absorptions are due to the formation of hydrogen bonded complexes in which
the BH3 group acts as a proton acceptor. They have suggested possible explanations
of the complex formation: the formation of weak B· · ·H–O bonds or the formation
of weak B–H· · ·H–O bonds between the hydridic H atoms attached to boron and the
protonic H atom of OH group or this kind of interaction is a property of the whole
162 S. J. Grabowski
Fig. 7.1 The structure of

ReH2 (CO)(NO)(PMe3 )2 with
indole; the H· · ·H contact
corresponding to the
N–H· · ·H–Re dihydrogen
bond is presented (broken
line), this structure was taken
from the Cambridge
Structural database [54]
followed ref. [53]
BH3 group. One can see that one of those explanations suggest the existence of the
B–H· · ·H–O connections, i.e. the dihydrogen bonds (Scheme 7.1).
The important studies justifying definitively the existence of the attractive interac-
tion between hydrogen atoms of the opposite charges concern crystal structures. The
first studies on structures containing such interactions were performed in nineties of
the last century. In 1990 R. C. Stevens and coworkers reported the hydridohydroxy
complex cis-[IrH(OH)(PMe3 )4 ]PF6 crystal structure analyzed at 20 K by single-
crystal neutron diffraction [46]. The neutron diffraction results confirm earlier X-ray
study on this crystal structure [47]. The authors have postulated, on the basis of
geometrical results, the existence of an attractive H−δ . . .+δ H interaction between the
hydridic Ir–H bond and the electron-deficient hydrogen of the OH group. In 1994
R.H.Morris and coworkers have found the Ir–H−δ . . . + δ H–N intramolecular link be-
tween an iridium hydride and the protonated nitrogen of a sulfur-bonded thiopyridine
ligand in the crystal structures analyzed and they have claimed that this link is a new
type of interaction [48]. It seems that in the middle of nineties of the last century two
groups of researchers have performed pioneering crystal structure analyses where
inter- and intramolecular attractive H· · ·H interactions were detected; the group of
Morris [48–50] and that one of Crabtree and coworkers [1, 4–6, 37, 51, 52]. One
should also mention the early experimental studies on dihydrogen bonds of Epstein
and Shubina and coworkers [25, 29, 30, 33, 53] and of others [7]. Figure 7.1 presents
fragment of the crystal structure where the dihydrogen bond interaction is presented.
The dihydrogen bond was usually defined as an attractive interaction between
the H-atoms of the opposite charge or as an interaction between a conventional
hydrogen bond donor such as an NH or OH bond as the weak acid component
and an element-hydride bond as the weak base component, where the element in
question can be a transition metal or boron [5]. One can see that two concepts are
presented here, the center-center interaction between the acidic and basic hydrogen
atoms and the concept of the unconventional hydrogen bond where the σ-electrons
of the hydridic bond play the role of the Lewis base. The latter kind of the hydrogen
bond may be designated as A–H· · ·σ, where A–H is the proton donating bond; such
understanding of the dihydrogen bond interaction was proposed by Crabtree and
coworkers [37]. The authors have also pointed out that there is a possibility that in a
case of M–H σ-bonds, where M designates the transition metal, the metals possess
nonbonding dπ electrons which can act as the alternative proton acceptors. Hence
they have analyzed the N–H· · ·H–B interactions (B designates boron here) to exclude
the possibility of the interaction with dπ electrons [37]. They have performed the
search through Cambridge Structural Database [54] which has shown twenty-six
intermolecular interactions of that kind in crystal structures with the corresponding
H· · ·H distances in the range 1.7–2.2 Å. The N–H· · ·H angle for those contacts tends
to be close to linearity with the majority of angles in the range 150–170◦ , B–H· · ·H
angle tends to be bent with the majority of angles in the range 95–115◦ . This is why the
authors have pointed out that for such interactions the N–H proton donating bond is
directed towards the σ-B–H hydridic bond but not to the negatively charged H-atom.
This is worth mentioning that the neutron diffraction results on the crystal structure
of ammonia-borane are in line with these findings since N–H· · ·H–B dihydrogen
bonds were found there with the N–H· · ·H and B–H· · ·H angles equal to 156(3)◦
and 106(1)◦ , respectively, and with the H· · ·H intermolecular distance amounting
2.02(3) Å [37]. However the mentioned above concept of the A–H· · ·σ hydrogen
bond was based only on the geometry of the structures considered. It was described
and discussed later [32] that there are other energetic and topological results which
do not support the idea of the B–H and M–H σ-bonds acting as a whole as proton
acceptors in hydrogen bonded systems.
The BH3 NH3 ammonia-borane may be treated as the model example for the
explaining of the significance of the dihydrogen bond interactions [37]. One can see
that ethane and ammonia-borane are isoelectronic but there are large differences in
their properties; the melting point is −181 C and +104 ◦ C for ethane and ammonia-
borane, respectively. Such a difference between ethane and ammonia-borane is partly
related to the polarity of BH3 NH3 (5.2 D). However the melting point for also polar
CH3 F (1.8D) and also isoelectronic with C2 H6 and BH3 NH3 , is equal to −140◦ C.
This means that for ammonia-borane the other factors influence additionally on its
properties, not only polarity, and that these factors are related to the existence of the
dihydrogen bond interactions. This situation is similar to that one of water where
the existence of hydrogen bonds influences its physical properties; water is often
compared to the isoelectronic species of methane.
The next studies have shown that the range of dihydrogen bonding interactions
may be extended; in one of the first reviews on DHBs, and especially on crystal
structures containing such interactions, Custelcean and Jackson have pointed out that
DHB is an interaction between hydridic hydrogens of M–H bonds, where M = Al,
B, Ga, Ir, Mo, Mn, Os, Re, Ru, W) and traditional X–H proton donors where
X = F, O, N, C [7]. Bakhmutov, in the monograph on dihydrogen bonded systems,
describes more types of the dihydrogen bond [36]. Even the C–H· · ·H–C dihydrogen
164 S. J. Grabowski
bonds were analyzed by Wolstenholme and Cameron in the crystal structures of

tetraphenylphosphonium squarate, bianthrone, and bis(benzophenone)azine [9].
7.3 First Theoretical Calculations on the Dihydrogen Bond
Liu and Hoffmann analyzed theoretically the [Ir{H(η1 -SC5 H4 NH)(PCy3 )}2 ]+ BF− 4
structure [55], synthesized earlier by Morris and coworkers [48], where Cy desig-
nates cyclohexyl. The BF− 4 ion was neglected in the theoretical study to simplify the
calculations; also the PCy3 group was replaced with the PH3 group in calculations
[55]. The calculations were performed with the use of the extended Hückel method.
The authors have found for the cation considered the weakly attractive intramolec-
ular Ir–H· · ·H–N interaction, with the H· · ·H distance amounting 1.75 Å. Liu and
Hoffmann analyzed also the linear F–H· · ·H–Li system performing calculations at
the RHF/6-31G* level and finding the H· · ·H interaction energy of 9.3 kcal/mol [55].
These are ones of the first calculations carried out for the intra- and intermolecular
dihydrogen bonded systems.
The BH3 NH3 crystal structure containing dihydrogen bonds was mentioned in
the previous section; also the unusual physical properties of BH3 NH3 which reflect
the existence of the DHB interactions were presented. The ammonia-borane and its
derivatives were the subject of numerous analyses. For example, in one of the first
theoretical studies on DHBs [1, 5], the BH3 NH3 dimer in the gas phase was consid-
ered theoretically since the PCI-80/B3LYP [56, 57] calculations were performed for
it. The authors have found the structure being in energetic minimum where molecules
are linked through two identical B–H· · ·H–N interactions. The energy corresponding
to each H· · ·H interaction in this structure was evaluated to amount 6.1 kcal/mol.
More extended and systematic calculations, up to the MP2/6-31G** level, were
performed on a series of A–H· · ·H–B dihydrogen bonds (A = B, Li, Be; B = N, C)
[10]. The interaction energies for the systems analyzed are in the range typical for
the hydrogen bonds. Even in cases of the BH− +
4 · · ·HCN and BeH2 · · ·NH4 complexes
strong charge assisted interactions are observed since the binding energies for them
are equal to 18.0 and 9.3 kcal/mol, respectively (BSSE correction included).
Ab initio calculations on intermolecular and intramolecular DHBs for Mo, W, Ru,
Re and Ir complexes have been carried out showing the H· · · H interaction energy
greater than 5 kcal/mol and sometimes greater than 10 kcal/mol [12–14], similarly
as for strong conventional O–H· · · O and N–H· · · O hydrogen bonds. There are other
early theoretical studies on dihydrogen bonded systems, for example the calculations
(MP2, MP4, QCISD(T) and B3LYP methods) were performed on the dihydrogen
bonded complexes between the hydrides LiH, NaH, BeH2 , MgH2 , CH4 , SiH4 , GeH4 ,
SnH4 and hydrofluoric acid and the correlations between H· · ·H distance and the
binding energy were presented [58, 59].
7.4 Dihydrogen Bonds as a Sub-Class of Hydrogen Bonds
It was pointed out in few early studies that the dihydrogen bond being the special
type of the hydrogen bond is mainly electrostatic in nature interaction. However
such findings generally do not concern the dihydrogen bonded systems but only the
samples of complexes analyzed. For example, in one of early studies, the decompo-
sition of the energy of interaction within the Kitaura-Morokuma scheme [60] was
performed for two dihydrogen bonded systems, LiH· · ·HF and LiH· · ·HOH [12].
The same type of the decomposition was also performed for two complexes linked
through conventional F–H· · ·O and O–H· · ·O hydrogen bonds in HNO· · ·HF and
HNO· · ·HOH complexes for comparison [12]. It was found that in both cases, of
DHBs and of the conventional hydrogen bonds, the electrostatic interaction energy
(ES) is the most important attractive term followed by the charge transfer term (CT)
and next by the polarization (PL). If one considers the absolute values of the men-
tioned above energies thus these results show that |CT| is less than 30 % of |ES| if
all four hydrogen and dihydrogen bonded systems are taken into account. If only di-
hydrogen bonded systems are taken into account thus |CT| is less than 25 % of |ES|.
This is probably the first study on DHBs where the decomposition of the energy of
interaction was performed to deepen the understanding of those interactions [12].
The IMPPT (intermolecular Møller–Plesset perturbation theory) decomposition
scheme [61] was applied for the H· · ·H interactions in LiH· · ·H2 , LiH· · ·CH4 ,
LiH· · ·C2 H6 and LiH· · ·C2 H2 complexes calculated up to the MP2/aug-cc-pVTZ
level [62]. The energy partitioning was carried out to the second order. This type of
decomposition was also performed for the water dimer linked through the O–H· · ·O
hydrogen bond for comparison [62]. The authors have pointed out that a very similar
picture is presented for the LiH· · ·C2 H2 complex and the water dimer because for
both complexes the main binding contribution comes from electrostatic interaction.
For these complexes the weights of the attractive electrostatic (negative) and the re-
pulsive exchange (positive) energies are such that their sum, i.e. the Heitler-London
interaction energy term, is negative. However for the remaining complexes, charac-
terized by weak interactions, LiH· · ·H2 , LiH· · ·CH4 and LiH· · ·C2 H6 , the picture is
different since the repulsive exchange contribution outweighs the electrostatic term
thus the Heitler-London energy is positive and the dispersion energy is the main
attractive contribution among the remaining attractive terms. The authors conclude
that these results justify the classification of the LiH· · ·C2 H2 complex to be linked
through the dihydrogen bond. However the remaining complexes characterized by
the H· · ·H contacts can be classified as weak van der Waals complexes.
The binary complexes of ammonia-borane, aminoborane and ammonia linked
through hydrogen and/or dihydrogen bonds were analyzed [63]. The Kitaura-
Morokuma decomposition of the energy of interaction [60] performed on the
MP2/aug-cc-pVDZ results shows differences between the dihydrogen and hydro-
gen bond. In the latter case there is much greater significance of ¨non-electrostatic¨
attractive terms than in a case of the dihydrogen bond. This means that for the
hydrogen bonds the polarization, charge transfer, correlation and the higher order
166 S. J. Grabowski
terms of the energy of interaction play more important role than for the dihydrogen
bonded systems where the stronger dominance of the electrostatic attraction is ob-
served. However in both cases of dihydrogen and hydrogen bonds the electrostatic
interaction is the most important attractive term [63].
One can state that the dihydrogen bond, similarly as the hydrogen bond, is an
electrostatic in nature interaction. However the complexes presented so far are linked
through weak or at most medium in strength interactions. The strongest dihydrogen
bond interaction where the interaction energy partitioning was performed presented
here so far occurs for the LiH· · ·HF complex [55, 58, 59]. The situation changes
significantly, sometimes drastically, for very strong dihydrogen bonds. Del Bene
et al. [64, 65] has carried out the MP2/aug’-cc-pVTZ calculations on the model
systems ranging from weak to strong dihydrogen bonds. For example they have
calculated the binding energy for the LiNCH+ · · ·HLi complex to be equal to 27.1
kcal/mol. This complex is characterized by the short H· · ·H intermolecular contact
of 1.309 Å [64].
One of the studies reveals strong dihydrogen bonds within the NH+ 4 · · ·HBeH,
NF3 H+ · · ·HBeH and NH+ 4 · · ·HBeF complexes [66]. For the NF 3 H +
· · ·HBeH system
optimized at the MP2/aug-cc-pVDZ level the short intermolecular H· · ·H contact
of 1.132 Å and the binding energy (corrected for BSSE) of 22.8 kcal/mol were
calculated. The binding energy for the latter complex calculated by the MP2 method
with aug-cc-pVXZ (X = 2,3) extrapolated to the complete basis set (CBS) amounts
21.6 kcal/mol.
The variation-perturbation approach of the decomposition of the energy of inter-
action [67, 68] was applied for the dihydrogen bonded complexes characterized by
very strong interactions [69]. Since the results of this decomposition are presented
later in this chapter and this approach is not commonly applied in other studies on in-
termolecular interactions thus it is described briefly here. The starting wave functions
of the subsystems are obtained in this decomposition in the dimer-centred basis set
[70]. In contrast to the Kitaura-Morokuma scheme [60] applied also in the analysis of
DHB systems [12, 63] the total interaction energy as well as all of its components in
the variation-perturbation approach are free of basis set superposition error (BSSE)
due to the full counterpoise correction [67, 68]. The following interaction energy
terms are the result of this decomposition.
ΔE = EEL (1) + EEX (1) + EDEL (R) + ECORR . (7.1)
EEL (1) is the first order electrostatic term describing the Coulomb interaction of static
charge distributions of both sub-systems of the complex analyzed; EEX (1) is the re-
pulsive first order exchange component resulting from the Pauli exclusion principle;
and EDEL (R) and ECORR correspond to higher order delocalization and correlation
terms. Delocalization term contains all classical induction, exchange-induction, etc.
from second order up to infinity. Strongly basis set dependent charge transfer term
is included in much less basis set sensitive delocalization contribution [67, 68]. The
correlation term includes dispersion interactions as well as intramolecular corre-
lated electrostatic, exchange, induction and dispersion contributions. The presented
decomposition scheme has been implemented within GAMESS package [71].
The MP2/aug-cc-pVTZ calculations have been performed for the

H2 OH+ · · ·HBeH, H2 OH+ · · ·HBeBeH, H2 OH+ · · ·HBeF, HClOH+ · · ·HBeH,
Cl2 OH+ · · ·HBeH and Cl2 OH+ · · ·HBeF complexes [69]. For all of them very short
H· · ·H intermolecular contacts (1.0–1.3 Å) were observed. These are the shortest
intermolecular distances which have been ever reported, with binding energies for
the corresponding complexes within the range of 13.7–24.3 kcal/mol. The decom-
position of the interaction energy (Eq. 7.1) was performed for these complexes and
for all of them the first order Heitler-London energy (ΔE HL = ΔE EX (1) + ΔE EL (1) )
is positive. For all those complexes linked by DHBs the delocalization term,
ΔE (R)
DEL , is the most important attractive interaction contribution and it is mostly
responsible for the stabilization of the complexes. The correlation energy term
ΔE CORR is much smaller in absolute value, about 3–4 times, in comparison with
the delocalization term. This is different than for the weak and medium in strength
hydrogen and dihydrogen bonds, usually known as electrostatic in nature, where
the electrostatic term outweighs the exchange energy and the ΔE HL term is negative
[72, 73]. However the significance of the delocalization energy is also observed
for the conventional strong hydrogen bonds [72]. This means that the hydrogen
and dihydrogen bonds possess similar characteristics since for both for weaker
interactions the electrostatic interaction is dominant while for stronger ones the
processes connected with the electron charge redistribution being the result of
complexation play the main role. It was pointed out that the hydrogen bond is an
interaction without borders [74] what means that there is the continuous passage
from covalent bonds to H-bonds and further to van der Waals interactions. And the
same is observed for the dihydrogen bonds [75].
The other characteristics of the A–H· · ·B hydrogen bonds are commonly known.
For the shortening of the H· · ·B distance the elongation of the A–H proton donating
bond is observed what is accompanied by the increase of the strength of the hydrogen
bond [76, 77]. The same relationships are observed for dihydrogen bonds [75]. For
example, the complexes linked through the H· · ·H interactions were analyzed where
the HF molecule plays the role of the Lewis acid and different hydrides are the proton
acceptors [58]. The non-linear correlations between the H· · ·H distance and the
binding energy were found, i.e. for shorter intermolecular H· · ·H contacts the stronger
interactions were observed. The elongation of the H–F bond was also observed for
these complexes as a result of dihydrogen bond formation. This elongation correlates
with the binding energy, i.e. for the stronger interactions the greater elongation of
the H–F bond is observed.
For some of hydrogen bonds the shortening of the proton donating bond is ob-
served as a result of complexation with the corresponding stretch shift to higher
frequencies (blue-shift); this is why such interactions are often named as the blue
shifted hydrogen bonds [78]. The blue shifted dihydrogen bonds were also predicted
for a number of complexes: F3 C–H· · ·H–Be–X, F3 C–H· · ·H–Mg–X, F3 C–H· · ·H4 Si,
and the analogues complexes where F3 C–H is replaced by F3 Si–H (X = H, F, Cl
and CH3 ). The calculations for these systems were performed up to the QCISD/6-
31+G(d) level of approximation [79]. One can also mention the study on N–H· · ·H–B
blue-shifted dihydrogen bonds [80] and on the coexistence of blue and red-shifted
168 S. J. Grabowski
H...H distance (kcal/mol)

0
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8
-5
-10
R = 0.97
-15
-20
-25
-30
-35 binding energy (kcal/mol)
Fig. 7.2 The relationship between H· · ·H distance and the binding energy for different dihydrogen
bonded complexes
hydrogen and dihydrogen bonds in valine [81]. However, similarly as for the hy-
drogen bonds, for the dihydrogen bonds the complexation, for the greater part of
interactions, is connected with the elongation of the proton donating bond with the
combined shift of the corresponding stretch frequency to lower frequencies. Even
the dependence between stretch frequency and the bond length for the proton donat-
ing bond was observed for a sample of complexes linked through N–H· · ·H–B and
N–H· · ·H–Al contacts [82].
Different types of DHBs were analyzed; from weak H· · ·H interactions classi-
fied as van der Waals interactions and with binding energies less than 2 kcal/mol,
through DHBs of the medium strength, to strong interactions possessing numerous
characteristics of covalent bonds [75]. 22 different complexes were analyzed where
C2 H2 , C2 FH, CH4 , CFH3 , CF2 H2 , CF3 H, CClH3 , CCl2 H2 , CCl3 H, H3 O+ , HClOH+ ,
Cl2 OH+ , NH+ +
4 , and F3 NH were chosen as the proton donating systems while dif-
ferent beryllium molecules as well as the LiH molecule as the species delivering
the H−δ negatively charged Lewis base center. The correlation between the H· · ·H
distance and the binding energy was found (see Fig. 7.2); it is the non-linear and
monotonic relationship (the second order polynomial dependence). This is worth
to mention that for numerous hydrogen bonded systems the excellent distance—
binding energy linear correlation is observed [83]. However for such studies the
homogenous samples are considered, i.e. the samples of complexes formed by re-
lated species. For the relation presented in Fig. 7.2 different classes of Lewis acid and
Lewis base centers were taken into account. For example, for different dihydrogen
bonded complexes analyzed early the distance—binding energy correlation was not
25
Heitler-London energy (kcal/mol)
20
15
H...H 1.7 Å
10
0
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8
-5
H...H distance (Å)
-10
Fig. 7.3 The relationship between H· · ·H distance and the Heitler-London energy for different
dihydrogen bonded complexes
found [10]. However in that case various Lewis acid and Lewis base centers were
also analyzed.
Figure 7.3 presents the dependence between the H· · ·H distance and the
Heitler-London energy (ΔE HL = ΔE EX (1) + ΔE EL (1) ) calculated within the variation-
perturbation approach (Eq. 7.1). The sample of 22 dihydrogen bonded complexes
[75] mentioned earlier here is considered in this relationship. This is interesting that
for the H· · ·H contacts shorter than 1.7 Å, i.e. for the charge assisted strong dihy-
drogen bonds, the ΔE HL term is positive since the electrostatic energy is outweighed
here by the exchange repulsion. This means that the delocalization (ΔE DEL (R) ) is
dominant for such strong interactions. And really, for short H· · ·H distances the
ΔE DEL (R) /ΔE EL (1) ratio is much greater than for the weaker not charge assisted DHBs
[72, 75]. It was also found within the variation-perturbation decomposition scheme
that in general for strong and very strong hydrogen and dihydrogen bonds the ΔE HL
term is positive and the delocalization interaction is much more important than the
electrostatic one [72, 73].
7.5 Could the Dihydrogen Bond Be Classified as the A–H. . . σ

Interaction?
It is interesting to analyze processes of the electron charge redistribution being a result

of the formation of the dihydrogen bond. Such processes were often analyzed for the
hydrogen bonded systems [83]. It was pointed out, within the Natural Bond Orbitals
170 S. J. Grabowski
scheme [84, 85], that the A–H· · ·B hydrogen bond formation is a combination of
two effects: the hyperconjugative A–H bond weakening and the rehybridization-
promoted A–H bond strengthening [84–87]. The first effect is very well known
and it was often analyzed in numerous studies [83–85]. It is connected with the
electron charge transfer from the lone pair orbital of the B Lewis base center to
the antibonding σ* orbital of the A–H bond. The second process of rehybridization
leads to the increase in the s-character of the A-atom hybrid orbital of the A–H bond
∗
[86]. The hyperconjugative effect is connected with the nB → σAH donor-acceptor
interaction. The second-order perturbation energy of this interaction is calculated
according to Eq. 7.2.
E (nB → σ AH * ) = −2 nB |F | σ AH * 2 /(ε (σ AH * ) – ε(nB )) (7.2)

∗ ∗
where nB |F|σAH is the Fock matrix element and (ε(σAH ) − ε(nB )) is the orbital en-
ergy difference. Equation 7.2 may be applied only for the conventional hydrogen
bonds where the Lewis base center is characterized by the electron lone pair. How-
ever it can be slightly modified to include broader spectrum of interactions recently
classified as hydrogen bonds as well as to include the other Lewis acid—Lewis base
interactions. For example, the πB → σ*AH orbital-orbital interaction corresponds
to the A–H· · ·π hydrogen bond where π-electrons play the role of the Lewis base
center [32].
It is interesting that for the A–H· · ·H–B dihydrogen bonds [32], the σB−H → σ*AH
donor-acceptor interaction within the Natural Bond Orbitals (NBO) method is the
most important term connected with the electron charge transfer from the Lewis base
to the Lewis acid. The Eq. (7.2) may be modified to be applied for the dihydrogen
bonded systems:
2
E (σ BH → σ AH * ) = −2 σ BH |F | σ AH ∗ /(ε(σ AH * ) − ε(σ BH )). (7.3)
The terms of Eq. 7.3 are defined in the similar way as those of Eq. 7.2. It seems that
the σB−H → σ*AH interaction energy may be treated as a measure of the strength
of dihydrogen bond. For example, the H· · ·H intramolecular contacts possessing
characteristics of the dihydrogen bond were analyzed recently theoretically (MP2/6-
311++G(d, p) level) [88]. Different classes of dihydrogen bonds were considered,
among them intramolecular C–H· · ·H–C interactions. However for two classes of in-
tramolecular interactions, O–H· · ·H–B and O–H· · ·H–Al the most important changes
being the result of the formation of dihydrogen bonds were detected. Figure 7.4
presents the relationships between the H· · ·H distance and the energy expressed by
Eq. 7.3 for two samples mentioned above and containing BH2 or AlH2 hydridic
group acting as the Lewis base. The excellent exponential correlations show that the
σB−H → σ*AH energy of interaction may be treated as the indicator of the strength
of dihydrogen bond here.
It was found recently that the hyperconjugative and rehybridization processes may
be considered as steering the formation of numerous non-covalent interactions such
as dihydrogen bond, halogen bond, dihalogen bond, hydride bond etc, not only the
hydrogen bond [89–91]. Let us analyze in detail the rehybridization process in a case
ENBO (kcal/mol)
3.0
2.5
2.0
1.5 Al-H...H-O
R² = 0.9888
1.0
B-H...H-O
R² = 0.9632
0.5
0.0
1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6
H...H distance (Å)
Fig. 7.4 The exponential dependences between the H· · ·H distance and the σB−H → σ*AH energy of
interaction (designated as ENBO in the figure) for two classes of intramolecular dihydrogen bonds;
O–H· · ·H–B and O–H· · ·H–Al; this figure was prepared on the basis of the results of ref. [88]
of the hydrogen bond formation. It leads to the increase in s-character of A hybrid or-
bital in the A–H bond; this is in line with Bent’s rule which states that atoms maximize
the s-character in hybrid orbitals aimed toward electropositive substituents and the
atoms minimize such character (maximize the p-character) toward more electroneg-
ative substituents [92]. The A–H· · ·B hydrogen bond formation usually leads to the
increase of the positive charge of the H-atom [83] thus this atom may be treated as
more electropositive in the hydrogen bond than for the A–H bond not involved in any
interaction. Hence the Bent rule may be applied for the hydrogen bonded systems and
really it was found that the hydrogen bond formation is connected with the increase
in the mentioned above s-character [93]. The same is observed for the A–H· · ·H–
B dihydrogen bonds. The N–H· · ·H–Be and N–H· · ·H–Mg dihydrogen bonds were
analyzed recently [91] in the NH+ +
4 · · ·HBeH and NH4 · · ·HMgH complexes, respec-
tively (the MP2(full)/6-311++G(3df,3pd) level calculations were performed). The
QTAIM atomic integrated charges were calculated for these complexes and for the
Lewis acid and Lewis base sub-units not involved in interactions. It was found that
the charge of H-atom in the free NH+ 4 cation is equal to 0.509 au while the charge
of the H-atom being in contact with the hydridic hydrogen in the NH+ 4 · · ·HBeH and
NH+ 4 · · ·HMgH complexes is equal to 0.528 au in both cases; thus the increase of
the positive charge (electropositivity) of H-atom is observed as a result of the for-
mation of the dihydrogen bond. According to the Bent rule the s-character of the
nitrogen hybrid orbital should increase. And this situation is really observed since
the s-character amounts 25.0 for free ammonia cation and it is equal to 28.6 and 30.2
in the NH+ +
4 · · ·HBeH and NH4 · · ·HMgH complexes, respectively.
Thus NBO results justify that the dihydrogen bond is the sub-class of the hydrogen
bond interactions. However the NBO method shows also the difference between those
172 S. J. Grabowski
Fig. 7.5 The molecular graph of the water dimer, (H2 O)2 , big circles correspond to attractors, small
ones to the bond critical points, solid and broken lines represent the bond paths; the broken one
corresponds to the H· · ·O intermolecular contact representing the existence of the hydrogen bond
∗
interactions, the nB → σAH and σB−H → σ*AH are the most important orbital-orbital
interactions for the hydrogen and dihydrogen bond, respectively [32]. These findings
also support the concept that the dihydrogen bond may be classified as the A–H· · ·σ
hydrogen bond since for the σB−H → σ*AH interaction the BH orbital represents here
σ-electrons acting as the Lewis base. However one should note that the mentioned
above orbital-orbital interaction is only the part of the charge transfer interaction
within the NBO approach. Besides there are the other attractive interactions, as it was
discussed earlier here, such as electrostatic, polarization and dispersion. And most
often, the electrostatic interaction is the most important attractive one for hydrogen
and dihydrogen bonded systems.
Crabtree and coworkers have pointed out that there are conventional A–H· · ·B
hydrogen bonds and that the dihydrogen bond may be classified as the A–H· · ·σ
unconventional hydrogen bond; similarly as the A–H· · ·π interaction which also
possesses characteristics typical for the hydrogen bond [37]. It seems that the QTAIM
approach [94, 95] may be useful to clarify the nature of interactions discussed here.
The bond path (BP), being the link between attractors, with the corresponding bond
critical point (BCP) is often treated as the evidence of the stabilizing interaction
[96]. It was pointed out that the molecular graph representing the structure analyzed
shows bond paths which are the indicators of the preferable interactions. In a case
of the A–H· · ·B hydrogen bond the H· · ·B bond path is usually detected, where
H and B designate the attractors of the hydrogen atom and of the B Lewis base
center, respectively. Figure 7.5 shows the molecular graph of the water dimer where
one can see the H· · ·O bond path corresponding to the O–H· · ·O hydrogen bond.
The distinct situation is observed for the A–H· · ·π interactions. For example, for
the C2 H2 · · ·C2 H4 complex the molecules are linked through the C–H· · ·π hydrogen
bond (Fig. 7.6) where the H-attractor of the C–H proton donating bond of acetylene
is connected with the BCP of the double bond of ethylene. One can say that the
C = C bond critical point of C2 H4 mimics here ¨ the point Lewis base center ¨; the
C2 H4 molecule possesses two-center (the C = C double bond) electron donating
system. However there are numerous studies where benzene or other multicenter
aromatic systems play the role of the proton acceptor in the A–H· · ·π hydrogen
bonds. Figure 7.7 shows the molecular graph of the Cl3 CH· · ·C6 H6 complex where
the C–H proton donating bond is directed to the center of the benzene molecule and
Fig. 7.6 The molecular graph of the C2 H2 · · ·C2 H4 complex, big circles correspond to attractors,
small ones to the bond critical points, solid and broken lines represent the bond paths; the broken
line—H· · ·BCP bond path, corresponds to the H· · ·π interaction representing the C–H· · ·π hydrogen
bond
Fig. 7.7 The molecular graph

of the Cl3 CH· · ·C6 H6
complex, big circles
correspond to attractors,
small ones to the bond, ring
and cage critical points, solid
and broken lines represent the
bond paths. Broken lines
represent six H· · ·C bond
paths which correspond to the
H· · ·π interaction
where six H· · ·C bond paths linking the H-atom attractor of the Cl3 CH molecule
with the carbon attractors of the benzene ring are observed.
The π → σ*AH orbital-orbital interaction is the main contribution to the charge
transfer interaction for the system linked through A–H. . . π contact with two-center
π-electron system. The situation is more complicated for many-center π-electron
systems like benzene and other aromatic species [32, 97]. The QTAIM and NBO
approaches were applied recently to discuss different types of the A–H· · ·π hydrogen
bonds [97].
174 S. J. Grabowski
Fig. 7.8 The molecular

graphs of the NH+ 4 · · ·HBeH
(up) and H2 OH+ · · ·HBeH
(below) complexes, big
circles correspond to
attractors, small ones to the
bond critical points, solid
lines represent the bond
paths; the electron density
isolines are also presented
This is interesting that for the A–H· · ·H–B dihydrogen bonds the molecular graphs
show the existence of H· · ·H bond paths with the corresponding bond critical points.
Figure 7.8 presents the molecular graphs of NH+ +
4 · · ·HBeH and H2 OH · · ·HBeH
dihydrogen bonded complexes. One can see that there are not H· · ·BCP bond paths
where BCP mimics ¨ the point ¨ Lewis base center of the Be–H σ-bond for those
complexes. Generally the H· · ·BCP bond paths do not exist for the A–H· · ·H–B
dihydrogen bonds; also two bond paths linking the protonic hydrogen with the at-
tractors of the B–H bond (H and B attractors) do not exist for DHBs; at least, to my
knowledge, they have not been reported in literature so far.
According to earlier studies the bond paths observed for the systems being in
the energetic minima correspond to the preferable interactions [96, 97]. For the
dihydrogen bonds the H· · ·H bond paths are observed which correspond to the mainly
electrostatic in nature interactions between the protonic and hydridic hydrogen atoms.
One may conclude that the QTAIM results show that the dihydrogen bonds with
contacts between H-atoms of the opposite charges could not be classified as the A–
H· · ·σ interactions. However the question arises if the A–H· · ·σ hydrogen bonds exist
at all. It is discussed in the next section of this chapter.
The Quantum Theory of Atoms in Molecules (QTAIM) often cited here is one of
the most powerful methods to analyze different inter- and intramolecular interactions
[94, 95]. Eight QTAIM criteria of the existence of the hydrogen bond were proposed
early on by Koch and Popelier [98]. In numerous studies on the A–H· · ·B hydrogen
bonds the criterion of the existence of the H· · ·B bond path with the corresponding
bond critical point is the most often checked one among the other criteria. The
Fig. 7.9 The (BH3 NH3 )2

dimer, BH2 · · ·H–N bifurcated
dihydrogen bonds are
presented (broken lines)
properties of this BCP are also often analyzed in numerous studies. Koch and Popelier
have pointed out that the electron density at the H· · ·B BCP should lie within the
range (0.002, 0.04) au and also the range for the laplacian was proposed. The decrease
of the electron charge of the hydrogen atom as an effect of the A–H· · ·B hydrogen
bond formation and the simultaneous decrease of the hydrogen atom volume are the
other QTAIM criteria. The latter criteria are in line with the NBO description of the
mechanism of the hydrogen bond formation since the same changes of the H-atom
charge and volume were found to be the result of the rehybridization process [84, 86].
This is also important that the criteria of ranges of the values of the electron density
and its laplacian concern medium in strength and weak hydrogen bonds. Very strong
interactions possess numerous features typical for covalent bonds while very weak
interactions possess features of van der Waals interactions [72, 74]. This is why in
such cases the characteristics of interactions may lie outside of ranges proposed by
Koch and Popelier [98].
The QTAIM parameters often applied to characterize hydrogen bonds were also
analyzed by Popelier for the (BH3 NH3 )2 dimer and it was found that this dimer is
linked through three H· · ·H contacts for which the corresponding bond paths with
BCPs exist [99]. The latter contacts may be classified as dihydrogen bonds since nu-
merous criteria proposed earlier for hydrogen bonds are also fulfilled here. However it
seems that energetically more stable (BH3 NH3 )2 dimer is characterized by the equiv-
alent BH3 NH3 molecules related by inversion symmetry where monomers are linked
through four H· · ·H contacts corresponding to two bifurcated DHBs (Fig. 7.9). Such
a structure was found by Cramer and Gladfelter with the use of the MP2/cc-pVDZ
level [100] and later it was analyzed by Scheiner (MP2/aug-cc-pVDZ) [63].
The other characteristics of BCP, not only ρBCP , and its laplacian, ∇ 2 ρBCP , men-
tioned earlier here are also often applied to deepen the understanding of the nature
of interactions. These are the total electron energy density at BCP (HBCP ) and the
components of the latter value: the kinetic electron energy density (GBCP —always
positive value) and the potential electron energy density (VBCP —always negative
value). There are the following relations between the mentioned above values (Eq.
7.4 in atomic units) [95].
∇ 2 ρ BCP = 2GBCP + VBCP where, HBCP = GBCP + VBCP . (7.4)

176 S. J. Grabowski
0.08
QTAIM parameters(au)
0.06 ∇ 2 ρBCP
GBCP
0.04
0.02
0
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8
-0.02 H...H distance (Å)

HBCP
-0.04
-0.06
-0.08
VBCP
-0.1
-0.12
Fig. 7.10 The dependencies between the H· · ·H distance and the characteristics of the corresponding
BCP; ∇ 2 ρBCP (triangles), HBCP (full circles), VBCP (squares) and GBCP (open circles)
It is worth mentioning that the negative value of laplacian is attributed to the

covalent interactions while its positive value is characteristic for van der Waals inter-
actions, ionic interactions and hydrogen bonds [95]. However for strong hydrogen
bonds HBCP is negative and even in extreme cases of very strong hydrogen bonds the
laplacian of the electron density at the H· · ·B BCP is negative, similarly as in the case
of covalent bonds [72, 73, 101, 102]. The same concerns dihydrogen bonds [75]. Fig-
ure 7.10 presents dependencies between the H· · ·H distance and the mentioned above
characteristics of the H· · ·H BCP for different dihydrogen bonded systems, the results
for the sample containing 22 complexes [75] which was analyzed earlier here are pre-
sented in Fig. 7.10. One can see three types of interactions here. Those where ∇ 2 ρ BCP
is negative may be classified as very strong DHBs, those where ∇ 2 ρ BCP is positive
and HBCP is negative belong to the class of strong interactions and finally the DHBs
where both ∇ 2 ρ BCP and HBCP are positive belong to medium in strength or weak
interactions. There are the following examples of the dihydrogen bonded complexes
belonging to the mentioned above classes; Cl2 OH+ · · ·HBeH, H3 NH+ · · ·HBeH and
HCCH· · ·HLi which correspond to very strong (∇ 2 ρBCP < 0 and HBCP < 0) strong
(∇ 2 ρBCP > 0 and HBCP < 0) and weak (∇ 2 ρBCP > 0 and HBCP > 0) dihydrogen bonds,
respectively. The same classification based on the signs of the ∇ 2 ρBCP and HBCP
values was proposed early on for the hydrogen bonds by Rozas et al. [103].
7.6 The A–H· · ·σ Interactions: From Complexes to Clusters
It was described in the previous sections that the A–H· · ·H–B dihydrogen bond is
not classified as the A–H· · ·σ interaction since the σ-electrons of the B–H bond do
not mimic the one center Lewis base, at least not in the same way as it was observed
for π-electrons in the A–H· · ·π hydrogen bonds [104]. For the latter interactions
few cases may be considered [97]. One can observe the attractor—bond critical
point bond paths, where attractor corresponds to the H-atom of the A–H proton
donating bond while BCP mimics the Lewis base center. That situation is observed
for the two-center π-electron systems like for the CC bond in acetylene, ethylene
and similar molecules. For multicenter π-electron system like in benzene molecule
and other aromatic systems the situation is more complicated [97]. The H-attractor
of the A–H bond is connected with single atom or with the bond critical point if the
unsymmetrical distribution of the electron charge density for the Lewis base sub-unit
is observed. The H-attractor may be also connected with few atomic attractors of
the Lewis base. ([97], see Fig. 7.7). The case of the existence of attractor—attractor
bond path should not be classified as the A–H. . . π interaction. Such a situation
was observed for the C5 H− 5 · · ·HF complex where the sub-units are connected by
the single H· · ·C bond path and the interaction may be classified as the F–H· · ·C
hydrogen bond; the C-atom is the most negatively charged atom of the C5 H− 5 ring
and it possesses the properties typical of the Lewis base center [97]. For the A–
H· · ·H–B interactions presented so far here the H-atom attractor of the A–H bond is
not connected through the bond path with the B–H bond critical point but the H· · ·H
bond path is observed.
It seems that the molecular hydrogen is a good candidate as a Lewis base σ-
electron system to form links with Lewis acids classified as the A–H· · ·σ hydrogen
bonds. Numerous experimental and theoretical studies are known where H2 molecule
interacts as the Lewis base with metals or cations through the σ-electrons [26].
The similar interaction was observed early for the NH+ 4 · · ·H2 complex analyzed
theoretically (MP2/aug-cc-pVTZ level) [105]. It was found that the H2 molecule is
approximately perpendicular to the N–H proton donating bond and that the binding
energy for this complex is equal to 2.5 kcal/mol. The authors have also analyzed
the NH+ 4 . . . (H2 )n clusters (n up to 8) and they have found that if the number of
H2 molecules in the cluster increases from 1–4 thus the following N–H bonds are
saturated by the N–H· · ·H2 interactions where hydrogen molecules are perpendicular
to N–H bonds [105]. Figure 7.11 presents the structure of the NH+ 4 · · ·H2 complex
as well as of the NH+ +
4 . . . (H2 )2 and the NH4 . . . (H2 )3 clusters. If the number of
H2 molecules exceeds 4 thus the additional interactions are observed between the
nitrogen center of NH+ 4 and H2 molecules [105].
The N–H· · ·H2 T-shaped arrangements were not discussed as possible types of
hydrogen bond in the mentioned here study [105]. However such interactions were
analyzed later since the more extended ab initio studies have been performed on
XH+ +
4 · · ·H2 complexes and XH4 . . . (H2 )5 clusters (X = N, P, As, Sb, Bi) where the
MP2/aug-cc-pVTZ calculations were carried out [106]. The N–H· · ·H2 interactions
178 S. J. Grabowski
Fig. 7.11 The structure of the NH+ + +

4 · · ·H2 complex and of the NH4 . . . (H2 )2 and NH4 . . . (H2 )3
clusters
were classified as A–H· · ·σ hydrogen bonds where σ-electrons of the molecular

hydrogen play the role of the proton acceptor. The molecular graphs show here
the H· · ·BCP bond paths where BCP corresponds to the σ-bond of the hydrogen
molecule. Figure 7.12 presents the molecular graphs of the NH+ 4 · · ·H2 complex and
of the NH+ 4 . . . (H2 )4 cluster.
The decomposition of the energy of interaction performed within the variation-
perturbation approach [67, 68] for the XH+ 4 · · ·H2 systems shows that the delocal-
ization term is the most important attractive one followed by the electrostatic and
dispersion contributions [106]. These findings are similar to those observed for the
A–H· · ·π hydrogen bonds. It was discussed earlier that for the latter interactions the
BCP of the π-electron system mimics the Lewis base center [97, 104]. This is ob-
served here for A–H· · ·σ hydrogen bonds where the BCP of H–H bond corresponds
to the Lewis base center. There are other similarities between A–H· · ·π and A–H· · ·σ
interactions. For example, the decomposition of the energy of interaction was per-
formed for numerous A–H· · ·π hydrogen bonded complexes and it was found that for
such interactions the delocalization term is an important attractive contribution [107].
The similar A–H· · ·σ interaction was detected for the T-shaped complex of F–
H· · ·H2 where the calculations performed at the MP2/6-311++G(3df,3pd) level
show the binding energy of 0.8 kcal/mol [108]. Figure 7.13 presents the molecular
Fig. 7.12 The molecular graph of the NH+ +

4 · · ·H2 complex (left) and of the NH4 . . . (H2 )4 cluster
(right); big circles correspond to attractors, small ones to the bond critical points, solid and broken
lines represent the bond paths

graph of the FH· · ·H2
T-shaped complex with the
isolines of Laplacian of the
electron density; positive
values are depicted in solid
lines and negative values in
broken lines (the laplacian
isodensity lines in the plane
of the complex)
graph of the complex with the H· · ·BCP bond path corresponding to the H· · ·σ
interaction. The laplacian of the electron density isolines in the plane of the complex
are presented indicating the regions of the concentration of electron density for the
negative laplacian values. The analysis of results of calculations led to the conclusion
that the F–H· · ·σ(H2 ) interaction possesses properties typical for the hydrogen bond
[108]. The following properties were observed there; for the HF and H2 molecules
in the complex there is the elongation of both H–F and H–H σ-bonds by ∼ 0.002 Å
as a result of the hydrogen bond formation. Both elongations are connected with the
electron charge transfer from the H2 molecule acting as the Lewis base to the HF
Lewis acid unit. It corresponds to the σ → σ∗ orbital-orbital interaction. The loss of
the electron charge of the H2 molecule results in the decrease of the occupation of the
σ(H–H) orbital and the weakening of the H–H bond while the transfer of the electron
charge to the HF molecule results in the increase of the occupation of σ∗ (H–F) orbital
and consequently the weakening and of H–F bond.
The other complexes linked through the A–H· · ·σ hydrogen bonds were also
analyzed and they were compared with the analogues complexes with A–H· · ·π
180 S. J. Grabowski
interactions; the complexes of acetylene and molecular hydrogen were compared.

For all complexes the C2 H2 and H2 molecules play the role of the proton acceptor,
i.e. of the Lewis base. It was described in this chapter that the characteristics of the
A–H· · ·σ interaction are different than those of the A–H· · ·H–B dihydrogen bond.
It is also worth to notify that usually there is the strong polarization of the B–H
bond in the A–H· · ·H–B system what is not observed for the molecular hydrogen
acting as the Lewis base in the A–H· · ·σ interaction. Such a strong polarization was
observed for the B–H bond (B designates boron here) in the N–H· · ·H–B contact
of the ammonia-borane dimer [37]. It seems that the mentioned above polarization
leads to the excess of the negative charge on the H(B)-atom and the directional
H−δ . . .+δ H interaction (dihydrogen bond) being mostly electrostatic in nature. In a
case of molecular hydrogen acting as the Lewis base the sites of H-atoms are char-
acterized by the positive electrostatic potential while in the direction perpendicular
(and nearly so) to the σ-bond the negative electrostatic potential is observed [110].
The atomic charges of the H2 molecule in the complex are slightly positive because
of the electron charge transfer to the Lewis acid and such charges are equal to each
other, or nearly so [32]. This is why for the A–H· · ·H2 interactions, if A–H is the
typical proton donating bond with the excess of the positive charge on the H-atom
the H· · ·BCP intermolecular bond paths are observed.
The hydrogen molecule is a very weak Lewis base, however sometimes strong
or at least medium in strength A–H· · ·σ interactions are observed [109]. For exam-
ple, for the H2 OH+ · · ·H2 complex (Fig. 7.14) the binding energy is equal to 5.2
kcal/mol (MP2/6-311++G(3df,3pd) level) [109], comparable with the binding en-
ergy for the water dimer linked through the typical O–H· · ·O hydrogen bond [3]. For
the H2 OH+ · · ·H2 complex there is also the meaningful electron charge transfer from
molecular hydrogen to the hydronium cation, 139 milielectrons, while for the water
dimer such a transfer from one water molecule to the other one amounts ∼ 20 mili-
electrons depending on the level of calculation. For the H2 OH+ · · ·H2 complex the
total electron energy density at the H· · ·σ(H2 ) BCP, HBCP , is negative. Note that the
negative HBCP value concerns the BCP situated at the bond path linking the H-atom
attractor of H3 O+ with the BCP of the H–H bond (Fig. 7.14). The latter means that
according to QTAIM the O–H· · ·σ interaction may be classified here as a strong one
and that such interaction possesses some properties of the covalent bond [83]. The
second-order perturbation energy of the σ → σ∗ orbital-orbital interaction (see Eq.
7.3) is equal to 16.5 kcal/mol for the H2 OH+ · · ·H2 complex. The analogues n(O) →
σ(OH)∗ orbital-orbital interaction energy for the water dimer, for the interaction typ-
ical for the conventional hydrogen bonds (see. Eq. 7.2), is equal to 6.5–7.0 kcal/mol
depending on the basis set applied.
This is worth to mention that the A–H· · ·σ interactions were firstly classified as the
hydrogen bonds for the XH+ +
4 · · ·H2 complexes and XH4 . . . (H2 )5 (X = N, P, As, Sb,
Bi) clusters [106]; next for the FH· · ·H2 complex [108] and next such interactions
were detected in the other complexes [109]. Even the A–H· · ·σ interactions are
classified as hydrogen bonds in new IUPAC definition of hydrogen bond [111].
The A–H· · ·σ interactions were also analyzed experimentally. For example, gas
phase measurements of dipole moment and vibrational predissociation lifetimes were

graph of the H2 OH+ · · ·H2
complex; the laplacian
electron isodensity lines in the
plane containing the O–H· · ·σ
interaction (the O–H· · ·H2
fragment) are also presented
performed for the F–H· · ·H2 complex [112] and the determination of rotational
constants confirmed its T-shaped structure. The high-resolution infrared spectra
were analyzed for the H2 –HF, D2 –HF and HD–HF systems solvated in helium
nanodroplets and also here the existence of the T-shaped structure was confirmed
[113–115].
The infrared vibrational predissociation spectra measurements and theoretical
calculations were performed for the H2 –HCO+ complex linked through C–H· · ·σ
interaction, where the σ-electrons are those of the molecular hydrogen while C–H
proton donating bond is of the HCO+ ion [116]. The authors explain that the most ex-
tensive vibrational bands arise from excitation of the C–H and H2 stretch vibrations
exhibiting rotational structure composed of sub-bands expected for the T-shaped
minimum energy structure. The QCISD(T)/6-311G(2df,2pd) calculations show that
the energy of the T-shaped conformer of the H2 -HCO+ complex is lower than the
energy of the separated monomers by 4.1 kcal/mol [116]. The authors explain that
the experimental spectra of H2 -HCO+ provide information on two intramolecular
vibrations; the H–H stretch localized on H2 (4060 cm−1 ) and the C–H stretch lo-
calized on HCO+ (2840 cm−1 ); the corresponding vibrations of isolated molecules
are equal to 4161 and 3089 cm−1 , respectively. This is worth to mention that the
decrease of the C–H stretching frequency connected with the elongation of the bond
is typical for the red-shift hydrogen bond. The authors explain in their study [116]
that the H–H stretch decrease is connected with the transfer of the electron charge
density from H2 molecule into HCO+ ion—this is typical for the hydrogen bond
[83] and in general for the Lewis acid—Lewis base interactions [117] where the
electron charge transfer from the Lewis base to the Lewis acid is observed. Thus the
experimental results are in line with the theoretical findings since they show that the
A–H· · ·σ interactions possess numerous characteristics typical for hydrogen bonds.
One can also see that the A–H· · ·σ and A–H· · ·H–B (dihydrogen bonds) interactions
are different sub-classes of the hydrogen bond.
182 S. J. Grabowski
7.7 The Dihydrogen Bonds and the A–H· · ·σ Interactions

as Types of the Hydrogen Bonds—Summary
The A–H· · ·H–B dihydrogen bond and the A–H· · ·σ interaction are classified as types
of the hydrogen bond. The protonic H-atom of the A–H bond is characterized by the
excess of the positive charge and there is the electron charge shift from the Lewis
base unit to the Lewis acid for both interactions. The latter phenomenon is common
for all Lewis acid—Lewis base interactions.
The energy of the σ → σ* orbital-orbital overlap is the most important part of
the charge transfer energy for the A–H· · ·H–B dihydrogen bond as well as for the
A–H· · ·σ interaction. This may suggest that both interactions belong to the same
sub-class of the hydrogen bond. However one can observe numerous differences
between them. For the dihydrogen bond the hydridic H-atom is the proton acceptor,
i.e. the Lewis base center, while for the A–H· · ·σ interaction the σ-electrons of the
molecular hydrogen play the electron donating role. In other words for DHBs there
is the öne-atom¨ Lewis base center (H−δ ) while for the A–H· · ·σ interactions there
is ¨ two atoms center ¨ of molecular hydrogen. In the latter case the bond critical
point corresponding to the H–H bond mimics the one center Lewis base.
For the dihydrogen bonded systems there is the H· · ·H bond path between hydro-
gen atoms characterized by opposite charges. This is worth to mention that the only
A–H· · ·σ interactions known so far are those where the σ-electrons of the molecular
hydrogen play the role of electron donors. For such interactions both H-atoms of H2
molecule are slightly positively charged within the complex as the result of the Lewis
acid—Lewis base electron charge shift. For the A–H· · ·H–B dihydrogen bonds the
polarization of the B–H bond is observed where H-atom is negatively charged.
The weak and medium in strength dihydrogen bonds are usually electrostatic in
nature and the significance of the interaction energy terms related to the electron
charge shift increases for stronger interactions. For the A–H· · ·σ interactions the
dominance of the electrostatic term is not observed, the delocalization interaction
energy term is often the most important attractive contribution.
One can also observe few analogues between A–H· · ·σ and A–H· · ·π interactions
on one hand and the analogues between A–H· · ·B conventional hydrogen bond and
the A–H· · ·H–B dihydrogen bond on the other hand. For the first pair of interactions
there is no one-atom Lewis base center and electrons play the role of the proton
acceptor, the Lewis base centers are simulated by the bond critical points. However
the situation is more diverse for the A–H· · ·π interactions. For the second pair, the
conventional hydrogen bonds and the dihydrogen bonds the one-atom negatively
charged Lewis base center is observed.
Acknowledgments Financial support comes from Eusko Jaurlaritza (GIC IT-588-13) and
the Spanish Office for Scientific Research (CTQ 2012-38496-C05-04). Technical and human
support provided by Informatikako Zerbitzu Orokora—Servicio General de Informatica de la
Universidad del Pais Vasco (SGI/IZO-SGIker UPV/EHU), Ministerio de Ciencia e Innovación
(MICINN), Gobierno Vasco Eusko Jaurlanitza (GV/EJ), European Social Fund (ESF) is gratefully
acknowledged.
References
1. Richardson TB, de Gala S, Crabtree RH (1995) Unconventional hydrogen bonds: intermolec-

ular B–H· · ·H–N interactions. J Am Chem Soc 117:12875–12876
2. Negative values represent binding and interaction energies for stable complexes. For the
convenience of discussion the absolute positive values are given in the text
3. Scheiner S (1994) Ab initio studies of hydrogen bonds: the water dimer paradigm. Annu Rev
Phys Chem 45:23–56
4. Wessel J, Lee JC Jr, Peris E, Yap GPA, Fortin JB, Ricci JS, Sini G, Albinati A, Koetzle TF,
Eisenstein O, Rheingold AL, Crabtree RH (1995) An unconventional intermolecular three-
center N–H· · ·H2 Re hydrogen bond in crystalline [ReH5 (PPh3 )3 ]-indole-C6 H6 . Angew Chem
Int Ed Engl 34:2507–2509
5. Crabtree RH, Siegbahn PEM, Eisenstein O, Rheingold AL, Koetzle TFA (1996) A new inter-
molecular interaction: unconventional hydrogen bonds with element-hydride bonds as proton
acceptor. Acc Chem Res 29:348–354
6. Crabtree RH, Eisenstein O, Sini G, Peris E (1998) New types of hydrogen bonds. J Organomet
Chem 567:7–11
7. Custelcean R, Jackson JE (2001) Dihydrogen bonding: structures, energetics, and dynamics.
Chem Rev 101:1963–1980
8. Robertson KN, Knop O, Cameron TS (2003) C–H· · ·H–C interactions in organoammonium
tetraphenylborates: another look at dihydrogen bonds. Can J Chem 81:727–743
9. Wolstenholme DJ, Cameron TS (2006) Comparative study of weak interactions in molecular
crystals: H–H bonds vs hydrogen bonds. J Phys Chem A 110:8970–8978
10. Alkorta I, Elguero J, Foces-Foces C (1996) Dihydrogen bonds (A–H· · ·H–B). Chem Commun
1633–1634
11. Remko M (1998) Thermodynamics of dihydrogen bonds (A–H· · ·H–B). Mol Phys 94:839–
842
12. Orlova G, Scheiner S (1998) Intermolecular MH· · ·HR bonding in monohydride Mo and W
complexes. J Phys Chem A 102:260–269
13. Orlova G, Scheiner S (1998) Intermolecular H· · ·H bonding and proton transfer in semisand-
wich Re and Ru complexes. J Phys Chem A 102:4813–4818
14. Orlova G, Scheiner S, Kar T (1999) Activation and cleavage of H–R bonds through inter-
molecular H· · ·H bonding upon reaction of proton donors HR with 18-electron transition
metal hydrides. J Phys Chem A 103:514–520
15. Kulkarni SA (1998) Dihydrogen bonding in main group elements: an ab initio study. J Phys
Chem A 102:7704–7711
16. Kulkarni SA, Srivastava AK (1999) Dihydrogen bonding in main group elements: a case study
of complexes of LiH, BH3 , and AlH3 with third-row hydrides. J Phys Chem A 103:2836–2842
17. Custelcean R, Jackson JE (1998) Topochemical control of covalent bond formation by
dihydrogen bonding. J Am Chem Soc 120:12935–12941
18. Matus MH, Anderson KD, Camaioni DM, Autrey ST, Dixon DA (2007) Reliable predictions
of the thermochemistry of boron-nitrogen hydrogen storage compounds: BxNxHy, x = 2, 3.
J Phys Chem A 111:4411–4421
19. Miranda CR, Ceder G (2007) Ab initio investigation of ammonia-borane complexes for
hydrogen storage. J Chem Phys 126:184703
20. Keaton RJ, Blacquiere JM, Baker RT (2007) Base metal catalyzed dehydrogenation of
ammonia–borane for chemical hydrogen storage. J Am Chem Soc 129:1844–1845
21. Staubitz A, Besora M, Harvey JN, Manners I (2008) Computational analysis of amine-borane
adducts as potential hydrogen storage materials with reversible hydrogen uptake. Inorg Chem
47:5910–5918
22. Custelcean R, Jackson JE (2000) Topochemical dihydrogen to covalent bonding transforma-
tion in LiBH4 · TEA: a mechanistic study. J Am Chem Soc 122:5251–5257
23. Custelcean R, Vlassa M, Jackson JE (2000) Toward crystalline covalent solids: crystal-to-
crystal dihydrogen to covalent bonding transformation in NaBH4 · THEC. Angew Chem Int
Ed Engl 39:3299–3302
184 S. J. Grabowski
24. Kenward AL, Piers WE (2008) Heterolytic H2 activation by nonmetals. Angew Chem Int Ed
Engl 47:38–41
25. Filippov OA, Filin AM, Tsupreva VN, Belkova NV, Lledós A, Ujaque G, Epstein LM, Shubina
ES (2006) Proton-transfer and H2 -elimination reactions of main-group hydrides EH4 —(E =
B, Al, Ga) with alcohols. Inorg Chem 45:3086–3096.
26. Kubas GJ (2001) Metal Dihydrogen and σ-bond complexes—structure, theory, and reactivity.
Kluwer Academic/Plenum Publishers: New York
27. Crabtree RH (2005) The organometallic chemistry of the transition metals. Wiley, Hoboken
28. Alcaraz G, Grellier M, Sabo-Etienne S (2009) Bis σ -bond dihydrogen and borane ruthenium
complexes: bonding nature, catalytic applications, and reversible hydrogen release. Acc Chem
Res 42:1640–1649
29. Epstein LM, Shubina ES (2002) New types of hydrogen bonding in organometallic chemistry.
Coord Chem Rev 231:165–181
30. Belkova NV, Shubina ES, Epstein LM (2005) Diverse world of unconventional hydrogen
bonds. Acc Chem Res 38:624–631
31. de Oliveira BG (2013) Structure, energy, vibrational spectrum, and Bader’s analysis of π· · ·H
hydrogen bonds and H−δ · · ·H+δ dihydrogen bonds. Phys Chem Chem Phys 15:37–79
32. Grabowski SJ (2013) Dihydrogen bond and X–H· · ·σ interaction as sub-classes of hydrogen
bond. J Phys Org Chem 26:452–459
33. Epstein LM, Belkova NV, Shubina ES (2001) Dihydrogen bonded complexes and proton
transfer to hydride ligands by spectral (IR, NMR) studies. In: Peruzzini M, Poli R (eds)
Recent advances in hydride chemistry (Chapter). Elsevier, Amsterdam pp 391–418
34. Grabowski SJ, Leszczynski J (2005) Is a dihydrogen bond a unique phenomenon? Chapter
in vol 9, a book series: computational chemistry: reviews of current trends. World Scientific
Publishing Co: Singapore, pp 195–235
35. Grabowski SJ, Leszczynski J (2009) Dihydrogen bonds: novel feature of hydrogen bond
interactions. In: Leszczynski J, Shukla M (eds) Practical aspects of computational chemistry,
methods, concepts and applications (Chapter). Springer: Heidelberg Dordrecht, London, New
York
36. Bakhmutov VI (2008) Dihydrogen bonds, principles, experiments, and applications. John
Wiley & Sons, Inc.: Hoboken, New Jersey
37. Klooster WT, Koetzle TF, Siegbahn PEM, Richardson TB, Crabtree RH (1999) Study of the
N–H· · ·H–B Dihydrogen bond including the crystal structure of bh3 nh3 by neutron diffraction.
J Am Chem Soc 121:6337–6343
38. Zachariasen WH, Mooney RCL (1934) The structure of hypophosphite group as determined
from the crystal lattice of ammonium hypophosphite. J Chem Phys 2:34–37
39. Burg AB (1964) Enhancement of P–H bonding in a phosphine monoborane. Inorg Chem
3:1325–1327
40. Titov LV, Makarova MD, Rosolovskii VY (1968) Guanidinium borohydride. Dokl Akad Nauk
180:381–382
41. Grabowski SJ, Krygowski TM (1999) The proton transfer path for C=O· · ·H–O systems
modelled from crystal structure data. Chem Phys Lett 305:247–250
42. Sobczyk L, Grabowski SJ, Krygowski TM (2005) Interrelation between H-bond and Pi-
electron delocalization. Chem Rev 105:3513–3560
43. Grabowski SJ (2014) Tetrel bond—σ-hole bond as a preliminary stage of the SN reaction.
44. Brown MP, Heseltine RW (1968) Co-ordinated BH3 , as a proton acceptor group in hydrogen
bonding. Chem Commun 23:1551–1552
45. Brown MP, Heseltine RW, Smith PA, Walker PJ (1970) An infrared study of coordinated BH3
and BH2 groups as proton acceptors in hydrogen bonding. J Chem Soc A 410–414
46. Stevens RC, Bau R, Milstein D, Blum O, Koetzle TF (1990) Concept of the H(δ+)· · ·H(δ−)
interaction. A low-temperature neutron diffraction study of cis- [IrH (OH)(PMe3 )4 ] PF6 . J
Chem Soc Dalton Trans 1429–1432
47. Milstein D, Calabrese JC, Williams ID (1986) Formation, structures, and reactivity of cis-
hydroxy-, cis-methoxy-, and cis-mercaptoiridium hydrides. Oxidative addition of water to
Ir(I). J Am Chem Soc 108:6387–6389
48. Lough AJ, Park S, Ramachandran R, Morris RH (1994) Switching on and off a new
intramolecular hydrogen-hydrogen interaction and the heterolytic splitting of dihydrogen.
Crystal and molecular structure of [Ir({H(ηl -SC5 H4 NH) } 2(PCy3 )2 ]BF4 .2.7CH2 C12 . J Am
Chem Soc 116:8356–8357
49. Gusev DG, Lough AJ, Morris RH (1998) New polyhydride anions and proton-hydride
hydrogen bonding in their ion pairs. J Am Chem Soc 120:13138–13147
50. Ramachandran R, Morris RH (1994) A new type of intramolecular H· · ·H· · ·H interac-
tion involving N–H· · ·H(Ir)· · ·H–N atoms. Crystal and molecular structure of [IrH(ηl -
SC5 H4 NH)2 (η2 -SC5 H4 N) (PCy3 )]BF4 .0.72CH2 C12 . J Chem Soc Chem Commun 2201–2202
51. Lee JC, Pens E, Rheingold AL, Crabtree RH (1994) An unusual type of H· · ·H interaction:
Ir–H· · ·H–O and Ir–H· · ·H–N hydrogen bonding and its involvement in σ-bond metathesis. J
Am Chem Soc 116:11014–11019
52. Lee JC, Rheingold AL, Muller B, Pregosin PS, Crabtree RH (1994) Complexation of an
amide to iridium via an iminol tautomer and evidence Ir–H· · ·H–O hydrogen bond. J Chem
Soc Chem Commun 1021–1022
53. Belkova NV, Shubina ES, Gutsul EI, Epstein LM, Eremenko IL, Nefedov SE (2000) Structural
and energetic aspects of hydrogen bonding and proton transfer to ReH2 (CO)(NO)(PR3 )2 and
ReHCl(CO)(NO)(PMe3 )2 by IR and X-ray studies. J Organomet Chem 610:58–70
54. Allen FH (2002) The Cambridge structural database: a quarter of a million crystal structures
and rising. Acta Cryst B58:380–388
55. Liu Q, Hoffmann R (1995) Theoretical aspects of a novel mode of hydrogen-hydrogen
bonding. J Am Chem Soc 117:10108–10112
56. Siegbahn PEM, Blomberg MRA, Svensson M (1994) PCI-X, a parametrized correlation
method containing a single adjustable parameter X. Chem Phys Lett 223:35–45
57. Siegbahn PEM, Svensson M, Boussard PJE (1995) First row bench mark tests of the PCI-X
scheme. J Chem Phys 102:5377–5386
58. Grabowski SJ (1999) Study of correlations for dihydrogen bonds by quantum-chemical
calculations. Chem Phys Lett 312:542–547
59. Grabowski SJ (2000) High-level ab initio calculations of dihydrogen-bonded complexes. J
Phys Chem A 104:5551–5557
60. Kitaura K, Morokuma K (1976) A new energy decomposition scheme for molecular
interactions within the Hartree-Fock approximation. Int J Quantum Chem 10:325–340
61. Cybulski SM, Chałasiński G, Moszyński R (1990) On decomposition of MP2 supermolecular
interaction energy and basis set effects. J Chem Phys 92:4357–4363
62. Cybulski H, Pecul M, Sadlej J (2003) Characterization of dihydrogen-bonded D–H· · ·H–A
complexes on the basis of infrared and magnetic resonance spectroscopic parameters. J Chem
Phys 119:5094–5104
63. Kar T, Scheiner S (2003) Comparison between hydrogen and dihydrogen bonds among
H3 BNH3 , H2 BNH2 , and NH3 . J Chem Phys 119:1473–1482
64. Alkorta I, Elguero J, Mó O, Yánez M, Del Bene JE (2002) Ab Initio study of the structural,
energetic, bonding, and IR spectroscopic properties of complexes with dihydrogen bonds. J
Phys Chem A 106:9325–9330
65. Del Bene JE, Perera SA, Bartlett RJ, Alkorta I, Elguero J, Mó O, Yánez M (2002) One-
bond (1d JH−H ) and three-bond (3d JX−M ) spin-spin coupling constants across X-H· · ·H–M
dihydrogen bonds. J Phys Chem A 106:9331–9337
66. Grabowski SJ, Robinson TL, Leszczynski J (2004) Strong dihydrogen bonds—ab initio and
atoms in molecules study. Chem Phys Lett 386:44–48
67. Sokalski WA, Roszak S, Pecul K (1988) An efficient procedure for decomposition of the
SCF interaction energy into components with reduced basis set dependence. Chem Phys Lett
153:153–159
186 S. J. Grabowski
68. Sokalski WA, Roszak S (1991) Efficient techniques for the decomposition of intermolecular
interaction energy at SCF level and beyond. J Mol Struct (Theochem) 234:387–400
69. Grabowski SJ, Sokalski WA, Leszczynski J (2005) How short can the H· · ·H intermolecular
contact be? New findings that reveal the covalent nature of extremely strong interactions. J
Phys Chem A 109:4331–4341
of separate total energies. Some procedures with reduced errors. Mol Phys 19:553–566
71. Schmidt MS, Baldridge KK, Boatz JA, Elbert ST, Gordon MS, Jensen JH, Koseki S, Matsunaga
N, Nguyen KA, Su SJ, Windus TL, Dupuis M, Montgomery JA (1993) General atomic and
molecular electronic structure system. J Comp Chem 14:1347–1363
72. Grabowski SJ, Sokalski WA, Dyguda E, Leszczynski J (2006) Quantitative classification of
covalent and noncovalent H-bonds. J Phys Chem B 110:6444–6446
73. Grabowski SJ (2009) Covalent character of hydrogen bonds enhanced by π-electron
delocalization. Croat Chem Acta 82:185–192
74. Desiraju GR (2002) Hydrogen bridges in crystal engineering: interactions without borders.
Acc Chem Res 35:565–573
75. Grabowski SJ, Sokalski WA, Leszczynski J (2007) Wide spectrum of H· · ·H interactions: van
der Waals contacts, dihydrogen bonds and covalency. Chem Phys 337:68–76
76. Jeffrey GA, Saenger W (1991) Hydrogen bonding in biological structures. Springer-Verlag:
Berlin
77. Scheiner S (1997) Hydrogen bonding: a theoretical perspective. Oxford University Press:
New York
79. Feng Y, Zhao S-W, Liu L, Wang J-T, Li X-S, Guo Q-X (2004) Blue-shifted dihydrogen bonds.
J Phys Org Chem 17:1099–1106
80. Yang Y, Zhang W (2007) Theoretical study of N–H· · ·H–B blue-shifted dihydrogen bonds. J
Mol Struct (Theochem) 814:113–117
81. Yu W, Lin Z, Huang Z (2006) Coexistence of dihydrogen, blue and red-shifting hydrogen
bonds in an ultrasmall system: valine. ChemPhysChem 7:828–830
82. Trung NT, Hue TT, Nguyen MT, Zeegers-Huyskens T (2008) Theoretical study of the inter-
action between HNZ (Z = O, S) and H2 XNH2 (X = B, Al). Conventional and dihydrogen
bonds. Phys Chem Chem Phys 10:5105–5113
83. Grabowski SJ (2011) What is the covalency of hydrogen bonding? Chem Rev 11:2597–2625
84. Weinhold F, Landis C (2005) Valency and bonding, a natural bond orbital donor—acceptor
perspective. Cambridge University Press: Cambridge, UK
86. Alabugin IV, Manoharan M, Peabody S, Weinhold F (2003) Electronic basis of improper
hydrogen bonding: a subtle balance of hyperconjugation and rehybridization. J Am Chem
Soc 125:5973–5987
87. Weinhold F, Klein R (2012) What is a hydrogen bond? Mutually consistent theoretical and
experimental criteria for characterizing H-bonding interactions. Mol Phys 110:565–579
88. Alkorta I, Elguero J, Grabowski SJ (2008) How to determine whether intramolecular H· · ·H
interactions can be classified as dihydrogen bonds. J Phys Chem A 112:2721–2727
89. Grabowski SJ (2011) Halogen bond and its counterparts: Bent’s rule explains the formation
of nonbonding interactions. J Phys Chem A 115:12340–12347
90. Grabowski SJ (2012) QTAIM characteristics of halogen bond and related interactions. J Phys
Chem A 116:1838–1845
91. Grabowski SJ (2013) Non-covalent interactions—QTAIM and NBO analysis. J Mol Model
19:4713–4721
92. Bent HA (1961) An appraisal of valence-bond structures and hybridization in compounds of
the first-row elements. Chem Rev 61:275–311
93. Grabowski SJ (2011) Red- and blue-shifted hydrogen bonds: the Bent rule from quantum
theory of atoms in molecules perspective. J Phys Chem A 115:12789–12799
94. Bader RFW (1985) Atoms in molecules. Acc Chem Res 18:9–15
95. Bader RFW (1990) Atoms in molecules, a quantum theory. Oxford University Press, Oxford
96. Bader RFW (2009) Bond paths are not chemical bonds. J Phys Chem A 113:10391–10396
97. Grabowski SJ, Ugalde JM (2010) Bond paths show preferable interactions: ab initio and
QTAIM studies on the X–H· · · π hydrogen bond. J Phys Chem A 114:7223–7229
98. Koch U, Popelier PLA (1995) Characterization of C–H–O hydrogen bonds on the basis of
the charge density. J Phys Chem 99:9747–9754
99. Popelier PLA (1998) Characterization of a dihydrogen bond on the basis of the electron
density. J Phys Chem A 102:1873–1878
100. Cramer CJ, Gladfelter WL (1997) Ab initio characterization of [H3 N.BH3 ]2 , [ H3 N.A1H3 ]2 ,
and [H3 N.GaH3 ]2 . Inorg Chem 36:5358–5362
101. Cremer D, Kraka E (1984) A description of the chemical-bond in terms of local properties of
electrodensity and energy. Croat Chem Acta 57:1259–1281
102. Jenkins S, Morrison I (2000) The chemical character of the intermolecular bonds of seven
phases of ice as revealed by ab initio calculation of electron densities. Chem Phys Lett
317:97–102
103. Rozas I, Alkorta I, Elguero J (2000) Behavior of ylides containing N, O, and C atoms as
hydrogen bond acceptors. J Am Chem Soc 122:1154–11161
104. Domagała M, Grabowski SJ (2009) X–H· · ·π and X–H· · ·N hydrogen bonds—acetylene and
hydrogen cyanide as proton acceptors. Chem Phys 363:42–48
105. Urban J, Roszak S, Leszczynski J (2001) Shellvation of the ammonium cation by molecular
hydrogen: a theoretical study. Chem Phys Lett 346:512–518
106. Szymczak JJ, Grabowski SJ, Roszak S, Leszczynski J (2004) H· · ·σ interactions—an ab initio
and “atoms in molecules” study. Chem Phys Lett 393:81–86
107. Grabowski SJ, Lipkowski P (2011) Characteristics of X–H· · ·π interactions: ab initio and
QTAIM studies. J Phys Chem A 115:4765–4773.
108. Grabowski SJ, Sokalski WA, Leszczynski J (2006) Can H· · ·σ, π· · ·H+· · ·σ and σ· · ·H+· · ·σ
interactions be classified as H-bonded? Chem Phys Lett 432:33–39
109. Grabowski SJ (2007) Hydrogen bonds with π and σ electrons as the multicenter proton
acceptors: high level ab initio calculations. J Phys Chem A 111:3387–3393
110. Grabowski SJ, Alkorta I, Elguero J (2013) Complexes between dihydrogen and amine, phos-
phine, and arsine derivatives. Hydrogen bond versus pnictogen interaction. J Phys Chem A
117:3243–3251
111. Arunan E, Desiraju GR, Klein RA, Sadlej J, Scheiner S, Alkorta I, Clary DC, Crabtree
RH, Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ (2011)
Definition of the hydrogen bond. Pure Appl Chem 83:1637–1641
112. Jucks KW, Miller RE (1987) Infrared Stark spectroscopy on the hydrogen-HF binary complex.
J Chem Phys 87:5629–5633
113. Moore DT, Miller RE (2003) Dynamics of hydrogen–HF complexes in helium nanodroplets.
J Chem Phys 118:9629–9636
114. Moore DT, Miller RE (2003) Solvation of HF by molecular hydrogen: helium nanodroplet
vibrational spectroscopy. J Phys Chem A 107:10805–10812
115. Moore DT, Miller RE (2004) Rotationally resolved infrared laser spectroscopy of (H2 )n –HF
and (D2 )n –HF (n = 2–6) in helium nanodroplets. J Phys Chem A 108:1930–1937
116. Bieske EJ, Nizkorodov SA, Bennett FR, Maier JP (1995) The infrared spectrum of the H2–
HCO1 complex. J Chem Phys 102:5152–5164
117. Lipkowski P, Grabowski SJ, Leszczynski J (2006) Properties of the halogen-hydride
interaction: an ab initio and “Atoms in Molecules” analysis. J Phys Chem A 110:10296–10302
Part II
Other Bridging Atoms
Chapter 8
The Pnicogen Bond in Review: Structures,
Binding Energies, Bonding Properties, and
Spin-Spin Coupling Constants of Complexes
Stabilized by Pnicogen Bonds
Janet E. Del Bene, Ibon Alkorta and José Elguero
Abstract Extensive ab initio MP2/aug’-cc-pVTZ studies have been carried out in our
laboratories to determine the structures, binding energies, bonding properties, and
EOM-CCSD spin-spin coupling constants of various series of complexes stabilized
by pnicogen bonds. These systematic studies provide insight into the nature of the
pnicogen bond, and how changes in this bond are reflected in the properties of these
complexes.
8.1 Introduction
The term “pnicogen” or “pnictogen” was originally proposed by Prof. A. E. van

Arkel to designate the group 15 elements nitrogen, phosphorus, arsenic, antimony,
and bismuth [1, 2]. The name comes from the Greek word meaning choke or choking
gas. The symbol Z is used to designate this family of elements, just as the symbols Y
and X refer to the group 16 chalcogens and group 17 halogens, respectively [3]. The
pnicogen bond which involves a group 15 element acting as an electron acceptor,
joins the list of other weak intermolecular interactions such as the hydrogen bond and
the halogen bond. The IUPAC definition of a hydrogen bond (HB) is “an attractive
interaction between a hydrogen atom from a molecule or a molecular fragment X–H
in which X is more electronegative than H, and an atom or a group of atoms in the
same or a different molecule, in which there is evidence of bond formation” [4]. The
hydrogen bond is often written as X-H...Y, with Y either a σ electron-pair donor or
a π electron donor.
J. E. Del Bene ()

Department of Chemistry, Youngstown State University,
44555 Youngstown, OH, USA
e-mail: jedelbene@ysu.edu
I. Alkorta () · J. Elguero
Instituto de Química Médica (IQM-CSIC),
Juan de la Cierva, 3, 28006 Madrid, Spain
e-mail: ibon@iqm.csic.es

192 J. E. Del Bene et al.
The IUPAC definition of a halogen bond (XB) is as follows: “A halogen bond

occurs when there is evidence of a net attractive interaction between an electrophilic
region associated with a halogen atom in a molecular entity and a nucleophilic region
in another, or the same, molecular entity” [5]. Politzer and Murray have described
the halogen atom using the concept of the σ-hole which corresponds to a positive
region on the molecular electrostatic potential (MEP) surface [6]. Our definition of
the pnicogen bond (ZB) states that this bond is a Lewis acid-Lewis base attractive
interaction in which the pnicogen atom is the Lewis acid. While all three bonds are
necessarily attractive interactions, there are similarities and differences among them.
In the hydrogen bond, the hydrogen atom is positively charged since it is bonded to a
more electronegative atom or group of atoms, and it acts as an electron acceptor from
an atom or group of atoms, as for example in the F-H...NH3 complex. The electron-
donating partner may donate a lone pair of electrons or π electrons. Similarly, the
halogen bond may have the halogen atom bonded to a more electronegative atom or
group of atoms as it acts as an electron acceptor through its σ-hole, as illustrated by
the complex F-Cl...NH3 . However, it is not necessary that the halogen be bonded
to a more electronegative atom, since complexes between I2 and amines are also
known, as well as complexes in which C-Br is the electron acceptor [7, 8]. In a
complex stabilized by a pnicogen bond, the pnicogen atom may be bonded to an
electronegative or electropositive atom or group of atoms. The pnicogen atom acts
as an electron acceptor either through a σ-hole or a π-hole depending on the nature
of the pnicogen-containing molecule. Both the σ-hole and the π-hole correspond
to positive regions on the MEP surface. In addition, since the pnicogen atom has
a lone pair of electrons and in some cases π electrons, it may simultaneous act as
an electron donor. Examples of pnicogen-bonded complexes are H2 FP:PFH2 and
H2 (CH3 )P:P(CH3 )H2 [9].
The timeline for research on pnicogen interactions can be divided in two eras,
with the year 2011 being the dividing line between them. In 1995 two relevant
papers on pnicogen bonds appeared, one by Pyykko et al. on dipnicogen dimers,
(H2 Z-ZH2 )2 [10], and the other by Corriu et al. on intramolecular coordination
of phosphorous [11], as in phosphatranes [12]. In 2007, Murray, Politzer, et al.
introduced the concept of the σ-hole which was used to describe the halogen bond, and
subsequently the pnicogen bond [13]. Two other papers, one on P...P interactions in
halo-phosphines [14], and the other on weak interactions between trivalent pnicogen
centers in X3 Z...ZX3 complexes also appeared during this period [15]. The modern
era of studies of the pnicogen bond began in 2011 with the publication of two papers,
one a landmark paper by Hey-Hawkins and coworkers, with the title “Pnicogen
Bonds: A New Molecular Linker?” [9], and the other by Scheiner with the title
“A New Noncovalent Force: Comparison of P...N Interactions with Hydrogen and
Halogen Bonds” [16]. The subsequent resurgence of interest in the pnicogen bond
led to the publication of about one hundred articles on this subject from 2011 to
mid-2014.
8 The Pnicogen Bond in Review 193
8.1.1 Experimental Studies
In their 2011 paper, Hey-Hawkins et al. discussed P...P through-space NMR spin-
spin couplings [17], and an X-ray structure which suggested the existence of an
attractive interaction between P atoms [18]. In this paper they cited as an antecedent
the work of Sundberg et al. on 1,2-dicarba-cloro-dodecaboranes [19], and based
their discussion on the presence of a P...P bond critical point. In addition to these
two papers, the most significant experimental data on complexes involving pnicogen
bonds is a study of the microwave spectrum, structure, and dipole moment of the
F3 P:OH2 complex [20]. Politzer, Murray et al. reported a survey of crystal structure
data which supported the concept of σ-hole interactions [21].
8.1.2 Theoretical Studies
Studies of selected series of pnicogen-bonded complexes which have been carried

out in our laboratories [22–44] will be discussed in detail in this review. In the present
section we survey the works of other authors who have made notable contributions
to the literature on the pnicogen bond. We begin with the work of Scheiner and his
group, who reported studies of a variety of pnicogen-bonded complexes, emphasizing
substituent effects on P...N interactions [45–47], the sensitivity of the pnicogen bond
to angular distortion [48], and the possibility of N...N interactions [49]. They also
compared P...P and P...N bonds with other non-covalent interactions in trimers and
a tetramer of PH3 [50]; compared hydrogen bonding with P...N interactions [51];
investigated the effect of the carbon chain (R) in RH2 P...NH3 complexes [52, 53];
and compared the sensitivity of noncovalent bonds including hydrogen, halogen,
chalcogen, and pnicogen bonds to intermolecular separation [54, 55]. Scheiner also
published two reviews on pnicogen bonds [56, 57].
The group of Frontera has published papers on pnicogen–π complexes and noted
some biological implications [58]; characterized complexes with YO2 Br with Y=N,
P, and As [59]; carried out a CSD survey of halogen, chalcogen and pnicogen bonds
[60]; and reported a study of pnicogen bonds using different computational meth-
ods [44]. Solimannejad has investigated pnicogen, chalcogen, and halogen bonds
involving substituted s-triazines [61]; cooperativity with lithium bonds [62]; and
complexes involving NO2 X [63]. Grabowski reported a study of pnicogen bonds
versus hydrogen bonds [64], including large clusters such as NH+ 4 :(NCH)1−8 , and
NH+ 4 :(N2 )1−8 [65]. Contributions from other groups include studies of pnicogen-
hydride interactions [66]; cooperativity [67–69]; competition among hydrogen,
halogen, and pnicogen bonds [70]; pnicogen interactions involving anions [71];
electron transfer in pnicogen bonds [72]; and characterization of pnicogen bonds
using the Laplacian of the electron density [73].
8.2 Methods
The structures of the monomers and complexes reported in this review were opti-
mized at second-order Møller-Plesset perturbation theory (MP2) [74–77] using the
aug’-cc-pVTZ basis set [78]. This basis set is the Dunning aug-cc-pVTZ basis set
[79, 80] with diffuse functions removed from H atoms. Transition structures were
computed for some systems which were found to have double minima along the
pnicogen-bonding coordinate. Frequencies were computed to confirm equilibrium
and transition structures. These calculations were performed using the Gaussian 09
program [81].
The electron densities of the complexes were analyzed using the Atoms in
Molecules (AIM) methodology [82–85] and the Electron Localization Function
(ELF) [86], employing the AIMAll [87] and Topmod programs [88]. The topo-
logical analysis of the electron density produces the molecular graph of the complex.
This graph identifies the location of electron density features of interest, including
the electron density (ρ) maxima associated with various nuclei, saddle points which
correspond to bond critical points (BCPs), and ring critical points which indicate a
minimum electron density within a ring. The zero gradient line which connects a
BCP with two nuclei is the bond path. The electron density at the BCP (ρBCP ), the
Laplacian of the electron density at the BCP (∇ 2 ρBCP ), and the total energy density
(HBCP ) are additional useful quantities for characterizing the nature of intermolecular
interactions [89]. The ELF function illustrates those regions of space at which the
electron density is high. All of these measures are useful for identifying bonds and
lone pairs, and characterizing bond types. The Natural Bond Orbital (NBO) method
[90] was employed to obtain MP2/aug’-cc-pVTZ atomic charges and to analyze the
stabilizing charge-transfer interactions in these complexes, employing the NBO-5
and NBO-6 programs [91, 92]. Since MP2 orbitals are nonexistent, the charge-
transfer interactions were computed using the B3LYP functional [93, 94] with the
aug’-cc-pVTZ basis set at the MP2/aug’-cc-pVTZ geometries, so that at least some
electron correlations effects could be included. NBO orbitals were represented with
the Jmol program [95] using the tools developed by Marcel Patek. [96]
Spin-spin coupling constants were evaluated using the equation-of-motion cou-
pled cluster singles and doubles (EOM-CCSD) method in the CI (configuration
interaction)-like approximation, [97, 98] with all electrons correlated. For these cal-
culations, the Ahlrichs [99] qzp basis set was placed on 13 C, 15 N, 17 O, and 19 F, and
the qz2p basis set on 31 P and 35 Cl. The Dunning cc-pVDZ basis set was placed on
all 1 H atoms except for a hydrogen-bonded 1 H, in which case the qz2p basis set was
used. All of the EOM-CCSD calculations were carried out using ACES II [100] on
the IBM Cluster 1350 (Glenn) at the Ohio Supercomputer Center.

electrostatic potential of
PH2 F showing regions of
positive and negative charge.
The color code is Red > 0.04
> Yellow > 0.02 > Green >
0.00 > Blue. The positions of
local minima and maxima are
indicated with light blue and
black dots, respectively.
from Ref. [22]. Copyright
(2012) Elsevier)
8.3 Discussion
The phosphorous atom has been the most studied participant in the pnicogen bond.
The molecular electrostatic potentials of molecules containing phosphorus exhibit a
region of negative charge associated with the P lone pair, and a region of positive
charge at the same P which is the σ-hole. This is illustrated in Fig. 8.1 for the H2 FP
molecule. Thus, the P atom can act as both an electron-pair donor and an electron
acceptor at P in complexes with pnicogen bonds.
In the following sections, we review some of the research carried out in our lab-
oratories on pnicogen-bonded complexes, with special emphasis on their structures,
binding energies, some bonding properties, and spin-spin coupling constants. The
first section focuses on complexes with σ-σ pnicogen bonds in dimers (PH2 X)2 [22]
and [H2 C=PX]2 [34], mixed binary complexes H2 C=(X)P:PXH2 [35], H2 XP:PCX
[36], H2 XP:NXH2 [23], X=PH3 :NY [37] and X=PH3 :PY [37], H2 FP:ClX [38], and
the molecular anions H2YP:X− [39], with X and Y a variety of substituents. On the
H2 C=(X)P:PXH2 [35] and H2 XP:PCX surfaces [36] there are also equilibrium struc-
tures stabilized by π-σ pnicogen bonds, and these complexes are examined in the
second section. Relationships between the properties of corresponding complexes
with σ-σ and π-σ bonds are discussed. Complexes YN:PO2 X [33] which are sta-
bilized by π-σ bonds are also included in this section. The final section focuses on
ternary and higher-order complexes in which a pnicogen-bonded complex is involved
in additional intermolecular interactions [26, 27, 31]. The interest here is to examine
how these interactions change the energy of the pnicogen bond, and the nonadditivi-
ties of interaction energies. More detailed discussions of the complexes reviewed in
this chapter may be found in the original references, in which the geometries, total
energies, and molecular graphs are available as Supporting Information. Finally, it
should be noted that there may be other equilibrium structures on the potential sur-
faces in addition to those discussed in this chapter, but these have not been included
in this review.
Fig. 8.2 Equilibrium structures of (PH2 F)2 and [PH2 (CCH)]2 illustrating the P-P-F and P-P-C
arrangements which approach linearity. (Reprinted with permission from Ref. [22]. Copyright
(2012) Elsevier)
8.3.1 σ-σ Pnicogen Bonds
A σ-σ pnicogen bond arises when an atom in one molecule acts as a σ lone-pair
electron donor to a group 15 atom which is the electron-pair acceptor through its
σ-hole. If both atoms are pnicogen atoms, then it is possible that each will be an
electron-pair donor and acceptor. In this section, we discuss some of the properties
of complexes stabilized by σ-σ pnicogen bonds.
8.3.1.1 (PH2 X)2
The first pnicogen-bonded complexes that were investigated in our laboratories are
those formed between two formally sp3 hybridized molecules, PH2 X, with the sub-
stituents X=F, Cl, OH, NC (bonded at either N or C), CCH, CH3 , and H [22]. Table 8.1
reports selected properties of the (PH2 X)2 complexes, including their binding en-
ergies, charge-transfer energies, P-P distances, A-P-P angles with A the atom of X
directly bonded to P, and spin-spin coupling constants 1p J(P-P) across the pnicogen
bond. The equilibrium structures of these complexes have C2h symmetry, with a
P-P-A arrangement that approaches linearity. This arrangement is illustrated in
Fig. 8.2 for (PH2 F)2 and [PH2 (CCH)]2 .
The intermolecular P-P distances in dimers (PH2 X)2 range from 2.471 to 3.589
Å, which are significantly longer than the P-P covalent bond distance of 2.219 Å in
H2 P-PH2 . The binding energies of the dimers lie between –7.1 kJ·mol−1 for (PH3 )2
and –34.0 kJ·mol−1 for (PH2 F)2 , a range which is comparable to that of neutral
hydrogen-bonded complexes. As evident from Table 8.1, the more electronegative
substituents form the more strongly bound dimers, while the more electropositive
substituents form the more weakly bound complexes. The binding energies of these
complexes are related to the nature of the stabilizing charge-transfer energies, which
arise as electron density is transferred from the lone pair of one P atom to the anti-
bonding σ* P-A orbital of the other. The largest charge-transfer energy is found for
the most strongly bound complex (PH2 F)2 , and the smallest for (PH3 )2 which has
the weakest pnicogen bond. The charge transfer energies and the binding energies
are related quadratically, with a correlation coefficient R2 of 0.982. There are two
factors which play a role in determining the magnitude of the charge-transfer energy.
The first is the dependence of the charge-transfer energy on the P-P distance. For
Table 8.1 MP2/aug’-cc-pVTZ P-P distances (R, Å), A-P-P angles ( , o ), binding energies (E,
kJ·mol−1 ), charge-transfer energies [Plp (1)→σ*P-A(2), kJ·mol−1 ], and spin-spin coupling con-
stants [1p J(P-P), Hz] for complexes (PH2 X)2 . (Reprinted with permission from Ref. [22]. Copyright
(2012) Elsevier)
Complex R(P-P) A-P-Pa E Plp (1)→σ*P-A(2) 1p
J(P-P)
(PH2 F)2 2.471 163 − 34.0 131.8 998.6
(PH2 Cl)2 2.771 167 − 22.1 59.9 1119.9
(PH2 (OH)]2 2.851 169 − 20.6 46.6 644.0
(PH2 (NC)]2 3.040 168 − 13.8 31.6 640.3
(PH2 (CCH)]2 3.353 174 − 12.2 11.3 281.9
(PH2 (CN)]2 3.375 171 − 8.4 11.5 300.0
(PH2 (CH3 )]2 3.481 178 − 8.9 11.3 160.9
(PH3 )2 3.589 179 − 7.1 5.6 130.9
a
A is the atom of group X that is directly bonded to P
Fig. 8.3 Overlap of the

lone-pair orbital on one P
atom with the σ* P-A orbital
of the other in (PH2 F)2 and
(PH3 )2
complexes (PH2 X)2 , charge transfer has a quadratic dependence on distance, with
a correlation coefficient R2 of 0.976. The second factor is the degree of overlap be-
tween the P lone pair and the σ* P-A orbital, as illustrated in Fig. 8.3 for (PH2 F)2 and
(PH3 )2 . The importance of the nature of A in determining the properties of pnicogen-
bonded complexes can be seen in two papers on (PHFX)2 complexes with F-P...P-F,
H-P...P-H, and A-P...P-A approaching linearity [24, 25].
The AIM analysis shows that a bond critical point connects the two P atoms in
these complexes. The electron density at the BCP correlates exponentially with the
intermolecular P-P distance, with a correlation coefficient R2 of 0.999. This corre-
lation is found for various types of intermolecular bonds. [101–110] The Laplacian
of the electron density and the total energy density at the BCP are positive for these
complexes, except for (PH2 F)2 , (PH2 Cl)2 , and [PH2 (OH)]2 , which have negative val-
ues of the energy density. This indicates that the P...P bonds in these three complexes
have some degree of covalency. There is a significant electron-density buildup in the
region between the two phosphorus and at the atoms A, as illustrated in Fig. 8.4 for
(PH3 )2 . It is the build-up of charge in the region between the two P nuclei that results
in the formation of the pnicogen bond.
One-bond spin-spin coupling constants 1p J(P-P) provide additional insight into
the pnicogen bond. For complexes (PH2 X)2 , 1p J(P-P) values are dominated by the
Fermi contact terms which are excellent approximations to total J. Values of 1p J(P-P)
are reported in Table 8.1, and can be seen to vary between 131 Hz for (PH3 )2 to 1000
Fig. 8.4 Electron density

shifts for (PH3 )2 at the 0.0001
au isosurface. Blue and
yellow indicate regions of
decreased and increased
electron densities,
respectively. (Reprinted with
permission from Ref. [22].
Copyright (2012) Elsevier)
Fig. 8.5 1p J(P-P) versus the

P-P distance for complexes
(PH2 X)2 . (Reprinted with
Copyright (2012) Elsevier)
and 1120 Hz for (PH2 F)2 and (PH2 Cl)2 , respectively, thus illustrating the sensitivity
of P-P coupling to the nature of X. 1p J(P-P) exhibits a quadratic dependence on
the P-P distance with a correlation coefficient of 0.887, as illustrated in Fig. 8.5.
From this figure is appears that the value of 1p J(P-P) for (PH2 Cl)2 is too large for its
intermolecular distance. Using plots such as that of Fig. 8.5, it would be possible to
obtain an estimate of the intermolecular distance from the experimental value of the
corresponding coupling constant.
8.3.1.2 (H2 C=PX)2
Complexes (H2 C=PX)2 with formally hybridized sp2 P atoms may also form
pnicogen-bonded complexes of C2h symmetry with P-P-A and/or P-P-C approaching
linearity [34]. Complexes (H2 C=PX)2 withA-P...P-A linear are designated conforma-
tion A complexes. Their intermolecular P-P distances, selected structural parameters,
and binding energies are given in Table 8.2, and representative complexes (H2 C=PH)2
and (H2 C=PF)2 are illustrated in Fig. 8.6. As evident from Table 8.2, the range of
binding energies is relatively narrow, from − 9.5 to − 13.9 kJ·mol−1 , much smaller
than the range of binding energies of complexes (PH2 X)2 . Moreover, the order of
decreasing binding energies with respect to the substituent X differs significantly
in the two series, since (PH2 F)2 has the highest binding energy among (PH2 X)2
complexes, but (H2 C=PF)2 is a relatively weakly bound complex.
Since the binding energies of (H2 C=PX)2 complexes do not correlate with the
P-P distances, there may be another factor in addition to the pnicogen bond which
Table 8.2 Binding energies (E), charge-transfer energies [Plp (1)→σ*P-A(2), kJ·mol−1 ], P-P
and P-Hb distances (R,Å), and coupling constants [1p J(P-P), Hz] for conformation A complexes.
(Reprinted with permission from Ref. [34]. Copyright (2013) American Chemical Society)
(H2 C=PX)2 E R(P-P) R(P-Hb )a 1p
J(P-P) Plp (1)→σ*P-A(2)
X=CCH − 13.9 3.510 3.230 140.3 5.4
Cl − 13.3 3.373 3.305 330.1 8.7
CN − 11.6 3.456 3.293 187.3 6.9
H − 11.2 3.618 3.220 74.7 3.6
NC − 11.1 3.422 3.362 264.1 9.4
CH3 − 10.8 3.701 3.217 58.5 1.8
F − 10.4 3.477 3.406 247.9 6.0
OH − 9.5 3.583 3.370 134.3 4.1
a
Hb is the H atom of the CH2 group of one molecule which is closer to the P atom of the other
Fig. 8.6 Structures and molecular graphs of (H2 C=PF)2 and (H2 C=PH)2 illustrating the two types
of conformation A dimers. (H2 C=PF)2 has only a P...P pnicogen bond, while (H2 C=PH)2 has a
pnicogen bond and two P...Hb interactions. (Reprinted with permission from Ref. [34]. Copyright
(2013) American Chemical Society)
stabilizes these complexes. It is possible to divide the (H2 C=PX)2 complexes into
two groups: those which are stabilized by the pnicogen bond, and those which
are stabilized by the P...P pnicogen bond and in addition by two significant P...Hb
interactions, with Hb the H atom of the CH2 group of one molecule that is closer
to the P atom of the other. (H2 C=PF)2 in Fig. 8.6 is typical of the first group, and
(H2 C=PH)2 of the second. A plot of the binding energies of (H2 C=PX)2 complexes
versus the intermolecular P-P distance is given in Fig. 8.7. The complexes stabilized
by pnicogen bonds are those with the more electronegative substituents Cl, CN, F,
and OH. These complexes have P-Hb distances of 3.29 Å or greater, and P-P-C
angles between 90 and 93o . The remaining complexes with the more electropositive
substituents CCH, H, and CH3 appear in Fig. 8.7 to have binding energies which are
too high for their P-P distances. Complexes in this group have shorter P-Hb distances
near 3.22 Å and P-P-C angles between 83 and 87o . Since there is a second stabilizing
interaction in these three complexes in addition to the P...P pnicogen bond, it should
15
14
CCH
Cl
13
ΔE, kJ/mol
12
CN
H
11 NC
CH3
F
10
OH
9
3.3 3.4 3.5 3.6 3.7 3.8
R(P-P), Å
Fig. 8.7 Binding energies of conformation A complexes (H2 C=PH)2 versus the P-P distance.
Complexes with P...P pnicogen bonds Complexes involving two P...Hb interactions in addition
to a P...P bond. (Reprinted with permission from Ref. [34]. Copyright (2013) American Chemical
Society)
not be surprising that the binding energies of (H2 C=PX)2 complexes do not correlate
with the P-P distances. However, a correlation does exist for the subset of complexes
stabilized only by a pnicogen bond, as can be seen in Fig. 8.7.
Table 8.2 also presents P(1)lp →σ*P-A(2) stabilizing charge-transfer energies for
complexes (H2 C=PX)2 , but these energies are relatively small, ranging from 1.8 to
9.4 kJ·mol−1 , compared to 100 kJ·mol−1 for (PH2 F)2 with F-P...P-F linear. Thus,
they reflect the weaker pnicogen bonds in (H2 C=PX)2 . Charge transfer energies tend
to be greater in complexes stabilized only by a pnicogen bond than in those with both
a pnicogen bond and two P...Hb interactions.
Although it might have been expected that complexes (H2 C=PX)2 with C2h
symmetry and C-P...P-C approaching linearity should also be pnicogen-bonded equi-
librium structures, such is not the case. The only equilibrium structure of this type
is (H2 C=POH)2 , but it is stabilized primarily by two distorted O-H...P hydrogen
bonds. However, pnicogen-bonded complexes do exist which have the carbon of
the CH2 group of one molecule and atom A of the substituent X of the other ap-
proaching a nearly linear C-P...P-A arrangement. (H2 C=POH)2 and (H2 C=PH)2
have Cs symmetry, while the remaining equilibrium complexes with C-P...P-A ap-
proaching linearity have C1 symmetry. These include (H2 C=PCl)2 , (H2 C=PCH3 )2 ,
Table 8.3 P-P Distances [(R(P-P), Å], binding energies (E, kJ·mol−1 ), and 31 P-31 P spin-spin

coupling constants [1p J(P-P), Hz] of equilibrium (H2 C=PX)2 structures B and B . (Reprinted with
permission from Ref. [34]. Copyright (2013) American Chemical Society)
Complex Typea Sym. R(P-P) E 1p
J(P-P)b
(H2 C=POH)2 B Cs 3.738 − 16.4 43.5

(H2 C=PCl)2 B C1 3.457 − 11.6
(H2 C=PCH3 )2 B C1 3.696 − 11.4
B Cs 3.712 − 11.4 66.1
(H2 C=PCN)2 B C1 3.519 − 11.0
(H2 C=PNC)2 B C1 3.254 − 10.0
(H2 C=PH)2 B Cs 3.657 − 8.9 99.3
(H2 C=PF)2 B C1 3.579 − 7.7 95.9
a
The arrangement which approaches linearity is C-P(1)...P(2)-A
b1p
J(P-P) calculations were not feasible for complexes with C1 symmetry except (H2 C=PF)2
2
2
1
1
Fig. 8.8 Conformations B (H2 C=POH)2 and B (H2 C=PF)2 . Labels identify P(1) and P(2).
(H2 C=PCN)2 , (H2 C=PNC)2 , and (H2 C=PF)2 , although (H2 C=PCH3 )2 with C1 sym-
metry is essentially identical structurally and energetically to the Cs dimer. Structural
and energetic data for equilibrium complexes with Cs (B) and C1 (B’) symmetry
are given in Table 8.3. Figure 8.8 illustrates the structures of (H2 C=POH)2 and
(H2 C=PF)2 .
Except for (H2 C=PCH3 )2 , B and B conformations have interestingly different
structures. (H2 C=POH)2 is stabilized by a P...P pnicogen bond as well as an O-H...P
hydrogen bond, with a binding energy of –16.4 kJ·mol−1 . The hydrogen bond charge-
transfer energy P(2)lp →σ*O-H(1) is 15.8 kJ·mol−1 , significantly more stabilizing
than the P(1)lp →σ*P-O(2) energy of 2.3 kJ·mol−1 . (H2 C=PH)2 conformation B is
stabilized by a pnicogen bond and one P...Hb interaction, and has a binding energy
of –8.9 kJ·mol−1 . The NBO analysis shows that both molecules are involved as
lone pair donors and acceptors for the pnicogen bond with P(1)lp →σ*P-H(2) and
P(2)lp →σ*P-C(1) energies of 3.6 kJ·mol−1 .
Table 8.4 NBO

charge-transfer stabilization Complex P(1)lp →σ*P-A(2) P(2)lp →σ*P-C(1)
energies (kJ·mol−1 ) for (H2 C=PCl)2 5.2 4.9
(H2 C=PX)2 conformation B
(H2 C=PCH3 )a2 2.1 1.5
complexes. (Reprinted with
permission from Ref. [34]. (H2 C=PCN)2 3.9 4.2
Copyright (2013) American
(H2 C=PNC)2 5.6 4.0
Chemical Society)
(H2 C=PF)2 4.0 3.3
a
(H2 C=PCH3 )2 also has a stabilizing P...Hb interaction
In the five conformation B complexes with C1 symmetry, the monomers essen-

tially retain their plane of symmetry, but the plane of one monomer is rotated about
the P...P bond relative to the other, as illustrated in Fig. 8.8 for (H2 C=PF)2 . The bind-
ing energies of these complexes decrease in the order (H2 C=PCl)2 > (H2 C=PCH3 )2
> (H2 C=PCN)2 > (H2 C=PNC)2 > (H2 C=PF)2 . Except for (H2 C=PCH3 )2 , this is
the same order found for the complexes with conformation A. The increased relative
stability of (H2 C=PCH3 )2 among the B complexes may be attributed to the retention
of the P...Hb interaction in its essentially planar structure. Table 8.4 reports the ener-
gies of the two charge-transfer interactions which stabilize these complexes. These
energies are similar, and except for (H2 C=PCN)2 , P(1)lp →σ*P-A(2) is slightly more
stabilizing than P(2)lp →σ*P-C(1).
Spin-spin coupling constants for conformation A, B, and (H2 C=PF)2 (B ) com-
plexes are reported in Tables 8.2 and 8.3. For conformation A, these values range from
59 Hz for (H2 C=PCH3 )2 to 330 Hz for (H2 C=PCl)2 . Consistent with binding energies
and charge-transfer energies, this range is much smaller than found for the dimers
(PH2 X)2 . Coupling constants for the conformation B complexes and (H2 C=PF)2 lie
between 43 and 100 Hz. Coupling constants for (H2 C=PX)2 complexes correlate
quadratically with the P-P distance, with a correlation coefficient R2 of 0.933.
8.3.1.3 H2 C=(X)P:PXH2
Just as the (H2 C=PX)2 complexes may exist as conformation A and B dimers, the bi-
nary complexes H2 C=(X)P:PXH2 may also exist as A conformers with Cs symmetry,
and B with C1 symmetry [35]. Intermolecular P-P and Ps -Hb distances, Pd -Ps -A and
Ps -Pd -A angles, and binding energies for conformation A complexes are reported in
Table 8.5, and a plot of E versus the P-P distance is given in Fig. 8.9. The relatively
low correlation coefficient arises from an additional secondary interaction in com-
plexes with the more electropositive substituents, which involves the phosphorus of
Ps H2 X and the H atom of the CH2 group of H2 C=PX which is closer to it (Hb ). The
structural differences between complexes with and without this interaction can be
seen in Fig. 8.10 by comparing H2 C=(F)P:PFH2 and H2 C=(CH3 )P:P(CH3 )H2 , and
by comparing Ps -Hb distances and Ps -Pd -A angles in complexes with this interaction
compared to complexes with F, Cl, OH, and NC as substituents.
Table 8.5 P-P and Ps -Hb distances (R, Å), binding energies (E, kJ·mol−1 ), and Pd -Ps -A and Ps -
Pd -A angles ( , deg) of conformation A complexes H2 C=(X)P:PXH2 with Cs symmetry. (Reprinted
with permission from Ref. [35]. Copyright (2013) American Chemical Society)
R(P-P) E R(Ps -Hb ) Pd -Ps -A Ps -Pd -A
X=F 3.017 − 17.1 3.548 180 149
Cl 3.099 − 16.5 3.420 179 155
OH 3.204 − 13.9 3.547 177 155
NC 3.228 − 12.2 3.472 178 159
CCH 3.417 − 13.4 3.306 180 168
CN 3.404 − 10.2 3.358 176 168
CH3 3.572 − 10.2 3.260 175 171
H 3.600 − 9.5 3.254 175 177
Fig. 8.9 The binding energy (E) versus the P-P distance of H2 C=(X)P:PXH2 complexes with Cs
symmetry. The second-order trendline has a correlation coefficient R2 of 0.859. (Reprinted with
Fig. 8.10 Structures and molecular graphs of conformation A complexes H2 C=(F)P:PFH2 and
H2 C=(CH3 )P:P(CH3 )H2 , indicating a Ps ...Hb interaction in the latter but not in the former.
Table 8.6 NBO charges (au) on H2 C=PX and Pd lp→σ*Ps -A and Ps lp→σ*Pd -A stabilizing charge-
transfer energies (kJ·mol−1 ) for conformation A complexes H2 C=(X)P:PXH2 a . (Reprinted with
H2 C=(X)P:PXH2 Charge on H2 C=PX Pd lp→σ*Ps -A Ps lp→σ*Pd -A
X=F 0.034 36.6 17.1
Cl 0.022 26.2 16.9
OH 0.019 21.7 9.8
NC 0.013 21.9 13.6
CCH 0.005 10.0 6.2
CN 0.005 10.8 7.6
CH3 0.003 5.6 3.4
H 0.003 2.6 0.2
a
Pd refers to the phosphorus of H2 C=PX and Ps to that of PH2 X. See Fig. 8.11. A is the atom of X
which is directly bonded to P
The charges on H2 C=PX and the stabilizing charge-transfer energies for A com-
plexes H2 C=(X)P:PXH2 are reported in Table 8.6. In all complexes, charge transfer
from the Pd lone pair to the σ*Ps -A orbital is always more stabilizing than charge
transfer from the lone pair of Ps to the σ*Pd -A orbital, with Ps and Pd the singly- and
doubly-bonded P atoms, respectively. This suggests that the preferred direction of
charge transfer is determined by the nature of the electron-pair acceptor σ* orbital.
The dominant direction of charge-transfer is consistent with the positive charges
on H2 C=PX in these complexes, as indicated in Table 8.6. For complexes not hav-
ing a Ps ...Hb interaction, Pd lp→σ*Ps -A charge-transfer energies range from 21.7
to 36.6 kJ·mol−1 , while Ps lp→σ*Pd -A charge-transfer energies range from 9.8 to
17.1 kJ·mol−1 . For those complexes having this secondary interaction, the charge-
transfer energies are much smaller, ranging from 2.6 to 10.8 kJ·mol−1 for
Pd lp→σ*Ps -A, and 0.2 and 7.6 kJ·mol−1 for Ps lp→σ*Pd -A.
The molecular graphs of these complexes show the presence of bond critical
points and associated P...P bond paths. The Laplacians at the BCPs are always
positive. However, complexes with the more electronegative substituents F, Cl, OH,
and NC have the shortest P-P distances, the largest binding energies, and negative
total energy densities at the BCPs, indicating that there is some degree of covalency
in these P...P bonds.
H2 C=(X)P:P(X)H2 conformation B complexes have C=Pd ...Ps -A approach-
ing linearity, with C the carbon of H2 C=PX. Similar to (H2 C=POH)2 , only
H2 C=(OH)P:P(OH)H2 has an equilibrium planar structure which is stabilized by
a P...P pnicogen bond and an O-H...Ps hydrogen bond. The five remaining σ-σ
complexes with X=F, Cl, NC, CCH, and CN are nonplanar with C1 symmetry,
and are stabilized only by pnicogen bonds. Figure 8.11 illustrates the structures
of H2 C=(OH)P:P(OH)H2 and H2 C=(F)P:PFH2 . Stable H2 C=(CH3 )P:P(CH3 )H2 and
H2 C=(H)P:PH3 B complexes do not exist on the potential surfaces.
The intermolecular P-P distances and binding energies of conformation B com-
plexes are reported in Table 8.7. Not surprisingly, H2 C=(OH)P:P(OH)H2 which is
Fig. 8.11 The structures of H2 C=(OH)P:P(OH)H2 and H2 C=(F)P:PFH2 with conformation B.

Table 8.7 Binding energies (E, kJ·mol−1 ), intermolecular P-P distances (R, Å), NBO charges (au)
on H2 C=PX, and Pd lp→σ*Ps -A and Ps lp→σ*Pd =C stabilizing charge-transfer energies (kJ·mol−1 )
of conformation B complexes H2 C=(X)P:PXH2 . (Reprinted with permission from Ref. [35].
Copyright (2013) American Chemical Society)
R E H2 C=PX charge Pd lp→σ*Ps -A Ps lp→σ*Pd =C
X=F 3.089 − 13.8 0.036 31.7 8.9
Cl 3.172 − 14.1 0.020 21.0 9.6
OHa 3.355 − 21.9 –0.006 13.4 4.3b
NC 3.297 − 10.8 0.011 17.2 5.8
CCH 3.423 − 12.2 0.001 8.5 4.5
CN 3.463 − 8.8 0.002 8.9 4.9
a
Has Cs symmetry. All other complexes have C1 symmetry
b
Has an additional stabilizing Ps (lp)→σ*O-H charge-transfer energy of 14.8 kJ·mol−1
stabilized by an O-H...Ps hydrogen bond as well as a P...P pnicogen bond has the
highest binding energy of −21.9 kJ·mol−1 . Figure 8.12 presents a plot of the binding
energies of conformation B complexes versus the intermolecular P-P distances. The
point for H2 C=(OH)P:P(OH)H2 dramatically illustrates that its binding energy is
much too high for its P-P distance, a result of the stabilizing effect of the hydrogen
bond. Excluding the point for H2 C=(OH)P:P(OH)H2 , the trendline is linear, but the
correlation coefficient is only 0.67. With the exception of H2 C=(OH)P:P(OH)H2 ,
conformation B complexes are less stable than the corresponding conformation A
complexes.
For complexes (H2 C=PX)2 conformations A and B, the Pd -Ps -A angles tend to-
ward linearity, varying from 175 to 180◦ in conformation A, and from 168 to 177◦
in conformation B. However, the Ps -Pd -A and Ps -Pd =C angles exhibit greater devia-
tions, as evident from Tables 8.5 and 8.8. The Ps -Pd -A angles vary between 149 and
177◦ in conformation A, while the Ps -Pd =C angles vary from 138 to 156◦ in con-
formation B. The notable exception is the Ps -Pd =C angle in H2 C=(OH)P:P(OH)H2
which is 179◦ , thereby facilitating the formation of the O-H...Ps hydrogen bond. The
deviation from linearity particularly of the Ps -Pd =C angles in conformation B may
contribute to the lack of correlation between binding energies and P-P distances.
The NBO charges on H2 C=PX and the stabilizing charge-transfer energies
for conformation B complexes are reported in Table 8.7. As for conformation A
Fig. 8.12 The binding energy (E) versus the P-P distance for conformation B complexes.
H2 C=(OH)P:P(OH)H2 . (Reprinted with permission from Ref. [35]. Copyright (2013) American
Chemical Society)
1p
H2 C=(X)P:PXH2 <Pd -Ps -A <Ps -Pd =C J(P-P)
Table 8.8 Pd -Ps -A and
Ps -Pd =C angles (<, deg) and X=F 173 138 364.8
31 31
P- P spin-spin coupling Cl 172 129 319.5
constants [1p J(P-P), Hz] for
conformation B complexes OH 177 179 174.6
H2 C=(X)P:PXH2 . (Reprinted NC 170 150 231.8
with permission from Ref.
[35]. Copyright (2013) CCH 168 144 —a
American Chemical Society) CN 172 156 181.0
a
Not computed because of computational expense
complexes, charge transfer from the Pd lone pair to the σ*Ps -A orbital is always
more stabilizing than charge transfer from the lone pair of Ps to the σ*Pd =C orbital.
This is consistent with the positive charges on H2 C=PX in these complexes, except
for H2 C=(OH)P:P(OH)H2 , since the dominant charge-transfer interaction is across
the hydrogen bond, Ps (lp)→σ*O-H.
The P...P BCPs of conformation B complexes have smaller electron densities than
those of the corresponding conformation A complexes. This is consistent with the
longer P-P distances in conformation B, and with their smaller binding energies, ex-
cept for H2 C=(OH)P:P(OH)H2 . Only the complexes with X=F and Cl have negative
values of HBCP , indicating that the P...P bonds in these two complexes have some
degree of covalency. The degree of covalency is less in conformation B than in the
corresponding A complex.
Table 8.8 also reports the spin-spin coupling constants 1p J(P-P) for conformation B
complexes, and Fig. 8.13 shows the expected correlation between 1p J(P-P) and the P-P
distance. It is interesting to note that although H2 C=(OH)P:P(OH)H2 is structurally
Fig. 8.13 Coupling constants 1p J(P-P) versus the P-P distance for H2 C=(X)P:P(X)H2 conformation
B complexes. (Reprinted with permission from Ref. [35]. Copyright (2013) American Chemical
Society)
and energetically quite distinct from the remaining conformation B complexes, these
differences do not influence the relationship between 1p J(P-P) and the P-P distance.
The second-order trendline shown in Fig. 8.13 has a correlation coefficient R2 of
0.961.
It is possible to compare the properties of the mixed binary complexes
H2 C=(X)P:PXH2 to those of the dimers (PH2 X)2 and (H2 C=PX)2 , all with con-
formation A. The binding energies of these complexes are reported in Table 8.9.
These data indicate that for complexes in which X is one of the more electroneg-
ative substituents, the binding energies decrease from left to right, in going from
(PH2 X)2 to H2 C=(X)P:PXH2 to (H2 C=PX)2 . When X is an electropositive sub-
stituent, binding energies increase from left to right. The ratios of the binding energies
E(PH2 X)2 /E(H2 C=PX)2 and E[H2 C=(X)P:PXH2 ]/E(H2 C=PX)2 , are also re-
ported in Table 8.9, and plotted in Fig. 8.14. The trendline is a second-order
polynomial with a correlation coefficient R2 of 0.988. This plot indicates that there
is a systematic relationship among the relative stabilities of the complexes (PH2 X)2 ,
H2 C=(X)P:PXH2 , and (H2 C=PX)2 as a function of the substituent X, despite the fact
that some complexes H2 C=(X)P:PXH2 and (H2 C=PX)2 have stabilizing interactions
in addition to the P...P bond.
The second property of interest in conformation A complexes (PH2 X)2 ,
H2 C=(X)P:PXH2 , and (H2 C=PX)2 with A-P...P-A approaching linearity is the one-
bond spin-spin coupling constant 1p J(P-P) as a function of the P-P distance. Table 8.10
reports the P-P distances for these complexes and 1p J(P-P) values, and Fig. 8.15 pro-
vides a plot of these variables. Although for a given X, the shorter P-P distance and
larger 1p J(P-P) are found in complexes (PH2 X)2 compared to (H2 C=PX)2 , there is
some overlap, as evident from Fig. 8.15. 1p J(P-P) values cover a large range, from 59
Hz for (H2 C=PCH3 )2 to 1000 Hz for (PH2 F)2 and 1120 Hz for (PH2 Cl)2 . The points
Table 8.9 Binding energies (E, kJ·mol−1 ) of complexes and the ratios
E(PH2 X)2 /E(H2 C=PX)2 and E[H2 C=(X)P:PXH2 ]/E(H2 C=PX)2 for complexes with
A-P...P-A approaching linearity. (Reprinted with permission from Ref. [35]. Copyright (2013)
(PH2 X)2 a H2 C=(X)P:PXH2 b (H2 C=PX)2 a (PH2 X)2 / H2 C=(X)P:PXH2 /
(H2 C=PX)2 (H2 C=PX)2
X=F − 34.0 − 17.1 − 10.4 3.263 1.644
Cl − 22.1 − 16.5 − 13.3 1.657 1.235
OH − 20.6 − 13.9 − 9.5 2.175 1.471
NC − 13.8 − 12.2 − 11.1 1.244 1.100
CCH − 12.2 − 13.4 − 13.9 0.882 0.965
CH3 − 8.9 − 10.2 − 10.8 0.821 0.945
CN − 8.4 − 10.2 − 11.6 0.721 0.880
H − 7.1 − 9.5 − 11.2 0.634 0.847
a
C2h symmetry
b
Cs symmetry
Fig. 8.14 E(PH2 X)2 /E(H2 C=PX)2 (Ratio 1) versus E[H2 C=(X)P:PH2 X]/E(H2 C=PX)2
(Ratio 2) for corresponding complexes with A-P...P-A approaching linearity. (Reprinted with
for the latter two complexes deviate most from the trendline. Nevertheless, Fig. 8.15
indicates that complexes in these three series with similar P-P distances have similar
values of 1p J(P-P).
Table 8.10 P-P distances (R, Å) and 31 P-31 P spin-spin coupling constants [1p J(P-P), Hz] for
(PH2 X)2 , H2 C=(X)P:PXH2 , and (H2 C=PX)2 conformation A complexes with A-P...P-A approach-
ing linearity. (Reprinted with permission from Ref. [35]. Copyright (2013) American Chemical
Society)
(PH2 X)a2 H2 C=(X)P:PXHb2 (H2 C=PX)a2
1p 1p 1p
X= R J(P-P) R J(P-P) R J(P-P)
F 2.471 1000 3.017 637 3.477 248
Cl 2.771 1120 3.099 617 3.373 330
OH 2.851 644 3.204 361 3.583 134
NC 3.040 640 3.228 428 3.422 264
CCH 3.353 282 3.417 210 3.510 140
CH3 3.481 161 3.572 106 3.701 59
CN 3.375 300 3.404 244 3.456 187
H 3.589 131 3.600 101 3.618 75
a
C2h symmetry
b
Cs symmetry
Fig. 8.15 1p J(P-P) versus the P-P distance for conformation A complexes (PH2 X)2 , H2 C
=(X)P:PXH2 , and (H2 C=PX)2 . The correlation coefficient R2 is 0.927. (Reprinted with
8.3.1.4 H2 XP:PCX
PH2 X and PCX form pnicogen-bonded complexes only when the substituents are
CCH, NC, CN, CH3 , and H [36]. The binding energies, P-P distances, Pt -Ps -A and
Ps -Pt -C angles, and charge-transfer energies of the equilibrium H2 XP:PCX com-
plexes are reported in Table 8.11. The binding energies range from − 3.1 kJ·mol−1
Table 8.11 Binding energies (E, kJ·mol−1 ), P-P distances (R, Å), Pt -Ps -A and Ps -Pt -C angles
( , deg), charge-transfer energies (kJ·mol−1 ), and spin-spin coupling constants [1p J(P-P), Hz] for
H2 XP:PCX conformation A complexesa . (Reprinted with permission from Ref. [36]. Copyright
(2014) Springer Science and Business Media)
H2 XP:PCX E R(P-P) Pt -Ps -A Ps -Pt -C Pt (lp)→σ*Ps -A Ps (lp)→σ*Pt -C 1p
J(P-P)
X=CCH − 7.4 3.594 163 179 6.4 1.4 157.6
NC − 4.2 3.521 166 179 9.4 2.5 209.4
CN − 3.1 3.649 167 176 5.6 1.5 150.4
CH3 − 5.7 3.705 160 176 3.4 1.5 100.8
H − 4.7 3.772 158 175 3.9 1.4 88.3
a
Complexes with X=F and Cl do not form complexes with σ-σ pnicogen bonds, and H2 (OH)P:PCOH
is a hydrogen-bonded complex
for H2 (CN)P:PCCN to –7.4 kJ·mol−1 for H2 (CCH)P:PCCCH. These binding ener-

gies are smaller than the binding energies of the corresponding complexes (PH2 X)2
and H2 C=(X)P:PXH2 . For a given X, the intermolecular P-P distances in these three
sets of complexes decrease in the order H2 XP:PCX > H2 C=(X)P:PXH2 > (PH2 X)2 .
The Ps -Pt -C alignment in conformation A complexes closely approaches linearity,
while the Pt -Ps -A alignment deviates from linearity to some extent. Pt and Ps are the
triply- and singly-bonded P atoms of PCX and PH2 X, respectively.
H2 XP:PCX complexes with pnicogen bonds are stabilized by charge transfer. The
more favorable charge-transfer interaction involves donation of the Pt lone pair to the
σ* Ps -A orbital. Charge-transfer energies range from 3.4 kJ·mol−1 when X=CH3 to
9.4 kJ·mol−1 when X=NC. In contrast, charge-transfer energies from the Ps lone pair
to the σ* P=C orbital are 2.5 kJ·mol−1 when X=NC, and 1.4 or 1.5 kJ·mol−1 for the
remaining complexes. The preference for charge transfer to PH2 X further supports
the suggestion that the factor which determines the direction of charge transfer is the
nature of the σ* P-A orbital. The Pt (lp)→σ*P-A charge-transfer energies do not cor-
relate with the binding energies of H2 XP:PCX complexes, but correlate quadratically
with the P-P distances, with a correlation coefficient R2 of 0.963.
Table 8.11 also reports the values of 1p J(P-P) for these conformation A com-
plexes. The coupling constants vary from 88 Hz for H3 P:PCH to 209 Hz for
H2 (NC)P:PCNC. A second-order trendline with a correlation coefficient of 0.961
illustrates the dependence of 1p J(P-P) on the P-P distance.
8.3.1.5 H2 XP:NXH2 , H2 FP:NXH2 , and H2 XP:NFH2
All of the complexes discussed thus far have P...P pnicogen bonds, but a number
of other atoms with lone pairs may act as electron donors to group 15 elements to
form pnicogen bonds. Among these is nitrogen, which is also a pnicogen atom that
forms a series of complexes H2 XP:NXH2 , for X=F, Cl, OH, CN (bonded at both C
and N), CCH, CH3 , and H [23]. The majority of these complexes have Cs symmetry
Fig. 8.16 Complexes H2 FP:NFH2 , H2 (CH3 )P:N(CH3 )H2 , H2 (CCH)P:N(CCH)H2 , and

H2 (CH3 )P:NFH2 with P...N pnicogen bonds. (Reprinted with permission from Ref. [23]. Copyright
with the H atoms of PXH2 and NXH2 trans with respect to the P-N axis. The com-
plexes with unsaturated substituents H2 (NC)P:N(NC)H2 , H2 (CCH)P:N(CCH)H2 ,
and H2 (CN)P:N(CN)H2 have C1 symmetry and a gauche arrangement of these atoms.
Figure 8.16 illustrates the structures of selected complexes, and Table 8.12 presents
the intermolecular N-P distances, binding energies, and A-P-N and P-N-A angles.
Binding energies range from − 8 to − 27 kJ·mol−1 . These binding energies are greater
than those of the corresponding complexes (PH2 X)2 , except for (PH2 F)2 which is
7 kJ·mol−1 more stable than H2 FP:NFH2 , [PH2 (OH)]2 which is 1 kJ·mol−1 more
stable than H2 (OH)P:N(OH)H2 , and [PH2 (CH3 )]2 and H2 (CH3 )P:N(CH3 )H2 which
have similar stabilities.
The intermolecular N-P distances for the entire set of complexes in Table 8.12
range from 2.448 Å in H2 FP:N(CH3 )H2 to 3.292 Å in H3 P:NH3 , and correlate well
with binding energies, as illustrated in Fig. 8.17. TheA-P-N angles approach linearity,
varying between 160 and 172◦ . This orientation provides for the overlap of the
nitrogen lone pair orbital with the σ* P-A orbital. The P-N-A angles are significantly
less than the A-P-N angles, a reflection of the tetrahedral arrangement of the N-A
bond and the lone pair at N. However, H2 (CH3 )P:NFH2 and H3 P:NFH2 have P-N-A
angles of 180 and 177◦ , respectively, and in these two complexes the dominant
charge-transfer interaction is from the lone pair on P to the σ* N-A orbital, as can be
seen in Table 8.13. H2 (CH3 )P:N(CH3 )H2 and H3 P:NH3 have P-N-A angles of 163
and 155◦ , respectively. These four complexes have relatively weak pnicogen bonds.
In contrast, complexes H2 FP:NXH2 have the largest binding energies of −29 to −39
kJ·mol−1 .
Bond critical points and corresponding bond paths link the phosphorous and nitro-
gen atoms in each complex. The Laplacians at the BCPs are positive, but the energy
densities for the complexes of PH2 F, PH2 Cl and PH2 (OH) are negative, indicating
that the N...P bonds in complexes involving these molecules have some degree of
covalent character [111].
Table 8.12 MP2/aug’-cc-pVTZ N-P distances (R, Å), angles A-P-N and P-N-A ( , o ), binding
energies (E, kJ·mol−1 ), and spin-spin coupling constants [1p J(N-P), Hz] for pnicogen-bonded
complexes with N...P bonds. (Reprinted with permission from Ref. [23]. Copyright (2011)American
Chemical Society)
Complexes H2 XP:NXH2 R(N-P) A-P-Na P-N-Aa E 1p
J(N-P)
H2 FP:NFH2 2.524 172 131 − 26.7 − 113.6
H2 ClP:NClH2 2.669 167 121 − 24.3 − 90.8
H2 (OH)P:N(OH)H2 2.750 160 112 − 19.2 − 52.0
H2 (NC)P:N(NC)H2 2.887 166 110 − 17.4 − 41.4
H2 (CCH)P:N(CCH)H2 3.140 165 95 − 15.1 − 19.4
H2 (CN)P:N(CN)H2 3.208 166 90 − 12.8 − 16.8
H2 (CH3 )P:N(CH3 )H2 3.257 172 163 − 8.6 − 19.4
H3 P:NH3 3.292 165 155 − 7.8 − 17.5
Complexes H2 XP:NFH2
H2 ClP:NFH2 2.695 171 130 − 21.0 − 105.0
H2 (CH3 )P:NFH2 3.120 168 180 − 12.8 − 67.9
H3 P:NFH2 3.225 165 177 − 10.7 − 33.1
Complexes H2 FP:NXH2
H2 FP:N(CH3 )H2 2.448 166 118 − 39.4 − 41.3
H2 FP:NH3 2.609 168 122 − 29.9 − 65.6
H2 FP:NClH2 2.543 169 126 − 29.1 − 95.1
a
A is the atom of X or X’ that is directly bonded to P or N, respectively
Fig. 8.17 Binding energies of complexes H2 XP:NXH2 , H2 FP:NXH2 , and H2 XP:NFH2 vs. the
N–P distance. The exponential relationship has a correlation coefficient of 0.925. (Reprinted with
Table 8.13 Stabilizing Complexes H2 XP:NXH2 N(lp)→σ*P-Aa P(lp)→σ*N-Aa

charge-transfer energies
(kJ·mol−1 ) for complexes H2 FP:NFH2 53.9 12.1
H2 XP:NXH2 , H2 XP:NFH2 , H2 ClP:NClH2 42.6 4.0
and H2 FP:NXH2 . (Reprinted
with permission from Ref. H2 (OH)P:N(OH)H2 33.7 1.6
[23]. Copyright (2011) H2 (NC)P:N(NC)H2 23.9 0.7
H2 (CCH)P:N(CCH)H2 7.9 0.3
H2 (CN)P:N(CN)H2 6.8 0.8
H2 (CH3 )P:N(CH3 )H2 4.9 1.5
H3 P:NH3 6.5 1.7
Complexes H2 XP:NFH2
H2 ClP:NFH2 42.7 5.7
H2 (CH3 )P:NFH2 4.7 6.4
H3 P:NFH2 0.9 1.1
Complexes H2 FP:NXH2
b b
H2 FP:N(CH3 )H2
H2 FP:NH3 11.6 0.9
H2 FP:NClH2 47.4 8.8
a
A is the atom of X or X bonded directly to P or N, respectively
b
The NBO method considers this complex to be a single molecule
Table 8.12 also reports coupling constants 1p J(N-P) for these complexes. 1p J(N-P)
varies from − 17 Hz for H2 (CN)P:N(CN)H2 to −114 Hz for H2 FP:NFH2 . Figure 8.18
presents a plot of 1p J(N-P) versus the N-P distance for complexes with the same
Fig. 8.18 1p J(P-N) vs. the P-N distance for the pnicogen-bonded complexes H2 XP:NXH2 ,
H2 XP:NFH2 , and H2 FP:NXH2 . (Reprinted with permission from Ref. [23]. Copyright (2011)
substituent bonded to both N and P. For these, there is a quadratic increase in 1p J(N-P)
as the intermolecular N-P distance decreases, with a correlation coefficient R2 of
0.973. The points for H2 FP:NXH2 lie below the trendline for the H2 XP:NXH2
complexes, while points for H2 XP:NFH2 lie above the trendline in Fig. 8.18.
8.3.1.6 Complexes X=PH3 :NY and X=PH3 :PY
The electrostatic potentials of molecules O=PH3 , S=PH3 , HN=PH3 , and H2 C=PH3

exhibit a region of positive charge on the side opposite the X=P bond [37]. These
regions are analogous to those generated by electronegative atoms, and are σ-holes
which are suitable for interaction with a Lewis base. The maximum values of the
MEPs decrease with respect to X in the order O > S > NH > CH2 . The MEP minima
for the nitrogen bases NH3 , NCH, N2 , and the phosphorus base PH3 indicate a region
of negative charge along the principal symmetry axis of each molecule at N for the
nitrogen bases and at P of PH3 . However, the MEP of PCH at P along its symmetry
axis is positive. This is consistent with the observation that PCH is a poor base,
although it does form linear hydrogen-bonded complexes with itself, FH, and ClH.
The P-N distances and binding energies of complexes X=PH3 :NY and X=PH3 :PY
are reported in Table 8.14. For a given X=PH3 , the binding energies of the stronger
bases decrease in the order NH3 > NCH > PH3 , while the weaker bases N2 and PCH
have binding energies between − 6 and − 7 kJ·mol−1 with all X=PH3 . Moreover,
for each of the nitrogen bases and PH3 , the binding energies decrease with respect
to the acid in the order O=PH3 > S=PH3 > HN=PH3 > H2 C=PH3 , but there is little
variation among binding energies when PCH is the base.
These observations are illustrated in Figs. 8.19 and 8.20. Figure 8.19 provides a
plot of the binding energies as a function of the minimum values of the MEPs of
Table 8.14 MP2/aug’-cc-pVTZ binding energies (E, kJ·mol−1 ) and intermolecular distances (R,
Å) of complexes X=PH3 :NY and X=PH3 :PY. (Reprinted with permission from Ref. [37]. Copyright
E R(P-N)
X=PH3 NH3 NCH N2 NH3 NCH N2
O=PH3 − 20.4 − 17.8 − 7.2 3.145 3.120 3.287
S=PH3 − 18.9 − 17.3 − 6.9 3.241 3.190 3.362
HN=PH3 − 16.2 − 14.3 − 6.3 3.249 3.201 3.357
H2 C=PH3 − 13.9 − 12.6 − 5.7 3.331 3.266 3.416
E R(P-N)
X=PH3 PH3 PCH PH3 PCH
O=PH3 − 12.6 − 6.6 3.629 3.684
S=PH3 − 11.7 − 6.8 3.732 3.729
HN=PH3 − 10.5 − 6.4 3.713 3.690
H2 C=PH3 − 9.3 − 6.2 3.767 3.752
Fig. 8.19 Binding energies of complexes of X=PH3 with P and N bases as a function of the MEP
minima of the bases. Vertical stacks of points are identified by the corresponding bases. (Reprinted
the N and P bases, which are identified along the MEP minima axis. The stronger
bases NH3 , NCH, and PH3 have binding energies which differentiate among the
acids X=PH3 , while the binding energies of complexes with the weaker bases N2
and PCH show little dependence on the nature of the acid. An alternate view of the
binding energies can be seen in Fig. 8.20 in a plot of the binding energies against the
maximum values of the MEPs for the X=PH3 acids, which are identified along the
MEP maxima axis. The trendlines indicate that the stronger bases NH3 , NCH, and
PH3 have binding energies which depend on the nature of the acid X=PH3 . This is
also true but to a much lesser extent for the weaker base N2 . However, the trendline
for PCH is flat, showing that the binding energy is essentially independent of the acid.
Table 8.15 reports the stabilizing charge-transfer energies for complexes
X=PH3 :NY and X=PH3 :PY. Charge transfer occurs from the lone pair of P or N to
the antibonding σ* P=A orbital of X=PH3 , where A is the atom of X bonded directly
to P. The charge-transfer energies for these complexes are within ± 3.2 kJ·mol−1 of
the charge-transfer energy of 6.5 kJ·mol−1 for the N lone pair to the σ* P-H orbital of
H3 P:NH3 . Figure 8.21a illustrates the lone-pair orbital of NH3 interacting with the
σ* P=O orbital of O=PH3 . For comparison, the N lone pair orbital and the σ* P-H
orbital for H3 P:NH3 are shown in Fig. 8.21b. In X=PH3 complexes with the nitrogen
bases, charge transfer energies are largest when O=PH3 is the acid, and smallest when
H2 C=PH3 is the acid. The energies for S=PH3 and HN=PH3 complexes are interme-
diate and similar. With all X=PH3 , NH3 has the largest charge-transfer energy, and
PCH and N2 have similar, relatively small charge-transfer energies. Charge-transfer
energies tend to increase as the binding energies of these complexes increase.
Fig. 8.20 Binding energies of complexes of X=PH3 with P and N bases as a function of the MEP
maxima of the acids. Vertical stacks of points are identified by the corresponding X=PH3 acids.
Table 8.15 Stabilizing charge-transfer energies (kJ·mol−1 ) from the lone pair of the P and N bases
to the σ* P=A orbital of X=PH3 a . (Reprinted with permission from Ref. [37]. Copyright (2014)
Acid/Base NH3 NCH N2 PH3 PCH
O=PH3 9.7 6.2 3.7 6.7 3.4
S=PH3 7.7 5.1 2.8 4.5 2.6
HN=PH3 7.8 5.1 3.2 5.4 3.5
H2 C=PH3 6.2 4.2 2.6 4.1 2.7
a
A is the atom of X directly bonded to P
In complexes of X=PH3 with the N and P bases, bond paths connect the N or P
of the base to the P and H atoms of the acid, or just to the H atoms or the P-H bonds.
This does not mean that the intermolecular bonds are not pnicogen bonds, since the
charge-transfer energies from the N or P lone pair to the σ* P-H orbitals of X=PH3
are negligibly small. The orbital representation of charge transfer for the complex
HN=PH3 :NH3 which has a bond path involving H atoms is illustrated in Fig. 8.21c.
The bond paths to H atoms may simply be a consequence of the diffuseness of the σ*
P=A orbital of X=PH3 , and should not be interpreted as an indication of the absence
of a pnicogen bond.
The total coupling constants 1p J(P-N) and 1p J(P-P) and the corresponding P-N
and P-P distances are reported in Table 8.16. Figure 8.22 presents a plot of 1p J(P-N)
versus the P-N distance for these complexes, which has a second-order trendline with
a correlation coefficient R2 of 0.868. In this plot there are four sets of two data points
with similar P-N distances and coupling constants. How these arise can be seen in
Fig. 8.23, which provides a plot of 1p J(P-N) versus the P-N distance as a function
Fig. 8.21 Depiction of the orbitals involved in charge-transfer interactions. NH3 lone pair with the
a σ*P=O orbital of O=PH3 ; b σ*P-H orbital of PH3 ; c σ*P=N orbital of HN=PH3 . (Reprinted with
Table 8.16 Intermolecular distances (R, Å) and 1p J(P-N) and 1p J(P-P) spin-spin coupling constants
(Hz) for complexes of X=PH3 with nitrogen and phosphorus bases. (Reprinted with permission
from Ref. [37]. Copyright (2014) American Chemical Society)
1p 1p
R(P-N) J(P-N) R(P-P) J(P-P)
O=PH3
Base=NH3 3.145 − 19.9 PH3 3.269 150.8
NCH 3.120 − 17.7 PCH 3.684 92.3
N2 3.287 − 8.7
S=PH3
Base=NH3 3.241 − 14.1 PH3 3.732 105.9
NCH 3.190 − 12.5 PCH 3.729 68.1
N2 3.362 − 5.5
HN=PH3
Base=NH3 3.249 − 13.9 PH3 3.713 107.2
NCH 3.201 − 12.6 PCH 3.690 75.4
N2 3.357 − 6.0
H2 C=PH3
Base=NH3 3.331 − 9.6 PH3 3.767 75.5
NCH 3.266 − 8.9 PCH 3.752 53.7
N2 3.416 − 4.1
Fig. 8.22 1p J(P-N) versus the P-N distance for complexes X=PH3 :NY with P...N bonds. (Reprinted
Fig. 8.23 1p J(P-N) versus the P-N distance for complexes X=PH3 :NY as a function of the nature of
the base. (Reprinted with permission from Ref. [37]. Copyright (2014) American Chemical Society)
of the nature of the base. For each base there is one point at a short distance, two at
similar intermediate distances, and one at a long distance. Thus, the pairs of points at
similar distances in Fig. 8.22 correspond to the same base with two different acids,
S=PH3 and HN=PH3 . The trendlines in Fig. 8.23 have correlation coefficients R2
between 0.983 and 0.997.
Coupling constants 1p J(P-P) do not correlate with the P-P distance as a function
of the nature of the acid, but the correlation is improved as a function of the base.
Fig. 8.24 1p J(P-P) versus the P-P distance for X=PH3 :PY complexes as a function of the base.
Figure 8.24 shows the variation of 1p J(P-P) with distance for these complexes. The
linear trendlines have correlation coefficients of 0.972 for PH3 and 0.857 for PCH.
8.3.1.7 H2 FP...ClX Pnicogen Bonds
Although both pnicogen bonds and halogen bonds may be formed in complexes
H2 FP:ClX, with X=F, Cl, CN (bonded at C and N), CCH, CH3 , and H, here we focus
only on pnicogen-bonded complexes [38]. Equilibrium complexes ZB-1 have Cs
symmetry with the Cl-A bond cis to the bisector of the H-P-H angle, with A the atom
of X directly bonded to Cl. Complexes ZB-2 may have either Cs or C1 symmetry,
with the Cl-A bond trans or gauche, respectively, to the bisector. H2 FP:ClCH3 ZB-1
and ZB-2 complexes are illustrated in Fig. 8.25. Pnicogen-bonded complexes exist on
all H2 FP:ClX surfaces except H2 FP:ClF and H2 FP:ClNC. The very electronegative
substituents F and NC withdraw sufficient electron density from Cl so that it can no
longer act as an electron-pair donor.
The structures and binding energies of the remaining H2 FP:ClX complexes are
reported in Table 8.17. The complexes with the most electropositive substituent,
H2 FP:ClCH3 ZB-1 and ZB-2, have the highest binding energies of − 15.1 and − 18.3
kJ·mol−1 , respectively. The binding energies of complexes with substituents H, Cl,
and CCH have binding energies between − 10.0 and − 11.6 kJ·mol−1 . H2 FP:ClCN
complexes have the smallest binding energies of about − 6.4 kJ·mol−1 . In each group,
the P-Cl distance is shortest in the most strongly bound complex and longest in the
most weakly bound. A linear correlation exists between the binding energies and
the P-Cl distances in ZB-2 complexes, with a correlation coefficient of 0.958. A
Fig. 8.25 H2 FP:ClCH3 complexes ZB-1 and ZB-2. (Reprinted with permission from Ref. [38].
Table 8.17 P-Cl distances (R, Å), F-P-Cl and P-Cl-A angles ( , deg), and binding energies (E,
kJ·mol−1 ) of pnicogen-bonded complexes H2 FP:ClX. (Reprinted with permission from Ref. [38].
H2 FP:ClX ZB-1 ZB-2
R F-P-Cl P-Cl-A E R F-P-Cl P-Cl-A E
X=CH3 3.157 163 100 − 15.1 3.121 169 90 − 18.3
H 3.280 167 110 − 10.0 3.277 169 80 − 11.6
Cl 3.169 168 110 − 11.5 3.265 166 89 − 10.8
CCH 3.295 162 93 − 11.0 3.320 165 85 − 10.7
CN 3.365 165 95 − 6.5 3.437 168 81 − 6.3
good correlation between these two variables is not found for the ZB-1 complexes,
suggesting that there may be secondary interactions between the substituents and the
H atoms of H2 FP in the cis orientation.
Values of the F-P-Cl and P-Cl-A angles indicate that H2 FP:ClX complexes are
stabilized by pnicogen bonds. In ZB-1 and ZB-2 complexes, the F-P-Cl angle char-
acteristically approaches linearity, varying between 162 and 169◦ . At the same time,
values of the P-Cl-A angle are between 93 and 110◦ in ZB-1 and between 81 and
90◦ in ZB-2. These angles indicate that no halogen bond exists in these complexes.
In ZB-1 and ZB-2 complexes, the stabilizing charge-transfer interaction occurs
from the lone pair of Cl to the σ* P-F orbital of H2 FP. The data of Table 8.18 indicate
that the weakest charge transfer interaction in each series occurs in H2 FP:ClCN, and
the strongest in H2 FP:ClCH3 . However, the charge-transfer energies do not correlate
with the binding energies of these complexes, but do correlate linearly with the P-Cl
distances, with correlation coefficients R2 of 0.950 and 0.941 for ZB-1 and ZB-2,
respectively. Figure 8.26 provides a pictorial description of the lone-pair orbital on Cl
and the σ* P-F orbital which are involved in charge transfer in H2 FP:ClCl complexes.
1p
J(P-Cl) values for these pnicogen-bonded complexes are reported in Table 8.18.
Values range from 25 to 57 Hz for complexes ZB-1, and from 16 to 29 Hz for ZB-2.
Values for complexes ZB-1 are always greater than the corresponding ZB-2 values,
although the P-Cl distances in ZB-1 complexes are not always shorter than those in
ZB-2. The tendency for 1p J(P-Cl) to increase as the P-Cl distance decreases can be
seen for all pnicogen bonded complexes, although the data are definitely scattered.
Table 8.18 Cl(lp)→σ*P-F charge-transfer energies (kJ·mol−1 ) and 1p J(Cl-P) spin-spin coupling
constants (Hz) for pnicogen-bonded complexes H2 FP:ClX ZB-1 and ZB-2. (Reprinted with
1p
Cl(lp)→σ*P-F J(Cl-P)
H2 FP:ClX ZB-1 ZB-2 ZB-1 ZB-2
X=CH3 19.4 20.4 40.6 29.0
H 13.6 13.7 36.8 22.4
Cl 17.7 13.6 56.8 20.2
CCH 9.6 7.9 26.8 18.5
CN 7.4 5.2 24.7 15.6
Fig. 8.26 Representation of the orbitals involved in charge-transfer interactions in H2 FP:ClCl com-
plexes ZB-1 and ZB-2. (Reprinted with permission from Ref. [38]. Copyright (2014) American
Chemical Society)
The correlation coefficients for the variation of 1p J(P-Cl) as a function of the P-Cl
distance are 0.705 for ZB-1 and 0.928 for ZB-2.
8.3.1.8 Pnicogen-Bonded Anionic Complexes
Anions are strong electron pair donors which can therefore form strong pnicogen
bonds [39]. The binding energies of 21 equilibrium structures H2YP:X− , for X,
Y=CH3 , CCH, F, CN (bonded through C and N), and Cl are given in Table 8.19.
The binding energies are defined as the negative of the dissociation energies relative
to the more stable H2YP and X− products. The diagonal elements of Table 8.19 are
the energies of the complexes (H2 XPX)− which have C2v symmetry and symmetric
X-P-X bonds. From the arrangement of complexes in Table 8.19, it is possible to
determine for any given X, Y pair, which substituent will be covalently bonded to
P, and which will form the ion-molecule bond. The anion X− is determined by the
ability of X to accommodate a negative charge, which is
Cl > NC > F > CCH > CH3 .
The complexes (H2 FPF)− and H2 (CH3 )P:F− are illustrated in Fig. 8.27.
Table 8.19 Binding energies (kJ·mol−1 ) of complexes H2YP:X−a . (Reprinted with permission from
Ref. [39]. Copyright (2014) American Chemical Society)
H2YP:X− H2 (CH3 )P H2 (CCH)P H2 FP H2 (CN)P H2 (NC)P H2 ClP
X=CH3 − 70.0
CCH − 24.0 − 73.1
F − 57.4 − 119.8 − 180.5
CN − 17.7 − 50.0 − 126.5 − 102.2
NC − 19.8 − 50.3 − 98.5 − 93.6 − 131.8
Cl − 19.4 − 50.2 − 85.4 − 93.8 − 120.9 − 113.1
−
a
Binding energies relative to the more stable monomers H2YP + X
Fig. 8.27 The symmetric molecular anion (H2 FPF)− and the anionic complex H2 (CH3 )P:F− .
To understand the variation in binding energies, it is advantageous to first examine

the structures of complexes H2YP:X− which have Cs symmetry. These complexes
can be described in terms of the P-A and P-A distances and the A-P-A angle, where
A and A are the atoms of X and Y, respectively, which are directly bonded to P. In
these complexes, the A-P-A arrangement tends toward linearity, a recurring feature
of complexes with pnicogen bonds. The angle varies between 162 and 169 degrees,
except for the complexes of P(CH3 )H2 with CN− , NC− , and Cl− , which have A-P-C
angles between 152 and 156 degrees.
Tables 8.20 and 8.21 give the P-A and P-A distances, respectively, in complexes
H2YP:X− . If the P-A bond is shorter than the P-A bond in the corresponding sym-
metric complex, or if it is longer but within 0.30 Å of the length of the P-A bond
in the symmetric structure, we have assumed that it has some covalent character.
Moreover, if the P-A bond is within 0.15 Å of the length of the P-A bond in the
corresponding symmetric complex, the P-A bond has a reduced degree of covalency
relative to other corresponding P-A bonds, and has some ion-molecule character.
The lengths these two bonds suggest that P-A bonds have some covalent charac-
ter and P-A bonds have some ion-molecule character in complexes H2 (CH3 )P:F− ,
H2 (CCH)P:F− , H2 FP:CN− , H2 FP:NC− , H2 FP:Cl− , and H2 (NC)P:Cl− , as well as in
the six symmetric complexes (H2 XPX)− .
That the binding energies of the H2YP:X− complexes are related to the nature
of the P-A and P-A bonds can be seen from the data of Table 8.19 by noting that
for a fixed H2YP, as the ability of X to accommodate a negative charge increases
Table 8.20 P-A distances (Å) in complexes H2YP:X−a,b . (Reprinted with permission from Ref.
[39]. Copyright (2014) American Chemical Society)
P-CH3 P-CCH P-F P-CN P-NC P-Cl
X=CH3 2.063
CCH 2.876 2.079
F 2.052 1.915 1.837
CN 3.115 2.662 2.021 2.095
NC 2.975 2.614 2.131 2.357 2.005
Cl 3.302 2.942 2.595 2.722 2.475 2.388
a
A is the atom of X directly bonded to P
b
P-A ion-molecule bonds with some covalent character are indicated in italics
Table 8.21 P-A distances (Å) in complexes H2YP:X−a,b . (Reprinted with permission from Ref.
P-CH3 P-CCH P-F P-CN P-NC P-Cl
X=CH3 2.063
CCH 1.903 2.079
F 1.960 1.944 1.837
CN 1.884 1.848 1.836 2.095
NC 1.880 1.829 1.756 1.886 2.005
Cl 1.881 1.827 1.733 1.874 1.924 2.388
a
A is the atom of Y directly bonded to P
b
Pnicogen P-A bonds with some ion-molecule character are in italics
down a column, the binding energies decrease or are similar with two exceptions,
H2 (CH3 )P:F− and H2 (CCH)P:F− , which have significantly greater binding energies
than the complexes directly above them. In addition, for a given substituent X− ,
binding energies tend to increase from left to right across the row to the symmetric
complex. There are three exceptions: H2 FP:CN− , H2 FP:NC− , and H2 (NC)P:Cl− .
The binding energies of these three complexes are greater than the binding energies
of the complexes to their immediate right. These five complexes are five of the
six complexes identified as having P-A bonds with partial covalent character and
P-A bonds with partial ion-molecule character. Thus, the binding energies of these
complexes reflect the degree of covalent and ion-molecule character of P-A and P-A
bonds, respectively.
Charge-transfer in these anionic complexes occurs from a lone pair on A to the
σ* P-A orbital of H2YP, and the charge-transfer energies are reported in Table 8.22.
No data are reported for the symmetric complexes (H2 XPX)− , or the complexes
H2 FP:NC− and H2 FP:Cl− , since the NBO program considers these as single molec-
ular ions. Neither is a charge-transfer energy reported for H2 FP:CN− since this
Table 8.22 Charge-transfer stabilization energies (kJ·mol−1 ) for complexes H2YP:X−a . (Reprinted
P:X− H2 (CH3 )P H2 (CCH)P H2 FP H2 (CN)P H2 (NC)P
X=CCH 46.4
F 150.3 222.3
b
CN 23.0 86.1
c
NC 15.0 103.2 107.9
c
Cl 16.3 49.4 96.2 230.2
a −
The NBO program considers the (H2 XPX) complexes as single molecular ions
b
The NBO program considers this complex as H2 (CN)P:F−
c
The NBO program considers this complex as a single molecular ion
complex is described as H2 (CN)P:F− . Complexes H2 (CH3 )P:X− and H2 (CCH)P:X−

have the largest charge-transfer energies when X− is F− . For a given X− , the charge-
transfer energies increase across the row from left to right as the P-A bond length
decreases, since charge transfer is more effective at shorter distances. The charge-
transfer energies are linearly related to the P-Cl distances in H2YP:Cl− , with a
correlation coefficient R2 of 0.992.
The electron densities at bond critical points are also indicators of the nature
of the P-A and P-A bonds. Thus, ρBCP values for the P-A bonds in the six sym-
metric complexes (H2 XPX)− , and the P-A bonds in H2 (CH3 )P:F− , H2 (CCH)P:F− ,
H2 FP:CN− , H2 FP:NC− , H2 FP:Cl− , and H2 (NC)P:Cl− have the largest electron den-
sities of 0.044e or greater. The only other P-A bond in this range is the P-N bond in
H2 (CN)P:NC− which has a BCP density of 0.045e. Moreover, for fixed Y, all P-A
bonds in these same complexes have electron densities at the BCPs that are smaller
than the P-A densities for the remaining complexes in the series. The six symmetric
complexes (H2 XPX)− and the complexes H2 (CH3 )P:F− , H2 (CCH)P:F− , H2 FP:CN− ,
H2 FP:NC− , H2 FP:Cl− , and H2 (NC)P:Cl− are those that have ion-molecule P-A
bonds with increased covalent character, and covalent P-A bonds with increased
ion-molecule character.
The Laplacians are negative for the P-A bonds in H2 FP:CN− and in the symmetric
(H2 XPX)− complexes with X=CH3 , CCH, CN, and NC. In addition, these same
bonds and the P-A bonds in (H2 FPF)− , (H2 ClPCl)− , H2 (CH3 )P:F− , H2 (CCH)P:F− ,
H2 FP:NC− , H2 FP:Cl− , and H2 (NC)P:Cl− have values of the energy density that
are also negative and greater in absolute value than 0.01 au, indicating that they
have some covalent character. In addition, for fixed Y, P-A bonds in these same
complexes have less negative energy densities than the other P-A bonds in the same
series, indicating that the former bonds have lost some covalency. Thus, the properties
of the electron densities at bond critical points of P-A and P-A bonds are consistent
with the characterization of these bonds as having partial covalent character and
partial ion-molecule character, respectively, based on binding energies and P-A and
P-A distances.
Table 8.23 Spin-spin

coupling constants 1p J(P-A) Complex 1p
J(P-A) Complex 1
J(P-A )
and 1 J(P-A ) (Hz) for H2YP:Cl − 1p
J(P-Cl) H2 (CH3 )P:X − 1
J(P-C)
complexes H2YP:X− .
(Reprinted with permission Y=CH3 21.0 X=CH3 − 32.9
from Ref. [39]. Copyright CCH 51.5 CCH 11.5
(2014) American Chemical
F 85.2 F 19.4
Society)
CN 70.6 CN 5.3
NC 88.3 NC 5.6
Cl 85.5 Cl 3.6
H2YP:NC− 1p
J(P–N) H2 (CCH)P:X− 1
J(P-C)
Y=CH3 − 26.0 X=CCH − 57.8
CCH − 43.4 F − 87.2
F − 3.8 CN − 46.5
CN − 36.2 NC − 38.0
NC 13.7 Cl − 41.5
H2YP:CN− 1p
J(P-C) H2 FP:X− 1
J(P–F)
Y=CH3 73.3 X=F 164.1
CCH 99.1 CN 137.2
F − 81.0 NC 9.4
CN − 36.9 Cl − 100.8
H2YP:F− 1p
J(P–F) H2 (CN)P:X− 1
J(P-C)
Y=CH3 273.7 X=CN − 36.9
CCH 230.4 NC − 62.5
F 164.1 Cl − 67.2
− 1p − 1
H2YP:CCH J(P-C) H2 (NC)P:X J(P–N)
Y=CH3 95.8 X=NC 13.7
CCH − 57.8 Cl 30.9
H2YP:CH−
3
1p
J(P-C) H2 ClP:X− 1
J(P-Cl)
Y=CH3 − 32.9 X=Cl 85.5
Since the Fermi contact terms are not good approximations to 1p J(P-A) and 1 J(P-

A ) for all of the molecular anions, all coupling constants reported for these anions are
total J values. Coupling constants 1p J(P-Cl) across the pnicogen bonds for complexes
H2YP:Cl− are reported in Table 8.23, and the P-Cl distances are given in Table 8.20.
The symmetric structure (H2 ClP:Cl)− has the shortest P-Cl distance, followed by the
P-Cl distances in H2 (NC)P:Cl− and H2 FP:Cl− . All of these P...Cl bonds have some
covalent character. The values of 1p J(P-Cl) for these three complexes are similar,
and are significantly greater than the values for H2 (CN)P:Cl− , H2 (CCH)P:Cl− , and
H2 (CH3 )P:Cl− . Figure 8.28 illustrates the very good correlation between 1p J(P-Cl)
Fig. 8.28 1p J(P-Cl) versus the P-Cl distance for complexes H2YP:Cl− . Points are identified at the
bottom of the graph by the nature ofY. (Reprinted with permission from Ref. [39]. Copyright (2014)
and the P-Cl distance.The linear relationship shown has a correlation coefficient R2
of 0.961, while a second-order curve gives a slightly better fit with a correlation
coefficient of 0.979.
Coupling constants 1p J(P-N) for complexes H2YP:NC− once again differenti-
ate between P...N pnicogen bonds with and without covalent character. Symmetric
[H2 (NC)P:NC]− has the shortest P-N distance, followed by the P-N distance in
H2 FP:NC− . These two complexes have the smallest absolute values of 1p J(P-N),
while the remaining three complexes have much larger absolute values. Sim-
ilarly, 1p J(P-C) values for H2 FP:CN− and [H2 (CN)P:CN]− are negative, while
H2 (CCH)P:CN− and H2 (CH3 )P:CN− have positive values of 1p J(P-C) at longer P-C
distances. Thus, coupling constants can distinguish between P-A pnicogen bonds
with increased covalent character and shorter bond lengths, and P-A bonds that are
essentially ion-molecule pnicogen bonds with longer P-A distances.
1
J(P-A ) values also differentiate between normal covalent bonds and bonds with
reduced covalency and increased ion-molecule character, as evident from Table 8.23.
Figure 8.29 presents a plot of 1 J(P-C) versus the P-C distance for H2 (CH3 )P:X−
complexes, and dramatically illustrates the effect of the loss of covalency and the
gain of ion-molecule character by the P-C bond.1 J(P-C) for [H2 (CH3 )P:CH3 ]− is –33
Hz, but then increases in absolute value as the P-C distance decreases, and exhibits
its maximum value for H2 (CH3 )P:F− . As this distance further decreases,1 J(P-C)
decreases in H2 (CH3 )P:CCH− , and decreases further as the P-C distance decreases
in the remaining three complexes. The correlation coefficient for the second-order
curve in Fig. 8.29 is 0.996. The variation of 1 J(P-C) would be difficult to understand
without some knowledge of the nature of the P-A bonds and their variation with
distance.1 J(P-C) as a function of the P-C distance for H2 (CCH)P:X− is illustrated
Fig. 8.29 1 J(P-C) versus the P-C distance for complexes H2 (CH3 )P:X− . Points are identified at
the bottom of the graph by the nature of X. (Reprinted with permission from Ref. [39]. Copyright
Fig. 8.30 1 J(P-C) versus the P-C distance for complexes H2 (CCH)P:X− . Points are identified at the
top of the graph by the nature of X. (Reprinted with permission from Ref. [39]. Copyright (2014)
in Fig. 8.30, and shows a similar pattern. Thus, the recognition of the partial ion-
molecule character of longer P-A bonds is essential for understanding the variation
of 1 J(P-A ).
8.3.2 π-σ Pnicogen Bonds
As noted in the previous section, some complexes H2 C=(X)P:PXH2 [35] and

H2 XP:P≡CX [36] are stabilized by σ-σ pnicogen bonds. However, H2 C=PX and
P≡CX contain an unsaturated P-C bond which can also act as a π electron donor to
the σ-hole of the P atom of H2 XP, while the σ system of that same P can function as
a lone-pair donor to the unsaturated molecule through its π-hole. In this section we
discuss properties of complexes stabilized by π-σ pnicogen bonds. Also included in
this section are YN:PO2 X complexes which are stabilized by π-σ pnicogen bonds as
the σ system of NY acts as an electron-pair donor to PO2 X through its π-hole at P.
8.3.2.1 H2 C=(X)P:PXH2
In addition to conformations A and B, a third set of H2 C=(X)P:PXH2 conformation

C complexes exist which are stabilized by π-σ pnicogen bonds [35]. In conformation
C, the H2 C=PX molecule is nearly perpendicular to the plane defined by the P...P
bond and the bisector of the H-Ps -H angle. Pd -Ps -A angles approach linearity, while
Ps -Pd =C angles are less than 90o , dramatically smaller than values in the corre-
sponding conformation A and B complexes. This geometry facilitates interaction
of H2 C=PX through its π system with PH2 X. Since the C=P π bond is polarized
toward C, Ps -Pd =C angles less than 90o promote electron donation by H2 C=PX. The
structures of conformation C complexes H2 C=(F)P:PFH2 and H2 C=(H)P:PH3 are
illustrated in Fig. 8.31.
The binding energies, intermolecular P-P distances, and Pd -Ps -A and Ps -Pd =C an-
gles for conformation C complexes are reported in Table 8.24. Binding energies range
from − 8.2 kJ·mol−1 for H2 C=(H)P:PH3 to − 20.0 kJ·mol−1 for H2 C=(F)P:PFH2 .
The order of binding energies of conformation C complexes is similar to that of the
corresponding conformation A complexes. Conformation C complexes usually have
the largest binding energies, with the order of binding energies being C > A > B with
two exceptions. Although the binding energies of conformation C complexes tend
to be greater than the corresponding A complexes, intermolecular distances in C are
usually longer than those in A. A plot of the binding energy versus the intermolecular
distance for C conformers has a second-order trendline with a correlation coefficient
R2 of 0.890.
Fig. 8.31 Structures of conformation C complexes H2 C=(F)P:PFH2 and H2 C=(H)P:PH3 .

R E Pd -Ps -A Ps -Pd =C
Table 8.24 Binding energies
(E, kJ·mol−1 ), X=F 3.029 − 20.0 173 76
intermolecular P-P distances
Cl 3.146 − 19.6 174 77
(R, Å), and Pd -Ps -A and
Ps -Pd =C angles ( , o ) for OH 3.236 − 17.6 175 77
conformation C complexes
NC 3.281 − 15.2 174 77
H2 C=(X)P:PXH2 . (Reprinted
with permission from Ref. CCH 3.456 − 15.3 173 78
[35]. Copyright (2013) CN 3.528 − 12.1 175 78
CH3 3.484 − 12.8 179 83
H 3.595 − 8.2 175 85
Table 8.25 NBO charges (au) on H2 C=PX, πPd =C→σ*Ps -A and Ps lp→π*Pd =C stabilizing charge-
transfer energies (kJ·mol−1 ), and spin-spin coupling constants [1p J(P-P), Hz] for conformation
C complexes H2 C=(X)P:PXH2 . (Reprinted with permission from Ref. [35]. Copyright (2013)
H2 C=(X)P:PXH2 Charge on H2 C=PX πPd =C→σ*Ps -A Ps lp→π*Pd =C 1p
J(P-P)
X=F 0.010 34.9 16.7 105.1
Cl 0.016 29.3 11.6 125.7
OH 0.006 21.0 9.1 85.3
NC 0.007 18.9 7.8 79.2
CCH − 0.002 8.9 4.6 —a
CN 0.002 8.0 3.3 44.0
CH3 − 0.003 7.9 5.1 45.0
H 0.001 6.8 4.4 40.2
a
Value not available because of computational expense
Table 8.25 reports the NBO charges on H2 C=PX and the stabilizing charge-
transfer energies πPd =C→σ*Ps -A and Ps lp→π*Pd =C. As observed for correspond-
ing conformation A and B complexes, charge transfer from H2 C=PX to PH2 X is
again favored. This is consistent with the positive charges on H2 C=PX in these
complexes, except for H2 C=(CCH)P:P(CCH)H2 and H2 C=(CH3 )P:P(CH3 )H2 which
have small negative charges on H2 C=PX. The charge-transfer energy for a confor-
mation C complex is greater than that for the corresponding conformation B, except
for H2 C=(CN)P:P(CN)H2 .
The electron density properties at the P...P BCPs have similar characteristics to
those observed for conformations A and B. Conformation C complexes with X=F, Cl,
OH, and NC have negative values of HBCP , indicating some covalent character of the
P...P bond. Moreover, there is an excellent correlation between the electron densities
at the P...P BCPs and the P-P distances in A, B, and C complexes. The Laplacian
contours and molecular graph, and the ELF of H2 C=(F)P:PFH2 are illustrated in
Fig. 8.32a and 8.32b, respectively. They present a pictorial representation of the
Fig. 8.32 a The Laplacian contours and molecular graph of H2 C=(F)P:PFH2 conformation C com-
plex. The contour plane is defined by the two P atoms and the BCP of the P...P bond. b The 0.8 au
ELF isosurface of the H2 C=(F)P:PFH2 conformation C complex. (Reprinted with permission from
Ref. [35]. Copyright (2013) American Chemical Society)
Fig. 8.33 Coupling constants 1p J(P-P) versus the P-P distance for conformations C , A , and
B complexes H2 C=(X)P:PXH2 . (Reprinted with permission from Ref. [35]. Copyright (2013)
bond path, and of the Ps lone pair and the π electrons which interact with the π- and
σ-holes, respectively, consistent with the NBO analysis.
Table 8.25 also reports the 31 P-31 P spin-spin coupling constants 1p J(P-P) for
conformation C complexes, and Fig. 8.33 presents a plot of 1p J(P-P) versus the
intermolecular P-P distance for conformers A, B, and C. The usual second-order

trendline has a correlation coefficient R2 of 0.875 for conformation C. However,
a decaying exponential, which has a correlation coefficient R2 of 0.919, provides a
better description of the relationship between these two variables. The ordering of the
trendlines for the three conformations is A > B > C. This ordering reflects the favor-
able A-P...P-A arrangement in A, followed by the A-P...P=C arrangement in B. Since
the pnicogen bond is formed through the π system of H2 C=PX in C, and 1p J(P-P)
is essentially equal to the FC term which depends on s electron densities in ground
and excited states, conformation C complexes have the smallest values of 1p J(P-P) at
each P-P distance.
8.3.2.2 H2 XP:PCX
In addition to the equilibrium conformation A complexes, H2 XP:PCX complexes

may also be stabilized by π-σ bonds in two different conformations B and C [36].
In B complexes the π-σ pnicogen bond has the A-Ps ...C alignment approaching
linearity, while in conformation C the A-Ps ...Pt alignment approaches linearity, as
The bond paths of conformation B complexes connect Ps with PCX through the
π system at the P≡C C atom. The binding energies, P-P and Ps -C distances, and
C-Ps -A angles are reported in Table 8.26. The binding energies of these complexes
range from − 8.7 kJ·mol−1 for H3 P:PCH to − 16.6 kJ·mol−1 for H2 FP:PCF. The Ps -C
distances are always shorter than the corresponding P-P distances, but there is no
correlation between the binding energies and either the P-P or the Ps -C intermolecular
distances.
Conformation B complexes are stabilized by charge-transfer interactions. Since
the P≡C bond is polarized toward C, charge transfer occurs from the π bond at
C through the σ-hole to Ps , and from the lone pair on Ps to Pt through the π-
hole. The dominant charge transfer interaction is πP=C→σ*P-A, as evident from
Table 8.27. The single exception occurs when X=H, in which case charge transfer
πP=C→σ*P-H is 0.2 kJ·mol−1 less stabilizing. The πP=C→σ*P-A charge-transfer
energies vary from 3.6 kJ·mol−1 for X=CH3 , to 19.8 kJ·mol−1 for X=F. Charge-
transfer energies from the lone pair on Ps to the π*P=C orbital range from 1.3
Fig. 8.34 H3 P:PCH complexes with conformations B and C. Complexes have Cs symmetry ex-
cept for H2 (CN)P:PCCN and H2 (CCH)P:PCCCH B which have C1 symmetry. (Reprinted with
permission from Ref. [36]. Copyright (2014) Springer Science and Business Media)

H2 XP:PCH E R(P-P) R(Ps -C) C-Ps -A
(E, kJ·mol−1 ), P-P and Ps -C
distances (R, Å), and C-Ps -A X=Cl − 16.4 3.476 3.326 175
angles ( , deg) for
F − 16.6 3.389 3.070 172
conformation B complexesa .
(Reprinted with permission CCH − 13.3 3.827 3.404 161
from Ref. [36]. Springer OH − 13.2 3.516 3.251 175
Science and Business Media)
NC − 12.0 3.557 3.241 177
CN − 9.6 3.941 3.426 165
CH3 − 12.6 3.850 3.439 177
H − 8.7 4.071 3.413 171
a
These complexes have Cs symmetry, except for
H2 (CN)P:PCCN and H2 (CCH)P:PCCCH which have C1
symmetry
kJ·mol−1 for X=CCH to 10.0 kJ·mol−1 for X=F. The πP=C→σ*P-A charge-transfer
energies do not correlate with the binding energies of conformation B complexes or
with the Ps -C distances, but do correlate with the intermolecular P-P distances, as
Conformation C complexes have bond paths that connect Ps to the π system of
PCX, usually but not always at Pt . Table 8.28 reports the Ps -C distances, which are
shorter than the P-P distances, although the difference between them is much less
than found for conformation B complexes due to the smaller values of the Ps -Pt -C
angles. Table 8.28 also reports the binding energies of these complexes, which range
from − 7.5 kJ·mol−1 for X=H to − 17.6 kJ·mol−1 for X=Cl. The binding energies
of conformation B and C complexes are similar, differing by 1 to 1.5 kJ·mol−1 . The
single exception is conformation C of H2 (CH3 )P:PCCH3 which is 2.6 kJ·mol−1 less
stable than B. The binding energies do not correlate with the P-P distances.
Table 8.27 Charge-transfer energies (kJ·mol−1 ) and spin-spin coupling constants [1p J(P-P), Hz] for
conformation B complexes H2 XPs :Pt ≡CX. (Reprinted with permission from Ref. [36]. Copyright
(2014) Springer Science and Business Media)
H2 XP:PCX πP=C→σ*Ps -A Ps (lp)→π*P=C 1p
J(P-P)
X=Cl 14.4 5.1 41.1
F 19.8 10.0 54.3
CCH 4.3 1.3 16.9
OH 11.5 5.1 41.9
NC 11.0 5.6 31.9
CN 3.7 1.5 14.9
CH3 3.6 3.4 11.9
H 4.2 4.4 7.3
21
A
B
C
14
Charge Transfer
0
3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1
R
Fig. 8.35 Charge-transfer energies from PCX to PH2 X (kJ·mol−1 ) versus the P-P distance (R, Å) for
complexes with conformations A , B , and C . Correlation coefficients R2 are 0.963, 0.992,
and 0.990, respectively. (Reprinted with permission from Ref. [36]. Copyright (2014) Springer

H2 XP:PCH E R(P-P) R(Ps -C) Pt -Ps -A
(E, kJ·mol−1 ), P-P and Ps -C
distances (R, Å), and Pt -Ps -A X=Cl − 17.6 3.323 3.313 174
angles ( , deg) for
F − 15.6 3.298 3.270 175
conformation C complexes.
(Reprinted with permission CCH − 14.7 3.610 3.431 172
from Ref. [36]. Springer OH − 14.2 3.432 3.360 171
NC − 13.5 3.407 3.317 175
CN − 10.6 3.639 3.428 176
CH3 − 10.0 3.747 3.571 166
H − 7.5 3.765 3.702 168
The charge-transfer energies reported in Table 8.29 exhibit a pattern similar to

that observed for conformation B. The πP=C→σ*Ps -A charge-transfer energies
are significantly greater than the Ps (lp)→π*P=C, and are also greater than the
corresponding πP=C→σ*Ps -A energies of conformation B complexes, except for
H2 FP:PCF which has the largest charge-transfer energy among these complexes.
Once again, the πP=C→σ*Ps -A charge-transfer energies correlate with the P-P
Table 8.29 Charge-transfer

energies (kJ·mol−1 ) and H2 XP:PCX πP=C→σ*Ps -A Ps (lp)→π*P=C 1p
J(P-P)
spin-spin coupling constants X=Cl 17.2 5.1 116.7
[1p J(P-P), Hz] for
conformation C complexes F 17.5 6.2 119.7
H2 XPs :Pt ≡CX. (Reprinted CCH 7.3 1.0 49.4
with permission from Ref.
OH 11.8 4.1 83.3
[36]. Copyright (2014)
Springer Science and NC 14.3 2.7 86.8
Business Media) CN 7.4 0.7 48.0
CH3 4.9 0.9 32.6
H 4.8 1.5 34.5
distance, as seen in Fig. 8.35. The net result of charge transfer is to make PH2 X
negatively charged in the complex, decrease the positive charge on Ps , and increase
the positive charge on Pt .
B and C complexes are energetically and structurally similar, since both are sta-
bilized by pnicogen bonds involving Ps of PH2 X and the PCX π system. Transition
structures for interconverting B and C for H3 P:PCH and H2 FP:PCF suggest that the
interconversion occurs via rotation of the PCH or PCF molecules about an axis which
connects Ps to the P=C π bond. This allows the complexes to remain intact during the
interconversion. Relative to the less stable conformer C, the barriers to converting C
to B are 2.6 and 9.7 kJ·mol−1 for H3 P:PCH and H2 FP:PCF, respectively.
Coupling constants for complexes with conformations B and C are given in
Tables 8.27 and 8.29, respectively. Interesting relationships can be found by com-
paring the values of 1p J(P-P) for H2 XP:PCX complexes with conformations A, B,
and C. For fixed X, the order of decreasing 1p J(P-P) is A > C > B. The large values
of 1p J(P-P) for conformation A complexes may be attributed primarily to the σ-σ na-
ture of the pnicogen bond, and the dependence of the dominant FC term on s electron
densities in both ground and excited states. The nature of the FC term is also consis-
tent with the reduced values of 1p J(P-P) for B and C complexes which are stabilized
by π-σ pnicogen bonds involving the π electrons of PCX. That 1p J(P-P) for a given
X is greater for the conformation C complex compared to B is also consistent with
the shorter P-P distances in C, and with the A-Ps -Pt arrangement which approaches
linearity. Figure 8.36 presents plots of 1p J(P-P) versus the P-P distance for A, B,
and C complexes. The good correlation between these two variables is evident, with
the second-order trendlines having correlation coefficients R2 of 0.961, 0.995, and
0.976, respectively.
8.3.2.3 YN:PO2 X
The molecular electrostatic potentials (MEPs) on the 0.001 au electron density iso-
surfaces of the monomers PO2 F and PO2 Cl are very negative around the oxygen
atoms, slightly negative around the halogen atoms, and positive above and below the
210
A
B
C
140
J(P-P), Hz
1p
70
0
3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1
R(P-P), Å
Fig. 8.36 1p J(P-P) versus the P-P distance for complexes with conformationsA, B, and C. (Reprinted
with permission from Ref. [36]. Copyright (2014) Springer Science and Business Media)
phosphorous atoms [33]. The MEP values at the π-holes are 0.099 and 0.078 au for
PO2 F and PO2 Cl, respectively. It is through the π-holes that N bases can donate a
pair of electrons to PO2 F and PO2 Cl to form P...N pnicogen bonds.
The P-N distances and the binding energies of complexesYN:PO2 X forYN=NH3 ,
H2 C=NH, NH2 F, NP, HCN, FCN, NF3 , and N2 , are reported in Table 8.30. There are
large variations among P-N distances, which range from 1.88 to 2.90 Å, and among
binding energies, which range from − 13 to − 149 kJ·mol−1 . Based on these two
properties, the complexes may be subdivided into three groups: the first containing the
strongest bases NH3 , H2 C=NH, and NH2 F; the second containing the sp hybridized
nitrogen bases PN, HCN, and FCN; and the third group made up of complexes with
the weak bases NF3 and N2 . From Table 8.30 it can also be seen that the binding energy
is greater and the P-N distance shorter in YN:PO2 F compared to the corresponding
YN:PO2 Cl, consistent with the MEP values for the two acids. Figure 8.37 presents a
plot of the binding energies of these complexes versus the P-N distance.
There are two interesting structural features of complexes with the stronger bases.
The first is the distortion from planarity of the PO2 X molecules. The degree of
distortion can be measured as the sum of the bond angles around P, which is 360◦
for a planar structure, and 328.4◦ for a tetrahedral structure. When the base is the
strong base H2 C=NH, the sums are 351.7◦ and 352.2◦ for complexes with PO2 Cl and
PO2 F, respectively. In contrast, this sum is 360.0◦ and 359.9◦ , respectively, when
the base is N2 . Moreover, in the complexes with the stronger nitrogen bases, the

(E, kJ·mol−1 ) and P-N PO2 F PO2 Cl
distances (R, Å) for E R E R
complexes YN:PO2 X.
(Reprinted with permission H3 N:PO2 X − 148.5 1.902 − 128.6 1.912
from Ref. [33]. Copyright H2 C=(H)N:PO2 X − 146.2 1.878 − 126.0 1.893
H2 FN:PO2 X − 105.1 1.953 − 90.2 1.962
Society)
PN:PO2 X − 87.4 1.926 − 70.3 1.949
HCN:PO2 X − 60.4 2.042 − 42.8 2.161
FCN:PO2 X − 52.0 2.096 − 36.7 2.295
F3 N:PO2 X − 21.8 2.381 − 16.5 2.646
N2 :PO2 X − 15.4 2.725 − 12.9 2.897
P-N distances are very short, and approach the computed MP2/aug’-cc-pVTZ P-N
distances of 1.63 Å in PO2 NH2 and 1.67 Å in PO2 NC.
The ELF representations illustrated in Fig. 8.38 for the most strongly and weakly
bound complexes H3 N:PO2 F and N2 :PO2 Cl, respectively, clearly illustrate lone pair
donation by the nitrogen base and acceptance through the π-hole of phosphorus.
Consistent with this picture, the AIM analysis of the electron density shows the
presence of one intermolecular BCP and a corresponding bond path connecting the
P and N atoms. The Laplacians at the BCP are always positive, but the total energy
densities are negative except for the four most weakly bound complexes with the
bases NF3 and N2 . The relatively large negative values of HBCP are indicative of the
covalent character of the stronger P...N bonds.
The NBO stabilizing intermolecular charge-transfer energies are reported in
Table 8.31. The evaluation of charge-transfer energies is only possible for com-
plexes with binding energies less than − 71 kJ·mol−1 , since the NBO method treats
complexes with greater binding energies as single molecules, thereby producing
unrealistic charge-transfer energies. In complexes with the weaker bases, charge
transfer occurs from the nitrogen lone pair to the antibonding σ* P-O and P-X orbital
through the π-hole. Thus, PO2 X gains electron density as the nitrogen base loses
electron density.
The spin-spin coupling constants 1p J(P-N) for complexes YN:PO2 F and
YN:PO2 Cl are given in Table 8.31, and plots of 1p J(P-N) versus the P-N distance
in Fig. 8.39 provide further insight into the nature of the N...P pnicogen bonds in
these complexes. The points for each trendline can be grouped in a similar way to the
groupings for distances and binding energies. Although the correlation coefficients of
the trendlines are not good, the trendlines do provide a clear indication of coupling-
constant patterns. At long distances, 1p J(P-N) values increase as the P-N distance
decreases, reach a maximum value at shorter distances, and then decrease as the P-N
distance further decreases. At the shorter distances, 1p J(P-N) approaches the values of
1
J(P-N) for the molecules PO2 NC (-0.3 Hz) and PO2 NH2 (− 26.6 Hz). 1p J(P-N)
values are not useful for estimating P-N distances for these complexes, since two
different distances can have the same value of 1p J(P-N).
150
100
-ΔE (kJ/mol)
50
0
1.8 2.0 2.2 2.4 2.6 2.8 3.0
P···N Distance (Å)
Fig. 8.37 The negative of the binding energy (–E) versus the P-N distance in complexes with
PO2 F (2) and PO2 Cl . The best-fit trendline has the form –E=1/[a + bR(P-N)], with a correla-
tion coefficient R2 =0.94. (Reprinted with permission from Ref. [33]. Copyright (2013) American
Chemical Society)
Fig. 8.38 ELF 0.75 au

isosurfaces of H3 N:PO2 F and
N2 :PO2 Cl illustrating the
interaction between the N
lone pair and the π-hole of P.
Society)
Table 8.31 Total charge transfer (e), charge-transfer stabilization energies [N(lp)→σ*P-Y,
kJ·mol−1 ] and spin-spin coupling constants [1p J(P-N), Hz] for complexes YN:PO2 X. (Reprinted
Complex Total charge transfer Total N(lp)→σ*P-Ya 1p
J(P-N)
X=F X=Cl X=F X=Cl X=F X=Cl
b b
H3 N:PO2 X 0.286 0.296 20.1 28.4
b b
H2 CHN:PO2 X 0.245 0.251 17.0 30.0
b b
FH2 N:PO2 X 0.246 0.244 29.1 38.6
b
PN:PO2 X 0.127 0.127 321.3 48.7 64.7
HCN:PO2 X 0.130 0.107 219.2 159.8 57.6 65.3
FCN:PO2 X 0.117 0.077 188.3 96.4 64.2 54.6
F3 N:PO2 X 0.079 0.043 69.5 32.8 39.9 18.8c
N2 :PO2 X 0.013 0.008 19.0 12.6 12.2 5.6
a
P-Y includes the P-X and two P-O bonds
b
The NBO method treats these complexes as molecules with an intramolecular P-N bond
c
Because of the computational cost, only the FC term has been computed for F3 N:PO2 Cl
Fig. 8.39 1p J(P-N) versus R(P-N) for complexes YN:PO2 F and YN:PO2 Cl . The points for
PO2 NH2 and PO2 NC at short distances have been included with both series. The correlation coef-
ficients R2 are 0.84 and 0.83, respectively. (Reprinted with permission from Ref. [33]. Copyright
8.3.3 Ternary and Higher-Order Complexes with σ -σ Pnicogen

Bonds
Binary complexes stabilized by σ-σ pnicogen bonds may form ternary and higher-
order complexes. In this section we examine the effect of the formation of other
intermolecular bonds on the properties of pnicogen bonds, and the cooperativity of
intermolecular binding energies. Complexes included in this section are the trimers
(PH2 X)3 [31], and pnicogen-bonded complexes that are also hydrogen-bonded,
nFH:(PH2 F)2 and nFH:(H2 FP:NFH2 ), for n=1− 3 [26, 27].
8.3.3.1 (PH2 X)3
The equilibrium pnicogen-bonded trimers (PH2 X)3 , for X=F, Cl, OH, NC, CCH,
CN, CH3 , and H have C3h symmetry, except for [PH2 (CH3 )]3 which has C1 sym-
metry, although the C1 structure is only 0.25 kJ·mol−1 more stable than C3h [31].
The structures and molecular graphs of trimers (PH2 F)3 , [PH2 (OH)]3 and (PH3 )3 are
illustrated in Fig. 8.40. P-P distances, A-P-P angles, and binding energies of these
trimers and the corresponding dimers are reported in Table 8.32. The trimer binding
energies span a large range, from − 17.1 kJ·mol−1 for (PH3 )3 to − 62.9 kJ·mol−1 for
(PH2 F)3, and are necessarily greater than those of the corresponding dimers. How-
ever, the energy per P...P bond is greater in the dimer compared to the corresponding
trimer.
The intermolecular P-P distances in the trimers are significantly longer than the
corresponding dimer distances, and vary from 2.97 Å in (PH2 F)3 to 3.81 Å in (PH3 )3 .
Trimer binding energies exhibit a quadratic dependence on the P-P distance, with a
correlation coefficient of 0.981. The A-P-P angles are between 159 and 171◦ , and
provide a P-P-A alignment which approaches linearity. This alignment facilitates the
stabilizing charge-transfer interactions from the lone pair orbital of one P to the σ*
P-A orbital of the P adjacent to the lone pair.
Fig. 8.40 Molecular graphs of trimers (PH2 F)3 , [PH2 (OH)]3 and (PH3 )3 . Green and red dots indicate
the positions of bond and ring critical points, respectively. The dashed lines connecting the P atoms
and the BCPs are the bond paths. (Reprinted with permission from Ref. [31]. Copyright (2013)
Table 8.32 Binding energies (E, kJ·mol−1 ), intermolecular P-P distances (R, Å) and A-P-P angles
( A-P-P,◦ ) for (PH2 X)3 and (PH2 X)2 . (Reprinted with permission from Ref. [31]. Copyright (2013)
(PH2 X)3 (PH2 X)2
Monomer E R A-P-P E R A-P-P
PH2 F − 62.9 2.974 171 − 34.0 2.471 163
PH2 Cl − 49.6 3.144 171 − 22.1 2.771 167
PH2 (OH) − 43.9 3.229 168 − 20.6 2.851 169
PH2 (NC) − 37.4 3.272 170 − 13.8 3.040 168
PH2 (CCH) − 30.3 3.549 166 − 12.2 3.353 174
PH2 (CN) − 23.7 3.541 168 − 8.4 3.375 171
PH2 (CH3 )a − 18.8 3.738 161 − 8.9 3.481 178
PH3 − 17.1 3.809 159 − 7.0 3.589 179
a
C3h structure
Table 8.33 MBIE components of the interaction energies (kJ·mol−1 ) for trimers with C3h symmetry.
2
ER (1)a ER 2 E(1,2)a E 3 E(1,2,3)
(PH3 F)3 0.90 2.70 − 21.48 − 64.44 − 1.16
(PH2 Cl)3 0.50 1.50 − 16.58 − 49.74 − 1.29
[PH2 (OH)]3 0.24 0.72 − 15.01 − 45.03 + 0.45
[PH2 (NC)]3 0.40 1.20 − 12.06 − 36.18 − 2.38
[PH2 (CCH)]3 0.12 0.36 − 10.12 − 30.36 − 0.29
[PH2 (CN)]3 0.19 0.57 − 7.64 − 22.92 − 1.32
[PH2 (CH3 )]3 0.05 0.15 − 6.13 − 18.39 − 0.53
(PH3 )3 0.05 0.15 − 5.61 − 16.83 − 0.36
a 2 2 2
For these trimers, ER (1)=ER (2)=ER (3), and E(1,2)= E(1,3)= E(2,3)
Trimer total binding energies have been decomposed in terms of many-body inter-
action energies (MBIE)[112, 113] which are reported in Table 8.33. By far, the dom-
inant energy term is the two-body interaction 2 E(1,2). The monomer relaxation en-
ergy ER (1) is destabilizing, while the three-body energy 3 E(1,2,3) is
stabilizing, ex-
cept for [PH2 (OH)]3 . An excellent linear correlation exists between 2 E and E.
In the trimers, charge transfer occurs from the lone pair of one P to the σ* P-A
orbital of the P atom which is adjacent to the lone pair, with the P…P-A alignment
approaching linearity. A linear correlation exists between the binding energies of the
trimers and the charge-transfer energies, with a correlation coefficient R2 =0.958. In
all trimers, electron population of the σ* P-A orbital leads to a lengthening of the
P-A bond relative to the monomer, as evident from the data of Table 8.34.
Table 8.34 P(lp)→σ*P-Aa

Trimers P(lp)→σ*P-A δR(P-A)
charge transfer energies
(kJ·mol−1 ) and changes in (PH3 F)3 50.5 0.006
P-A bond lengths [δR(P-A),
(PH2 Cl)3 30.6 0.009
Å] in trimers (PH2 X)3 .
(Reprinted with permission [PH2 (OH)]3 26.4 0.004
from Ref. [31]. Copyright [PH2 (NC)]3 26.9 0.001
Society) [PH2 (CCH)]3 11.0 0.003
[PH2 (CN)]3 11.8 0.001
[PH2 (CH3 )]3 5.6 0.002
(PH3 )3 5.3 0.002
a
A is the atom of X which is directly bonded
to P, with the P-P-A angle approaching
linearity
The ELF isosurface of each trimer shows the location of the P lone pairs, as
illustrated in Fig. 8.41a for (PH2 F)3 . The MEP on the 0.001 au isosurface of isolated
PH2 F in Fig. 8.41b identifies the lone pair and σ-hole for a single molecule, and their
interactions in the trimer are depicted schematically in Fig. 8.41c. The molecular
graph of (PH2 F)3 in Fig. 8.42 shows the three intermolecular bond critical points
that connect the phosphorous atoms, and a ring critical point. The electron densities
at P...P BCPs correlate exponentially with the P-P distance with an R2 value of
0.9996.The P...P bond paths show a significant curvature, especially in the more
strongly bound complexes. The simultaneous representation of the Laplacians and
bond paths in Fig. 8.42 indicates that each path leads from one P atom where the
concentration of lone pair electron density is high and the Laplacian is negative, to
a b c
Fig. 8.41 a ELF 0.75e isosurface of (PH2 F)3 illustrating the P lone pair basins. b The MEP on
the 0.001 au electron density isosurface of the isolated PH2 F molecule. The location of the σ-hole
maxima (0.060 au) is indicated with a black dot. Color code scale (au): Red > 0.03 > Yellow >
0.015 > Green > 0.0 > Blue. c Schematic representation of the (PH2 F)3. lone pairs (shaded lobes)
and σ-holes (open circles). (Reprinted with permission from Ref. [31]. Copyright (2013) American
Chemical Society)
Fig. 8.42 Laplacian contours

in the symmetry plane and the
molecular graph of (PH2 F)3 .
Negative values of the
Laplacian are indicated with
dashed lines. (Reprinted with
Copyright (2013) American
Chemical Society)
Table 8.35 Computed

Fermi-contact terms and Trimer Computed FC terms Estimated 1p J(P-P)a
estimated spin-spin coupling (PH2 F)3 502.0 503
constants [1p J(P-P), Hz] for
(PH2 Cl)3 384
trimers (PH2 X)3 . (Reprinted
with permission from Ref. [PH2 (OH)]3 331.6 332
[31]. Copyright (2013)
[PH2 (NC)]3 306
[PH2 (CCH)]3 172
[PH2 (CN)]3 175
[PH2 (CH3 )]3 103.6b 106
c
(PH3 )3 86.8 86
a
Estimated from the equation of the FC trendline for the four
trimers
b
C3h structure
c
The computed total 1p J(P-P) is 86.9 Hz
the σ-hole of the adjacent P atom, avoiding the region of electron concentration on the
latter P. The sign of the Laplacians at the BCPs are always positive, while the energy
densities are positive for the more weakly bound trimers but negative for the five
most strongly bound trimers. Thus, some partial covalent character can be ascribed
to the pnicogen bonds in the five trimers with the more electronegative substituents.
Because of the computational cost associated with EOM-CCSD calculations, it
was not possible to compute coupling data for all trimers. Since Fermi contact terms
have been found to be good approximations to P-P coupling constants in pnicogen-
bonded complexes, these terms are reported in Table 8.35 for two strongly bound
trimers (PH2 F)3 and [PH2 (OH)]3 , and the weakly bound trimers [PH2 (CH3 )]3 and
(PH3 )3 . Total 1p J(P-P) was computed for (PH3 )3 . Figure 8.43 provides a plot of the
FC terms for the four trimers and 1p J(P-P) for the corresponding dimers (PH2 X)2
Fig. 8.43 1p J(P-P) total coupling constants versus the P-P distance for dimers (PH2 X)2 , and the
FC terms for four trimers (PH2 X)3 , versus the P-P distance. The points belonging to the trimers
(PH3 )3 , (PH2 F)3 , (PH2 OH)3 ,and (PH2 CH3 )3 are enclosed in red boxes. The solid green trendline
refers to all complexes, while the dashed red trendline refers only to the trimers. (Reprinted with
against the intermolecular P-P distance. The trendline has a correlation coefficient of
0.967, indicating that the trimer points fit very well with the dimer points. However,
the range of 1p J(P-P) for the dimers is very large, and the curvature of the trendline is
certainly influenced by the value of 1p J(P-P) for (PH2 F)2 at its short P-P distance. The
correlation coefficient of the trendline for the four trimers is 1.000, and this trendline
has a reduced curvature. The equation of this trendline has been used to obtain
estimates of 1p J(P-P) for the remaining trimers, and these values are also reported in
Table 8.35.
8.3.3.2 nFH:(PH2 F)2
The (PH2 F)2 dimer can form higher polymeric structures when the F atoms donate
one, two, or three pairs of electrons to FH molecules to form F-H...F hydrogen bonds
in ternary (t), quaternary (q), and penternary (p) nFH:(PH2 F)2 complexes, for n = 1–3
[26]. To assist in the analysis of the binding energies of these complexes, binding
energies of the corresponding nFH:PH2 F binary (b) ternary (t), and quaternary (q)
complexes have also been computed. Possible positions of the FH molecule acting
as a proton donor to PH2 F have been denoted as A, B, C, and D in Scheme 8.1. These
orientations are cis, trans, and gauche at two positions, respectively, to the bisector
of the H-P-H angle. In order to distinguish complexes in which FH molecules are
bonded to the same or the other PH2 F molecule, a prime has been added to the
Scheme 8.1 Positions of FH molecules interacting with FPH2
Fig. 8.44 Structures, designations, and symmetries of complexes nFH:PH2 F. (Reprinted with
position (A’, B’, C’, D’) when the interaction is with the second PH2 F molecule.
For complexes having a different number of FH molecules bonded to the two PH2 F
molecules, (1) indicates the PH2 F molecule with the greater number of hydrogen
bonds, and (2) refers to the molecule with fewer F-H...F bonds. The details of the
structures of these complexes and the optimization challenges encountered are not
reiterated in this review, but can be found in the original reference. Figs. 8.44 and
8.45 provide the names, symmetries, and pictorial descriptions of the complexes
nFH:(PH2 F) and nFH:(PH2 F)2 , respectively.
Table 8.36 shows that formation of F-H...F hydrogen bonds in nFH:(PH2 F)2 com-
plexes leads to a decrease in the P-P distance, and that the P-P distance continues
to decrease as the number of F-H...F hydrogen bonds increases. Complex qCD with
Fig. 8.45 Structures, designations, and symmetries of complexes nFH:(PH2 F)2 . (Reprinted with
two FH molecules bonded to F(1) has the shortest P-P distance among the quaternary
complexes, while pACD with three FH molecules bonded to F(1) has the shortest
P-P distance among the penternary complexes.
The computed P...P binding energies of ternary, quaternary, and penternary com-
plexes are also reported in Table 8.36. The addition of one, two, or three FH molecules
hydrogen-bonded to the same or different F atoms of (PH2 F)2 always increases the
strength of the pnicogen bond, but the number and positions of the FH molecules
determine the extent of that increase. If one FH molecule is bonded to (PH2 F)2 ,
hydrogen bonding at position B is slightly favored. If two FH molecules are bonded
to the same F as in qCD, the binding energy further increases, but if the two FH
molecules are bonded to different F atoms as in qAA and qBB , the binding energies
are similar to but slightly less than the binding energies of tA and tB. If three FH
molecules form hydrogen bonds with the same F atom as in pACD, the energy of the
P...P bond increases dramatically, and is 15 kJ·mol−1 greater than the bond energy
of other penternary complexes. Bonding two FH molecules to F(1) and one FH to
F(2) leads to complexes with binding energies only slightly greater than those of tA
and tB.
Table 8.36 reports the nonadditivities of binding energies (δE), computed in the
standard way given as footnote c. In these complexes the nonadditivities are negative
(synergistic), that is, the binding energy of the complex is greater than the sum of
the binding energies of the two corresponding components. The penternary complex
Table 8.36 P-P distances [R(P-P), Å], changes in P-P distances [δR(P-P), Å], pnicogen bond
energies [E(P...P), kJ·mol−1 ], nonadditivities (δE, kJ·mol−1 ), and spin-spin coupling constants
[1p J(P-P), Hz] for complexes nFH:(PH2 F)2 . (Reprinted with permission from Ref. [26]. Copyright
Complex R(P-P) δR(P-P)a E(P...P)b δEc 1p
J(P-P)
(PH2 F)2 2.471 − 34.0 1008
tA 2.425 − 0.047 − 36.5 − 2.5 1021
tB 2.422 − 0.049 − 37.3 − 3.4 1006
qCDd 2.363 − 0.108 − 44.1 − 4.1 948

qAA (C2 ) 2.401 − 0.070 − 34.7 − 0.8 1080

qAA (Ci ) 2.400 − 0.072 − 34.9 − 0.9 1079
qBB (C2 ) 2.395 − 0.076 − 36.5 − 2.6 1058

qBB (Ci ) 2.395 − 0.076 − 36.5 − 2.6 1058
pACD 2.301 − 0.170 − 54.6 − 20.6 782
pABC 2.353 − 0.118 − 39.6 − 5.6 —e
pCDA 2.356 − 0.115 − 38.7 − 4.8 —e
pCDB 2.354 − 0.117 − 39.3 − 5.4 1072
a
δR(P-P)=R(P-P) for nFH:(PH2 F)2 -R(P-P) for (PH2 F)2
b
E(P...P)=E(complex) – {E(1) (nFH:PH2 F) + E(2) (n FH:PH2 F)}, where n and n indicate the num-
ber of FH molecules bonded to molecules (1) and (2), respectively, with n ≥ n , and n + n ≤ 3
c
δE=E[nFH:(PH2 F)2 ] – i Ei (binary)
d
Not an equilibrium structure on the potential surface
e
EOM–CCSD coupling constants could not be computed for complexes of C1 symmetry
pACD is unique, having a significantly greater nonadditivity than any other complex.
This suggests that the nature of the intermolecular interactions may be changing in
this complex.
The topological analysis of the electron density shows the presence of in-
termolecular bond critical points and corresponding bond paths connecting the
pnicogen-bonded P...P atoms and the hydrogen-bonded H...F atoms. Both the Lapla-
cian of the electron density at the P...P BCP and the total energy density at the BCP
are negative for all complexes except the parent (PH2 F)2 , indicating that these P...P
bonds have covalent character. The degree of covalency can be seen in the variation
of the Laplacians with the intermolecular distance, which is illustrated in Fig. 8.46.
In general, covalency increases as the P-P distance decreases and the number of FH
molecules increases. However, complexes qCD and pACD which have two and three
FH molecules, respectively, hydrogen bonded to the same F(1) atom, have P...P bonds
with significantly greater covalent character than the remaining complexes with the
same number of hydrogen bonds. Values of ∇ 2 ρBCP vary quadratically with the P-P
distance, with a correlation coefficient of 0.986.
Fig. 8.46 The Laplacian of the electron density versus the P-P distance for complexes (PH2 F)2 ,
FH:(PH2 F)2 , 2FH:(PH2 F)2 , and 3FH(PH2 F)2 . (Reprinted with permission from Ref. [26].
The default parameters for the NBO analysis usually describe these complexes
as single molecules with a formal P...P bond. In order to quantify the orbital inter-
action energies, it was necessary to impose a Lewis structure with no P...P bond.
Unfortunately, the charge-transfer energies obtained are unrealistically high. How-
ever, they do show that charge transfer occurs from the PH2 F molecule with fewer
F-H...F hydrogen bonds to the molecule with the greater number of hydrogen bonds.
Moreover, since charge transfer also occurs from F of PH2 F to the F-H molecule
hydrogen-bonded to it, charge-transfer effects will be synergistic in complexes that
have a greater number of F-H...F hydrogen bonds at one of the PH2 F molecules.
31 31
P- P spin-spin coupling constants are also reported in Table 8.36, and are
plotted in Fig. 8.47 as a function of the P-P distance. These data exhibit significant
scatter. In contrast, the data for a subgroup of complexes including the parent and
complexes with one, two or three F-H...F hydrogen bonds at the same F atom, exhibit
a very interesting pattern. The trendline for 1p J(P-P) versus the P-P distance for this
subset decreases quadratically with decreasing distance, with a correlation coefficient
of 0.997.
Some insight into this behavior may be gained by examining the structure of
pACD in detail. In this complex, the P-P distance is short, the P(1)-F(1) distance is
very long, and the three FH molecules hydrogen bonded to F(1) resemble an anionic
cluster 3(FH)F− with approximately local C3v symmetry. This suggests that pACD is
approaching the ion-pair complex H3 F− +
4 : (H2 P:PFH2 ). For comparison, Table 8.37
reports J(P-P) for the cations derived from (PH2 F)2 and pACD, and1 J(P-P) for the
1p
Fig. 8.47 1p J(P-P) vs R(P-P) for complexes nFH:(PH2 F)2 . Complexes with 0, 1, 2, and 3 FH
molecules bonded to the same P-F. (Reprinted with permission from Ref. [26]. Copyright (2012)
Table 8.37 P-P distances [R(P-P), Å] and 31 P-31 P coupling constants [1p J(P-P), Hz] for neutral com-
plexes and corresponding cations (H2 P:PH2 F)+ , and P-P distances and 1 J(P-P) for P2 H4 . (Reprinted
1p
Neutral R(P-P) J(P-P)a Cation R(P-P) 1p
J(P-P)a
+b
(PH2 F)2 2.471 1008 (H2 P:PH2 F) 2.471 160
+b
pACD 2.301 783 (H2 P:PH2 F) 2.301 − 147
1
P2 H4 R(P-P) FC J(P-P)
C2 2.225 − 165 − 115
C2v 2.266 − 230 − 132
C2h 2.238 − 45 + 18
a
Approximated by the FC term
b
At the geometry of the corresponding neutral complex
P2 H4 molecule. These data indicate that the decrease in 1p J(P-P) for qCD and pACD
reflects the increasing covalent character of the P...P bonds and the short P-P distances
that approach the P-P distance in P2 H4 . However, it should be noted that although the
FC term is generally an excellent approximation to 1p J(P-P) in binary complexes, this
may not be the case for pACD. At the short P-P distances in P2 H4 , both the PSO and
SD terms make non-negligible contributions to the total coupling constant1 J(P-P),
which differs significantly from the FC term, as seen from the data of Table 8.37.
A
H
H
D B'
F P
C
D'
N F
B C'
H H
A'
Scheme 8.2 Positions of FH molecules interacting with the binary complex (H2 FP:NFH2 )
8.3.3.3 nFH:H2 FP:NFH2
Scheme 8.2 defines the hydrogen-bonding positions for complexes of FH with

H2 FP:NFH2 [27]. These are denoted as A, B C, and D when hydrogen bonding
occurs at P-F, and A , B , C , and D when hydrogen bonding occurs at N-F. b, t,
q, and p are defined as for the complexes nFH:PH2 F and nFH:(PH2 F)2 . Complexes
nFH:NFH2 and nFH:H2 FP:NFH2 are illustrated in Figs. 8.48 and 8.49, respectively.
Four equilibrium structures FH:(H2 FP:NFH2 ) and four 2FH:(H2 FP:NFH2 ) with
FH molecules hydrogen-bonded to P-F and/or N-F have been found on the potential
surfaces. In addition, one complex 3FH:(H2 FP:NFH2 ) pACD which is structurally
similar to 3FH:(PH2 F)2 pACD, has also been determined. Some complexes with
PH2 F or NH2 F needed for energetic comparisons may not be equilibrium structures,
but are constrained structures.
Table 8.38 reports the P-N distances in complexes nFH:(H2 FP:NFH2 ).
Hydrogen-bond formation at P-F always leads to a decrease of the P-N dis-
tance, with the decrease increasing as the number of FH molecules increases.
In contrast, hydrogen-bond formation at N-F tends to lengthen the P-N bond,
but the changes are smaller than changes in this bond length when hydrogen
bonding occurs at P-F. The complex pACD with three FH molecules hydro-
gen bonded to P-F has the shortest P-N distance of 2.143 Å. However, this
distance is significantly longer than the P-N distance of 1.721 Å computed
at the same level of theory for the H2 P-NH2 molecule. P-F bond distances
increase relative to the parent complex when hydrogen bonding occurs at P-F,
with the degree of lengthening increasing as the number of F-H...F hydrogen bonds
increases. The N-F bond lengthens slightly when hydrogen bonding occurs at N-F.
The P...N binding energies of complexes nFH:(H2 FP:NFH2 ) are also reported in
Table 8.38. Hydrogen-bond formation only at P-F increases the energy of the P...N
bond relative to the parent complex (H2 FP:NFH2 ), and the binding energy increases
as the number of FH molecules increases. In contrast, the energy of the P...N bond
decreases slightly when FH molecules hydrogen bond at N-F. This decrease is not
very sensitive to the number of FH molecules present, as these binding energies are
Fig. 8.48 Structures, designations, and symmetries of complexes FH:NFH2 and 2FH:NFH2 .
all approximately − 23 kJ·mol−1 . The binding energies of the complexes with one
FH molecule hydrogen bonded at P-F and one at N-F are intermediate between the
ternary complexes with one FH at P-F and those with one FH at N-F.
Cooperative effects in complexes nFH:(H2 FP:NFH2 ) which are stabilized by both
pnicogen bonds and hydrogen bonds are reported in Table 8.39. From these data it is
apparent that binding energies are synergistic when hydrogen bond formation occurs
at the P-F bond. However, the synergistic effect for 3FH:(H2 FP:NFH2 ) pACD is
less than that for 2FH:(H2 FP:NFH2 ) qCD, and significantly less than that for the
corresponding complex 3FH:(PH2 F)2 pACD. In contrast, when hydrogen bonding
occurs at N-F, binding energies are diminutive.
Insight into the nonadditivites of binding energies can be gained by examining the
charge-transfer energies in these complexes. However, the problem encountered with
the NBO method in describing charge-transfer energies in nFH:(PH2 F)2 complexes
is also encountered for nFH:(H2 FP:NFH2 ), so Lewis structures with no P...N bonds
were imposed. The resulting orbital interaction energies are much more realistic than
those obtained for the nFH:(PH2 F)2 complexes, and are reported in Table 8.40. In
the parent complex (H2 FP:NFH2 ), both N and P have lone pairs of electrons, so
Fig. 8.49 Structures, designations, and symmetries of complexes nFH:(H2 FP:NFH2 ). (Reprinted
Table 8.38 P-N, P-F, and N-F distances (Å) and P...N binding energies (kJ·mol−1 ) for com-
plexes nFH:(H2 FP:NFH2 ). (Reprinted with permission from Ref. [27]. Copyright (2012) American
Chemical Society)
Structures
Complex FH at P-F Same no. of FH at P-F FH at N-F
and N-F
P-N P-F N-F P-N P-F N-F P-N P-F N-F
H2 FP:NFH2 2.524 1.638 1.420
tA 2.398 1.678 1.415
tB 2.397 1.677 1.416
tB 2.599 1.629 1.435

tC 2.546 1.635 1.434
qCD 2.255 1.732 1.411
qAC 2.401 1.677 1.425

qBC 2.401 1.675 1.427

qC D 2.532 1.634 1.445
pACD 2.143 1.800 1.407
Binding Energiesa
FH at P-F Same no. of FH at P-F FH at N-F
and N-F
H2 FP:NFH2 − 26.7
tA − 34.4
tB − 35.3
tB − 22.6

tC − 23.7
qCD − 45.2
qAC − 30.6

qBC − 31.1
qC D − 23.2
pACD − 56.8
a
E(P...N)=E(complex) – {E(nFH:PH2 F) + E(n FH:NH2 F)}, where n and n indicate the number
of FH molecules bonded to PH2 F and NH2 F, respectively, with n + n ≤ 3
both may act as electron-pair donors. However, since NH2 F is the stronger base,
the stabilizing charge-transfer energy for N(lp)→σ*P-F is significantly greater than
that for P(lp)→σ*N-F. The N(lp)→σ*P-F charge-transfer energy is very sensitive to
the hydrogen-bonding scheme, and in complexes with hydrogen bonding at P-F,
this energy increases as the number of hydrogen bonds increases. In contrast,
N(lp)→σ*P-F charge-transfer energies decrease when hydrogen bonding occurs at
Table 8.39 Cooperativity of

Complex FH at P-F Same no. of FH FH at N-F
binding energies (δE,
at P-F and N-F
kJ·mol−1 )a for complexes
nFH:(H2 FP:NFH2 ). tA − 7.7
(Reprinted with permission tB − 8.6
(2012) American Chemical tB + 4.7
Society)
tC + 3.0
qCD − 12.5
qAC − 4.0

qBC − 4.4
qC D + 9.9
pACD − 9.2
a
δE=E[nFH:(H2 FP:NFH2 )] – i Ei (binary)
N-F. If hydrogen bonding occurs at both P-F and N-F, the N(lp)→σ*P-F charge-
transfer energy is similar to that for complexes with one FH at P-F. By comparison,
N(lp)→σ*P-F charge-transfer energies are not very sensitive to hydrogen bonding.
There are two different charge-transfer interactions in nFH:H2 FP:NFH2 com-
plexes, one across the P...N pnicogen bond and the other across the F-H...F hydrogen
bond. Analysis of the electron densities associated with the molecules PH2 F and
NH2 F indicates that when charge-transfer occurs across the pnicogen bond, not all
of the charge remains on the PH2 F and NH2 F molecules in nFH:(H2 FP:NFH2 ), but
Table 8.40 P(lp)→σ*N-F and N(lp)→σ*P-F charge-transfer energies (kJ·mol−1 ) for

nFH:(H2 FP:NFH2 ) complexes. (Reprinted with permission from Ref. [27]. Copyright (2012)
P (lp)→σ*N-F N (lp)→σ*P-F
Complex FH at P-F Same no. FH FH at N-F FH at P-F Same no. FH FH at N-F
at each at each
H2 FP:NFH2 12.3 58.7
tA 9.1 91.3
tB 9.5 91.4

tB 15.4 43.5
tC 13.9 53.5
qCD 12.0 137.2

qAC 9.2 90.3
qBC 9.9 89.9

qC D 15.9 54.5
pACD 15.3 195.2
Fig. 8.50 Charge-transfer across pnicogen bonds and hydrogen bonds in complexes
FH:(H2 FP:NFH2 ) with hydrogen bonding at P-F and N-F. (Reprinted with permission from Ref.
some charge is transferred to the FH molecules. Moreover, charge-transfer across

hydrogen bonds appears to be the dominant charge-transfer interaction. As a result,
when hydrogen bonding occurs at P-F, the direction of charge flow is that preferred
in the parent complex across the pnicogen bond from NH2 F to PH2 F, and this is in the
same direction as that from PH2 F to FH across hydrogen bonds. These complexes
are then more stable than the parent, and the nonadditivities reported in Table 8.39
are synergistic. In contrast, when hydrogen bonding occurs at N-F, the direction
of charge flow is from NH2 F to FH, which is opposite that preferred in the parent
complex from NH2 F to PH2 F. As a result, the complexes become less stable, and
nonadditivities are antagonistic, or diminutive. These two different charge-transfer
schemes are illustrated in Fig. 8.50.
Table 8.41 provides intermolecular P-N distances and values of 1p J(P-N), and
Fig. 8.51 presents a graphical representation of these data. It is apparent that
like 1p J(P-P) for complexes nFH:(PH2 F)2 , the usual distance dependence of increas-
ing absolute value of 1p J with decreasing distance is not found. However, the parent
complex (H2 FP:NFH2 ) and those with FH molecules hydrogen bonded only at P-F
do exhibit a pattern. The trendline which connects 1p J(P-N) for these complexes has
a correlation coefficient of 0.981. The changes in the P-N distance and 1p J(P-N) in
3FH:(H2 FP:NFH2 ) pACD are reminiscent of changes observed for 3FH:(PH2 F)2
pACD, but are much smaller. While the P-N bond length in 3FH:(H2 FP:NFH)2
approaches the P-N bond length in the Cs and C1 structures of H2 P-NH2 , it is signifi-
cantly greater than the computed P-N bond lengths of 1.766 and 1.721 Å, respectively,
for this molecule. The FC terms for P-N coupling in these structures are 12.9 and
44.5 Hz, respectively. Thus, the FC term does decrease systematically in absolute
value in complexes with 1, 2, and 3 FH molecules at P-F. Its value in pACD appears
to be approaching the value for the covalent bond in H2 P-NH2 , as can be seen in
Fig. 8.51. The decrease in 1p J(P-N) in pACD is a reflection of the changing nature of
the P...N pnicogen bond as it acquires increased covalent character.
Table 8.41 R(P-N) (Å) and 1p J(P-N) (Hz) for complexes nFH:(H2 FP:NFH2 ) with n=1–3. (Reprinted
FH at PH2 F FH at PH2 F and NH2 F FH at NH2 F
1p 1p 1p
Complex R(P-N) J(P-N) R(P-N) J(P-N) R(P-N) J(P-N)
(H2 FP:NFH2 ) 2.524 − 113.5
tA 2.398 − 114.5
tB 2.397 − 115.0

tB 2.599 − 124.4
tC 2.546 − 120.7
qCD 2.255 − 111.4

qAC 2.401 − 117.3
qBC 2.401 − 119.7

qC D 2.532 − 128.2
pACD 2.143 − 99.6
Fig. 8.51 1p J(P-N) vs. the P-N distance for complexes () nFH:(H2 FP:NFH2 ). Complexes with
0, 1, 2, and 3 FH molecules hydrogen bonded at P-F and the H2 P-NH2 molecule at two different
symmetries. The correlation coefficient R2 =0.951. (Reprinted with permission from Ref. [27].
8.4 Summary
The pnicogen bond is a Lewis acid–Lewis base interaction in which a group 15

atom acts as the Lewis acid. Complexes stabilized by pnicogen bonds that have
been investigated in our laboratories and included in this review are the dimers
(PH2 X)2 and (H2 C=PX)2 , the binary complexes H2 C=(X)P:PXH2 , H2 XP:PCX,
H2 XP:NXH2 , X=PH3 :NY, X=PH3 :PY, H2 FP:ClX, and YN:PO2 X, and the molec-
ular anions H2YP:X− , for a variety of substituents X and Y. Trimers (PH2 X)3 , and
ternary and high-order complexes nFH:(PH2 F)2 and nFH:(H2 FP:NFH2 ) for n = 1–
3, have been included as well. Some of the general features of pnicogen-bonded
complexes which have emerged from our studies are summarized below.
A σ-σ pnicogen bond arises between two pnicogen atoms when the first atom
donates a σ lone pair of electrons to the second through its σ-hole, and that atom in
turn donates a σ lone pair of electrons to the first through its σ-hole. If the two pnicogen
atoms are both phosphorus, then the structures of the resulting complexes tend to
have linear A-P...P-A alignments, with A and A the atoms of the substituents X and
Y that are directly bonded to the two phosphorus atoms. This alignment provides for
the overlap of the lone-pair orbital of one P with the σ* P-A orbital of the other, and
facilitates P(lp)→σ*P-A charge-transfer across the pnicogen bond. Charge-transfer
energies usually correlate with the binding energies of these complexes, provided that
there are no other secondary interactions. They may also correlate with the distance
across the pnicogen bond, since charge transfer is more efficient at shorter distances.
Moreover, since different atoms may be bonded to the two P atoms, the identities
of A and A can vary, which gives rise to different stable conformers with the same
molecular formula, but with different binding energies and charge-transfer energies.
σ-σ pnicogen bonds may also form between P and N in H2 XP:NXH2 , between
P and either P or N in X=PH3 :NY and X=PH3 :PY, between P and Cl in complexes
H2 FP:ClX, and even between two N atoms in the complex (NH2 F)2 . In complexes
H2 XP:NXH2 , the preferred direction of charge transfer is N(lp)→σ*P-A. However,
as the binding energies of the H2 XP:NXH2 complexes decrease, the two charge-
transfer energies approach each other. In the two complexes which have the smallest
binding energies, the preferred direction of charge-transfer is P(lp)→σ*N-A. In
complexes X=PH3 :NY and X=PH3 :PY, the only charge-transfer interaction occurs
from the N or P base to the X=PH3 acid. Similarly, in complexes H2 FP:ClX, the
only stabilizing charge-transfer interaction is Cl(lp)→σ*P-F. In the molecular anions
H2YP:X− charge-transfer always occurs from the lone pair onA to the σ* P-A orbital,
with A and A the atoms of X and Y, respectively, which are directly bonded to P.
Molecular anions have relatively large binding energies and charge-transfer energies.
Depending on the nature of X and Y, the P-A bonds in these complexes may have
reduced ion-molecule character and increased covalent character, and the P-A bonds
may have reduced covalent character and increased ion-molecule character. These
bonding characteristics are reflected in other properties of these complexes.
If one of the two molecules containing a pnicogen atom is unsaturated, then a

π-σ pnicogen bond can be formed. In these complexes, the first molecule donates π
electrons to the second through its σ-hole, and that atom in turn donates a σ lone pair
of electrons to the first through its π-hole. In general, complexes with π-σ pnico-
gen bonds are more stable than corresponding complexes with σ-σ pnicogen bonds.
Charge-transfer also stabilizes π-σ pnicogen bonds. In complexes H2 C=(X)P:PXH2 ,
charge transfer from the π P=C bond of H2 C=PX to the σ* P-A orbital of PXH2 is
more stabilizing than charge transfer from the lone pair on P of PXH2 to the π* P=C
orbital. Similarly, for complexes H2 XP:PCX, charge-transfer from the π orbital of
PCX to the σ* P-A orbital of PXH2 is more stabilizing than the reverse transfer. It
appears to be the nature of the σ* P-A orbital which is the dominant factor in deter-
mining the direction of charge transfer. PO2 X:NY and PO2 X:PY are also stabilized
by π-σ pnicogen bonds. Since the molecules PO2 F and PO2 Cl have equilibrium
planar structures, NY and PY act as σ electron-pair donors to P through its π-hole.
In the more strongly bound complexes, PO2 F and PO2 Cl distort from their planar
structures.
A second intermolecular interaction by a pnicogen-bonded binary complex can al-
ter the strength of the pnicogen bond. Pnicogen-bonded trimers (PH2 X)3 have greater
total binding energies than the corresponding dimers, but the binding energies per
pnicogen bond are reduced and the P-P distances are longer in the trimers compared
to the corresponding dimers. The A-P...P alignment approaches linearity, and this
facilitates charge transfer from the lone pair orbital of one P to the σ* P-A orbital of
the P atom which is adjacent to the lone pair. The charge-transfer energies correlate
with the trimer binding energies.
Dimers (PH2 F)2 can also form ternary and higher-order complexes with FH,
nFH:(PH2 F)2 , for n = 1–3, in which the FH molecules are hydrogen bonded to P-F,
forming F-H...F hydrogen bonds. The addition of one, two, or three FH molecules
at the same or different P-F bonds always increases the binding energies, with the
greatest increase occurring in the complex with three FH molecules bonded to the
same P-F bond. In this complex, the P...P bond acquires significant covalent character
as it shortens and approaches the P-P distance in isolated P2 H4 . The pnicogen and
hydrogen bond energies are synergistic. Charge-transfer stabilizes these complexes,
and occurs from the molecule with fewer F-H...F hydrogen bonds to the molecule
which forms the greater number of hydrogen bonds, and from that molecule to the
FH molecules.
In contrast, hydrogen-bond formation may increase or decrease the energy of
the P...N bond in nFH:(H2 FP:NFH2 ) complexes depending on the number of FH
molecules and their interaction sites. Hydrogen bonding at P-F decreases the P-N
distance and increases the energy of the P...N bond, but not nearly to the extent seen
for nFH:(PH2 F)2 complexes. Hydrogen bonding at N-F increases the P-N distance
and decreases the energy of the P...N bond. Thus, binding energies are synergistic
in nFH:(PH2 F)2 but may be either synergistic or diminutive in nFH:(H2 FP:NFH2 )
depending on the hydrogen-bonding scheme. In nFH:(H2 FP:NFH2 ) with hydrogen
bonding at P-F, charge transfer occurs from the lone-pair orbital on N to the σ* P-F
orbital, and then from the lone pair of that F to the σ* F-H orbital. These two charge-
transfer interactions are in the same direction, and have a synergistic effect on the
binding energies. In the nFH:(H2 FP:NFH2 ) complexes with hydrogen bonding at
N-F, the two charge transfers are in opposite directions, from the lone-pair orbital of
N to the σ* P-F orbital of PH2 F, and from the lone pair on F of NH2 F to the σ* F-H
orbital. This leads to a diminutive effect on binding energies.
Spin-spin coupling constants across pnicogen bonds provide further insight into
pnicogen-bonded complexes. In general, 1p J is dominated by the Fermi contact term,
which is a very good approximation to total J. Coupling constants 1p J(P-P) for com-
plexes with σ-σ bonds are greater in absolute value than 1p J(P-P) for corresponding
complexes with π-σ bonds. This may be attributed to the dominance of the Fermi
contact terms which depend on s electron densities in ground and excited states. For
a related series of complexes, coupling constants usually increase in absolute value
as the distance between the pnicogen-bonded atoms decreases. Thus, intermolecular
distances could be extracted from these coupling constants. However, if the char-
acter of the pnicogen bond changes within a series of complexes, then the distance
dependence of 1p J also changes. At long distances, 1p J may increase with decreasing
distance, exhibit a maximum absolute value at an intermediate distance, and then
decrease as the distance further decreases and approaches the covalent bond distance
of a related isolated molecule.
Acknowledgments This work was supported by the Ministerio de Economía y Competitivi-

dad (Project No. CTQ2012–35513-C02–02) and the Comunidad Autónoma de Madrid (Project
MADRISOLAR2, ref. S2009/PPQ1533). The authors are also grateful to the CTI of CSIC and
the Ohio Supercomputer Center for computational support. Without all of this support, the work
reported in this review would not have been possible.
References
1. Heyding RD, Calvert LD (1961) Arsenides of the transition metals. IV. A note on the platinum
metal arsenides. Can J Chem 39:955–957
2. Girolami GS (2009) Origin of the terms pnictogen and pnictide. J Chem Educ 86:1200–1201
3. Cranton GE, Heyding RD (1965) Synthesis in aqueous solutions of binary compounds of the
transition metals with group VA and VIA elements. Can J Chem 43:2027–2032
4. Arunan E, Desiraju GR, Klein RA, Sadlej J, Scheiner S, Alkorta I, Clary DC, Crab-
tree RH, Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ
(2011) Definition of the hydrogen bond (IUPAC Recommendations 2011). Pure Appl Chem
83:1637–1641
5. Desiraju GR, Ho PS, Kloo L, Legon AC, Marquardt R, Metrangolo P, Politzer P, Resnati G,
Rissanen K (2013) Definition of the halogen bond (IUPAC Recommendations 2013). Pure
Appl Chem 85:1711–1713
6. Politzer P, Murray JS, Clark T (2013) Halogen bonding and other σ-hole interactions: a
perspective. Phys Chem Chem Phys 15:11178–11189
7. Mulliken RS, Person WB (1969) Molecular complexes. Wiley, New York
8. Metrangolo P, Resnati G, Pilati T, Biella S (2008) Halogen bonding in crystal engineering.
Struct Bond 126:105–136
9. Zahn S, Frank R, Hey-Hawkins E, Kirchner B (2011) Pnicogen bonds: a new molecular

linker? Chem Eur J 17:6034–6038
10. Klinkhammer KW, Pyykko P (1995) Ab Initio Interpretation of the closed-shell intermolecular
E...E attraction in dipnicogen (H2 E-EH2 )2 and dichalcogen (HE-EH)2 hydride model dimers.
Inorg Chem 34:4134–4138
11. Carré F, Chuit C, Corriu RJP, Monforte P, Nayyar NK, Reyé C (1995) Intramolecular coordi-
nation at phosphorus: donor-acceptor interaction in three- and four-coordinated phosphorus
compounds. J Organomet Chem 499:147–154
12. Galasso V (2004) Theoretical study of the structure and bonding in phosphatrane molecules.
J Phys Chem A 108:4497–4504
13. Murray JS, Lane P, Politzer P (2007) A predicted new type of directional noncovalent
interaction. Int J Quantum Chem 107:2286–2292
14. Ganesamoorthy C, Balakrishna MS, Mague JT, Tuononen HM (2008) New tetraphosphane
ligands {(X2 P)2 NC6 H4 N(PX2 )2 } (X=Cl, F, OMe, OC6 H4 OMe-o): synthesis, derivatiza-
tion, group 10 and 11 metal complexes and catalytic investigations. DFT calculations on
intermolecular P...P interactions in halo-phosphines. Inorg Chem 47:7035–7047
15. Moilanen J, Ganesamoorthy C, Balakrishna MS, Tuononen HM (2009) Weak interactions
between trivalent pnictogen centers: computational analysis of bonding in dimers X3 E...EX3
(E=Pnictogen, X=Halogen). Inorg Chem 48:6740–6747
16. Scheiner S (2011) A new noncovalent force: comparison of P...N interaction with hydrogen
and halogen bonds. J Chem Phys 134:094315
17. Bauer S, Tschirschwitz S, Lönnecke P, Frank R, Kirchner B, Clarke ML, Hey-Hawkins E
(2009) Enantiomerically pure bis(phosphanyl)carbaborane(12) compounds. Eur J Inorg Chem
2009:2776–2788
18. Tschirschwitz S, Lonnecke P, Hey-Hawkins E (2007) Aminoalkylferrocenyldichlorophos-
phanes: facile synthesis of versatile chiral starting materials. Dalton Trans 1377–1382
19. Sundberg MR, Uggla R, Vias C, Teixidor F, Paavola S, Kivekäs R (2007) Nature of in-
tramolecular interactions in hypercoordinate C-substituted 1,2-dicarba-cloro-dodecaboranes
with short P...P distances. Inorg Chem Commun 10:713–716
20. LaBarge MS, Andrews AM, Taleb-Bendiab A, Hillig KW, Kuczkowski RL, Bohn RK (1991)
Microwave spectrum, structure, and dipole moment of the phosphorus trifluoride-water
complex. J Phys Chem 95:3523–3527
21. Politzer P, Murray J, Janjić G, Zarić S (2014) σ-Hole interactions of covalently-bonded
nitrogen, phosphorus and arsenic: a survey of crystal structures. Crystals 4:12–31
22. Del Bene JE, Alkorta I, Sanchez-Sanz G, Elguero J (2011) 31 P–31 P spin–spin coupling
constants for pnicogen homodimers. Chem Phys Lett 512:184–187
23. Del Bene JE, Alkorta I, Sanchez-Sanz G, Elguero J (2011) Structures, energies, bonding, and
NMR properties of pnicogen Complexes H2 XP:NXH2 (X=H, CH3 , NH2 , OH, F, Cl). J Phys
Chem A 115:13724–13731
24. Del Bene JE, Alkorta I, Sanchez-Sanz G, Elguero J (2012) Structures, binding energies, and
spin–spin coupling constants of geometric isomers of pnicogen homodimers (PHFX)2 , X=F,
Cl, CN, CH3 , NC. J Phys Chem A 116:3056–3060
25. Del Bene JE, Sanchez-Sanz G, Alkorta I, Elguero J (2012) Homo- and heterochiral dimers
(PHFX)2 , X=Cl, CN, CH3 , NC: to what extent do they differ? Chem Phys Lett 538:14–18
26. Alkorta I, Sánchez-Sanz G, Elguero J, Del Bene JE (2012) Influence of hydrogen bonds on
the P...P pnicogen bond. J Chem Theory Comput 8:2320–2327
27. Del Bene JE, Alkorta I, Sánchez-Sanz G, Elguero J (2012) Interplay of F–H...F hydrogen
bonds and P...N pnicogen bonds. J Phys Chem A 116:9205–9213
28. Blanco F, Alkorta I, Rozas I, Solimannejad M, Elguero J (2011) A theoretical study of the
interactions of NF3 with neutral ambidentate electron donor and acceptor molecules. Phys
Chem Chem Phys 13:674–683
29. Alkorta I, Sánchez-Sanz G, Elguero J, Del Bene JE (2013) Exploring (NH2 F)2 , H2 FP:NFH2 ,
and (PH2 F)2 potential surfaces: hydrogen bonds or pnicogen bonds? J Phys Chem A 117:183–
191
30. Del Bene JE, Alkorta I, Sánchez-Sanz G, Elguero J (2013) Phosphorus as a simultaneous
electron-pair acceptor in intermolecular P...N pnicogen bonds and electron-pair donor to lewis
acids. J Phys Chem A 117:3133–3141
31. Alkorta I, Elguero J, Del Bene JE (2013) Pnicogen-bonded cyclic trimers (PH2 X)3 with X=F,
Cl, OH, NC, CN, CH3 , H, and BH2 . J Phys Chem A 117:4981–4987
32. Sánchez-Sanz G, Alkorta I, Trujillo C, Elguero J (2013) Intramolecular pnicogen interactions
in PHF-(CH2 )n -PHF (n=2–6) systems. ChemPhysChem 14:1656–1665
33. Alkorta I, Elguero J, Del Bene JE (2013) Pnicogen bonded complexes of PO2 X (X=F, Cl)
with nitrogen BASES. J Phys Chem A 117:10497–10503
34. Del Bene JE, Alkorta I, Elguero J (2013) Characterizing complexes with pnicogen bonds
involving sp2 hybridized phosphorus atoms: (H2 C=PX)2 with X=F, Cl, OH, CN, NC, CCH,
H, CH3 , and BH2 . J Phys Chem A 117:6893–6903
35. Del Bene JE, Alkorta I, Elguero J (2013) Properties of complexes H2 C=(X)P:PXH2 , for X=F,
Cl, OH, CN, NC, CCH, H, CH3 , and BH2 : P...P pnicogen bonding at σ-holes and π-holes. J
Phys Chem A 117:11592–11604
36. Del Bene J,Alkorta I, Elguero J (2014) σ–σ and σ–π pnicogen bonds in complexes H2 XP:PCX,
for X=F, Cl, OH, NC, CN, CCH, CH3 , and H. Theor Chem Acc 133:1–9
37. Alkorta I, Sánchez-Sanz G, Elguero J, Del Bene JE (2014) Pnicogen bonds between X=PH3
(X=O, S, NH, CH2 ) and phosphorus and nitrogen bases. J Phys Chem A 118:1527–1537
38. Del Bene JE, Alkorta I, Elguero J (2014) Influence of substituent effects on the formation of
P...Cl pnicogen bonds or halogen bonds. J Phys Chem A 118:2360–2366
39. Del Bene JE, Alkorta I, Elguero J (2014) Pnicogen-bonded anionic complexes. J Phys Chem
A 118:3386–3392
40. Azofra LM, Alkorta I, Elguero J (2014) Chiral discrimination in dimers of diphosphines (PH2 -
PH2 and PH2 -PHF). ChemPhysChem in press 15:33663–3670. doi:10.1002/cphc.201402086
41. Sánchez-Sanz G, Trujillo C, Alkorta I, Elguero J (2014) Intramolecular pnicogen interactions
in phosphorus and arsenic analogues of proton sponges. Phys Chem Chem Phys in press
16:15900–15909. doi:10.1039/C4CP01072H
42. Grabowski SJ, Alkorta I, Elguero J (2013) Complexes between dihydrogen and amine, phos-
phine, and arsine derivatives. Hydrogen bond versus pnictogen interaction. J Phys Chem A
117:3243–3251
43. Alkorta I, Elguero J, Solimannejad M (2014) Single electron pnicogen bonded complexes. J
Phys Chem A 118:947–953
44. Bauzá A, Alkorta I, Frontera A, Elguero J (2013) On the reliability of pure and hybrid DFT
methods for the evaluation of halogen, chalcogen, and pnicogen bonds involving anionic and
neutral electron donors. J Chem Theory Comput 9:5201–5210
45. Solimannejad M, Gharabaghi M, Scheiner S (2011) SH...N and SH...P blue-shifting H-bonds
and N...P interactions in complexes pairing HSN with amines and phosphines. J Chem Phys
134:024312
46. Scheiner S (2011) Effects of substituents upon the P...N noncovalent interaction: the limits of
Its STRENGTH. J Phys Chem A 115:11202–11209
47. Adhikari U, Scheiner S (2012) Substituent effects on Cl...N, S...N, and P...N noncovalent
bonds. J Phys Chem A 116:3487–3497
48. Adhikari U, Scheiner S (2012) Sensitivity of pnicogen, chalcogen, halogen and H-bonds to
angular distortions. Chem Phys Lett 532:31–35
49. Scheiner S (2011) Can two trivalent N atoms engage in a direct N...N noncovalent interaction?
50. Adhikari U, Scheiner S (2011) Comparison of P...D (D=P, N) with other noncovalent bonds
in molecular aggregates. J Chem Phys 135:184306
51. Scheiner S (2011) Weak H-bonds. Comparisons of CHO to NHO in proteins and PHN to
direct PN interactions. Phys Chem Chem Phys 13:13860–13872
52. Scheiner S (2011) Effects of multiple substitution upon the P...N noncovalent interaction.
Chem Phys 387:79–84
53. Scheiner S (2012) Extrapolation to the complete basis set limit for binding energies of
noncovalent interactions. Comp Theor Chem 998:9–13
54. Scheiner S (2011) On the properties of X...N noncovalent interactions for first-, second-, and
third-row X atoms. J Chem Phys 134:164313
55. Scheiner S (2013) Sensitivity of noncovalent bonds to intermolecular separation: hydrogen,
halogen, chalcogen, and pnicogen bonds. CrystEngComm 15:3119–3124
56. Scheiner S (2012) The pnicogen bond: its relation to hydrogen, halogen, and other noncovalent
57. Scheiner S (2013) Detailed comparison of the pnicogen bond with chalcogen, halogen, and
hydrogen bonds. Int J Quantum Chem 113:1609–1620
58. Bauza A, Quinonero D, Deya PM, Frontera A (2012) Pnicogen-π complexes: theoretical study
and biological implications. Phys Chem Chem Phys 14:14061–14066
59. Bauzá A, Ramis R, Frontera A (2014) A combined theoretical and cambridge structural
database study of π-hole pnicogen bonding complexes between electron rich molecules and
both nitro compounds and inorganic bromides (YO2 Br, Y=N, P, and As). J Phys Chem A
118:2827–2834
60. Bauza A, Quinonero D, Deya PM, Frontera A (2013) Halogen bonding versus chalcogen
and pnicogen bonding: a combined Cambridge structural database and theoretical study.
CrystEngComm 15:3137–3144
61. Solimannejad M, Gholipour A (2013) Revealing substituent effects on the concerted interac-
tion of pnicogen, chalcogen, and halogen bonds in substituted s-triazine ring. Struct Chem
24:17051711
62. Solimannejad M, Bayati E, Esrafili MD (2014) Enhancement effect of lithium bonding on the
strength of pnicogen bonds: XH2 P...NCLi...NCY as a working model (X=F, Cl; Y=H, F, Cl,
CN). Mol Phys 112:2058–2062
63. Solimannejad M, Ramezani V, Trujillo C, Alkorta I, Sánchez-Sanz G, Elguero J (2012) Com-
petition and interplay between σ-Hole and π-Hole interactions: a computational study of 1:1
and 1:2 complexes of nitryl halides (O2 NX) with Ammonia. J Phys Chem A 116:5199–5206
64. Grabowski SJ (2013) σ-Hole bond versus hydrogen bond: from tetravalent to pentavalent N,
P, and As atoms. Chem Eur J 19:14600–14611
65. Grabowski SJ (2014) Clusters of ammonium cation-hydrogen bond versus σ-hole bond.
66. Li Q-Z, Li R, Liu X-F, Li W-Z, Cheng J-B (2012) Pnicogen–hydride interaction between FH2 X
(X=P and As) and HM (M=ZnH, BeH, MgH, Li, and Na). J Phys Chem A 116:2547–2553
67. Li Q-Z, Li R, Liu X-F, Li W-Z, Cheng J-B (2012) Concerted interaction between pnicogen and
halogen bonds in XCl-FH2 P-NH3 (X=F, OH, CN, NC, and FCC). ChemPhysChem 13:1205–
1212
68. An X-L, Li R, Li Q-Z, Liu X-F, Li W-Z, Cheng J-B (2012) Substitution, cooperative, and
solvent effects on π pnicogen bonds in the FH(2)P and FH(2)As complexes. J Mol Model
18:4325–4332
69. Li Q, Zhuo H,Yang X, Cheng J, Li W, Loffredo RE (2014) Cooperative and diminutive effects
of pnicogen bonds and cation–π interactions. ChemPhysChem 15:500–506
70. Liu X, Cheng J, Li Q, Li W (2013) Competition of hydrogen, halogen, and pnicogen bonds in
the complexes of HArF with XH2 P (X=F, Cl, and Br). Spectrochim Acta Part A Mol Biomol
Spectros 101:172–177
71. ChenY,Yao L, Lin X (2014) Theoretical study of (FH2 X)n ·Y (X=P and As, n=1–4, Y=F− , Cl− ,
Br− , I− , NO− 2−
3 and SO4 ): the possibility of anion recognition based on pnicogen bonding.
Comput Theor Chem 1036:44–50
72. Guan L, Mo Y (2014) Electron transfer in pnicogen bonds. J Phys Chem A 118:8911–8921.
73. Eskandari K, Mahmoodabadi N (2013) Pnicogen bonds: a theoretical study based on the
laplacian of electron density. J Phys Chem A 117:13018–13024
74. Pople JA, Binkley JS, Seeger R (1976) Theoretical models incorporating electron correlation.
Int J Quantum Chem Quantum Chem Symp 10:1–19
75. Krishnan R, Pople JA (1978) Approximate fourth-order perturbation theory of the electron
correlation energy. Int J Quantum Chem 14:91–100
76. Bartlett RJ, Silver DM (1975) Many–body perturbation theory applied to electron pair corre-
lation energies. I. closedâŁłshell first–row diatomic hydrides. J Chem Phys 62:3258–3268
77. Bartlett RJ, Purvis GD (1978) Many–body perturbation theory, coupled-pair many-electron
theory, and the importance of quadruple excitations for the correlation problem. Int J Quantum
Chem 14:561–581
78. Del Bene JE (1993) Proton affinities of ammonia, water, and hydrogen fluoride and their
anions: a quest for the basis-set limit using the dunning augmented correlation-consistent
basis sets. J Phys Chem 97:107–110
79. Dunning TH (1989) Gaussian basis sets for use in correlated molecular calculations. I. The
atoms boron through neon and hydrogen. J Chem Phys 90:1007–1023
80. Woon DE, Dunning TH (1995) Gaussian basis sets for use in correlated molecular calculations.
V. Core–valence basis sets for boron through neon. J Chem Phys 103:4572–4585
81. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman, JR, Scalmani G,
Barone V, Mennucci B, Petersson GA et al (2009) Gaussian, Inc.: Wallingford CT,. Gaussian-
09, Revision A.01
91:893–928
83. Bader RFW (1990) Atoms in molecules, A quantum theory. Oxford University, Oxford
84. Popelier PLA (2000) Atoms In molecules. An introduction. Prentice Hall, Harlow, England
85. Matta CF, Boyd RJ (2007) The quantum theory of atoms in molecules: from solid state to
DNA and drug design. Wiley-VCH, Weinheim
86. Silvi B, Savin A (1994) Classification of chemical bonds based on topological analysis of
electron localization functions. Nature 371:683–686
87. Todd A, Keith TK (2011) Gristmill Software, Overland Park KS, USA.
http://aim.tkgristmill.com. Accessed 1 Aug 2013 (AIMAll (Version 11.08.23))
88. Noury S, Krokidis X, Fuster F, Silvi B (1997) TopMod Package
89. Rozas I, Alkorta I, Elguero J (2000) Behavior of ylides containing N, O, and C atoms as
hydrogen bond acceptors. J Am Chem Soc 122:11154–11161
91. Glendening ED, Badenhoop JK, ReedAE, Carpenter JE, Bohmann JA, Morales CM, Weinhold
F (2001) NBO 5.0. University of Wisconsin Press, Madison
92. Glendening ED, Badenhoop JK, Reed AE, Carpenter JE, Bohmann JA, Morales CM, Landis
CR, Weinhold F (2013) NBO 6.0. University of Wisconsin Press. Madison
93. Becke AD (1993) DensityâŁłfunctional thermochemistry. III. The role of exact exchange. J
Chem Phys 98:5648–5652
94. Lee C, Yang W, Parr RG (1988) Development of the colle-salvetti correlation-energy formula
into a functional of the electron density. Phys Rev B 37:785–789
95. Jmol: an open-source Java viewer for chemical structures in 3D, version 13.0.
http://www.jmol.org/. Accessed 26 Sept 2013
96. Patek M “Jmol NBO Visualization Helper” program. http://www.marcelpatek.com/nbo/
nbo.html.Accessed 26 Sept 2013
97. Perera SA, Nooijen M, Bartlett RJ (1996) Electron correlation effects on the theoreti-
cal calculation of nuclear magnetic resonance spin–spin coupling constants. J Chem Phys
104:3290–3305
98. Perera SA, Sekino H, Bartlett RJ (1994) Coupled–cluster calculations of indirect nuclear
coupling constants: the importance of non–fermi contact contributions. J Chem Phys
101:2186–2196
99. Schäfer A, Horn H, Ahlrichs R (1992) Fully optimized contracted gaussian basis sets for
atoms Li to Kr. J Chem Phys 97:2571–2577
100. Stanton JF, Gauss J, Watts JD, Nooijen M, Oliphant N, Perera SA, Szalay PS, Lauderdale WJ,
Gwaltney SR, Beck S, Balkova A, Bernholdt DE, Baeck KK, Tozyczko P, Sekino H, Huber
C, Bartlett RJ ACES II is a program product of the Quantum Theory Project, University of

Florida. Integral packages included are VMOL (Almlöf J, Taylor PR); VPROPS (Taylor PR);
ABACUS (Helgaker T, Jensen Aa HJ, Jørgensen P, Olsen J, Taylor PR). Brillouin-Wigner
perturbation theory was implement by Pittner J
101. Knop O, Boyd RJ, Choi SC (1988) Sulfur-sulfur bond lengths, or can a bond length be
estimated from a single parameter? J Am Chem Soc 110:7299–7301
102. Gibbs GV, Hill FC, Boisen MB, Downs RT (1998) Power law relationships between bond
length, bond strength and electron density distributions. Phys Chem Miner 25:585–590
103. Alkorta I, Barrios L, Rozas I, Elguero J (2000) Comparison of models to correlate electron
density at the bond critical point and bond distance. Theochem 496:131–137
104. Espinosa E, Alkorta I, Elguero J, Molins E (2002) From weak to strong interactions: a compre-
hensive analysis of the topological and energetic properties of the electron density distribution
involving H...F systems. Part I: the transit region between pure closed-shell and shared-shell
interactions. J Chem Phys 117:5529–5542
105. Alkorta I, Elguero J (2004) Fluorine-fluorine interactions: a NMR and AIM analysis. Struct
Chem 15:117–120
106. Tang TH, Deretey E, Knak Jensen SJ, Csizmadia IG (2006) Hydrogen bonds: relation between
lengths and electron densities at bond critical points. Eur Phys J D 37:217–222
107. Vener MV, Manaev AV, Egorova AN, Tsirelson VG (2007) QTAIM study of strong H-bonds
with the O–H...A fragment (A=O, N) in three-dimensional periodical crystals. J Phys Chem
A 111:1155–1162
108. Castillo N, Robertson KN, Choi SC, Boyd RJ, Knop O (2008) Bond length and the electron
density at the bond critical point: X-X, Z-Z, and C-Z bonds (X=Li-F, Z=Na-Cl). J Comput
Chem 29:367–379
109. Mata I, Alkorta I, Molins E, Espinosa E (2010) Universal features of the electron density
distribution in hydrogen-bonding regions: a comprehensive study involving H...X (X=H, C,
N, O, F, S, Cl, π) interactions. Chem Eur J 16:2442–2452
110. Zeng Y, Li X, Zhang X, Zheng S, Meng L (2011) Insight into the Nature of the interactions
of furan and thiophene with hydrogen halides and lithium halides: Ab INITIO and QTAIM
studies. J Mol Model 17:2907–2918
111. Cremer D, Kraka E (1984) A description of the chemical bond in terms of local properties of
electron density and energy. Croat Chem Acta 57:1259–1281
112. Hankins D, Moskowitz JW, Stillinger FH (1970) Water molecule interactions. J Chem Phys
53:4544–4554
113. Xantheas S (1994) Ab Initio studies of cyclic water clusters (H2 O)N , N=1–6. II. Analysis of
many–body interactions. J Chem Phys 100:7523–7534
Chapter 9
Chalcogen Bonds in Protein Architecture
Michio Iwaoka
Abstract Nonbonded interactions between a divalent sulfur atom and polar func-
tional groups, i.e., S· · · X (X = O, N, and S) interactions, have recently been
demonstrated to stabilize protein structures to some extent and play putative roles in
their function and evolution, thanks to the interplay of statistical analysis of protein
structure databases and theoretical calculation using simple molecular cluster mod-
els. The directional features observed between the interacting S and X atoms suggest
that they can also be called S· · · X chalcogen bonds in analogy to halogen bonds.
While the existence of chalcogen bonds in proteins may be accepted today by the pro-
tein scientist community, there are still several issues that should be addressed clearly
before going forward to applications of the interactions to protein engineering and
the ligand design. Herein, the current status of the research on the S· · · X chalcogen
bonds in proteins is reviewed from several points of view, including the historical
aspects and the analytical methods for mining chalcogen bonds in protein struc-
tures. Statistical, directional, and energetic features of four typical S· · · X chalcogen
bonds in proteins are presented. Possibility of analogous Se· · · X chalcogen bonds
in selenoproteins is also pointed out. Implications of S· · · X chalcogen bonds in the
structural control and in the function and evolution of some particular proteins are
subsequently summarized from the recent literature. Finally, it is proposed that the
concept of S· · · X chalcogen bonds will be a useful tool for fully understanding not
only protein structures but also the biological aspects. Thus, the chalcogen bonds,
though their interaction energy is smaller than that of classical hydrogen bonds, can
be an element of weak noncovalent interactions that control chemical and biological
properties of protein architecture.
9.1 Introduction
Sulfur is involved in almost all proteins as cysteine (Cys) and methionine (Met)
amino acid residues. In particular, it is contained at high concentrations in keratin
[1, 2], a major component of wool, and metallothionein [3], a representative metal-
M. Iwaoka ()
Department of Chemistry, School of Science,
Tokai University, Kitakaname, Hiratsuka-shi, Kanagawa 259-1292, Japan
e-mail: miwaoka@tokai.ac.jp
266 M. Iwaoka
loprotein capable of binding many zinc, cadmium, and copper atoms per one pro-
tein molecule. Roles of sulfur atoms are diverse in proteins. For example, the sulfur
atoms of Cys and Met residues can bind the prosthetic groups of several enzymes,
like iron-sulfur clusters of ferredoxins [4] and zinc ions of zinc fingers [5], as soft
ligands. The sulfanyl (SH) group of a Cys residue is sometimes utilized as an ac-
tive center of oxidoreductases, such as thioredoxin (Trx) [6] and protein disulfide
isomerase (PDI) [7, 8, 9], and cysteine proteases, such as papain and caspase [10].
Besides these roles, the disulfide (SS) bonds that are formed between two SH groups
of Cys residues effectively stabilize protein structures [11]. Thus, sulfur atoms in
proteins are important structural and functional elements for protein architecture. In
the most cases, the sulfur atoms of Cys residues play more significant roles than
those of Met residues. This is in accord with the observation that Cys residues are
strongly conserved in the amino acid sequence during the evolution [11]. Indeed,
site-directed mutation of Cys to other amino acids or the deletion should result in
serious damage on the structure and function of a protein, while the consequence of
similar modification on the Met residues would be much more limited.
Most sulfur atoms of proteins are present in a divalent state, having two cova-
lent bonds and two lone pairs, and are contained in a SS bond formed between Cys
residues or a methylsulfanyl group (MeS) of Met residues. These sulfur-containing
functional groups, i.e., SS and MeS groups, are ubiquitous in proteins, but apart
from the aforementioned distinct roles, they were usually considered as hydropho-
bic moieties, showing no apparent interaction with nearby polar functional groups.
However, with accumulation of precise structural data of proteins [12], existence of
weak noncovalent interactions around the divalent sulfur atoms has become evident
thanks to the interplay of elaborate structural database analysis and sophisticated the-
oretical calculation. These interactions in protein architecture contains NH· · · S and
OH· · · S hydrogen bonds [13, 14], S· · · π interactions [15, 16], and S· · · X (X = O,
N, and S) interactions [15, 17–19]. The last two can be categorized to “non-classical”
interactions in the sense that the divalent sulfur atom is formally accepting electron
donation from the π system or the X atom to the σ-hole, adapting a pseudo hyperva-
lent coordination state [20]. The S· · · X interactions that have pertinent directional
propensity can also be called “chalcogen bonds” [21] in analogy to halogen bonds
[22–24]. This contemporary terminology is adapted throughout this chapter, but the
term, S· · · X interaction, is also used to represent the interaction that has not yet
been characterized by the directional propensity, and the term, chalcogen bond, is
intentionally used for the interactions that have distinct directionalities.
While the concept of chalcogen bonds in proteins may be accepted today by the
protein scientist community, there are still several issues that should be addressed
clearly before going forward to applications of the interactions to protein engineering
and the ligand design. The following questions would be particularly important to
be understood with consensus.
Q1. Do the S· · · X chalcogen bonds really exist in proteins?
Q2. Are the directional propensities similar to those observed for small organic
molecules?
9 Chalcogen Bonds in Protein Architecture 267
Q3. How do the potential surfaces look like?

Q4. Can the S· · · X chalcogen bonds control protein structure?
Q5. Are the S· · · X chalcogen bonds relevant to protein function and evolution?
Q6. Are there similar Se· · · X chalcogen bonds in selenoproteins too?
In this chapter, to provide useful information as to above questions, the current status
of the research on the S· · · X chalcogen bonds in proteins is reviewed from several
points of view, based mostly on the research carried out in the author’s laboratory
since 2001. First, the historical aspects including the discovery are overviewed briefly
(Sect. 9.2). Then, the analytical methods for mining chalcogen bonds in protein
structures are introduced (Sect. 9.3). Nonbonded S· · · X interactions found in proteins
are subsequently discussed in detail based on the probabilities, the directionalities,
and the energetics in order to characterize four typical S· · · X chalcogen bonds in
proteins. Possibility of analogous Se· · · X chalcogen bonds in selenoproteins is also
pointed out (Sect. 9.4). Their implications in the structural control and in the function
and evolution of some particular proteins are summarized from the recent literature
in Sect. 9.5. Finally, the possible answers to the above six questions are addressed
based on the data having been obtained so far (Sect. 9.6).
9.2 Historical Aspects
9.2.1 Nonbonded Interactions Involving a Divalent Sulfur Atom
Evidence of attractive S· · · X interactions between a divalent sulfur atom (S) and

nearby heteroatoms (X) in the crystals of small organic compounds [25, 26] was
already known in late 1970s according to a pioneer work by Rosenfield and
Parthasarathy [27], who analyzed close atomic contacts around S in Cambridge
crystallographic database (CSD) [28] and found that X tends to approach S in the
backside of one of the two covalent bonds (i.e., in the σ*S direction). Subsequently,
Row and Parthasarathy [29] demonstrated that nonbonded S· · · S interactions in or-
ganic crystals may be stabilized by the orbital interaction between the lone pair of
one sulfur atom (nS ) and the anti-bonding orbital of the other sulfur atom (σ*S ).
This orbital model of the weak S· · · S interactions was later supported by Desiraju
and Nalini [30], whereas Dahaoui et al. [31] pointed out the importance of disper-
sion force for the S· · · S interactions. On the other hand, the directional preference
of S· · · O interactions around the O atom was studied in detail by Kucsman and
Kapovitz [32]. For intramolecular 1,4- and 1,5-S· · · O = C interactions, the S atom
tended to lie in the direction of the O lone pairs (i.e., the nO direction) rather than the
π electrons (i.e., the πO direction), suggesting the importance of nO → σ*S orbital
interaction for the stability. The presence of S· · · π interactions between a divalent
S atom and an aromatic ring in organic crystals was also suggested by Zauhar et al.
[33].
268 M. Iwaoka
180 180
150 150
120 120
90 90
60 60
30 30
0 0
0 30 60 90 120 150 180 0 30 60 90 120 150 180
a
θ 1 (degree)
b
θ 3 (degree)
Fig. 9.1 Directionality of intermolecular S· · · O=C interactions with d ≡ r – vdw(S) – vdw(O) = r

– 1.80–1.52 ≤ 0.2 Å. Open diamonds represent intermolecular S· · · O(amide) interactions. Filled
circles represent the other types of intermolecular S· · · O interactions. a Spatial distribution of O
relative to S by using angles θ1 and θ2 . b Spatial distribution of S relative to O by using angles θ3
and θ4 . This figure was modified from ref [34]
We [34] previously carried out statistical analysis of intermolecular S· · · O = C

interactions found in CSD in order to clearly characterize nonbonded S· · · O inter-
actions in organic crystals in comparison with those in proteins. Figure 9.1 shows
the observed directionalities around S and O atoms for the intermolecular S· · · O
contacts. It is obvious that the O atom approaches to the S atom in the backside
of the S–Y (Y = C or S) bond (θ1 ∼ 90◦ and θ2 ∼ 130◦ in Fig. 9.1a) in accord with
previous observations for S· · · O [27, 32] and S· · · S [29, 30] interactions. However,
no directional preference was observed around the O atom irrespective of whether it
is involved in an amide group or other carbonyl groups (Fig. 9.1b). This is in strong
contrast to the observation of Kucsman and Kapovitz [32] for the intramolecular 1,4-
and 1,5-S· · · O = C interactions, suggesting that the directional preference around
the O atom in intermolecular S· · · O = C interactions should be very week and hence
it will be easily affected by other steric and energetic factors. On the other hand, the
linearity of the Y–S· · · O alignment was proved to be robust against the packing force
of organic crystals.
As mentioned above, database analyses for various types of nonbonded S· · · X
(X = O, S, etc.) interactions have demonstrated that specific directional preferences
are obvious for the S· · · X interactions in organic crystals, although in the case of
intermolecular S· · · O = C interactions the preference is vanished. The directionali-
ties strongly suggested the importance of orbital interactions, such as nO → σ*S and
nS → σ*S . However, the nonbonded distances between the two interacting atoms
are sometimes only marginally shorter than the sum of the van der Waals radii, and,
in such cases, the directionality becomes a little subtle. Therefore, there should be
other electronic factors that are responsible for formation of the S· · · X contacts.
Such factors would involve the dispersion force (or electron correlation effects)
[31] and the electrostatic interaction between the positively charged S atom and the
negatively charged O atom as suggested by Burling and Goldstein [35]. Recently,
sophisticated theoretical methodologies have been applied for better understanding
such noncovalent interactions. Nakanishi [36] attempted to describe S· · · X interac-
tions more exactly by using the total electron energy density and Laplacian of the
electron density at the bond critical points of the interaction in the framework of
atoms-in-molecules (AIM) theory [37, 38], while symmetry-adapted perturbation
theory (SAPT) [39] was applied by Scheiner [40, 41] to decomposing the energetic
components to electrostatic, induction, and dispersion terms.
Relevance of S· · · O interactions to the biological functions of several organic
molecules has been reported. For example, the 1,4-S· · · O interaction of thiazole
nucleoside analogues was found to be important for the antitumor activity [35]. Sim-
ilarly, 1,5-S· · · O interactions were shown to play important roles in the antagonism
of (acylimino)thiadiazoline derivatives to an angiotensin II receptor [42] and in the
antitumor activity of leinamycin [43]. Analogous interactions between a divalent se-
lenium atom (Se) and X (i.e., Se· · · X interactions) have also been well characterized
in organic selenium compounds [44–46] and have been suggested to be relevant to
the biological activity of selenium [47, 48].
9.2.2 Discovery of S· · · X Chalcogen Bonds in Proteins
Weak noncovalent interactions are important physicochemical forces that control the
structure of proteins [49]. Ionic interaction, hydrogen bond, van der Waals force, and
hydrophobic interaction are mainly involved in this class of interactions, but some
novel interaction patterns, such as C–H· · · O hydrogen bond [50, 51, 52], cation-π
interaction [53–56], and CH/π hydrogen bond [57–59], have also been characterized
in folded protein structures. Importance of these new interactions for the stability
and function of proteins has already been pointed out.
The SS and MeS groups of Cys and Met residues, however, were usually consid-
ered just as hydrophobic moieties in folded protein structures until recently, except
for S· · · π interactions [16] and weak NH· · · S and OH· · · S hydrogen bonds [13, 14].
The S· · · π interactions in proteins were first pointed out by Morgan and co-workers
[60], who analyzed the close atomic contact between S and a π-ring system in eight
protein structures and found that the S atoms have a propensity to come over the
π-plane. On the other hand, Reid et al. [61] suggested by using a larger set of protein
structures that the close S· · · π contact in proteins can also be explained by C–H· · · S
interactions because the S atoms access to the π-plane from the side rather than the
top. According to elaborate experimental and theoretical studies [15, 16, 62, 63],
however, the nature of S· · · π interactions in proteins would be well rationalized by
the interaction between the aromatic π electrons and the S atom [15, 16]. Mean-
while, NH· · · S and OH· · · S hydrogen bonds were suggested to play some roles in
particular proteins [13, 14], but the interactions can be found with low frequency in
270 M. Iwaoka
protein structures. The S atoms of Cys and Met may have only a weak character as
a hydrogen-bond acceptor.
Stereochemistry of the close S· · · X contacts involving a Met sulfur atom (a MeS
group) was statistically analyzed in protein structures by Carugo in 1999 [17], but
distinct directional preferences of the contacts were not observed. It was therefore
suggested that nonbonded S· · · X interactions in proteins are either very weak or
physicochemically different from those in small molecules. Two years after Carugo’s
first attempt, we [18] and Pal and Chakrabarti [19] reported more extensive database
analysis of S· · · X interactions using a larger set of protein structures. These studies
led us to the discovery of distinct directionality in the interactions. Thus, the presence
of S· · · X chalcogen bonds in proteins was suggested for the first time. The chalcogen
bonds can be characterized by two directional propensities. One is the linearity of
C–S· · · X (or S–S· · · X) alignment, and the other is the directional preference of S
around X. These features are discussed in detail in the following sections.
9.3 Analytical Methods
Characterization of weak noncovalent interactions is still a difficult task in any

molecular systems. Usually, several experimental and theoretical techniques are
collaboratively employed for the detection. For the study of S· · · X chalcogen bonds
in proteins, only limited techniques may be applicable not only because the struc-
ture of a protein is very complicated but also because the interactions are somehow
elusive. Therefore, the existence of S· · · X chalcogen bonds in proteins had not been
recognized longtime. However, since high-resolution protein structural data [12] are
now available, we can use the data for mining the chalcogen bonds.
Another useful technique available to characterize weak interactions is theoreti-
cal calculation. Sophisticated calculation techniques, such as ab initio calculation,
cannot be directly applied for a whole protein molecule even today. However, the
calculation for small model systems still provides us valuable information as shown
later. Indeed, from such simple calculation we know the intrinsic interaction energy
and directional preference of the interactions, which will be compared with what we
have obtained from the statistical database analysis of protein structures. In this way,
we can reasonably assume the nature of S· · · X chalcogen bonds and the structural
perturbations in protein architecture. Other state-of-the-art spectroscopic techniques
could also be applied for direct detection of the chalcogen bonds in proteins. De-
velopment of such techniques will have large impacts in versatile fields of protein
research in future.
PDB
PDB-SELECT
(n = 604, R ≤ 2.0 Å)
TOP500
A B (n = 500, R ≤ 1.8 Å)
n = 283 n = 321
C
Cys = 347 Cys = 443
n = 179
Met = 1081 Met = 1043
Cys = 275
Met = 821
Fig. 9.2 The numbers of protein structures (n), cysteine residues (Cys), and methionine residues
(Met) involved in PDB-SELECT and TOP500 databases. R represents the resolution. The numbers
for each zone (A, B, or C) are shown
9.3.1 Protein Structure Database
All protein structures that have been determined to date are available from protein
databank (PDB) [12], which we have utilized for characterization of S· · · X chalco-
gen bonds in proteins [18, 34, 64–67]. However, PDB contains a lot of redundant
structural data for one protein. Therefore, filtering of the data was necessary to ob-
tain reliable statistical information about the relevant interactions from the database.
Several filtering algorithms are known in order to choose nonredundant and heteroge-
neous protein structures from PDB. PDB_SELECT database [68, 69] collects 1085
nonredundant X-ray structures, which consist of 604 high resolution (≤ 2.0 Å) and
481 low resolution (>2.0 Å) structures, from PDB with a criterion of the structural ho-
mology less than 25 %. The high-resolution data were employed in refs [18, 34, 64].
TOP500 database [70], which collects 500 heterogeneous protein structures with
high resolution (≤ 1.8 Å) from PDB for drawing Ramachandran-plot distributions,
was applied in ref [67]. These two databases contain 321 common structures as il-
lustrated in Fig. 9.2. Comparing the structural features of S· · · X chalcogen bonds
obtained from PDB-SELECT with those obtained from TOP500, we can guarantee
reliability of the database analysis.
On the other hand, redundant structural data for one particular protein are useful
for the analysis on the functional and evolutional aspects of specific S· · · X chalco-
gen bonds because such data contain a variation of the ligands and mutations. By
comparing the structural features of the S· · · X chalcogen bonds among the redun-
dant data, the plausible roles in the protein function and evolution can be inferred
[65, 66].
A computer program was coded for extracting all non-hydrogen atoms (X) that
have a close atomic contact in proteins to a divalent sulfur atom (S) of Cys (a SS
group) or Met (a MeS group). The data of S· · · X contacts obtained were sorted by
272 M. Iwaoka
the use of several structural parameters, such as an atomic distance between S and
X (r), access angles of X to S (θ1 and θ2 ), and access angles of S to X (θ3 and θ4 )
(see Fig. 9.1). The relative atomic distance to the van der Waals radii was define by
d ≡ r − vdw(S) − vdw(X), where vdw(S) = 1.80 Å and vdw(X) = 1.70 Å for C, 1.55
Å for N, 1.52 Å for O, etc. Details of the algorithm were described elsewhere [64].
9.3.2 Theoretical Calculation
The purpose of theoretical calculation includes quantitative characterization of the

structural features of S· · · X chalcogen bonds, estimation of the interaction energies,
and identification of the energetic components of the interactions. To accomplish
this mission, ab initio calculation has been carried out for simple cluster models,
CH3 SSCH3 + CH3 CONHCH3 (cluster I) and CH3 SCH3 + CH3 CONHCH3 (cluster
II), at various calculation levels, such as HF/6-31G(d), MP2/6-31G(d), and MP2/6-
311G(d,p) [18, 34, 64]. The truncated models employed have an advantage that they
will provide important information about the interactions in the absence of steric per-
turbations from the protein structure. Thus, intrinsic nature of the interactions can be
deduced from the calculation. On the other hand, comparison of the results obtained
at different calculation levels are applicable to addressing the energetic components
because the second order Møller-Plesset perturbation theory (MP2) sufficiently in-
corporates electron correlation effects whereas the Hartree-Fock method (HF) does
not [71]. The potential energy surfaces of the S· · · X chalcogen bonds were drawn
by using the energies obtained as a function of access angles θ1 –θ4 for the clusters
at MP2/6-31(G) level [64].
9.4 Characterization of Chalcogen Bonds in Proteins
9.4.1 Statistics of S· · · X Interactions
Results of the statistical analysis of the nonbonded S· · · X contacts that were found
in protein structure databases are summarized in Fig. 9.3. The S· · · X contacts of the
SS group extracted from PDB-SELECT only (zone A of Fig. 9.3a) show that the
probabilities of S· · · O and S· · · S contacts increase in a short distance range (d ≤ 0.0
Å), while those of S· · · N and S· · · C contacts decrease contrarily. The same trends are
observed for zone B (common in PDB-SELECT and TOP500) and zone C (TOP500
only), suggesting the presence of unique S–S· · · O and S–S· · · S interactions around
the SS bonds in proteins. For the MeS group, on the other hand, the probabilities of
S· · · O and S· · · N contacts increase in a short distance range, while that of S· · · C
contacts decreases and S· · · S contacts are very rare from the beginning. Again,
common trends were observed in zones A-C, suggesting the presence of unique C–
S· · · O and C–S· · · N interactions around the MeS groups in proteins. It is important
b
Fig. 9.3 Probabilities of S· · · X (X = O, N, S, and C) contacts observed for SS groups a and
MeS groups b in the total S· · · X contacts extracted from the databases. Zones A–C correspond to
PDB-SELECT only, common, and TOP500 only, respectively. See also Fig. 9.2
to notice that the probabilities do not change significantly depending on the databases
employed. Thus, the existence of S· · · O, S· · · N, and S· · · S interactions should not
be due to an artifact.
9.4.2 Directionality of S· · · X Interactions
Four types of S· · · X interactions were suggested by statistical analysis of the relative

atomic distances (d) for the S· · · X contacts found in protein structure databases.
These interactions were further analyzed by using access angles θ1 –θ4 around the
interacting atoms to characterize the directional propensity in proteins.
The S· · · O contacts observed for the SS group contained the interactions with
various types of O, such as amide, hydroxyl, and water O atoms. Among these, the
main-chain amide O atom was most frequently found (more than 50 %). Therefore,
274 M. Iwaoka
a b
Fig. 9.4 Directionality of SS· · · O(amide) interactions found in PDB-SELECT database. Filled
squares represent the interactions with d ≤ 0.0 Å. Open circles represent the interactions with
0.0 < d ≤ 0.2 Å. a Spatial distribution of O relative to S by using angles θ1 and θ2 . b Spatial
distribution of S relative to O by using angles θ3 and θ4 . This figure was modified from ref [64]
the directionality of the S· · · O contacts between a SS group and a main-chain amide

O atom was analyzed. Figure 9.4 shows the directionality of SS· · · O(amide) interac-
tions found in PDB-SELECT database [64]. It is clearly seen that the O atom tends
to approach the S atom in the backside of the S–S (θ1 ∼ 90◦ and θ2 ∼ 230◦ ) or S–C
(θ1 ∼ 90◦ and θ2 ∼ 130◦ ) covalent bond. It should be noted that the definitions of the
access angles are slightly different from those used in Fig. 9.1. On the other hand,
the S atom tends to come above or below the amide plane (θ3 ∼ 90◦ and θ4 ∼ 180
or 360◦ ). The linearity of S–S· · · O or C–S· · · O alignment as well as the perpen-
dicularity of S· · · O = C alignment allows an effective charge transfer through the
πO → σ*S orbital interaction. Thus, the existence of two types of S· · · X chalcogen
bonds, i.e., S–S· · · O = C (type 1) and C–S· · · O = C (type 2), is strongly suggested.
It is notable that not lone pairs but π electrons of the O atom participates in these
chalcogen bonds. This feature is in significant difference from the trend observed in
organic crystals (Fig. 9.1b) [34].
Similar but slightly less obvious directionalities were observed for the S· · · O
contacts between a MeS group and a main-chain amide O atom (Fig. 9.5). The direc-
tional preferences, i.e., linear C–S· · · O alignment and perpendicular approach of S
to the amide plane, provide a strong evidence for the existence of type-2 chalcogen
bond (C–S· · · O = C) in proteins.
a b
Fig. 9.5 Directionality of MeS· · · O(amide) interactions found in PDB-SELECT database. Filled
squares represent the interactions with d ≤ 0.0 Å. Open circles represent the interactions with
0.0 < d ≤ 0.3 Å. a Spatial distribution of O relative to S by using angles θ1 and θ2 . b Spatial
distribution of S relative to O by using angles θ3 and θ4 . This figure was modified from ref [64]
The S· · · N contacts found around the MeS group of Met residues (Fig. 9.3b)
mostly contained the N atoms of main-chain and side-chain amide groups. To char-
acterize directional features of these S· · · N contacts, the access angles between
the S and the main-chain amide N atoms were analyzed in the range of d ≤ 0.3 Å
(Fig. 9.6). The scattergrams clearly indicate that the S· · · N contacts contain two
types of interactions.
One is a S· · · N chalcogen bond (type 3) (open circles in Fig. 9.6) with linear
C–S· · · N alignment and perpendicular approach of S to the amide plane, allow-
ing effective πN → σ*S orbital interaction. The directionality around the nitrogen
atom (Fig. 9.6b) was more explicit than that of type-1 and type-2 chalcogen bonds
(Figs. 9.4b and 9.5b), whereas the directionality around the S atom (Fig. 9.6a) was
less explicit. The other type of the S· · · N interaction is a N–H· · · S hydrogen bond
(filled squares in Fig. 9.6) with linear N–H· · · S alignment and roughly perpendicular
approach of N to S. N–H· · · S hydrogen bonds can be experimentally characterized
by gas-phase IR/UV spectroscopy for short model peptides [14].
Statistical analysis of short S· · · S contacts in PDB-SELECT suggested the pres-
ence of S· · · S interactions around the SS groups in proteins although the frequency
was low. Indeed, the directional analysis around the interacting S atoms (data no
shown) revealed the unique direction preferences, which are similar to those of the
nonbonded S· · · S interactions found in organic crystals [29, 30]. Namely, one S atom
is located in the backside of a S–S or S–C bond of another sulfur atom, and the latter
276 M. Iwaoka
a b
Fig. 9.6 Directionality of MeS· · · N(amide) interactions (d ≤ 0.3 Å) found in PDB-SELECT
database. a Spatial distribution of the main-chain N atoms relative to the S atom. b Spatial dis-
tribution of the S atoms relative to the main-chain N atom. Open circles (), crosses (+) and
filled squares () represent the contacts with θ3 ≤ 120◦ , 120 < θ3 ≤ 150◦ , and 150 < θ3 ≤ 180◦ ,
respectively. This figure was modified from ref [64]
S atom is approaching the former S atom in the lone pair directions. Thus, the S· · · S
interactions should have an advantage of an effective nS → σ*S orbital interaction.
This interaction can be categorized to type-4 chalcogen bond (S–S· · · S–S).
According to the statistical and directional analysis of the S· · · X contacts found
in protein databases, existence of four typical S· · · X chalcogen bonds (Fig. 9.7) was
suggested in proteins. These chalcogen bonds are characterized by three structural
features. (1) The nonbonded S· · · X distance is only slightly less than or nearly equal
to the sum of the van der Waals radii. (2) The X atom always comes to the backside
of the S–S or S–C bond with preference of the extension of the S–S bond rather than
the S–C bond, probably due to a more distinct σ-hole in the backside of the SS bond.
These features are compatible with the definition of chalcogen bond [21]. (3) The S
atom tends to approach the amide O or N atom in the directions perpendicular to the
amide plane, while in the S· · · S chalcogen bond the S atom tends to approach the
other S atom in the direction of the lone pairs. Energetic aspects of these chalcogen
bonds are shown in the subsequent section.
Fig. 9.7 Typical chalcogen bonds in proteins
C2
C2
S2 C1 C1 C3’
H1 C2
S1 C3’ S1 S2 S1
C3’ C1 N1’
N1’ C2’ S1 H1’ O1’
C2 O1’
C2’
C2’ O1’
N1’ H1 C1 N1’
C1’ C3’
O1’ H1 C2’ C1’
C1’
C1’ H1’
a b c d
Fig. 9.8 Stable structures obtained for molecular clusters CH3 SSCH3 + CH3 CONHCH3 (a and b)
and CH3 SCH3 + CH3 CONHCH3 (c and d) at MP2/6-31G(d). This figure was modified from ref
[64]
9.4.3 Potential Surfaces of S· · · X Interactions
In order to elucidate the intrinsic structural preferences as well as the strength of

S· · · O chalcogen bonds, ab initio calculation was performed for molecular clusters of
CH3 SSCH3 + CH3 CONHCH3 (cluster I) and CH3 SCH3 + CH3 CONHCH3 (cluster
II) [18, 34, 64]. The representative structures obtained as a global or local energy
minimum at MP2/6-31G(d) are shown in Fig. 9.8.
Structure a is the global energy minimum of cluster I, which has a vertical S· · · O
interaction to the amide plane with keeping a linear S–S· · · O alignment. The relative
S· · · O distance (d) was 0.02 Å. The structural features are in good agreement with
those observed for type-1 chalcogen bond in proteins. The total stabilization energy
due to the complexation with the correction for basis set superposition errors (BSSE)
[72] was 3.21 kcal/mol including the contribution from a coexisting C–H· · · O hy-
drogen bond. The interaction distance was slightly decreased to d = − 0.02 Å when
a water molecule was hydrogen-bonded to the amide O atom [18]. Another stable
structure located for cluster I (Structure b) has a horizontal S· · · O interaction to the
amide plane. The relative S· · · O distance (d) was 0.06 Å, and the total complexation
energy with BSSE correction was 2.18 kcal/mol, which is about 1 kcal/mol smaller
278 M. Iwaoka
than that of Structure a. The preference of the vertical interaction to the horizontal
one observed for cluster I can be explained by the relative energy levels of the π and
lone pair (n) orbitals for the amide group: the π orbital takes HOMO in an amide
group, while in other carbonyl groups it yields HOMO to the n orbital [34].
On the other hand, two characteristic structures were obtained for cluster II.
Structure c, which was located as the global energy minimum, has a vertical S· · · N
interaction between CH3 SCH3 and CH3 CONHCH3 . The relative S· · · N distance
(d) was 0.22 Å, and the total complexation energy with BSSE correction was 2.88
kcal/mol, which is slightly smaller than that of Structure a. The structural features are
consistent with type-3 S· · · N chalcogen bond, supporting its existence in proteins.
Structure d has a linear N–H· · · S hydrogen bond with d = 0.24 Å. The relative energy
of structure d to structure c, however, was + 0.78 kcal/mol. These calculation results
are consonant with the observation for the S· · · N contacts in proteins (Fig. 9.6).
Type-2 chalcogen bond could not be characterized as an energy minimum structure
in the above ab initio calculation. However, when the S· · · O distance of the complex
was fixed to r = 3.30 Å (i.e., d = − 0.02 Å), the structure consonant with type-
2 chalcogen bond was obtained with the complexation energy of 2.47 kcal/mol.
Although it is not a local energy minimum structure, the result suggests that type-2
chalcogen bond in proteins is almost as strong as type-3 chalcogen bond.
The stable structures obtained by ab initio calculation (Fig. 9.8) strongly suggest
the importance of orbital interactions, such as πO → σ*S and πN → σ*S , for the
chalcogen bonds in proteins. Indeed, such components were quantitatively evaluated
by using the natural bond orbital (NBO) method [73]. However, the estimated value
for structure A (0.64 kcal/mol) was not enough to explain the total interaction energy
(3.21 kcal/mol) [18, 34]. On the other hand, a significant contribution of dispersion
force was evident because the S· · · O atomic distance for cluster I became larger when
optimized at the HF/6-31G(d) level (d = 0.37 Å) and the stabilization energy reduced
to 2.18 kcal/mol. Within the framework of the MP2 theory, the difference between
MP2 and HF calculation results is assigned to the effects of high-order electrostatic
interactions such as a dispersion force [71]. The S· · · O chalcogen bonds in proteins
may be primarily stabilized by the dispersion and/or long-range electrostatic forces
but the directional propensity would be controlled by the HOMO–LUMO-type orbital
interaction [64].
The potential surfaces of the type-1 S· · · O chalcogen bond with a fixed S· · · O
distance (d = 0.18 Å) (Fig. 9.9) were then calculated for cluster I at MP2/6-31G(d)
level [64]. The potential energy minimum is located at θ1 ∼ 90◦ and θ2 ∼ 240◦ around
S and at θ3 ∼ 0◦ and θ2 ∼ 90◦ around O as indicated with x. The position corresponds
nearly to the global energy minimum (i.e., structure a). The potential surface around
O (Fig. 9.9b) is very flat compared to that around S (Fig. 9.9a), suggesting that
the directionality around O is intrinsically weaker than that around S. It is more
interesting to notice that the distribution patterns of the S· · · O chalcogen bonds in
proteins (Fig. 9.4) almost perfectly follow the potential surfaces calculated for the
isolated model cluster. Thus, it is likely that the S· · · O chalcogen bond is one of
the important forces that determine protein structures. The attractive interaction was
sustained at least in the range of −0.22 < d < 0.58 Å at the same calculation level,
d = 0.18 Å
-2.5
d = 0.18 Å
-1.0
x -2.5
-2.0
-1.5 -2.25 -1.5
-1.0 -2.0
-0.5
0.0
-2.0
-2.5
x
θ1 (degree)
a b θ 3 (degree)
Fig. 9.9 Potential energy surfaces of type-1 S· · · O chalcogen bond calculated for molecular cluster
CH3 SSCH3 + CH3 CONHCH3 at MP2/6-31G(d). The relative distance (d) was fixed at 0.18 Å. This
figure was modified from ref [64]
suggesting that the potential energy surface of S· · · O chalcogen bond is also very
flat along the axis of the interaction distance. Similar features were observed for the
potential surfaces of type-2 chalcogen bond (C–S· · · O = C) [64].
9.4.4 Possibility of Se· · · X Chalcogen Bonds in Proteins
Selenium atoms are seldom found in protein structure databases. Therefore, statis-
tical analysis of Se· · · X chalcogen bonds in proteins is not feasible at this moment.
However, in some particular selenoproteins, such as glutathione peroxidase [74–77],
thioredoxin reductase [78, 79], the chalcogen bonds may exist because a divalent
selenium atom has a distinct σ-hole and nonbonded Se· · · X interactions (or Se· · · X
chalcogen bonds) [44–46] are generally stronger than the S· · · X chalcogen bonds.
Therefore, mining Se· · · X chalcogen bonds must be easier if high-resolution struc-
tures of selenoproteins will be accumulated. In the catalytic cycle of glutathione
peroxidase, significant roles of Se· · · O and Se· · · N interactions at the active site
have been suggested [47, 48]. It would be an interesting trial in future works to re-
place selenocysteine (Sec) for the Cys residue that is involved in a S· · · O chalcogen
bond of a particular protein and see the change in the thermodynamic stability and
the function.
280 M. Iwaoka
9.5 Implications in Protein Science

9.5.1 Structural Control
The comparison of the potential energy surface calculated for a simple cluster model
(Fig. 9.9) with the directional preferences of S· · · O chalcogen bond observed in
organic crystals (Fig. 9.1) and protein structures (Fig. 9.4) allows us for following
considerations.
1. The directional preferences of the S· · · O chalcogen bond observed in protein
structures are in accord with the profiles of the potential surfaces calculated for
the truncated cluster. The directionalities around both O and S atoms are not
significantly perturbed in proteins. This remarkable agreement strongly suggests
that the S· · · O chalcogen bond is an important structural determinant for protein
architecture.
2. In organic crystals, the directional preference around S follows the profile of the
potential surface calculated for the isolated cluster, but the directional preference
is not present around O. This suggests that crystal packing force can overwhelm
the intrinsic directional preference around O, while it cannot disturb the intrinsic
directionality around S.
Thus, the order of strength for the structural factors that are present in organic crystals
and protein structures can be estimated to be: linearity of S–S· · · O and C–S· · · O >
crystal packing force > verticality of S· · · O = C > protein structure [34, 66]. This or-
der will be useful for protein engineering and molecular design of functional organic
sulfur compounds.
9.5.2 Enzymatic Function
Importance of S· · · X interactions in enzymatic functions has already been pointed

out for some particular proteins [80–84]. For example, Taylor and Markham [80]
reported that in the active site of S-adenosylmethionine synthetase the electrostatic
S· · · O interaction between the S atom of Met substrate and the carboxylate O atom
of Asp118 plays a significant role. Brandt et al. [81] suggested that cleavage of a
disulfide bond in the extracellular region of G-protein receptors is catalyzed by the
S· · · O interaction between Cys121 and the carboxylate side chain of Asp288. In
most of these cases, the S· · · O interaction is present between a divalent S atom and a
negatively charged O atom. As for the electrically neutral S· · · X interaction having
pertinent directional features to S· · · X chalcogen bonds, on the other hand, only a
few examples [65, 66, 85] have been suggested to play roles in enzymatic functions
of proteins.
After a success in structural characterization of S· · · X chalcogen bonds in pro-
teins, particular protein families or domains were employed for mining the specific
S· · · X chalcogen bonds that are commonly present in a wide range of homogeneous
Cys84
Met8
Cys94
3.35
3.23
Arg100
Phe94
3.35
3.30 Cys98
Cys61
Cys44
Ala55 3.30
Glu40
Active site
Fig. 9.10 Structure model of porcine phospholipase A2 (PDB code, 1HN4) with indication of five
S· · · X interactions. The atomic distances are shown in Å
structures. For this purpose, redundant protein structures registered in PDB were
very useful. Examples are shown below for the case of phospholipase A2 (PLA2 )
[65], ribonuclease A (RNase A) [66], insulin [66], and lysozyme [66]. A number of
high-resolution structures can be found in PDB for these typical proteins
Phospholipase A2 (PLA2 ) [86, 87] is a small globular protein consisting of about
130 amino acid residues. This SS-rich protein catalyzes hydrolysis of the 2-acyl
ester bond of phosphoglycerides in the presence of a calcium ion. Comprehensive
search for S· · · X chalcogen bonds in PLA2 domain group revealed the presence
of four common S· · · O and one common S· · · N interactions (Fig. 9.10). Among
these, S(Cys61)· · · O(Ala55), S(Cys84)· · · O(Cys96), and S(Met8)· · · N(Arg100)
interactions can be clearly assigned to type-1, type-2, and type-3 chalcogen bonds,
respectively, based on the directional characters. The other two have rather linear
S· · · O = C alignment. Most of the five S· · · X interactions were present in vicinity
of the active site, structurally supporting the conformational stability. They were
not disturbed by the conformational changes caused by the binding of various sub-
strates. Indeed, the enzymatic activity of porcine PLA2 was significantly decreased
in the Met8,20Leu mutant [88], which cannot hold the chalcogen bond of type-2
S(Met8)· · · N(Arg100). Thus, this chalcogen bond was supposed to be important for
the enzymatic activity of PLA2 [65].
A similar functional role of S· · · O interaction was proposed by Palenchar et al.
[85] for Bacillus subtilis adenylosuccinate lyase, in which the MeS group of Met10
is located close to the main-chain amide O atom of Thr317 (Fig. 9.11). This S· · · O
interaction would effectively hold the N-terminal region close to the helix bundle.
Indeed, the thermodynamic stability was significantly decreased upon substituting
Met10 with leucine (Leu), resulting in a loss of the enzymatic activity to some extent.
282 M. Iwaoka
Fig. 9.11 Structure model of

Thermotoga maritima
adenylosuccinate lyase (PDB
code, 1C3C) with indication
of the S· · · O interaction N-terminal
between Met10 and Thr317
Met10
4.38
Thr317
RNase A [89] is a small monomeric globular protein consisting of 124 amino

acid residues having four SS bonds in the native state. For this domain, 43
high-resolution (≤ 2.0 Å) structures were extracted from PDB, and close S· · · X
contacts were thoroughly sought. As a result, two common S· · · O and one com-
mon S· · · N interactions, i.e., S(Cys26)· · · Oγ(Thr99), S(Cys65)· · · O(Gln69), and
S(Cys58)· · · N(Pro117), were characterized [66]. Their locations are shown in
Fig. 9.12. The S(Cys26)· · · Oγ(Thr99) interaction has linear S–S· · · Oγ(side-chain)
alignment. The other S(Cys65)· · · O(Gln69) and S(Cys58)· · · N(Pro117) interactions
can be assigned to type-1 and type-3 chalcogen bonds, respectively, based on almost
linear S–S· · · O and C–S· · · N alignment. The S(Cys65)· · · O(Gln69) chalcogen bond
is located in the Cys65–Cys72 disulfide loop, which is one of the important sites that
fold in the beginning of the oxidative folding [90] and gives significant thermo-
dynamic stability to the native structure [90]. Therefore, structural and functional
importance of the type-1 chalcogen bond was suggested [66].
Insulin [92] is a peptide hormone composed of two peptides (A and B chains),
which are connected together by two SS bridges. Another SS bond is present
in the A chain. For this protein domain, two common S· · · O interactions, i.e.,
S(Cys20)· · · O(Glu17) and S(Cys19)· · · O(Leu15), were detected within A and B
chains, respectively, in the 23 high-resolution (≤ 2.0 Å) structures. Figure 9.13 shows
their locations in the native structure. The two S· · · O interactions can be assigned to
type-1 chalcogen bonds, forming a little bent O· · · S–S· · · O alignment. Interestingly,
these chalcogen bonds were not disrupted by complexation with various ligands,
while new S· · · O interactions were emerged in the A chain; S(Cys6)· · · O(Ile2),
S(Cys7)· · · O(Val3), or S(Cys11)· · · O(Gln5). For the case of 1ZNI (Fig. 9.13), the
former two S· · · O interactions were formed by binding zinc chloride. The allosteric
effects may possibly contribute to the function of insulin [66].
Lysozyme [93] consists of 129 or 130 amino acid residues with four SS bonds in
the native state. A large number of the structures with various ligand have been

bovine ribonuclease A (PDB
code, 9RAT) with indication
of three S· · · X interactions.
The atomic distances are
shown in Å Pro117
Cys58
3.28
Cys26
3.01
3.38
Cys65
Thr99
Gln69

porcine insulin (PDB code,
1ZNI, A and B chains) with
indication of four S· · · O Cys7
Cys6
A chain
Glu17
interactions. ZnCl2 is omitted 3.77
for clarity. The atomic 3.18 Ile2 3.50
distances are shown in Å Leu15 Cys20

Val3
3.33
Cys19
y B chain
registered in PDB. We [66] selected 28 high-resolution (≤ 1.5 Å) structures of

the lysozyme domain and analyzed close S· · · X contacts involved in the struc-
tures. One common S· · · O interaction, i.e., S(Cys127)· · · O(Ile124), and three
common S· · · N interactions, i.e., S(Cys30)· · · Nε(Trp123), S(Cys80)· · · N(Asn65),
and S(Cys127)· · · Nη(Arg5), were characterized as shown in Fig. 9.14. Among these,
S(Cys127)· · · O(Ile124) and S(Cys127)· · · Nη(Arg5) make bifurcated interactions.
However, functional roles of these interactions are not yet clear.
9.5.3 Evolution
The evolutional aspect of S· · · X interactions was analyzed for the snake PLA2 do-
main [65]. For this domain group, the phylogenetic dendrogram was already obtained
by Ohno et al. [94] using the amino acid sequences of various PLA2 in venom of
snakes inhabiting the southern islands of Japan. Mapping the S· · · O and S· · · N in-
teractions commonly observed by the database analysis, it was found that most of
the S· · · X interactions (including three chalcogen bonds shown in Fig. 9.10) make
clusters on the dendrogram. The evolutional conservation suggests a possible role of
S· · · X chalcogen bonds in molecular evolution of proteins [65].
284 M. Iwaoka

chicken lysozyme (PDB code,
194 L) with indication of four
S· · · X interactions
3.29 Trp123
Cys30
Arg5
3 29
3.29 Ile124
3.25
Cys127
Cys80
3.50
Asn65
9.6 Outlook
According to the elaborate analysis of close S· · · X contacts in protein structure

databases, four typical chalcogen bonds have been characterized as shown in Fig. 9.7
[64]. The S–S· · · O = C chalcogen bond (type 1) is characterized by linear S–S· · · O
alignment and vertical approach of S to the amide plane, allowing effective πO
→ σ*S orbital interaction. The C–S· · · O = C chalcogen bond (type 2) has similar
structural features to type 1 except for the replacement of C–S for the S–S bond.
The C–S· · · N(amide) chalcogen bond (type 3) is characterized by linear C–S· · · N
alignment and vertical approach of S to the amide plane, allowing effective πN
→ σ*S orbital interaction. The last S–S· · · S–S chalcogen bond (type 4) has linear
S–S· · · S or C–S· · · S alignment and roughly vertical approach of S to the other S
atom, allowing nS → σ*S orbital interaction. The strength of these chalcogen bonds
would decrease in the order of type 1 > type2 ≈ type 3 > type 4, according to ab
initio calculation for simple molecular cluster models. The maximum interaction
energy was estimated to be 3.21 kcal/mol for type1. The potential energy surfaces
are significantly flat around the O atom and along the interaction distance, while the
directional preference around the S atom is distinct. For some particular proteins,
putative roles of the chalcogen bonds in their enzymatic functions and evolution has
been suggested. Thus, the chalcogen bonds, though their interaction energy is smaller
than that of classical hydrogen bonds (∼ 5 kcal/mol), can be an important element
of weak noncovalent interactions that control chemical and biological properties of
protein architecture.
On the basis of the data obtained from statistical database analysis and theoretical
calculation on the S· · · X chalcogen bonds in proteins, possible answers to the six
questions raised earlier would be as follows.
Q1. Do S· · · X chalcogen bonds really exist in proteins?
A1. Yes. According to the similarity between the statistical analysis results ob-
tained by using two different structure databases (PDB-SELECT and TOP500),
S· · · X chalcogen bonds should exist widely in proteins.
Q2. Are the directional propensities similar to those observed for small organic
molecules?
A2. No. In the S· · · O = C interactions, the O atom approaches the S atom in
the backside of a S–S or S–C bond (σ*S direction) in both organic crystals and
proteins. However, the directionality around the O atom is in significant difference.
There is no directional preference in organic crystals due probably to the diversity
of the C = O functional groups as well as the crystal packing force, whereas in
proteins the S atom comes over the carbonyl O atom in the πO direction due to
the elevated πO orbital energy level in the amide group compared to other types
of carbonyl groups, such as ketone and ester.
Q3. How do the potential surfaces look like?
A3. Profiles of the potential surfaces around the S and O atoms for both S–S· · · O =
C and C–S· · · O = C interactions match well with the directional distributions of
the S· · · O interactions observed in proteins. There are distinct potential holes in
the σ*S and πO orbital directions. However, the potential surface around the O
atom is rather flat compared to that around the S atom.
Q4. Can the S· · · X chalcogen bonds control protein structure?
A4. Yes. Since the directionalities around S and O atoms of type-1 S· · · O chalco-
gen bonds in proteins are not disturbed from their intrinsic propensities on the
potential surfaces calculated for the truncated cluster models, S· · · O chalcogen
bonds should control the protein structures to some extent. Similar situation may
be applied for the other S· · · X chalcogen bonds in proteins.
Q5. Are the S· · · X chalcogen bonds relevant to protein function and evolution?
A5. Probably yes. There are several examples in which specific S· · · X chalcogen
bonds are suggested to control protein functions (for PLA2 , RNase A, insulin,
etc.) and evolution (for PLA2 ).
Q6. Are there similar Se· · · X chalcogen bonds in selenoproteins too?
A6. Probably yes. Se· · · X chalcogen bonds should be stronger than S· · · X chalco-
gen bonds because a Se atom has larger propensity than a S atom to adapt a
hypervalent state. Therefore, Se· · · X chalcogen bonds may possibly be found in
selenoproteins if the high-resolution structures will be accumulated.
Thus, the concept of S· · · X chalcogen bonds will be a useful tool for fully understand-
ing the structure, function, and evolution of proteins. Their applications to protein
engineering and the ligand design will be interesting and promising challenges in
the fields of protein science.
Acknowledgment I thank Ms. Natsuki Babe for her assistance in preparation of graphical
materials.
286 M. Iwaoka
References
1. Tonin C, Aluigi A, Varesano A, Vineis, C (2010) Keratin-based nanofibers. In: Nanofibers KA

(ed) InTech, chapter 7. pp 139–158
2. Fraser RDB, Parry DAD (2011) The structural basis of the filament-matrix texture in the
avian/reptilian group of hard β-keratins. J Struct Biol 173:391–405
3. Coyle P, Philcox JC, Carey LC, Rofe AM (2002) Metallothionein: the multipurpose protein.
Cell Mol Life Sci 59:627–647
4. Moura JJG, Macedo AL, Palma PN (1994) Ferredoxins. Methods Enzymol 243:165–188
5. Laity JH, Lee BM, Wright PE (2001) Zinc finger proteins: new insights into structural and
functional diversity. Curr Opin Struct Biol 11:39–46
6. Nordberg J, Arner ESJ (2001) Reactive oxygen species, antioxidants, and the mammalian
thioredoxin system. Free Rad Biol Med 31:1287–1312
7. Ellgaard L, Ruddock LW (2005) The human protein disulphide isomerase family: Substrate
interactions and functional properties. EMBO Rep 6:28–32
8. Gruber CW, Čemažar M, Heras B, Martin JL, Craik DJ (2006) Protein disulfide isomerase: the
structure of oxidative folding. Trends Biochem Sci 31:455–464
9. Hatahet F, Ruddock LW (2009) Protein disulfide isomerase: a critical evaluation of its function
in disulfide bond formation. Antioxid Redox Signal 11:2807–2850
10. Rao MB, Tanksale AM, Ghatge MS, Deshpande VV (1998) Molecular and biotechnological
aspects of microbial proteases. Microbiol Mol Biol Rev 62:597–635
11. Ventura S, Chang RJY (2011) Oxidative folding: coupling conformational folding and disulfide
formation. In Folding of disulfide proteins, Chang RJY, Ventura S eds; Springer: New York,
chapter 1, pp 1–22
12. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne
PE (2000) The protein data bank. Nucl Acids Res 28:235–242
13. Ippolito JA, Alexander RS, Christianson DW (1990) Hydrogen bond stereochemistry in protein
structure and function. J Mol Biol 215:457–471
14. Biswal H, Gloaguen E, Loquais Y, Tardivel B, Mons M (2012) Strength of NH· · · S hydrogen
bonds in methionine residues revealed by gas-phase IR/UV spectroscopy. J Phys Chem Lett
3:755–759
15. Bhattacharya R, Pal D, Chakrabarti P (2004) Disulfide bonds, their stereospecific environment
and conservation in protein structures. Protein Eng Des Sel 17:795–808
16. Chakrabarti P, Bhattacharyya R (2007) Geometry of nonbonded interactions involving planar
groups in proteins. Prog Biophys Mol Biol 95:83–137
17. Carugo O (1999) Stereochemistry of the interaction between methionine sulfur and the protein
core. Biol Chem 380:495–498
18. Iwaoka M, Takemoto S, Okada M, Tomoda S (2001) Statistical characterization of nonbonded
S· · · O interactions in proteins. Chem Lett 132–133
19. Pal D, Chakrabarti P (2001) Non-hydrogen bond interactions involving the methionine sulfur
atom. J Biomol Struct Dynam 19:115–128
20. Murray JS, Lane P, Politzer P (2008) Simultaneous σ-hole and hydrogen bonding by sulfur-
and selenium-containing heterocycles. Int J Quant Chem 108:2770–2781
21. Wang W, Ji B, Zhang Y (2009) Chalcogen bond: a sister noncovalent bond to halogen bond. J
Phys Chem A 113:8132–8135
22. Metrangolo P, Neukirch H, Pilati T, Resnati G (2005) Halogen bonding based recognition
processes: a world parallel to hydrogen bonding. Acc Chem Res 38:386–395
23. Parisini E, Metrangolo P, Pilati T, Resnati G, Terraneo G (2011) Halogen bonding in halocarbon-
protein complexes: a structural survey. Chem Soc Rev 40:2267–2278
24. Politzer P, Murray JS, Clark T (2010) Halogen bonding: an electrostatically-driven highly
directional noncovalent interaction. Phys Chem Chem Phys 12:7748–7757
25. Ciuffarin E, Guaraldi G (1970) Nucleophilic substitution at dicoordinated sulfur. Effect of the
leaving group on the reaction between triphenylmethyl sulfenyl derivatives and n-butylamine.
J Org Chem 35:2006–2010
26. Lu J, Lu Y, Yang S, Zhu W (2011) Theoretical and crystallographic data investigations of

noncovalent S· · · O interactions. Struct Chem 22:757–763
27. Rosenfield Jr RE Parthasarathy R Dunitz JD (1977) Directional preferences of nonbonded
atomic contacts with divalent sulfur. 1. Electrophiles and nucleophiles. J Am Chem Soc
99:4860–4862
28. Allen FH (2002) The Cambridge structural database: a quarter of a million crystal structures
and rising. Acta Crystallogr Sect B 58:380–388
29. Row TNG, Parthasarathy R (1981) Directional preferences of nonbonded atomic contacts with
divalent sulfur in terms of its orbital orientations. 2. S· · · S interactions and nonspherical shape
of sulfur in crystals. J Am Chem Soc 103:477–479
30. Desiraju GR, Nalini V (1991) Database analysis of crystal-structure-determining interactions
involving sulphur: implications for the design of organic metals. J Mater Chem 1:201–203
31. Dahaoui S, Pichon-Pesme V, Howard JAK, Lecomte C (1999) CCD charge density study on
crystals with large unit cell parameters: the case of hexagonal L-cystine. J Phys Chem A
103:6240–6250
32. Kucsman A, Kapovits I (1985) Nonbonded sulfur-oxygen interaction in organic sulfur com-
pounds. In: Bernardi F, Csizmadia IG, Mangini A (eds) Organic sulfur chemistry: theoretical
and experimental advances. Elsevier, Amsterdam, pp 191–245
33. Zauhar RJ, Colbert CL, Morgan RS, Welsh WJ (2000) Evidence for a strong sulfur-aromatic
interaction derived from crystallographic data. Biopolymers 53:233–248
34. Iwaoka M, Takemoto S, Tomoda S (2002) Statistical and theoretical investigations on the
directionality of nonbonded S· · · O interactions. Implications for molecular design and protein
engineering. J Am Chem Soc 124:10613–10620
35. Burling FT, Goldstein BM (1992) Computational studies of nonbonded sulfur-oxygen and
selenium-oxygen interactions in the thiazole and selenazole nucleosides. J Am Chem Soc
114:2313–2320
36. Nakanishi W, Hayashi S, Narahara K (2008) Atoms-in-molecules dual parameter analysis
of weak to strong interactions: behaviors of electronic energy densities versus Laplacian of
electron densities at bond critical points. J Phys Chem A 112:13593–13599
37. Bader RFW (1990) Atoms in molecules: a quantum theory. Oxford University Press: NewYork
38. Gillespie RJ, Popelier PLA (2001) Chemical bonding and molecular geometry. Oxford
University Press: New York
39. Jeziorski B, Moszynski R, Szalewicz K (1994) Perturbation theory approach to intermolecular
potential energy surfaces of van der Waals complexes. Chem Rev 94:1887–1930
40. Scheiner S (2011) On the properties of X· · · N noncovalent interactions for first-, second-, and
third-row X atoms. J Chem Phys 134:164313
41. Adhikari U, Scheiner S (2011) The S· · · N noncovalent interaction: comparison with hydrogen
and halogen bonds. Chem Phys Lett 514:36–39
42. Nagao Y, Hirata T, Goto S, Sano S, Kakehi A, Iizuka K, Shiro M (1998) Intramolecular non-
bonded S· · · O interaction recognized in (acylimino)thiadiazoline derivatives as angiotensin II
receptor antagonists and related compounds. J Am Chem Soc 120:3104–3110
43. Wu S, Greer A (2000) Attractive through-space S–O interaction in the DNA-cleaving antitumor
antibiotic leinamycin. J Org Chem 65:4883–4887
44. Iwaoka M, Tomoda S (1996) Nature of the intramolecular Se· · · N nonbonded interaction of
2-selenobenzylamine derivatives. An experimental evaluation by 1H, 77Se, and 15N NMR
spectroscopy. J Am Chem Soc 118:8077–8084
45. Iwaoka M, Komatsu H, Katsuda T, Tomoda S (2004) Nature of nonbonded Se· · · O interac-
tions characterized by 17O NMR spectroscopy and NBO and AIM analyses. J Am Chem Soc
126:5309–5317
46. Iwaoka M, Komatsu H, Katsuda T, Tomoda S (2002) Quantitative evaluation of weak non-
bonded Se· · · F interactions and their remarkable nature as orbital interactions. J Am Chem
Soc 124:1902–1909
47. Mugesh G, Panda A, Singh HB, Punekar NS, Butcher RJ (2001) Glutathione peroxidase-like
antioxidant activity of diaryl diselenides: a mechanistic study. J Am Chem Soc 123:839–850
288 M. Iwaoka
48. Takei T, Urabe Y, Asahina Y, Hojo H, Nomura T, Dedachi K, Arai K, Iwaoka M (2014) Model
study using designed selenopeptides on the importance of the catalytic triad for the antioxidative
functions of glutathione peroxidase. J Phys Chem B 118:492–500
49. Dobson CM, Šali A, Karplus M (1998) Protein folding: a perspective from theory and
experiment. Angew Chem Int Ed 37:868–893
50. Derewenda ZS, Lee L, Derewenda U (1995) The occurrence of C–H· · · O hydrogen bonds in
proteins. J Mol Biol 252:248–262
51. Bella J, Berman HM (1996) Crystallographic evidence for C(α)–H· · · O = C hydrogen bonds
in a collagen triple helix. J Mol Biol 264:734–742
52. Vargas R, Garza J, Dixon DA, Hay BP (2000) How strong is the C(α)–H· · · O = C hydrogen
bond? J Am Chem Soc 122:4750–4755
53. Burley SK, Petsko GA (1986) Amino-aromatic interactions in proteins. FEBS Lett 203:139–
143
54. Flocco MM, Mowbray SL (1994) Planar stacking interactions of arginine and aromatic side-
chains in proteins. J Mol Biol 235:709–717
55. Gallivan JP, Dougherty DA (1999) Cation-π interactions in structural biology. Proc Nat Acad
Sci U S A 96:9459–9464
56. Gallivan JP, Dougherty DA (2000) A computational study of cation-π interactions vs salt
bridges in aqueous media: Implications for protein engineering. J Am Chem Soc 122:870–874
57. Umezawa Y, Nishio M (1998) Ch/π interactions as demonstrated in the crystal structure of
guanine- nucleotide binding proteins, src homology-2 domains and human growth hormone in
complex with their specific ligands. Bioorg Med Chem 6:493–504
58. Takahashi O, KohnoY, Nishio M (2010) Relevance of weak hydrogen bonds in the conformation
of organic compounds and bioconjugates: evidence from recent experimental data and high-
level ab initio MO calculations. Chem Rev 110:6049–6076
59. Brandl M, Weiss MS, Jabs A, Sühnel J, Hilgenfeld R (2001). C-H· · · π-interactions in proteins.
J Mol Biol 307:357–377
60. Morgan RS, Tatsch CE, Gushard RH, McAdon J, Warme PK (1978) Chains of alternating
sulfur and pi-bonded atoms in eight small proteins. Int J Peptide Protein Res 11:209–217
61. Reid KSC, Lindley PF, Thornton JM (1985) Sulphur-aromatic interactions in proteins. FEBS
Lett 190:209–213
62. Viguera AR, Serrano L (1995) Side-chain interactions between sulfur-containing amino acids
and phenylalanine in α-helices. Biochemistry 34:8771–8779
63. Cheney BV, Schulz MW, Cheney J (1989) Complexes of benzene with formamide and
methanethiol as models for interactions of protein substructures. Biochim Biophys Acta
996:116–124
64. Iwaoka M, Takemoto S, Okada M, Tomoda S (2002) Weak nonbonded S· · · X (X = O, N, and S)
interactions in proteins. Statistical and theoretical studies. Bull Chem Soc Jpn 75:1611–1625
65. Iwaoka M, Isozumi N (2006) Possible roles of S· · · O and S· · · N interactions in the functions
and evolution of phospholipase A2. Biophysics 2:23–34
66. Iwaoka M, Isozumi N (2012) Hypervalent nonbonded interactions of a divalent sulfur atom.
Implications in protein architecture and the functions. Molecules 17:7266–7283
67. Iwaoka M, Babe N (2015) Mining and structural characterization of S···X chalcogen bonds in
protein database. Phosphorus Sulfur Silicon Relat Elem, in press.
68. Hobohm U, Scharf M, Schneider R, Sander C (1992) Selection of representative protein data
sets. Protein Sci 1:409–417
69. Hobohm U, Sander C (1994) Enlarged representative set of protein structures. Protein Sci
3:522–524
70. Richardson Laboratory, 3D analysis:: 500 filtered structures. http://kinemage.biochem.duke.
edu/databases/top500.php. Accessed Aug 2014
71. Jaffe RL, Smith GD (1996) A quantum chemistry study of benzene dimer. J Chem Phys
105:2780–2788
72. Schwenke DW, Truhlar DG (1985) Systematic study of basis set superposition errors in the
calculated interaction energy of two HF molecules. J Chem Phys 82:2418–2426
74. Flohé L, Günzler WA, Schock HH (1973) Glutathione peroxidase: a selfnoenzyme. FEBS Lett
32:132–134
75. Savaskan NE, Ufer C, Kühn H, Borchert A (2007) Molecular biology of glutathione peroxi-
dase 4: from genomic structure to developmental expression and neural function. Biol Chem
388:1007–1017
76. Conrad M, Schneider M, Seiler A, Bornkamm GW (2007) Physiological role of phospholipid
hydroperoxide glutathione peroxidase in mammals. Biol Chem 388:1019–1025
77. Toppo S, Flohé L, Ursini F, Vanin S, Maiorino M (2009) Catalytic mechanisms and specificities
of glutathione peroxidases: variations of a basic scheme. Biochim Biophys Acta 1790:1486–
1500
78. Holmgren A, Lu J (2010) Thioredoxin and thioredoxin reductase: current research with special
reference to human disease. Biochem Biophys Res Commun 396:120–124
79. Prast-Nielsen S, Huang HH, Williams DL (2011) Thioredoxin glutathione reductase: its role in
redox biology and potential as a target for drugs against neglected diseases. Biochim Biophys
Acta 1810:1262–1271
80. Taylor JC, Markham GD (1999) The bifunctional active site of S-adenosylmethionine
synthetase. Roles of the active site aspartates. J Biol Chem 274:32909–32914
81. Brandt W, Golbraikh A, Täger M, Lendeckel U (1999) A molecular mechanism for the cleavage
of a disulfide bond as the primary function of agonist binding to G-protein-coupled receptors
based on theoretical calculations supported by experiments. Eur J Biochem 261:89–97
82. NagaoY, Honjo T, Iimori H, Goto S, Sano S, Shiro M,Yamaguchi K, SeiY (2004) Intramolecular
nonbonded S· · · O interaction in acetazolamide and thiadiazolinethione molecules in their
dimeric crystalline structures and complex crystalline structures with enzymes. Tetrahedron
Lett 45:8757–8761
83. Singh AK, Singh N, Sharma S, Shin K, Takase M, Kaur P, Srinivasan A, Singh TP
(2009) Inhibition of lactoperoxidase by its own catalytic product: Crystal structure of the
hypothiocyanate-inhibited bovine lactoperoxidase at 2.3-Å resolution. Biophys J 96:646–654
84. Nakamura T, Yamamoto T, Abe M, Matsumura H, Hagihara Y, Goto T, Yamaguchi T, Inoue T
(2008) Oxidation of archaeal peroxiredoxin involves a hypervalent sulfur intermediate. Proc
Nat Acad Sci U S A 105:6238–6242
85. Palenchar JB, Crocco JM, Colman RF (2003) The characterization of mutant Bacillus subtilis
adenylosuccinate lyases corresponding to severe human adenylosuccinate lyase deficiencies.
Protein Sci 12:1694–1705
86. Dennis EA (1994) Diversity of group types, regulation, and function of phospholipase A2.
J Biol Chem 269:13057–13060
87. Jain MK, Gelb MH, Rogers J, Berg OG (1995) Kinetic basis for interfacial catalysis by
phospholipase A2. Methods Enzymol 249:567–614
88. Yu BZ, Pan YH, Janssen MJW, Bahnson BJ, Jain MK (2005) Kinetic and structural properties
of disulfide engineered phospholipase A2: insight into the role of disulfide bonding patterns.
Biochemistry 44:3369–3379
89. Raines RT (1998) Ribonuclease A. Chem Rev 98:1045–1065
90. Xu X, Rothwarf DM, Scheraga HA (1996) Nonrandom distribution of the one-disulfide
intermediates in the regeneration of ribonuclease A. Biochemistry 35:6406–6417
91. Li YJ, Rothwarf DM, Scheraga HA (1995) Mechanism of reductive protein unfolding. Nat
Struct Biol 2:489–494
92. De Meyts P (2004) Insulin and its receptor: structure, function and evolution. BioEssays
26:1351–1362
93. Merlini G, Bellotti V (2005) Lysozyme: a paradigmatic molecule for the investigation of protein
structure, function and misfolding. Clin Chim Acta 357:168–172
94. Ohno M, Chijiwa T, Oda-Ueda N, Ogawa T, Hattori S (2003) Molecular evolution of myotoxic
phospholipases A2 from snake venom. Toxicon 42:841–854
Chapter 10
A Unified View of Halogen Bonding, Hydrogen
Bonding and Other σ-Hole Interactions
Peter Politzer and Jane S. Murray
Abstract The term “σ-hole” refers to a region of diminished electronic density along
the extension of a covalent single bond to a hydrogen or an atom of Groups IV—VII.
This region often has a positive electrostatic potential through which the atom can in-
teract attractively with a negative site (such as a lone pair of a Lewis base, π electrons
or an anion) to form a noncovalent complex. Hydrogen bonding and halogen bonding
are the most prominent examples of such σ-hole interactions, although they have long
been known experimentally for Groups IV—VI as well (but without the σ-hole label).
σ-Holes result from the anisotropic charge distributions of covalently-bonded atoms.
It follows from the Hellmann-Feynman theorem that σ-hole interactions can be un-
derstood and described as Coulombic, which includes polarization and dispersion.
In the context of noncovalent interactions, charge transfer is simply a mathematical
formulation of polarization.
10.1 Noncovalent Interactions: The Chemistry

of the Twenty-First Century
Recent years have seen a remarkable surge of research activity in the area of nonco-
valent interactions. One indication of this is the number of review and perspective
papers that have appeared just since the year 2000; some of them (certainly not all)
are cited here [1–17]. Indeed, an observation made by Schneider appears thus far to
be borne out: “. . . one might assert that the chemistry of the last century was largely
the chemistry of covalent bonding, whereas that of the present century is more likely
to be the chemistry of noncovalent binding.” [18].
Much of this recent interest has focused upon what is known as “halogen bond-
ing,” a highly-directional attractive interaction that is found to occur between many
covalently-bonded halogen atoms and negative sites (lone pairs of Lewis bases, π
electrons, anions, etc.). The interest is due in part to the applications of halogen
bonding, existing and potential, in fields such as crystal engineering, molecular bi-
ology and pharmacology [5, 8, 10, 15], but it also reflects the intriguing fact that an
P. Politzer () · J. S. Murray

Department of Chemistry, University of New Orleans, 70148 New Orleans, LA, USA
e-mail: ppolitze@uno.edu

292 P. Politzer and Jane S. Murray
electronegative halogen is attracted to a negative site. This was sometimes described

as an enigma!
Our primary focus in this chapter will be upon halogen bonding. We emphasize,
however, that this is simply a subset of a much larger category of noncovalent in-
teractions, “σ-hole bonding,” that can involve covalently-bonded atoms of Groups
IV—VI as well as Group VII (the halogens) and also includes hydrogen bonding.
The nature of the interaction is essentially the same in all of these cases, as shall be
discussed. The formation of noncovalent complexes between Group IV—VII atoms
and negative sites has been known to experimentalists for decades, although the
unifying σ-hole interpretation is of more recent origin.
10.2 Some Historical Background
A very early example of what we now call halogen bonding was the report, in 1814,
of an interaction between iodine and ammonia [19]. The product was later purified
by Guthrie [20] and formulated as I2 ·NH3 . Additional complexes of Cl2 , Br2 and
I2 with amines were subsequently observed [21, 22], as well as an adduct between
iodoform and quinoline [23].
The twentieth century saw many more R-X—B systems being identified, where
RX is a halide molecule and B is a Lewis base; a number of these studies are
cited by Blackstock et al [24]. Of particular note was the spectral characterization
of complexes of molecular iodine with benzene and other aromatic hydrocarbons
by Benesi and Hildebrand [25, 26]. Their observations were a factor in Mulliken’s
development of his “charge-transfer” formalism [27], which is frequently invoked
as a mathematical (but not physical) description of noncovalent bonding. However
Mulliken incorrectly suggested that in the I2 —C6 H6 complex, the I2 axis is parallel
to the benzene plane.
Important advances in elucidating the structure of halogen bonding have come
through crystallography. In the 1960’s, Hassel and his colleagues determined the
crystal structures of a number of complexes between organic halides and oxy-
gen/nitrogen Lewis bases; this work contributed to his receiving the Nobel Prize
in 1969. These studies showed the crystals to be composed of chain-like structures
held together by weak X–B bonds, e.g. 1 [28]. For a review, see Bent [29]. At about the
same time, attractive Br—O intermolecular interactions were found in solid POBr3
[30] and Cl—O in solid POCl3 [31].
Br Br
N N Br C=C Br N N Br C=C Br
Br Br
1
10 A Unified View of Halogen Bonding, Hydrogen Bonding . . . 293
A few years later, Murray-Rust et al conducted extensive surveys of halide (RX;

X = Cl, Br, I) crystal structures, using the Cambridge Structural Database [32–34].
Their objective was to identify close contacts between halogen atoms and atoms of
neighboring molecules. Close contact was defined as being less than the sum of the
respective van der Waals radii, and was taken to reflect an attractive interaction. A very
significant pattern was revealed: Close contacts of halogen atoms with electrophilic
sites, such as metal ions, were primarily at angles of 90–120◦ relative to the R–X
bond (2); close contacts with nucleophilic sites, e.g. nitrogens and oxygens, tended
to be nearly linear, the angles being 160–180◦ (3).
electrophile (positive site)
R X R X nucleophile (negative site)

2 3
Thus a given halogen atom can interact favorably, but in different directions, with
both positive and negative sites.
The near-linear interactions of covalently-bonded halogens with nucleophiles, as
in 3, are what we call halogen bonding, R-X—B. To our knowledge, this term was
first used in 1976 [35, 36]. In the past, halogen bonding was sometimes viewed
as very puzzling, since univalent halogens are generally regarded as being negative
in character; how then can they be attracted to negative sites? This question was
answered in 1992, as shall be discussed in Sect. 10.3.2.
During the last two decades, the practical significance and potential applica-
tions of halogen bonding have increasingly been recognized. Imakubo et al prepared
semiconducting crystals and a superconductor by linking iodine-containing organic
sulfides via negative ions (acting as the bases) [37, 38]. An important development
was the realization that the role of halogen bonding in formulating new materials is en-
hanced if the halogen-containing molecule also contains strongly electron-attracting
groups, e.g. perfluorination [39]. This helped to stimulate a great deal of activity in
the area of crystal engineering, with applications in electronics, nonlinear optical
activity, magnetic materials, liquid crystals, etc. [5, 8, 10, 40–42]. Catalysis through
halogen bonding is also being explored [43, 44].
In chemical biology as well, there is a growing awareness of the importance of
halogen bonding. A milestone was the survey of the Protein Data Bank by Auffinger
et al [45], revealing C-X—O close contacts (X = Cl, Br, I); it was subsequently
expanded by Lu et al to C-X—Y (X = Cl, Br, I; Y = O, N, S) [46]. Such interactions
are now known to affect protein-ligand binding, recognition and assembly processes,
docking, conformational stability, protein folding, etc. [5, 15, 41, 47], and are being
exploited in drug design.
During roughly the same years as the surveys of halide crystal structures by
Murray-Rust et al [32–34], discussed above, analogous studies of close contacts in
organic sulfides and selenides were being carried out [48–50]. Very similar results
were obtained. Electrophilic sites interacted with the lateral sides of the sulfur or
selenium (4), while interactions with nucleophiles were close to linear, along the
extensions of the R1 –S(Se) and/or R2 –S(Se) bonds (5). Some of the close contacts
were between a sulfur in one molecule and the same sulfur in another identical
molecule [49], i.e. “like attracting like” (6); the interactions were linear for one of
the sulfurs and lateral for the other.
electrophile (positive site)

R1 R1 (Se)
S(Se) S
R2 R2 nucleophile (negative site)
4 5
R1
S
R2
R1
S
R2 6
10.3 The Electrostatic Potential and the σ-Hole
All of these observations were explained and the “enigma” of halogen bonding was
resolved in 1992—through the analysis of molecular electrostatic potentials. We will
accordingly give some brief background concerning this property.
10.3.1 The Electrostatic Potential
The nuclei and electrons of a molecule (or other system) create an electrostatic
potential V(r) at each point r in the surrounding space:
ZA
ρ(r )dr
V(r) = − (10.1)
A
|RA − r| |r − r|
In Eq. (10.1), which is simply one form of Coulomb’s Law, ZA is the charge on
nucleus A, located at RA , and ρ(r) is the molecule’s electronic density distribution.
V(r) is positive or negative in any given region depending upon whether the effects
of the positive nuclei or the negative electrons are dominant there.
If a point charge Q is placed at r, then its interaction energy E with the molecule’s
nuclei and electrons is,
E = QV(r) (10.2)
If Q and V(r) have the same sign (positive or negative), then E > 0 and the inter-
action is repulsive; if they have opposite signs, then E < 0 and the interaction is
attractive.
The electrostatic potential is accordingly an effective means of predicting close
contacts and noncovalent interactions [51–53]. In general, regions of positive V(r)
tend to interact favorably (at least initially) with negative sites and negative V(r) with
positive sites. (A more detailed analysis needs to take polarization into account, as
shall be discussed.)
We want to emphasize that the electrostatic potential is a real property of a system,
a physical observable. It can be determined experimentally by diffraction methods
[54–56], as well as computationally.
When V(r) is used for interpreting and predicting noncovalent interactions,
it is typically computed on the molecular “surface” defined by the 0.001 au
(electrons/bohr3 ) contour of the molecule’s electronic density ρ(r). This was sug-
gested by Bader et al [57] as a reasonable representation of a molecular surface;
it encompasses roughly 97 % of the electronic charge and reflects specific features
such as lone pairs, π electrons and atomic anisotropies. V(r) on the 0.001 au surface
is labeled VS (r), and its locally most positive and most negative values (of which
there may be several) are designated by VS,max and VS,min , respectively.
10.3.2 The σ-Hole
In 1992, Brinck et al reported something quite unexpected [58]: covalently-bonded

halogens with regions of positive electrostatic potential on their outer sides, on the
extensions of the covalent bonds. The lateral sides of the halogens were negative, as
anticipated. Additional studies in the next two years produced similar results [59, 60].
See, for instance, Fig. 10.1; the outer sides of the bromines, on the extensions of the
C–Br bonds, have positive VS (r) while the lateral sides of the bromines are negative
(although less so after perfluorination).
Such positive outer regions appear to contradict the common view that univalent
halogens are negative in character. However, as Brinck et al pointed out [58, 59],
these findings do provide an immediate explanation of halogen bonding and of the
patterns of crystallographic close contacts observed by Murray-Rust et al [32–34],
i.e. structures 2 and 3. Halogen bonding is the attractive interaction between the
positive outer portion of the halogen and a negative site. Since the positive region
on the halogen surface is on the extension of the covalent bond to the halogen, the
same will be true of the interaction with the negative site; it will be near-linear, as
shown in 3. The halogen can also interact favorably with positive sites through its
negative lateral sides, as in 2 (except for occasional instances in which the halogen’s
entire surface is positive). Some years later, Auffinger et al [45] and Awwadi et al
[61] presented analogous intepretations of halogen bonding.
In 2007, the term “σ-hole” was introduced to denote the outer regions of positive
potential on univalent halogens [62, 63]. The name reflects their being along the
Fig. 10.1 Computed

electrostatic potentials on
0.001 au molecular surfaces
of a 1,4-dibromo-n-butane
and b 1,4-dibromoperfluoro-
n-butane. Bromines are at far
left and far right. Black
hemispheres indicate positive
potential maxima on
bromines, on extensions of
C–Br bonds. Color ranges, in
kcal/mol: red, greater than 25;
yellow, between 25 and 15;
green, between 15 and 0;
blue, less than 0 (negative).
Note that bromines become
more positive after
perfluorination.
Computational level:
M06-2X/6-311G(d, p)
extensions of σ bonds. As will be pointed out, however, σ-holes can sometimes also
have negative potentials.
At about the same time that regions of positive potential were found on the exten-
sions of covalent bonds to halogens [58–60], an analogous discovery was made for
divalent sulfur. In a computational analysis of the heterocycle 7 [64], Burling and
Goldstein showed that the sulfur has positive potentials on the extensions of the C–S
bonds, with negative ones on its lateral sides. It was noted that this is consistent with
the crystallographically-revealed patterns of close contacts in divalent sulfides and
selenides [48–50], structures 4–6; it also explains intramolecular S—O and Se—O
close contacts that stabilize certain biologically-active thiazole and selenazole con-
formations [64]. (See also a later computational study involving other substituted
sulfur—and selenium-containing heterocycles [65].) The basis for “like attracting
like,” 6, now becomes apparent—for halides as well as for sulfides and selenides:
The positive outer region on one of these atoms interacts favorably with a negative
side of the same atom in another, identical molecule, as in 6.
Fig. 10.2 Computed electrostatic potential on 0.001 au molecular surface of HSeCN. The selenium
is in the foreground, cyano group at the right. Black hemispheres indicate selenium positive σ-
holes on extensions of C–Se and H–Se bonds, the former being more positive. Color ranges, in
kcal/mol: red, greater than 25; yellow, between 25 and 15; green, between 15 and 0; blue, less than
0 (negative). Computational level: M06-2X/6-311G(d, p)
What Burling and Goldstein described are simply more examples of what are now
called positive σ-holes: Regions of positive electrostatic potential on the outsides of
singly-bonded atoms, along the extensions of the covalent bonds. Between 2007 and
2009, numerous positive σ-holes were found computationally on atoms of Groups
IV—VI as well as Group VII (the halogens); this work in summarized in several
reviews [11, 15, 16]. Through these regions, the atoms can interact favorably with
negative sites (lone pairs of Lewis bases, π electrons, anions, etc.), forming non-
covalent complexes. This is σ-hole bonding, of which halogen bonds are a subset.
The interactions are highly directional, close to linear, along the extensions of the
covalent single bonds to the atoms (see 3 and 5).
Since Group IV—VI atoms can form four, three and two single bonds (or more if
hypervalent [66, 67]), they can have the same numbers of σ-holes on the extensions
of these bonds. This can be seen in Figs. 10.2, 10.3 and 10.4, which show that the
σ-holes on a given atom can have different potentials, depending upon the partners
in the bonds that gave rise to the σ-holes.
Some reports in the recent literature describe noncovalent interactions between
Group IV—VI atoms and negative sites as new discoveries. This is incorrect (an error
of which we have also been guilty [68]). Experimentalists have been familiar with
Fig. 10.3 Computed

electrostatic potential on
0.001 au molecular surface of
PF2 Br. The phosphorus is in
the foreground, bromine at
top. Black hemispheres
indicate phosphorus positive
σ-holes on extensions of Br–P
and F–P bonds, the former
being more positive. Color
ranges, in kcal/mol: red,
greater than 25; yellow,
between 25 and 15; green,
between 15 and 0; blue, less
than 0 (negative).
M06-2X/6-311G(d, p)
them for decades, as is well documented in two reviews [16, 69]. What is new is the
unifying σ-hole interpretation of a great many of these interactions [11,14–17].
10.4 Origins of σ-Holes
10.4.1 Anisotropies of Covalently-Bonded Atoms
The electronic density of a free neutral atom is, on the average, spherically-
symmetrical [70]. The electrostatic potential V(r) due to its nucleus and electrons
is positive for all r < ∞ [71], the effect of the nucleus dominating over that of the
dispersed electrons. When two atoms begin to interact, at large separations, the elec-
tronic density of each of them becomes somewhat polarized toward the other [72], in
response to the electric field of the latter. This results in the electronic density being
less on the outer side of each atom (along the internuclear axis) than on its lateral
sides. These diminished outer electron densities are incipient σ-holes!
As the atoms continue to approach and to interact in forming a covalent single
bond, there is further redistribution of electronic densities, depending upon the atoms’
relative polarizabilities and electronegativities. However the σ-holes continue to have
lesser electronic densities than the surrounding portions of the atoms. It is indeed
well established that covalently-bonded atoms have anisotropic charge distributions
[14, 17, 49, 50, 61, 73–79]; their “radii” are less along the extensions of single bonds
than perpendicular to them. For example, in the bromines in Fig. 10.1a, the distances
to the 0.001 au surface are 2.06 Å along the extensions of the C–Br bonds and 2.28
Å in the perpendicular directions. In Fig. 10.1b, the same distances are 2.02 Å and
Fig. 10.4 Two views of the

computed electrostatic
potential on the 0.001 au
molecular surface of SiH3 CN.
In a the silicon is in the
foreground, cyano group at
right. In b a hydrogen is in
the foreground, cyano group
at top. Black hemispheres
indicate silicon positive
σ-holes on extensions of C–Si
and H–Si bonds, the former
being more positive. Color
ranges, in kcal/mol: red,
greater than 25; yellow,
between 15 and 0; blue, less
than 0 (negative).
M06-2X/6-311G(d, p)
2.24 Å. These values show the effect of the fluorines in attracting electronic density
from the bromines.
The electronic density in a σ-hole is often sufficiently low that a positive electro-
static potential VS (r) results (Figs. 10.1, 10.2, 10.3 and 10.4). This is then a positive
σ-hole, and favorable interactions with negative sites can be anticipated. In some in-
stances, on the other hand, the electronic density in the σ-hole is high enough that its
VS (r) is negative, i.e. a negative σ-hole, although less negative than its surroundings;
see Fig. 10.5a. Examples of the VS,max of both positive and negative σ-holes are in
Table 10.1 for the halogens [80, 81] and in Table 10.2 for Group IV—VI atoms [81].
The tables also give the most negative potentials (VS,min ) on these atoms.
It is very important to keep in mind, however, that the VS (r) computed for a free
molecule in its unperturbed ground state, e.g. Figs. 10.1, 10.2, 10.3, 10.4 and 10.5,
does not reflect the polarization induced by the electric field of another molecule
Fig. 10.5 Two views of the

computed electrostatic
potential on the 0.001 au
molecular surface of FOH. In
a the fluorine is in the
foreground; the black
hemisphere indicates its
σ-hole, which is negative, on
the extension of the O–F
bond. In b the hydrogen is to
the left; the black hemisphere
on the hydrogen indicates the
positive σ-hole on the
extension of the O–H bond.
Color ranges, in kcal/mol:
red, greater than 30; yellow,
between 0 and—8; blue, more
negative than—8.
M06-2X/6-311G(d, p)
when they begin to interact. It has been demonstrated [82, 83] that polarization due
to an external field can strengthen or weaken a σ-hole, and can sometimes convert it
from negative (positive) to positive (negative). Thus one can occasionally find that
a σ-hole complex forms even though the σ-hole may have been negative or near-
neutral prior to the interaction; the negative site induced the positive σ-hole [84].
Some examples are H3 C-Cl—O = CH2 [85] and H3 P—NSH [86], in which positive
σ-holes were induced on the chlorine of H3 C–Cl and the phosphorus of H3 P.
Table 10.1 Most positive σ-hole potentials (VS,max ) and most negative potentials (VS,min ) on 0.001
au surfaces of the halogens shown in bold type, in kcal/mol. Computational method:M06–2X/6-
311G(d, p)
Molecule Bond producing σ-hole VS,max (σ-hole) VS,min Reference
HO–F O–F − 6.9 − 19.7 Present work
F–F F–F 11.2 − 2.6 17
F–Cl F–Cl 45.1 + 0.3 80
F–Br F–Br 53.2 − 0.1 80
Cl–Cl Cl–Cl 25.4 − 2.7 80
Br–Br Br–Br 29.0 − 4.1 80
H3 C–F C–F − 24.6 − 25.3 Present work
H3 C–Cl C–Cl − 0.9 − 15.6 Present work
H3 C–Br C–Br 5.6 − 14.8 Present work
Br–(CH2 )4 –Br C–Br 5.8 − 13.5 Present work
F3 C–F C–F − 1.3 − 2.8 81
F3 C–Cl C–Cl 19.9 − 0.8 81
F3 C–Br C–Br 25.3 − 2.0 81
Br–(CF2 )4 –Br C–Br 26.3 − 1.7 Present work
F3 C–I C–I 31.9 − 1.9 81
NC–F C–F 12.8 10.9 17
NC–Cl C–Cl 36.0 10.3 17
NC–Br C–Br 42.7 8.6 17
NC–I C–I 48.7 7.1 17
H3 Si–Br Si–Br 0.2 − 11.9 Present work
H3 Ge–Br Ge–Br − 2.1 − 13.7 Present work
H2 P–Br P–Br 3.5 − 14.3 Present work
F2 P–Br P–Br 9.6 − 7.6 80
Cl–C≡C–Cl C–Cl 23.7 − 0.6 Present work
Br–C≡C–Br C–Br 30.1 − 2.1 17
C6 H5 –Cl C–Cl 4.5 − 13.7 Present work
C6 H5 –Br C–Br 10.2 − 13.7 Present work
10.4.2 Trends in σ-Hole Potentials
Since σ-holes are created by the shifting of electronic charge away from the outer
portion of a singly-bonded atom, it is reasonable that the σ-hole electrostatic poten-
tial should become more positive as the atom is more polarizable and as it is less
electronegative relative to the remainder of the molecule (particularly its bonding
Table 10.2 Most positive σ-hole potentials (VS,max ) and most negative potentials (VS,min ) on 0.001
au surfaces of the Group IV—VI atoms shown in bold type, in kcal/mol. Computational method:
M06-2X/6-311G(d, p). The data are from the present work except as otherwise indicated
Molecule Bond producing σ-hole VS,max (σ-hole) VS,min
FOBr F–O 10.5 − 11.7
Br–O −3.4 − 11.7
FSBr F–S 35.9 − 4.0
Br–S 27.2 − 4.0
FSeBr F–Se 43.5 − 4.9
Br–Se 35.2 − 4.9
HSeCN H–Se 26.4 − 2.3
C–Se 43.5 − 2.3
F2 NBr F–N 13.2 − 9.4
Br–N 11.8 − 9.4
H2 PBr H–P 18.6 − 9.3
Br–P 35.0 − 9.3
F2 PBr F–P 32.3 11.4
Br–P 35.1 11.4
F2AsBr F–As 39.7 20.9
Br–As 38.6 20.9
F3 CBr F–C 16.0a nonea
a
Br–C 21.6 nonea
F3 SiBr F–Si 39.0a nonea
a
Br–Si 45.0 nonea
F3 GeBr F–Ge 45.3a nonea
a
Br–Ge 44.5 nonea
H3 SiCN H–Si 26.8 none
C–Si 37.7 none
a
Reference 81
partner). Two common generalizations are [11,15–17]: (1) Within a given Group and
for a particular molecular framework, the σ-hole potential (VS,max ) tends to become
more positive in going from the lighter to the heavier (more polarizable and less
electronegative) atoms, and (2) for a given atom, VS,max increases as the remainder
of the molecule becomes more electron-attracting. These are useful empirical gen-
eralizations; supporting evidence can be seen in Tables 10.1 and 10.2. They explain,
for example, why much of the early crystal engineering involving halogen bonding
focused upon iodine, but in perfluorinated molecules [5, 8, 39]. Figure 10.1 and
Table 10.1 show that perfluorination of the carbon chain of 1,4-dibromo-n-butane
increases the bromine VS,max from 5.8 to 26.3 kcal/mol. The empirical generaliza-
tions also account for the fact that changing the bonding partner of bromine from
the more electronegative carbon to silicon and germanium (in H3 C–Br, H3 Si–Br
and H3 Ge–Br) causes the bromine σ-hole VS,max to become less positive and even
negative (Table 10.1).
The first-row atoms N, O and F, being least polarizable and most electronegative,
often have negative σ-holes (Fig. 10.5). It was indeed believed at one time that
fluorine does not halogen bond. However it is well established since 2007 that fluorine
can have a positive σ-hole and can participate in halogen bonding when it is in a
sufficiently electron-attracting molecular environment [87–89]. In FCN, in fact, the
portion of the molecular surface that is associated with the fluorine is entirely positive
[87]; this is true as well of the other halocyanides (Table 10.1), demonstrating that
positive σ-holes are not always surrounded by negative regions but sometimes simply
by less positive ones. It has also been known since 2007–2009 that carbon [67],
nitrogen [68] and oxygen [90] can also have positive σ-holes in appropriate molecular
environments, and participate in σ-hole interactions. In the case of tetravalent Group
IV atoms (which have no lone pairs), we have customarily found their entire exposed
surfaces to be positive [67, 81].
It must be acknowledged that the generalizations concerning trends in σ-hole po-
tentials, mentioned above, are somewhat oversimplified (although frequently valid).
Consider, for instance, the molecule F2 PBr in Table 10.2 and Fig. 10.3. The phospho-
rus σ-hole produced by the bond to the bromine is more positive than that due to the
bond to the fluorine, despite the fluorine’s greater electronegativity. A detailed anal-
ysis [80] has shown that it is not only the polarizability and relative electronegativity
of the σ-hole atom that are involved, but also the polarizability and electron-attracting
power of the remainder of the molecule, plus factors such as overlapping with the
electrostatic potentials of other parts of the molecule.
10.4.3 Covalently-Bonded Hydrogen
σ-Holes are not necessarily limited to singly-bonded atoms of Groups IV—VII. In

particular, we must consider covalently-bonded hydrogen. Its electronic charge dis-
tribution, which involves just a single electron, is certainly anisotropic; it is centered
in the internuclear region. This is why standard crystallographic techniques under-
estimate the lengths of covalent bonds to hydrogens [91, 92]. Accordingly, there is
generally a positive σ-hole on the outer side of a hydrogen. This was in fact ob-
served for the hydroxyl hydrogen in ethanol already in 1991 [93], although it was
not labeled a σ-hole. Figure 10.5b clearly shows the hydrogen σ-hole in FOH. In
some instances, as in Fig. 10.4b, a VS,max is not found on a hydrogen surface because
the hydrogen’s positive potential is overlapped by that of a larger neighboring atom.
In Fig. 10.1a, the hydrogens do have VS,max but they are not shown explicitly. The
interactions of such positive regions with negative sites readily explains hydrogen
bonding as a σ-hole interaction [14, 15, 94, 95]; indeed, electrostatic interpretations
of hydrogen bonding long preceded the discovery of σ-holes [55,93,96–100].
Since hydrogen has only the one valence electron, its lateral sides have relatively
low electronic densities and the positive σ-hole usually covers a larger area than is
typical of the Group V—VII atoms. Thus hydrogen σ-holes are less focused, as can
be seen in Fig. 10.5b and in earlier reports [11, 15, 17, 94, 101]. It is accordingly
not surprising that hydrogen bonds are overall less directional than are other σ-hole
bonds [94, 102, 103].
10.5 σ-Hole Interactions
10.5.1 Interaction Energies
Attractive interactions between positive σ-holes and negative sites can give rise to
numerous noncovalent complexes. Some examples are presented in Table 10.3 for
the complexes R-X—B, R-Y—B and R-H—B, where X is a halogen, Y is a Group
IV—VI atom and B represents the negative sites; in Table 10.3, these are the lone
pairs of NH3 , HCN, (H3 C)2 O and H2 S, the π electrons of C6 H6 , and the Br− ion. The
table includes the VS,max of the σ-holes, the VS,min of the negative sites, the interaction
energies E, the angles R-X—B, R-Y—B and R-H—B and the separations X—B,
Y—B and H—B. The interaction energies were obtained with Eq. (10.3),

2
E = E(complex) − E(component i) (10.3)
i= 1
Table 10.3 confirms that the X—B, Y—B and H—B distances are less than or similar
to the sums of the respective van der Waals radii [104, 105]. The interactions are
approximately linear, i.e. along the extensions of the R–X, R–Y and R–H bonds. De-
viations from linearity generally reflect secondary interactions involving neighboring
atoms; this is illustrated in Figs. 10.6 and 10.7.
In view of the structural similarity between halogen and hydrogen bonds (both
are due to positive σ-holes on univalent atoms), it is natural to compare them. For
a given R and B, the R-H—B interaction is generally stronger than the R-X—B
(E more negative) when X = F or Cl, but they become competitive when X =
Br or I, with halogen bonding sometimes being stronger [106–108]. Compare, for
instance, the E values for F3 C-H—NH3 (−4.2 kcal/mol), F3 C-Cl—NH3 (−2.5
kcal/mol) and F3 C-Br—NH3 (−3.7 kcal/mol) in Table 10.3. Instances of halogen
bonding dominating over hydrogen bonding have long been observed experimentally
[109, 110].
Table 10.3 Computed data for various types of σ-hole complexes. Interaction energies E and
geometries were obtained by optimizations with the M06-2X/aug-cc-pVTZ procedure. σ-Hole
VS,max and negative site VS,min , prior to interaction, were calculated at the M06-2X/6-311G(d) level
Complex Negative Separationa (Å) Angle (degrees)
σ-Hole site
E VS,max VS,min
(kcal/mol)
Halogen-bonded:
F-F—NH3 − 1.5 11.2 − 46.7 F—N: 2.60 (3.05) F-F—N: 177
F3 C-Cl—NCH − 1.6 19.9 − 33.5 Cl—N: 3.13 (3.35) C-Cl—N: 179
F-F—O(CH3 )2 − 1.8 11.2 − 34.8 F—O: 2.48 (3.00) F-F—O: 169
Br2 C = − 2.1 24.1 − 33.5 Br—N: 3.09 (3.46) C-Br—N: 178
CBr(Br)—NCH
F3 C-Cl—NH3 − 2.5 19.9 − 46.7 Cl—N: 3.03 (3.35) C-Cl—N: 180
Br-C≡C-Br—NCH − 2.7 30.1 − 33.5 Br—N: 3.05 (3.46) C-Br—N: 180
Cl-Cl—SH2 − 2.9 25.4 − 20.2 Cl—S: 3.13 (3.53) Cl-Cl—S: 178
F3 C-Br—NH3 − 3.7 25.3 − 46.7 Br—N: 3.07 (3.46) C-Br—N: 180
N≡C-Cl—O(CH3 )2 − 4.0 36.0 − 34.8 Cl—O: 2.77 (3.30) C-Cl—O: 169
Br-C≡C-Br— − 4.1 30.1 − 34.8 Br—O: 2.86 (3.41) C-Br—O: 172
O(CH3 )2
Br-C≡C-Br—NH3 − 4.2 30.1 − 46.7 Br—N: 2.99 (3.46) C-Br—O: 180
F-Br—NCH − 7.1 53.2 − 33.5 Br—N: 2.60 (3.46) F-Br—N: 180
−
F3 C-Cl—Br − 8.8 19.9 − 136.4 Cl—Br: 3.19 C-Cl—Br: 179
(3.59)
F3 C-Br—Br− − 12.9 25.3 − 136.4 Br—Br: 3.09 C-Br—Br: 180
(3.70)
Group IV-VI σ -hole-bonded:
Cl2 Se—NCH − 4.0 36.1 − 33.5 Se—N: 2.92 (3.51) Cl-Se—N: 179
H2 FP—SH2 − 4.1 38.8 − 20.2 P—S: 3.25 (3.59) F-P—N: 166
H3 FSi—NCH − 4.2 34.7 − 33.5 F—Si: 2.89 (3.54) F-Si—N: 180
F2 S—NCH − 4.5 40.2 − 33.5 S—N: 2.79 (3.40) F-S—N: 174
H2 FP—NCH − 4.7 38.8 − 33.5 P—N: 2.81 (3.41) F-P—N: 164
H3 FGe—NCH − 4.9 43.0 − 33.5 Ge—N: 2.89 (–) F-Ge—N: 180
HFS—SH2 − 5.0 45.9 − 20.2 S—S: 3.08 (3.58) F-S—S: 170
H2 FAs—NCH − 5.7 44.5 − 33.5 As—N: 2.82 (3.46) F-As—N: 163
H2 FP—NH3 − 7.2 38.8 − 46.7 P—N: 2.71 (3.41) F-P—N: 167
HFS—NH3 − 8.4 45.9 − 46.7 S—N: 2.48 (3.40) F-S—N: 170
H2 FAs—NH3 − 8.7 44.5 − 46.7 As—N: 2.73 (3.46) F-As—N: 165
ClF3 Si—NH3 − 10.5 47.6 − 46.7 Si—N: 2.07 (3.71) Cl-Si—N: 180
HFSe—NH3 − 11.3 50.9 − 46.7 Se—N: 2.51 (3.51) F-Se—N: 170
Table 10.3 Continued

Complex Negative Separationa (Å) Angle (degrees)
σ-Hole site
E VS,max VS,min
(kcal/mol)
Hydrogen-bonded:
NC–H—SH2 − 2.6 51.3 − 20.2 H—S: 2.71 (2.88) C-H—S: 178
F3 C-H—NCH − 3.0 34.8 − 33.5 H—N: 2.39 (2.70) C-H—N: 178
H3 CO-H—SH2 − 3.1 44.4 − 20.2 H—S: 2.53 (2.88) O-H—S: 168
H3 CO-H—NCH − 4.1 44.4 − 33.5 H—N: 2.13 (2.70) O-H—N: 160
F3 C-H—NH3 − 4.2 34.8 − 46.7 H—N: 2.31 (2.70) C-H—N: 178
NC-H—NCH − 4.7 51.3 − 33.5 H—N: 2.21 (2.70) C-H—N: 180
H3 CO-H—C6 H6 − 4.8 44.4 − 19.4 b b
NC-H—NH3 − 6.6 51.3 − 46.7 H—N: 2.11 (2.70) C-H—N: 180
H3 CO-H—NH3 − 6.7 44.4 − 46.7 H—N: 1.96 (2.70) O-H—N: 176
F3 C-H—Br− − 14.0 34.8 − 136.4 H—Br: 2.45 (2.94) C-H—Br: 180
−
H3 CO-H—Br − 14.5 44.4 − 136.4 H—Br: 2.34 (2.94) O-H—Br: 160
−
NC-H—Br − 19.5 51.3 − 136.4 H—Br: 2.29 (2.94) C-H—Br: 180
a
Values in parentheses are the sums of the van der Waals radii of the interacting atoms, references
104 and 105. No van der Waals radius was found for germanium
b
The hydroxyl hydrogen is above the center of the benzene ring, and thus is not directly in line with
any particular atom
10.5.2 Interaction Energy Relationships
Since we consider σ-hole interactions to be electrostatic, does it follow that the

E values should correlate directly with the magnitudes of the VS,max and VS,min ?
Not necessarily, for several reasons. First, the attractive interactions are not simply
between a single point on the σ-hole (the VS,max ) and a single point on the negative
site (the VS,min ); the whole σ-hole and negative regions can participate, as was shown
by Shields et al [94]. Furthermore, interactions involving the remaining portions of
the molecules may be significant (Figs. 10.6 and 10.7) [65, 81, 111]. Second, as
pointed out in Sect. 4.1, the VS,max and VS,min are for the unperturbed molecules,
prior to interaction, and do not reflect their mutual polarizing effects; these can be
quite important, as shall be seen.
Despite these potential complications, which can be real, E has been related—
perhaps surprisingly well—to VS,max and to combinations of VS,max and VS,min . When
the negative site is kept the same, good correlations have been obtained between E
and the VS,max of the σ-holes [15, 16, 83, 111, 112]; as VS,max is more positive, E
tends to become more negative (stronger interaction). For instance, plots of E vs.
VS,max for two series of Group IV—VII complexes with NH3 and with HCN had R2 of
Fig. 10.6 Optimized geometries of Br-C≡C-Br—O(CH3 )2 (top) and H3 CO-H—NCH (bottom).

Carbons are gray, nitrogen blue, oxygens red, bromines burgundy and hydrogens white. The C-
Br—O and O-H—N angles are 172.4 and 160.4◦ , respectively. Possible reasons for the deviations
of these angles from 180◦ are the secondary interactions of methyl hydrogens with the negative
sides of the bromine in the top structure and with the nitrogen lone pair in the bottom structure.
Computational level: M06-2X/aug-cc-pVTZ
0.95 and 0.98, respectively [16]. When different negative sites are involved, then both
the VS,max and the VS,min must be taken into account. Thus when the aforementioned
complexes with NH3 and HCN were taken together and E was plotted against the
product of the σ-hole VS,max and the negative site VS,min , an R2 of 0.96 was obtained.
We have now carried out a double regression analysis encompassing all 39 com-
plexes in Table 10.3, thereby including σ-hole interactions of hydrogens, halogens
and Groups IV—VI with six different negative sites. E was expressed as,
E = c1VS,max + c2VS,min + c3 (10.4)
The relationship between the predicted and the computed E is presented in Fig. 10.8;
the R2 is 0.91, which is noteworthy considering the varied nature of the database.
(When the E in Table 10.3 are correlated with the product of VS,max and VS,min , R2
is 0.88. While using the product may appear to be more consistent with Coulomb’s
Law, this is actually not the case, since VS,max and VS,min are potentials, not charges.)
Fig. 10.7 Optimized

geometries of HFS—SH2
(top) and H2 FAs—NH3
(bottom). Fluorines are light
blue, nitrogen royal blue,
sulfurs yellow, arsenic
lavender and hydrogens
white. The F-S—S and
F-As—N angles are 169.5
and 165.0◦ , respectively.
Possible reasons for the
deviations of these angles
from 180◦ are the secondary
interactions of the HFS
hydrogen with a negative side
of the sulfur in the top
structure and the hydrogens
of H2 FAs with the nitrogen
lone pair in the bottom
structure. Computational
level: M06-2X/aug-cc-pVTZ
In a variation of these studies, Shields et al looked at E as a function of the

positive potentials that would be created by the isolated R–H and R–Br molecules
at the positions of the nitrogens in two series of complexes: (a) R-H—NCH and
R-Br—NCH, and (b) R-H—NH3 and R-Br—NH3 [94]. This was done for R-H—
N and R-Br—N angles ranging from 100 to 180◦ . For each series, E correlated
linearly and extremely well with the positive potentials at the nitrogen positions at
the different angles; R2 was 0.986 for the NH3 complexes and 0.990 for the HCN.
All of the interactions had negative E, confirming that the entire σ-hole region can
participate.
10.5.3 Thermodynamic Stability
We have discussed the energetics of σ-hole complex formation in terms of the inter-
action energies E, as is customary. From a thermodynamic standpoint, however,
stability is governed by the free energy change G; at a given absolute temperature T,
G = H − TS (10.5)
where H and S are the changes in enthalpy and entropy that accompany the
formation of the complex. For thermodynamic stability, G must be negative.
For a noncovalent interaction A + B → A—B in the gas phase at T = 298 K, H
is normally within 1 kcal/mol of E [16, 113], and if the interaction is attractive,
as in Table 10.3, then E < 0 and H < 0. S is also typically negative, because
-5
-10
-15
-20
-20 -15 -10 -5 0
Computed Δ E (kcal/mol)
Fig. 10.8 Plot of E predicted by double regression analysis, Eq. (10.4), vs. E computed with
Eq. (10.3) for the σ-hole complexes in Table 10.3. R2 = 0.91
forming the complex diminishes the degrees of freedom of A and B. This can result
in |TS| > |H|, so that G > 0 and the complex is thermodynamically unstable
despite the interaction being attractive and having E < 0 and H < 0.
It has been found that G > 0 for many σ-hole complexes in the gas phase at 298
K [16, 113, 114]; it is only the relatively strong interactions that have |H| > |TS|
and therefore G < 0. It should be kept in mind, however, that G > 0 does not
completely preclude an interaction; it simply means that the equilibrium constant is
less than one.
When the interaction takes place in solution or in the solid state, additional factors
are involved that may lead to G < 0 even when G > 0 for the gas phase com-
plex. For example, the halogen-bonded solid —Cl-C(O)-C(O)-Cl—1,4-dioxane—
was characterized crystallographically already in 1965 [115], although its gas phase
G(298 K) has been computed to be + 6.0 kcal/mol [16]. For further discussion of
thermodynamic stability in relation to σ-hole interactions, see Politzer et al [16, 113].
10.6 The Nature of σ-Hole Interactions

10.6.1 The Hellmann-Feynman Theorem
We have presented a physical interpretation of σ-hole interactions, which include

hydrogen bonding, in terms of electrostatics/polarization (the two are inextricably
linked, unless only point charges are involved). This interpretation, which many
theoreticians find unacceptably straightforward, has rigorous support: The potential
energy terms in the Hamiltonian operator are Coulombic. This leads, via the rigorous
Hellmann-Feynman theorem [116–119], to the conclusion that the forces acting
upon the nuclei in a molecule or complex can be determined purely classically as
Coulombic interactions with the electrons and with the other nuclei. (As an interesting
aside, note that the Hellmann-Feynman theorem was originally derived neither by
Hellmann nor by Feynman.)
As Levine put it, “. . . there are no ‘mysterious quantum-mechanical forces’ acting
in molecules.” [120] A knowledge of the electronic density distribution, the nuclear
positions and Coulomb’s Law suffices to determine the forces within the system—in
fact the electronic density also locates the positions of the nuclei, and indeed all of
the system’s properties, according to the Hohenberg-Kohn theorem [121].
Notwithstanding all of the above, theoreticians have happily argued for years about
the relative roles of covalency and electrostatics in hydrogen bonding [96, 122, 123].
The argument is unimpeded by the fact that covalent bonds are themselves Coulom-
bic (Hellmann-Feynman theorem) nor by the absence of any rigorous definition of
covalency. The latter point is indeed very advantageous: since covalency cannot be
measured, no one can be proved wrong and the argument can (and doubtless will)
continue indefinitely. To make matters even better, new victims have appeared—
halogen bonding and other σ-hole interactions—and can be subjected to the same
intense scrutiny!
How can the distressingly simple Coulombic explanation of noncovalent bond-
ing (i.e. electrostatics/polarization) be reconciled with the formidable array of
complexities that are typically invoked? These generally consist of some subset
of electrostatics, exchange, Pauli repulsion, polarization (or induction), charge
transfer (or donor-acceptor interaction), dispersion, orbital overlap, etc. (Different
researchers emphasize different subsets.)
We will begin by separating exchange and Pauli repulsion from the others. They
refer to mathematical requirements that must be satisfied by the wave function. Elec-
trons are indistinguishable, which is handled by introducing exchange, and the wave
function must be antisymmetric, which results in the Pauli exclusion principle (Pauli
“repulsion”). These are mathematical effects [120, 124, 125], not physical forces.
In the same context, orbitals are a valuable means of expressing wave functions but
they have no physical existence [126]—nor does, therefore, orbital overlap.
Next, it should be recognized that analyses of noncovalent bonding commonly
use the term “electrostatics” in a restricted and misleading manner, as referring
to the Coulombic interaction between the unperturbed molecules prior to forming
the complex. This is physically unrealistic [84]; it ignores the polarization of each
molecule’s electronic density distribution by the electric field of the other. This is
shown schematically for the σ-hole complexes R-X—B, R-Y—B and R-H—B in
8–10 (X = halogen, Y = Group IV—VI atom). Such polarization is an intrinsic part
of the Coulombic interaction [14–16, 85, 86, 95].
R-X---B R-Y---B R-H---B

← ← ← ← ← ←
8 9 10
The error incurred in treating the electrostatic interaction as involving only the un-
perturbed isolated molecules and ignoring their mutual polarization is graphically
illustrated by computing the difference between the electron density of the complex
and the sum of the electronic densities of the isolated molecules placed at the same
separation as in the complex. Such density difference plots for σ-hole complexes
show exactly the features depicted in 8–10 [14, 85, 127, 128]: The electric field of
the σ-hole polarizes the electronic charge of B toward the σ-hole, while the electric
field of the negative site on B polarizes the electronic charge near the σ-hole away
from B.
Another demonstration of the importance of polarization was provided by Henne-
mann et al [82]. They showed computationally that the electric field of a point charge
Q placed at a distance of 1.90 Å from one of the hydrogens of a water molecule,
Q—H-OH, caused the σ-hole potential of that hydrogen to vary linearly as a function
of the charge on Q. In the absence of the point charge, the hydrogen’s VS,max was
57 kcal/mol. As Q was made increasingly negative, it repelled electronic charge
from the hydrogen σ-hole and made it more positive; thus for Q = −0.4 au, the
σ-hole VS,max was 75 kcal/mol. As Q was made increasingly positive, on the other
hand, it attracted electronic charge to the hydrogen σ-hole and made it less positive;
for Q = 0.4 au, VS,max = 38 kcal/mol. This shows how polarization can affect the
electronic density distribution of a molecule, and in particular, that it can strengthen,
weaken or even induce a positive σ-hole. It explains why the complexes H3 C-Cl—
O = CH2 [85] and H3 P—NSH [86] were found to form despite the chlorine and
the phosphorus σ-holes in H3 C–Cl and H3 P being near-neutral or negative; positive
σ-holes were induced by the electric fields of O = CH2 and NSH (Sect. 4.1).
The polarization depicted in 8–10 and clearly evident in density difference plots
[14, 83, 127, 128] readily explains why the formation of an initial σ-hole bond may
promote a second one (“cooperativity”) or might inhibit it [11, 15, 79, 89, 129]. Thus,
it can be anticipated that 11 will have an enhanced likelihood for further interactions
through both the terminal nitrogen and the terminal hydrogen. In 12, on the other
hand, the terminal atoms should be less prone to forming additional σ-hole bonds.
NC-Br---NC-H Br-C≡C-Br---NC-CN
11 12
It was shown in Sect. 5.2 that the interaction energies of σ-hole complexes correlate
quite well with the σ-hole VS,max and the negative site VS,min , even though these are
computed for the isolated molecules prior to interaction and therefore do not reflect
their mutual polarization. An example of such a correlation is Fig. 10.8, which is
based upon Table 10.3. Some σ-hole complexes, however, have been found to have
unusually negative E and short separations [130, 131], indicating atypically strong
interactions. This simply means that polarization, particularly of the negative site, is
very significant, and the polarizabilities of the σ-hole and especially of the negative
site must be taken explicitly into account along with their VS,max and VS,min . When this
is done, via regression analyses, good correlations between predicted and computed
E are again obtained. For more extensive discussions of this, see Politzer et al
[16, 17, 131].
Proceeding to dispersion, this is a very useful concept in interpreting noncova-
lent interactions since it can always be invoked when all else fails. The relative
roles of dispersion and electrostatics are a favorite subject for debate, even though
the Hellmann-Feynman theorem tells us that dispersion is part of the Coulombic
interaction.
Dispersion is commonly linked to electronic correlation [101, 132, 133]. This
refers to the instantaneous correlated movements of electrons as they respond to
their mutual electrostatic repulsions. The usual view is that these movements create
temporary dipoles, and it is the attraction between these dipoles that accounts for the
stabilizing effect of dispersion.
Another view, widely overlooked, was proposed by Feynman [72]. He argued
that the attraction is between the nuclei of each molecule and its own electronic
charge, which has shifted to some extent into the intermolecular region. A study by
Hirschfelder and Eliason [134] and a proof by Hunt [135] supported Feynman.
Whichever explanation one prefers, dipole-dipole or nuclear-electronic attrac-
tions, the point is that both are Coulombic. Thus dispersion is fully encompassed by
the Hellmann-Feynman theorem.
10.6.2 Charge Transfer or Polarization?
Charge transfer from an “electron donor” to an “electron acceptor” has been widely
invoked as an important factor in hydrogen and halogen bonding, and in σ-hole
interactions in general. Some small fraction of an electron is supposedly transferred
from the negative site B (the donor) into an antibonding orbital of the molecule R–X,
R–Y or R–H that has the σ-hole (the acceptor). This weakens the R–X, R–Y or R–H
bond and accounts for its stretching frequency often (but not always) being lower in
the complex than in the isolated molecule (a red shift).
While this scenario can sometimes be effective as a purely mathematical model, it
should not be confused with physical reality. Electrons are indivisible; a small portion
of one cannot be plucked away. Orbitals—bonding, antibonding or otherwise—do
not really exist [126], however useful and convenient the concept may be. It is in-
creasingly being recognized [136–140] that charge transfer theory is a mathematical
attempt to represent a physical process, which is the mutual polarization of the inter-
acting molecules, the so-called “donor” and “acceptor.” Thus it is redundant to cite
both charge transfer and polarization as separate factors in noncovalent interactions
[83, 84].
Hermansson [141] and Qian and Krimm [142] have derived formalisms that ex-
plain and predict both the red and the blue shifts in R–X, R–Y and R–H stretching
frequencies in terms of just the electric field of B and the permanent and induced
dipole moments of the R–X, R–Y and R–H molecules. These procedures have been
extensively applied to σ-hole interactions [143, 144].
A simple demonstration of how polarization can produce either a red or a blue shift
was provided by Hennemann et al [84]. They put a point charge Q near the hydroxyl
hydrogen of methanol, Q—H-O-CH3 , and computed the O–H stretching frequency
as a function of the charge on Q. Starting with Q = 0, as Q was made increasingly
negative, the O–H frequency steadily decreased; when Q was made increasingly
positive, the frequency increased until Q = 0.3 and then began to decrease. Thus both
red and blue shifts could be produced by varying the charge and hence the electric
field of Q, thereby polarizing the methanol molecule. The lower O–H frequencies
cannot be due to the transfer of electronic charge into an antibonding orbital of the
methanol because Q has no electronic charge to transfer. Both the red shifts and the
blue shifts are purely polarization effects.
The mathematical rather than physical nature of the charge transfer concept is
illustrated by an example due to Stone and Misquitta [145]. The usual quantum
chemical description of a noncovalent complex, for example R-Y—B, is in terms of
basis orbitals on both R–Y and B. However it could also be done quite satisfactorily
using orbitals on only R–Y or only B, if enough of them were used. The polarization
shown in 9 would be adequately described, by the computed electronic density
distribution of the complex. However the charge transfer from an orbital of B into
an orbital of R–Y would necessarily be zero, since either B or R–Y has no orbitals.
The physical reality is maintained (by the electronic density distribution, which is
an observable), but the mathematical model (charge transfer) fails.
To summarize, “charge transfer” in noncovalent interactions is, physically, po-
larization and polarization is Coulombic. The Hellmann-Feynman theorem lives
on!
10.7 The Fallacy and Inadequacy of Atomic Charges
The preceding discussion is clearly linked to the question: Is it meaningful to assign

quantitatively a charge to an atom in a molecule? In the laboratory, experience indi-
cates the usefulness of viewing some atoms as having positive or negative characters
relative to others. It is tempting to try to quantify this (a temptation to which one
of the current authors yielded in his misspent youth [146]). However the concept
of an atom in a molecule, while very useful in practice, does not have a rigorous
basis—and therefore neither does giving it a numerical charge [53]. Since there is no
correct way to do this, everyone is free to invent his/her own scheme, and many have
done so. By 1994, more than 30 different procedures for assigning atomic charges
had been proposed [147]. They sometimes produce remarkably varied results; for
instance, Wiberg and Rablen cited calculated charges for the carbon in H3 C–NO2
that ranged from—0.478 to + 0.564 [148].
Furthermore, attributing a single positive or negative charge to an atom in a
molecule ignores the well-known anisotropies of its charge distribution and its
electrostatic potential (Sect. 4.1). Price has referred to this rather bluntly as “a
travesty of bonding theory” [149]. As was discussed in earlier sections, it is these
anisotropies that account for the observed abilities of many covalently-bonded atoms
to interact favorably with both positive and negative sites and for the phenomenon
of “like attracting like.” Calculated atomic charges cannot explain such behavior
[45, 84, 150].
Auffinger et al [45] pointed out that since many force fields used in molecular
mechanics and molecular dynamics do use atomic charges, they may miss some non-
covalent interactions; an example was presented by Dobeš et al [151]. Accordingly
several research groups have sought to develop more realistic force fields [152–155].
10.8 Concluding Comments
The avalanche of studies that has descended upon the area of noncovalent interactions
in recent years has subjected them to minute dissection and extensive compartmental-
ization. For instance, the literature now mentions at least a dozen types of hydrogen
bonding!
Our emphasis, however, has been upon the unifying principle that a great many
noncovalent interactions—involving atoms of Groups IV—VII as well as hydrogen—
fit under the umbrella of σ-hole bonding (illustrated on the front cover of Physical
Chemistry Chemical Physics, volume 15, issue 27, 2013): A region of positive
electrostatic potential, on the extension of a single covalent bond to the atom, interacts
attractively with a negative site. Closely related to this is so-called “π-hole” bonding,
in which the positive potential is perpendicular to an atom or region in a planar portion
of a molecular framework. This is examined in detail elsewhere [16, 156, 157].
We have also invoked the rigorous Hellmann-Feynman theorem (which seems to
often be overlooked) to point out that these noncovalent interactions can be fully
understood and described as being Coulombic, which includes polarization and dis-
persion. If this seems simplistic to some, we argue that it is because of a tendency to
confuse mathematical modeling with physical reality. The latter can be annoyingly
straightforward. Newton said that, “Nature is pleased with simplicity” [158]. Einstein
seemed to agree: “Nature is the realization of the simplest conceivable mathematical
ideas” [158].
References
1. Müller-Dethlefs K, Hobza P (2000) Noncovalent interactions: a challenge for experiment and

theory. Chem Rev 100:143–167
2. Engkvist O, Åstrand P-O, Karlström G (2000) Accurate intermolecular potentials obtained
from molecular wave functions: bridging the gap between quantum chemistry and molecular
simulations. Chem Rev 100:4087–4108
3. Chałasiński G, Szczês̀niak (2000) State of the art and challenges of the ab initio theory of
intermolecular interactions. Chem Rev 100:4227–4252
4. Braga D, Grepioni F (2000) Intermolecular interactions in nonorganic crystal engineering.
Acc Chem Res 33:601–608
5. Metrangolo P, Neukirch H, Pilati T, Resnati G (2005) Halogen bonding based recognition
processes: a world parallel to hydrogen bonding. Acc Chem Res 38:386–395
6. Belkova NV, Shubina ES, Epstein LM (2005) Diverse world of unconventional hydrogen
7. Hobza P, Zahradník R, Müller-Dethlefs K (2006) The world of non-covalent interactions:
2006. Collect Czech Chem Commun 71:443–531
8. Metrangolo P, Resnati G, eds (2008) Halogen bonding: fundamentals and applications.
Springer, Berlin
9. Stone AJ (2008) Intermolecular potentials. Science 321:787–789
10. Legon AC (2010) The halogen bond: an interim perspective. Phys Chem Chem Phys 12:7736–
7747
11. Politzer P, Murray JS, Clark T (2010) Halogen bonding: an electrostatically-driven highly
directional noncovalent interaction. Phys Chem Chem Phys 12:7748–7757
12. Hobza P, Müller-Dethlefs K (2010) Non-Covalent Interactions: Theory and Experiment. Royal
13. Scheiner S (2011) Weak H-Bonds. comparisons of CH–O to NH–O in proteins and PH–N to
direct P–N interactions. Phys Chem Chem Phys 13:13860–13872
14. Politzer P, Riley KE, Bulat FA, Murray JS (2012) Perspectives on halogen bonding and other
σ-hole interactions: lex parsimoniae (Occam’s Razor). Comput Theoret Chem 998:2–8
15. Politzer P, Murray JS (2013) Halogen bonding: an interim discussion. Chem Phys Chem
14:278–294
16. Politzer P, Murray JS, Clark T (2013) Halogen bonding and other σ-hole interactions: a
17. Politzer P, Murray JS, Clark T (2014) σ-Hole Bonding: A Physical Interpretation. Topics Curr
Chem (in press).
18. Schneider H-J (2009) Binding mechanisms in supramolecular complexes. Angew Chem Int
Ed 48:3924–3977
19. Colin M (1814) Ann Chim 91:252
20. Guthrie F (1863) XXVIII. On the Iodide of Iodammonium. J Chem Soc 16:239–244
21. Mallet JW (1881) Chem News 44:188
22. Remsen I, Norris JF (1896) The action of the halogens on the methylamines. Am Chem J
18:90–95
23. Rhoussopoulos O (1883) Einwirkung von Chinolin auf Chloroform and Jodoform. Berichte
16:202–203
24. Blackstock SC, Lorand JP, Kochi JK (1987) Charge-transfer interactions of amines with
tetrahalomethanes. X-ray crystal structures of the donor-acceptor complexes of quinuclidine
and diazabicyclo[2.2.2]octane with carbon tetrabromide. J Org Chem 52:1451–1460
25. Benesi HA, Hildebrand JH (1948) Ultraviolet absorption bands of iodine in aromatic
hydrocarbons. J Am Chem Soc 70:2832–2833
26. Benesi HA, Hildebrand JH (1949) A Spectrophotometric investigation of the interaction of
iodine with Aromatic Hydrocarbons. J Am Chem Soc 71:2703–2707
27. Mulliken RS (1952) Molecular compounds and their spectra. II. J Am Chem Soc 74:811–824
28. Dahl T, Hassel O (1966) A close relationship between the crystal structure of an acceptor and
that of an addition compound. Acta Chem Scand 20:2009
29. Bent HA (1968) A structural chemistry of donor-acceptor interactions. Chem Rev 68:587–648
30. Olie K, Mijlhoff FC (1969) The crystal structure of POBr3 and intermolecular bonding. Acta
Cryst B 25:974–977
31. Olie K (1971) The crystal structure of POCl3 . Acta Cryst B 27:1459–1460
32. Murray-Rust P, Motherwell WDS (1979) Computer retrieval and analysis of molecular
geometry. 4. intermolecular interactions. J Am Chem Soc 101:4374–4376
33. Murray-Rust P, Stallings WC, Monti CT, Preston RK, Glusker JP (1983) Intermolecular
interactions of the carbon-fluorine bond: the crystallographic environment of fluorinated
carboxylic acids and related structures. J Am Chem Soc 105:3206–3214
34. Ramasubbu N, Parthasarathy R, Murray-Rust P (1986) Angular preferences of intermolec-
ular forces around halogen centers: preferred directions of approach of electrophiles and
nucleophiles around carbon-halogen bond. J Am Chem Soc 108:4308–4314
35. Dumas J-M, Kern M, Janier-Dubry JL (1976) Cryoscopic and calorimetric study of mx4 -polar
organic base interactions (M = C, Si, X = Cl, Br)—influence of element and of halogen. Bull
Soc Chim Fr 11–1:1785–1790
36. Martire DE, Sheridan JP, King JW, O’Donnell SE (1976) Thermodynamics of molecular
association. 9. An NMR study of hydrogen bonding of chloroform and bromoform to Di-n-
octyl ether, Di-n-octyl thioether, and Di-n-Octylmethylamine. J Am Chem Soc 98:3101–3106
37. Imakubo T, Sawa H, Kato R (1995) Novel radical cation salts of organic π-donors containing
iodine atom(s): the first application of strong intermolecular I–X (X = CN, halogen atom)
interaction to molecular conductors. Synth Metals 73:117–122
38. Imakubo T, Tajima N, Tamura M, Kato R, Nishio Y, Kajita K (2003) Supramolecular
organic conductor θ-(DIETS)2 [Au(CN)4 ]: unique crystal structure and superconductivity un-
der Uniaxial Strain (DIETS = Diodo(ethylenedithio) diselenadithiafulvalene). Synth Metals
133–134:181–183
39. Amico V, Meille SV, Corradi E, Messina MT, Resnati G (1998) Perfluorocarbon-hydrocarbon
self-assembly. 1D infinite chain formation driven by nitrogen—iodine interactions. J Am
Chem Soc 120:8261–8262
40. Rissanen K (2008) Halogen bonded supramolecular complexes and networks. Cryst Eng
Comm 10:1107–1113
41. Metrangolo P, Meyer F, Pilati T, Resnati G, Terraneo G (2008) Halogen bonding in
supramolecular chemistry. Angew Chem Int Ed 47:6114–6127
42. Priimagi A, Cavallo G, Metrangolo P, Resnati G (2013) The halogen bond in the design of
functional supramolecular materials: recent advances. Acc Chem Res 46:2686–2695
43. Bruckmann A, Pena MA, Bolm C (2008) Organocatalysis through halogen-bond activation.
Synlett 6:900–902
44. Kniep F, Jungbauer SH, Zhang Q, Walter SM, Schindler S, Schnapperelle I, Herdtweck E,
Huber SM (2013) Organocatalysis by neutral multidentate halogen-bond donors. Angew
Chemie Int Ed 52:7028–7032 (and references cited)
45. Auffinger P, Hays FA, Westhof E, Shing Ho P (2004) Halogen bonds in biological molecules.
Proc Natl Acad Sci 48:16789–16794
46. Lu Y, Wang Y, Zhu W (2010) Nonbonding interactions of organic Halogens in biological

systems: implications for drug discovery and Biomolecular design. Phys Chem Chem Phys
12:4543–4551
47. Murray JS, Riley KE, Politzer P, Clark T (2010) Directional weak intermolecular interactions:
σ-Hole bonding. Aust J Chem 63:1598–1607.
48. Rosenfield RE Jr, Parthasarathy R, Dunitz JD (1977) Directional preferences of nonbonded
atomic contacts with divalent sulfur. 1. Electrophiles and nucleophiles. J Am Chem Soc
99:4860–4862
49. Guru Row TN Parthasarthy R (1981) Directional preferences of nonbonded atomic contacts
with divalent sulfur in terms of its orbital orientations. 2. S–S interactions and nonspherical
shape of sulfur in crystals. J Am Chem Soc 103:477–479
50. Ramasubbu N, Parthasarathy R (1987) Stereochemistry of incipient electrophilic and nu-
cleophilic reactions at divalent selenium center: electrophilic-nucleophilic pairing and
anisotropic shape of Se in Se–Se interactions. Phosphorus Sulfur 31:221–229
51. Politzer P, Murray JS (1998) Statistical analysis of the molecular surface electrostatic poten-
tial: an approach to describing noncovalent interactions in condensed phases. J Mol Struct
(Theochem) 425:107–114
52. Politzer P, Murray JS (2002) The fundamental nature and role of the electrostatic potential in
atoms and molecules. Theor Chem Acc 108:134–142
53. Murray JS, Politzer P (2011) The electrostatic potential: an overview. WIREs Comp Mol Sci
1:153–163
54. Stewart RF (1979) On the mapping of electrostatic properties from bragg diffraction data.
55. Politzer P, Truhlar DG, eds (1981) Chemical applications of atomic and molecular electrostatic
potentials. Plenum Press, New York
56. Klein CL, Stevens ED (1988) Charge density studies of drug molecules. In Liebman JF,
Goldberg A (eds) Structure and reactivity. VCH Publishers, New York, ch 2, pp 25–64
57. Bader RWF, Carroll MT, Cheeseman JR, Chang C (1987) Properties of atoms in molecules:
atomic volumes. J Am Chem Soc 109:7968–7979
58. Brinck T, Murray JS, Politzer P (1992) Surface electrostatic potentials of halogenated
methanes as indicators of directional intermolecular interactions. Int J Quantum Chem
44(Suppl 19):57–64
59. Brinck T, Murray JS, Politzer P (1993) Molecular electrostatic potentials and local ionization
energies of group V—VII hydrides and their anions. Relationships for aqueous and gas-phase
acidities. Int J Quantum Chem 48(Suppl 20):73–88
60. Murray JS, Paulsen K, Politzer P (1994) Molecular surface electrostatic potentials in the
analysis of non-hydrogen-bonding noncovalent interactions. Proc Indian Acad Sci (Chem
Sci) 106:267–275
61. Awwadi FF, Willett RD, Peterson KA, Twamley B (2006) The nature of halogen-halogen
synthons: crystallographic and theoretical studies. Chem Eur J 12:8952–8960
62. Clark T, Hennemann M, Murray JS, Politzer P (2007) Halogen bonding: the σ-hole. J Mol
Model 13:291–296
63. Politzer P, Lane P, Concha MC, Ma Y, Murray JS (2007) An overview of halogen bonding. J
Mol Model 13:305–311
64. Burling FT, Goldstein BM (1992) Computational studies of nonbonded sulfur-oxygen and
selenium-oxygen interactions in the thiazole and selenazole nucleosides. J Am Chem Soc
114:2313–2320
65. Murray JS, Lane P, Politzer P (2008) Simultaneous σ-hole and hydrogen bonding by sulfur-
and selenium-containing heterocycles. Int J Quantum Chem 108:2770–2781
66. Clark T, Murray JS, Lane P, Politzer P (2008) Why are dimethyl sulfoxide and dimethyl
sulfone such good solvents? J Mol Model 14:689–697
67. Murray JS, Lane P, Politzer (2009) Expansion of the σ-hole concept. J Mol Model 15:723–729
interaction. Int J Quantum Chem 107:2286–2292
69. Politzer P, Murray JS, Janjić GV, Zarić SD (2014) σ-hole interactions of covalently-bonded
nitrogen, phosphorus and arsenic: a survey of crystal structures. Crystals 4:12–31
70. Delgado-Barrio G, Prat RF (1975) Deformed hartree-fock solutions for atoms. III. convergent
iterative process and results for O−− . Phys Rev A 12:2288–2297.
71. Sen KD, Politzer P (1989) Characteristic features of the electrostatic potentials of singly-
negative monoatomic ions. J Chem Phys 90:4370–4372
72. Feynman RP (1939) Forces in Molecules. Phys Rev 56:340–343
73. Pauling L (1960) The nature of the chemical bond, 3rd edn. Cornell University Press, Ithaca
74. Stevens ED (1979) Experimental electron density distributions of molecular chlorine. Mol
Phys 37:27–45
75. Nyburg SC, Faerman CH (1985) A revision of van der Waals atomic Radii for molecular
crystals: N, O, F, S, Cl, Se, Br and I bonded to carbon. Acta Cryst B 41:274–279
76. Ikuta S (1990) Anisotropy of electron density distribution around atoms in molecules: N, P,
O and S atoms. J Mol Struct (Theochem) 205:191–201
77. TsirelsonVG, Zou PF, Tang T-H, Bader RFW (1995) Topological Definition of crystal structure
determination of the bonded interactions in solid molecular chlorine. Acta Cryst A 51:143–153
78. Torii H (2003) The role of atomic quadrupoles in intermolecular electrostatic interactions of
polar and nonpolar molecules. J Chem Phys 119:2192–2198
79. Bilewicz E, Rybarczyk-Pirek AJ, Dubis AT, Grabowski SJ (2007) Halogen bonding in crystal
structures of 1-methylpyrrol-2-yl trichloromethyl ketone. J Mol Struct 829:208–211
80. Murray JS, Macaveiu L, Politzer P (2014) Factors affecting the strengths of σ-hole electrostatic
potentials. J Comput Sci 5:590–596
81. Bundhun A, Ramasami P, Murray JS, Politzer P (2013) Trends in σ-hole strengths and
interactions of F3 MX molecules (M = C, Si, Ge and X = F, Cl, Br, I). J Mol Model
19:2739–2746
82. Hennemann M, Murray JS, Politzer P, Riley KE, Clark T (2012) Polarization-induced σ-holes
and hydrogen bonding. J Mol Model 18:2461–2469
83. Clark T, Murray JS, Politzer P (2014) Role of polarization in halogen bonds. Aust J Chem
67:451–456
84. Clark T, Murray JS, Politzer P (2015) Correct electrostatic treatment of non-covalent
interactions: the importance of polarisation. WIREs Comp Mol Sci, DOI: 10.1002/wcms.1210
85. Riley KE, Hobza P (2008) Investigations into the nature of halogen bonding including
symmetry adapted perturbation theory analyses. J Chem Theory Comp 4:232–242
86. Solimannejad M, Gharabaghi M, Scheiner S (2011) SH–N and SH–P blue-shifting H-bonds
and N–P interactions in complexes pairing HSN with amines and phosphines. J Chem Phys
134:24312(1–6)
87. Politzer P, Murray JS, Concha MC (2007) Halogen bonding and the design of new materials:
organic bromides, chlorides and even fluorides as donors. J Mol Model 13:643–650
88. Chopra D, Guru Row TN (2011) Role of organic fluorine in crystal engineering. Cryst Eng
Comm 13:2175–2186
89. Metrangolo P, Murray JS, Pilati T, Politzer P, Resnati G, Terraneo G (2011) Fluorine-centered
halogen bonding: a factor in recognition phenomena and reactivity. Cryst Growth Des
11:4238–4246
90. Murray JS, Lane P, Clark T, Politzer P (2007) σ-hole bonding: molecules containing group
VI atoms. J Mol Model 13:1033–1038
91. Wilson CC (2000) Single crystal neutron diffraction from molecular materials.World
Scientific, Singapore
92. Parkin A, Harte SM, Goeta AE, Wilson CC (2004) Imaging proton migration from X-rays
and neutrons. New J Chem 28:718–721
93. Murray JS, Politzer P (1991). Correlations between the solvent hydrogen-bond-donating
parameter α and the calculated molecular surface electrostatic potential. J Org Chem 56:6715–
6717
94. Shields ZP-I, Murray JS, Politzer (2010) Directional tendencies of halogen and hydrogen
bonding. Int J Quantum Chem 110:2823–2832
95. Clark T (2013) σ-holes. WIREs Comput Mol Sci 3:13–20

96. Kollman P, McKelvey J, Johansson A, Rothenberg S (1975) Theoretical studies of hydrogen-
bonded dimers. Complexes involving HF, H2 O, NH3 , CH1 , H2 S, PH3 , HCN, HNC, HCP,
CH2 NH, H2 CS, H2 CO, CH4 , CF3 H, C2 H2 , C2 H4 , C6 H6 , F− and H3 O+ . J Am Chem Soc
97:955–965
97. Politzer P, Daiker KC (1981) Models for chemical reactivity. In: Deb BM (ed) The Force
Concept in Chemistry. Van Nostrand Reinhold Co, New York, ch 6.
98. Legon AC, Millen DJ (1982) Determination of properties of hydrogen-bonded dimers by
rotational spectroscopy and a classification of dimer geometries. Faraday Discuss Chem Soc
73:71–87
99. Buckingham AD, Fowler PW (1983) Do electrostatic interactions predict structures of Van
der Waals complexes? J Chem Phys 79:6426–6428
100. Buckingham AD, Fowler PW (1985) A Model for the geometries of Van der Waals complexes.
Can J Chem 63:2018–2025
101. Riley KE, Murray JS, Fanfrlík J, Řezáč J, Solá RJ, Concha MC, Ramos FM, Politzer P (2013)
Halogen bond tunability II: the varying roles of electrostatic and dispersion contributions to
attraction in halogen bonds. J Mol Model 19:4651–4659
102. Legon AC (1999) Prereactive complexes of dihalogens XY with Lewis bases B in the gas
phase: a systematic case for the halogen analogue B–XY of the hydrogen bond B–HX. Angew
Chem Int Ed 38:2686–2714
103. Legon AC (2010) The halogen bond: an interim perspective. Phys Chem Chem Phys 12:7736–
7747
104. Bondi A (1964) van der Waals volumes and Radii. J Phys Chem 68:441–451
105. Rowland RS, Taylor R (1996) Intermolecular nonbonded contact distances in organic crystal
structures: comparison with distances expected from van der Waals Radii. J Phys Chem
100:7384–7391
106. Politzer P, Murray JS, Lane P (2007) σ-hole bonding and hydrogen bonding: competitive
interactions. Int J Quantum Chem 107:3046–3052.
107. Aakeröy CB, Fasulo M, Shultheiss N, Desper J, Moore C (2007) Structural competition
between hydrogen bonds and halogen bonds. J Am Chem Soc 129:13772–13773
108. Alkorta I, Blanco F, Solimannejad M, Elguero J (2008) Competition of hydrogen bonds and
halogen bonds in complexes of hypohalous acids with nitrogenated bases. J Phys Chem A
112:10856–10863
109. Di Paolo T Sandorfy C (1974) On the biological importance of the hydrogen bond breaking
potency of fluorocarbons. Chem Phys Lett 26:466–469
110. Corradi E, Meille SV, Messina MT, Metrangolo P, Resnati G (2000) Halogen bonding versus
hydrogen bonding in driving self-assembly processes. Angew Chem Int Ed 39:1782–1786
111. Riley KE, Murray JS, Fanfrlík J, Řezáˇc J, Solá RJ, Concha MC, Ramos FM, Politzer P (2011)
Halogen bond tunability I: the effects of aromatic fluorine substitution on the strengths of
halogen-bonding interactions involving chlorine, bromine and iodine. J Mol Model 17:3309–
3318
112. Riley KE, Murray JS, Politzer P, Concha MC, Hobza P (2009) Br–O complexes as probes
of factors affecting halogen bonding: interactions of bromobenzenes and bromopyrimidines
with acetone. J Chem Theory Comput 5:155–16
113. Politzer P, Murray JS (2013) Enthalpy and entropy factors in gas phase halogen bonding:
compensation and competition. Cryst Eng Comm 15:3145–3150
114. Lu X, Li H, Zhu X, Zhu W, Liu H (2011) How does halogen bonding behave in solution? A
Theoretical study using implicit solvation model. J Phys Chem A 115:4467–4475
115. Damm E, Hassel O, Romming C (1965) X-ray analysis of the (1:1) addition compounds of
1,4-Dioxan with oxalyl chloride resp. Oxalyl bromide. Acta Chem Scand 19:1159–1165
116. Güttinger P (1932) Das Verhalten von Atomen in magnetischen Drehfeld. Z Phys 73:169–184
117. Pauli W (1933) Principles of Wave Mechanics, Handbuch der Physik, 24. Springer, Berlin,
p. 162
118. Hellmann H (1933) Zur Rolle der kinetischen Electronenenergie für die zwischenatomaren
Kräfte. Z Phys 85:180–190
119. Feynman RP (1939) Forces in molecules. Phys Rev 56:340–343
120. Levine IN (2000) Quantum chemistry, 5th ed. Prentice-Hall, Upper Saddle River
121. Hohenberg P, Kohn W (1964) Inhomogeneous electron gas. Phys Rev 136:B864–B871
122. Ratajczak H, Orville-Thomas WJ (1976) Charge transfer theory of hydrogen bonds: relations
between vibrational, spectra and energy of hydrogen bonds. Chem Phys 17:197–216 (and
references cited)
Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ (2011) Defining
the hydrogen bond: an account. Pure Appl Chem 83:1619–1636 (and references cited)
124. Berlin T (1951) Binding regions in diatomic molecules. J Chem Phys 19:208–213
125. Bader RFW (2006) Pauli repulsions exist only in the eye of the beholder. Chem Eur J 12:2896–
2901
126. Scerri ER (2000) Have orbitals really been observed? J Chem Ed 77:1492–1494
127. Solimannejad M, Malekani M, Alkorta I (2010) Cooperative and dimunitive unusual weak
bonding in F3 C–X–HMgH–Y and F3 C–X–Y–HMgH Trimers (X = Cl, Br; Y = HCN and
HNC). J Phys Chem A 114:12106–12111
128. Scheiner S (2011) On the Properties of X–N Noncovalent Interactions for First-, Second-,
and Third-Row X Atoms. J Chem Phys 134:164313(1–9)
129. Grabowski SJ, Bilewicz E (2006) Cooperative halogen bonding effect—Ab Initio calculations
on H2 CO–(ClF)n complexes. Chem Phys Lett 427:51–55
130. Del Bene JE Alkorta I Elguero J (2010) Do traditional, chlorine-shared and ion-pair halogen
bonds exist? An Ab Initio investigation of FCl:CNX complexes. J Phys Chem A 114:12958–
12962
131. Politzer P, Murray JS (2012) Halogen bonding and beyond: factors influencing the nature of
CN–R and SiN–R complexes with FCl and Cl2 . Theor Chem Acc 131:1114(1–10)
132. Hobza P, Zahradnik R (1992) An essay on the theory and calculations of intermolecular
interactions. Int J Quantum Chem 42:581–590
133. Cramer CJ (2002) Essentials of computational chemistry. Wiley, Chichester
134. Hirschfelder JO, Eliason MA (1967) Electrostatic Hellmann-Feynman theorem applied to the
long-range interaction of two hydrogen atoms. J Chem Phys 47:1164–1169
135. Hunt KLC (1990) Dispersion dipoles and dispersion forces: proof of Feynman’s “Conjecture”
and generalization to interacting molecules of arbitrary symmetry. J Chem Phys 92:1180–1187
136. Stone AJ, Price SL (1988) Some new ideas in the theory of intermolecular forces: anisotropic
atom-atom potentials. J Phys Chem 92:3325–3335
138. Sokalski WA, Roszak SM (1991) Efficient techniques for the decomposition of intermolecular
interaction energy at SCF level and beyond. J Mol Struct (Theochem) 234:387–400
139. Chen J, Martínez TJ (2007) QTPIE: charge transfer with polarization current equalization: a
fluctuating charge model with correct asymptotics. Chem Phys Lett 438:315–320
140. Clark T (2014) Directional electrostatic bonding. In: Frenking G (ed) The chemical bond:
chemical bonding across the periodic table. Wiley-VCH, Weinheim, Germany, ch. 18.
141. Hermansson K (2002) Blue-shifting hydrogen bonds. J Phys Chem A 106:4695–4702
142. Qian W, Krimm S (2002) Vibrational spectroscopy of hydrogen bonding. Origin of the
different behavior of the C–H–O hydrogen bond. J Phys Chem A 106:6628–6636
143. Wang W, Wang NB, Zheng W, Tian A (2004) Theoretical study on the blueshifting halogen
bond. J Phys Chem A 108:1799–1805
144. Murray JS, Concha MC, Lane P, Hobza P, Politzer P (2008) Blue shifts vs. red shifts in σ-hole
bonding. J Mol Model 14:699–704
145. Stone AJ, Misquitta AJ (2009) Charge-transfer in symmetry-adapted perturbation theory.
146. Politzer P, Harris RR (1970) Properties of atoms in molecules. I. A proposed definition of the
charge of an atom in a molecule. J Am Chem Soc 92:6451–6454
147. Meister J, Schwarz WHE (1994) Principal components of ionicity. J Phys Chem 98:8245–8252
148. Wiberg KB, Rablen PR (1993) Comparison of atomic charges derived via different procedures.
J Comput Chem 14:1504–1518
149. Price SL (1996) Applications of realistic electrostatic modelling to molecules in complexes,
solids and proteins. J Chem Soc Faraday Trans 92:2997–3008
150. Politzer P, Murray JS, Concha MC (2008) σ-hole bonding between like atoms. A fallacy of
atomic charges. J Mol Model 14:659–665
151. Dobeš P, Řezáˇc J, Fanfrlík J, Otyepka M, Hobza P (2011) Semiempirical quantum
mechanical method PM6-DH2X describes the geometry and energetics of CK2-inhibitor
complexes involving halogen bonds well, while the empirical potential fails. J Phys Chem B
115:8581–8589
152. Ibrahim MAA (2011) Molecular mechanical study of halogen bonding in drug discovery. J
Comput Chem 32:2564–2574
153. Kolař M, Hobza P (2012) On extension of the current biomolecular empirical force field for
the description of halogen bonds. J Chem Theory Comput 8:1325–1333
154. Carter M, Rappé AK, Shing Ho P (2012) Scalable anisotropic shape and electrostatic models
for biological bromine halogen bonds. J Chem Theory Comput 8:2461–2473
155. Jorgensen WL, Schyman P (2012) Treatment of halogen bonding in the opls-aa force field:
application to potent anti-hiv agents. J Chem Theory Comput 8:3895–3901
156. Murray JS, Lane P, Clark T, Riley KE, Politzer P (2012) σ-holes, π-holes and electrostatically-
driven interactions. J Mol Model 18:541–548
157. Murray JS, Shields ZP-I, Seybold PG, Politzer P (2015) Intuitive and counterintuitive nonco-
valent interactions of aromatic π regions with the hydrogen and nitrogen of HCN. J Comput
Sci DOI: 10.1006/j.jocs.02.001
158. Isaacson W (2007) Einstein: his life and universe. Simon and Schuster, New York, p. 549
Chapter 11
The X-C· · ·Y Carbon Bond
Devendra Mani and E. Arunan
Abstract The CH3 X (X more electronegative group/atom than C) molecules have

their C atom partially positive which enables it to bond with another atom/group
of atoms having an electron rich region, Y. This leads to a penta-coordinate carbon
having X-C· · ·Y ‘carbon bond’ in addition to the four covalent bonds around the
central carbon. The X-C· · ·Y carbon bond is very similar to the X-H· · ·Y hydrogen
bond. The X-C bond lengthens and there is a red-shift in X-C stretching frequency.
The X-C· · ·Y is linear. There is an overlap between X-C σ* orbital and the donor
orbital in Y, which leads to the weakening X-C bond and red-shift. Analysis of
the electron density topology shows a bond critical point between C and Y. It also
shows that there is mutual penetration of electron clouds of the C and Y atoms,
leading to bond formation. The X-C· · ·Y carbon bond is an intermediate in SN 2
reactions, very much like the X-H· · ·Y hydrogen bond being an intermediate in proton
transfer reactions. Moreover, the carbon bonded H2 O· · ·CH3 OH geometry is a text-
book example for the ill-defined hydrophobic interactions, as the H2 O approaches
methanol from the hydrophobic CH3 side. Similar interactions involving Si and Ge
containing molecules exist and these have all been named ‘tetrel bonds’, extending
halogen bonding (Group 17), chalcogen bonding (Group 16) and pnicogen bonding
(Group 15) to Group 14. While such interactions involving elements from all groups
are likely, we expect that the carbon bonding would be next only to hydrogen bonding
in its importance, as C and H are the elements chosen by nature and life as we know
is largely based on these elements. This chapter summarizes the results on carbon
bonding since it was proposed about a year ago.
11.1 Introduction
Non-covalent interactions are important binding forces utilized by nature to produce

the condensed phases of neutral molecules [1, 2]. This term has been made popular
E. Arunan () · D. Mani

Department of Inorganic and Physical Chemistry,
Indian Institute of Science, 560012
Bangalore, India
e-mail: arunan@ipc.iisc.ernet.in

324 E. Arunan and D. Mani
to denote intermolecular interactions, which were initially called van der Waals
interactions following his well known equation. Without such interactions, there is
no earth or life and hence the enormous interest in this field is not surprising. In the
early part of twentieth century, stunning differences in intermolecular interactions
between some molecules, such as H2 O, and other molecules such as H2 S, warranted a
further classification of intermolecular interactions. Water is a high boiling liquid and
H2 S is a gas at ambient conditions. In solid H2 O (ice), each molecule is surrounded by
four molecules and in solid H2 S, each molecule is surrounded by twelve molecules.
Only a sphere could accommodate 12 neighbours and it appeared that H2 S molecules
were essentially spherical having isotropic interactions, implying that the H atoms
in H2 S had no specific orientations. In ice, on the other hand each H2 O molecule has
specific interactions with its four neighbours, with the O atom pointing towards two
H atoms of two of the neighbouring molecules and each H atom pointing towards
an O atom of a neighbouring molecule all with an O. . . H distance of ≈ 1.8 Å. This
may be compared to the O–H covalent bond distance of ≈ 1.0 Å within the H2 O
molecule. The longer intermolecular contact was called ‘hydrogen bonding’ and it
makes the intermolecular interaction in ice very directional and anisotropic. It was
decided that the anisotropic intermolecular interactions as in H2 O would be called
‘hydrogen bonding’ and isotropic intermolecular interactions as in H2 S would be left
as van der Waals interaction. Pauling’s classic book on The Nature of Chemical Bond
was perhaps the first book to popularize hydrogen bonds [3] and it concluded that
H2 O can form hydrogen bonds and H2 S cannot. Since then there have been numerous
attempts to differentiate hydrogen bonding and van der Waals interactions in terms
of the physical forces involved and it only led to a lot of discussion on semantics.
Now we know that both H2 O and H2 S can form directional hydrogen bonds but at
different conditions. Intermolecular forces in both these molecules originate from
electrostatic interactions, induction/polarization, dispersion, exchange correlation
and charge-transfer/covalency.
The recent IUPAC recommendation on the definition of hydrogen bonding states
clearly that no unique physical forces can be attributed to hydrogen bonding and the
forces acting between the interacting molecules should have contributions beyond
dispersion [4]. Hydrogen bonding, often denoted as X-H· · ·Y, is clearly a subset of
van der Waals interaction. Its unique features may be summarized as follows: (i)
it is directional and the angle X-H· · ·Y tends to be linear, (ii) H is bonded to two
(or more) atoms and is ‘hypervalent’ as it has a normal valency of 1 and (iii) X is
more electronegative than H giving the H a partial positive charge. The uniqueness
of hydrogen among all the elements of the periodic table made it easy to identify ‘hy-
drogen bonding’ as a well-defined intermolecular interaction. The fact that it controls
the properties of the molecules of life such as H2 O and DNA ensured that enormous
attention was paid to it resulting in almost one paper on it being published every hour
in the last several years [4]. Its importance in biological sciences has been specifically
addressed in some recent books [5, 6]. However, it was only a matter of time before
similar interactions involving other elements in the periodic table were found.
Many specific non-covalent interactions, similar to hydrogen bonding, have now
been recognised and classified. The obvious choices were interactions involving
11 The X-C· · ·Y Carbon Bond 325
atoms in the alkali (Group 1) and halogen (Group 17) groups, as H has similarities
to both these group elements. As shown in a recent comprehensive study, lithium
bonding is dominated by electrostatics as LiX molecules are distinctly ionic as op-
posed to HX and XY molecules (X, Y halogens), both of which have significant
covalent nature [7]. Indeed, halogen bonding has been characterized as a distinct in-
termolecular interaction similar to hydrogen bonding now [8]. In the last few years,
chalcogen bonding (Group 16) [9] and pnicogen bonding (Group 15) [10] have also
been characterized. Clearly all these interactions could be classified as van der Waals
interactions. However, these new names are most certainly reflective of the detailed
understanding of these interactions as observed in these specific groups.
Highly directional halogen bonding interactions have found great applications
in supramolecular chemistry, crystal engineering and molecular recognition process
[11]. The bonding in these interactions could be explained using the sigma-hole con-
cept which was first introduced by Politzer and coworkers [12, 13] to explain halogen
bonding. This was useful as there was some confusion about why the electronegative
halogen atoms were interacting with other electronegative atoms. Even though, the
presence of a sigma hole in the molecules containing Group 14 elements had also
been predicted [13], work on such interactions involving carbon atom was particu-
larly lagging. There had been quite a few studies for the complexes containing more
polarisable Si and Ge atoms [13, 14] and in one of the studies complexes containing
carbon atom were also considered [15]. This interaction involving C is much weaker
compared to the interactions involving Si and Ge atoms, perhaps explaining the
initial lack of interest. Another difficulty in identifying this interaction involving C
atoms is that, in most molecules a tetravalent carbon is protected in all directions and
penta-coordinate carbons have not been easy to identify. The fact that this interaction
with C containing molecules is relatively weaker is important in nature, vide infra.
Our interest in carbon bonding similar to hydrogen bonding, arose following a
microwave spectroscopic investigation on Ar· · ·propargyl alcohol complex, which
may not attract widespread interest [16, 17]. Propargyl alcohol could exist in gauche
and trans conformations, gauche being more stable than trans. We attempted to ob-
serve the rotational spectra of the Ar complexes with both these forms, but could
succeed in observing only the complex formed by the gauche conformer. Experimen-
tal microwave spectrum showed that the equilibrium structure had Ar positioned in
between the OH and C≡C groups exhibiting both O–H· · ·Ar and Ar· · ·π interactions.
Atoms in molecules (AIM) theoretical analysis on the ab initio optimized structure
showed bond critical points confirming both these interactions. These resembled
CH3 OH· · ·Ar and HC≡CH· · ·Ar complexes, as might be expected. While the ex-
periment on Ar· · ·trans-propargyl alcohol complex did not succeed, theory posed no
such challenge. The trans conformer formed a complex with Ar exhibiting O–H· · ·Ar
interaction very similar to what is observed in the CH3 OH· · ·Ar complex. The AIM
calculations also revealed evidence for O–H· · ·Ar interaction but showed another
bond critical point connecting Ar to C! This was unexpected and given the contra-
dicting claims of the relation between ‘bonds’ and ‘bond critical points’ [18, 19], it
may have been prudent to ignore this result. However, the AIM analysis was repeated
for CH3 OH· · ·Ar complex, which also exhibited both OH· · ·Ar and Ar· · ·C bonds.
We had previously shown that the tetrahedron faces of CH4 are electron rich, based on
its electrostatic potential, and it could accept a hydrogen bond as in H4 C· · ·HX, lead-
ing to a penta-coordinate carbon [19]. How would the tetrahedron face of CH3 OH,
opposite to the OH group, look? When we looked at it, see next section, carbon bond
was born.
We showed that a specific interaction exists between the carbon atom of a CH3 X
molecule, where X is an electron withdrawing atom or group, and electron rich
centre (atom) of the other binding unit; like oxygen of H2 O [17]. Such interactions
were recognised as unique X-C· · ·Y interactions and were called carbon bond after
confirming their various bonding characteristics. Interactions with other atoms of
the group IV have also been explored in detail recently and were found to be similar
to the carbon bonding [20–23]. The interactions involving group 14 elements were
collectively called tetrel bond by Grabowski and Bauza et al., almost simultaneously
and independently [20, 24]. The very fact that nature prefers carbon over other tetrel
atoms convinces us that carbon bonding would remain the most important subset of
tetrel bonding.
In addition to ‘hydrogen bonding’ and van der Waals interactions, intermolecular
interactions have often been classified as electrostatic interactions and hydrophobic
interactions [2]. While the term ‘electrostatic interactions’ is easy to understand as
the interactions between charges, dipoles, quadrupoles etc., the term hydrophobic
interactions is not as well defined. A careful look at carbon bonding interactions
revealed that they are a specific part of the ill-defined hydrophobic interaction [17].
Carbon bonding is quite young and is just a little older than 1 year. However, in
this short span of time, there have been many studies on carbon bonds and much
more is known about these now [20–23, 25–31]. Importance of carbon bonding in
SN 2 reactions was pointed out in the original paper [17] and was later explored in
detail by Grabowski [20]. Frontera et al. [27, 28] have shown that small substituted
cycloalkane molecules are potential carbon bond donor. Alkorta and co-workers
have pointed out the presence of intramolecular carbon bonding in carbohydrate
molecules [26]. Varadwaj et al. [25] have studied complexes of CO with a number
of CH3 X molecules. It was observed that both the ends of CO molecule can interact
with the carbon atom of the CH3 X molecule leading to the formation of O· · ·C-X
and C· · ·C-X carbon bonds, the latter can be called a dicarbon bond [25], in analogy
to the ‘dihydrogen bonds’ [4b]. Moreover, carbon bonds have also been observed
experimentally using charge density analysis [32].
In this chapter, we have tried to cover most of the important results on carbon
bonding till date. The chapter begins with an introduction to the carbon bond interac-
tions and then explores it in detail by using various theoretical tools. We summarize
the results from Reference 17 and extend it for complexes with CH3 Cl and CH3 Br
as carbon bond donors and also, π electrons, rare gases and sigma electrons in H2
as carbon bond acceptors. We then summarize results from other groups on carbon
bonds with unpaired electrons and anions as acceptors. Experimental results on car-
bon bonds are discussed next. This includes the experiments that were carried out
after the carbon bond was proposed and also the earlier experiments that had given
structures which could now be classified as ‘carbon bonded complexes’.
Fig. 11.1 Molecular

electrostatic potential for
methane molecule. Relative
color code bar is shown on
right
11.1.1 What is a Carbon Bond?
Carbon bond can be understood in terms of sigma hole concept which has already
been used to explain other interactions like halogen bonding [8, 11, 12, 33]. Analysis
of the molecular electrostatic potential (MESP) surfaces helps in locating regions
of electronic charge depletion and accumulation in molecules [34, 35]. Regions
with charge depletion can receive electronic charge from other electron rich species
and may lead to bonding interactions. Similarly, electron rich regions interact with
nucleophiles and may lead to bonding interactions. A sigma hole is one such electron
depleted region along the direction opposite to a polar sigma bond.
We start our discussion here with the methane molecule. Its MESP is shown in
Fig. 11.1. Red regions are electronegative and green regions are electropositive which
represent regions of electron accumulation and electron depletion, respectively. The
tetrahedral face of methane is electronegative and thus is nucleophilic in nature. The
nucleophilic tetrahedral face interacts with molecules like HF, HCl, HCN, H2 O etc.,
in such a way that the hydrogen atom of these molecules points towards the tetra-
hedral face of methane. Such structures have been experimentally observed for the
complexes of CH4 with HF/HCl/HCN/H2 O using microwave spectroscopy [36–39].
Atoms in molecules (AIM) theoretical analysis showed a bond critical point between
carbon atom of methane and the hydrogen atom of HX/H2 O molecule confirming
the presence of a X-C· · ·H hydrogen bond through the tetrahedral face [19].
What happens to the bonding nature of methane when one of its hydrogen atoms
is replaced by some atom or group which is more electronegative than carbon? Is the
CH3 face of a molecule like CH3 OH hydrogen bond acceptor like that of CH4 ? Let
us have a look at the MESP of the CH3 OH molecule shown in Fig. 11.2.
In this molecule, the CH3 tetrahedral face is electropositive in contrast to the
electronegative tetrahedral face of methane. Being electropositive, it interacts with
the electron rich sites like a lone-pair of electrons and leads to the stable geometries
like the one shown for H2 O· · ·CH3 OH complex Fig. 11.3.
Fig. 11.2 Molecular

electrostatic potential
surfaces for CH3 OH
molecule, viewed from the
CH3 side. Relative color code
bar is also shown on right
The AIM analysis confirms the presence of a bond critical point between the
electron rich atom and the carbon atom of CH3 X molecules as shown in Fig. 11.3b.
This bond critical point follows several criteria used to classify hydrogen bonding
and halogen bonding [17, 40] which confirms the formation of an X-C· · ·Y carbon
bond. Here H2 O is the carbon bond acceptor unit and methanol is the donor unit. Such
interactions exist for a series of CH3 X complexes which are discussed in the next
sections. It is worth pointing out again that both in H4 C· · ·H2 O and in H2 O· · ·CH3 OH,
we have a penta-coordinate carbon. In the former, CH4 is the electron donor, bonding
with the electron deficient H in H2 O and it is an O–H· · ·C hydrogen bonded complex.
In the later, H2 O is the electron donor, bonding with the electron deficient C in
CH3 OH and it is an O–C· · ·O carbon bonded complex. This convention in naming
intermolecular bonding is needed today for two reasons. (1) The ‘hydrogen bond’
terminology is well established in science for nearly a century. (2) Such specific
interactions involving other atoms have been widely recognized now. Without this
convention, text book examples for hydrogen bonds, such as HF dimer might be
classified as ‘halogen bonded’ and H2 O dimer might be classified as ‘chalcogen
bonded’.
Fig. 11.3 a Optimized geometry of H2 O· · ·CH3 OH complex, b Electron density topology for the
complex; small green dots represent bond critical points
11.2 Computational Details
The geometry optimization was performed at MP2/6-311+G(3df,2p) as well as

MP2/Aug-cc-pVTZ level of theories. Frequency calculations were performed on
the optimized geometries to confirm that these geometries are minima on potential
energy hypersurface as well as to find out the zero-point correction to the energies.
Optimization as well as frequency calculations were performed using Gaussian 09
software suite [41]. Binding energies were corrected using the counterpoise method
[42] inbuilt in Gaussian 09. The binding energies were also corrected for the zero-
point energy. The MESP surfaces were plotted using GaussView software [43].
Electrostatic potential values were extracted using Multiwfn software [44]. The
AIM analysis was performed on the optimized geometries using AIM-ALL [45]
and AIM2000 [46] software suites. The NBO analysis was performed using NBO
6.0 software [47].
11.3 Lone Pairs as Carbon Bond Acceptor
11.3.1 Complexes with Methanol
As mentioned earlier, presence of the electron withdrawing hydroxyl group in

methanol leads to electron depletion at its CH3 face and facilitates the formation
of a carbon bond. Hydrogen bonds were first observed with lone pairs of electrons
as acceptors and not surprisingly the same is true for carbon bonds. This section
focuses on lone pair acceptors.
Complexes of methanol with various lone pair donor molecules; H2 O, H2 S, HX
(X = F/Cl/Br), LiX (X = F/Cl/Br), ClF, NH3 and PH3 were considered to explore
these interactions in detail [17]. In each of these molecules, central atom bears
one or more lone pair of electrons which can take part in intermolecular bonding.
Geometry optimization was performed starting with guess geometries in which the
atoms with lone-pair(s) of the carbon bond acceptor molecule faces the CH3 face of
methanol molecule. The optimized geometries for all these complexes are shown in
Fig. 11.4. All these geometries were confirmed to be minima on the potential energy
hypersurface by confirming the presence of real frequencies for all the normal modes
of vibrations. Important geometrical parameters for the optimized geometries are
given in Table 11.1.
Intermolecular bond lengths in all the complexes are shorter at MP2/Aug-cc-
pVTZ level than at MP2/6-311+G(3df,2p) level. However, at both the levels of
theory, the bond lengths for all the complexes appear long. The values are compa-
rable to the sum of van der Waals radii of Y and C atoms in the case of complexes
with H2 O, HF and ClF; are larger in case of complexes with H2 S, HCl, HBr, NH3
and PH3 ; and are much shorter in case of complexes with LiF, LiCl and LiBr. How-
ever, as pointed out in the IUPAC recommendation on hydrogen bonding [4], use
of van der Waals ‘radii’ to confirm/rule out specific bonding interactions could be
Fig. 11.4 Optimized geometries for a H2 O· · ·CH3 OH, b H2 S· · ·CH3 OH, c HF· · ·CH3 OH, d
HCl· · ·CH3 OH, e HBr· · ·CH3 OH, f LiF· · ·CH3 OH, g LiCl· · ·CH3 OH, h LiBr· · ·CH3 OH, i
ClF· · ·CH3 OH, j H3 N· · ·CH3 OH, k H3 P· · ·CH3 OH complexes at MP2/6-311+G(3df,2p) level of
theory. (Reproduced with permission from RSC, Reference 17)
Table 11.1 Intermolecular bonding parameters for various complexes of CH3 OH at MP2/6-
311+G(3df,2P) and MP2/Aug-cc-pVTZ levels of theory. Distances are in angstroms and angles in
degrees
Complex MP2/6-311+G(3df,2p) MP2/Aug-cc-pVTZ
d (Y· · ·C) Angle (Y· · ·C-X) d (Y· · ·C) Angle (Y· · ·C-X)
H2 O· · ·CH3 OH 3.167 176.5 3.167 176.5
H2 S· · ·CH3 OH 3.713 177.6 3.692 177.6
HF· · ·CH3 OH 3.131 180.0 3.091 177.8
HCl· · ·CH3 OH 3.568 177.4 3.561 177.1
HBr· · ·CH3 OH 3.770 176.1 3.610 175.7
ClF· · ·CH3 OH 3.113 176.3 3.115 175.4
LiF· · ·CH3 OH 2.985 179.5 2.983 179.5
LiCl· · ·CH3 OH 3.434 179.2 3.434 179.2
LiBr· · ·CH3 OH 3.657 179.8 3.522 178.0
H3 N· · ·CH3 OH 3.377 175.7 3.384 174.9
H3 P· · ·CH3 OH 3.803 174.4 3.602 166.3
misleading. For all the complexes the Y· · ·C-X angle is greater than 175◦ except for
H3 P· · ·CH3 OH complex. The atoms in molecules theoretical analysis, vide infra,
does indicate that there is penetration of electron cloud between the two molecules
Table 11.2 Binding energies for the methanol complexes calculated at different levels of theory.
The BSSE corrected, E(BSSE) , and BSSE as well as zero-point corrected, E(BSSE + ZPC) values are
given. All the values are given in kJ.mol−1 . (Reproduced with permission from RSC, Reference 17)
Complex At MP2 At MP2 At CCSD-T
/6-311+G(3df,2p) /Aug-cc-pVTZ /6-311+G(3df,2p)
E(BSSE) E(BSSE + ZPC) E(BSSE) E(BSSE + ZPC) E(BSSE)
H2 O· · ·CH3 OH 4.2 1.9 4.5 2.2 4.4
H2 S· · ·CH3 OH 2.5 1.0 3.5 2.3 2.7
HF· · ·CH3 OH 2.9 1.2 3.1 1.7 3.1
HCl· · ·CH3 OH 2.4 1.1 3.0 1.8 2.2
HBr· · ·CH3 OH 2.5 1.4 2.9 1.5 2.6
ClF· · ·CH3 OH 2.4 1.1 2.9 1.5 2.3
LiF· · ·CH3 OH 9.0 7.6 9.7 8.4 9.7
LiCl· · ·CH3 OH 7.2 6.0 8.1 7.0 7.0
LiBr· · ·CH3 OH 6.9 6.0 7.5 6.6 6.7
H3 N· · ·CH3 OH 4.3 2.2 4.7 2.7 4.5
H3 P· · ·CH3 OH 2.8 1.6 3.3 1.7 2.6
confirming the ‘carbon bond’ formation. Binding energies at both the levels of theory
are given in Table 11.2. Values obtained from single point counterpoise calculations
at CCSD(T)/6-311+G(3df,2p) level are also given.
Intermolecular bond distance pattern is reflected in the binding energies also
for these complexes. Complexes are slightly more stable at the MP2/Aug-cc-
pVTZ level as compared to the MP2/6-311+G(3df,2p) level. This is due to the
shorter bond distances observed at the MP2/Aug-cc-pVTZ level. Values calculated
at CCSD(T)/6-311+G(3df,2p) level are not very different from those obtained at
MP2/6-311+G(3df,2p) level. Comparing the binding energies shows that the com-
plexes with HY are much weaker than those with LiY (Y = F/Cl/Br). The LiY
molecules have larger dipole moment than HY molecules and it can explain the
difference in bond energies.
The Atoms in molecule theoretical analysis shows presence of a bond critical point
between theY atom and C atom in all the complexes. For the H2 O· · ·CH3 OH complex
the electron density topology is shown in Fig. 11.2. Intermolecular bond paths were
found to be present in each of the complexes, which connect the intermolecular bond
critical point to the Y and C atoms, suggesting bonding between the two. Electron
density and Laplacian of electron density values at the intermolecular bond critical
points are given in Table 11.3. The table also lists the values of mutual penetration
between Y and C atoms. It is defined as the difference between the sum of the non-
bonded ‘radii’ of the two atoms and the sum of their bonded ‘radii’. Non-bonded
‘radii’ are calculated as the distance from the atom to the point where the electron
density reduces to 0.001, along the bond path. Mutual penetration value for all the
complexes is positive which confirms the bonding between the two atoms. We recall
Table 11.3 Electron density (ρ) and Laplacian of electron density (∇ 2 ρ) values at the intermolecular
bond critical point for various complexes of methanol. Mutual penetration between the carbon and
Y atoms are also given. (Reproduced with permission from RSC, Reference 17)
Complex ρ(r) in a.u. ∇ 2 ρ in a.u. Mutual penetration (Å)
H2 O· · ·CH3 OH 0.0050 0.0248 0.41
H2 S· · ·CH3 OH 0.0037 0.0164 0.30
HF· · ·CH3 OH 0.0037 0.0230 0.28
HCl· · ·CH3 OH 0.0038 0.0174 0.34
HBr· · ·CH3 OH 0.0034 0.0140 0.31
ClF· · ·CH3 OH 0.0050 0.0299 0.39
LiF· · ·CH3 OH 0.0063 0.0350 0.54
LiCl· · ·CH3 OH 0.0049 0.0228 0.48
LiBr· · ·CH3 OH 0.0042 0.0181 0.42
H3 N· · ·CH3 OH 0.0044 0.0201 0.43
H3 P· · ·CH3 OH 0.0035 0.0143 0.34
that the distance between the two atoms is more than the sum of van der Waals radii
in some cases, even though there is mutual penetration. This is because the van der
Waals ‘radii’ assumes the atoms to be spherical, when they are not. The carbon bonds
in these complexes follow most of the criteria proposed by Koch and Popelier [48]
for C-H· · ·O hydrogen bonds and these will be discussed in the next section.
11.3.2 Complexes with Halomethanes (CH3 F, CH3 Cl and CH3 Br)
In this section, we consider the complexes between CH3 X (X = F/Cl/Br) molecules

with the same 11 carbon bond acceptor molecules considered in the preceding section.
The halogen atoms are more electron withdrawing than the OH group and therefore
the sigma hole on the CH3 face of halomethane molecules is more dominant in
comparison to that of methanol. The ESP surface for CH3 F molecule is shown in
Fig. 11.5. The sigma hole at the face centre is clearly visible in this case. The ESP
values corresponding to maxima at the CH3 face centres are + 97.8, 82.3 and 78.6
kJ.mol−1 for CH3 F, CH3 Cl and CH3 Br respectively. These values are in accordance
to the electronegativity order for F, Cl and Br.
Complexes of halomethanes were optimized following the methods as employed
for the methanol complexes. Similar geometries, suggesting interaction between
the atom with lone pair(s) of electrons and CH3 face of halomethane molecules,
were obtained on optimization. Binding energies for these complexes, obtained from
different levels of theories are collected in Table 11.4. The values show that these
complexes are stronger compared to the corresponding methanol complexes. Among
Fig. 11.5 Electrostatic

potential surface for CH3 F
molecule mapped on the
0.001 a.u. electron density
surface
the halomethanes, the complexes with CH3 F are the strongest. The general trend of
binding energy for a given acceptor is CH3 F > CH3 Cl > CH3 Br. For a particular
halomethane, the complexes with LiY (Y = F/Cl/Br) molecules are the strongest
ones as was the case with methanol complexes.
Table 11.4 Binding energies for the halomethane complexes calculated at different levels of theory.
The BSSE corrected, E(BSSE) , and BSSE as well as zero-point corrected, E(BSSE + ZPC) values
are given. All the values are given in kJ.mol−1
Complex MP2 /6-311+G(3df,2p) MP2 /Aug-cc-pVTZ CCSD-T /6-
311+G(3df,2p)
Complexes with CH 3 F
H2 O· · ·CH3 F 8.6 6.0 7.5 4.7 7.3
H2 S· · ·CH3 F 3.7 1.6 4.8 2.8 3.8
HF· · ·CH3 F 4.7 2.7 5.1 3.0 5.0
HCl· · ·CH3 F 5.0 3.5 3.7 2.2 2.8
HBr· · ·CH3 F 4.1 2.8 3.6 2.0 3.2
ClF· · ·CH3 F 2.6 1.0 3.7 2.2 3.1
LiF· · ·CH3 F 15.9 14.0 16.8 15.0 16.7
LiCl· · ·CH3 F 11.6 10.1 12.6 11.3 11.3
LiBr· · ·CH3 F 10.4 9.3 11.3 10.2 10.0
H3 N· · ·CH3 F 7.6 4.6 8.2 5.2 7.9
H3 P· · ·CH3 F 4.1 2.5 4.8 3.1 3.8

Complex MP2 /6-311+G(3df,2p) MP2 /Aug-cc-pVTZ CCSD-T /6-
311+G(3df,2p)
Complexes with CH 3 Cl
H2 O· · ·CH3 Cl 6.7 4.2 7.1 4.8 6.8
H2 S· · ·CH3 Cl 3.6 1.9 4.7 3.1 3.6
HF· · ·CH3 Cl 4.6 2.6 4.9 3.2 4.8
HCl· · ·CH3 Cl 3.1 1.8 3.7 2.4 2.7
HBr· · ·CH3 Cl 3.2 2.1 3.7 2.4 3.1
ClF· · ·CH3 Cl 3.2 1.9 3.8 2.5 3.0
LiF· · ·CH3 Cl 15.8 14.2 16.6 15.2 16.5
LiCl· · ·CH3 Cl 12.1 10.9 13.1 12.1 11.5
LiBr· · ·CH3 Cl 11.1 10.1 11.9 11.0 10.4
H3 N· · ·CH3 Cl 7.1 4.6 7.5 5.1 7.3
H3 P· · ·CH3 Cl 4.1 2.7 4.7 3.5 3.7
Complexes with CH 3 Br
H2 O· · ·CH3 Br 6.4 4.2 6.8 4.6 6.4
H2 S· · ·CH3 Br 3.5 1.5 4.6 3.1 3.3
HF· · ·CH3 Br 4.4 2.9 4.8 3.0 4.5
HCl· · ·CH3 Br 3.1 2.0 3.7 2.7 2.5
HBr· · ·CH3 Br 3.2 2.3 3.7 2.5 2.1
ClF· · ·CH3 Br 3.2 2.1 3.8 2.6 2.7
LiF· · ·CH3 Br 15.4 14.0 16.2 14.7 16.0
LiCl· · ·CH3 Br 12.0 10.9 13.0 12.2 11.3
LiBr· · ·CH3 Br 11.1 10.2 11.9 11.1 9.7
H3 N· · ·CH3 Br 6.8 4.4 7.2 5.0 6.8
H3 P· · ·CH3 Br 4.0 2.8 4.7 3.6 3.4
The AIM theoretical analysis was performed on each of these complexes as well.
The analysis reveals the presence of an intermolecular bond critical point between
the Y atom and C atom in all the complexes. Electron density (ρ), Laplacian of
electron density (∇ 2 ρ) values, corresponding to intermolecular bond critical point
connecting Y and C atoms, for all the complexes are given in Table 11.5. The table
also lists values of mutual penetration between the Y and C atoms for each of the
molecules. The positive mutual penetration values for all of these complexes confirm
the intermolecular bonding.
Table 11.5 Electron density (ρ) and Laplacian of electron density (∇ 2 ρ) values at the intermolecular
bond critical point for various complexes of methanol. Mutual penetration between the carbon and
Y atoms are also given
Complex ρ(r) (a.u.) ∇ ( ρ ) (a.u.) Mutual penetration (Å)
Complexes with CH 3 F
H2 O· · ·CH3 F 0.0056 0.0307 0.50
H2 S· · ·CH3 F 0.0042 0.0184 0.45
HF· · ·CH3 F 0.0048 0.0298 0.38
HCl· · ·CH3 F 0.0041 0.0196 0.39
HBr· · ·CH3 F 0.0036 0.0164 0.32
ClF· · ·CH3 F 0.0051 0.0305 0.41
LiF· · ·CH3 F 0.0081 0.0459 0.67
LiCl· · ·H3 F 0.0058 0.0287 0.58
LiBr· · ·CH3 F 0.0050 0.0227 0.49
H3 N· · ·H3 F 0.0056 0.0262 0.55
H3 P· · ·CH3 F 0.0040 0.0163 0.42
Complexes with CH 3 Cl
H2 O· · ·CH3 Cl 0.0055 0.0282 0.52
H2 S· · ·CH3 Cl 0.0042 0.0172 0.46
HF· · ·CH3 Cl 0.0044 0.0260 0.37
HCl· · ·CH3 Cl 0.0041 0.0186 0.54
HBr· · ·CH3 Cl 0.0038 0.0157 0.36
ClF· · ·CH3 Cl 0.0049 0.0281 0.40
LiF· · ·CH3 Cl 0.0081 0.0442 0.69
LiCl· · ·CH3 Cl 0.0061 0.0287 0.46
LiBr· · ·CH3 Cl 0.0052 0.0227 0.53
H3 N· · ·CH3 Cl 0.0053 0.0232 0.53
H3 P· · ·CH3 Cl 0.0039 0.0150 0.40
Complexes with CH 3 Br
H2 O· · ·CH3 Br 0.0055 0.0264 0.51
H2 S· · ·CH3 Br 0.0042 0.0167 0.46
HF· · ·CH3 Br 0.0044 0.0252 0.39
HCl· · ·CH3 Br 0.0042 0.0185 0.43
HBr· · ·CH3 Br 0.0039 0.0153 0.37
ClF· · ·CH3 Br 0.0057 0.0321 0.41
LiF· · ·CH3 Br 0.0082 0.0435 0.70

Complex ρ(r) (a.u.) ∇ ( ρ ) (a.u.) Mutual penetration (Å)
LiCl· · ·CH3 Br 0.0062 0.0284 0.64
LiBr· · ·H3 Br 0.0054 0.0227 0.57
H3 N· · ·H3 Br 0.0052 0.0222 0.55
H3 P· · ·CH3 Br 0.0041 0.0148 0.43
11.4 Characteristics of Carbon bond
11.4.1 The AIM Descriptors for Carbon Bond
In the previous section, some of the most commonly used AIM descriptors for a
bond, such as electron density at the BCP, its Laplacian and mutual penetration
were presented along with the geometries and binding energies of the carbon bonded
complexes with lone pair acceptors. In this section, we explore all the criteria based on
AIM theoretical analysis given by Koch and Popelier [48] for the C-H· · ·O hydrogen
bonds and discuss their application to carbon bonds. Complexes of CH3 OH and
CH3 F were considered for this purpose and the results are summarized below:
1. The first criterion says that a bond critical point must be present between the two
interacting atoms and also bond paths should connect the intermolecular bond
critical point to both the interacting atoms. It is clear from the afore mentioned
analysis that the CH3 OH as well as CH3 F complexes follow this criterion
2. The second criterion defines the range of acceptable values of electron density
(ρ) at the bond critical point between the H and O atoms in case of C-H· · ·O
hydrogen bonds. It suggests that the electron density value at the bond critical
point should be within the range of 0.002–0.034 a.u. Note that small deviations
are expected from this range due to the change in basis sets and level of theory
used for the calculations. Values for the bond critical point corresponding to the
carbon bond in CH3 OH as well as CH3 F complexes show that this criterion is
followed by these complexes, Tables 11.3 and 11.5.
3. The third criterion defines the range for Laplacian of electron density (∇ 2 ρ) at
the bond critical point between H and O atoms for C-H· · ·O hydrogen bonds.
The acceptable range was given to be 0.024–0.139 a.u. for such hydrogen bonds.
Note that again small deviations from this range are expected depending on the
basis sets used. The Laplacian values for the carbon bond complexes of CH3 OH
and CH3 F do lie within this range, see 11.3 and 11.5.
4. The fourth criterion says that mutual penetration between the two interacting
atoms should be positive. This criterion is considered to be the necessary and
sufficient criterion for intermolecular bonding. As mentioned before, the mutual
penetration between the Y atom and C atom is positive for both the CH3 OH as
well as CH3 F complexes. Thus, this most important criterion is followed by these
complexes.
5. The criteria five to eight deal with the change in property of the hydrogen atom tak-
ing part in hydrogen bonding. The fifth criterion states that for hydrogen bonding,
there should be a loss of charge on the hydrogen atom upon complex formation.
This is calculated as the difference between the population on the hydrogen atom
in monomer, NH (monomer) and the population on the atom in the complex, NH
(complex) i.e. according to this criterion
NH = NH (complex) − NH (monomer) should be negative
The corresponding NC values for the C atom were extracted from the AIM cal-
culations and are complied in Table 11.6 for CH3 OH and CH3 F complexes. It is
clear from the table that this criterion is followed by all the complexes except the
H2 O· · ·CH3 OH complex for which the NC value is nearly zero.
6. The sixth criterion deals with the energy of the hydrogen atom in the C-H· · ·O
hydrogen bonds and states that the hydrogen atom should be destabilized on
complex formation i.e.
EH = EH (complex) − EH (monomer) should be positive.
EH (complex) and EH (monomer) are the energies of the hydrogen atom in the
hydrogen bonded complex and in monomer respectively. The E values for the C
atom (EC ) extracted from the AIM calculations are given in Table 11.6 for the
CH3 OH and the CH3 F complexes. This criterion is followed by all the complexes
except the H2 O· · ·CH3 OH and the LiCl· · ·CH3 OH complexes for which E value
is negative.
7. The seventh criterion states that the magnitude of dipolar polarization (M) of
hydrogen should decrease on complex formation i.e.
|MH | = |MH (complex)| − |MH (monomer)| should be negative
MH (complex) and MH (monomer) are the dipolar polarization values of the hydrogen
atom in the hydrogen bonded complex and in monomer respectively. The |MH | is the
difference between the two. The |MC | values for the C atom are given in Table 11.6
for all the CH3 OH and CH3 F complexes. It is clear from the table that many of the
CH3 OH complexes (with HF, HCl, HBr, ClF) do not follow this criterion and for
them the |M| values are positive. However, most of the CH3 F complexes follow
this criterion except the ClF· · ·CH3 F complex.
8. According to the eighth and final criterion, atomic volume of the H atom should
decrease on complex formation i.e.
VH = VH (complex) − VH (monomer) should be negative
Changes in the atomic volume of the carbon atom (VC ) for all the complexes are
given in Table 11.6. It can be noticed that all the complexes follow this criterion.
338
Table 11.6 Change in different properties of the carbon atom upon complex formation. Change in population, NC , energy EC , dipolar polarization
|MC |, and volume VC is given for the different complexes of CH3 OH and CH3 F. These properties were calculated using AIM theory. All the
properties are given in atomic units. (Reproduced with permission from RSC, Reference 17)
CH3 OH complexes NC EC |MC | VC CH3 F complexes NC EC |MC | VC
H2 O· · ·CH3 OH 0.002 − 0.0270 − 0.030 − 3.44 H2 O· · ·CH3 F − 0.163 0.1906 − 0.025 − 8.92
H2 S· · ·CH3 OH − 0.204 0.2323 − 0.011 − 8.79 H2 S· · ·CH3 F − 0.163 0.2309 − 0.029 − 7.96
HF· · ·CH3 OH − 0.204 0.2323 0.006 − 9.09 HF· · ·CH3 F − 0.164 0.2026 − 0.006 − 8.62
HCl· · ·CH3 OH − 0.204 0.2046 0.004 − 8.22 HCl· · ·CH3 F − 0.164 0.2366 − 0.009 − 8.04
HBr· · ·CH3 OH − 0.205 0.2370 0.000 − 8.56 HBr· · ·CH3 F − 0.163 0.2279 − 0.013 − 7.59
ClF· · ·CH3 OH − 0.205 0.2249 0.018 − 9.89 ClF· · ·CH3 F − 0.168 0.2303 0.007 − 9.37
LiF· · ·CH3 OH − 0.193 0.1919 − 0.046 − 9.74 LiF· · ·CH3 F − 0.168 0.1866 − 0.036 − 9.41
LiCl· · ·CH3 OH − 0.190 − 0.4515 − 0.040 − 9.00 LiCl· · ·CH3 F − 0.151 0.2156 − 0.050 − 8.35
LiBr· · ·CH3 OH − 0.193 0.2259 − 0.038 − 9.16 LiBr· · ·CH3 F − 0.148 0.2184 − 0.054 − 8.11
H3 N· · ·CH3 OH − 0.203 0.1937 − 0.020 − 9.24 H3 N· · ·CH3 F − 0.162 0.1848 − 0.043 − 8.64
H3 P· · ·CH3 OH − 0.202 0.2325 − 0.004 − 7.57 H3 P· · ·CH3 F − 0.162 0.2315 − 0.033 − 7.56
E. Arunan and D. Mani
10 18
9 16
8 14
R² = 0.960
7 R² = 0.998
12
ΔE (kJ.mol-1)
ΔE (kJ.mol-1)
6
10
5
8
4 R² = 0.778
6 R² = 0.809
3
4
2
2
1
0 0
0.003 0.005 0.007 0.003 0.004 0.005 0.006 0.007 0.008 0.009
ρ(r) (electron .bohr-3) ρ(r) (electron .bohr-3)
CH3OH Complexes CH3F Complexes

a b
18 18
16 16
14 14
R² = 0.989 R² = 0.993
12 12
ΔE (kJ.mol-1)
10 10
ΔE (kJ.mol-1)
8 8
R² = 0.897
6 6
4 R² = 0.897 4
2 2
0 0
0.003 0.005 0.007 0.009 0.003 0.005 0.007 0.009
ρ(r) (electron .bohr-3) ρ(r) (electron .bohr-3)
CH3Cl Complexes CH3Br Complexes

c d
Fig. 11.6 Plot of stabilization energy, E, against electron density, ρ, at the carbon bond critical
point for a CH3 OH complexes, b CH3 F complexes, c CH3 Cl complexes and d CH3 Br complexes
11.4.2 Stabilization Energy vs. Electron Density
In the case of hydrogen bonding, there is a linear correlation between stabilization

energy due to the hydrogen bonding and electron density at the bond critical point
corresponding to the hydrogen bond. Are these two quantities correlated for the
carbon bonds as well?
Plots between the stabilization energy and electron density at the intermolecular
bond critical point are shown in Fig. 11.6 for the carbon bonded (a) CH3 OH, (b)
CH3 F, (c) CH3 Cl and (d) CH3 Br complexes. In each of the correlation plots, two
different fits are shown. The upper fits show the correlation for complexes with LiY
(Y = F/Cl/Br) molecules. The lower fit shows the correlation for all other complexes
except those with ClF molecule. In all the four cases, the stabilization energy is almost
linearly correlated with the electron density for complexes with LiY molecules.
The correlation for complexes with the other molecules is also quite good and the
correlation coefficient is 0.778, 0.809, 0.897 and 0.897 for CH3 OH CH3 F, CH3 Cl
and CH3 Br complexes, respectively. In none of the cases, complexes with ClF were
included in the fit. The fits were worsening in doing so.
It can be seen that the stabilization energy does correlate with the electron density
at the bond critical point in case of carbon bonding as well. However, the cor-
relation depends on the nature of carbon bond acceptor molecules. We note that
Koch and Popelier in their original paper on C-H· · ·O hydrogen bonds observed that
correlations were better if the acceptors were similar.
11.4.3 Natural Bond Orbital Analysis on Carbon Bonded

Complexes
To understand the nature of interaction and orbitals involved in the carbon bonding,
the NBO analysis was performed on all the complexes under consideration using
wavefunctions for the optimized geometries at MP2/6-311+G(3df,2p) level as well
as MP2/Aug-cc-pVTZ level. The NBO Charge transfer (Q) from the carbon bond
acceptor unit (Y) to the carbon bond donor unit (CH3 X) was calculated from the
natural population analysis and is complied in Table 11.7. The negative value of Q
represents that the charge is transferred to the CH3 X unit which is the case in all the
complexes.
The table also contains second order perturbation energies (E2 ) resulting from the
second order perturbation analysis of Fock matrix in the NBO basis at HF level of
theory. The analysis shows a stabilization due to interaction between the lone pair
orbital ofY with the anti-bonding σ∗ (C-O/C-F/C-Cl/C-Br) orbital. In few cases, more
than one lone pair of electrons contribute towards the stabilization and in such cases
the reported value is sum of the energies from individual contributions, e.g. in case of
LiBr· · ·CH3 F complex all three lone-pairs of Br interact with σ∗ (C-F) orbital and the
second order perturbation energies are 2.3, 1.8 and 0.8 kJ/mol and the reported value
4.9 kJ/mol is the sum of these three contributions. It is clear that the stabilization is
due to the charge transfer from the lone pair orbital(s) of Y to the C-X anti-bonding
orbital. It is similar to the lone-pair to anti-bonding (H-X) charge transfer found in
the case of hydrogen bonding. The charge transfer from the lone-pair(s) of Y to the
C-X anti-bonding orbital of CH3 X is also reflected in C-X vibrational frequency shift
after complex formation as discussed below.
11.4.4 Red Shift in the C-X Vibrational Frequency
It is well known that in the case of hydrogen bonding generally there is a red shift in
H-X stretching frequency. However, cases with blue shift hydrogen bonds [49] and
‘no’ shift hydrogen bonds [50] have also been reported.
Table 11.7 Charge transfer (Q) from the carbon bond acceptor molecule to the donor CH3 X
molecule. Second order perturbation energies for the intermolecular lone-pair and antibonding C-X
orbital interactions are also given. Charges are given in atomic units and energies in kJ.mol−1
Q E2 [l.p. → σ∗ (C-X)] Q E2 [l.p. → s* (C-X)]
CH 3 OH complexes
H2 O· · ·CH3 OH − 0.0014 2.6 − 0.0021 2.8
H2 S· · ·CH3 OH − 0.0011 2.4 − 0.0016 3.0
HF· · ·CH3 OH − 0.0010 1.5 − 0.0016 1.6
HCl· · ·CH3 OH − 0.0007 2.0 − 0.0013 2.3
HBr· · ·CH3 OH − 0.0007 2.0 − 0.0018 3.3
ClF· · ·CH3 OH − 0.0018 2.3 − 0.0027 2.2
LiF· · ·CH3 OH − 0.0022 3.6 − 0.0027 3.6
LiCl· · ·CH3 OH − 0.0021 3.0 − 0.0036 3.1
LiBr· · ·CH3 OH − 0.0019 2.9 − 0.0040 3.9
H3 N· · ·CH3 OH − 0.0017 3.0 − 0.0019 2.8
H3 P· · ·CH3 OH − 0.0009 1.9 − 0.0006 1.8
CH 3 F complexes
H2 O· · ·CH3 F − 0.0020 3.6 − 0.0026 3.6
H2 S· · ·CH3 F − 0.0017 3.3 − 0.0023 4.0
HF· · ·CH3 F − 0.0014 2.3 − 0.0018 2.5
HCl· · ·CH3 F − 0.0009 2.8 − 0.0015 3.1
HBr· · ·CH3 F − 0.0010 2.5 − 0.0022 4.1
ClF· · · CH3 F − 0.0020 3.3 − 0.0026 2.9
LiF· · ·CH3 F − 0.0035 5.9 − 0.0038 5.6
LiCl· · ·CH3 F − 0.0029 4.1 − 0.0029 4.1
LiBr· · ·CH3 F − 0.0033 3.6 − 0.0050 4.9
H3 N· · ·CH3 F − 0.0027 5.1 − 0.0029 4.9
H3 P· · ·CH3 F − 0.0016 2.8 − 0.0020 3.5
CH 3 Cl complexes
H2 O· · ·CH3 Cl − 0.0016 2.5 − 0.0031 2.9
H2 S· · ·CH3 Cl − 0.0011 2.6 − 0.0029 3.3
HF· · ·CH3 Cl − 0.0015 1.5 − 0.0033 2.3
HCl· · ·CH3 Cl − 0.0007 2.1 − 0.0027 2.8
HBr· · ·CH3 Cl − 0.0006 2.0 − 0.0035 3.8
ClF· · ·CH3 Cl − 0.0027 2.3 − 0.0066 3.1
LiF· · ·CH3 Cl − 0.0027 4.3 − 0.0046 5.0

Q E2 [l.p. → σ∗ (C-X)] Q E2 [l.p. → s* (C-X)]
LiCl· · ·CH3 Cl − 0.0024 3.1 − 0.0061 4.0

LiBr· · ·CH3 Cl − 0.0026 2.5 − 0.0073 4.0
H3 N· · ·CH3 Cl − 0.0023 3.6 − 0.0032 4.3
H3 P· · ·CH3 Cl − 0.0009 1.9 − 0.0024 2.6
CH 3 Br complexes
H2 O· · ·CH3 Br − 0.0017 2.7 − 0.0034 3.1
H2 S· · ·CH3 Br − 0.0016 3.0 − 0.0037 3.9
HF· · ·CH3 Br − 0.0017 1.3 − 0.0032 1.6
HCl· · ·CH3 Br − 0.0009 2.3 − 0.0030 2.8
HBr· · ·CH3 Br − 0.0008 2.2 − 0.0039 4.0
ClF· · ·CH3 Br − 0.0028 2.5 − 0.0068 2.6
LiF· · ·CH3 Br − 0.0031 4.8 − 0.0051 5.6
LiCl· · ·CH3 Br − 0.0029 3.4 − 0.0064 4.0
LiBr· · ·CH3 Br − 0.0031 2.7 − 0.0079 4.1
H3 N· · ·CH3 Br − 0.0026 3.9 − 0.0035 4.4
H3 P· · ·CH3 Br − 0.0014 2.2 − 0.0029 2.9
To know the effect on C-X stretch frequency due to carbon bonding, frequency
change corresponding to C-X stretching motion was calculated for all the complexes
from the normal mode analysis. The frequency change (ν) values at both the level
of theories are given in Table 11.8. Negative value of ν implies that the stretching
frequency decreases on complex formation. In case of CH3 F, CH3 Cl and CH3 Br
molecules, the C-X stretching mode is a local mode of vibration. On the other hand,
in CH3 OH molecule, no mode, which is a pure C-O stretching mode, exists. How-
ever, there are two normal modes with significant contribution from C-O stretching.
Frequency shifts on complex formation for both these modes were very similar and
one of these values is reported in the Table 11.8. In all the cases, the ν value is neg-
ative which shows that carbon bonding red shifts the C-X stretching frequency. This
observation is again similar to hydrogen bonding. The frequency shift is largest in
case of complexes with LiY which is in accordance with the maximum stabilization
observed in these complexes.
11.4.5 Correlations Between the Frequency Shift and NBO

Charge Transfer/Second Order Perturbation Energy
Hyperconjugation is considered to be one of the major reasons for the observed red
shift in the X-H stretching frequency upon hydrogen bond formation. The molecular
Table 11.8 Frequency shifts, ν, upon complex formation for CH3 OH, CH3 F, CH3 Cl and CH3 Br complexes. Values from both,
MP2/6-311+G(3df,2p) and MP2/Aug-cc-pVTZ, levels are given
Complex ν/cm−1 Complex ν/cm−1

MP2 MP2 MP2 MP2
/6-311+G(3df,2p) /Aug-cc-pVTZ /6-311+G(3df,2p) /Aug-cc-pVTZ
11 The X-C· · ·Y Carbon Bond
CH3 OH complexes CH3 F complexes

H2 O· · ·CH3 OH − 9.1 − 7.8 H2 O· · ·CH3 F − 16.3 − 14.8
H2 S· · ·CH3 OH − 5.5 − 5.2 H2 S· · ·CH3 F − 10.4 − 11.0
HF· · ·CH3 OH − 6.3 − 6.5 HF· · ·CH3 F − 10.2 − 7.8
HCl· · ·CH3 OH − 4.0 − 3.2 HCl· · ·CH3 F − 6.5 − 4.3
HBr· · ·CH3 OH − 3.3 − 4.1 HBr· · ·CH3 F − 6.3 − 5.7
ClF· · ·CH3 OH − 2.8 − 3.1 ClF· · ·CH3 F − 4.9 − 3.1
LiF· · ·CH3 OH − 20.9 − 20.3 LiF· · ·CH3 F − 40.4 − 39.0
LiCl· · ·CH3 OH − 14.8 − 14.5 LiCl· · ·CH3 F − 29.1 − 26.3
LiBr· · ·CH3 OH − 12.5 − 14.3 LiBr· · ·CH3 F − 24.3 − 24.8
H3 N· · ·CH3 OH − 18.8 − 9.9 H3 N· · ·CH3 F − 22.9 − 20.3
H3 P· · ·CH3 OH − 3.9 − 1.0 H3 P· · ·CH3 F − 9.9 − 8.9
343
344
Complex ν/cm−1 Complex ν/cm−1

MP2 MP2 MP2 MP2
/6-311+G(3df,2p) /Aug-cc-pVTZ /6-311+G(3df,2p) /Aug-cc-pVTZ
CH 3 Cl complexes CH 3 Br complexes
H2 O· · ·CH3 Cl − 9.2 − 9.9 H2 O· · ·CH3 Br − 6.6 − 6.8
H2 S· · ·CH3 Cl − 6.3 − 7.0 H2 S· · ·CH3 Br − 5.5 − 5.5
HF· · ·CH3 Cl − 5.3 − 5.6 HF· · ·CH3 Br − 3.4 − 3.0
HCl· · ·CH3 Cl − 2.4 − 2.8 HCl· · ·CH3 Br − 2.0 − 1.7
HBr· · ·CH3 Cl − 3.4 − 3.9 HBr· · ·CH3 Br − 2.4 − 2.6
ClF· · ·CH3 Cl − 1.3 − 1.3 ClF· · ·CH3 Br − 0.2 0.3
LiF· · ·CH3 Cl − 27.9 − 27.8 LiF· · ·CH3 Br − 19.8 − 21.3
LiCl· · ·CH3 Cl − 18.7 − 18.7 LiCl· · ·CH3 Br − 13.6 − 13.9
LiBr· · ·CH3 Cl − 16.6 − 17.6 LiBr· · ·CH3 Br − 11.5 − 13.4
H3 N· · ·CH3 Cl − 14.1 − 14.9 H3 N· · ·CH3 Br − 10.5 − 11.5
H3 P· · ·CH3 Cl − 6.4 − 7.1 H3 P· · ·CH3 Br − 5.3 − 5.6
E. Arunan and D. Mani
50
25
40
Δν/cm-1
20
R² = 0.848 R² = 0.892
30
Δν/cm-1
15
20
10
5 10
0 0
0.0005 0.0010 0.0015 0.0020 0.0025 0.0005 0.0015 0.0025 0.0035
ΔQ ΔQ
a CH3OH Complexes b CH3F Complexes
30
25 25
Δν/cm-1
20 R² = 0.830 Δν/cm-1 20
R² = 0.832
15 15
10 10
5 5
0 0
0.0005 0.0015 0.0025 0.0035
0.0005 0.0015 0.0025
ΔQ ΔQ
c CH3Cl Complexes d CH3Br Complexes
Fig. 11.7 Plots between NBO charge transfer (Q) and C-X frequency shift (ν) for the a CH3 OH,
b CH3 F, c CH3 Cl and d CH3 Br complexes. Data points for ClF· · ·CH3 X complexes are marked
with brown squares and these were not included in the fit
orbitals involved in the hyperconjugation are usually lone pair orbitals of the hydro-
gen bond acceptor and the sigma anti-bonding orbital of the hydrogen bond donor
[σ*(H-X)]. The electron transfer from the lone pair orbital(s) to the anti-bonding or-
bital results in the weakening of the H-X sigma bond and thus, results in the decrease
of the stretching frequency. It is clear from the NBO data given in Table 11.7, the
orbitals involved in the X-C· · ·Y carbon bond interactions are l.p.(s) of Y and anti-
bonding C-X sigma orbital (σ*(C-X)). Correlations between (i) total charge transfer
from acceptor AY moiety to CH3 X moiety (Q) and C-X frequency shift (ν) and
(ii) second order perturbation energy for the l.p.(s) – σ*(C-X) orbital interaction and
the C-X frequency shift (ν), are discussed below.
11.4.5.1 Charge Transfer (Q) vs. C-X Stretching Frequency Shift (ν)
The C-X stretching frequency shifts were plotted against the charge transfer (Q)
obtained from the NBO analysis. The plots are shown in Figs. 11.7a–11.7d for the
different CH3 X (X = OH/F/Cl/Br) complexes. Correlation coefficients resulting
from the linear fits are also given in the plots. In none of the plots, data for the
a b
c d
Fig. 11.8 Plots between second order perturbation energy, E2 (l.p. → σ*), and C-X frequency shift
(ν) for the a CH3 OH, b CH3 F, c CH3 Cl and d CH3 Br complexes. Data points for the ClF· · ·CH3 X
complexes are marked with brown squares and these were not included in the fit
ClF· · ·CH3 X complexes was included since the frequency shift for these complexes
are very small and the fit was worsening on including the data. In all the cases,
correlation coefficient is more than 0.8 which shows the correlation between these
two properties is reasonable.
11.4.5.2 Perturbation Energy vs. C-X Stretching Frequency Shift
Other set of correlation diagrams were plotted between the second order perturbation
energy, E2 (l.p. → σ*(C-X)), and the C-X stretch frequency shift (ν). The plots are
shown in Fig. 11.8a–11.8d for the different CH3 X (X = OH/F/Cl/Br) complexes.
Correlation coefficients are also given in each case. Again the data points for the ClF
complexes were not included in the fit. The correlation coefficient for the CH3 OH
and the CH3 F complexes is nearly 0.8 but is poorer for the CH3 Cl and the CH3 Br
complexes, ∼ 0.74.
11.4.6 The ESP Value at CH3 Face Centre vs. Binding Energy
In this section, we discuss the correlation between the ESP values at CH3 face centre
of the CH3 X molecules and binding energies, keeping the carbon bond acceptor same.
Values at the MP2/6-311+G(3df,2p) level were considered for these correlations.
The correlation plots are given in Fig. 11.9a–11.9c. Figure 11.9a shows correlation
fits for the H2 O· · ·CH3 X (X = OH/F/Cl/Br) and H2 S· · ·CH3 X (X = OH/F/Cl/Br)
complexes. For both the sets the correlation is quite good and the correlation coeffi-
cients are 0.980 and 0.914, respectively. Figure 11.9b shows the correlation for the
HF· · ·CH3 X (X = OH/F/Cl/Br), HCl· · ·CH3 X (X = OH/F/Cl/Br) and HBr· · ·CH3 X
(X = OH/F/Cl/Br) complexes. The fits are reasonable in the case of HF and HBr
complexes, correlation coefficients being 0.905 and 0.920 respectively. However, in
the case of the HCl complexes the correlation is poor and the correlation coefficient
is 0.762. Figure 11.9c shows the correlation for the LiF· · ·CH3 X (X = OH/F/Cl/Br),
LiCl· · ·CH3 X (X = OH/F/Cl/Br) and LiBr· · ·CH3 X (X = OH/F/Cl/Br) complexes.
The correlation coefficients in these cases are not very good and these are 0.860,
0.754 and 0.685, respectively. Figure 11.9d shows the correlation plots for the
H3 N· · ·CH3 X (X = OH/F/Cl/Br) and H3 P· · ·CH3 X (X = OH/F/Cl/Br) complexes.
In these cases the correlation coefficients are 0.954 and 0.857, respectively.
In general, there is some correlation between the ESP value at the CH3 face centre
and the binding energy if the acceptor is kept fixed. The correlation coefficients are
reasonable in some cases and not very satisfactory in others. These observations
indicate that the electrostatics does play an important role in carbon bond but it is
not all.
It is clear from the above discussion that carbon bonds follow most of the es-
tablished criteria for intermolecular bonding. Various properties show expected
correlation in these complexes.
In the following sections we extend our discussion to carbon bonding with other
types of acceptor molecules.
11.5 Other Acceptors/Donors for Carbon Bonds
11.5.1 Radicals as Carbon Bond Acceptor
Radicals can act as hydrogen bond acceptor and single electron hydrogen bonds
[51, 52] have been reported in the literature. Li et al. [22] have recently shown that
single electron carbon bonds also exist. They studied complexes of CH3 F, CH3 CN,
CH3 NC and CH3 NO2 with methyl radical and found a minimum geometry in which
the methyl radical faces the CH3 face of CH3 X molecule as shown in Fig. 11.10.
The intermolecular C· · ·C distance and binding energy for this complex were found
to 3.486 Å and 2.8 kJ.mol−1 respectively at UMP2/Aug-cc-pVTZ level. The AIM
analysis confirms bonding between the two carbon atoms in this complex.
9.0
5.0 R² = 0.905
R² = 0.980
HF_complexes
8.0
H2O_Complexes HCl_complexes
4.5
H2S_complexes HBr_complexes R² = 0.762
7.0 Linear (H2O_Complexes) Linear (HF_complexes)
Linear (H2S_complexes) 4.0 Linear (HCl_complexes)
Linear (HBr_complexes)
6.0
Δ EBSSE Δ EBSSE 3.5 R² = 0.920
(kJ/mol) (kJ/mol)
5.0
3.0
R² = 0.914
4.0
2.5
3.0
2.0
2.0 45.0 55.0 65.0 75.0 85.0 95.0
45.0 55.0 65.0 75.0 85.0 95.0 ESP (kJ/mol)
ESP (kJ/mol)
a b
9.0
18.5
LiF_complexes
R² = 0.860 8.0 R² = 0.954
NH3_complexes
16.5 LiCl_complexes
LiBr_complexes PH3_complexes
7.0 Linear (NH3_complexes)
Linear (LiF_complexes)
14.5 Linear (LiCl_complexes) Linear (PH3_complexes)
Linear (LiBr_complexes) 6.0

Δ EBSSE R² = 0.754 Δ EBSSE
12.5 (kJ/mol)
(kJ/mol) 5.0
R² = 0.857
10.5
R² = 0.685 4.0
8.5
3.0
6.5 2.0
45.0 55.0 65.0 75.0 85.0 95.0 45.0 55.0 65.0 75.0 85.0 95.0
ESP (kJ/mol) ESP (kJ/mol)
c d
Fig. 11.9 a Correlation plot between ESP at the CH3 face centre of CH3 X and binding energy for
H2Y· · ·CH3 X (Y = O/S and X = OH/F/Cl/Br) complexes. b Correlation plot between ESP at the
CH3 face centre of CH3 X and binding energy for HY· · ·CH3 X (Y = F/Cl/Br and X = OH/F/Cl/Br)
complexes. c Correlation plot between ESP at the CH3 face centre of CH3 X and binding energy for
LiY· · ·CH3 X (Y = F/Cl/Br and X = OH/F/Cl/Br) complexes. d Correlation plot between ESP at
the CH3 face centre of CH3 X and binding energy for H3Y· · ·CH3 X (Y = N/P and X = OH/F/Cl/Br)
complexes
Fig. 11.10 Optimized geometry for CH3 · · ·CH3 F complex at UMP2/Aug-cc-VTZ level (Repro-
duced with permission from RSC, reference 22)
It was found that subsequent substitutions of the hydrogen atoms of the accep-
tor methyl radical with electron donating CH3 groups result in stronger bonding.
The binding energies for the complexes of CH3 , CH2 CH3 , CH(CH3 )2 and C(CH3)3
radicals with CH3 F were found to be 2.8, 4.6, 6.1 and 7.3 kJ.mol−1 respectively.
11.5.2 Anions as Carbon Bond Acceptors
Grabowski considered complexes of carbon bond donors with atomic anions to

explore SN 2 reactions [20]. McDowell and Joseph [21] studied these interactions
in more detail. As expected the binding energies in these complexes were much
higher than those for the complexes with the neutral species, e.g. binding energy
for F(−) · · ·CH3 F complex was found to be 54.2 kJ.mol−1 which is much higher in
comparison to the values 5.1 and 16.8 kJ.mol−1 for the HF· · ·CH3 F and LiF· · ·CH3 F
complexes respectively. Interestingly, in case of carbon bonding with anions the in-
teraction energy follows the pattern CH3 Br complexes > CH3 Cl complexes > CH3 F
complexes. Note that this pattern is opposite to the pattern observed in case of car-
bon bonding with lone pair of electrons as acceptors. However, this pattern could
be explained on the basis of the dipolar polarizability of the C-X bonds. The dipolar
polarizability for these bonds decreases in the order C-Br > C-Cl > C-F. Thus, the
polarization of the C-X bond by anion increases with increase in the size of the X
atom, from F to Br. This results in stronger bonding.
More recently, Li et al. [31] have studied complexes of metal hydrides (LiH,
BeH, NaH and MgH) with carbon bond acceptor CH3 F molecule as well as with
other tetrel molecules. For these complexes, the binding energy increases in the
order BeH· · ·CH3 F < LiH· · ·CH3 F < NaH· · ·CH3 F.
11.5.3 π-Electrons, σ-Electrons and Neutral Atoms as Carbon

Bond Acceptor
The π-electrons are well known acceptors for hydrogen bonds [4–6]. In a recent pub-
lication [53], we have shown that the π-electrons can act as carbon bond acceptor too
and lead to the formation of X-C· · · π carbon bonds. Ethylene and acetylene were
considered as the π-electron donors. Both the molecules, ethylene and acetylene,
have no dipole moments and therefore the carbon bonded complexes with CH3 X
were dominated by dispersion forces. All the complexes follow the AIM and NBO
bonding criteria and also show red shifts in the C-X stretching frequencies on com-
plexation. Complexes with molecules containing other tetrel atoms, Si and Ge, were
also considered and it was found that the interaction energy increases in the order
CH3 X < SiH3 X < GeH3 X, which is similar to that observed for other cases. How-
ever, electron density topology for most of the SiH3 X complexes was found to be
very different than the CH3 X and GeH3 X complexes and no bond critical point was
found connecting the π-electrons (C-atom) to the silicon atom.
Can σ-electrons and neutral atoms also act as the carbon bond acceptors? This was
explored considering complex formation between H2 molecule and Ar atom with
CH3 F molecule. The H2 · · ·CH3 F complex was optimised taking initial geometry
in which the σ-electrons of H2 bond critical point towards the CH3 face of CH3 F.
Similarly, theAr· · ·CH3 F complex was optimised taking the initial geometry in which
Fig. 11.11 Electron density topologies for the a H2· · ·CH3 F and b Ar· · ·CH3 F complexes
Ar atom points towards the CH3 face of the CH3 F molecule. Geometry optimisation
was performed at the MP2/Aug-cc-pVTZ level. The BSSE corrected binding energies
for the H2 · · ·CH3 F and Ar· · ·CH3 F complexes are 1.3 and 1.9 kJ.mol−1 ,respectively.
The AIM analysis was performed on both the complexes and the resulting electron
density topologies are shown in Fig. 11.11a and 11.11b. The topologies confirm
interaction between the σ-electrons (H-atom)/Ar and carbon atom in the H2 · · ·CH3 F
and Ar· · ·CH3 F complexes, respectively. The (ρ, ∇ 2 ρ) value at the intermolecular
bond critical point are (0.0030, +0.0145) for the H2 · · ·CH3 F complex and (0.0037,
+0.0190) for the Ar· · ·CH3 F complex.
11.5.4 Carbon Bonding in Cycloalkanes
Perfluoro cycloalkanes (CFn ) and percyano cycloalkanes (CCN n ) have been shown to
be potential carbon bond donors by Bauzá et al. [27, 28]. Large cycloalkanes, like
perfluorocyclohexane (CF6 ), possess very strong sigma hole at the centre of the ring in
their planar conformation. It is similar to the sigma hole found in case of molecules
like perfluorobenzene (C6 F6 ). However, the planar conformations are not the most
stable ones for the large ring cycloalkanes and the interaction with nucleophiles
can not compensate for the conformational instability. The more stable non-planar
conformers of these molecules, like chair form of perfluorocyclohexane, have much
weaker sigma hole. Moreover, such sigma holes are not accessible to electron rich
centres due to their proximity to the electron repelling F or CN groups. However,
the smaller cycloalkane rings, like C3 and C4 are conformarily more rigid. Perfluro
and percyano, cyclopropanes and cyclobutanes form multiple carbon bonds through
the edges as well as ring faces.
11.6 Manifestations of Carbon Bonding

11.6.1 Carbon Bonding and Hydrophobic Interaction
Carbon bonding is very closely related to hydrophobic interactions. Let us consider

methanol molecule as an example. The hydroxyl group of methanol is hydrophilic
and attracts H2 O molecule to form a hydrogen bond. However, the methyl group in
this molecule is hydrophobic in nature and any interaction such as the one shown
in Fig. 11.3 will be considered as hydrophobic interaction. Hydrophobic effect is
usually considered to be an entropic phenomenon, and the enthalpic contribution
was considered as the generic van der Waals interaction. However, the discovery
of carbon bond shows that there is a specific, small but non negligible enthalpic
contribution from carbon bond to the hydrophobic effect.
11.6.2 Carbon Bonding and SN 2 Reactions
The carbon bonded complexes resemble the entrance channel complexes of an SN 2

reaction. We pointed out [17] that the carbon bonding could be the stabilizing factor
of the intermediates to the SN 2 reactions. Grabowski [20] explored the role of carbon
bonding in SN 2 reactions in detail. Carbon bonding interactions in which anions act
as the acceptor are quite close to the SN 2 reaction intermediates. These interactions
are quite strong as well, as in such complexes the CH3 face is significantly deformed
towards planarity. We recognize that the ‘carbon bonded geometries’ as stable min-
imum along the reaction coordinate for SN 2 reactions have been noted earlier by
theoretical chemists exploring the potential surface and reaction dynamics of these
reactions [54, 55]. However, their main focus was reaction dynamics and hence the
similarity of these structures with that of hydrogen bonded complexes was not con-
sidered. The fact that SN 2 reactions happen offers the most convincing proof that
there is electron communication between the Y and C atoms leading to the formation
of the carbon bond. Such structures could have been misidentified in the past as
‘trifurcated hydrogen bonds’.
11.6.3 Carbon Bonding in Biological Systems
Alkyl group bonded to electronegative atoms are present in abundance in the bio-
logical systems. Such abundance implies that the carbon bonding interactions could
be present in these systems and could contribute to the functioning of the biolog-
ical molecules. For example, in amino acids (NH2 RCHCOOH) the -RCH- group
is sandwiched between two electron withdrawing groups, NH2 and COOH. The
middle carbon atom is positively charged and can participate in the carbon bond-
ing interactions. This possibility was indeed checked in the glycine· · ·FH complex.
Calculations showed that the geometry in which fluorine of HF interacts with CH2
carbon of glycine is a minimum. Alanine (R = CH3 ) has a CH3 group and thus is
closer to the CH3 X molecules, already discussed. The ESP calculations show the
presence of a maximum at the CH3 face of alanine, with ESP value 19.6 kJ.mol−1 .
It shows that the CH3 face of alanine should form carbon bonds with molecules like
HF, similar to the other CH3 X molecules. We optimized complex of alanine with HF
at MP2/6-311+G(3df,2p) level of theory. The initial geometry was taken such that
the F atom of HF faces the CH3 face of alanine molecule. The optimized structure
was confirmed to be a minimum by frequency calculation.
Fig. 11.12 X-C· · ·Y carbon

bonding interaction in the
HF· · ·alanine complex
The stabilization energy for this complex is 2.0 kJ.mol−1 . The AIM calculations
show the presence of a BCP between F and C and a bond path connecting them,
Fig. 11.12. The values of ρ(r) and ∇ 2 ρ(r) at the BCP are 0.0039 and +0.0205 a.u.
respectively. This example shows that the X-C· · ·Y interactions could also be present
in the biological systems and can play an important role in the biological processes
such as protein folding.
Azofra et al. [26] have analysed carbohydrate molecules in gas phase using Density
Functional Theory. They found intra-molecular O· · ·C-O carbon bonds in many of
the structures, cyclic as well as open chain.
11.7 Tetrel Bond
Interactions similar to carbon bonding exist for all the elements of the IV group (Si,
Ge, Sn). These interactions were named tetrel bond by Grabowski [20] and Bauzá
et al. [24], simultaneously but independently. Tetrel bonds have been explored in
detail now and like carbon bonds, tetrel bonds with radical [22] as well as anion
acceptors [21, 23, 27, 29] have been reported in the literature. The interaction energy
increases as one goes down the group from C to Sn. The reason follows: as the
atomic size increases from C to Sn the atom becomes more polarizable and thus the
sigma-hole becomes more prominent in the TH3 X (T = C, Si, Ge, Sn and X is an
electron withdrawing atom or group) molecules. As an example, binding energies
for the complexes of LiCN with CH3 F, SiH3 F and GeH3 F are 15.9, 32.2 and 35.9
kJ.mol−1 , respectively.
11.8 Experimental Evidence for Carbon Bonding
Following the first publication on carbon bond [17], Guru Row and co-workers car-
ried out a systematic charge density analysis on specific crystals expected to show
carbon bonding [32]. They chose two molecular systems in which carbon bonding
Fig. 11.13 Structure of a fenobam and b dimethylammonium-4-hydroxybenzoate, having potential

for carbon bonds. Electron density analysis shows C-H· · ·Cl hydrogen bond in the former (I) and
N-C· · ·O carbon bonding in the later (II). (Reproduced with permission from RSC, Reference 31)
could be found and solved their crystal structures, following it with charge den-
sity analysis. These were (i) the anxiolytic drug candidate fenobam (Fig. 11.13a)
for a possible C· · ·Cl carbon bond, and (ii) dimethylammonium-4-hydroxybenzoate
(Fig. 11.13b). Charge density analysis confirmed the N-C· · ·O carbon bond in the
crystal of dimethylammonium-4-hydroxybenzoate (Fig. 11.13 II) but in the fenobam,
it revealed the formation of C-H· · ·Cl hydrogen bond (Fig. 11.13 I). Indeed, ex-
perimental and theoretical charge density analyses are important in confirming the
presence of carbon bond.
We note that a recent microwave spectroscopic investigation [56] on pyridine-CF4
complex has led to a structure in which one of the tetrahedron face of CF4 is pointing
towards the N of pyridine. This could indeed be an example of an F-C· · ·N carbon
bond. The difference in electron density distribution found between hydrocarbon
and perfluorohydrocarbons have been noted for long. For example, benzene has an
electron rich centre whereas the perfluorobenzene has its centre, electron depleted.
This manifests in their interaction with other molecules. The H2 O interacts with
benzene through O–H· · · π hydrogen bond(s) [57] but it interacts with C6 F6 with
the O in H2 O pointing towards the electron depleted π centre [58]. Indeed, this
interaction would bear similarities to the ‘carbon bond’. Bauzá et al. [59] have
recently found similar interaction with C6 H5 CF3 molecule using a combined CSD
and ab intio analysis. The analysis showed many structures in which electron rich
atoms interact with the carbon atom of CF3 group through the face centre.
11.9 Conclusions
Carbon bond is an emerging field of research. Much is to be known about these

interactions, especially in terms of their presence in nature as well as their utilization
in synthetic approaches. Other tetrel bonds, with silicon and germanium, are much
stronger than the carbon bonds. However, abundance of carbon in biological systems
is likely to make carbon bonding as the most important tetrel bond. From a synthetic
point of view, hydrogen bonds mediate proton transfer reactions and carbon bonds
mediate SN 2 reactions. Carbon bonds, no doubt, contribute to hydrophobic interac-
tions. Strong carbon bonding interactions with molecules like LiF or anions might
be utilized in the field of supramolecular chemistry and these carbon bond acceptors
might prove to be useful synthons in the field.
Acknowledgements Authors thank financial support from the Indo-French Center for Promotion
ofAdvanced Scientific Research and Indian Institute of Science. DM thanks Council of Scientific and
Industrial Research for a fellowship. Authors thank Abhishek Shahi and Sharon Priya Gnanasekhar
for helpful discussions.
References
1. Hobza P, Müller-Dethlefs K (2010) Non-covalent interactions: theory and experiment. Royal

2. Karshikoff A (2006) Non-covalent interactions in proteins. World Scientific
3. Pauling L (1960) The nature of the chemical bond and the structure of molecules and crystals:
an introduction to modern structural chemistry. Cornell University Press
4. Arunan E, Desiraju GR, Klein RA et al (2011) Definition of the hydrogen bond. Pure Appl
Chem 83:1637–1641
5. Jeffrey GA, Saenger W (1994) Hydrogen bonding in biological structures. Springer-Verlag,
Berlin
6. Desiraju GR, Steiner T (1999) The weak hydrogen bond in structural chemistry and biology.
Oxford University Press, Oxford
7. Shahi A, Arunan E (2014) Hydrogen bonding, halogen bonding and lithium bonding: an atoms
in molecules and natural bond orbital perspective towards conservation of total bond order,
inter-and intra-molecular bonding. Phys Chem Chem Phys 16:22935–22952
8. Politzer P, Lane P, Concha M et al (2007) An overview of halogen bonding. J Mol Model
13:305–311
9. Wang W, Ji B, Zhang Y (2009) Chalcogen bond: a sister noncovalent bond to halogen bond. J
Phys Chem A 113:8132–8135
10. Scheiner S (2011) A new noncovalent force: comparison of P [center-dot] [center-dot] [center-
dot] N interaction with hydrogen and halogen bonds. J Chem Phys 134:094315–094319
11. Metrangolo P, Resnati G (eds) (2008) Halogen bonding: fundamentals and applications, of
structure and bonding, vol 126. Springer, Berlin
12. Clark T, Hennemann M, Murray JS et al (2007) Halogen bonding: the σ-hole. J Mol Model
13:291–296
13. Murray J, Lane P, Politzer P (2009) Expansion of the σ-hole concept. J Mol Model 15:723–729
14. Politzer P, Murray JS (2013) Halogen bonding: an interim discussion. ChemPhysChem 14:278–
294
15. Bundhun A, Ramasami P, Murray J et al (2013) Trends in σ-hole strengths and interactions of
F3 MX molecules (M = C, Si, Ge and X = F, Cl, Br, I). J Mol Model 19:2739–2746
16. Mani D, Arunan E (2013) Microwave spectroscopic and atoms in molecules theoretical investi-
gations on the Ar· · ·Propargyl Alcohol Complex: Ar· · ·H–O, Ar· · ·π, and Ar· · ·C interactions.
17. Mani D, Arunan E (2013) The X-CY (X = O/F, Y = O/S/F/Cl/Br/N/P) ‘carbon bond’ and
hydrophobic interactions. Phys Chem Chem Phys 15:14377–14383
18. Farrugia LJ, Evans C, Tegel M (2006) Chemical bonds without “chemical bonding”? A
combined experimental and theoretical charge density study on an iron trimethylenemethane
complex. J Phys Chem A 110:7952–7961
19. Raghavendra B, Arunan E (2008) Hydrogen bonding with a hydrogen bond: the methane–water
complex and the penta-coordinate carbon. Chem Phys Lett 467:37–40
20. Grabowski SJ (2014) Tetrel bond-σ -hole bond as a preliminary stage of the SN 2 reaction.
21. Mcdowell SA, Joseph JA (2014) The effect of atomic ions on model σ-hole bonded complexes
of AH3Y (A = C, Si, Ge; Y = F, Cl, Br). Phys Chem Chem Phys 16:10854–10860
22. Li Q, Guo X, Yang X et al (2014) A σ-hole interaction with radical species as electron donors:
does single-electron tetrel bonding exist? Phys Chem Chem Phys 16:11617–11625
23. Bauzá A, Ramis R, Frontera A (2014) Computational study of anion recognition based on
tetrel and hydrogen bonding interaction by calix [4] pyrrole derivatives. Comput Theor Chem
1038:67–70
24. Bauzá A, Mooibroek TJ, Frontera A (2013) Tetrel-bonding interaction: rediscovered
supramolecular force? Angew Chem Int Ed 52:12317–12321
25. Varadwaj PR, Varadwaj A, Jin B (2014) Significant evidence of CO and CC long-range contacts
in several heterodimeric complexes of CO with CH3 -X, should one refer to them as carbon and
dicarbon bonds! Phys Chem Chem Phys 16:17238–17252
26. Azofra LM, Quesada-Moreno MM, Alkorta I et al (2014) Carbohydrates in the gas phase:
conformational preference of d-ribose and 2-deoxy-d-ribose. New J Chem 38:529–538
27. Bauza A, Mooibroek TJ, Frontera A (2014) Influence of ring size on the strength of car-
bon bonding complexes between anions and perfluorocycloalkanes. Phys Chem Chem Phys
16:19192–19197
28. Bauzá A, Mooibroek TJ, Frontera A (2014) Small cycloalkane (CN) 2C–C (CN) 2 structures
are highly directional non-covalent carbon-bond donors. Chem-A Eur J 20:10245–10248
29. Mcdowell SA (2014) Sigma-hole cooperativity in anionic [FX· · · CH3 · · ·YF]− (X, Y = Cl, Br)
complexes. Chem Phys Lett 598:1–4
30. Arunan E (2013) Hydrogen bond seen, halogen bond defined and carbon bond proposed:
intermolecular bonding, a field that is maturing! Curr Sci 105:892–894
31. Li Q-Z, Zhuo H-Y, Li H-B et al (2014) Tetrel–hydride interaction between XH3 F (X = C, Si,
Ge, Sn) and HM (M = Li, Na, BeH, MgH). J Phys Chem A. doi:10.1021/jp503735u
32. Thomas SP, Pavan MS, Row TG (2014) Experimental evidence for ‘carbon bonding’in the
solid state from charge density analysis. Chem Commun 50:49–51
33. Politzer P, Murray JS, Clark T (2013) Halogen bonding and other [sigma]-hole interactions: a
34. Gadre SR, Shirsat RN (2000) Electrostatics of atoms and molecules. Universities Press,
Hyderabad
35. Sjoberg P, Politzer P (1990) Use of the electrostatic potential at the molecular surface to interpret
and predict nucleophilic processes. J Phys Chem 94:3959–3961
36. Suenram RD, Fraser GT, Lovas FJ et al (1994) The microwave spectrum of methane. . . water.
J Chem Phys 101:7230–7240
37. Legon AC, Roberts BP, Wallwork AL (1990) Rotational spectra and geometries of the gas-phase
dimers methane. . . HF and methane. . . HCl. Chem Phys Lett 173:107–114
38. Ohshima Y, Endo Y (1990) Rotational spectrum and internal rotation of a methane–HCl
complex. J Chem Phys 93:6256–6265
39. Legon AC, Wallwork AL (1992) Methane as a proton acceptor: rotational spectrum and internal
dynamics of a weakly bound dimer of methane and hydrogen cyanide. J Chem Soc Faraday
Trans 88:1–9
40. Mani D (2013) Microwave spectroscopic and atoms in molecules theoretical investigations on
weakly bound complexes: from Hydrogen Bond to ‘Carbon Bond’. Ph.D. thesis. In: Department
of Inorganic and Physical Chemistry Indian Institute of Science Bangalore
41. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Scalmani G,
Barone V, Mennucci B, Petersson GA, Nakatsuji H, Caricato M, Li X, Hratchian HP, Izmaylov
AF, Bloino J, Zheng G, Sonnenberg JL, Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa
J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Vreven T, Montgomery JA, Peralta
JE, Ogliaro F, Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN, Kobayashi R,
Normand J, Raghavachari K, Rendell A, Burant JC, Iyengar SS, Tomasi J, Cossi M, Rega
N, Millam JM, Klene M, Knox JE, Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts
R, Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Martin RL,
Morokuma K, Zakrzewski VG, Voth GA, Salvador P, Dannenberg JJ, Dapprich S, Daniels AD,
Farkas, Foresman JB, Ortiz JV, Cioslowski J, Fox DJ (2009) Gaussian 09, Revision B.01. In:
Wallingford CT
of separate total energies. Some procedures with reduced errors. Mol Phys 19:553
43. Dennington R, Keith T, Millam J (2009) GaussView, version 5. Semichem Inc., Shawnee
Mission KS
44. Lu T, Chen F (2012) Multiwfn: a multifunctional wavefunction analyzer. J Comput Chem
33:580–592
45. Aimall K (Version 13.02.26), Todd A. AIMAll (Version 12.09.23). In: TK Gristmill Software,
Overland Park KS, USA, 2012 (aim.tkgristmill.com)
46. Biegler-König F, Schönbohm J (2002) Update of theAIM2000-Program for atoms in molecules.
J Comput Chem 23:1489–1494
47. Glendening ED, Landis CR, Weinhold F (2013) NBO 6.0: natural bond orbital analysis
program. J Comput Chem 34:1429–1437
48. Koch U, Popelier PLA (1995) Characterization of C-H-O hydrogen bonds on the basis of the
charge density. J Phys Chem 99:9747–9754
50. Joseph J, Jemmis ED (2007) Red-, blue-, or no-shift in hydrogen bonds: a unified explanation.
J Am Chem Soc 129:4620–4632
51. Alkorta I, Rozas I, Elguero J (1998) Radicals as hydrogen bond acceptors. Ber Bunsengesells
Physikalische Chem 102:429–435
52. Raghavendra B, Arunan E (2007) Unpaired and σ bond electrons as H, Cl, and Li
bond acceptors: an anomalous one-electron blue-shifting chlorine bond. J Phys Chem A
111:9699–9706
53. Mani D, Arunan E (2014) The X-C· · ·π (X = F, Cl, Br, CN) carbon bond. J Phys Chem A
118:10081–10089
54. Manikandan P, Zhang J, Hase WL (2012) Chemical dynamics simulations of X– + CH3Y →
XCH3 + Y– Gas-phase SN2 nucleophilic substitution reactions. Nonstatistical dynamics and
nontraditional reaction mechanisms. J Phys Chem A 116:3061–3080
55. Gonzales JM, Cox RS, Brown ST et al (2001) Assessment of density functional theory for
model SN2 Reactions: CH3X + F– (X = F, Cl, CN, OH, SH, NH2, PH2). J Phys Chem A
105:11327–11346
56. Maris A, Favero LB, Velino B et al (2013) Pyridine-CF4: a molecule with a rotating cap. J
Phys Chem A 117:11289–11292
57. Gutowsky HS, Emilsson T, Arunan E (1993) LowâŁłJ rotational spectra, internal rotation, and
structures of several benzene–water dimers. J Chem Phys 99:4883–4893
58. Mishra BK, Sathyamurthy N (2007) Van der Waals Complexes of Small Molecules with Ben-
zenoid Rings: influence of Multipole Moments on Their Mutual Orientation. J Phys Chem A
111:2139–2147
59. Bauza A, Mooibroek TJ, Frontera A (2014) Non-covalent sp3 carbon bonding with ArCF3 is
analogous to CH-[small pi] interactions. Chem Commun 50:12626–12629
Chapter 12
Interplay of Hydrogen, Halogen, Lithium and
Beryllium Bonds in Complexes of Thiirane
Sean A. C. McDowell and Jerelle A. Joseph
Abstract Hydrogen, halogen, lithium and beryllium bonding are briefly surveyed as
a prelude to a report of a computational study of the interplay between these various
non-covalent interactions. Our study used model dimers and trimers involving the
thiirane molecule, (CH2 )2 S, complexed with small molecules like HF, ClF, BrF,
LiF and BeH2 to assess and investigate the interplay between the different non-
covalent interactions. The model trimer systems show positive cooperative effects
when thiirane is one of the terminal molecules, whereas a negative cooperative
effect is evident when it is at the center of the trimer. The changes in selected
molecular properties, including the redistribution of charge densities obtained by the
natural population analysis (NPA), implemented in the natural bond orbital (NBO)
procedure, and an Atoms in Molecules (AIM) topological analysis, were useful in
understanding these cooperative effects.
12.1 Introduction
12.1.1 Overview of Non-Covalent Interactions
Non-covalent interactions in molecular systems is a topic of much scientific interest as

it spans a number of scientific subdisciplines. They are responsible for the existence
of condensed phases, for which the forces operative between interacting neutral
molecules are strong enough to limit the volume of samples of a particular material
but weak enough to ensure the fluidity of such materials in bulk.
Such interactions are usually dominated by classical electrostatic forces, through
the interaction between the permanent multipole moments of the interacting molecu-
lar species. These electrostatic forces are usually augmented by classical polarization
and quantum-mechanical dispersion forces, and their long-range behavior dependent
on some inverse power of their molecular separation (usually measured by the dis-
tance between the center of masses of the interacting molecules). At equilibrium
S. A. C. McDowell () · J. A. Joseph

Department of Biological and Chemical Sciences,
The University of the West Indies, Cave Hill Campus, Wanstead, Barbados
e-mail: sacm@mail.com

358 S. A. C. McDowell and J. A. Joseph
separations, the electron-electron (Pauli) exchange repulsion plays a more promi-

nent role and helps to modify the closeness of approach of the interacting subunits,
as well as the shape of the resulting molecular complexes and clusters.
12.1.1.1 Hydrogen Bonding
There are a number of non-covalent interactions which are of scientific, as well as,
technological interest. The most well-known and well-studied of these interactions
is hydrogen bonding, which may be represented schematically by A-H. . .Y, with A
being an atom or fragment which is more electron-withdrawing than the covalently-
attached proton, which enjoys an attractive interaction with the electron-rich regions
(e.g. lone pairs, π-electrons etc.) on a proton acceptor Y [1–7].
The hydrogen bond has been extensively studied and documented for more than
a century now. This important concept continues to be relevant especially since
considerable interest has developed in recent times for a variety of other non-covalent
bonds, most prominently, for example, in halogen bonding [8–10], but also for other
interactions with exotic names like chalcogen [11, 12], pnicogen [13] and tetrel
bonding [14, 15]. The latter four interactions can all be considered as sub-sets of
the more general concept of “σ-hole” bonding, with the differences between each
type determined by the type of atom which gives rise to the σ-hole—Group IV atom
(tetrel), Group V (pnicogen), Group VI (chalcogen) and Group VII (halogen).
A number of books have been written over the years about the hydrogen bond,
with the seminal work by Pimentel and McClellan in 1960 [1] setting the stage for the
future advances and current understanding of this concept. More recent monographs
include Theoretical Treatments of Hydrogen Bonding [2]; Hydrogen Bonding: A
Theoretical Perspective [4]; An Introduction to Hydrogen Bonding [3]; The Weak
Hydrogen Bond [5]; a Frontiers article in 2008 also gives a future outlook on this
topic [6].
Although the hydrogen bond derives its strength from the different relative con-
tributions of the electrostatic, induction, charge transfer, dispersion and exchange
repulsion energies to its overall energy, a good description of the attractive compo-
nent of conventional hydrogen bonds (in gas and condensed phases) is possible with a
combined electrostatic-plus-induction energy model [6]. The normal hydrogen bond
is usually characterized by an increase in the A-H bond length due to the attraction
of the electron-rich site of Y for the proton of A-H which is accompanied by a de-
crease of its vibrational frequency (a red shift) and increase in its IR intensity [4].
The transfer of electron density from the proton acceptor Y to the antibonding A-H
σ∗ orbital is also thought to occur and has the effect of weakening the A-H bond,
resulting in its elongation.
However, in some instances there is an increase in the A-H stretching frequency
on complex formation (a blue shift), usually accompanied by compression of the A-
H bond and a decrease in IR intensity of the A-H stretching frequency. The reversal
of the normal features of hydrogen bonding in the so-called blue-shifting hydrogen
bonds received considerable attention about a decade ago, and provoked a lively
12 Interplay of Hydrogen, Halogen, Lithium and Beryllium Bonds . . . 359
and sometimes controversial discourse in the literature about its nature and origin
and especially whether it was fundamentally different from the normal red-shifting
analogue.
Numerous theoretical papers have been devoted to elucidation of the blue-shift
phenomenon, with some researchers favouring an electrostatic model, while others
advocate charge-transfer mechanisms [6]. A theoretical analysis using a perturbation
theory model has been found useful in describing both red- and blue-shifting behavior
in a set of model hydrogen-bonded complexes and this work suggests that both
types of behavior can be described by the same theory [6]. Although there are many
hydrogen-bonded complexes, the great majority are of the usual red-shifting variety
and it would appear that the blue-shifting phenomenon is dependent on specific
characteristics of the individual interacting molecular subunits which pre-dispose
them to blue-shifting behavior. For example, the change in the permanent dipole
moment of the proton donor when the A-H bond length increases and the chemical
“hardness” of the atom of the proton acceptor to which the A-H molecule is directly
bonded, appear to be two important parameters that determine whether the A-H. . .Y
hydrogen bond is red- or blue-shifted [6].
12.1.1.2 Halogen Bonding
As mentioned earlier non-covalent interactions such as halogen bonds have been more
extensively studied in recent years. Halogen bonds refer to the interaction between a
covalently bonded halogen atom in a molecule A-X (A = electron-withdrawing atom
or group; X = F, Cl, Br, I) and a nucleophilic site on Y (e.g a lone pair or anion)
[8–10, 14, 16–22]. As is the case for hydrogen bonds, halogen bond formation
normally leads to elongation of the A-X bond with a concomitant reduction in the
A-X vibrational stretching frequency, though cases of ‘blue-shifting’ halogen bonds
have also been reported in the literature [23, 24].
At first glance this interaction seems counterintuitive since halogen atoms are
electronegative species and as such are not expected to participate in non-covalent
bonding with electron-rich sites. Through the pioneering work of Brinck and others,
it has been shown that when bonded to other electronegative species (A), an electron-
deficient region develops on the surface of the halogen atom (adjacent to and collinear
with the A-X molecular axis), termed a σ-hole, which allows for such interactions to
occur [25–29].
A σ-hole on X is often characterized by computed molecular surface electrostatic
potential plots for A-X and when A is a strong electron-withdrawing group, the σ-
hole(s) on X acquires a positive electrostatic potential [25–29]. Several studies have
reported good correlations between the σ-hole strength (i.e. the magnitude of the
positive potential on X) and the stability of the resulting halogen-bonded complexes
A-X. . .Y [30, 31], highlighting the electrostatic nature of halogen bonds. Since the
polarizability of X increases as X gets larger (i.e. going from F to I), the σ-hole
strength increases (for fixed A and Y) in the same order, generally leading to stronger
halogen-bonded complexes [30, 31].
Halogen bonds are noted for their marked directionality (i.e. the halogen bond
angle is invariably linear or very close to linearity) which is a result of the localized
nature of σ-holes [8, 32, 33]. Another important feature of halogen bonding is the
noted Lewis acid—Lewis base duality of X which allows for the simultaneous inter-
action of X with the Lewis base Y and with an electrophile [32, 34, 35]. This feature
stems from the anisotropic charge distribution of X, in which regions along the A-X
bonding axis are depleted of electronic charge (σ-hole), while those regions roughly
perpendicular to the A-X bonding axis are electron-rich (lone pairs of X) [16, 36–38].
These and other interesting features of halogen bonds have stimulated the interest of
researchers seeking to exploit the properties of halogen bonding in fields as diverse
as biological modelling [39–42], crystal engineering and drug design [43–52], to
name a few.
12.1.1.3 Lithium Bonding
The lithium bond is similar to the hydrogen bond and likewise may be denoted
as A-Li. . .Y. The term ‘lithium bond’ is generally used to describe the interaction
of a chemically bonded Li atom in a molecule A-Li with an electron-rich site in
another molecule Y. One of the earliest theoretical studies of lithium bonding was
undertaken by Kollman and coworkers who investigated a series of lithium-bonded
dimers (HF. . . LiF, LiF. . . LiLi, LiF. . . LiF) and their hydrogen-bonded counterparts
(HF. . . HF, LiF. . . HH, LiF. . . HF) [53]. Later, Ault et al obtained matrix isolation
infrared spectra for A-Li. . .Y (A = Cl, Br; Y = N(CH3 )3 , NH3 , O(CH3 )2 , OH2 ) com-
plexes and provided the first experimental evidence of lithium halides participating
in lithium bonding [54].
In a detailed review of lithium bonding, Sannigrahi et al described the lithium
bond as a case where a lithium atom is di-coordinated [55]. For a fixed A, A-Li tends
to be more ionic than A-H, since the Li atom is more electropositive than the H
atom. Accordingly, the lithium bond (A-Li. . .Y) is stronger than the corresponding
hydrogen bond (A-H. . .Y) [55, 56]. The relatively high electropositive character
of the Li atom also accounts for the dominance of electrostatic forces in lithium
bonds [56, 57].
Intermolecular charge transfer is thought to play a significant role in hydrogen
bond formation, but is less significant for lithium bonding [56]. Vila and coworkers
used Bader’s Atoms in Molecules (AIM) theory [58] to perform topological analyses
for several lithium- and hydrogen-bonded complexes. The relative electron density
at bond critical points (ρbcp ), for a given set of bonds, usually correlates well with the
relative bond stability. Interestingly, it was found that the ρbcp for the lithium bonds
was half that for the hydrogen bonds, although the lithium bond is stronger. This
further emphasized the role of electrostatics in the formation of lithium bonds [57].
A recent study alluded to a difference in the mechanism for lithium and hydrogen
bond formation. It was found that while the positive charge on the H atom of A-
H. . .Y increases upon complex formation, with a concomitant decrease in H atomic
volume, the Li atom of A-Li. . .Y showed a general reduction in positive charge and
an increase in its volume [59]. It has also been reported that, unlike hydrogen bonds
where the A-H bond extends (and its vibrational frequency decreases) or contracts
(with its vibrational frequency increasing), the bond length changes and frequency
shifts for A-Li in A-Li. . .Y do not conform to these ‘rules’ [60–62].
12.1.1.4 Beryllium Bonding
Beryllium bonds are formed when a Be atom in a AA Be molecule (A may or may
not be the same atom as A) interacts with an nucleophilic site in another molecule Y;
i.e. AA Be. . .Y [63–65]. The Be atom, like H and Li, is also highly electropositive
and therefore the interaction energies of beryllium bonds tend to have a substantial
electrostatic component and are generally comparable in strength to very strong
hydrogen bonds.
However, due to the presence of vacant low-lying 2p orbitals on Be in AA Be,
considerable charge transfer into these orbitals may occur upon AA Be. . .Y complex
formation. Electronic charge may also be transferred into the A-Be/A -Be antibond-
ing orbital (σ∗ A-Be/A -Be ). The shift in charge densities greatly increase the covalent
character of beryllium bonds [63, 64] and may result in relatively high ρbcp for these
bonds compared to their hydrogen-bonded analogues [63].
Significant charge redistribution which occurs when beryllium bonds are formed
usually causes bending of the linear AA Be (A = A ) molecule, as well as elongation
of the A-Be/A -Be bond and large red or blue shifts in the stretching frequencies of
the A-Be/A -Be bond. It may also lead to substantial changes in the electron density
of the Lewis base to the extent that the acidity of the base increases [63, 65].
12.1.2 Interplay Between Non-Covalent Interactions
The study of the interplay between the hydrogen bond and other less well-known
interactions in molecular complexes is of widespread interest and the insight to
be gained from such investigations is strong motivation for contemporary research
on intermolecular forces, including the present contribution. Cooperativity is one
of the most interesting features of non-covalent interactions and the study of this
phenomenon may provide an understanding of cluster growth, aggregation and the
formation of condensed phases.
Generally, cooperativity is manifested as an increase in the binding strength of
a complex beyond the sum of its individual interactions taken a pair at a time. For
example, a trimer system consisting of molecular subunits AX. . .Y. . . AX is said to
exhibit positive cooperativity when the interaction energy of the trimer is greater than
the sum of the separate AX. . .Y andY. . . AX pair interaction energies. Conversely, a
negative cooperative or a diminutive effect is associated with a decrease in the trimer
interaction energy relative to sum of the pair interaction energies. Cooperativity in
a trimer can be assessed from the relative contribution of the three-body pairwise
non-additive energy to the total interaction energy.
Cooperativity results mainly from the mutual polarization of the constituent
molecules of the cluster or complex—the energy of each pair of molecules in a
trimer, for example, is modified by the presence of the third interacting molecule,
especially if the molecules are both highly polar and polarizable. Hydrogen-bonded
complexes commonly show positive cooperative effects, but negative cooperative ef-
fects are observed if one of the molecules in a trimer system acts as a double proton
acceptor or a double proton donor [4].
Numerous studies have been devoted to the investigation of cooperative effects
of hydrogen bonding and other non-covalent interactions. In particular, great em-
phasis has been placed on the interplay between multiple hydrogen bonds [66–77],
hydrogen and dihydrogen bonds [78–80] and, most recently, hydrogen and halogen
bonds [34, 81–91]. Parra et al investigated the cooperative effects of bifurcated hy-
drogen bonding. Interestingly, the hydrogen bond strength in the decamer was about
160 % higher than in the dimer, despite the hydrogen bonds being 3-center hydrogen
bonds as opposed to being typical 2-center hydrogen bonds [70]. Arauju et al showed
that the cooperative effect in hetero hydrogen-bonded chains was greater than that
for homo chains for a series of (HCN)n -HF and (HCN)n (n = 1, 2, 3) clusters [71].
Additionally, as the chains become longer there was a corresponding increase in the
cooperative effects observed [71].
Cooperative effects in hydrogen-bonded systems may play a significant role in
the self-assembly of certain complexes. In 2005, Planas and co-workers demon-
strated this property for mercaptane-metallacarborane complexes, which contained
both hydrogen and dihydrogen intermolecular bonds [78]. It appears that the self-
assembly of these complexes is driven by the unusual cooperative effects between
the aforementioned non-covalent bonds [78]. Li et al further studied the coopera-
tivity between hydrogen and dihydrogen bonds for a set of LiH-(HCN)n (n = 2, 3)
complexes and found that the non-additive energy in complexes with various types of
hydrogen bonds was greater than those with the same type of hydrogen bonds [79].
Furthermore, they reported that charge transfer and polarization forces were major
contributors to the cooperativity between these non-covalent bonds [79].
With regards to halogen bonding, it has been shown that halogen and hydrogen
bonds can exhibit strong cooperative or diminutive effects on each other in molecular
clusters. Zhou et al explored the cooperativity between these bonds for a series
of ring-shaped complexes [A-(HF)X. . . NH3 (X = Cl, Br and A = F, Cl, Br)] that
contained two hydrogen bonds and one halogen bond [34]. A larger non-additive
energy was reported for the AX = Br2 cluster compared to the AX = Cl2 cluster.
This observation again showed the significant dependence of cooperative effects on
the relative polarizabilities of the interacting species [34]. Grabowski et al conducted
an investigation of Cl− . . . HCCH. . . HF, Cl− . . . ClCCH. . . HF, F− . . . ClCCH. . . HF
clusters and partitioned the total interaction energy for each cluster into various
energy contributions. For all of the complexes which showed positive cooperative
effects (i.e. Enon-add < 0), the most important attractive energy term was found to be
the polarization term [88]. Additionally, it was reported that charge transfer forces
were the most significant contributor to the formation of the hydrogen bonds [88].
Jing et al described a diminutive relationship between hydrogen and halogen
bonds in model systems of HNC and HOBr when the HOBr molecule simultaneously
acted as the hydrogen and halogen bond donor [83]. Conversely, in systems where
HNC acted as both the proton donor and the halogen bond acceptor, the cooperative
energy exceeded 20 % of the interaction energy [83]. Solimannejad and Malekani
observed similar results for a series of complexes of HCCX (X = Cl, Br) with HCN
and HNC [87].
More often than not, when both hydrogen and halogen bonds exist in molecular
systems, one bond is strengthened at the expense of the other; usually the weaker bond
experiences a positive cooperative effect. Zhao et al demonstrated this relationship
for trimeric clusters of carbon tetrabromide, a halide, and a solvent molecule [91].
For clusters containing the smaller halide ions, the halogen bonds were elongated in
the trimer relative to the dimer, while the hydrogen bond contracted [91]. Alkorta et al
found that the hydrogen-bonded complexes were stronger for the hypohalous acids
containing the smaller halogen atoms. The interaction energies for the dimers with the
higher halogens, I and At, were similar or larger than that of their hydrogen-bonded
counterparts [89].
Cooperative effects between lithium bonds and other non-covalent interactions
have received increased attention in recent years. Studies on the interplay be-
tween lithium bonds and hydrogen [92–96], dihydrogen [95], halogen [93, 97, 98],
chalcogen [99], pnicogen [100] and cation- π [101] bonds, as well as in lithium
clusters [102–104] have been documented. Solimannejad showed that for a series
of F3YLi. . . NCH. . . HMH and F3YLi. . . HMH. . . HCN trimers (Y = C, Si; M = Be,
Mg), involving both lithium and dihydrogen bonds, the overall interaction was co-
operative or diminutive, depending on whether HMH was located at the end or in
the center of the trimer chains [95].
A similar study on the interplay between lithium and halogen bonds also found
that the ordering of monomers had a notable impact on the cooperativity observed
in the resulting trimers [98]. McDowell and Yarde, showed that for a series of
model lithium-, halogen- and hydrogen-bonded trimers, the lithium bond was most
significant in producing positive cooperative effects. They also reported that the
non-additive energies were higher in the cyclic clusters than in their linear counter-
parts [93]. A study involving linear chains consisting of LiCN subunits bound by
multiple lithium bonds revealed that electrostatic and polarization forces were the
main contributors to the stability of the chains [102].
The mutual influence between beryllium bonds and other non-covalent bonds has
also been investigated in recent years [105–107], though not as extensively as the
preceding interactions. Mó et al carried out investigations on the effect of beryllium
bonds on inter- and intra-hydrogen bonds [107]. The beryllium bonds and the inter-
hydrogen bonds showed positive cooperative effects in the imidazole-BeX2 (X = H,
F) trimers. While in the malonaldehyde-BeX2 trimers, the cooperative effects of the
beryllium bonds and the intra-hydrogen bonds were heavily dependent on whether
the Be atom was attached to the hydrogen bond donor or acceptor; the former being
cooperative and the latter diminutive [107]. Another study involving beryllium bonds
and multiple hydrogen bonds revealed that it was energetically more favorable to
attach a BeX2 molecule to H2 O dimers or trimers than to solvate BeX2 with the
individual H2 O molecules [106].
12.2 Cooperative Effects in Unusual Thiirane Complexes
Our theoretical study of cooperativity is restricted to its manifestation in molecular

trimers, with dimeric complexes of the thiirane molecule (CH2 )2 S being the com-
mon motif throughout. A recent high-level computational study of the thiirane. . . ClF
complex found an unusually strong interaction in this dimer accompanied by sig-
nificant charge transfer from thiirane to the ClF molecule [108]. The author of this
study suggested that this type of dimer be classified as a Mulliken “inner complex”
of the form [A-H]+ . . . B, rather than that of a more weakly-bound “outer complex”
of the usual A-H. . .Y form, where the charge transfer is smaller and the interaction
more electrostatic in nature [108].
Mulliken and Person’s work on molecular complexes [109] established this
way of classifying donor-acceptor complexes, with specific reference to hydrogen-
bonded complexes. The considerable interest in halogen bonding [18, 21] has also
prompted some suggestion of classifying halogen-bonded complexes as either inner
complexes [A-X]+ . . .Y− or outer complexes A-X. . .Y, where the former involve
more substantial charge transfer between the interacting molecules than in the
latter [108, 110].
The thiirane. . . ClF dimer has been investigated experimentally by gas-phase
rotational spectroscopy [111] and its characterization through the explicitly corre-
lated coupled cluster CCSD(T)-F12b/CBS method, established a relatively large
value of −55.4 kJ mol−1 for this strong intermolecular interaction, compared
with an interaction energy of −32.8 kJ mol−1 for the hydrogen-bonded analogue
thiirane. . . HCl [108]. Furthermore, it was found that significantly more charge was
transferred (by a factor of about four), mainly from the sulfur lone pair to the Cl-F
antibonding orbital, than the corresponding charge transfer in thiirane. . . HCl.
The natural partial atomic (NPA) charges on the sulfur atom of thiirane. . . ClF,
obtained from a natural bond orbital (NBO) analysis [112], found that remark-
ably the sulfur atom was positively charged, compared with the oxygen atom of
oxirane. . . ClF, which has a negative partial charge. This finding, along with the
large interaction energy and significant charge transfer, prompted Hill to suggest
that thiirane. . . ClF be classified as the first “inner-type” halogen-bonded complex,
in Mulliken’s terminology [108].
Decomposition of the interaction energy of thiirane. . . ClF using symmetry-
adapted perturbation theory (SAPT) [113] showed that the four main components of
the interaction energy: the electrostatic (Eelec ), exchange repulsion (Eexch ), induc-
tion (Eind ), and dispersion (Edisp ) energies, were all substantially larger than in the
hydrogen-bonded thiirane. . . HCl and oxirane. . . HCl complexes, and in the halogen-
bonded oxirane. . . HCl complex [108]. These SAPT results provide further support
for the notion that thiirane. . . ClF should be considered as an “inner-type” complex.
12.2.1 Thiirane. . . XF/BeH2 Dimers
We use Hill’s computational study of thiirane. . . ClF as the starting point for this
study of the cooperative effect of non-covalent interactions involving the familiar
hydrogen bond and the less-familiar, but no less important, halogen bond, as well
as two other unusual interactions—the lithium bond and the beryllium bond. Our
approach is to optimize the thiirane. . . XF/BeH2 (XF = HF, ClF, BrF, LiF) series of
dimers using density functional theory (DFT) with the widely-used B3LYP func-
tional as our computational method of choice. The large 6-311++G(3df,3pd) basis
set, which is of triple-zeta quality and contains sets of diffuse functions and higher
angular momentum functions, should be capable of producing reliable results with
sufficiently high accuracy. The available experimental and computational data for
thiirane. . . ClF is useful in gauging the accuracy of the present work.
The set of XF/BeH2 molecules can attach to the S atom of thiirane via hydrogen,
two types of halogen bonds (Cl, Br), lithium or beryllium bonds and this allows for
an assessment of the relative strengths of these various intermolecular interactions.
Computed parameters of the optimized dimer, which give useful insight into the
nature of the operative non-covalent forces, include the interaction energy, the S. . . X
intermolecular separation, the S. . . X-F bond angle, the change in the X-F bond length
and the complexation-induced shift of the X-F harmonic vibrational stretching mode.
The optimization and subsequent computation of these properties at the B3LYP/6-
311++G(3df,3pd) level of theory for the thiirane. . . XF series was performed with
the Gaussian 03 suite of programs [114].
The interplay between hydrogen and halogen bonding in sulfur-containing sys-
tems has potential relevance for thyroid chemistry [115, 116] and as such a study of
the cooperative effects between these and other non-covalent interactions in model
systems may provide a better understanding of the structure and properties of re-
lated biomolecular complexes, especially those that may govern important biological
activities in living systems.
Table 12.1 shows the computed properties for the thiirane. . . XF dimer systems.
Figures 12.1, 12.2 and 12.3 show the optimized geometries of the thiiriane. . . ClF,
thiirane. . . LiF and thiirane. . . BeH2 complexes and span the range of molecular
Table 12.1 Interaction energies (E/kJ mol−1 ), bonding distances (R/Å), bondlength changes
(r/Å), frequency shifts (ω/cm−1 ) and bond angles (</◦ ) for (CH2 )2 S. . . XF/BeH2 (X = H, Cl,
Br, Li) complexes at the B3LYP/6-311++G(3df,3pd) level of theory
Complex E R(S. . . X/Be) r(X-F)/ ω(X-F)/ < S. . . X-F
r(Be-H) ω(Be-H)
(CH2 )2 S. . . HF − 32.5 2.150 0.024 − 539.8 166.1
(CH2 )2 S. . . ClF − 56.2 2.459 0.127 − 258.3 177.5
(CH2 )2 S. . . BrF − 64.6 2.582 0.102 − 171.6 177.9
(CH2 )2 S. . . LiF − 69.4 2.440 0.038 − 88.3 111.8
(CH2 )2 S. . . BeH2 − 53.6 2.200 0.021 − 185.5asymm –
0.033 − 122.0symm
Fig. 12.1 Optimized

geometry for (CH2 )2 S. . . ClF
Fig. 12.2 Optimized

geometry for (CH2 )2 S. . . LiF
structures that these dimers adopt. The S atom binds directly to the X atom (or Be
atom) and sits at the apex of a 3-atom C-S-C cycle comprising the thiirane ring. The
HF, ClF and BrF molecules adopt a similar orientation in the dimer structure, with
the S. . . ClF/BrF orientation close to linearity (about 178◦ ). The HF orientation, on
the other hand, deviates by about 14◦ from linearity, consistent with its weaker bond
strength relative to the halogen analogues. The LiF molecule is almost parallel to
the C-S-C ring of thiirane (see Fig. 12.2) and the F atom appears to benefit from an
attractive interaction with the H atoms of thiirane. The BeH2 molecule is oriented with
its hydridic H atoms bent away from the thiirane molecule (as shown in Fig. 12.3),
which indicates a substantial perturbation of its structure due to the interaction with
thiirane.
Fig. 12.3 Optimized

geometry for
(CH2 )2 S. . . BeH2
The interaction energies for the thiirane. . . XF/BeH2 dimers are in the order LiF >
BrF > ClF > BeH2 > HF. The thiirane binds most strongly with LiF, presumably
through an S. . . Li contact, as well as interaction of the F atom with the nearby
protons of thiirane. This is inferred by the almost parallel orientation of LiF relative
to thiirane. By contrast, the other XF molecules appear to maximize their interaction
with the S atom by polarization of the thiirane molecule—this molecule has large
dipole polarizabilities and the electric field arising from the XF molecules facilitates
the shift of electron density from the sulfur lone pair towards the antibonding σ*
orbital of XF. The bonding in the BeH2 complex is different from the XF complexes
since the Be atom is oriented in such a way that the vacant 2p orbitals can overlap
with the S lone pair thereby facilitating electron density transfer between thiirane
and BeH2 . It is noteworthy that the hydrogen bond has the weakest non-covalent
interaction with thiirane.
The order of the intermolecular S. . . X/Be separation in the complexes is HF <
BeH2 < LiF < ClF < BrF and does not correlate with the trend for interaction ener-
gies above. It seems that the intermolecular separation is dependent on the electron
repulsion between S and X/Be, so that the H atom has the closest approach, while the
large Br atom has the furthest, with the other atoms falling in between these limits.
The order of the intermolecular separation roughly correlates with the number of
electrons on the X atom. However, for the small Be atom, the two valence electrons
are effectively localized on the terminal H atoms of BeH2 allowing the tightly-held 1s
core electrons to approach the S atom quite closely. It is noteworthy that the trend for
the S. . . X/Be separation follows the order of the H/Be/Li ionic radius (H < Be < Li)
and the Cl/Br atomic radius (Cl < Br); i.e. the respective values of the ionic or atomic
radius for H+ , Be2+ , Li+ , Cl, Br are 0.00066, 0.34, 0.78, 0.99, 1.14 Å [117].
For all five dimers, there is an increase in the X-F (or Be-H) bond length and a
corresponding red shift of the X-F (H-Be-H) stretching frequency. This elongation
can be viewed as a consequence of the electrostatic attraction between the positive
potential on X and the lone pair on S, as well as a consequence of the charge transfer
from the S lone pair orbitals into the antibonding σ∗ orbitals of X-F or Be-H. The
frequency shift of the X-F vibration is dependent not only on the magnitude of the
electrostatic forces acting on the covalent X-F bond or the extent of charge transfer
from thiirane, but also on the strength (or stretching force constant) of the X-F bond.
The red shift and bond extension associated with the X-F molecule are usually
correlated in a qualitative sense (as is the unconventional blue-shifting behaviour,
which is usually correlated with a bond contraction in hydrogen-bonded com-
plexes). However, a more quantitative (and proportionate) relationship between the
frequency shift and the bond length change would depend on the relative sign and
magnitude of the force exerted on the X-F bond by the electric field originating
from the thiirane molecule and the “stiffness” of the X-F bond (as measured by its
stretching force constant).
For example, the HF stretching frequency suffers the largest red shift
(of −540 cm−1 ) even though the thiirane. . . HF complex is the most weakly-bound
complex and has the smallest bond extension (0.024 Å). This large red shift is ob-
viously due to the small size of the H atom, compared with the other X atoms. The
largest bond extension (0.127 Å) is observed in thiirane. . . ClF, which also has the
second highest computed red shift (−258 cm−1 ). These relatively large values appear
to be consistent with the characteristics of a strongly-bound “inner type” complex,
as noted by Hill in his recent study [108].
It is interesting to note that the second largest bond extension and red shift were
obtained for thiirane. . . BrF, even though the complex has a larger interaction energy
and BrF produces a larger σ-hole (and therefore, stronger halogen bond) than ClF.
Do these results indicate that thiirane. . . BrF should also be considered as an inner
complex? Comparison of the changes in electron density distribution, before and
after complexation, in these two different halogen-bonded dimers should provide
an answer to this intriguing question. These changes were investigated through the
use of the electron density partitioning schemes implemented by the NBO in the
Gaussian03 suite of programs and AIM analyses using the AIMAll software [118]—
to be discussed later.
In the next sub-sections we shall see how the properties of the thiirane. . . XF/BeH2
are modified when a third molecule interacts with the S atom or when it interacts
with the negatively charged F atom of XF or the hydridic H atoms of BeH2 . The
cooperative effects of the different non-covalent interactions can be assessed from
these computational results.
Table 12.2 Interaction

Complex Edimer E R(S. . . X)
energies (E/kJ mol−1 ) and
bonding distances (R/Å) for (CH2)2S. . . HF 2.150
FX /H2 Be. . . (CH2 )2 S. . . XF FH. . . (CH2 )2 S. . . HF − 65.0 − 55.9 2.203
(X = H, Cl, Br, Li; X = H,
Cl, Br) complexes at the FCl. . . (CH2 )2 S. . . HF − 88.7 − 70.2 2.278
B3LYP/6-311++G(3df,3pd)
FBr. . . (CH2 )2 S. . . HF − 97.1 − 78.0 2.296
level. The Edimer values
represent the sum of the FLi. . . (CH2 )2 S. . . HF − 101.9 − 92.3 2.199
FX /H2 Be. . . (CH2 )2 S and H2 Be. . . (CH2 )2 S. . . HF − 86.1 − 69.7 2.281
(CH2 )2 S. . . XF dimer
(CH2)2S. . . ClF 2.459
interaction energies. The
(CH2 )2 S. . . XF dimer results FH. . . (CH2 )2 S. . . ClF − 88.7 − 70.2 2.529
are shown for the sake of FCl. . . (CH2 )2 S. . . ClF − 112.4 − 81.1 2.593
comparison
FBr. . . (CH2 )2 S. . . ClF − 120.8 − 87.5 2.612
FLi. . . (CH2 )2 S. . . ClF − 125.6 − 106.5 2.517
H2 Be. . . (CH2 )2 S. . . ClF − 109.8 − 77.9 2.620
(CH2)2S. . . BrF 2.582
FH. . . (CH2 )2 S. . . BrF − 97.1 − 78.0 2.627
FCl. . . (CH2 )2 S. . . BrF − 120.8 − 87.5 2.672
FBr. . . (CH2 )2 S. . . BrF − 129.2 − 93.7 2.681
FLi. . . (CH2 )2 S. . . BrF − 134.0 − 115.0 2.615
H2 Be. . . (CH2 )2 S. . . BrF − 118.2 − 84.5 2.693
12.2.2 FX /H2 Be. . . Thiirane. . . XF/BeH2 Trimers
Table 12.2 shows selected properties for the FX /H2 Be. . . thiirane. . . XF (X = H, Cl,
Br, Li; X = H, Cl, Br) trimers which have the typical structure shown in Fig. 12.4.
In these complexes the S atom acts as a double acceptor and all of the complexes
show a negative cooperative or diminutive effect, as reflected by their reduced inter-
action energies, relative to the sum of the constituent dimer energies. This diminutive
effect is also observed for hydrogen-bonded complexes involving a double proton
acceptor [4].
The other computed parameters reflect this diminutive effect; i.e. all inter-
molecular S. . . X distances are increased, while all X-F bond extensions are
smaller in the trimers, compared with the corresponding dimer values. The order
of the diminutive effect due to the FX molecule as reflected by R(S. . . X/Be) is
FBr > H2 Be > FCl > FH > FLi.
The FX molecule attaches to the free S lone pair of the thiirane. . . XF dimer to
form the trimer and in so doing polarizes the S atom by pulling electron density
towards itself, thereby reducing the lone pair density, both available for bonding to
the X atom of XF, as well as reducing the charge transferred to XF, and vice versa.
Consequently, the net binding weakens, the S. . . X separation increases, the XF bond
Fig. 12.4 Optimized geometry for FCl. . . (CH2 )2 Si. . . ClF
extension decreases and the XF red shift decreases, relative to the thiirane. . . XF
dimer (note the bond extensions and frequency shifts are not shown in Table 12.2).
Despite the negative cooperative effect, all complexes are strongly-bound
(−E > 55 kJ/mol) and all XF stretching frequencies are strongly red-shifted. The
diminutive effect of FX on E, R(S. . . X) and r(X-F) is always larger for FCl, FBr
and BeH2 . This is because both halogen and beryllium bonding will shift electron
density away from the central S atom, thereby diminishing the S. . . X interaction.
This is manifested as an increase in the electron density transferred from S to the
antibonding σ∗ orbital of the FX molecule as a result of complex formation—see the
NBO partial atomic charge results, to be discussed later. The least diminutive effect
occurs for LiF and is probably because the repulsion between the Li 1s core electrons
and the S lone pairs hinders charge transfer from S to the LiF molecule.
Table 12.3 shows selected properties for FX /H2 Be. . . thiirane. . . LiF (Fig. 12.5)
as well as for FX /H2 Be. . . thiirane. . . BeH2 (Fig. 12.6). Addition of FX /H2 Be to
form FX /H2 Be. . . thiirane. . . LiF causes a net reduction of E, increase in R(S. . . Li)
and diminished LiF bond extension (values not shown in Table 12.3), as is the case
for the other XF trimers, with FCl/FBr producing the largest diminutive effects.
A diminutive effect is also evident for FX /H2 Be. . . thiirane. . . BeH2 ; the net E
is smaller than the sum of the interaction energies of the constituent dyads, the
S. . . Be separation is larger and the r/ω values (not shown in Table 12.3) are
smaller than the corresponding dimer values, with the additional halogen-bonded
and beryllium-bonded interactions yielding the largest diminutions. These results
suggest that the displacement of the S lone pair electron density into the vacant p
orbitals of the adjacent Be atom is reduced due to competition with the FX molecule.
In percentage terms, the largest diminutive effects in the FX /H2 Be. . .
thiirane. . . BeH2 trimers are for those involving ClF, BrF and BeH2 ; this means
that the halogen and beryllium bonds have the most significant diminutive effect.
However, the most strongly-bound dimers and trimers are those which involve LiF,
except FLi. . . thiirane. . . LiF, where the competition for the S lone pairs renders the
whole complex unstable since no local minima were located. Perhaps these results
indicate that the electron density transfer occurring in the ClF, BrF and BeH2 trimers
is the most significant cause of the negative cooperativity, whereas in the complexes
Table 12.3 Interaction energies (E/kJ mol−1 ) and bonding distances (R/Å) for FX /H2 Be. . .
(CH2)2 S. . . LiF/BeH2 (X = H, Cl, Br, Li) complexes at the B3LYP/6-311++G(3df,3pd) level.
The Edimer values represent the sum of the FX /H2 Be. . . (CH2 )2 S and (CH2 )2 S. . . LiF/BeH2
dimer interaction energies. The (CH2 )2 S. . . XF dimer results are shown for the sake of comparison

Complex Edimer E R(S. . . X/Be)
(CH2)2S. . . LiF 2.440
FH. . . (CH2 )2 S. . . LiF − 101.9 − 92.3 2.478
FCl. . . (CH2 )2 S. . . LiF − 125.6 − 106.5 2.539
FBr. . . (CH2 )2 S. . . LiF − 134.0 − 115.0 2.537
FLi. . . (CH2 )2 S. . . LiF − 138.8 2 imag freq
H2 Be. . . (CH2 )2 S. . . LiF − 123.0 − 108.3 2.516
(CH2)2S. . . BeH2 2.200
FH. . . (CH2 )2 S. . . BeH2 − 86.1 − 69.7 2.231
FCl. . . (CH2 )2 S. . . BeH2 − 109.8 − 77.9 2.282
FBr. . . (CH2 )2 S. . . BeH2 − 118.2 − 84.5 2.295
FLi. . . (CH2 )2 S. . . BeH2 − 123.0 − 108.3 2.206
H2 Be. . . (CH2 )2 S. . . BeH2 − 107.2 − 75.6 2.285
Fig. 12.5 Optimized geometry for FCl. . . (CH2 )2 Si. . . LiF
containing LiF, this charge transfer is reduced by the Li 1s core electrons, as well
as its orientation relative to thiirane.
Table 12.4 shows the pairwise additive and non-additive energy contributions to
the interaction energy. The non-additiveenergy (Enon-add ) is a measure of coopera-
tivity and is defined as Enon-add = E − Eij , where E is the interaction energy
of the trimer and Eij is the pair interaction energy for a dimer subunit ij, computed
at the geometry that this subunit adopts in the optimized trimer.
As expected, the pair interaction of the terminal FX and XF molecules is sig-
nificantly smaller than the other pair interactions. The diminutive effect of FX is
Fig. 12.6 Optimized

geometry for
FCl. . . (CH2 )2 Si. . . BeH2
generally evident in the FX . . . thiirane. . . XF complexes, except those involving

either LiF or BeH2 . Recall that a diminutive effect was previously noted for all the
trimers (relative to their respective thiirane. . . XF/BeH2 dimers). The results in Ta-
ble 12.4 does not contradict this observation. However, in Table 12.4 the interaction
between the c. . . b pair (which is not accounted for in Edimer ) is incorporated in
the calculation of the non-additive energy. The c. . . b pair interaction is repulsive in
all trimers and so modifies their estimated non-additive energies. Additionally, the
trimers involving LiF or BeH2 have a different configuration than the hydrogen- and

Table 12.4 Sum of the pairwise additive ( Eij ) and percentage non-additive contributions
(%Enon-add ) of E for [FX /H2 Be] . . . [(CH2 )2 S]a . . . [XF/BeH2 ]b (X , X = H, Cl, Br, Li) complexes
c
at the B3LYP/6-311++G(3df,3pd) level. All values are measured in kJ/mol

Complexes E Eij %Enon-add
c. . . . . . . . . ..a. . . . . . . . . ..b
FH. . . (CH2 )2 S. . . HF − 55.9 − 58.9 − 5.4
FCl. . . (CH2 )2 S. . . HF − 70.2 − 72.0 − 2.6
FBr. . . (CH2 )2 S. . . HF − 78.0 − 83.1 − 6.5
FLi. . . (CH2 )2 S. . . HF − 92.3 − 91.1 1.3
H2 Be. . . (CH2 )2 S. . . HF − 69.7 − 53.7 2.3
FCl. . . (CH2 )2 S. . . ClF − 81.1 − 88.9 − 9.6
FBr. . . (CH2 )2 S. . . ClF − 87.5 − 101.2 − 15.7
FLi. . . (CH2 )2 S. . . ClF − 106.5 − 100.3 5.8
H2 Be. . . (CH2 )2 S. . . ClF − 77.9 − 73.3 5.9
FBr. . . (CH2 )2 S. . . BrF − 93.7 − 112.0 − 19.5
FLi. . . (CH2 )2 S. . . BrF − 115.0 − 111.0 3.5
H2 Be. . . (CH2 )2 S. . . BrF − 84.5 − 85.5 − 1.2
H2 Be. . . (CH2 )2 S. . . LiF − 108.3 − 83.2 23.2
H2 Be. . . (CH2 )2 S. . . BeH2 − 75.6 − 55.5 26.6
halogen-bonded counterparts and this may explain why there is a positive coopera-
tive effect in these complexes—this configuration allows the negative F (of LiF) or
H of (BeH2 ) to interact favourably with the protons of the CH2 subunits of thiirane.
12.2.2.1 NBO Analysis of Thiirane. . . XF/BeH2 and

FX /H2 Be. . . Thiirane. . . XF/BeH2
Table 12.5 shows that the positive charge on the S atom of the isolated thiirane
molecule increases on complexation with HF, ClF and BrF and correlates with trans-
fer of electron density from one of the S lone pairs into the antibonding σ∗ orbital
of XF, with the largest transfer to the ClF in accordance with Hill’s findings [108].
Interestingly the EDT for thiirane. . . BrF is comparable to that of ClF. Therefore it
is possible that this dimer is an inner complex as well.
The decrease in the positive charge on the S atom in the LiF dimer complex is
due to the substantial polarization of the thiirane molecule by the large LiF dipole
which is oriented in such a way that it forces electron density towards the S atom
(its dipole is almost parallel to the C-S-C plane and in the opposite direction to the
thiirane dipole). For the thiirane. . . BeH2 dimer, there is evidence of electron transfer
from the BeH2 bonding orbitals into the intermolecular S. . . Be region.
Going from dimer to trimer, there is a decrease in charge transfer from thiirane
to XF due to competition between FX and XF for the S lone pairs. This supports
the observed diminutive effect for the FX /H2 Be. . . thiirane. . . XF/BeH2 complexes.
Though diminished, the EDT for XF = Cl/Br is still considerably larger, relative
to the trimers of the other XF molecules, and so these complexes may exhibit the
characteristics of a Mulliken’s inner complex [108].
12.2.3 Thiirane. . . XF/BeH2. . . X F/BeH2 Trimers
Table 12.6 shows selected properties for the thiirane. . . XF/BeH2 . . . X F/BeH2 (X =
H, Cl, Br, Li; X = H, Cl, Br, Li) trimers which span a range of various structures,
some of which are shown in Figs. 12.7–12.10. In these complexes, the interaction
of the S atom of thiirane with the X atom of XF (or Be atom of BeH2 ) is modified
by the attachment of the X F/BeH2 molecule to the negative F atom of X F (or the
negative H atom of BeH2 ).
Generally, the interaction energy E for each set of trimers (with fixed XF) seems
to depend on the polarizing strength of the terminal X F molecule. Hence, the lithium
bond attached to the F of XF produces the most strongly-bound complexes, followed
by the beryllium bond, then the hydrogen bond, with the halogen bonds either pro-
ducing the most-weakly bound complexes or yielding no optimized structures at all.
For the latter halogen-bonded complexes, the repulsion between the F lone pairs (of
XF) and the lone pairs of the halogen atom X , is likely to be the main destabilizing
factor, especially for the Br atom. We shall now consider each set of trimers in turn.
374
Table 12.5 Changes in NPA partial atomic charges (q in e) w.r.t isolated monomers for (CH2 )2 S. . . XF/BeH2 , FX /H2 Be. . . (CH2 )2 S. . . XF/BeH2
(X , X = H, Cl, Br, Li) complexes and electron density transfer (EDT in e) determined using the B3LYP/6-311++G(3df,3pd) procedure. Partial
atomic charges for monomers are in parentheses
Complex q (S) EDT (CH2 )2 S → XF Complex q (S) EDT (CH2 )2 S → XF
(CH2 )2 S. . . HF 0.003 (0.085) 0.070 FH. . . (CH2 )2 S. . . BrF 0.130 0.246
(CH2 )2 S. . . ClF 0.246 0.345 FCl. . . (CH2 )2 S. . . BrF 0.248 0.207
(CH2 )2 S. . . BrF 0.200 0.304 FBr. . . (CH2 )2 S. . . BrF 0.229 0.193
(CH2 )2 S. . . LiF − 0.108 0.037 FLi. . . (CH2 )2 S. . . BrF 0.053 0.256
(CH2 )2 S. . . BeH2 − 0.061 – H2 Be. . . (CH2 )2 S. . . BrF 0.042 0.195
FH. . . (CH2 )2 S. . . HF − 0.026 0.048 FH. . . (CH2 )2 S. . . LiF − 0.130 –
FCl. . . (CH2 )2 S. . . HF 0.168 0.033 FCl. . . (CH2 )2 S. . . LiF 0.083 –
FBr. . . (CH2 )2 S. . . HF 0.130 0.029 FBr. . . (CH2 )2 S. . . LiF 0.053 –
FLi. . . (CH2 )2 S. . . HF −0.130 0.047
H2 Be. . . (CH2 )2 S. . . HF − 0.098 0.029 H2 Be. . . (CH2 )2 S. . . LiF − 0.196 –
FH. . . (CH2 )2 S. . . ClF 0.168 0.267 FH. . . (CH2 )2 S. . . BeH2 − 0.098 0.065
FCl. . . (CH2 )2 S. . . ClF 0.270 0.209 FCl. . . (CH2 )2 S. . . BeH2 0.060 0.049
FBr. . . (CH2 )2 S. . . ClF 0.248 0.192 FBr. . . (CH2 )2 S. . . BeH2 0.042 0.044
FLi. . . (CH2 )2 S. . . ClF 0.083 0.276 FLi. . . (CH2 )2 S. . . BeH2 − 0.196 0.063
H2 Be. . . (CH2 )2 S. . . ClF 0.060 0.198 H2 Be. . . (CH2 )2 S. . . BeH2 − 0.148 0.047
S. A. C. McDowell and J. A. Joseph
Table 12.6 Interaction Complex E R(S. . . X/Be)

energies (E/kJ mol−1 ) and
bonding distances (R/Å) for (CH2 )2 S. . . HF. . . HF − 72.0 2.033
CH2 )2 S. . . XF/BeH2 . . . X F/BeH2 (CH2 )2 S. . . HF. . . ClF − 52.0 2.076
(X, X = H, Cl, Br, Li)
(CH2 )2 S. . . HF. . . BrF − 58.2 2.053
complexes at the
B3LYP/6-311++G(3df,3pd) (CH2 )2 S. . . HF. . . LiF − 136.7 1.918
level
(CH2 )2 S. . . HF. . . BeH2 − 97.8 1.857
(CH2 )2 S. . . ClF. . . HF − 104.8 2.260
(CH2 )2 S. . . ClF. . . ClF − 85.1 2.278
(CH2 )2 S. . . ClF. . . LiF − 213.1 2.106
(CH2 )2 S. . . ClF. . . BeH2 − 192.9 2.072
(CH2 )2 S. . . BrF. . . HF − 110.8 2.455
(CH2 )2 S. . . BrF. . . ClF − 91.9 2.470
(CH2 )2 S. . . BrF. . . LiF − 199.2 2.325
(CH2 )2 S. . . BrF. . . BeH2 − 165.0 2.325
(CH2 )2 S. . . LiF. . . HF − 169.1 2.448
(CH2 )2 S. . . LiF. . . BrF − 151.3 2.432
(CH2 )2 S. . . LiF. . . LiF − 302.7 2.248
(CH2 )2 S. . . LiF. . . BeH2 − 273.6 2.453
(CH2 )2 S. . . BeH2 . . . LiF − 152.0 2.122
(CH2 )2 S. . . BeH2 . . . BeH2 − 206.5 2.173
12.2.3.1 Thiirane. . . HF. . . X F/BeH2
1. All of these trimers are stable, with E increasing as X F = ClF < BrF < HF <
BeH2 < LiF, consistent with the order of decreasing R(S. . . H), which is ClF >
BrF > HF > LiF > BeH2 , except that the last two molecules are switched around.
The Be atom of BeH2 makes the closest contact with the F atom of HF—in
this complex, the interatomic electron repulsion is minimized since the 1s core
electrons of Be are tightly held and its valence electrons are displaced towards
the adjacent hydridic H atoms, which are bent away from the S atom. The F. . . X
separation (not shown in Table 12.6) in these trimers increases accordingly with
the interatomic repulsion in the order X = H < Li < Cl < Br.
2. The H-F bond extensions and vibrational red shifts are strongly correlated and
generally scale with the binding strength.
3. The cooperativity between the S. . . H and the F. . . X non-covalent interactions is
evident from the large non-additive energies displayed in Table 12.7, which are
estimated to be more than 25 % of E for those complexes. This (positive) coop-
erativity is also apparent in the shorter S. . . H distances for thiirane. . . HF. . . X F
shown in Table 12.6 compared with thiirane. . . HF (see Table 12.1).
4. For the thiirane. . . HF. . . X F series, the cooperative effect of the S. . . H hydrogen
bond and the non-covalent F. . . X interactions is significant, with the lithium and
Fig. 12.7 Optimized

geometries for a
(CH2 )2 S. . . HF. . . LiF and b
(CH2 )2 S. . . HF. . . BeH2
beryllium bonds having the largest, and the halogen bonds, the weakest effects.
The optimized structures for thiirane. . . HF. . . LiF and thiirane. . . HF. . . BeH2
(Fig. 12.7) show short contacts between the F of LiF (or one of the H atoms
of BeH2 ) and one of the nearby protons of thiirane, suggesting a cyclization
involving the atoms of the three molecules constituting each trimer. This would
suggest a substantial association involving all three molecules and is consistent
with the large non-additive energies shown in Table 12.7.
12.2.3.2 Thiirane. . . ClF. . . X F/BeH2
1. A similar pattern to that observed for thiirane. . . HF. . . X F is evident. The lithium
and beryllium bonds produce more strongly-bound trimers (and shorter S. . . Cl
contacts) than the halogen bonds, with the hydrogen bond falling in between. The
largest ClF bond extensions and red shifts are also obtained with the lithium and
beryllium bonds.
2. The substantial electron density transfer from the S atom of thiirane to ClF
noted by Hill [108] is also evident from the NPA results in Table 12.8, which
show an increase of 0.246e mainly into the antibonding σ∗ orbital of ClF in
the thiirane. . . ClF dimer. This transfer causes a significant increase in the ClF
bondlength (by 0.127 Å) and appears to be further enhanced by the addition of
the X F or BeH2 molecule to ClF.
Fig. 12.8 Optimized

geometries for a
(CH2 )2 S. . . ClF. . . LiF and b
(CH2 )2 S. . . ClF. . . BeH2
For example, Table 12.8 shows an increase in the positive charge on the S atom
(corresponding to greater charge transfer to ClF) going from the less polarizing
HF and ClF molecules to the more polarizing LiF and BeH2 molecules. The
S. . . Cl distance also decreases from 2.459 Å in the thiirane. . . ClF dimer to
2.278 Å for the most weakly-bound trimer (thiirane. . . ClF. . . ClF) to 2.072 Å
for the strongly-bound thiirane. . . ClF. . . BeH2 , see Table 12.6.
3. Hence, the addition of a more polarizing X F/BeH2 molecule to the thiirane. . . ClF
dimer leads to a stronger inner complex trimer. Table 12.7 also shows substantial
cooperativity in the thiirane. . . ClF. . . X F series, to a much greater extent than
is evident for the thiirane. . . HF. . . X F set. The F. . . Li and F. . . Be distances
in the optimized structures of thiirane. . . ClF. . . LiF and thiirane. . . ClF. . . BeH2
are short (1.757 and 1.444 Å, respectively). These complexes are more
strongly-bound than the other members of this set of trimers; there is evidence
of the cyclization of atoms from each of the three molecules constituting the
trimer—the F atom (of LiF) and the H atom (of BeH2 ) make close contacts
(1.680 and 1.456 Å, respectively) with one of the nearby protons of the CH2
fragment of thiirane. There is little evidence of strong association between the
terminal thiirane and X F molecules in the other members of this set.
Fig. 12.9 Optimized

geometries for a
(CH2 )2 S. . . LiF. . . LiF and b
(CH2 )2 S. . . LiF. . . BeH2
Fig. 12.10 Optimized

geometry for
(CH2 )2 S. . . BeH2 . . . LiF

Table 12.7 Sum of pairwise additive ( Eij ) and percentage non-additive contributions
(%Enon−add ) of E for [CH2 )2 S]a . . . [XF/BeH2 ]b . . . [X F/BeH2 ]c (X, X = H, Cl, Br, Li) complexes
at the B3LYP/6-311++G(3df,3pd) level. All values are measured in kJ/mol

Complexes E Eij %Enon−add
a. . . . . . . . . ..b. . . . . . . . . ..c
(CH2 )2 S. . . HF. . . HF − 72.0 − 50.1 30.4
(CH2 )2 S. . . HF. . . ClF − 52.0 − 38.2 26.5
(CH2 )2 S. . . HF. . . BrF − 58.2 − 40.1 31.1
(CH2 )2 S. . . HF. . . LiF − 136.7 − 86.3 36.9
(CH2 )2 S. . . HF. . . BeH2 − 97.8 4.0 104.1
(CH2 )2 S. . . ClF. . . HF − 104.8 0.3 100.3
(CH2 )2 S. . . ClF. . . ClF − 85.1 9.8 111.5
(CH2 )2 S. . . ClF. . . LiF − 213.1 90.9 142.7
(CH2 )2 S. . . ClF. . . BeH2 − 192.9 299.9 255.5
(CH2 )2 S. . . BrF. . . HF − 110.8 − 41.7 62.4
(CH2 )2 S. . . BrF. . . ClF − 91.9 − 34.0 63.0
(CH2 )2 S. . . BrF. . . LiF − 199.2 10.1 105.1
(CH2 )2 S. . . BrF. . . BeH2 − 165.0 21.3 112.9
(CH2 )2 S. . . LiF. . . HF − 169.1 − 106.7 36.9
(CH2 )2 S. . . LiF. . . BrF − 151.3 − 139.6 7.7
(CH2 )2 S. . . LiF. . . LiF − 302.7 − 279.3 7.7
(CH2 )2 S. . . LiF. . . BeH2 − 273.6 − 197.8 27.7
(CH2 )2 S. . . BeH2 . . . LiF − 152.0 − 56.3 63.0
(CH2 )2 S. . . BeH2 . . . BeH2 − 206.5 − 44.3 78.5
12.2.3.3 Thiirane. . . BrF. . . X F/BeH2
1. The trends observed for thiirane. . . ClF. . . X F/BeH2 are mirrored by this set
of trimers, except that there is less charge transfer from S to Br (Table 12.8),
smaller non-additive energies (Table 12.7), longer S. . . Br distances (Table 12.6)
and F. . . X /Be distances, and smaller X-F/BeH bond extensions/red shifts (not
shown). The larger Br atom probably introduces more electron density in the inter-
molecular S. . . Br region (relative to the S. . . Cl region of thiirane. . . ClF. . . X F)
and thereby reduces the electron density transferred from thiirane.
2. The lithium and beryllium bonds due to LiF and BeH2 produce the most
strongly-bound trimers and the close interatomic contacts in these complexes
suggest an association between the terminal molecules (as was the case for
thiirane. . . ClF. . . X F/BeH2 ) which leads to relatively large interaction and
non-additive energies, compared with the other members of this set.
Table 12.8 Changes in NPA partial atomic charges (q in e) w.r.t isolated monomers for
(CH2 )2 S. . . XF/BeH2 . . . X F/BeH2 (X , X = H, Cl, Br, Li) complexes and electron density transfer
(EDT in e) determined using the B3LYP/6-311++G(3df,3pd) procedure. Partial atomic charges
for monomers are in parentheses
Complex q (S) Complex q (S)
(CH2 )2 S. . . HF 0.003 (0.085) (CH2 )2 S. . . BrF. . . HF 0.276
(CH2 )2 S. . . ClF 0.246 (CH2 )2 S. . . BrF. . . ClF 0.270
(CH2 )2 S. . . BrF 0.200
(CH2 )2 S. . . LiF − 0.108 (CH2 )2 S. . . BrF. . . LiF − 0.336
(CH2 )2 S. . . BeH2 − 0.061 (CH2 )2 S. . . BrF. . . BeH2 0.372
(CH2 )2 S. . . HF. . . HF 0.009 (CH2 )2 S. . . LiF. . . HF − 0.078
(CH2 )2 S. . . HF. . . ClF 0.016
(CH2 )2 S. . . HF. . . BrF 0.021 (CH2 )2 S. . . LiF. . . BrF − 0.073
(CH2 )2 S. . . HF. . . LiF 0.009 (CH2 )2 S. . . LiF. . . LiF − 0.059
(CH2 )2 S. . . HF. . . BeH2 − 0.278 (CH2 )2 S. . . LiF. . . BeH2 − 0.057
(CH2 )2 S. . . ClF. . . HF 0.378
(CH2 )2 S. . . ClF. . . ClF 0.350
(CH2 )2 S. . . ClF. . . LiF 0.473 (CH2 )2 S. . . BeH2 . . . LiF − 0.118
(CH2 )2 S. . . ClF. . . BeH2 0.578 (CH2 )2 S. . . BeH2 . . . BeH2 − 0.012
12.2.3.4 Thiirane. . . LiF. . . X F/BeH2
1. The thiirane. . . LiF dimer shows the least electron density transfer from thiirane
to XF (Table 12.7) probably because this charge shift is likely to be retarded by
repulsion from the tightly held core electrons on Li. On the other hand, the binding
in this most strongly-bound dimer is dominated by the polarization induced in
thiirane by the large LiF dipole moment.
2. The thiirane. . . LiF. . . X F/BeH2 complexes are the most strongly-bound of all
the trimer sets and their structures suggest that the LiF and X F/BeH2 molecular
subunits form strong aggregates or clusters, which interact with thiirane through
a S. . . Li contact. This idea is supported by the fact that the LiF. . . X F/BeH2
pair energies were found to be the most dominant attractive component of the
interaction, followed by the thiirane. . . LiF pair energy.
3. The non-additive energies are relatively low for these trimers, with the largest
cooperative effect estimated for the X F = HF species (representing 37 % of E).
The thiirane. . . LiF. . . LiF is the most strongly bound trimer (E = −303 kJ/mol)
and has the closest contact between the F atom of LiF and one of the protons of
thiirane.
Table 12.9 B3LYP/6- Complex (ρ,L)S. . . X or (ρ,L)S. . . Be

311++G(3df,3pd) QTAIM
electron density (ρ/ a.u) and (CH2 )2 S. . . HF (0.034, − 0.045)
negative Laplacian of electron (CH2 )2 S. . . ClF (0.064, − 0.090)
density (L/a.u) at the
(CH2 )2 S. . . BrF (0.058, − 0.072)
S. . . X/Be and X-F bond
critical points for (CH2 )2 S. . . LiF (0.020, − 0.083)
(CH2 )2 S. . . XF/BeH2 , (CH2 )2 S. . . BeH2 (0.042, − 0.134)
FX /H2 Be. . . (CH2 )2 S. . . XF/BeH2
(X , X = H, Cl, Br, Li) FH. . . (CH2 )2 S. . . HF (0.029, − 0.048)
complexes FH. . . (CH2 )2 S. . . ClF (0.055, − 0.092)
FH. . . (CH2 )2 S. . . BrF (0.052, − 0.075)
FH. . . (CH2 )2 S. . . LiF (0.018, − 0.075)
FH. . . (CH2 )2 S. . . BeH2 (0.038, − 0.126)
12.2.3.5 Thiirane. . . BeH2 . . . X F/BeH2
1. These complexes are the most unstable set of trimers. In the thiirane. . . BeH2
dimer, the small and highly positive Be atom can approach the S atom quite
closely. However, this close approach causes the hydridic H atoms to be repelled
by the S lone pair thereby pushing the H atoms away from the S atom and bending
the H-Be-H angle. This creates a dipole moment which can then polarize the S
atom, shifting electron density onto the S atom and making it less positive and
the Be atom more positive (Table 12.8).
Attachment of a hydrogen bond or halogen bonds to one of the H atoms of BeH2
apparently destabilizes the complex, possibly because the electron cloud of the
H atom is too diffuse to form a strong attachment to the proton of HF or repels
the electron density on the halogen atom (Cl or Br).
2. The thiirane. . . BeH2 . . . LiF trimer shows cyclization of the atoms from each of
the three monomers, with a short contact of 1.829 Å between the H (of thiirane)
and the F atom of LiF joining the two terminal molecules together. The non-
additive energy that results from this interaction is estimated to be about 63 % of
E (Table 12.7) and indicates significant positive cooperativity.
3. By contrast, the thiirane. . . BeH2 . . . BeH2 trimer is more strongly bound and
benefits from a strong aggregation between the two BeH2 molecules. The
BeH2 . . . BeH2 pair energy and the non-additive energy dominate the interaction
and suggest the existence of a sizeable Be. . . Be interaction.
12.2.4 AIM Analysis of Thiirane. . . XF/BeH2 ,

FX /H2 Be. . . Thiirane. . . XF/BeH2 , and
Thiirane. . . XF/BeH2 . . . X F/BeH2
For the thiirane. . . XF/BeH2 dimer, ρ varies between 0.020 and 0.064 a.u and L
between − 0.045 and − 0.134 a.u (Table 12.9) for the S. . . X/Be bcp, indicating
Table 12.10 B3LYP/6-311++G(3df,3pd) QTAIM electron density (ρ) and negative Laplacian of
electron density (L) at the S. . . X/Be and X-F bond critical points for (CH2 )2 S. . . XF/BeH2 and
(CH2 )2 S. . . XF/BeH2 . . . X F/BeH2 (X , X = H, Cl, Br, Li) complexes
Complex (ρ,L)S. . . X or Complex (ρ,L)S. . . X or
(ρ,L)S. . . Be (ρ,L)S. . . Be
(CH2 )2 S. . . HF (0.034, − 0.045) (CH2 )2 S. . . BrF. . . HF (0.075, − 0.061)
(CH2 )2 S. . . ClF (0.064, − 0.090) (CH2 )2 S. . . BrF. . . ClF (0.073, − 0.063)
(CH2 )2 S. . . BrF (0.058, − 0.072)
(CH2 )2 S. . . LiF (0.020, − 0.083) (CH2 )2 S. . . BrF. . . LiF (0.099, − 0.023)
(CH2 )2 S. . . BeH2 (0.042, − 0.134) (CH2 )2 S. . . BrF. . . BeH2 (0.098, − 0.029)
(CH2 )2 S. . . HF. . . HF (0.045, − 0.029) (CH2 )2 S. . . LiF. . . HF (0.020, − 0.081)
(CH2 )2 S. . . HF. . . ClF (0.040, − 0.036)
(CH2 )2 S. . . HF. . . BrF (0.043, − 0.031) (CH2 )2 S. . . LiF. . . BrF (0.021, − 0.083)
(CH2 )2 S. . . HF. . . LiF (0.059, − 0.005) (CH2 )2 S. . . LiF. . . LiF (0.018, − 0.072)
(CH2 )2 S. . . HF. . . BeH2 (0.068, − 0.032) (CH2 )2 S. . . LiF. . . BeH2 (0.020, − 0.080)
(CH2 )2 S. . . ClF. . . HF (0.096, − 0.079)
(CH2 )2 S. . . ClF. . . ClF (0.094, − 0.069)
(CH2 )2 S. . . ClF. . . LiF (0.136, − 0.029) (CH2 )2 S. . . BeH2 . . . LiF (0.052, − 0.145)
(CH2 )2 S. . . ClF. . . BeH2 (0.143, − 0.061) (CH2 )2 S. . . BeH2 . . . BeH2 (0.044, − 0.149)
non-covalent interactions. In the FX /H2 Be. . . thiirane. . . XF/BeH2 trimers, for fixed
XF/BeH2 , there is little variation in ρ and L values for S. . . X/Be bonds. In going
from dimer to trimer, there is a decrease in ρ due to the FX interaction, consistent
with the competition for the S lone pair electron density by both FX and XF. This
decrease in ρ is also consistent with the previously discussed negative cooperative
effect.
Compared with FX . . . thiirane. . . XF, there is a greater variation in ρ and L for
the S. . . X bcp for the thiirane. . . XF/BeH2 . . . X F/BeH2 trimers. Going from dimer
to trimer, ρ(S. . . X/Be) increases, which is consistent with the positive cooperative
effect due to FX . Generally, the most polarizing molecules, LiF and BeH2 , produce
the largest ρ(S. . . X/Be) values since these molecules would shift more electron
density into the S. . . X intermolecular region. (Table 12.10)
12.3 Conclusions
a. Our study supports and extends Hill’s work on thiirane. . . ClF, where he proposed
that this species be considered as an example of a Mulliken inner complex.
b. The binding strength of the thiirane. . . XF/BeH2 dimers increases in the order:
lithium bond > halogen bond > beryllium bond > hydrogen bond.
c. For the FX /H2 Be. . . (CH2 )2 S. . . XF/BeH2 trimers, where the thiirane acts as a
double acceptor, negative cooperativity is evident, with the largest effects due to
the halogen and beryllium bonds.
d. For the (CH2 )2 S. . . XF/BeH2 . . . X F/BeH2 trimers, the binding strength depends
on the polarizing power of the terminal X F/BeH2 molecules and large coopera-
tive effects are evident, with some complexes showing cyclization involving the
terminal molecules.
e. The diminutive effect in FX /H2 Be. . . (CH2 )2 S. . . XF/BeH2 and the coopera-
tive effect in (CH2 )2 S. . . XF/BeH2 . . . X F/BeH2 are mirrored by the respective
decrease and increase in the electron density in the intermolecular S. . . X region.
f. Generally, the complexes with relatively small S to XF charge transfer are domi-
nated by strong electrostatic interactions (e.g. XF = LiF and HF), whereas greater
charge transfer occurs in the halogen-bonded systems, where the electrostatic
contribution to the binding is presumably relatively smaller than in the more
strongly-bound analogues.
Acknowledgement We would like to thank the School for Graduate Studies and Research of the
University of the West Indies, Cave Hill Campus, for financial support.
References
1. Pimentel G, McClellan A (1960) The hydrogen bond. W. H. Freeman & Co., San Francisco
2. Henri-Rousseau O, Blaise P, Hadzi D (1997) Theoretical treatments of hydrogen bonding.
Wiley, New York
3. Jeffrey GA (1997) An introduction to hydrogen bonding, vol 12. Oxford University Press,
New York
4. Scheiner S (1997) Hydrogen bonding: a theoretical perspective. Oxford University Press,
New York
5. Desiraju GR (1999) Weak hydrogen bond. Oxford University Press, New York
6. Buckingham AD, Del Bene JE, McDowell SAC (2008) The hydrogen bond. Chem Phys Lett
463(1–3):1-10. doi:10.1016/j.cplett.2008.06.060
Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ (2011) Defini-
tion of the hydrogen bond (iupac recommendations 2011). Pure Appl Chem 83(8):1637–1641.
doi:10.1351/pac-rec-10–01-02
8. Metrangolo P, Resnati G (2001) Halogen bonding: a paradigm in supramolecular chem-
istry. Chem-a Eur J 7(12):2511–2519. doi:10.1002/1521-3765(20010618)7:12<2511::aid-
chem25110>3.0.co;2-t
9. Clark T, Hennemann M, Murray JS, Politzer P (2007) Halogen bonding: the sigma-hole. J
Mol Model 13(2):291–296. doi:10.1007/s00894-006-0130-2
10. Politzer P, Murray JS (2013) Halogen bonding: an interim discussion. Chemphyschem
14(2):278–294. doi:10.1002/cphc.201200799
11. Bleiholder C, Werz DB, Koppel H, Gleiter R (2006) Theoretical investigations on chalcogen-
chalcogen interactions: what makes these nonbonded interactions bonding? J Am Chem Soc
128(8):2666–2674. doi:10.1021/ja056827g
12. Murray JS, Lane P, Clark T, Politzer P (2007) Sigma-hole bonding: molecules containing
group vi atoms. J Mol Model 13(10):1033–1038. doi:10.1007/s00894-007-0225-4
interaction. Int J Quantum Chem 107(12):2286–2292. doi:10.1002/qua.21352
14. Murray JS, Lane P, Politzer P (2009) Expansion of the sigma-hole concept. J Mol Model
15(6):723–729. doi:10.1007/s00894-008-0386-9
15. Bauza A, Mooibroek TJ, Frontera A (2013) Tetrel-bonding interaction: re-
discovered supramolecular force? Angew Chem-Int Ed 52(47):12317–12321.
doi:10.1002/anie.201306501
16. Auffinger P, Hays FA, Westhof E, Ho PS (2004) Halogen bonds in biological molecules. Proc
Natl Acad Sci U S A 101(48):16789–16794. doi:10.1073/pnas.0407607101
17. Politzer P, Lane P, Concha MC, Ma YG, Murray JS (2007) An overview of halogen bonding.
J Mol Model 13(2):305–311. doi:10.1007/s00894-006-0154-7
18. Metrangolo P, Resnati G (2008) Halogen bonding: fundamentals and applications, vol 126.
Springer,
19. Legon AC (2010) The halogen bond: an interim perspective. Phys Chem Chem Phys
12(28):7736–7747. doi:10.1039/c002129f
20. Politzer P, Murray JS, Clark T (2010) Halogen bonding: an electrostatically-driven
highly directional noncovalent interaction. Phys Chem Chem Phys 12(28):7748–7757.
doi:10.1039/c004189k
21. Desiraju GR, Ho PS, Kloo L, Legon AC, Marquardt R, Metrangolo P, Politzer P, Resnati G,
Rissanen K (2013) Definition of the halogen bond (iupac recommendations 2013). Pure Appl
Chem 85(8):1711–1713. doi:10.1351/pac-rec-12-05-10
22. Politzer P, Murray JS, Clark T (2013) Halogen bonding and other sigma-hole interactions: a
perspective. Phys Chem Chem Phys 15(27):11178–11189. doi:10.1039/c3cp00054k
23. Wang WZ, Wong NB, Zheng WX, Tian AM (2004) Theoretical study on the blueshifting
halogen bond. J Phys Chem A 108(10):1799–1805. doi:10.1021/jp036769q
24. Murray JS, Concha MC, Lane P, Hobza P, Politzer P (2008) Blue shifts vs red shifts in
sigma-hole bonding. J Mol Model 14(8):699–704. doi:10.1007/s00894-008-0307-y
25. Brinck T, Murray JS, Politzer P (1992) Surface electrostatic potentials of halogenated
methanes as indicators of directional intermolecular interactions. Int J Quantum Chem 57–64
26. Murray JS, Paulsen K, Politzer P (1994) Molecular-surface electrostatic potentials in the
analysis of non-hydrogen-bonding noncovalent interactions. Proc Indian Acad Sci-Chem Sci
106(2):267–275
27. Murray JS, Politzer P (1998) Statistical analysis of the molecular surface electrostatic poten-
tial: an approach to describing noncovalent interactions in condensed phases. Theochem-J
Mol Struct 425(1–2):107–114
28. Politzer P, Murray JS (2002) The fundamental nature and role of the electrostatic potential in
atoms and molecules. Theor Chem Acc 108(3):134–142. doi:10.1007/s00214-002-0363-9
29. Murray JS, Politzer P (2011) The electrostatic potential: an overview. Wiley Interdiscip Rev-
Comput Mol Sci 1(2):153–163. doi:10.1002/wcms.19
30. Riley KE, Murray JS, Fanfrlik J, Rezac J, Sola RJ, Concha MC, Ramos FM, Politzer P
(2011) Halogen bond tunability i: the effects of aromatic fluorine substitution on the strengths
of halogen-bonding interactions involving chlorine, bromine, and iodine. J Mol Model
17(12):3309–3318. doi:10.1007/s00894-011-1015-6
31. Bundhun A, Ramasami P, Murray JS, Politzer P (2013) Trends in sigma-hole strengths and
interactions of F3MX molecules (M = C, Si, Ge and X = F, Cl, Br, I). J Mol Model 19(7):2739–
2746. doi:10.1007/s00894-012-1571-4
32. Murray-Rust P, Motherwell WDS (1979) Computer retrieval and analysis of molecular
geometry. 4. Intermolecular interactions. J Am Chem Soc 101(15):4374–4376
33. Hassel O, Hvoslef J (1954) The structure of bromine 1, 4-dioxanate. Acta Chem Scand
8(5):873–873
34. Zhou P-P, Qiu W-Y, Liu S, Jin N-Z (2011) Halogen as halogen-bonding donor and hydrogen-
bonding acceptor simultaneously in ring-shaped H3NX(Y)HF (X = Cl, Br and Y =
F, Cl, Br) complexes. Phys Chem Chem Phys 13(16):7408–7418. doi:10.1039/c1cp00025j
35. McDowell SAC, Joseph JA (2012) Communication: an unusual halogen-bonding motif: the
LiBr. . .BrF dimer as a model system. J Chem Phys 137(17). doi:10.1063/1.4766932
36. Bilewicz E, Rybarczyk-Pirek AJ, Dubis AT, Grabowski SJ (2007) Halogen bonding in crystal
structure of 1-methylpyrrol-2-yl trichloromethyl ketone. J Mol Struct 829(1–3):208–211.
doi:10.1016/j.molstruc.2006.06.032
37. Politzer P, Murray JS, Concha MC (2007) Halogen bonding and the design of new materials:
organic bromides, chlorides and perhaps even fluorides as donors. J Mol Model 13(6–7):643–
650. doi:10.1007/s00894-007-0176-9
38. Price SL, Stone AJ, Lucas J, Rowland RS, Thornley AE (1994) The nature of -Cl. . . Cl-
intermolecular interactions. J Am Chem Soc 116(11):4910–4918. doi:10.1021/ja00090a041
39. Voth AR, Hays FA, Ho PS (2007) Directing macromolecular conformation through halogen
bonds. Proc Natl Acad Sci U S A 104(15):6188–6193. doi:10.1073/pnas.0610531104
40. Lu Y, Shi T, Wang Y, Yang H, Yan X, Luo X, Jiang H, Zhu W (2009) Halogen bonding-a novel
interaction for rational drug design? J Med Chem 52(9):2854–2862. doi:10.1021/jm9000133
41. Voth AR, Khuu P, Oishi K, Ho PS (2009) Halogen bonds as orthogonal molecular interactions
to hydrogen bonds. Nat Chem 1(1):74–79. doi:10.1038/nchem.112
42. Hardegger LA, Kuhn B, Spinnler B, Anselm L, Ecabert R, Stihle M, Gsell B, Thoma R,
Diez J, Benz J, Plancher J-M, Hartmann G, Banner DW, Haap W, Diederich F (2011) Sys-
tematic investigation of halogen bonding in protein-ligand interactions. Angew Chem-Int Ed
50(1):314–318. doi:10.1002/anie.201006781
43. Imakubo T, Sawa H, Kato R (1995) Novel radical-cation salts of organic pi-donors contain-
ing iodine atom(s)—the first application of strong intermolecular -I-X- (X = CN, halogen
atom) interaction to molecular conductors. Synth Met 73(2):117–122. doi:10.1016/0379-
6779(95)03322-x
44. Amico V, Meille SV, Corradi E, Mesina MT, Resnati G (1998) Perfluorocarbon–hydrocarbon
self-assembling. 1d infinite chain formation driven by nitrogen· · ·iodine interactions. J Am
Chem Soc 120(32):8261–8262. doi:10.1021/ja9810686
45. Iwaoka M, Takemoto S, Tomoda S (2002) Statistical and theoretical investigations on the
directionality of nonbonded (SO)-O- Interactions. Implications for molecular design and
protein engineering. J Am Chem Soc 124(35):10613–10620. doi:10.1021/ja026472q
46. Metrangolo P, Neukirch H, Pilati T, Resnati G (2005) Halogen bonding based recogni-
tion processes: a world parallel to hydrogen bonding. Acc Chem Res 38(5):386–395.
doi:10.1021/ar0400995
47. Saha BK, Nangia A, Jaskolski M (2005) Crystal engineering with hydrogen bonds and halogen
bonds. Crystengcomm 7:355–358. doi:10.1039/b501693b
48. Syssa-Magale JL, Boubekeur K, Schollhorn B (2005) First molecular self-assembly of 1,4-
diiodo-tetrafluoro-benzene and a ketone via (o. . . ) non-covalent halogen bonds. J Mol Struct
737(2–3):103–107. doi:10.1016/j.molstruc.2004.10.008
49. Lucassen ACB, Karton A, Leitus G, Shimon LJW, Martin JML, van der Boom ME (2007) Co-
crystallization of sym-triiodo-trifluorobenzene with bipyridyl donors: consistent formation
of two instead of anticipated three N. . .I halogen bonds. Cryst Growth Des 7(2):386–392.
doi:10.1021/cg0607250
50. Metrangolo P, Meyer F, Pilati T, Resnati G, Terraneo G (2008) Halogen bonding in supramolec-
ular chemistry. Angew Chem-Int Ed 47(33):6114–6127. doi:10.1002/anie.200800128
51. Rissanen K (2008) Halogen bonded supramolecular complexes and networks. Crystengcomm
10(9):1107–1113. doi:10.1039/b803329n
52. Moilanen J, Ganesamoorthy C, Balakrishna MS, Tuononen HM (2009) Weak interactions
between trivalent pnictogen centers: computational analysis of bonding in dimers X3E. . .EX3
(E = pnictogen, X = halogen). Inorg Chem 48(14):6740–6747. doi:10.1021/ic900635f
53. Kollman PA, Liebman JF, Allen LC (1970) Lithium bond. J Am Chem Soc 92(5):1142–1150
54. Ault BS, Pimentel GC (1975) Matrix isolation infrared studies of lithium bonding. J Phys
Chem 79(6):621–626
55. Sannigrahi A, Kar T, Niyogi BG, Hobza P, Schleyer PvR (1990) The lithium bond reexamined.
Chem Rev 90(6):1061–1076
56. Ammal SSC, Venuvanalingam P (1998) Ab initio and dft investigations of lithium/hydrogen
bonded complexes of trimethylamine, dimethyl ether and dimethyl sulfide. J Chem Soc-
Faraday Trans 94(18):2669–2674
57. Vila A, Vila E, Mosquera RA (2006) Topological characterisation of intermolecular lithium
bonding. Chem Phys 326(2–3):401–408. doi:10.1016/j.chemphys.2006.02.032
58. Bader RFW (1990) Atoms in molecules—a quantum theory. Oxford University Press, Oxford
59. Lipkowski P, Grabowski SJ (2014) Could the lithium bond be classified as
the sigma-hole bond?—qtaim and nbo analysis. Chem Phys Lett 591:113–118.
doi:10.1016/j.cplett.2013.11.017
60. Feng Y, Liu L, Wang JT, Li XS, Guo QX (2004) Blue-shifted lithium bonds. Chem Commun
(1):88–89. doi:10.1039/b310723j
61. Li QZ, Wang YF, Li WZ, Cheng JB, Gong BA, Sun JZ (2009) Prediction and characterization
of the HMgHLiX (X = H, OH, F, CCH, CN, and NC) complexes: a lithium–hydride lithium
bond. Phys Chem Chem Phys 11(14):2402–2407. doi:10.1039/b820309a
62. McDowell SAC, St Hill JAS (2011) A theoretical study of hydrogen- and lithium-bonded
complexes of F-H/Li and Cl-H/Li with NF3, NH3, and NH2(CH3). J Chem Phys 135(16).
doi:10.1063/1.3653476
63. Yanez M, Sanz P, Mo O, Alkorta I, Elguero J (2009) Beryllium bonds, do they exist? J Chem
Theory Comput 5(10):2763–2771. doi:10.1021/ct900364y
64. Eskandari K (2012) Characteristics of beryllium bonds; a qtaim study. J Mol Model
18(8):3481–3487. doi:10.1007/s00894-012-1360-0
65. Yanez M, Mo O, Alkorta I, Elguero J (2013) Can conventional bases and unsaturated
hydrocarbons be converted into gas-phase superacids that are stronger than most of the
known oxyacids? The role of beryllium bonds. Chem-a Eur J 19(35):11637–11643.
doi:10.1002/chem.201300808
66. Suhai S (1994) Cooperative effects in hydrogen-bonding—4th-order many-body perturbation-
theory studies of water oligomers and of an infinite water chain as a model for ice. J Chem
Phys 101(11):9766–9782. doi:10.1063/1.467942
67. Tsuzuki S, Houjou H, Nagawa Y, Goto M, Hiratani K (2001) Cooperative enhancement of
water binding to crownophane by multiple hydrogen bonds: analysis by high level ab initio
calculations. J Am Chem Soc 123(18):4255–4258. doi:10.1021/ja0037264
68. Szczesna B, Urbanczyk-Lipkowska Z (2002) Cooperative effect of multiple hydrogen bonding
involving the nitro group: Solid state dimeric self-assembly of o-, m- and p-hydroxyphenyl-
2,4-dinitrophenylhydrazones. New J Chem 26(2):243–249. doi:10.1039/b105498h
69. Thallapally PK, KatzAK, Carrell HL, Desiraju GR (2002) Unusually long cooperative chain of
seven hydrogen bonds. An alternative packing type for symmetrical phenols. Chem Commun
(4):344–345. doi:10.1039/b110036j
70. Parra RD, Bulusu S, Zeng XC (2003) Cooperative effects in one-dimensional chains
of three-center hydrogen bonding interactions. J Chem Phys 118(8):3499–3509.
doi:10.1063/1.1535441
71. Araujo RCMU, Soares VM, Oliveira BG, Lopes KC, Ventura E, Do Monte SA, Santana OL,
Carvalho AB, Ramos MN (2006) Theoretical study of cooperative effects in the homo- and
heteromeric hydrogen bond chains (HCN)(n)–HF. with n = 1, 2, and 3. Int J Quantum Chem
106(13):2714–2722. doi:10.1002/qua.21132
72. Larsen RW, Suhm MA (2006) Cooperative organic hydrogen bonds: the librational modes of
cyclic methanol clusters. J Chem Phys 125(15). doi:10.1063/1.2358349
73. Song H-J, Xiao H-M, Dong H-S (2006) Cooperative effects, strengths of hydrogen bonds,
and intermolecular interactions in circular cis, trans-cyclotriazane clusters (n = 3-8). J Chem
Phys 125(7). doi:10.1063/1.2336209
74. Znamenskiy VS, Green ME (2007) Quantum calculations on hydrogen bonds in cer-
tain water clusters show cooperative effects. J Chem Theory Comput 3(1):103–114.
doi:10.1021/ct600139d
75. Fradelos G, Kaminski JW, Wesolowski TA, Leutwyler S (2009) Cooperative effect of
hydrogen-bonded chains in the environment of pi → pi* chromophore. J Phys Chem A
113(36):9766–9771. doi:10.1021/jp906483z
76. Rodziewicz P, Rutkowski KS, Melikova SM, Koll A (2009) Cooperative effects in blue-shifted
hydrogen bonded cluster of CF3H. . . HF)(1 < = n < = 3) from first principles simulations.
Chem Phys 361 (3):129–136. doi:10.1016/j.chemphys.2009.05.017
77. Roztoczynska A, Kozlowska J, Lipkowski P, Bartkowiak W (2014) Does the spatial confine-
ment influence the electric properties and cooperative effects of the hydrogen bonded systems?
Hcn chains as a case study. Chem Phys Lett 608:264–268. doi:10.1016/j.cplett.2014.05.102
78. Planas JG, Vinas C, Teixidor F, Comas-Vives A, Ujaque G, Lledos A, Light ME, Hursthouse
MB (2005) Self-assembly of mercaptane–metallacarborane complexes by an unconventional
cooperative effect: a C–H. . .S–H. . .H–B hydrogen/dihydrogen bond interaction. J Am Chem
Soc 127(45):15976–15982. doi:10.1021/ja055210w
79. Li Q-Z, Hu T, An X-L, Gong B-A, Cheng J-B (2008) Cooperativity between the dihydrogen
bond and the N. . .HC hydrogen bond in lih–(hcn)n complexes. Chemphyschem 9(13):1942–
1946. doi:10.1002/cphc.200800320
80. Solimannejad M, Rabbani M, Ahmadi A, Esrafili MD (2014) Cooperative and diminu-
tive interplay between the sodium bonding with hydrogen and dihydrogen bondings in
ternary complexes of NaC3N with HMgH and HCN (HNC). Mol Phys 112(15):2017–2022.
doi:10.1080/00268976.2013.879496
81. Lankau T, WuY-C, Zou J-W,Yu C-H (2008) The cooperativity between hydrogen and halogen
bonds. J Theor Comput Chem 7(1):13–35. doi:10.1142/s0219633608003563
82. Li Q, Lin Q, Li W, Cheng J, Gong B, Suo J (2008) Cooperativity between the halogen bond
and the hydrogen bond in H3N. . .XY. . .HF complexes (X, Y = F, Cl, Br). Chemphyschem
9(15):2265–2269. doi:10.1002/cphc.200800467
83. Jing B, Li Q, Li R, Gong B, Liu Z, Li W, Cheng J, Sun J (2011) Competition and cooperativity
between hydrogen bond and halogen bond in HNC. . .(HOBr)n and (HNC)n. . .HOBr (n = 1
and 2) systems. Comput Theor Chem 963(2–3):417–421. doi:10.1016/j.comptc.2010.11.006
84. Solimannejad M, Malekani M, Alkorta I (2011) Cooperativity between the hydrogen bonding
and halogen bonding in F3CX . . . NCH(CNH) . . . NCH(CNH) complexes (X = Cl, Br). Mol
Phys 109(13):1641–1648. doi:10.1080/00268976.2011.582050
85. Zhao Q, Feng D, Hao J (2011) The cooperativity between hydrogen and halogen bond
in the XY. . .HNC. . .XY (X, Y = F, Cl, Br) complexes. J Mol Model 17(11):2817–2823.
doi:10.1007/s00894-011-0974-y
86. Li Q, Li R, Zhou Z, Li W, Cheng J (2012) S. . .X halogen bonds and H. . .X hydrogen bonds in
H2CS–XY (XY = FF, ClF, ClCl, BrF, BrCl, and BrBr) complexes: cooperativity and solvent
effect. J Chem Phys 136(1). doi:10.1063/1.3673540
87. Solimannejad M, Malekani M (2012) Cooperative and diminutive interplay between the hy-
drogen bonding and halogen bonding in ternary complexes of HCCX (X = Cl, Br) with hcn
and hnc. Comput Theor Chem 998:34–38. doi:10.1016/j.comptc.2012.05.021
88. Grabowski SJ (2013) Cooperativity of hydrogen and halogen bond interactions. Theor Chem
Acc 132(4). doi:10.1007/s00214-013-1347-7
89. Alkorta I, Blanco F, Deya PM, Elguero J, Estarellas C, Frontera A, Quinonero D
(2010) Cooperativity in multiple unusual weak bonds. Theor Chem Acc 126(1–2):1–14.
doi:10.1007/s00214-009-0690-1
90. McDowell SAC, Joseph JA (2012) Cooperative effects of noncovalent bonds to the Br
atom of halogen-bonded H3N. . . BrZ and HCN. . . BrZ (Z = F, Br). J Chem Phys 137(7).
doi:10.1063/1.4745838
91. Zhao XR, Wu YJ, Han J, Shen QJ, Jin WJ (2013) Theoretical study of the triangular bonding
complex formed by carbon tetrabromide, a halide, and a solvent molecule in the gas phase. J
Mol Model 19(1):299–304. doi:10.1007/s00894-012-1518-9
92. Li Q, Hu T, An X, Li W, Cheng J, Gong B, Sun J (2009) Theoretical study of the inter-
play between lithium bond and hydrogen bond in complexes involved with hli and hcn.
Chemphyschem 10(18):3310–3315. doi:10.1002/cphc.200900549
93. McDowell SAC, Yarde HK (2012) Cooperative effects of hydrogen, lithium and halo-
gen bonding on f–h/lioh2 complexes. Phys Chem Chem Phys 14(19):6883–6888.
doi:10.1039/c2cp40203c
94. Solimannejad M, Rezaei Z, Esrafili MD (2013) Competition and interplay between the lithium
bonding and hydrogen bonding: R3c center dot center dot center dot hy center dot center dot
center dot liy and r3c center dot center dot center dot liy center dot center dot center dot
hy triads as a working model (r = h, ch3; y = cn, nc). J Mol Model 19(11):5031–5035.
doi:10.1007/s00894-013-2006-6
95. Solimannejad M (2012) Cooperative and diminutive interplay between lithium and dihydrogen
bonding in f3yli center dot center dot center dot nch center dot center dot center dot center dot
hmh and f3yli center dot center dot center dot hmh center dot center dot center dot hcn triads
(y = c, si; m = be, mg). Chemphyschem 13(13):3158–3162. doi:10.1002/cphc.201200333
96. McDowell SAC, Joseph JA (2013) Cooperative effects in novel lif/hf· · ·lif· · ·xf (x = f, cl, br)
clusters. J Chem Phys 138(16). doi:10.1063/1.4801863
97. Li Q, Li R, Liu Z, Li W, Cheng J (2011) Interplay between halogen bond and lithium bond
in mcn-licn-xcch (m = h, li, and na; x = cl, br, and i) complex: The enhancement of halogen
bond by a lithium bond. J Comput Chem 32(15):3296–3303. doi:10.1002/jcc.21916
98. Esrafili MD, Esmailpour P, Mohammadian-Sabet F, Solimannejad M (2013) Theoretical study
of the interplay between halogen bond and lithium-pi interactions: cooperative and diminutive
effects. Chem Phys Lett 588:47–50. doi:10.1016/j.cplett.2013.10.009
99. Esrafili MD, Mohammadian-Sabet F, Solimannejad M (2014) A theoretical evidence
for mutual influence between s· · ·n(c) and hydrogen/lithium/halogen bonds: Competition
and interplay between π-hole and σ-hole interactions. Struct Chem 25(4):1197–1205.
doi:10.1007/s11224-014-0392-8
100. Solimannejad M, Bayati E, Esrafili MD (2014) Enhancement effect of lithium bonding on
the strength of pnicogen bonds: Xh2p· · ·ncli· · ·ncy as a working model (x = f, cl; y = h, f, cl,
cn). Mol Phys 112(15):2058–2062. doi:10.1080/00268976.2014.884250
101. Li Q, Li R, Liu X, Cheng J, Li W (2012) Ab initio study of synergetic effects of two strong
interactions of cation–π interaction and lithium bond in m+· · ·phenyl lithium· · · n (m = li, na,
k; n = h2o and nh3) complex. Mol Phys 110(8):457–465. doi:10.1080/00268976.2012.655793
102. Esrafili MD, Fatehi P, Solimannejad M (2013) Cooperative effects in cyclic licn
and hcn clusters: a comparative study. Comput Theor Chem 1022:115–120.
doi:10.1016/j.comptc.2013.08.011
103. Solimannejad M, Ghafari S, Esrafili MD (2013) Theoretical insight into cooperativity in
lithium-bonded complexes: linear clusters of licn and linc. Chem Phys Lett 577:6–10.
doi:10.1016/j.cplett.2013.05.023
104. Esrafili MD, Esmailpour P, Mohammadian-Sabet F, Solimannejad M (2014) Substituent
effects on cooperativity between lithium bonds. Int J Quantum Chem 114(4):295–301.
doi:10.1002/qua.24560
105. Albrecht L, Boyd RJ, Mo O, Yanez M (2014) Changing weak halogen bonds into strong
ones through cooperativity with beryllium bonds. J Phys Chem A 118(23):4205–4213.
doi:10.1021/jp503229u
106. Albrecht L, Boyd RJ, Mo O, Yanez M (2012) Cooperativity between hydrogen bonds and
beryllium bonds in (h2o)(n)bex2 (n = 1-3, x = h, f) complexes. A new perspective. Phys
Chem Chem Phys 14(42):14540–14547. doi:10.1039/c2cp42534c
107. Mo O, Yanez M, Alkorta I, Elguero J (2012) Modulating the strength of hydrogen bonds
through beryllium bonds. J Chem Theory Comput 8(7):2293–2300. doi:10.1021/ct300243b
108. Hill JG (2014) The halogen bond in thiirane. . . Clf: an example of a mulliken inner complex.
Phys Chem Chem Phys 16(36):19137–19140. doi:10.1039/c4cp03412k
109. Mulliken RS, Person WB (1969) Molecular complexes: a lecture and reprint volume. Wiley-
Interscience, New York
110. Legon AC (1999) Prereactive complexes of dihalogens xy with lewis bases b in the gas
phase: a systematic case for the halogen analogue b—xy of the hydrogen bond b—hx. Angew
Chem-Int Ed 38(18):2687–2714
111. Evans CM, Holloway JH, Legon AC (1996) Nature and angular geometry of the pre-reactive
complex thiirane-chlorine monofluoride from its rotational spectrum. Chem Phys Lett 255(1–
3):119–128. doi:10.1016/0009-2614(96)00340-5
orbital, donor-acceptor viewpoint. Chem Rev 88(6):899–926
113. Jeziorski B, Moszynski R, Szalewicz K (1994) Perturbation-theory approach to intermolec-
ular potential-energy surfaces of van-der-waals complexes. Chem Rev 94(7):1887–1930.
doi:10.1021/cr00031a008
114. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Montgomery
JA, Vreven T, Kudin KN, Burant JC, Millam JM, Iyengar SS, Tomasi J, Barone V, Mennucci
B, Cossi M, Scalmani G, Rega N, Petersson GA, Nakatsuji H, Hada M, Ehara M, Toyota
K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Klene M, Li
X, Knox JE, Hratchian HP, Cross JB, Adamo C, Jaramillo J, Gomperts R, Stratmann RE,
Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Ayala PY, Morokuma K, Voth GA,
Salvador P, Dannenberg JJ, Zakrzewski VG, Dapprich S, Daniels AD, Strain MC, Farkas O,
Malick DK, Rabuck AD, Raghavachari K, Foresman JB, Ortiz JV, Cui Q, Baboul AG, Clifford
S, Cioslowski J, Stefanov BB, Liu G, Liashenko A, Piskorz P, Komaromi I, Martin RL, Fox
DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A, Challacombe M, Gill PMW, Johnson
B, Chen W, Wong MW, Gonzalez C, Pople JA (2003) Gaussian 03, revision b.05. Gaussian,
Inc., Pittsburgh
115. Arman HD, Gieseking RL, Hanks TW, Pennington WT (2010) Complementary halogen and
hydrogen bonding: sulfur center dot center dot center dot iodine interactions and thioamide
ribbons. Chem Commun 46(11):1854–1856. doi:10.1039/b925710a
116. Li QZ, Jing B, Li R, Liu ZB, Li WZ, Luan F, Cheng JB, Gong BA, Sun JZ (2011) Some
measures for making halogen bonds stronger than hydrogen bonds in h2cs-hox (x = f, cl, and
br) complexes. Phys Chem Chem Phys 13(6):2266–2271. doi:10.1039/c0cp01543a
117. Emsley J (1998) The elements, 3rd edn. Clarendon Press, Oxford
118. Keith TA (2012) Aimall (version 12.11.09). TK Gristmill Software, Overland Park
Chapter 13
Understanding Lone Pair-π Interactions
from Electrostatic Viewpoint
Shridhar R. Gadre and Anmol Kumar
Abstract Over the last two decades, studies on lone pair-π interaction have attracted
lot of attention of experimental as well as theoretical chemists due to its intriguing
nature and its suspected presence in biological systems. The present Chapter be-
gins with a brief overview of the earlier theoretical and experimental work done in
this area. This is followed by exploration of the nuances of bonding in lone pair-π
interaction, employing the tool of molecular electrostatic potential (MESP) since
such weak interactions are mainly dominated by electrostatic features of host and
guest molecules. The critical points associated with the scalar field of MESP are
exploited for scrutinizing the directionality and bonding sites involved in the lone
pair-π complexes. Furthermore, the electrostatic potential for intermolecular com-
plexation (EPIC) model developed by Gadre et al., has been employed for finding
out the electrostatically optimized structures and interaction energies of these com-
plexes. The outcomes of EPIC model are compared with the results obtained from
quantum chemical calculations of the complexes employing M06L/6-311++G(d,p)
level of theory. The present study details out four different cases of lone pair-π
complexes, which are currently in vogue. Hexafluorobenzene, one of the most ex-
plored π-deficient host in the present context, is initially taken up to demonstrate
various facets of MESP for gaining insights into this interaction. This is followed by
the scrutiny of special classes of recently synthesized highly π-deficient molecules,
viz. tetraoxacalix [2]arene[2]triazine and naphthalenediimide, which are known to
have specificity and large affinity, respectively, towards the electron rich species.
The chapter ends with the description of lone pair-π interaction in the case of urate
oxidase, an enzyme present in biological systems.
S. R. Gadre () · A. Kumar

Department of Chemistry, Indian Institute of Technology Kanpur, Kanpur 208016, India
e-mail: gadre@iitk.ac.in
A. Kumar
e-mail: anmol@iitk.ac.in

392 S. R. Gadre and A. Kumar
13.1 Introduction
Supramolecular chemistry is generally governed by non-covalent forces. The non-

covalent interactions are known to play an indispensable role in ensuring functionality
of most of the chemical, physical and especially biological processes. Their relevance
is brought out by several examples, e.g. the existence of water in liquid form is solely
due to non-covalent interaction [1] viz. the hydrogen bond. Among the various types
of non-covalent interactions, there exists a category, which involves interactions
with aromatic rings such as in π-π, CH-π, cation-π interactions. Several biochemi-
cal phenomenon, either mechanical or cognitive such as protein folding, DNA/RNA
stacking, drug-receptor interactions etc. are the consequence of the harmonic in-
terplay between many such non-covalent interactions. Synthetic chemists exploit
such weak interactions in the field of biomimetics [2], crystal engineering, medicine
[3, 4], catalysts [5], material science etc. A recent addition to the aforementioned list
of non-covalent aromatic interactions is the interaction of π-deficient aromatic rings
with lone pair-containing species, popularly known as anion-π interaction.
A glimpse at this interaction may leave one amused as the aromatic ring is gener-
ally perceived to be endowed with a π electron cloud and the very idea of interaction
between lone pair and π cloud may seem antithetical. However, theoretical as well
as experimental investigations are suggestive of the attractive nature of this interac-
tion. Due to its rather mysterious nature and its possibility of existence in biological
systems, lone pair-π interaction has recently captured the attention of many theoret-
ical and experimental chemists. Being neutral in nature, π-deficient aromatic ring
also offers a promising alternative to cationic anion-receptors. Although some active
research on the anion-π interaction is being pursued, its cationic counterpart viz.
cation-π interaction is widely recognized as an important non-covalent interaction
by the scientific community since long [6–13]. The significant role of the latter in
biological systems can be depicted through many examples such as protein crystal
structures, enzyme catalysis, molecular signaling etc. to name a few. It has been ver-
ified that the nucleic acid bases in double-stranded DNA exhibit cation-π interaction
with amino acid side chains of proteins at the interface between protein and DNA.
In the recent past, several reviews on the anion-π interaction have been published
[14–21], providing quite comprehensive details of research done in this field. The
present Chapter is aimed at providing a brief summary of the earlier research done on
lone pair-π interactions and then scrutinize the nature of such interactions employing
the tools of molecular electrostatics and ab initio calculations to provide mechanistic
insights.
13 Understanding Lone Pair-π Interactions from Electrostatic Viewpoint 393
Before proceeding further, we would like to clarify the usage of the terms, “lone
pair-π interaction” and “anion-π interaction” in the present Chapter. The term
lone pair-π interaction seems more generalized than anion-π interaction due
to three reasons:
1. Every anion has at least one lone pair and the interaction with π-deficient
aromatic system generally occurs through the lone pair of anion.
2. Even the neutral molecules possessing lone pairs can interact with π-
deficient aromatic system.
3. The usage of the term “lone pair” brings in the sense of directionality of
the interaction.
In view of this, the present Chapter will employ both the terms viz. “lone
pair-π ” and “anion-π”, depending on the nature of the species involved.
Although the term “anion-π interaction” was first coined by Quiñonero et al. [22]
in 2002, its existence was already noted long back in 1973 by Boden et al. [23] in
X ray analysis of solid hexafluorobenzene (HFB) at 120 K. This showed fluorine
atom of one HFB molecule lying above the centroid of the other one in a unit cell
of monoclinic lattice. Later in 1980, Vrbancich et al. [24] experimentally verified
the quadrupole moment of hexafluorobenzene, benzene and other non-dipolar
aromatic molecules using electric field-gradient birefringence method. Unlike
benzene, which was found to possess a large negative quadrupole moment of (−33.3
± 2.1) × 10−40 Cm2 , hexafluorobenzene showed a highly positive quadrupole
moment of ( + 31.7 ± 1.7) × 10−40 Cm2 explaining the possible cause of interaction
between the fluorine atom of one HFB and the centroid of the other. Interestingly, the
actual experimental verification of existence of lone pair-π interaction was achieved
by Schneider et al. [25–27] in the early 90’s. NMR shifts indicated attractive
interaction between anions and organic neutral π-deficient arene systems, the
change in the free energy for which, was estimated (through force field calculations)
to be about 2 kJ mol−1 .
Research in this area till early nineties had been sporadic and accelerated only
after the simultaneous pioneering theoretical studies done by the groups of Alkorta
[28, 29] and Dougherty [30] in the late 1990’s. In their works, Alkorta and coworkers
[28, 29, 31] studied the ability of perfluoro aromatic rings to bind to electron donor
molecules/anions such as HF, :CH−2 , F− , Cl− , CN− . Since all the latter systems
exhibit diffuse electronic distribution, Alkorta et al. performed quantum chemical
calculations on them at HF, MP2 and DFT levels of theory employing Pople basis
sets with diffuse and polarization functions. The results confirmed the presence of
local minimum in PES for complexes formed between perfluoro aromatic rings and
electron donor molecules. For the systems studied by them, the interaction energy
was found to vary widely in the range of −1 to −20 kcal/mol depending on electron
donating tendency of anions/lone pair-containing molecule and π deficiency of the
aromatic ring. In a similar quest, Gallivan et al. [30] determined the attractive nature
of interaction between lone pair of water and HFB ring employing quantum chemical
calculations and the interaction energy was found to be quite comparable with that
of global minimum of water-benzene complex.
Significant contributions into understanding the nature of anion-π interaction, the
role of anions and π systems were made by Frontera et al. [22, 32–34]. Molecular
potential with polarization model (MIPp) was employed by them for separating out
components of interaction energy into electrostatic, classical dispersion-repulsion
and perturbationally derived polarization terms. Critical analysis of weak interac-
tions may also be carried out by employing symmetry adapted perturbation theory
(SAPT) [35, 36] which separates out physically meaningful components of a given
interaction viz. electrostatics, induction and dispersion interaction. To achieve this
decomposition, the corresponding Hamiltonian is partitioned into monomer Fock
operators, Møller-Plesset fluctuation operators and the intermolecular interaction
operators. It was first employed by Kim et al. [37] for scrutinizing the separate con-
tributions of interaction energy components in anion-π interaction. The optimized
geometries of weak complexes between anions (F− , Cl− , Br− , CN− , NC− , NO3 −
and CO3 −2 ) and π systems (tetrafluoroethene, hexafluorobenzene and 1,3,5-triazine)
were obtained [37] using supermolecular calculations. SAPT was then employed on
those optimized geometries using hybrid basis viz. 6-31G** for fluorine atoms in
perfluoroaromatic ring and 6-31++G* for rest of the atoms. The results showed
interaction energies of anion-π systems to be nearly comparable to those in cation-π
systems and the contributions from dispersion energies was found to be higher in
former, especially in the case of complexes with organic anions. However, in the
case of halide anions, the interactions were observed to be dominated by electro-
static and induction energies. Hence both MIPp and SAPT studies are indicative of
electrostatic dominance in anion-π interaction.
The pioneering work of specifically synthesizing complexes exhibiting anion-π
interaction was reported by Meyer and coworkers [38] in 2004. They synthesized
stable crystals of a complex wherein the chloride ions were found to attractively in-
teract with ligand containing electron deficient triazine ring. Crystallographic studies
showed that in a single molecule of the above mentioned compound, there exist two
triazine-containing ligands stacked parallely and held together by three Cu+2 ions
forming a carousel structure. Interestingly, the counteranions Cl− and CuCl−2 4 were
seen to lie ∼ 3.15 Å above the centroid of the two parallely stacked triazine rings.
Electron spray ionization (ESI) mass spectrum also showed the presence of two Cl−
ions interacting with the two triazine rings of the cationic part in methanol solution.
Later in the same year, Reedijk et al. [39] synthesized an octadentate dendritic

ligand, (N,N ,N ,N -tetrakis{2,4-bis(di-2-pyridylamino)-1,3,5-triazinyl}-1,4,8,11-
tetraazacyclotetradecane (azadendtriz)) the cationic tetranuclear copper moiety of
which could bind Cl− in a cavity created by four pyridine rings. The distance/s
between Cl− and pyridine ring/s varied in the range of 3.5 Å to 3.7 Å.
Interestingly, Cambridge Structural Database (CSD) search of many previously
synthesized compounds and biological systems reveals [22] the evidence of lone
pair-π interactions. As pointed out by Quiñonero et al. [22], a search of CSD
shows HOHJUJ structure originally synthesized by Frohn et al. [40] possessing
tetrafluoroborate anion (BF4 − ) with fluorine atoms facing aromatic ring of three
pentafluorophenyl moieties. It may be noted that Meyer and Reedijk [38, 39] suc-
ceeded in their above described work to synthesize compounds that distinctly show
anion-π interactions, but the species interacting with anions were cationic in nature.
The surplus positive charge could enhance the π deficiency of the π acidic rings
making the potential more positive, ensuring the high binding energy with anions.
The hope that a π-deficient aromatic ring could prove to be a promising anion
receptor has prompted the experimental chemists to practically achieve hosts that
are specific to anion reception. Designing and synthesizing anion receptors has been
a challenging field of research due to larger size of anions as compared to their
isoelectronic cationic counterparts, their wide range of shape and size and their pH
sensitivity [41]. The anion receptors generally bind anions through salt bridges, hy-
drogen bonds, ion pair or hydrophobic effects. The π aromatic ring adds to the
category of neutral anion receptors which ensures lesser sensitivity of host towards
pH and thus provides a wider pH window to remain operational for anion reception.
Although there has been increasing experimental evidence of neutral π-deficient
anion receptors in solid-state and solution, such an evidence for anion-π interac-
tion in gas phase seems to be conspicuous by its absence from the literature. In
a study carried out in solution phase, Rosokha et al. [42] have clearly illustrated
anion-π complex formation through their work on electron deficient olefinic and
aromatic centers interacting with halides in acetonitrile solution. Job’s method based
on continuous variation of absorbance with respect to mole fraction ensured 1:1 com-
plexation between tetracyanopyrazine ring (TCP) and Br− . New absorbance band
in UV-vis spectra at about 400 nm in addition to characteristic absorption band of
TCP (λ = 220–300 nm) confirmed the charge transfer in the complex. X-ray analy-
sis showed short halide-C contacts (∼ 0.4 Å shorter than the sum of respective van
der Waals radii) in crystal structure of salt-admixtured complexes formed by simple
electron deficient molecules viz. TCP with halides. All these analyses confirmed the
presence of anion-π interactions and the concomitant charge-transfer.
Due to the low G0 value and weak nature of interaction, anion-π interac-
tion is a poor candidate to bring specificity in anion receptors on its own, but
can pave a way to challenges of designing anion receptors assisted by the coop-
erative effects of other non-covalent interactions. However, the additive nature of
anion-π interactions with an increase in number of π-deficient rings may allow
the synthetic chemists to design hosts which could be specific and bind mainly
through anion-π interaction [43]. Li et al. [44] recently synthesized a triazine-
based host (Bis(phenylcarbamoyl)-functionalized tetraoxacalix[2]arene[2]triazine)
bearing H-bonding motifs and anion-π domains. This host could bind selectively to
H2 PO− 4 using both H-bonding and anion-π interaction. Spectroscopic titration of the
host with variety of tetrabutylammonium salt of anions was performed in acetonitrile
solution at the room temperature. The use of solvents with high dielectric constant
may increase the G0 value mainly due to ion pair formation with complex [45–48].
H2 PO− 4 showed selective quenching of fluorescence spectrum at ∼ 308 nm (signa-
ture of host system) and a formation of a new fluorescence emission band at nearly
412 nm, whereas all other anions such as Cl− , Br− , SCN− , NO3 − , BF4 − could not
vary either UV or fluorescence spectrum of the host.
Recently, synthetic chemists are constantly engaged in employing anion-π inter-

action to tune the fragile chemical pathways. One such example is the determination
of particular structural motif of Ag(I) complexes by modulating interaction between
anion and π-acidic aromatic rings during self-assembly reactions [49]. Matile and
Mareda [50] have come up with synthesis of a highly electron deficient class of
molecule, i.e. substituted naphthalenediimide (NDI) with the quadrupole moment
as high as +55.5 B. A whole new synthetic branch has emerged with the varied use
of NDI and its derivatives in terms of application to anions [50–54]. An example is
the use of oligomers of NDIs for making anion-π slides where the host possesses a
gradient of electron deficiency along its length and they have been synthetically ex-
ploited to transfer anions through membrane [50]. Another application of NDI shows
that the catalysis of the Kemp elimination is achieved by utilizing anion-π interac-
tions [51]. The features of molecular electrostatic potential (MESP) for NDI have
been briefly explored by Matile and coworkers [50]. In the later part of the present
Chapter, this unique set of molecules will be scrutinized through the topographical
features of MESP.
We conclude by quoting an example of anion-π interaction in biological sys-
tems which has put it at the center stage in the realm of non-covalent interactions.
In their recent work, Estarellas et al. [34] have reported experimental findings re-
trieved from Protein Data Bank (PDB) [55] unveiling relevant anion-π interaction
operative in urate oxidase enzyme (UOX). UOX is an antioxidant enzyme (absent in
humans) which oxidizes uric acid (π-deficient ring) into (S)-allantoin in the presence
of molecular oxygen. Although attempts to crystallize UOX with molecular oxygen
have been unsuccessful, its probable position is guessed by the location acquired by
CN− which inhibits UOX functionality in solution. Experimental results revealed
that CN− is located above the plane of uric acid (URC) in UOX/URC/CN− complex
wherein the distances of the carbon and nitrogen atoms of CN− from the mean plane
of the URC molecule are 3.2 Å and 3.0 Å respectively. All these experimental re-
sults were supported by quantum chemical calculations at a high level of theory and
sufficiently large basis, viz. RI-MP2/CBS. The interaction energy of uric acid and
CN− was found to be of the order of −17 kcal/mol. This case has been elaborated
in the forthcoming parts of the present Chapter in terms of justification to location
and orientation of anions above uric acid ring.
A critical understanding of the energetic and geometric features of lone pair-π
interactions is essential to the development of this field, particularly the science of
biomimetics and the development of new synthetic applications in catalysis, ma-
terials science, and medicine. The present chapter mainly focuses on the aspects
governing lone pair-π interactions. It seems from the search of literature that there is
a need to acquire better understanding of location and orientation of anions or lone
pair-containing molecules above the π-deficient ring. This Chapter is aimed at of-
fering an electrostatic view of the lone pair π interactions. The following subsection
summarizes the features of molecular electrostatic potential (MESP) and its critical
points.
13.2 Use of Molecular Electrostatic Potential for Probing Lone

Pair-π Interactions
We first elaborate the basic features of the scalar field of molecular electrostatic
potential. Molecular electrostatic potential (MESP) and its topography [56–65] is a
well-established tool for studying molecular reactivity [59, 66–73] (especially weak
interactions) and other chemical phenomena such as existence of lone pairs [74, 75],
aromatic π cloud [76], resonance and inductive effect [77] etc. The research in this
area was pioneered by Tomasi and Pullman where the MESP was exploited to under-
stand sites of electrophilic attack [56–58, 78]. The works of Politzer et al. [59, 66–71]
popularized the use of electrostatics in understanding various interactions between
molecules. Extensive work by Gadre et al. on electrostatics and its topography has
made the use of MESP more quantitative. MESP can be mathematically written as
sum of the nuclear and electronic potential, viz.

N
ZA ρ(r )d 3 r
V (r) = − (13.1)
A
|r − RA | |r − r |
where {ZA } are the nuclear charges centered at position {RA } and ρ(r ) is the contin-
uous electron density. Thus MESP can attain both positive and negative values which
ensure richer topographical features than electron density. The values of MESP can
be estimated theoretically using standard ab initio packages. Since MESP is electron
density derived quantity, its experimental determination is possible using X-ray crys-
tallography. Lecomte et al. have reported experimental techniques of determining
MESP in molecular crystals [79–82]. The features of topography of three dimen-
sional scalar field such as MED, MESP etc. is succinctly brought out in terms of
its critical points (CPs), an area of research extensively explored by Gadre et al.
[60–65, 83]. Critical points are locations, P in space, where all the first order partial
derivatives of MESP with respect to the coordinates vanish, i.e.

∂V ∂V ∂V
|∇V (r)|P = i +j +k =0 (13.2)
∂x ∂y ∂z P
Classification of these CP’s is done on the basis of signs of the eigenvalues of the
corresponding Hessian matrix at the CP. Hessian matrix is formed by second order
partial derivatives, the terms of which can be given as
2
∂ V (r)
Hij = (13.3)
∂ri ∂rj r=rCP
where V is the electrostatic field, ri and rj denote Cartesian coordinate i.e. x, y or z

and the Hessian is evaluated at any given critical point (CP). A CP is represented in
terms of R and S viz. (R, S) where R denotes the rank i.e. the number of non-zero
eigenvalues of the Hessian matrix and S denotes the signature i.e. algebraic sum of the
signs of these eigenvalues. If all the eigenvalues at a CP are non-zero, then such a CP
is called non-degenerate CP. On the other hand, if any one or more eigenvalue is zero,
the associated CP is called degenerate CP. Degenerate CPs are unstable and either
vanish with slightest conformational change in nuclear position or bifurcate into non-
degenerate CPs. There exist four distinct type of non-degenerate CPs for MESP viz.
(3, −3), (3, −1), (3, +1) and (3, +3). A (3, +3) CP is referred to as local minimum
as all three eigenvalues are positive. Negative-valued (3, +3) CP corresponds to a
point where the negative potential due to the electrons, dominates maximally over
the positive potential generated by the nuclei, thus indicating electron-dense region
in a molecule. The (3, +1) CP is a saddle point with two positive eigenvalues. All
chemical structures possessing topograhical rings show (3, +1) type of CP as their
signature. A (3, +1) CP also appears between two (3, +3) CPs as a topographical
artifact. The (3, −1) CP is another type of non-degenerate saddle, with two negative
eigenvalues. Two bonded atoms are marked by the presence of a (3, −1) CP. Lastly
a (3, −3) CP is a local maximum which explicitly represents a nucleus. Gadre et al.
[84] proved that non-nuclear maximum in MESP cannot exist. The mathematical
argument put for this condition is following: In the case of MESP, the eigenvalues
of the Hessian matrix of MESP satisfy the Poisson relation, given as Eq. 13.4
∇ 2 V (r) = (λ1 + λ2 + λ3 ) = 4πρ(r) (13.4)
However, for a non-nuclear maximum to exist, all the three eigenvalues and thus
their sum should be negative, whereas the right hand side of the above equation is
always non-negative. Due to non-existence of non-nuclear MESP maxima, Politzer
came up with the method of locating most positive potential on van der Waals surface
of the the molecule [67] so as to justify the nucleophillic attack.
MESP isosurfaces of a typical π-rich aromatic ring, viz. benzene and a π-deficient
system viz. hexafluorobenzene (HFB) are illustrated in Fig. 13.1, along with corre-
sponding MESP CPs. Benzene ring clearly shows negative regions of π electrons
spread above as well as below the ring plane (represented by blue MESP isosurface
in Fig. 13.1). On each side of the π cloud, a total of six (3, +3) CP’s (red dots) me-
diated by six (3, +1) CP’s (green dots) forms a ring of CPs with a (3, −1) CP located
at the center of the ring formed by these CPs. The typical MESP value at (3, +3)
CPs in aromatic π cloud of benzene is −18.7 kcal/mol. Hexafluorobenzene, on the
other hand, does not possess any such negative region on either side of the ring plane.
Rather, the lone pairs on fluorine atoms in HFB are disposed in the plane of the ring,
forming a zone of negative MESP around HFB. However, the MESP value associated
with the CP signifying the lone pair of fluorine is unusually low (−6.6 kcal/mol) as
compared to lone pair strength of fluorine in HF (−29.0 kcal/mol). Other commonly
used π-deficient systems [75] viz. s-triazine (STZ), trinitrobenzene (TNB), tetra-
cyanobenzene (TCB) and cyanuric acid (CUA) are displayed in Fig. 13.2. Similar to
HFB, absence of π electrons and presence of negative MESP only around the periph-
ery of the ring, are common features of these systems too. The complete absence of
the electrostatic π cloud in these molecules (such as HFB) is in conflict with the usual
connotation of the term “π” in “lone pair-π interaction”. This electrostatic picture
brings out the reason why electron rich species interact attractively with π-deficient
aromatic systems.
Fig. 13.1 Positive (red) and negative (blue) molecular electrostatic potential (MESP) isosurfaces
(isovalue indicated under the molecules) of benzene and hexafluorobenzene (HFB). Small red,
green and gray spheres denote (3,+3), (3,+1) and (3, −1) CPs respectively. See text for details
Fig. 13.2 Positive (red) and negative (blue) molecular electrostatic potential (MESP) isosurfaces
(isovalue indicated under corresponding molecules) for π-deficient host systems viz. tetracyanoben-
zene (TCB), trinitrobenzene (TNB), cyanuric acid (CUA) and s-triazine (STZ). Color coding of CPs
identical to that in Fig. 13.1. See text for details
Based on electrostatics and its topographical features, Gadre et al. [85, 86] devel-
oped a model several years’ ago, named as electrostatic potential for intermolecular
complexation (EPIC), which provides a simple tool for estimating the interaction
energy and geometry of the weakly interacting molecular species. In this model, the
interaction energies are calculated by summing the MESP driven point charges qA,i
of the guest molecule multiplied with the electrostatic potential value of the host at
the location of the point charge and vice versa.
⎧ ⎫
1 ⎨ ⎬
EEP I C = VA,i qB,i + VB,j qA,j (13.5)
2⎩ i j
⎭
where VA , VB , qA and qB are the MESP and the MESP-derived charges on the
interacting moieties A and B, respectively.
This interaction energy is minimized by rotating and translating the guest molecule
such that the van der Waals surfaces of the guest and the host do not interpenetrate. The
EPIC model was applied successfully to many systems such as DNA base dimers and
trimers [86], hydration of crown ether [87] and stacked and H-bonded cytosine dimer
[88]. Previous studies performed in our laboratory generally demonstrated striking
similarities between the geometries obtained from ab initio and EPIC calculations.
This model works quite well for the systems where the interaction energy has a large
contribution from electrostatic component including the cases of cation-π and anion-
π complexes. In the case of a system where induction and dispersion interactions
are predominant (such as benzene dimer), EPIC model does not well-capture the
energetics. However, the essence of directionality and orientation of host and guest
is still brought out quite reasonably by electrostatic model.
13.3 Methodology
The quantum chemical calculations of the systems reported in this Chapter are ex-
ecuted by employing M06L DFT functional subjected to 6-311++G(d,p) basis set
without any symmetry constraints. M06L is a local meta-GGA exchange correla-
tion functional especially recommended for studying non-covalent interactions [89].
Frequency calculations are performed for optimized structures to ensure if they corre-
spond to true minima on the corresponding potential energy surface. The Gaussian09
suite of programs [90] is employed for all the ab initio calculations. The calcula-
tions of interaction energies of the complexes are done considering the correction
for basis set superposition error (BSSE) by employing the Boys and Bernardi coun-
terpoise technique [91]. Wavefunctions and density matrices of all the systems are
extracted for topographical analysis at M06L(DFT) level of theory employing large
Pople basis viz. 6-311++G(d,p). However, it is an established fact that the nature
of the critical points, their location and the corresponding MESP value are nearly
constant beyond Hartree-Fock level of theory applied to reasonably large basis viz.
6-31G(d,p) [92]. The topographical analysis of MESP has been carried out on a large
variety of molecular systems comprising neutral molecules and anions. The location
and characterization of the MESP CPs is carried out employing the rapid topogra-
phy mapping Fortran code developed recently by Yeole et al. [93]. This code uses
a hierarchy of the scalar field viz. bare nuclear potential (BNP), molecular electron
density (MED) and MESP for building the topography i.e. critical points of BNP,
serve as guess points for generating MED CPs, which in-turn serve as guess points
for generating MESP CPs. To ensure that no CP is left unidentified in a given re-
gion, the guess points are enriched with randomly generated points inside a cube
around each nucleus. All these guess points are optimized using the L-BFGS code.
The topography of MESP is much richer than that of BNP and MED in terms of the
number of CPs. The underlying code for evaluation of MED, MESP and the corre-
sponding gradients is based on the deformed atoms in molecules (DAM) procedure
proposed by Rico et al. [94]. DAM method, based on partitioning of scalar field into
atomic contributions, provides a rapid and sufficiently accurate function and gradient
calculation.
All the EPIC calculations are performed using the integrated EPIC program re-
cently put together by Yeole and Gadre [95] which optimizes the starting geometry
into electrostatically most favorable geometry, such that the van der Waals surfaces
of the interacting species do not interpenetrate. Method of interaction energy calcu-
lation mentioned in the Eq. 13.5 is replaced by a simpler equation, as described in
the Eq. 13.6.

EEP I C = VA,i qB,i (13.6)
i
where VA is MESP of one of the interacting species (preferably electron rich species)
and qB are the MESP-derived point charges of the other. In the case of anion-π com-
plexes, the initial guess geometry is provided to EPIC model by placing the (3, +3)
CP of electron-rich anion/molecule over the π-deficient region of host molecule.
13.4 Results and Discussion
Several examples showing lone pair-π interactions have been briefly mentioned in
the Introduction Section. The results of four representative case studies of lone pair-π
interaction are discussed below from the electrostatic perspective. For this purpose,
we select hosts which are widely studied in the literature on this subject.
13.4.1 Hexafluorobenzene
HFB is one of the most commonly employed π-deficient systems in theoretical as

well as experimental studies for understanding lone pair-π interactions. As previously
discussed in Introduction Section, one of the reasons behind this molecule being in
vogue is its quadrupole moment, which is roughly equal to that of the benzene, but
opposite in sign. We limit ourselves to the investigation of complexation between
HFB and lone pair-containing neutral species and anions. In one of our previous
works [75] we also studied interaction between radicals and the π-deficient hosts.
Topographical features of hexafluorobenzene (cf. Fig. 13.1), reveal that the ring
centroid of HFB is electron deficient due to the electron withdrawing property of
fluorine atoms. This is the physical reason why the lone pair of a guest molecule
orients along the C6 axis of HFB.
We have analyzed the topographical features of several molecules including neu-
tral species and anions. These molecules/anions interact with π-deficient system
mostly through their electron rich regions, which are brought out by (3, +3) MESP
CPs. It has been observed in our earlier works [75] that the value of MESP at a
(3, +3) CP reflects the strength of the lone pair and it correlates remarkably well
with the interaction energy of the lone pair-π complex. This primarily indicates the
large electrostatic contribution to the energetics of such complexes.
Table 13.1 Relevant MESP topographical parameters and quantum chemically calculated inter-
action energies of neutral species and anions with Hexafluorobenzene. Dist implies distance of
CP (MESP minima) from lone pair-containing atom. D R and D S denotes degenerate ring and
degenerate spheres respectively. For explanation of the angles θ , θ and θ, see text and Fig. 13.4
System Nature and Dist MESP Eint θ θ θ
no. of CPs (Å) (kcal/mol) (kcal/mol) (o ) (o ) (o )
PH3 (3,+3), 1 1.88 −23.8 −2.1 123.3 123.9 0.6
Pyridine (3,+3), 1 1.28 −61.2 −4.3 121.6 121.5 −0.1
H2 O (3,+3), 2 1.24 −54.0 −3.1 119.8 124.3 4.5
H2 CO (3,+3), 2 1.29 −39.2 −2.7 131.3 174.0 42.7
O2 D R, 2 2.09 −1.8 −1.9a – – –
HCl D R, 1 1.90 −12.3 −1.7 – – –
NO−
3 (3,+3), 6 1.20 −167.7 −14.3 107.1 86.4 −20.7
H2 PO−
4 (3,+3), 4 1.19 −168.4 −15.0 124.2 93.7 −30.5
−
CN (3,+3), 2 1.23 −186.9 −14.5 180.0 85.8 −94.2
N−
3 D R, 2 1.25 −169.9 −15.9 – – –
−
F D S, 1 1.08 −240.3 −21.1 – – –
Cl− D S, 1 1.57 −171.0 −14.5 – – –
a
Eint for O2 is calculated without BSSE correction. See text for more details
Topographical features of some prototype examples of lone pair-containing

molecules and anions are summarized in Table 13.1 and Fig. 13.3. The inter-
action energies reported in Table 13.1 refer to the complexes of corresponding
molecules/anions with hexafluorobenzene ring. As an example, we elaborate the
case of H2 O wherein the two lone pairs on oxygen are characterized by two (3, +3)
CP, disposed tetrahedrally with respect to the hydrogen atoms in H2 O molecule. The
MESP value at both the CPs is −54.0 kcal/mol showing the electron rich nature of
the oxygen atoms. These two (3, +3) CPs are joined by an intervening (3, +1) CP
possessing comparable MESP value. It is a case where all the CPs are non-degenerate
in nature. However, there are certain molecular systems exhibiting degenerate CPs
in the electron rich region. This suggests a delocalized distribution of lone pair elec-
trons. For instance, Fig. 13.3 shows O2 , HCl and N3 − each endowed with degenerate
ring/s of the negative potential. Similarly, MESP topography of Cl− reveals a sphere
of negative minimum around the ion. Further, as may be expected, it turns out that
the MESP minima in anions are much deeper as compared to those in the neutral
species (see Table 13.1), making them better guests for lone pair-π interaction than
the latter ones. Among the anions presented in Table 13.1, fluoride ion (F− ) exhibits
the deepest minimum of value −240.3 kcal/mol.
A closer look at the nature of these CPs reveals the fact that the one of the three
eigenvalues associated with the (3, +3) CP is significantly larger compared to the
other two. For instance, in the case of H2 O, the three eigenvalues of the Hessian
matrix of second derivative of MESP are 0.0064, 0.0327 and 0.1304. In a recent
Fig. 13.3 Representation of neutral species and anions with the corresponding MESP CPs wrapped
around by MESP isosurfaces. Small red, green and gray spheres denote (3,+3), (3,+1) and (3, −1)
CPs respectively. See text for details
work [74], the authors have proposed that the largest among the three eigenvalues
enables one to distinguish a CP associated with lone pair from other CPs associated
with aromatic ring. Thus, the characterization of lone pairs using MESP topography
reveals the location and strength of the lone pairs. This information indeed provides
insights into lone pair-π interactions wherein the electron rich lone pair of the guest
molecule interacts with the electron deficient aromatic rings.
The complexation between HFB and the lone pair-containing species is scrutinized
below using the results obtained from MESP topography, EPIC as well as ab initio
calculations. Topographical analysis is performed on the ab initio optimized struc-
tures of HFB:X complex (X = lone pair-containing species). MESP topographical
features of these complexes may be analyzed from the following two perspectives.
Directionality of Lone Pair During Complexation During lone pair π complex-
ation, it is expected that the lone pair gets directed towards the electron deficient
region of the HFB ring. But there exists no signature of the interacting lone pair in
the complex, as the nature of (3, +3) CP present in isolated species changes when
Fig. 13.4 Representation of θ

and θ in a model lone pair-π
complex. Black dot represents
(3, +3) CP associated with
lone pair. See text for details
the host approaches the guest. The method employed by us for scrutinizing the ori-
entation of the lone pair [75] is demonstrated in Fig. 13.4. A comparison is drawn
between the spatial disposition of lone pair in the free molecules/anions to that in lone
pair-π complex. In the case of isolated molecule/anion containing a lone pair, the
angle made by the (3, +3) CP, atom possessing it and a nearby atom in the monomer
is represented by θ . Similarly, in the case of the complex, the angle made by the ring
centroid of the HFB, lone pair bearing atom and the nearby atom chosen to define
θ , is represented by θ . The change in orientation of lone pair upon the formation of
the complex is thus measured by calculating the difference between these two angles
(θ − θ ) designated as θ. Table 13.1 lists the parameters θ , θ and θ for all the
guest molecules studied here.
It is evident from above study that the molecules bearing only one lone pair, such
as pyridine and PH3 , orient their lone pair almost perfectly above the ring centroid
of HFB. Similarly, the monoatomic anions like F− and Cl− possessing spherically
degenerate ring CP and molecules like HCl and O2 with degenerate ring of CPs also
align themselves such that the region of lone pair falls directly on top of the centroid
of HFB ring. However, the cases wherein the guest molecule bears two or more
competing minima such as H2 CO and H2 PO− 4 , prefer off-center geometry over the
HFB ring so as to maximize the electrostatic interaction of minima with the positive
potential. The angle θ for such systems is found to be as large as 20–40o displaying
the fact the concerned species binds closer to the periphery of the ring of HFB. The
angle of 94.2o in case of CN− is due to two competing minima on the C-N axis, with
the latter lying parallel to the HFB plane. The off-center geometry is also preferred in
molecules which have hydrogen atoms such that it can form weak H-bond with the
fluorine atoms of HFB, thus distorting the usual expected orientation and maximizing
the attractive forces.
Since the lone pair plays a central role in lone pair-π interaction, its strength,
as provided by the MESP value, is expected to decide the interaction energy of
the complex. This is in affirmation of the fact that MESP value at the (3, +3) CP
Fig. 13.5 Plot of the interaction energies (Eint ) of the HFB:X complexes versus the MESP value
(V min ) at the (3, +3) CP which interacts with HFB. Set of points on upper end are related to anion-π
complexes, while the points located in the lower region of plot are associated with neutral species-π
complexes. See text for details
correlates extremely well with the interaction energy of the complex with correlation
coefficient R2 = 0.985. The species included in this plot are those given in Table 13.1,
augmented by few more molecules such as CH3 OH, O(CH3 )2 , N(CH3 )2 , imidazole,
Br− and CH3 CO− 2 . Figure 13.5 depicts the correlation between interaction energy
(Eint ) of HFB:X complex and the MESP value V min at the most negative lone pair
CP of the neutral molecules/anions. It is incisive from the plot that higher is the
MESP value, larger is the interaction energy. This suggests that the directionality
and interaction energy in the lone pair-π interaction is determined by the number,
location and strength of the lone pairs.
Appearance of New CPs Between the Host and the Guest The lone pair-π in-
teraction is marked by the appearance of new CP/s in the region connecting the van
der Waals extreme of the two interacting species. As pointed out in the Introduction
Section, a (3, −1) CP located between two interacting nuclei may be indicative of the
bond between them. Elaborating further, one of the eigenvector of this (3, −1) CP
corresponding to positive eigenvalue (directional minimum), points in the direction
of nuclei connected through that CP. Let us designate the eigenvector associated with
positive eigenvalue as PEV. Interaction between the lone pair and π-deficient system
can be rather viewed as an interaction of a MESP minimum of electron rich species
with a region of diffuse positive potential of electron deficient host. Experimentalists
have identified two different binding motifs in lone pair-π interaction, the one where
electron rich center is located near the centroid of the ring and other where it is
positioned above the periphery [42]. The latter is commonly called as σ interaction.
The appearance of MESP CPs located between the anion and the electron deficient
host is affirmative of the existence of weak interaction. Additionally, the direction of
PEV of these newly appeared CPs provides a succinct information about the bonding
motifs of the two species involved.
Especially in the case of lone pair-π interaction, where the CP does not necessarily
connect two nuclei, the newly appearing CPs can also be of (3, +1) in nature. Still,
the PEV at this CP points in the direction of the two topographical features of the
two interacting species. The RCP in HFB possesses an eigenvector associated with
negative eigenvalue which is aligned perpendicular to the ring plane. This provides
it the topographical property of directional maximum atleast in one direction. Figure
13.6 shows a representative set of complexes along with the newly appeared CPs
and the PEV of the corresponding CPs. In most of the complexes there exists a CP
of nature (3, −1), lying almost on the line joining electron rich region of neutral
species/anions to the RCP of HFB ring confirming the existence of lone pair-π
interaction. For example, a (3, −1) CP is developed when Cl− interacts with the
HFB ring (see Fig. 13.6), which is a signature of interaction of the anion with the
π-deficient host. The PEV points in the direction of Cl− and the RCP of HFB,
indicating that Cl− interacts with the ring centroid of HFB. On the other hand, the
cases where there are more than one competing minima, the CPs can be generated
elsewhere in the region connecting the two species e.g. HFB:CN− complex, where
interaction is marked by two (3, −1) CPs mediated by a (3, +1) CP (see Fig. 13.6).
The PEVs of the two CPs indicate that the nitrogen atom of CN− is interacting with
the RCP of HFB, whereas the carbon end of CN− interacts with peripheral carbon
atom of HFB. The introduction of negative charge on complexation with the anion,
especially monoatomic anion viz. Cl− atom brings out yet another intriguing feature
i.e. appearance of symmetric pattern of CPs resembling those in the π cloud of
benzene (see Fig. 13.6). These CPs are located on one side of the HFB ring opposite
to approach of anion, indicating the phenomenon of charge transfer in the complex.
The case of HFB:HCl is interesting because geometrical alignment of HCl over the
HFB ring puts an impression as if the Cl atom is directly interacting with the ring
centroid. However, the newly appeared CPs show that chlorine interacts with the
two oppositely located peripheral carbons of HFB. Thus the topographical analysis
of these complexes succinctly brings out the nature as well as different modes of
interaction between electron rich species and π-deficient system.
The partitioning between positive and negative electrostatic potential is indicated
by zero isosurface. Gadre et al. [83] have shown that the negative potential region of
the anions lies outside the zero isosurface encapsulating the positive nuclear potential.
This theorem can be further applied to HFB:anion complexes, as they carry extra
negative charge/s. Indeed the pattern of zero isosurface in HFB:anion complex can
be easily distinguished from that in HFB:neutral species complex (cf. Fig. 13.7), as
the former possesses zero isosurface enclosing both the anionic nuclear framework
and π-deficient host separately, leaving all the negative potential to lie outside this
surface.
Fig. 13.6 Representation of HFB:X (X = lone pair-containing species) complexes along with rel-
evant MESP CPs. Small red, green and gray spheres denote (3,+3), (3,+1) and (3, −1) CPs
respectively. CPs associated with fluorine lone pairs are removed for clarity. See text for details.
The arrows indicate the directions of eigenvectors (PEV)
Fig. 13.7 Representation of HFB:X (X = lone pair-containing species) complexes wrapped around
by zero valued isosurfaces of MESP. See text for details.
It is evident from above discussion that electrostatic forces mainly govern the ener-
getics and geometries the lone pair-π complexes. In view of this, EPIC calculations
are performed to obtain the electrostatically favored conformations and their interac-
tion energies. The EPIC optimized geometries serve as very good initial guess for the
follow-up ab initio calculations. Our previous work [75] on lone pair-π interaction
has demonstrated an excellent correlation of EPIC energies with the correspond-
ing ab initio energies. The comparison of magnitude of EEP I C with respect to Eint
shows that electrostatic component in HFB:X complexation is nearly 65 % of the
total interaction energy. To this end, it can be incisively concluded that electrostatic
effects play a dominant role in governing the energetics and structural orientation
of such complexes. However, the difference in EPIC and ab initio calculations is
indicative of the fact that other forces such as dispersion and induction forces may
play a significant role in lone pair-π interaction.
Fig. 13.8 Representation of a tetraoxacalix[2]arene[2]triazine (CTZ) b optimized structure of CTZ

along with its MESP critical points enveloped by an isosurface of value −12.5 kcal/mol. c MESP
textured on van der Waals surface of the molecule. See text for details
13.4.2 Tetraoxacalix[2]arene[2]triazine
A myriad of cationic anion receptors already exist, but experimental chemists have
always thrived for alternatives, especially for achieving selective recognition of
polyatomic anions. The fact that the lone pair-containing species interacts with the
π-deficient system through a weak non-covalent interaction, empowers the hosts
with the property of specificity and reversibility towards the anionic guest. Wang et
al. [96] have recently identified tetraoxacalix[2]arene[2]triazine (CTZ) to have the
ability to chelate structurally diverse anionic guests, by self tuning the shape of its
π-deficient cavity. In their experimental investigation, the interaction between vari-
ous polyatomic anions and tetraoxacalix[2]arene[2]triazine was studied both in the
solution and in solid state employing the tool of spectrometric titrations and X-ray
crystallography, respectively.
The present subsection provides insights into the bonding motifs involved in the
recognition of four different polyatomic anions of varying dimensionality, employ-
ing the tool of MESP. The host system i.e. tetraoxacalix[2]arene[2]triazine (CTZ)
and anionic systems viz. SCN− , BF− − −
4 , PF6 and NO3 are optimized using M06L
DFT functional with a sufficiently large basis i.e. 6-311++G(d,p). Moreover, the
coordinates of the complexes, obtained from crystal structure [96] are used for ab
initio calculation with a view to explain the nature of bonding therein. The optimized
structure of CTZ, along with its MESP CPs enveloped by negative valued MESP iso-
surface is illustrated in Fig. 13.8. Figure shows the three dimensional orientation of
the benzene and triazine rings in the host molecule. The two benzene rings are posi-
tioned parallel to each other, whereas the two oppositely located triazine rings tend
to form a V-shaped cleft. An intriguing feature of MESP is revealed by projecting
the negative MESP isosurface along with the CTZ molecule, showing an isolated
region of π electrons sandwiched between the two benzene rings. The π-deficient
nature of this V-shaped cleft is prominently brought out by mapping the MESP on
van der Waals surface of the molecule (see Fig. 13.8 c). The red region on the van
der Waals surface spread all over the large cleft shows highly positive surface po-
tential. This brings out the fact that unlike simpler systems such as HFB, the present
system can readily bind larger anions. Upon complexation with polyatomic anions,
the orientation of the rings adjust so as to accommodate varying sizes of the anions.
Figure 13.9 displays the topographical features of the four complexes viz.
CTZ:SCN− , CTZ:NO− − −
3 , CTZ:BF4 and CTZ:PF6 . A large number of newly gen-
erated CPs, in addition to individual CPs of host and the guest are quite noticeable
in this Figure. These newly generated CPs are indicative of the existence of non-
covalent bonding between CTZ and the anions. The dispersed distribution of these
CPs is suggestive of the fact that anions do not bind merely due to lone pair-π inter-
action, rather the whole cincture of the positive region in the cleft offers binding sites
to the anions. The discussion on the nature of bonding within the complexes is taken
up individually. In the case of CTZ:SCN− complex, the SCN− is aligned over CTZ
such that the nitrogen atom in SCN− lies 3.0 Å away from the centroid of one of the
triazine rings, while the sulphur atom at the other end is located 3.6 Å away from
another triazine moiety. Amongst the newly appeared CPs of CTZ:SCN− complex,
there exists a (3, −1) CP between the nitrogen and the triazine ring, indicative of the
lone pair-π interaction between them. One of the eigenvector of this CP, correspond-
ing to positive eigenvalue, reveals the off-center weak σ interaction of nitrogen with
the π-deficient ring. Another (3, −1) CP can be observed between the sulphur atom
of SCN− and the second triazine ring, the eigenvector of which, indicates interaction
with a peripheral carbon of the triazine ring. Nonetheless the positive eigenvalue
associated with the eigenvector is very low (∼ 0.06 au) indicating this interaction to
be extremely weak. Another significant attractive interaction is provided by the two
hydrogen atoms located on the two benzene rings.
CTZ:NO− 3 complex distinctly shows a (3, −1) CP connecting one of the oxygen’s
of NO− 3 and centroid of the triazine ring. However, the other two oxygen atoms of
NO− 3 do not show any signature of anion-π interaction with the second triazine ring.
Nevertheless the other newly appeared saddle points show a dispersed electrostatic
interaction with the bed of positive potential lying underneath the negative potential
of NO− 3.
The lone pair-π interaction is exquisitely brought out in CTZ:BF− 4 complex, where
the interaction between two fluorine of BF− 4 and the two triazine rings are represented
by the two (3, −1) CPs connecting them. A closer look at the eigenvectors associated
with the two CPs reveals that one of the fluorine atom interacts with the centroid
of the triazine ring, while the other one possesses an off-center σ interaction with
another triazine ring. Weaker interaction with hydrogen atoms of benzene ring is
also brought out by the mediating (3, −1) CPs.
In the case of CTZ:PF− −
6 complex, the PF6 is positioned symmetrically in the
cavity formed by two triazine rings. Interestingly, the significant interaction to hold
the anion is provided only by the two hydrogen atoms of the benzene ring. However,
the anion-π interaction is very weak and does not contribute significantly to the total
interaction energy.
The discussion on bonding motifs present in CTZ:anion complex can be con-
cluded by stating that anions bind to the cavity of tetraoxacalix[2]arene[2]triazine
molecule, not just through anion-π interaction but also utilizing complementary
positive potential stretched along the cavity.
Fig. 13.9 Optimized structures of CTZ:X complexes along with its MESP critical points. Small
red, green and gray spheres denote (3,+3), (3,+1) and (3, −1) CPs respectively. Only those CPs
indicative of interaction are retained for clarity. See text for details. The arrows indicate the directions
of eigenvectors (PEV)
13.4.3 Naphthalenediimide
Naphthalenediimide (NDI) and its derivatives are currently getting widespread atten-
tion of theoretical as well as experimental chemists due to the exceptional π-deficient
nature of the former. Matile and Mareda [50–54] have recently reported the synthesis
of few molecules of this class such as N-tetramethylphenyl 2,6-dicyano-NDI and N-
tetramethylphenyl −2,3,6,7-tetracyano-NDI, with spectacular quadrupole moment
values of +39.2 B and +55.5B, respectively. They suggested in their work, the use
of linear chain of these molecules to design anion-π slides, which can selectively
transfer the anions. Use of such anion-π slide to create electroneutral photosystems
have also been explored by them.
The large quadrupole moments of these molecules can be attributed to cyano
groups attached to NDI scaffold which drastically pulls out the π electrons. Addi-
tionally, the tetramethylphenyl ring attached in plane perpendicular to the plane of
NDI contributes locally to the positive quadrupole moment. Figure 13.10 shows the
structure and the highly positive electrostatic potential textured on the van der Waals
Fig. 13.10 a Structure of N-tetramethylphenyl −2,3,6,7-tetracyano-Naphthalenediimide (TNDI),

b MESP values textured on van der Waals surface of the molecule. Arrows indicate the location of
most positive potential on the surface. See text for details
surface of N-tetramethylphenyl −2, 3, 6, 7-tetracyano-NDI (TNDI). The potential

at the two most positive locations on the surface is as high as ~66kcal/mol, marked
by black arrows in the Figure.
The quantum chemical calculations have been performed herein at similar level of
theory and basis as discussed in Methodology Section. The initial guess geometries
for the ab initio optimization are obtained from EPIC calculations, which provide
electrostatically most favorable geometries of the complexes. Following discussion
brings out the accuracy of EPIC method achieved in predicting the structure and
energetics of the complex formed between TNDI and lone pair-containing species.
The structures obtained from both EPIC and ab initio calculations are illustrated in
Fig. 13.11. The gross similarity of the structures obtained by the two methods is
indicative of the electrostatic dominance of energy in the interaction of lone pair-
containing species with TNDI. Most of the species studied here can be observed
to align their lone pair towards either of the two positive MESP location over the
π-deficient surface. Figure 13.12 shows the correlation between the interaction en-
ergies of TNDI:X (X = lone pair-containing species) obtained from EPIC and ab
initio calculations. Anions like HCO− − −
2 , NO3 and H2 PO4 show numerically large in-
teraction energy of −40 kcal/mol, higher than strongest H-bond. Nonetheless these
polyatomic anions interact through more than one lone pair, thus additive effect of
lone pair-π interaction comes into play. The average electrostatic contribution to the
total interaction energy is nearly 70 %, which justifies the use electrostatics as a tool
for studying anion-π interaction.
13.4.4 Anion-π Interaction in Urate Oxidase
To the best of our knowledge, present example is the first clear experimental verifi-
cation of presence of lone pair-π interaction in biological systems. The active center
of urate oxidase possibly carries out oxidation of uric acid employing lone pair-π
interaction with molecular oxygen as suggested by Estarellas et al. [34]. This exam-
ple has been briefly introduced in previous part of Chapter highlighting the essential
Fig. 13.11 Comparative representation of EPIC and ab initio optimized structures of TNDI:anion
complexes. See text for details
features. Figure 13.13 shows the enzymatic cascade reaction for uric acid oxidation
in UOX, as proposed in the literature.
Uniqueness of UOX activity lies in the fact that unlike other oxidase enzymes,
it does not require any metal atom or organic co-factor for catalysis. The sub-
strate (molecular oxygen) is expected to be aligned over the π-deficient ring of
URC through anion-π interaction. As mentioned earlier, the inference of conforma-
tion of UOX/URC/O2 is drawn indirectly [34] using the X-ray crystal structure of
UOX/URC/CN− .
In what follows, we investigate the bonding motifs and the nature of interaction
involved in uric acid oxidation inside UOX by essentially carrying out MESP analysis
of the host, guest and complex. Although the main substrate which binds to uric acid
Fig. 13.12 Correlation between the quantum chemically calculated interaction energy (Eint ) and
EPIC (EEP I C ) energy for TNDI:X complexes. See text for details
Fig. 13.13 Enzyme cascade for uric acid oxidation in UOX
in UOX is molecular oxygen, the other lone pair-containing anions viz. CN− and N3 −
are known to inhibit the activity of UOX. As evident from Fig. 13.3, all these three
systems do possess similar topographical features, i.e. existence of two competing
minima at both ends, on the internuclear axis. The O2 molecule shows two rings of
degenerate CPs with a very low MESP value of −1.8 kcal/mol, whereas the anions,
CN− and N3 − possess very deep MESP minimum of value −186.9 kcal/mol and
−169.9 kcal/mol respectively. The relative strength of negative ESP of latter species
provides enough reason why they could possibly bind to π-deficient surface of URC
replacing O2 and thus inhibit the UOX activity.
The MESP features of electron deficient counterpart of the complex i.e. uric acid
are reflected by mapping the MESP values on the van der Waals surface of URC as
illustrated in Fig. 13.14. The electrostatic potential mapping of URC shows that the
most positive values lie in the molecular plane near the α hydrogen of URC, still the
Fig. 13.14 a Structure of uric acid (URC), b projection of MESP values on van der Waals surface
of uric acid. Arrows indicate the location of most positive potential above the plane. See text for
details
guest molecule prefers to align on top of the ring (apparent from X ray structure).
This is due to the fact that the molecular plane is occupied by amino acids that leaves
the region above the molecular plane available for interaction with negative charges.
Apart from the most positive potential located near α hydrogen, the location of next
most positive potential are above the ring as indicated in Fig. 13.14 by black arrows.
The topographical analysis of the URC/CN− complex is performed in order to
probe the bonding motifs involved in the complexation. Figure 13.15 illlustrates
the geometrical alignment of CN− over the uric acid along with MESP CPs of
URC/CN− complex. Both nitrogen and carbon are roughly positioned above the two
ring centroids of uric acid molecule. A slight tilt of CN− with respect to URC plane is
due to the difference in the strength of the two minima, where the minima associated
with nitrogen atom being deeper, causes the latter to get closer to the uric acid ring.
As can be observed, the two newly generated (3,+1) CPs between CN− and uric acid
confirms the existence of non-covalent bonding between the two. The direction of
PEVs at the corresponding (3, +1) CPs reveals that both carbon and nitrogen strongly
interact with the two RCPs of the uric acid molecule. Since the topographical features
of O2 molecule bears compelling qualitative similarity with CN− , it is expected to
bind to uric acid in a similar fashion, but with a weaker interaction energy. This is in
accordance with the fact that O2 helps in enzymatic process of the enzyme cascade,
while CN− ion is the inhibitor. Thus, the electrostatic features incisively emphasizes
the evidence of existence of lone pair π interaction operational in the active center
of urate oxidase.
13.5 Concluding Remarks
For more than a decade now, the intriguing nature of lone pair-π interaction has cap-
tivated experimental as well as theoretical chemists. Several theoretical studies have
employed quantum chemical calculations on complexes formed between electron
rich species and model π-deficient hosts such as hexafluorobenzene and s-triazine
Fig. 13.15 The optimized

geometry of URC/CN−
complex along with its MESP
CPs. Small red, green and
gray spheres denote (3,+3),
(3,+1) and (3, −1) CPs
respectively. The arrows
indicate the direction of
eigenvectors (PEV). Only
those CPs indicative of
interaction are retained for
clarity. See text for details
in order to verify the attractive nature of the interaction. Experimental chemists have
been successful in synthesizing stable complexes which exhibit lone pair-π interac-
tion. Even the long existing skepticism on its biological role [42, 50] was cleared
with the study on urate oxidase enzyme, which distinctly reveals the importance of
such interaction in carrying out the cascade reaction at its active center. The methods
employed to understand such interaction have been based on simpler criteria such
as contact distances between the atoms of the interacting species, mapping of elec-
tron density followed up by atoms in molecules (AIM) analysis. Attempts using the
tool of electrostatic potential have also been explored in the literature, but their use
is mainly limited to depict the complementarity of the species involved. In one of
our recent works [75],we have used electrostatics in wider sense i.e. to understand
MESP features of host and guest in terms of CPs and their role in determining the
conformation of the complexes.
In the present chapter, we have elaborated the role of electrostatics for scrutinizing
the nature of interaction and the bonding motifs involved in the interaction. We could
successfully explain the scenario of lone pair-π interaction in certain experimentally
determined complexes, by analyzing the MESP CPs generated due to complexation.
Literature reports that the electron rich species binds either to the ring centroid or
the periphery of π-deficient rings, the latter disposition being known in the literature
as σ interaction. This fact is succinctly brought by the eigenvector (PEV) associated
with the CP located between the interacting species. Molecules possessing one lone
pair are usually observed to bind to RCP of the host, as evident from the PEV at
the intermediate (3, −1) CP, whereas cases where more than one competing minima
exist in the molecule, tend to lie at peripheral ends so as to minimize the interaction
energies. EPIC model has been employed to predict the electrostatically favorable
structures as well as to estimate the electrostatic contribution to total interaction
energies in such complexes. The resemblance of structures obtained from EPIC
model and ab initio calculations clearly brings out the dominance of electrostatics in
the formation of lone pair-π complexes. However, subtle differences are indicative
of the role of dispersion and inductive forces in tuning the final conformation of the
complexes. In summary, the essence of lone pair-π interaction is vividly brought
out by the molecular electrostatic potential and its critical points, proving it to be a
valuable tool for this purpose.
Acknowledgements Authors are thankful to Dr. C. H. Suresh and Dr. P. Balanarayan for fruit-
ful discussions. Anmol Kumar thanks the Council of Scientific and Industrial Research (CSIR)
for research fellowship. Professor Shridhar Gadre is grateful to the Department of Science and
Technology (DST), New Delhi for the award of J. C. Bose National Fellowship.
References
1. Némethy G, Scheraga HA (1962) J Chem Phys 36:3382–3400

2. Tzalis D, Tor Y (1996) Tetrahedron Lett 37:8293–8296
3. Müller-Dethlefs K, Hobza P (2000) Chem Rev 100:143–168
4. Duncan R, Kopec̆ek J (1984) Soluble synthetic polymers as potential drug carriers. In: Polymers
in medicine, vol. 57. Springer, Berlin
5. Huck WTS, Prins LJ, Fokkens RH, Nibbering NMM, van Veggel FCJM, Reinhoudt DN (1998)
J Am Chem Soc 120:6240–6246
6. Zhong W, Gallivan JP, Zhang Y, Li L, Lester HA, Dougherty DA (1998) Proc Nat Acad Sci
USA 95:12088–12093
7. Mecozzi S, West AP, Dougherty DA (1996) J Am Chem Soc 118:2307–2308
8. McCurdy A, Jimenez L, Stauffer DA, Dougherty DA (1992) J Am Chem Soc 114:10314–10321
9. Wheeler SE, Houk KN (2009) J Am Chem Soc 131:3126–3127
10. Sayyed FB, Suresh CH (2012) J Phys Chem A 116:5723–5732
11. Sayyed FB, Suresh CH (2012) Chem Phys Lett 523:11–14
12. Sayyed FB, Suresh CH (2011) J Phys Chem A 115:9300–9307
13. Mahadevi AS, Sastry GN (2013) Chem Rev 113:2100–2138
14. Egli M, Sarkhel S (2007)Acc Chem Res 40:197–205
15. Caltagirone C, Gale PA (2009) Chem Soc Rev 38:520–563
16. Wenzel M, Hiscock JR, Gale PA (2012) Chem Soc Rev 41:480–520
17. Gale PA, Busschaert N, Haynes CJE, Karagiannidis LE, Kirby IL (2014) Chem Soc Rev
43:205–241
18. Schottel BL, Chifotides HT, Dunbar KR (2008) Chem Soc Rev 37:68–83
19. Chifotides HT, Dunbar KR (2013) Acc Chem Res 46:894–906
20. Frontera A, Gamez P, Mascal M, Mooibroek TJ, Reedijk J (2011) Angew Chem Int Ed 50:
9564–9583
21. Mooibroek TJ, Black CA, Gamez P, Reedijk J (2008) Cryst Growth Des 8:1082–1093
22. Quiñonero D, Garau C, Rotger C, Frontera A, Ballester P, Costa A, Deyà PM (2002) Angew
Chem Int Ed 41:3389–3392
23. Boden N, Davis P, Stam C, Wesselink G (1973) Mol Phys 25:81–86
24. Vrbancich J, Ritchie GLD (1980) J Chem Soc Faraday Trans 2 76:648–659
25. Schneider H-J, Werner F, Blatter T (1993) J Phys Org Chem 6:590–594
26. Schneider H-J, Blatter T, Palm B, Pfingstag U, Ruediger V, Theis I (1992) J Am Chem Soc
114:7704–7708
27. Schneider H-J (1991) Angew Chem Int Ed 30:1417–1436
28. Alkorta I, Rozas I, Elguero J (1997) J Org Chem 62:4687–4691
29. Alkorta I, Rozas I, Elguero J (2002) J Am Chem Soc 124:8593–8598
30. Gallivan JP, Dougherty DA (1999) Org Lett 1:103–106
31. Alkorta I, Elguero J (2003) J Phys Chem A 107:9428–9433
32. Garau C, Frontera A, Quiñonero D, Ballester P, Costa A, Deyà PM (2003) Chem Phys Chem
4:1344–1348
33. Estarellas C, Frontera A, Quiñonero D, Deyà PM (2011) Chem Asian J 6:2316–2318
34. Estarellas C, Frontera A, Quiñonero D, Deyà PM (2011) Angew Chem Int Ed 50:415–418
35. Jeziorski B, Moszynski R, Szalewicz K (1994) Chem Rev 94:1887–1930
36. Lao KU, Herbert JM (2014) J Chem Phys 140:044108–8
37. Kim D, Tarakeshwar P, Kim KS (2004) J Phys Chem A 108:1250–1258

38. Demeshko S, Decher S, Meyer F (2004) J Am Chem Soc 126:4508–4509
39. de Hoog P, Gamez P, Mutikainen I, Turpeinen U, Reedijk J (2004) Angew Chem Int Ed
43:5815–5817
40. Frohn H, Giesen M, Welting D, Henkel G (1996) Eur J Solid State Inorg Chem 33:841–853
41. Sessler JL, Gale PA, Cho W-S (2006) Anion receptor chemistry, vol 8. Royal Society of
Chemistry
42. Rosokha YS, Lindeman SV, Rosokha SV, Kochi JK (2004) Angew Chem Int Ed 43:4650–4652
43. Müller M, Albrecht M, Gossen V, Peters T, Hoffmann A, Raabe G, Valkonen A, Rissanen K
(2010) Chem Eur J 16:12446–12453
44. Li S, Wang D-X, Wang M-X (2012) Tetrahedron Lett 53:6226–6229
45. Arranz-Mascarós P, Bazzicalupi C, Bianchi A, Giorgi C, Godino-Salido M-L, Gutiérrez-Valero
M-D, Lopez-Garzón R, Savastano M (2013) J Am Chem S 135:102–105
46. Giese M, Albrecht M, Krappitz T, Peters M, Gossen V, Raabe G, Valkonen A, Rissanen K
(2012) Chem Commun 48:9983–9985
47. Ballester P (2013) Acc Chem Res 46:874–884
48. Gil-Ramírez G, Escudero-Adán EC, Benet-Buchholz J, Ballester P (2008) Angew Chem
120:4182–4186
49. Schottel BL, Chifotides HT, Shatruk M, Chouai A, Pérez LM, Bacsa J, Dunbar KR (2006) J
Am Chem Soc 128:5895–5912
50. Mareda J, Matile S (2009) Chem Eur J 15:28–37
51. ZhaoY, DomotoY, Orentas E, Beuchat C, Emery D, Mareda J, Sakai N, Matile S (2013) Angew
Chem Int Ed 52:9940–9943
52. Gorteau V, Julliard MD, Matile S (2008) J Membr Sci 321:37–42
53. Vargas Jentzsch A, Emery D, Mareda J, Metrangolo P, Resnati G, Matile S (2011) Angew Chem
Int Ed 50:11675–11678
54. Lin N-T, Vargas Jentzsch A, Guenee L, Neudorfl J-M, Aziz S, Berkessel A, Orentas E, Sakai
N, Matile S (2012) Chem Sci 3:1121–112
55. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne
PE (2000) Nucleic Acids Res 28:235–242
56. Scrocco E, Tomasi J (1973) The electrostatic molecular potential as a tool for the interpretation
of molecular properties. In New concepts II, vol 42. Springer, Berlin, pp 95–170
57. Tomasi J, Mennucci B, Cammy M (1996) Molecular electrostatic potentials: concepts and
applications. Elsevier, Amsterdam
58. Tomasi J, Mennucci B, Cammi R (1996) Theor Comp Chem 3:1–103
59. Politzer P, Murray JS (2002) Theor Chem Acc 108:134–142
60. Yeole SD, Gadre SR (2011) J Phys Chem A 115:12769–12779
61. Balanarayan P, Kavathekar R, Gadre SR (2007) J Phys Chem A 111:2733–2738
62. Gadre SR, Shirsat RN (2000) Electrostatics of atoms and molecules. Universities Press,
Hyderabad
63. Gadre SR, Kulkarni SA, Shrivastava IH (1992) J Chem Phys 96:5253–5260
64. Balanarayan P, Gadre SR (2003) J Chem Phys 119:5037–5043
65. Shirsat RN, Bapat SV, Gadre SR (1992) Chem Phys Lett 200:373–378
66. Politzer P, Murray JS, Peralta-Inga Z (2001) Int J Quantum Chem 85:676–684
67. Politzer P, Landry SJ, Waernheim T (1982) J Phys Chem 86:4767–4771
68. Murray JS, Lane P, Politzer P (2007) Int J Quantum Chem 107:2286–2292
69. Politzer P, Murray JS, Lane P (2007) Int J Quantum Chem 107:3046–3052
70. Shields ZP, Murray JS, Politzer P (2010) Int J Quantum Chem 110:2823–2832
71. Politzer P, Murray JS, Clark T (2010) Phys Chem Chem Phys 12:7748–7757
72. Suresh CH, Koga N, Gadre SR (2000) Organometallics 19:3008–3015
73. Elango M, Subramanian V, Rahalkar AP, Gadre SR, Sathyamurthy N (2008) J Phys Chem A
112:7699–7704
74. Kumar A, Gadre SR, Mohan N, Suresh CH (2014) J Phys Chem A 118:526–532
75. Mohan N, Suresh CH, Kumar A, Gadre SR (2013) Phys Chem Chem Phys 15:18401–18409
76. Suresh CH, Gadre SR (2007) J Phys Chem A 111:710–714

77. Suresh CH, Alexander P, Vijayalakshmi KP, Sajith P, Gadre SR (2008) Phys Chem Chem Phys
10:6492–6499
78. Pullman B (1990) Int J Quantum Chem 38:81–92
79. Pichon-Pesme V, Lecomte C (1998) Acta Crystallogr Sect B 54:485–493
80. Benabicha F, Pichon-Pesme V, Jelsch C, Lecomte C, Khmou A (2000) Acta Crystallogr Sect
B 56:155–165
81. Bouhmaida N, Ghermani N-E, Lecomte C, Thalal A (1997) Acta Crystallogr Sect A 53:556–563
82. Bouhmaida N, Thalal A, Ghermani Ne, Lecomte C (1999) Acta Crystallogr Sect A 55:729–738
83. Gadre SR, Shrivastava IH (1991) J Chem Phys 94:4384–4390
84. Gadre SR, Pathak RK (1990) Proc Ind Acad Sci (Chem Sci) 102:189–192
85. Gadre SR, Pundlik SS (1997) J Phys Chem B 101:3298–3303
86. Pundlik SS, Gadre SR (1997) J Phys Chem B 101:9657–9662
87. Pingale SS, Gadre SR, Bartolotti LJ (1998) J Phys Chem A 102:9987–9992
88. Sivanesan D, Babu K, Gadre SR, Subramanian V, Ramasami T (2000) J Phys Chem A
104:10887–10894
89. Zhao Y, Truhlar DG (2006) J Chem Phys 125:194101–194118
90. Frisch MJ et al. (2009) Gaussian09 Revision D.01. Gaussian Inc. Wallingford
91. Boys S, Bernardi F (1970) Mol Phys 19:553–566
92. Gadre SR, Kulkarni SA, Suresh C, Shrivastava IH (1995) Chem Phys Lett 239:273–281
93. Yeole SD, López R, Gadre SR (2012) J Chem Phys 137:074116–074117
94. Rico JF, López R, Ramírez G, Ema I, Ludeña EV (2004) J Comput Chem 25:1355–1363
95. Yeole SD, Gadre SR (2011) J Chem Phys 134:084111–084118
96. Wang D-X, Wang M-X (2013) J Am Chem Soc 135:892–897
Part III
Aromatic Systems
Chapter 14
Unraveling the Origin of Substituents Effects
in π-Stacking Interactions
Steven E. Wheeler
Abstract Non-covalent interactions involving aromatic rings, which include π-

stacking, anion-π, and cation-π interactions, among others, are central to modern
chemical research. These interactions play vital roles in everything from protein-
DNA interactions and the properties of organic electronic materials to stereoselective
organocatalyzed reactions. We discuss recent efforts to understand the impact of sub-
stituents on the strength of π-stacking interactions through the application of modern
tools of computational quantum chemistry. We first provide an account of previous
efforts to develop qualitative physical models of these interactions, followed by ap-
parent flaws in these previous models. We then present our local, direct interaction
model of substituent effects in π-stacking interactions, and discuss recent extensions
of this model based on the examination of electric fields of arenes. We also discuss
related misconceptions regarding molecular electrostatic potentials (ESPs), and of-
fer a simpler view of the origin of ESP differences arising from the incorporation of
heteroatoms or substituents into aromatic rings. Finally, we conclude with an outlook
for future advances in our understanding of π-stacking interactions.
14.1 Introduction
Non-covalent interactions involving aromatic rings play vital roles in a myriad

of contexts ranging from supramolecular chemistry and crystal engineering to
organocatalysis [1–5]. Understanding the factors that impact both the strength and ge-
ometry of these non-covalent interactions has proved vital to harnessing their power
in practical applications. π-stacking interactions are traditionally defined as attractive
non-covalent interactions between aromatic rings (see Fig. 14.1), and their existence
has been known for many years [6]. These interactions play key roles in many areas
of chemistry and molecular biology. For example, π-stacking interactions between
DNA bases are the primary stabilizing factor in double-helical DNA, [7–10] and are
also responsible for the intercalation of drugs and various carcinogens into the DNA
S. E. Wheeler ()
Department of Chemistry, Texas A&M University,
College Station, TX 77843, USA
e-mail: wheeler@chem.tamu.edu
422 S. E. Wheeler
Fig. 14.1 Prototypical configurations of the benzene dimer. Of these, only the parallel-displaced
and T-shaped dimers are energy minima. Both the parallel displaced and T-shaped dimers are bound
by about 2.8 kcal mol−1 in the gas phase, whereas the sandwich dimer is bound by 1.8 kcal mol−1
[17–18]
double helix [9, 11, 12]. These interactions play central roles in many other areas
of chemistry, including the packing of polycyclic systems in organic electronic ma-
terials. Finally, in the last few years, there has been increasing attention afforded
to organocatalytic transformations in which π-stacking and other non-covalent
interactions play key roles in both catalytic activity and selectivity [13–16].
There has been some discussion recently [19–21] regarding the preferred name for
π-stacking interactions, which are alternatively referred to as “π-π interactions”, “π-
π stacking interactions”, “aromatic stacking interactions”, “aromatic donor-acceptor
interactions”, “charge-transfer interactions”, or simply “stacking interactions”. Each
of these terms is loaded with connotations, and the existence and widespread use
of these disparate terms reflects the storied history of our understanding of these
non-covalent interactions. Along these lines, we note that even though π-stacking
interactions are often defined as occurring between aromatic molecules, there is
growing evidence that aromaticity plays no direct role in these interactions. For
example, in 2008, Grimme [19] showed that, at least for aromatic systems smaller
than anthracene, there does not appear to be anything “special” about aromatic π-
π interactions. More recently, Bloom and Wheeler [20] showed more directly that
aromatic π-electron delocalization actually hinders π-stacking interactions in model
complexes. Finally, Martinez and Iverson [21] recently reviewed the experimental
and theoretical literature on π-stacking interactions, concluding that the term “π-
stacking interaction” does not accurately describe the forces responsible for the
association of most aromatic molecules. Ultimately, they suggested that the terms “π-
stacking interaction” and “π-π stacking interaction” should be avoided altogether!
We view “stacking interactions” as the safest choice of terminology for these non-
covalent interactions, since this term can be used purely as a geometric descriptor.
That is, “stacking” invokes a relatively clear physical picture of planar molecules in
roughly parallel, overlapping orientations, devoid of the added implications of the
14 Unraveling the Origin of Substituents Effects in π-Stacking Interactions 423
other widely-used terms. However, in the present work we will stick with the more
popular moniker “π-stacking interaction”, with the caveat that the “π-” refers to the
fact that the stacking interactions we are discussing involve aromatic “π-systems.”
14.2 Computational Methods for Studying π-Stacking

Interactions
The study of non-covalent interactions involving aromatic rings, including π-

stacking interactions, provides a number of challenges for standard electronic
structure methods. In particular, gas-phase π-stacking interactions are dominated
by dispersion effects whose description requires robust correlated ab initio methods
paired with large basis sets [17, 18, 22, 23]. Previously, this requirement limited
computational studies of non-covalent interactions involving aromatic rings to a
small model complexes containing 10–20 non-hydrogen atoms [4, 23]. However,
recent years have witnessed an explosion in the development of density functional
theory (DFT) methods capable of describing dispersion-dominated interactions [24–
43]. As a result, methods are now readily available that enable routine studies of
π-stacking interactions in relatively large systems (100–200 atoms), or studies of
massive numbers of non-covalent complexes involving relatively small systems. Per-
haps the simplest, yet most popular, of the various means of treating dispersion-bound
complexes using DFT is through the inclusion of empirical dispersion corrections.
This was popularized by Grimme, [25, 26] who introduced so-called DFT-D2 (and
later DFT-D3) approaches [25–27, 29–31]. We have relied heavily on the B97-D ap-
proach, [26, 44] which, when paired with triple-ζ quality basis sets, provides accurate
π-stacking interaction energies at a very modest computational cost [45, 46].
Of course, any application of DFT, particularly highly-paramterized functionals
or those including empirical corrections, must be done with care. To this end, our
strategy has been to apply B97-D/TZV(2d,2p) to a large collection of π-stacked
dimers, and then also compute robust coupled cluster energies using large basis sets
[e.g. CCSD(T)/aug-cc-pVTZ] for selected systems. In this way, we have been able
to confirm the performance of B97-D for a subset of the systems under consideration
while still examining a broad collection of π-stacked dimers. In other words, we can
use B97-D to study a large enough collection of systems to draw general and con-
vincing conclusions, while anchoring these computations to robust ab initio results
for selected systems.
Another powerful tool for the study of non-covalent interactions is symmetry-
adapted perturbation theory (SAPT) [47, 48]. SAPT not only provides rigorous
interaction energies for non-bonded dimers, but simultaneously decomposes these
interaction energies into electrostatic, exchange-repulsion, induction, and dispersion
contributions. By exploiting density-fitting techniques, Sherrill and co-workers have
implemented [23, 49–52] an incredibly efficient SAPT code in the open-source
software package Psi4 [53]. At lowest-order (SAPT0), the SAPT code in Psi4 can
routinely be applied to systems of several hundred atoms, enabling detailed studies of
424 S. E. Wheeler
the various components underlying non-covalent interactions in quite large systems.

We have relied heavily on SAPT0/jun-cc-pVDZ energies to study substituent effects
in π-stacking interactions in diverse model stacked sandwich dimers [54]. Through
error-cancelation, this level of theory provides surprisingly robust predictions at very
low computational cost [52].
14.3 Models of π-Stacking Interactions
There have been a number of efforts to develop predictive physical models of π-

stacking interactions over the years. Chief among these is the seminal work by
Hunter and Sanders, published in 1990 [6]. Models of substituent effects in π-
stacking interactions have been around nearly as long [55–63]. However, starting in
the early 2000s, apparent flaws in these models were identified based on gas-phase
computations, and, in recent years, these models have been challenged by new views
of substituent effects in simple π-stacked systems [18, 64–77]. Below, we provide
a brief overview of the most popular “conventional” views of substituent effects in
π-stacking interactions, followed by the most recent developments in this area.
The benzene dimer is typically used as a model system for π-stacking interactions,
for which three limiting configurations are typical considered (see Fig. 14.1). Among
these configurations, the sandwich dimer is a saddle-point on the potential energy
surface, whereas both the parallel-displaced and T-shaped dimers are energy minima
with roughly equal interaction energies. However, we do not consider the T-shaped
interactions to be an example of π-stacking interactions. First, this dimer does not
resemble a “stacked” geometry (we challenge the reader to try “stacking” plates in this
configuration!). Perhaps more importantly, the non-covalent interactions operative in
the T-shaped and edge-to-face dimers are qualitatively different from those in either
the sandwich or parallel displaced benzene dimer [18].
We also note that, for simplicity, most models of substituent effects in π-stacking
interactions are cast in terms of the benzene sandwich dimer, even though this config-
uration is not a stable minimum in the unsubstituted case. This raises some concern
that conclusions drawn from studies of model sandwich dimers are not necessarily
applicable to “real world” π-stacking interactions. However, we have shown [73]
that substituent effects in the parallel-displaced dimer are strongly correlated with
those in the sandwich dimer, meaning that results arising from studies of substituent
effects in the benzene sandwich dimer are directly transferable to more realistic
parallel-displaced configurations.
14.3.1 Conventional Views of Substituent Effects in π-Stacking

Interactions
The pioneering work of Hunter and Sanders [6] provided an intuitive physical model
of π-stacking interactions. In particular, Hunter and Sanders [6] showed that the
strength and preferred orientations of π-stacking interactions between disparate

aromatic rings could be understood and predicted based on a simple electrostatic
model, in which the aromatic rings are viewed as a collection of atom-centered local
quadrupole moments. That is, the basic behavior of stacked aromatic rings is recov-
ered by representing each non-hydrogen atom as a positive charge of +2q flanked
by charges of − q located at distances ± d above and below the molecular plane.
This simple model provided simple and intuitive explanations of the preference for
aromatic rings to form parallel-displaced and T-shaped configurations.
Among other things, Hunter and Sanders [6] also emphasized that the “donor-
acceptor concept can be misleading when used to describe π-π interactions: it is
the properties of the atoms in the regions of intermolecular contact that control the
strength and geometry of interactions, rather than the overall molecular oxidation or
reduction potentials.” In other words, the still all too common view of π-stacking
interactions as donor-acceptor complexes is, in many cases, incorrect. Furthermore,
Hunter and Sanders emphasized [6] that the charge-transfer transitions in the UV
spectra of π-stacked complexes are a consequence of the π-π interaction, not the
cause!
The Hunter-Sanders model [6] also laid the groundwork for the development of
an intuitive physical model of the impact of substituent effects in π-stacking interac-
tions. This was carried out primarily by Hunter, who published a series of papers on
this subject over the ensuing decades [55–57, 78–80]. The substituent effect model of
Hunter and co-workers, depicted in Fig. 14.2, is purely electrostatic in nature. Intrigu-
ingly, whereas the original Hunter-Sanders model represented the π-electron system
using atom-centered local quadrupoles, in describing substituent effects Hunter cast
his model in terms of the overall “π-electron system.” In particular, Hunter proposed
that electron-withdrawing substituents (e.g. –CN, –NO2 , etc.) enhance π-stacking
interactions by depleting the π-electron density of the substituted ring and reliev-
ing the π-π repulsion between the two rings. On the other hand, electron-donating
groups (e.g. –CH3 , –OCH3 , etc.) hinder π-stacking interactions through the opposite
mechanism. More generally, Hunter and co-workers [55, 57] emphasized that the
strongest π-stacking interactions will occur between aromatic rings with comple-
mentary electrostatic character. That is, “electron-rich” rings stack most favorably
with “electron-poor” rings. In terms of computed electrostatic potentials (ESPs),
this means that strong π-stacking interactions occur when one arene has a negative
ESP above the center and the other a largely positive ESP. This intuitive model is
widespread in the literature, and seems to jive with the basic concept that, when it
comes to electrostatic interactions, opposites attract.
This intuitive model of substituent effects in π-stacking interactions is entrenched
in the chemical literature, and has ruled over the field for decades. Moreover, this
view of π-stacking interactions has enabled the rational exploitation of π-stacking
interactions in many contexts and powered the growing fields of supramolecular
chemistry and crystal engineering. However, in recent years, high-accuracy compu-
tational studies of model π-stacking interactions have pinpointed what appear to be
deep-rooted flaws in this venerable model [18, 72, 74, 81].
426 S. E. Wheeler
Fig. 14.2 Depiction of the electrostatic model of substituent effects in π-stacking interactions from
Hunter and co-workers [55–57]
14.3.2 Unexpected Substituent Effects in the Benzene Dimer
Starting in the early 2000s, Sherrill and co-workers, [17, 18, 22, 23, 45, 46, 64–69,
82–85] as well as others, [5, 75–77, 86] began publishing systematic, high-accuracy
gas-phase interaction energies for prototypical stacked dimers of benzene and sub-
stituted benzenes. The data quickly revealed unexpected substituent effects for these
gas-phase dimers, [18, 64–67] kicking off a frenzy of computational and experi-
mental studies of substituent effects in π-stacking interactions. In particular, these
revelations engendered lively debate regarding the relative importance of electro-
static and dispersion effects in substituent effects in π-stacking interactions [18, 57,
65–69, 77, 79].
As noted above, the conventional view is that electron-withdrawing substituents
enhance π-stacking interactions, whereas electron-donating groups hinder these in-
teractions. Sherrill and co-workers [18, 64–67] found qualitatively different behavior
for model benzene dimers. In particular, they predicted enhanced π-stacking interac-
tions for all substituted dimers, regardless of the nature of the substituent. Moreover,
SAPT computations revealed that these substituent effects are not driven solely by
electrostatic effects, but result from a mélange of other effects (dispersion, induction,
etc.). Indeed, Sherrill found no correlation between computed interaction energies
and Hammett σ-constants. This flew in the face of conventional views of substituent
effects in π-stacking interactions, [55–57] as well as a vast body of experimental
data showing a correlation of π-stacking interactions with Hammett σ-constants!
14.3.3 The Importance of Direct Interactions
In 2008, Wheeler and Houk [70] entered the fray with a short communication on
the substituted benzene sandwich dimer. Unlike Sherrill et al. [18, 64–67] who
utilized high-level ab initio methods and were therefore limited in the number of
substituted benzene dimers that could reasonably be examined, Wheeler and Houk
[70] used the then recently developed M05-2X DFT functional [32]. This allowed
the examination of a far broader range of substituents. In particular, in contrast to
the four substituents examined by Sherrill and co-workers, [18, 65, 66] (OH, CH3 ,
F, and CN), Wheeler and Houk [70] considered a diverse set of 25 substituents. This
broader view afforded a much clearer picture of the overall trend in substituent effects
in gas-phase benzene sandwich dimers. In particular, Wheeler and Houk [70] showed
that there is a correlation between computed gas-phase interaction energies and σm
(see Fig. 14.3a). However, there was one outlier—the unsbubstituted benzene dimer.
In particular, when X = H was excluded, computed gas-phase interaction energies
are correlated with σm (r = 0.91), but the best-fit line had a y-intercept of − 0.4
kcal mol−1 . That is, the substituted sandwich dimers interact more favorably by
0.4 kcal mol−1 compared to the value predicted simply based on the Hammett σm
constant. This non-zero y-intercept was attributed to dispersion interactions, which
were postulated to stabilize all substituted dimers compared to the unsubstituted case.
In other words, Wheeler and Houk showed that the origin of Sherrill’s observation
that even electron-donating substituents enhance π-stacking interactions is due to
the stabilizing effects of dispersion interactions on all substituted benzene sandwich
dimers.
More importantly, Wheeler and Houk [70] showed that substituent effects in
the benzene sandwich dimer could be captured without the substituted benzene!
More precisely, they showed that interaction energies of H-X. . . C6 H6 dimers, in
which the substituent (X) was positioned exactly as it is in the corresponding
C6 H5 X. . . C6 H6 dimer, are also correlated with σm . Moreover, the best-fit line
through the HX. . . C6 H6 data has the same slope and intercept as the C6 H5 X. . . C6 H6
results (see Fig. 14.3b). This provided compelling evidence that the factors that are
responsible for the substituent effects in the HX. . . C6 H6 complexes are the same
as those in the C6 H5 X. . . C6 H6 sandwich dimers. Moreover, since hydrogen has no
π-electrons, these substituent effects in both complexes must not be the result of the
π-resonance effects that are central to Hunter’s model! Instead, Wheeler and Houk
proposed that substituent effects in the benzene sandwich dimer were due primarily
to direct interactions between the local dipole moment associated with the substituent
(and captured by H–X) and the other ring. Similarly, the dispersion interactions re-
sponsible for the 0.4 kcal mol−1 “extra” stabilization of the substituted dimers arose
from dispersion interactions between the substituents and the other ring.
Thus, the overall picture that emerged was one in which the substituent effect trend
is due to electrostatic effects, but all substituted dimers are stabilized by dispersion
interactions by an “extra” 0.4 kcal mol−1 , on average. However, the electrostatic
effects giving rise to the substituent effect trend were not the ones proposed by
428 S. E. Wheeler
Fig. 14.3 Interaction energies (kcal mol−1 ), relative to X = H, for a C6 H5 X. . . C6 H6 sandwich

dimers and b HX. . . C6 H6 dimers versus Hammett σm constants for the substituents X. The open
circles represent the X = H case. (Reprinted with permission from [70]. Copyright 2008 American
Chemical Society)
Hunter and co-workers, [55–57] but were instead due to direct interactions between
the substituents and the other ring!
In subsequent work, Sherrill and co-workers [68] and Lewis et al. [77] contin-
ued to debate the relative role of dispersion and electrostatic effects in substituent
effects in the benzene sandwich dimer. For example, Sherrill and co-workers [68]
examined polysubstituted benzene sandwich dimers, for which the correlation of
interaction energies with Hammett σm constants again breaks down. This was ex-
plained shortly thereafter in a remarkable paper by Sherrill and co-workers, [69]
in which they showed that at the close intermolecular distances characteristic of
these highly-substituted benzene sandwich dimers, electrostatic effects were always
stabilizing because of interpenetration effects. At larger inter-ring distances, the cor-
relation with Hammett σm constants reappeared. Meanwhile, Lewis and co-workers
[77] also examined a broad range of mono- and polysubstituted benzene sandwich
dimers using SAPT. Their results revealed that even though the impact of dispersion
interactions varies greatly among substituted dimers, these variations are almost ex-
actly cancelled by complementary variations in the exchange-repulsion and induction
components of these interactions. Consequently, the trends in substituent effects in
diverse substituted stacked dimers are almost completely due to electrostatic effects;
changes in the other components of the interaction tend to always cancel. Thus, the
overall picture that emerged is one in which electrostatic interactions are the dom-
inant source of substituent effects in π-stacking interactions, but these electrostatic
effects are often counterintuitive due to penetration effects.
14.3.4 The Local, Direct Interaction Model
In 2011, we introduced the local, direct interaction model of substituent effects in

π-stacking interactions, [73] which can be viewed as an extension of the direct inter-
action model from Wheeler and Houk [70]. The underlying concept is that substituent
effects in π-stacking interactions are due primarily to the direct, through-space inter-
actions of the substituent with the proximal vertex of the other ring (see Fig. 14.4).
In other words, the impact of a given substituent will depend only on what is in
its local environment. This model was motivated by the observation that substituent
effects in the benzene sandwich dimer could be reproduced by only considering HX
interacting with propene (with all conserved atoms in the same position as in the
intact sandwich dimer). That is, not only is the substituted phenyl ring unnecessary
to capture the substituent effects, but neither is half of the unsubstituted benzene.
This local nature of substituent effects in π-stacking interactions has a number of
important implications for understanding substituent effects in general π-stacking
interactions. For example, substituent effects in π-stacking interactions of substi-
tuted benzenes with 1,2,3-trifluorobenzene are drastically different depending on
whether the substituent sits over the fluorinated or non-fluorinated vertex of the tri-
fluorobenzene. If the substituent is over the fluorinated end of trifluorobenzene, then
the substituent effects are essentially identical to those in the C6 H5 X. . . C6 F6 dimer.
On the other hand, when the substituent is positioned over the non-fluorinated end
of trifluorobenzene, the substituent effects are simply those of the benzene dimer.
That is, the presence of the fluorines has no impact on the effect of substituent X
unless they are in the local environment of the substituent. Similarly, the local, direct
430 S. E. Wheeler
Fig. 14.4 Depiction of the

local, direct interaction model
of Wheeler and co-workers
[73]. (Reprinted with
permission from. Copyright
2011 American Chemical
Society)
interaction model also explains why 1,3,5-trifluorobenzene engages in p-stacking

interactions (− 5.9 kcal mol−1 ) that are nearly as strong as those between benzene
and perfluorobenzene (− 6.3 kcal mol−1 ).
This local nature of substituent effects also provides a simple means of under-
standing substituent effects in π-stacking interactions involving N-heterocycles. In
fact, unless the heteroatom is located in the local environment of the substituent,
it does not change the substituent effect at all. This means that the impact of sub-
stituents on pyridine-benzene and pyridine-pyridine dimers is the same as those in
the benzene dimer as long as the substituent is not located in the local environment
of one of the nitrogens. As seen below, the insensitivity of substituent effects to the
presence of heteroatoms extends well beyond a single nitrogen atom.
Overall, the local, direct interaction model implies that substituent effects in π-
stacking interactions will be additive, orientationally dependent, and transferable.
Many of these features have been previously observed. For example, Sherrill and
co-workers [66] showed the additivity of substituent effects in several configurations
of the benzene dimer. However, at the time, there was no straightforward explanation
of this additivity, since this behavior is not expected based on Hunter’s view of sub-
stituent effects. The local, direct interaction model provides a simple and physically
sound explanation of the additivity of substituent effects in π-stacking interactions.
Similarly, it has been observed experimentally that substituent effects are sensitive
to relative orientations in cases where both rings are substituted [87, 88]. This is not
consistent with a straightforward application of Hunter’s substituent effect model,
[55–57] which formally depends only on the π-electron density. This dependence
on relative orientation is a direct result of the local nature of substituent effects. This
orientational dependence also provides a potentially powerful tool for controlling
the local orientation of stacked discotic systems, which have potential applications
in organic electronic materials. For example, we recently showed [89] that the local
orientations of model stacked coronenes and circumcoronenes can be controlled by
exploiting the local, direct interactions of substituents on the stacked rings.
Finally, the transferability of substituent effects in π-stacking interactions means
that the impact of a given substituent will be the same across diverse π-stacked
dimers as long as the local environment is unchanged. This was demonstrated above
for π-stacking interactions with partially-fluorinated rings, as well as N-heterocycles.
In 2013, we tested the limits of this transferability by considering monosubstituted
sandwich dimers involving benzene, borazine, and triazine [54]. Ultimately, sub-
stituent effects were independent of the identity of the substituted ring. Moreover,
substituent effects were the same across any pair of dimers in which the unsubsti-
tuted ring was unchanged. For example, in Fig. 14.5a, relative interaction energies
for sandwich dimers of boron-substituted borazine are plotted again substituted ben-
zene sandwich dimers. These interaction energies are strongly correlated, and the
slope of the best-fit line is very close to unity. This confirms that substituent effects
in π-stacking interactions are independent of the identity of the substituted ring.
However, there were also instances of correlations among substituent effects in
π-stacked dimers in which the unsubstituted ring was not the same [54]. For exam-
ple, interaction energies for stacked dimers of substituted triazine with borazine are
correlated with those in stacked dimers of boron-substituted borazine with benzene.
Explaining unexpected correlations such as this, which were not predicted by the
local, direct interaction model, required an extension of the model.
14.3.5 Substituent Effects and Electric Fields
The unexpected correlations between substituent effects in π-stacked dimers in which

the unsubstituted ring is different (e.g. Fig. 14.5b) can be explained by examining the
electric fields of the unsubstituted rings. This is because the local, direct interactions
responsible for substituent effects in π-stacking interactions are dominated by the
electrostatic interaction of the local dipole moment associated with the substituent
and the electric field of the other ring.
The electric field of benzene in the plane bisecting the ring is shown in Fig. 14.6a,
superimposed over the ESP in this same plane. Also shown is a typical electron
density isosurface of benzene (r = 0.001 e/au3 ), along with the silhouette of a stacked
benzene at 3.65 Å. The electric field at the position of the substituent on this stacked
ring will interact with the parallel component of the local dipole associated with
the substituent. The electric field surrounding borazine is shown in Fig. 14.6b. Even
though the electric fields of benzene and borazine are different globally, the two
electric fields are parallel at the position of the substituents on the other ring. This
is demonstrated more quantitatively in Fig. 14.6c, where we plot the dot product
of the normalized electric fields of benzene and borazine. As seen in Fig. 14.6c,
the electric fields of benzene and borazine are parallel across much of this region,
particularly at the position of substituents above either the boron or nitrogen vertices
of borazine. Consequently, the electrostatic interaction with a substituent will be
similar, and π-stacked dimers involving unsubstituted benzenes and borazine will
exhibit the same substituent trends. Overall, substituent effect trends are the same in
two stacked dimers not only in cases in which the unsubstituted ring is unchanged,
but in any case where the direction of the electric field is parallel at the position
of the substituent. This explains the expected correlation in Fig. 14.5b, as well as
all other sandwich dimers considered in which either benzene or borazine was the
unsubstituted ring.
432 S. E. Wheeler
Fig. 14.5 SAPT0/jun-cc-pVDZ interaction energies, relative to the unsubstituted case for sandwich
dimers of a boron-substituted borazine with benzene vs. substituted benzene with benzene and b
substituted triazine with borazine and boron-substituted borazine with benzene. (Reprinted with
permission from [73]. Copyright 2011 American Chemical Society)
Fig. 14.6 Electric fields (arrows) and electrostatic potentials (solid colors) of a benzene and b bo-
razine in a plane bisecting the two rings. c Dot product of the normalized electric fields surrounding
benzene and borazine in the plane bisecting the rings. At a given point, blue indicates that the two
electric fields are parallel, while red signifies anti-parallel electric fields. Green indicates regions
where the two electric fields are perpendicular. d Strength of the electric field of borazine relative
to benzene in the plane bisecting the rings. Darker colors indicate that the electric field of borazine
is stronger than that of benzene at that point, while lighter colors indicate that the electric field of
borazine is weaker than that of benzene. In all four panels, the gray shaded region is an electron
density isosurface (ρ = 0.001 e/bohr3 ) and the black silhouette is of a sandwich stacked benzene at
R = 3.65 Å. (Reprinted with permission from [54]. Copyright 2013 American Chemical Society)
However, even if the electric fields above two rings are parallel, their strengths are
not necessarily the same. This is true for borazine and benzene, which underlies the
non-unit slope of the best-fit line in Fig. 14.5b. In particular, the electric field above
the vertex of borazine is weaker than that above benzene, as indicated in Fig. 14.6d.
434 S. E. Wheeler
Fig. 14.7 a ESPs of benzene, toluene, benzonitrile, and nitrobenzene. b ESPs of toluene, ben-
zonitrile, and nitrobenzene constructed by adding the ESP due to a hydrogen capped substituent
(i.e. H-CH3 , H-CN, and H-NO2 ) to the ESP of benzene. c ESP of 9-dicyanomethylene-2,4,7-
trinitrofluorine (far right) as well as the an ESP constructed by adding the ESPs of three H-NO2 s
to the ESP of 9-dicyanomethylenefluorene
Examining electric fields also provides a simple explanation of many other aspects
of substituent effects in π-stacking interactions. For instance, the electric field above
propene is very similar to that above a vertex of benzene, which explains why sub-
stituent effects are the same in HX. . . propene dimers as in the intact monosubstituted
benzene sandwich dimer [73]. Similarly, we have shown that substituent effects in
π-stacking interactions of substituted benzenes with carbon nanotubes (CNTs) and
a small graphene model depend only on the polarizability of the substituents, not
any electrostatic character [90]. This is a consequence of the weak strength of the
electric field above the center of CNTs and graphene.
14.4 What About Molecular Electrostatic Potentials?
Central to discussions of π-stacking and many other non-covalent interactions are

molecular electrostatic potentials (ESPs, e.g. see Fig. 14.7a). These colorful plots can
provide key insights into both the strength and preferred orientation of non-covalent
complexes, and have emerged as powerful tools for the analysis and prediction of
non-covalent interactions. Unfortunately, these ESP plots are widely misinterpreted
among chemists.
At a given point in space in the vicinity of a molecule, the ESP is the electrostatic
interaction that a positive test charge would experience at that position. In other
words, the ESP reflects the balance between the repulsion of a positive test charge
by the nuclei and the attraction of a test charge by the electron density. As such,
differences in the ESPs between two molecules indicate differences in both the elec-
tron density distribution and the nuclear charge distribution. Furthermore, the ESP
at a given point depends on the electron density everywhere in space. Consequently,
even a remote change in the electron density distribution will impact the ESP at a
given point. Practically, this means that a change in ESP in a given region is not
necessarily an indication of a local change in the electron density.
However, it is common among chemists to associate differences in ESPs between
two molecules with differences in the local electron density. This is seemingly
the basis of popular terms such as “electron-rich” and “electron-poor’ in describ-
ing aromatic systems in particular. For example, the ESPs above benzonitrile
and nitrobenzene are significantly more positive than the ESP above benzene (see
Fig. 14.7a). This difference is widely explained as arising from the π-resonance-
based delocalization of a formal positive charge throughout the π-system of the
aromatic rings bearing the electron-withdrawing substituents, since the presence of
this positive charge in the π-system of the ring would lead to repulsion of a positive
test charge. That is, the positive ESPs above these rings seem to reinforce the classic
view of benzonitrile and nitrobenzene as having an “electron-deficient π-system.”
This is not actually the case.
In 2009, Wheeler and Houk [91] showed that the ESPs above the faces of a diverse
collection of substituted arenes can be reproduce by simply adding the through-space
effects of the substituents to the ESPs of the corresponding unsubstituted arenes. For
example, in Fig. 14.7b we show the ESPs of a fictitious versions of toluene, benzoni-
trile, and nitrobenzene in which we simply added the ESP arising from hydrogen
capped-substituents to the ESP of benzene. These “additive” ESPs, in which both
the σ- and π-electron density of the arene itself are identical to that of benzene,
are strikingly similar to the ESP of the intact substituted systems. In other words,
the difference in the ESPs between these substituted phenyl rings and benzene can
be reproduced without making any changes in either the σ- or π-electron distri-
bution of the ring! Instead, the changes in the ESPs above these substituted arenes
primarily reflect the through-space effects of the local dipole associated with the sub-
stituents. Indeed, Wheeler and Houk [91] found for nearly all substituted arenes that
substituent-induced changes in the ESPs are primarily the result of through-space
effects of the substituents, not changes in the π-electron-density. This is shown
explicitly for 9-dicyanomethylene-2,4,7-trinitrofluorene in Fig. 14.7c.
These results underscore the local nature of substituent effects in π-stacking inter-
actions. That is, even though substituents result in drastic changes in the ESP above
aromatic systems, these changes do not reflect changes in the π-electron density.
Instead, these changes result from the electrostatic potential associated with the sub-
stituent, which will typically be dominated by the local dipole moment associated
with that substituent.
Similar misconceptions underlie common discussion of the ESPs of N-
heterocycles. For example, the ESP of benzene and five azines are shown in Fig. 14.8.
Generally, the positive ESPs above these “electron-deficient” N-heterocycles are
436 S. E. Wheeler
Fig. 14.8 Electrostatic potentials of benzene and five azines—pyridine, pyrazine, s-triazine, s-
tetrazine, and the hypothetical molecule hexazine
attributed to the depletion of π-electron density above the carbon atoms due to the
presence of the electronegative nitrogens. However, this is not the case! We recently
examined the differences in ESPs between benzene and these five azines. These
N-heterocycles exhibit ESPs that are drastically different from that of benzene, yet
are all both σ- and π-isoelectronic with benzene. This allowed direct comparisons
among the various contributions to the ESP differences.
In particular, we showed that the positive ESP above the center of the azines does
not arise from any change in the π-electron system. Instead, it simply reflects the
greater proximity of nuclear charge to the ring center in the azines, compared to
benzene. For example, the ESPs of benzene and triazine in the plane bisecting these
rings are shown in Fig. 14.9a. Figure 14.9b shows the difference in these ESPs in this
plane, as well as the contribution of the nuclei, σ-electrons, and π-electrons to this
difference. Notably, the reason that the ESP above triazine is positive is the position
of the nuclei—there simply is a greater amount of positive nuclear charge near the
center of tetrazine than the center of benzene. Moreover, the impact of both the σ-
and π-electron densities of triazine depress the ESP value above the ring, compared
to benzene. That is, compared to benzene, there is a more favorable electrostatic
interaction between the σ- and π-electron density and a positive test charge above
this so-called “electron-deficient” N-heterocycles! Similar results were found for the
other azines.
Although the implications of these most recent findings have only been examined
for anion-π interactions, [92] they have important implications for our understanding
of the impact of heteroatoms on all non-covalent interactions in which electrostatic
interactions play a role. This includes π-stacking interactions, and we think that this
is an area poised for a substantial shift in our understanding.
14.5 Conclusions and Outlook
In the late 1990s, π-stacking interactions seemed relatively well-understood [6, 55,
93, 94]. These interactions were abundant in chemical and biochemical systems,
and there were well-trodden physical models to describe π-stacking interactions
in both unsubstituted and substituted π-stacked dimers [55–57]. This all changed
Fig. 14.9 a Electrostatic potential in the planes bisecting benzene and triazine. b The difference
in ESPs between triazine and benzene (ESP) in this plane, as well as the nuclear (ESPN ),
σ-electronic [ESPe (σ)], and π-electronic [ESPe (π)] contributions to ESP
in the early 2000s, when high-accuracy gas-phase computations revealed a much

richer collection of forces at play in even simple π-stacked systems (i.e. the benzene
dimer). Moreover, studies of substituted π-stacked dimers revealed what seemed to
be significant weaknesses in the prevailing models of these interactions. Although
some of these weaknesses stemmed from differences in gas-phase versus solution-
phase interactions, the door had been opened for a dramatic shift in our understanding
of these interactions.
Consequently, there have been significant advances over the last 10 years in our
understanding of substituent effects in π-stacking interactions. Although it is still
not fully settled, there is mounting evidence that local, direct interactions between
the substituents and the other rings play a significant, and previously unrecognized,
role in tuning the strength of π-stacking interactions [73, 74, 81, 89]. We now
understand these local, direct interactions as arising primarily from the interaction
438 S. E. Wheeler
of the local dipole moment associated with the substituents and the electric field of
the other ring [54].
Ultimately, the local nature of substituent effects in π-stacking interactions reveals
that these effects are much simpler than previously thought. That is, even though
the π-resonance-based picture of Hunter and Sanders was intuitive (for chemists),
effects transmitted through π-resonance are inherently more complicated than direct,
through-space electrostatic interactions. Moreover, the additivity and transferability
implied by the local nature of substituent effects means that the net impact of multiple
substituents in larger, more complicated π-stacking interactions can be understood
quite simply. Finally, the local nature of substituent effects in π-stacking interactions
provides a potentially powerful tool for controlling the local orientation of π-stacking
interactions, which was formally not possible under a strict interpretation of the
original model championed by Hunter (i.e. Fig. 14.2).
Finally, there still exists a minor crisis with regard to nomenclature for these and
related non-covalent interactions. That is, the work from Martinez and Iverson, [21]
among others, [19, 20] have raised serious questions about the merits of referring to
these interactions as “π-stacking” or “π-π stacking” interactions. Similarly, it is not
uniformly established whether the phrase “π-stacking interaction” should be reserved
for aromatic, or even cyclic conjugated species, or if two ethylenes, for example, can
engage in “π-stacking interactions”. Although these linguistic issues might appear
trivial or superficial, language and the underlying concepts they convey form the
foundation for discussions of chemistry and chemical phenomena. Consequently,
these fundamental terms must be well-understand and universally defined.
Regardless, the last decade has been an exciting time to be studying π-stacking
interactions, and has also demonstrated that significant progress can be made under-
standing these enigmatic interactions. Ultimately, the prospect for continued progress
unraveling the origin of substituent effects in π-stacking interactions appears bright.
Acknowledgment This work was supported by the National Science Foundation (Grant CHE-
1254897) and the Welch Foundation (GrantA-1776). We also thank the TexasA&M Supercomputing
Center for providing computational resources and J. W. G. Bloom, R. K. Raju, K. N. Houk, D. A.
Dougherty, C. Corminboeuf, H. M. Jaeger, F. A. Evangelista, B. L. Iverson, and J. S. Siegel for
many fruitful discussions about π-stacking interactions.
References
and biological recognition. Angew Chem Int Ed 42:1210–1250
recognition: energetics and structures. Angew Chem Int Ed 50:4808–4842
3. Schneider H-J (2009) Binding mechanisms in supramolecular complexes. Angew Chem Int Ed
48:3924–3977
5. Kim KS, Tarakashwar P, Lee JY (2000) Molecular clusters of π-systems: theoretical studies
of structures, spectra, and origin of interaction energies. Chem Rev 100:4145–4186
6. Hunter CA, Sanders JKM (1990) The nature of π-π interactions. J Am Chem Soc 112:
5525–5534
7. Yakovchuk P, Protozanova E, Frank-Kamenetskii MD (2006) Base-stacking and base-pairing
contributions into thermal stability of the DNA double helix. Nucleic Acids Res 34:564–574
8. Kool ET (2002) Replacing the nucleobases in DNA with designer molecules. Acc Chem Res
35:936–943
9. Riley KE, Hobza P (2013) On the importance and origin of aromatic interactions in chemistry
and biodisciplines. Acc Chem Res 46:927–936
10. Cooper VR, Thonhauser T, Puzder A, Schröder E, Lundqvist BI, Langreth DC (2008) Stacking
interactions and the twist of DNA. J Am Chem Soc 130:1304–1308
11. Riley KE, Hobza P (2011) Noncovalent interactions in biochemistry. WIREs Comp Mol Sci
1:3–17
12. Hargis JC, Schaefer HF, Houk KN, Wheeler SE (2010) Non-covalent interactions of a
Benzo[a]pyrene diol epoxide with DNA base pairs: insight into the formation of adducts of
(+)-BaP DE-2 with DNA. J Phys Chem A 114:2038–2044
13. Knowles RR, Jacobsen EN (2010) Attractive noncovalent interactions in asymmetric catal-
ysis: links between enzymes and small molecule catalysts. Proc Natl Acad Sci U S A 107:
20678–20685]
14. Takenaka N, Chen JS, Captain B (2011) Helical chiral 2,2’-Bipyridine N-monoxides as cata-
lysts in the enantioselective propargylation of aldehydes with allenyltrichlorosilane. Org Lett
13:1654–1657
15. Lu T, Zhu R, An Y, Wheeler SE (2012) Origin of enantioselectivity in the propargylation of
aromatic aldehydes catalyzed by helical N-Oxides. J Am Chem Soc 134:3095–3102
16. Krenske EH, Houk KN (2013) Aromatic interactions as control elements in stereoselective
organic reactions. Acc Chem Res 46:979–989
17. Sinnokrot MO, Valeev EF, Sherrill CD (2002) Estimates of the Ab initio limit for π–π
interactions: the benzene dimer. J Am Chem Soc 124:10887–10893
18. Sinnokrot MO, Sherrill CD (2006) High-accuracy quantum mechanical studies of π-π
interactions in benzene dimers. J Phys Chem A 110:10656–10668
19. Grimme S (2008) Do special noncovalent π-π stacking interactions really exist? Angew Chem
Int Ed 47:3430–3434
20. Bloom JWG, Wheeler SE (2011) Taking aromaticity out of aromatic interactions. Angew Chem
Int Ed 50:7847–7849
21. Martinez CR, Iverson BL (2012) Rethinking the term “π-stacking”. Chem Sci 3:2191–2201
22. Sherrill CD, Takatani T, Hohenstein EG (2009) An assessment of theoretical methods for
nonbonded interactions: comparison to complete basis set limit coupled-cluster potential energy
curves for the benzene dimer, the methane dimer, benzene–methane, and benzene–H2 S. J Phys
Chem A 113:10146–10159
23. Hohenstein EG, Sherrill CD (2012) Wavefunction methods for noncovalent interactions.
WIREs Comp Mol Sci 2:304–326
24. Dion M, Rydberg H, Schröder E, Langreth DC, Lundqvist BI (2004) Van der Waals density
functional for general geometries. Phys Rev Lett 92:246401
25. Grimme S (2004) Accurate description of van der Waals complexes by density functional theory
including empirical corrections. J Comp Chem 25:1463–1473
26. Grimme S (2006) Semiempirical GGA-type density functional constructed with a long-range
dispersion correction. J Comp Chem 27:1787–1799
27. Grimme S, Antony J, Schwabe T, Mück-Lichtenfeld C (2007) Density functional theory with
dispersion corrections for supramolecular structures, aggregates, and complexes of (bio)organic
molecules. Org Biomol Chem 5:741–758
28. Schwabe T, Grimme S (2008) Theoretical thermodynamics for large molecules: walking the
thin line between accuracy and computational cost. Acc Chem Res 41:569–579
29. Grimme S, Antony J, Ehrlich S, Krieg H (2010) A consistent and accurate ab initio parametriza-
tion of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J Chem
Phys 132:154104
30. Grimme S (2011) Density functional theory with London dispersion corrections. WIREs Comp
Mol Sci 1:211–228
440 S. E. Wheeler
31. Ehrlich S, Moellmann J, Grimme S (2013) Dispersion-corrected density functional theory for
aromatic interactions in complex systems. Acc Chem Res 46:916–926
32. ZhaoY, Schultz NE, Truhlar DG (2006) Design of density functionals by combining the method
of constraint satisfaction with parametrization for thermochemistry, thermochemical kinetics,
and noncovalent interactions. J Chem Theory Comput 2:364–382
33. Zhao Y, Truhlar DG (2006) A new local density functional for main-group thermochemistry,
transition metal bonding, thermochemical kinetics, and noncovalent interactions. J Chem Phys
125:194101
34. Zhao Y, Truhlar DG (2007) Density functionals for noncovalent interaction energies of
biological importance. J Chem Theory Comput 3:289–300
35. Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group thermo-
chemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition
elements: two new functionals and systematic testing of four m06 functionals and twelve other
functionals. Theor Chem Acc 120:215–241
36. Zhao Y, Truhlar DG (2008) Density functionals with broad applicability in chemistry. Acc
Chem Res 41:157–167
37. Vydrov OA, Van Voorhis T (2009) Nonlocal van der Waals density functional made simple.
Phys Rev Lett 103:063004
38. Vydrov OA, Van Voorhis T (2012) Benchmark assessment of the accuracy of several van der
Waals density functionals. J Chem Theory Comput 8:1929–1934
39. Vydrov OA, Van Voorhis T (2010) Dispersion interactions from a local polarizability model.
Phys Rev B 81:062708
40. Becke A; Johnson ER (2005) Exchange-hole dipole moment and the dispersion interaction. J
Chem Phys 122:154104
41. Johnson ER, Mackie ID, DiLabio GA (2009) Dispersion interactions in density-functional
theory. J Phys Org Chem 22:1127–1135
42. Steinmann SN, Corminboeuf C (2010) A system-dependent density-based dispersion correc-
tion. J Chem Theory Comput 6:1990–2001
43. Tkatchenko A, Scheffler M (2009) Accurate molecular van der Waals interactions from ground-
state electron density and free-atom reference data. Phys Rev Lett 102:073005
44. Becke A (1997) Density-functional thermochemistry. V. systematic optimization of exchange-
correlation functionals. J Chem Phys 107:8554–8560
45. Vázquez-Mayagoitia Á, Sherrill CD, Apra E, Sumpter BG (2010) An assessment of density
functional methods for potential energy curves of nonbonded interactions: the XYG3 and
B97-D approximations. J Chem Theory Comput 6:727–734
46. Burns LA, Vázquez- Mayagoitia Á, Sumpter BG, Sherrill CD (2011) Density-functional
approaches to noncovalent interactions: a comparison of dispersion corrections (DFT-D),
exchange-hole dipole moment (XDM) theory, and specialized functionals. J Chem Phys
134:084107
47. Jeziorski B, Moszyński R, Szalewicz K (1994) Perturbation theory approach to intermolecular
potential energy surfaces of van der Waals complexes. Chem Rev 94:1887–1930
48. Szalewicz K (2012) Symmetry-adapted perturbation theory of intermolecular forces. WIREs
Comp Mol Sci 2:254–272
49. Hohenstein EG, Sherrill CD (2010) Density fitting and Cholesky decomposition approxima-
tions in symmetry-adapted perturbation theory: implementation and application to probe the
nature of π-π interactions in linear acenes. J Chem Phys 132:184111
50. Hohenstein EG, Sherrill CD (2010) Efficient evaluation of triple excitations in symmetry-
adapted perturbation theory via second-order Møller-Plesset perturbation theory natural
orbitals. J Chem Phys 133:104107
51. Hohenstein EG, Sherrill CD (2010) Density fitting of intramonomer correlation effects in
symmetry-adapted perturbation theory. J Chem Phys 133:014101
52. Parker TM, Burns LA, Parrish RM, Ryno AG, Sherrill CD (2014) Levels of symmetry adapted
perturbation theory (SAPT). I. Efficiency and performance for interaction energies. J Chem
Phys 140:094106
53. Turney JM, Simmonett AC, Parrish RM, Hohenstein EG, Evangelista FA, Fermann JT, Mintz
BJ, Wilke JJ, Abrams ML, Russ NJ et al (2012) Psi4: an open-source ab initio electronic
structure program. WIREs Comp Mol Sci 2:556–565
54. Raju RK, Bloom JWG, Wheeler SE (2013) Broad transferability of substituent effects in π-
stacking interactions provides new insights into their origin. J Chem Theory Comput 9:3479–
3490
55. Hunter CA, Lawson KR, Perkins J, Urch CJ (2001) Aromatic interactions. J Chem Soc Perkin
Trans 2:651–669
56. Cockroft SL, Hunter CA, Lawson KR, Perkins J, Urch CJ (2005) Electrostatic control of
aromatic stacking interactions. J Am Chem Soc 127:8594–8595
57. Cockroft SL, Perkins J, Zonta C, Adams H, Spey SE, Low CMR, Vinter JG, Lawson KR, Urch
CJ, Hunter CA (2007) Substituent effects on aromatic stacking interactions. Org Biomol Chem
5:1062–1080
58. Cozzi F, Cinquini M, Annunziata R, Dwyer T, Siegel JS (1992) Polar/π interactions between
stacked aryls in 1,8-Diarylnaphthalenes. J Am Chem Soc 114:5729–5733
59. Cozzi F, Cinquini M,Annunziata R, Siegel JS (1993) Dominance of polar/π over charge-transfer
effects in stacked phenyl interactions. J Am Chem Soc 115:5330–5331
60. Cozzi F, Ponzini F, Annunziata R, Cinquini M, Siegel JS (1995) Polar interactions between
stacked π systems in fluorinated 1,8-Diarylnaphthalenes: importance of quadrupole moments
in molecular recognition. Angew Chem Int Ed 34:1019–1020
61. Cozzi F, Siegel JS (1995) Interaction between stacked aryl groups in 1,8-Diarylnaphthalenes:
dominance of polar/π over charge-transfer effects. Pure Appl Chem 67:683–689
62. Cozzi F, Annunziata R, Benaglia M, Cinquini M, Raimondi L, Baldridge KK, Siegel JS (2003)
Through-space interactions between face-to-face, center-to-edge oriented arenes: importance
of polar-π effects. Org Biomol Chem 1:157–162
63. Cozzi F, Annunziata R, Benaglia M, Baldridge KK, Aguirre G, Estrada J, Sritana-Anant Y,
Siegel JS (2008) Through-space interactions between parallel-offset arenes at the van der
Waals distance: 1,8-diarylbiphenylene syntheses, structure and QM computations. Phys Chem
Chem Phys 10:2686–2694
64. Sinnokrot MO, Sherrill CD (2003) Unexpected substituent effects in face-to-face π-stacking
interactions. J Phys Chem A 107:8377–8379
65. Sinnokrot MO, Sherrill CD (2004) Substituent effects in π-π interactions: sandwich and T-
shaped configurations. J Am Chem Soc 126:7690–7697
66. Ringer AL, Sinnokrot MO, Lively RP, Sherrill CD (2006) The effect of multiple substituents
on sandwich and T-Shaped π-π interactions. Chem Eur J 12:3821–3828
67. Arnstein SA, Sherrill CD (2008) Substituent effects in parallel-displaced π-π interactions.
68. Ringer AL, Sherrill CD (2009) Substituent effects in sandwich configurations of multiply
substituted benzene dimers are not solely governed by electrostatic control. J Am Chem Soc
131:4574–4575
69. Hohenstein EG, Duan J, Sherrill CD (2011) Origin of the surprising enhancement of electro-
static energies by electron-donating substituents in substituted benzene sandwich dimers. J Am
Chem Soc 133:13244–13247
70. Wheeler SE, Houk KN (2008) Substituent effects in the benzene dimer are due to di-
rect interactions of the substituents with the unsubstituted benzene. J Am Chem Soc
130:10854–10855
71. Wheeler SE, McNeil AJ, Müller P, Swager TM, Houk KN (2010) Probing substituent effects
in aryl–aryl interactions using stereoselective Diels–Alder cycloadditions. J Am Chem Soc
132:3304–3311
72. Raju RK, Bloom JWG,AnY, Wheeler SE (2011) Substituent effects in non-covalent interactions
with aromatic rings: insights from computational chemistry. ChemPhysChem 12:3116–3130
73. Wheeler SE (2011) Local nature of substituent effects in stacking interactions. J Am Chem Soc
133:10262–10274
442 S. E. Wheeler
74. Wheeler SE (2013) Understanding substituent effects in non-covalent interactions involving

aromatic rings. Acc Chem Res 46:1029–1038
75. Lee EC, Kim D, Jurečka P, Tarakeshwar P, Hobza P, Kim KS (2007) Understanding of assembly
phemona by aromatic-aromatic interactions: benzene dimer and the substituted systems. J Phys
Chem A 111:3446–3457
76. Seo J-I, Kim I, Lee YS (2009) π-π interaction energies in monosubstituted-benzene dimers in
parallel- and antiparallel-displaced conformations. Chem Phys Lett 474:101–106
77. Watt M, Hardebeck LKE, Kirkpatrick CC, Lewis M (2011) Face-to-face arene-arene bidning
energies: dominated by dispersion but predicted by electrostatic and dispersion/polarizability
substituent constants. J Am Chem Soc 133:3854–3862
78. Hunter CA (2004) Quantifying intermolecular interactions: guidelines for the molecular
recognition toolbox. Angew Chem Int Ed 43:5310–5324
79. Cockroft SL, Hunter CA (2007) Chemical double-mutant cycles: dissecting non-covalent
interactions. Chem Soc Rev 36:172–188
80. Cockroft SL, Hunter CA (2009) Desolvation and substituent effects in edge-to-face aromatic
interactions. Chem Commun 3961–3963
81. Wheeler SE, Bloom JWG (2014) Toward a more complete understanding of non-covalent
interactions involving aromatic rings. J Phys Chem A 118:6133–6147
82. Tauer T, Sherrill CD (2005) Beyond the benzene dimer: an investigation of the additivity of
π–π interactions. J Phys Chem A 109:10475–10478
83. Hohenstein EG, Sherrill CD (2009) Effects of heteroatoms on aromatic π-π interactions:
benzene-pyridine and pyridine dimer. J Phys Chem A 113:878–886
84. Sherrill CD (2009) Computations of noncovalent pi interactions. In: Lipkowitz KB Cundari
TR (eds) Reviews computational chemistry, vol 26. Wiley-VCH, New York, pp 1–38
85. Marshall MS, Burns LA, Sherrill CD (2011) Basis set convergence of the coupled-cluster
correction, δ(MP2)(CCSD(T)): best practices for benchmarking non-covalent interactions and
the attendant revision of the S22, NBC10, HBC6, and HSG databases. J Chem Phys 135:194102
86. Beg S, Waggoner K, Ahmad Y, Watt M, Lewis M (2008) Predicting face-to-face arene-arene
binding energies. Chem Phys Lett 455:98–102
87. Gung BW, Emenike BU, Alvereza CN, Rakovan J, Kirschbaum K, Jain N (2010) Relative
substituent position on the strength of π-π stacking interactions. Tetrahedron Lett 51:1648–
1650
88. Benitex Y, Baranger AM (2011) Control of the stability of a protein-RNA Complex by the
position of fluorine in a base analog. J Am Chem Soc 133:3687–3689
89. Wheeler SE (2012) Controlling the local arrangements of π-stacked polycyclic aromatic
hydrocarbons through substituent effects. CrystEngComm 14:6140–6145
90. Munusamy E, Wheeler SE (2013) Endohedral and exohedral complexes of cyclohexane and
substituted benzenes with carbon nanotubes and graphene. J Chem Phys 139:094703
91. Wheeler SE, Houk KN (2009) Through-space effects of substituents dominate molecular
electrostatic potentials of substituted arenes. J Chem Theory Comput 5:2301–2312
92. Wheeler SE, Bloom JWG (2014) Anion-π interactions and positive electrostatic potentials of
N-heterocycles arise from the positions of the nuclei, not changes in the π-electron distribution.
Chem Commun 50:11118–11121
93. Hunter CA (1994) Meldola lecture. The role of aromatic interactions in molecular recognition.
Chem Soc Rev 23:101
94. Carver FJ, Hunter CA, Seward EM (1998) Structure-activity relationship for quantifying
aromatic interactions. Chem Commun 775–776
Chapter 15
Noncovalent Interactions of Organic Ions with
Polar Molecules in the Gas Phase
M. Samy El-Shall, Isaac K. Attah and Sean P. Platt
Abstract The chapter is focused on noncovalent interactions of organic ions with

small polar molecules in the gas phase. The organic ions studied include cyclic
C3 H + •+ •+
3 and the radical cations of benzene (C6 H6 ), pyridine (C5 NH5 ), pyrimi-
•+ •+ •+
dine (C5 N2 H4 ), fluorobenzene (C6 H5 F ), phenylacetylene (C8 H6 ), benzonitrile
(C7 NH•+ •+ +
5 ) and naphthalene (C10 H8 ). In addition, protonated pyridine (pyridine.H )
+
and protonated pyrimidine (pyrimidine.H ) are also included for comparison with the
radical cations. The solvent molecules include water (H2 O), hydrogen cyanide (HCN)
and acetonitrile (CH3 CN). The results presented include experimental thermochem-
ical data (H◦ and S◦ ) for the stepwise association of the solvent molecules with
the organic ions and theoretically calculated binding energies and structures. The
four major topics discussed are: (1) Trends in binding energies and entropy changes,
(2) Ionic hydrogen bonds with organic ions, (3) Internal vs. external solvation of the
organic ions, and (4) Intracluster proton transfer and deprotonation of the organic
ions.
15.1 Introduction
Intermolecular forces, including hydrogen bonds and ion-molecule interactions, [1–

5] are important in many biological, chemical, and astrochemical processes such as
the conformation and folding of proteins, base pair stacking in DNA, drug design,
macromolecular assemblies, molecular recognition, clathrate hydrate formation, and
the formation of complex organics and ices in interstellar space [1–7].
A special class of hydrogen bonding interactions, usually referred to as ionic
hydrogen bonds (IHBs), involves hydrogen bonding between radical ions or pro-
tonated molecules and neutral polar molecules such as water, methanol, ammonia,
and hydrogen cyanide [1]. IHBs have bond strengths higher than the typical
M. S. El-Shall () · I. K. Attah · S. P. Platt

Department of Chemistry, Virginia Commonwealth University, 23284–2006 Richmond, VA, USA
e-mail: mselshal@vcu.edu
M. S. El-Shall
Department of Chemistry, Faculty of Science, King Abdulaziz University,
Jeddah 21589, Saudi Arabia
444 M. S. El-Shall et al.
conventional hydrogen bond in neutral systems and could reach up to 35 kcal/mol,

nearly a third of the strength of covalent bonds [1]. These strong interactions are
critical in many fields such as ion induced nucleation, ionic clusters, ion solvation,
radiation chemistry, electrochemistry, acid-base chemistry, and self-assembly in
supramolecular chemistry [1–5]. IHBs are also important in biological systems
including protein folding, proton transport, membranes, enzyme active centers, and
molecular recognition [8, 9] Organic ions can form hydrogen bonds with solvents
in nature, for example, in icy grains doped with polycyclic aromatic hydrocarbons
that are subjected to ionizing radiation in interstellar dust grains [10–13].
Unconventional carbon-based IHBs are formed when the hydrogen donors are
ionized hydrocarbons containing CH groups and the hydrogen acceptors are elec-
tron lone pairs on hetero atoms such as O and N or π electrons in olefin double bonds
and aromatic systems [1]. For example, carbon-based CHδ+ –O IHBs appear in the
hydration of ionized aromatics such as benzene (C6 H•+ +
6 ), cyclic C3 H3 , cyclobutadi-
•+ •+ •+
ene (C4 H4 ), phenylacetylene (C8 H6 ), and naphthalene (C10 H8 ) [14–19]. Organic
ions can also interact with water molecules by stronger conventional hydrogen bonds,
such as in protonated pyridine or protonated pyrimidine where the NH+ –O hydrogen
bonds are formed [1, 20, 21].
In addition to water, other polar molecules containing lone pairs of electrons
such as hydrogen cyanide, methanol, and acetonitrile can participate in hydrogen
bonding interactions with the ring hydrogen atoms (CHδ+ ) of ionized aromatics.
Hydrogen cyanide is a useful probe of non-covalent interactions because it is highly
polar (dipole moment = 2.98 D), [22] and it can serve both as a hydrogen donor
and as a lone-pair hydrogen acceptor in hydrogen bonds [1]. HCN has received
considerable attention because of its role in atmospheric chemistry as a result of its
release by biomass burning, [23] and it is also one of the main interstellar/nebula
molecules [13]. HCN can be produced in space from the reactions of ammonia
and methane, and could play a very important role in the formation of nitrogen-
containing polycyclic aromatic hydrocarbons (NPAHs) [24]. Since molecules in
outer space are subjected to ionizing radiation from stars, radioactive decay and
cosmic rays, ionized aromatics such as benzene and PAHs may act as nucleation
centers for the condensation of astrochemical molecules such as hydrogen cyanide to
form clusters that could undergo intracluster reactions leading to nitrogen containing
PAHs [6, 24]. Nitrogen-containing aromatics are of interest because biologically
significant molecules such as DNA, RNA, and certain amino acids and proteins
contain nitrogen-substituted rings. As a result, a fundamental understanding of the
reactions and structures of nitrogen-containing species in space may provide insights
into the origin of life [24–26]. In fact, HCN polymers have been shown to exist in
meteorites, comets, planets, moons, and in circumstellar envelopes [24, 27, 28].
Insight into the basic molecular interactions can be obtained from the energies and
structures of the key species involved in the stepwise association of polar molecules
with organic ions. These data can be obtained experimentally by measuring sequen-
tial binding energies of solvent molecules to organic ions in the gas phase, and
computationally by calculating the structures and binding energies of the hydrated
and solvated organic ions [1,14–21, 29–32]. One of the most established methods for
15 Noncovalent Interactions of Organic Ions with Polar Molecules in the Gas Phase 445
measuring binding energies of solvent molecules to organic ions in the gas phase is
the mass-selected ion mobility (MSIM) method which allows measurements of ther-
mochemical equilibria between a mass-selected ion and several solvent molecules
(typically 1–6) [14–21, 30–32]. These measurements, when conducted at different
temperatures, yield enthalpy and entropy changes associated with the stepwise as-
sociation of solvent molecules with the organic ions. A good deal of work over the
last decade or two has been conducted using equilibrium measurements to obtain
thermochemical data on the stepwise solvation of organic ions with a variety of sol-
vent molecules including water [1, 33–42]. This work continues to this day, with a
focus on trends in binding energies and correlations with the properties of the solvent
molecule such as dipole moment and proton affinity, and the nature of the organic ion
including size, charge distribution and degree of charge delocalization. Our group
has been active in studying the stepwise hydration and solvation of different types
of organic ions, and this chapter provides an overall review of these recent studies
[14–21, 43–48]. In particular, we focus on the interactions of organic ions with three
solvent molecules: water, hydrogen cyanide and acetonitrile. We specifically address
three issues: (1) variation of binding energies with the structures and properties of the
solvent molecules, (2) the onset of ion solvation involving a small number (four to six)
of solvent molecules and whether internal or external solvation is preferred, and (3)
the degree of intracluster proton transfer within the solvated ions and the ability of a
number of solvent molecules to abstract a proton from the organic ion to form a more
stable protonated solvent cluster and covert the organic radical cation into a radical
which could be more reactive in addition reactions with small organic molecules.
The review begins with a brief description of the MSIM technique to carry out
the thermochemical equilibrium measurements to determine the sequential binding
energies and entropy changes of the stepwise solvation of organic ions. The next
section describes the main experimental and computational results for the solvation
of organic ions. Attention is focused on: (1) Trends in Binding Energies & Entropy
Changes, (2) Structures of the Hydrated/Solvated Ions, (3) Internal vs. External
Solvation, and (4) Intracluster Proton Transfer & Deprotonation Reactions.
15.2 Experimental Measurements of Sequential Enthalpy and

Entropy Changes
Figure 15.1 illustrates the essential components of the ion mobility system at Virginia
Commonwealth University (VCU) [17, 18, 21]. In the experiments, mass-selected
ions (generated by electron impact ionization of the neutral molecules or clusters)
are injected (in 20–30 μs pulses) into the drift cell containing a helium carrier gas
typically at 1 Torr containing a known partial pressure of the solvent vapor. Flow
controllers are used to maintain a constant pressure inside the drift cell. The drift cell
temperature can be set within the range of 78–773 K, and can be controlled to ± 1
K using four temperature controllers, using liquid nitrogen flowing through actuated
solenoid valves to cool down the drift cell for temperature experiments below room
Lens C
Lens A B
Fig. 15.1 VCU mass-selected ion mobility system [18]
temperature. The reaction products can be identified by scanning a second quadrupole

mass filter located coaxially after the drift cell. The arrival time distributions (ATDs)
are collected by monitoring the intensity of each ion as a function of time. The reaction
time can be varied by varying the drift voltage. Most of the ion thermalization occurs
outside the cell entrance by collisions with the solvent vapor escaping from the cell
entrance orifice. At a cell pressure of 0.2 Torr, the number of collisions that the ion
encounters with the solvent molecules within the 1.5 ms residence time inside the
cell is about 104 collisions, which is sufficient to ensure efficient thermalization of
most of the organic ions.
Figures 15.2a and 15.2b show typical examples of mass spectra obtained fol-
lowing the injection of the mass-selected pyrimidine and benzene radical cations
into the drift cell containing water and HCN vapors, respectively [21, 45]. In the
absence of water in the drift cell (Fig. 15.2a), only the mass-selected pyrimidine
radical cation is observed as shown in Fig. 15.2a (top). In the presence of water,
both the (C4 N2 H•+ +
4 )(H2 O)n with n = 1–5 and the protonated water series H (H2 O)n
with n = 4–9 are observed [21]. They consistently shift towards higher n as the tem-
perature of the drift cell decreases. For example, at 238 K the observed ions are
(C4 N2 H•+ +
4 )(H2 O)n with n = 2−5 and H (H2 O)n with n = 5−9 as shown in Fig. 15.2a
(bottom). The formation of protonated water clusters H+ (H2 O)n with n ≥ 4 is at-
tributed to the dissociative proton transfer reactions represented by Reaction 1 below
Eq. (15.1):
[C4 H4 N2•+ (H2 O)n−1 ] + H2 O → C4 H3 N2 • + [H + (H2 O)n ]; n ≥ 4 (15.1)
Similar dissociative proton transfer reactions have been observed in the hydration of
benzene and the cyclic C3 H3 cations [14–16]. These reactions are further discussed
in Sect. 15.2.4.
In Fig. 15.2b, the injection of the mass-selected C6 H•+ 6 ion into the drift cell
containing 0.1 Torr HCN vapor at 298 K, results in the formation of the association
products C6 H•+6 (HCN)n with n = 1 and 2 [45]. At 178 K (the lowest achievable
temperature before the condensation of HCN), the cluster population is dominated
by the C6 H•+
6 (HCN)n series from n = 2 to 6 with n = 3 and 4 being the major ions
observed as shown in Fig. 15.2b (bottom) [45].
Fig. 15.2 a Mass spectra resulting from the injection of the mass-selected pyrimidine radical cation
(C4 N2 H4•+ ) into a helium (He)—water (W ) vapor mixture at different temperatures using 12.7 eV
injection energy (laboratory frame) and 2.5 V/cm drift field [21]. b Mass spectra obtained following
the injection of the mass-selected C6 H•+
6 (B) into the drift cell containing He and HCN vapor at
different pressures and temperatures as indicated [45]
A good test of equilibrium is the observation of identical ATDs of the reactant and
product ions. If the I•+ (S)n−1 and I•+ (S)n ions are in equilibrium, their ATDs must
be identical [15–17, 22]. This is evident from the ATDs of the (C4 N2 H•+ 4 )(H2 O)n
ions with n = 1–5 as shown in Fig. 15.3a. The ion intensity ratio I•+ (S)n /I•+ (S)n−1 is
measured from the integrated peak areas of the ATDs as a function of decreasing cell
drift field corresponding to increasing reaction times, and equilibrium is achieved
when a constant ratio is obtained. The equilibrium constants are then obtained using
Eq. (15.2).
Keq = [I•+ (S)n ]/[I•+ (S)n−1 ]P(S) (15.2)
Here, [I•+ (S)n ] and [I•+ (S)n−1 ] are the intensities of the peaks taken from the in-
tegrated ATDs, and P(S) is the partial pressure of the solvent (S) in atm. The
equilibrium constants measured as a function of temperature yield H◦ and S◦
from the slopes and intercepts, respectively of the van’t Hoff plots as illustrated by
the typical examples shown in Fig. 15.3b.
Fig. 15.3 a (Left) ATDs of the pyrimidine•+ (H2 O)n ions with n = 1–3 collected following the
injection of mass-selected pyrimidine•+ into the drift cell containing 0.21 Torr water vapor + 0.22
Torr helium buffer gas at 268 K. (Right) ATDs of the pyrimidine•+ (H2 O)n ions with n = 3–5 collected
following the injection of mass-selected pyrimidine+ into the drift cell containing 0.19 Torr water
vapor and 0.22 Torr of helium buffer gas at 243 K [21]. b van’t Hoff plots of the temperature
dependence of the equilibrium constant of the stepwise hydration of the pyrimidine radical cation
and formation of pyrimidine•+ (H2 O)n with n = 1–4 [21]
Table 15.1 Ionization energies and proton affinities of the studied aromatic and polar molecules
[22]. The dipole moments (μ) of the polar molecules are also included
Compound Ionization energy (eV) Proton affinity (kcal/mol)
Benzene 9.2 179.3
Phenylacetylene 8.8 198.9
Fluorobenzene 9.2 180.7
1,4 di-Fluorobenzene 9.2 171.8
Benzonitrile 9.7 194.0
Naphthalene 8.1 191.9
Pyridine 9.3 222.0
Pyrimidine 9.3 211.7
H2 O (μ = 1.9 D) 12.6 165.0
CH3 OH (μ = 1.7 D) 10.8 180.3
HCN (μ = 2.98 D) 13.6 179.4
CH3 CN (μ = 3.9 D) 12.2 186.2
15.2.1 Trends in Binding Energies & Entropy Changes
Table 15.1 lists molecular properties relevant to the organic ions and solvent
molecules discussed in this chapter [22]. Tables 15.2, 15.3, 15.4 list the enthalpy
and entropy changes corresponding to the association of water, hydrogen cyanide,
and acetonitrile, respectively with different organic ions.
Table 15.2 Thermochemistry of hydrated organic cations

Cation (Ref) −Ho (kcal/mol) −So (cal mol−1 K−1 ) E (Calc) (kcal/mol)
Cyclic C3 H+
3 [16] 11.7 18.8 11.8
•+
Benzene [15] 9.0 19.5 8.5
•+
Phenylacetylene [17] 8.0 17.8 7.8
Benzonitrile•+a 11.5 29.4 7.7
•+
Naphthalene [19] 7.8 19.5 7.7
Pyridine•+ [20] 15.2 33.1 15.4 (distonic ion)
10.2 (conventional)
Protonated pyridine [20] 15.6 27.0 14.5
•+
Pyrimidine [21] 11.9 23.6 10.8
Protonated pyrimidine [21] 16.7 38.6 16.9
a
Unpublished results
Table 15.3 Thermochemistry of HCN complexes with organic cations

Cation (Ref) −Ho (kcal/mol) −So (cal mol−1 K−1 ) E (Calc) (kcal/mol)
•+
Benzene [45] 9.2 19.1 9.4
Phenylacetylene•+ [46] 10.5 24.6 9.5
•+
F-benzene [47] 11.2 24.4 9.6
1,4-di 11.2 22.5 9.8
F-benzene•+ [47]
Benzonitrile•+ [47] 9.6 19.1 9.5
Naphthalene•+a 6.8 15.3 7.8
•+
Pyridine [48] 11.4 21.8 11.9
Protonated pyridine 14.0 26.6 16.0
[48]
Pyrimidine•+ [48] 12.0 23.3 12.7
a
Unpublished results
Figure 15.4a compares the hydration energies and the sequential binding ener-
gies of several water molecules to the c-C3 H+ •+
3 , benzene (C6 H6 ), phenylacetylene
•+ •+
(C6 H5 CCH ) and naphthalene (C10 H8 ) ions. The hydration energy is the largest
for the c-C3 H+3 (11.7 kcal/mol) followed by the benzene and the phenylacetylene
cations (9.0 and 8.0 kcal/mol, respectively), and then the naphthalene cation (7.8
kcal/mol). The range of the hydration energies is consistent with the usual strength
of ionic hydrogen bonds of the type CHδ+ ··· OH2 which are typically about 9 kcal/mol
[1]. The IHB in the hydrated c-C3 H+ 3 is nearly 50 % stronger than in the respective
hydration of the naphthalene cation due to the higher charge density on the C3 H+ 3
Table 15.4 Thermochemistry of CH3 CN complexes with organic cations

Cation (Ref) −Ho −So (cal mol−1 K−1 ) E (Calc) (kcal/mol)
(kcal/mol)
C3 H+
3 [16] 15.3 21.5 –
Benzene•+ [45] 14.0 19.9 13.0
Protonated Pyrimidine – – 23.7
[21]
Naphthalene•+a 11.2 20.3 10.8
a
Unpublished results
Fig. 15.4 a Sequential binding energies of water molecules to the c-C3 H+ •+

3 , benzene (C6 H6 ),
phenylacetylene (C6 H5 CCH•+ ) and naphthalene (C10 H•+8 ) cations [15–17, 19]. b Sequential bind-
ing energies of water molecules to the benzene (C6 H•+ •+
6 ) and pyrimidine (C4 H4 N2 ) radical cations
+
and to protonated pyrimidine (C4 H4 N2 H ) [15, 21]
ion as compared to the C10 H•+ 8 ion. The charge delocalization over the bicyclic naph-
thalene cation decreases the charge-dipole interaction of the CHδ+ · · · OH2 bond in
the naphthalene•+ (H2 O) complex.
For both the c-C3 H+ 3 and the benzene cation, an increase in the binding energy
of the fourth water molecule is observed and for the phenylacetylene cation this
increase is observed for the addition of the third water molecule. This could suggest
the formation of geometrically stable structures of the first solvent shell around the
ion as shown from the DFT calculations discussed in Sect. 15.2.3.
Another interesting result is the apparent increase in the sequential binding ener-
gies associated with the formation of the benzene•+ (H2 O)7 , benzene•+ (H2 O)8 , and
phenylacetylene•+ (H2 O)7 clusters. The formation of these large clusters is also asso-
ciated with large entropy changes (−25.5, −32.6 and −32 cal/mol K, respectively) as
compared to the nearly constant −So value of 20 ± 3 cal/mol K for the n = 1–5 clus-
ters [15, 17]. This could suggest strong orientational restraint of water in the larger
clusters [49]. In fact, three-dimensional cage-like structures involving multiple rings
sharing edges are the lowest energy conformers of the water heptamer and octamer
[49, 50]. The observed large negative entropy of the phenylacetylene•+ (H2 O)7 and
benzene•+ (H2 O)8 clusters could be explained by the formation of cage-like structures
by 7–8 H2 O molecules similar to neutral water clusters [49, 50]. This observation
could be related to the common bulk view of hydrophobic hydration where water
molecules tend to organize around small hydrophobic units without sacrificing hy-
drogen bonds [51]. The entropy loss associated with this organization is the major
reason for the low solubility of nonpolar organic molecules in water [51]. However,
in the hydration of large organic ions such as benzene and phenylacetylene, it is
not theoretically confirmed that the organization of water molecules into cage-like
structures would incorporate such organic ions within the cages.
Figure 15.4b compares the sequential hydration energies of the pyrimidine rad-
ical cation (C4 H4 N•+ +
2 ) and protonated pyrimidine (C4 H4 N2 H ) with those of the
benzene radical cation. With the protonated pyrimidine, water is bonded by a con-
ventional NH+ · · ·OH2 bond which is significantly stronger (16.7 kcal/mol) than the
unconventional CHδ+ · · ·OH2 bond (8–9 kcal/mol) [21]. Also, the hydration energy
of the pyrimidine cation (11.9 kcal/mol) is higher than that of the benzene cation (9.0
kcal/mol) although in both cases the water binds to a CHδ+ site by an unconventional
IHB (CHδ+ · · ·OH2 ) as indicated by the DFT calculations of the lowest energy struc-
tures [15, 21]. The sequential binding energies of the pyrimidineH+ (H2 O)n clusters
decrease with n and approach the limiting macroscopic value of 10.5 kcal/mol, the
condensation energy of water at n = 3 [21]. This follows the usual trend in systems
with conventional ionic hydrogen bonds [1]. However, this trend is clearly different
from the hydration of the benzene cation as shown in Fig. 15.4b. In relation to these
data, the C-H hydrogens of the pyrimidine radical cation (C5 H4 N•+ 2 ) can only form
carbon-based CHδ+ ··OH2 bonds to water, similar to those formed by the benzene•+
ion [14, 15]. In fact, DFT calculations below also show that the hydrogen bond
strength of the pyrimidine•+ ion to water is 10.8 kcal/mol, in good agreement with
the measured value of 11.9 kcal/mol [21].
The hydration energy of protonated pyrimidine measured by the MSIM method
(16.7 kcal/mol) is consistent with the hydration energies of other aromatic ions re-
cently measured using energy-resolved collision-induced dissociation (CID) [29].
For example; the CID hydration energies of protonated aniline, acetophenone and
phenol were measured as 14.4, 15.6 and 17.5 kcal/mol, respectively [29]. The hydra-
tion energy of protonated pyrimidine is higher than that of aniline and slightly lower
than of that phenol. The measured hydration energies appear to correlate well with
the proton affinity of the aromatic molecules as shown in Tables 15.1 and 15.2 [1].
Figure 15.5a compares the sequential binding energies of HCN molecules with
protonated pyridine and the radical cations of pyrimidine, pyridine, phenylacetylene,
and benzene. The binding of multiple HCN molecules to the benzene cation is mostly
due to unconventional CHδ+ –NCH hydrogen bonds directly connected to the CHδ+
sites of the benzene cation and also hydrogen bonding chains (HCN–HCN) among
the HCN molecules [45]. The small difference in the bond strength of the two types
of interactions (CHδ+ –NCH and HCN–HCN) results in relatively small changes of
(−H◦n − 1,n ) for n = 1–4 as shown in Fig. 15.5a. HCN binds more strongly to the
pyridine and pyrimidine radical cations due to ion-dipole interactions in addition to
the CHδ+ –NCH hydrogen bonding interactions [48]. The strongest binding of HCN
Fig. 15.5 a Sequential binding energies of HCN molecules to the benzene (C6 H•+6 ), phenylacetylene
(C6 H5 CCH•+ ), pyridine (C5 H5 N•+ ), and pyrimidine (C4 H4 N•+
2 ) radical cations, and protonated
pyridine (C5 H5 NH+ ) [45–48]. b Sequential binding energies of CH3 CN molecules to c-C3 H+ 3,
benzene (C6 H•+
6 ), naphthalene (C10 H•+
8 ) and protonated pyrimidine (C4 H 4 N2 H+
) [16, 45, 48]
is observed with protonated pyridine as shown in Fig. 15.5a. This bonding interac-
tion is mainly due to an IHB that forms between the NH+ group of the protonated
pyridine and the N atom of HCN (NH+ –NCH bond). A significant drop in the bind-
ing energy (31 %) is observed upon the addition of the second HCN molecule to the
protonated pyridine in contrast to the smaller changes observed upon the addition
of the second HCN molecule to the pyridine or pyrimidine radical cations (23, and
15 %, respectively) as shown in Fig. 15.5a [48]. Despite the strong bonding of HCN
to protonated pyridine, the interaction decreases sharply by further addition of HCN
molecules and the binding of the fourth HCN molecule is only 6.1 kcal/mol [48].
In fact, in all of the studied HCN clusters around ionized or protonated aromatics,
[45–48] the binding energies converge with the addition of 4–5 HCN molecules to
the enthalpy of vaporization of HCN liquid (Ho vap ), which is 6.0 kcal/mol at 298
K [22].
Figure 15.5b compares the sequential binding energies of acetonitrile molecules to
protonated pyrimidine, c-C3 H3 cation and benzene and naphthalene radical cations.
DFT calculations indicate that acetonitrile forms a strong proton-bound dimer with
protonated pyrimidine (calculated binding energy = 23.7 kcal/mol) [46]. This is a
consequence of the large dipole moment of acetonitrile (3.9 D) as compared to 1.6 D
for water. As a result, the ion-dipole interaction term is stronger in the case of acetoni-
trile, and the overall binding energy is significantly higher in pyrimidineH+ (NCCH3 )
(23.7 kcal/mol) than in pyrimidineH+ (OH2 ) (16.7 kcal/mol) [21, 48].
Figure 15.5b also shows a significant drop in the binding energy (43 %)
upon the addition of the second acetonitrile molecule to the proton-bound dimer
(C4 N2 H4 )H+ (NCCH3 ) as compared to the water interactions where the correspond-
ing drop in binding energy is 24 % as shown in Fig. 15.4b. The transition from strong
ionic hydrogen bonding in the proton-bound dimer to weaker CHδ+ –N type of bonds
is responsible for the sharp drop in the binding energy of the (C4 N2 H4 )H+ (CH3 CN)2
cluster. However, in the case of water, extended hydrogen bonding networks can be
formed as shown in the calculated low energy structures of the (C4 N2 H4 )H+ (H2 O)2
clusters discussed in Sect. 15.2.3.
The effect of charge localization on the organic ion is clearly observed by compar-
ing the binding energies of c-C3 H+ •+
3 (CH3 CN)n and naphthalene (CH3 CN)n clusters
•+
shown in Fig. 15.5b. The naphthalene (CH3 CN)n clusters show the weakest se-
quential binding energies among the ions shown in Fig. 15.5b. This is again a result
of charge delocalization on the naphthalene radical cation which leads to weaker
charge-dipole interactions in comparison with the c-C3 H+ 3 (CH3 CN)n clusters.
15.2.2 CHδ+ . . . .O and CHδ+ . . . .N Ionic Hydrogen Bonds with

Organic Ions
The calculated lowest energy structures of several hydrated ions along with their
calculated binding energies (kcal/mol) are shown in Fig. 15.6. Except for the pro-
tonated pyrimidine, protonated pyridine and distonic pyridine ion, all the ions show
unconventional IHBs of the type CHδ+ · · ·OH2 . For the c-C3 H+ 3 (H2 O), a single hy-
drogen bond between the water oxygen atom and one of the C3 H+ 3 hydrogen atoms
is formed. The charges on c-C3 H+ 3 are equally distributed and localized among the
carbon atoms (90 % of the charge) [16]. Upon clustering with one water molecule,
the charge distribution changes with significant localizations on the hydrogen atoms
of the c-C3 H+3 , and one C-H bond shows elongation resulting from the linear H-bond
formation with the oxygen atom of the water molecule [16].
For the benzene•+ , phenylacetylene•+ , naphthalene•+ , and pyrimidine•+ radical
cations, the lowest energy hydrated ion has a bifurcated structure with H2 O bond-
ing to two CH hydrogens as shown in Fig. 15.6. This structure was also found in
several ab initio studies of the benzene.+ (H2 O) cluster [52–57]. The MP2 corrected
binding energy of 8.5 kcal/mol (Table 15.1) matches well the experimental value of
9.0 ± 1 kcal/mol, [15] and that reported from the IR photodissociation experiment
of 9.4 ± 0.3 kcal/mol of the benzene•+ (H2 O) cluster [58].
For the hydrated phenylacetylene cation, C8 H•+ 6 (H2 O), another structure where
the water molecule is attached to the hydrogen of the acetylene group is only 1
kcal/mol higher in energy than the lowest energy isomer (at the MP2//ROHF/6-
31+G** level) [17]. This indicates that the hydration of the organic ions could
involve several structures with energetically similar binding sites. Experimentally,
the observed equilibrium could involve different structural isomers depending on the
relative stability of these isomers and the temperature of the experiment. As a result,
the measured binding energy may represent an average value for water binding to
different hydrogen atoms on the ion.
For the hydrated naphthalene cation, the two lowest energy isomers (at the
B3LYP/6-311++G** level) have bifurcated structures with H2 O bonding to two CH
hydrogens as shown in Fig. 15.6d [19]. The bifurcated structures have significantly
larger binding energies than the ion-dipole structure where the water molecule lies
above the plane of the naphthalene cation [19]. Also the calculated binding energies
Fig. 15.6 Lowest energy structures of hydrated organic cations. ΔE is the calculated binding energy
in kcal/mol. The calculated structures were obtained at the MP2//ROHF/6-31+G(d, p) level for the
c-C3 H+ •+ •+ •+
3 , benzene , phenylacetylene , pyridine , distonic pyridine
•+
and protonated pyridine
cations, [15–17,20] at the B3LYP/6-311++G(d, p) level for the naphthalene•+ and benzonitrile•+ ,
[19] and at the M06-2X/6-311++G(d, p) level for the pyrimidine•+ and protonated pyrimidine [21]
of the bifurcated structures (7.7–6.8 kcal/mol) are in good agreement with the
experimentally determined value (7.8 ± 1 kcal/mol) [19].
For the hydration of the pyridine radical cation, the measured binding energy of
the pyridine•+ (H2 O) cluster (15.2 kcal/mol) was similar to that of the protonated
pyridine-water cluster (C5 H5 NH+ )(H2 O) (15.6 kcal/mol) that involves a NH+ ··OH2
bond, and different from those of the hydrated benzene radical cation (9.0 kcal/mol)
(Table 15.2) [20]. These relationships indicated that the hydrated pyridine•+ ions
have the distonic • C5 H4 NH+ structure that can form NH+ ··OH2 bonds [20]. The
calculated CHδ+ ··OH2 binding energy for the conventional pyridine ion to one water
molecule is 10.2 kcal/mol, which does not agree with the experimental measured
value of 15.2 kcal/mol, while the calculated NH+ ··OH2 binding energy of the dis-
tonic pyridine ion to one water molecule is 15.4 kcal/mol, in excellent agreement with
the experimental measured values of 15.2 and 15.6 kcal/mol for the(C5 H5 N•+ )(H2 O)
and (C5 H5 NH+ )(H2 O) clusters, respectively [20]. However, the hydrated pyrimidine
Fig. 15.7 Lowest energy structures of the organic cation −(HCN) clusters. ΔE is the calculated
binding energy in kcal/mol. The calculated structures were obtained at the B3LYP/6-311++G(d, p)
level for the benzene•+ , phenylacetylene•+ , F-benzene•+ , 1,4-di F-benzene•+ , benzonitrile•+ , and
naphthalene•+ radical cations [45–47], and at the M06-2X/6-311++G(d, p) level for the pyridine•+ ,
pyrimidine•+ and protonated pyridine [48]
ions were found to have the conventional radical cation structures similar to the hy-
drated benzene cation. For the hydrated pyrimidine radical cation (C4 N2 H•+
4 )(H2 O)n ,
the lowest energy isomer has a bifurcated structure with H2 O bonding to two CH
hydrogens with relatively larger distances (2.4 Å and 2.2 Å) than typical H-bonds as
shown in Fig. 15.6f. However, the lowest energy isomer of the hydrated protonated
pyrimidine has the water molecule directly attached to the NH+ center via a NH+ –O
hydrogen bond of 1.67 Å (Fig. 15.6g), similar to the structure of the hydrated pro-
tonated pyridine [21]. This is mediated by the stronger interaction that involves a
NH+ ··OH2 bond similar to other ionic hydrogen bonds where the proton is shared
between two centers containing lone pairs of electrons.
The calculated lowest energy structures of HCN complexes with several organic
ions are shown in Fig. 15.7. The complexes of HCN with ionized aromatics are
bonded primarily by electrostatic forces both in planar hydrogen-bonded structures

and in vertical L-shaped structures as shown in Fig. 15.7. The lowest energy isomers
of the HCN complexes with the benzene, benzonitrile, pyridine and naphthalene
radical cations calculated at the B3LYP/6-311++G(d, p) level have bifurcated struc-
tures with HCN binding to two CH hydrogen atoms as shown in Fig. 15.7. For the
C8 H•+
6 (HCN) complex, the lowest energy isomer shows the N-atom of the HCN
molecule interacting with the acetylenic CH by a 2.0 Å hydrogen bond as shown in
Fig. 15.7b [46]. This indicates that the attachment of HCN to the acetylene group is
more favorable than the interaction with the phenyl ring.
The lowest energy structure of the C5 H5 NH+ (HCN) complex shows a conven-
tional IHB between the NH+ group of the protonated pyridine and the nitrogen atom
of HCN as shown in Fig. 15.7h. The lowest energy structure of the C4 H4 N•+ 2 (HCN)
complex has a T-shaped ion-dipole structure with the N atom of HCN pointing toward
the center of the pyrimidine cation ring [48].
The HCN molecule binds to the fluorobenzene radical cation at an angle of 87◦
to the plane of the fluorobenzene•+ ion with a bond length of 2.8 Å as shown in
Fig. 15.7c. A similar geometry is found with the difluorobenzene radical cation, with
the ion plane-to-molecule angle and bond length of 88◦ and 2.9 Å, respectively as
shown in Fig. 15.7d [47]. The atomic charges on the ions are also of interest. In
fluorobenzene•+ , the electronegative F atom is significantly negative and its carbon
has a high partial positive charge (0.60) in the ion [47]. The negative N end of
the HCN neutral molecule (− 0.31) can interact attractively with the large positive
charge at this carbon. On the other hand, in hydrogen bonding to the ring hydrogens,
especially the para hydrogen, HCN interacts with a smaller positive charge (0.24),
but this interaction is not destabilized by nearby negative charges. These effects
balance out fortuitously to lead to practically equal binding energies in the planar
and vertical isomers [47]. Similar charges, interactions and binding energies apply
in 1,4-difluorobenzene•+ . However, in benzonitrile•+ , although the CN group is
electron withdrawing, there are no large positive charges on the ring carbons, and
a planar bifurcated hydrogen-bonded structure has the lowest energy similar to the
benzene•+ (HCN) complex [47].
15.2.3 “Internally” & “Externally” Solvated Ions
It is of interest to compare “internally solvated” structures where the solvent

molecules are bonded directly to the ion with “externally solvated” structures where
the solvent molecules are bonded to each other and the ion is hydrogen bonded to
the exterior of a solvent cluster. Figure 15.8 displays the lowest energy structures of
the hydrated organic ions by two water molecules. It is interesting to note that all the
structures shown in Fig. 15.8, with the exception of the c-C3 H+3 and pyrimidine ions,
form “externally hydrated” structures where the two water molecules are bonded to
each other. This indicates that the formation of “internally hydrated” structures in
Fig. 15.8 Lowest energy structures of the organic cation-(H2 O)2 clusters. E is the calculated bind-
ing energy in kcal/mol. The calculated structures were obtained at the MP2//ROHF/6-31+G(d, p)
level for the c-C3 H+ •+ •+ •+
3 , benzene , phenylacetylene , pyridine , distonic pyridine
•+
and protonated
pyridine, [15–17,20] at the B3LYP/6-311++G(d, p) level for the naphthalene•+ and benzonitrile•+
radical cations, [19] and at the M06-2X/6-311++G(d, p) level for the pyrimidine•+ and protonated
pyrimidine [21]
which the water molecules are bonded to the organic ion may require more than two
water molecules.
Clustering of the second water molecule on the C3 H+ 3 (H2 O) ion gives a symmetric
internally hydrated structure as shown in Fig. 15.8a [16]. A similar structure is found
for the addition of the second water molecule to the pyrimidine (C4 N2 H•+ 4 ) ion as
shown in Fig. 15.8f [21]. However, the calculated structures of the hydrated proto-
nated pyrimidine (C4 N2 H+ 5 )(H2 O)n clusters are significantly different from those of
the hydrated radical cation [21]. The lowest energy pyrimidineH+ (H2 O) cluster, has
the water molecule directly attached to the NH+ center via a NH+ –O hydrogen bond
of 1.67 Å (Fig. 15.6g) similar to the structure of the hydrated protonated pyridine
Fig. 15.9 Lowest energy structures of the organic cation-(H2 O)4 clusters. ΔE is the calculated
binding energy in kcal/mol. The calculated structures were obtained at the MP2//ROHF/6-31+G(d,
p) level for the c-C3 H+ •+
3 and benzene , [15–17,20] at the B3LYP/6-311++G(d, p) level for
•+ •+
the benzonitrile and naphthalene , [19] and at the M06-2X/6-311++G(d, p) level for the
pyrimidine•+ [21]. The structure of the Naphthalene•+ -(H2 O)6 (Naphtalene•+ -W6 ) is also shown
shown in Fig. 15.6i [21]. For the protonated pyrimidine, the second water molecule
binds to the first water molecule by a 1.73 Å hydrogen bond while the NH+ –O bond
shortens to 1.59 Å as shown in Fig. 15.8g.
Figure 15.9 displays the lowest energy structures of the hydrated organic ions by
four water molecules. The sequential addition of four water molecules to the C3 H+ 3
ion results in a cyclic water tetramer bound with two long hydrogen bonds to the C3 H+3
hydrogens [16]. It is interesting that the lowest energy isomer of the C3 H+ 3 (H2 O)3
cluster involves the formation of a cyclic water trimer [16]. This propensity of water
molecules to form cyclic structures in the presence of the c-C3 H+ 3 ion is consistent
with the structural behavior of neutral water clusters [49, 59].
The lowest energy structure of C6 H•+ 6 (H2 O)4 is symmetrical with a water dimer on
each side of the ion, both forming bifurcated bonds with the ion [15]. IR spectroscopy
and ab initio calculations indicate that the first water molecule forms two IHBs with
two adjacent hydrogen atoms of the benzene cation as shown in Figs 15.6b [55, 60].
For the C6 H•+ 6 (H2 O)n with n = 2–4 both “internally solvated” isomers where the
water molecules are bonded directly to the benzene cation and “externally solvated”
isomers where the benzene cation is attached to a cluster of water molecules have
been predicted [54, 55, 60].
The lowest energy isomers of the pyrimidine•+ (H2 O)4 cluster show both “inter-
nally solvated” and “externally solvated” structures with similar binding energies
[21]. However, the calculated structures of the hydrated protonated pyrimidine
(C4 N2 H+5 )(H2 O)n clusters are different from those of the hydrated radical cation
[21]. The pyrimidineH+ (H2 O)3 shows a symmetric structure where a central water
molecule binds by 1.75 Å hydrogen bonds to two other water molecules; and to
protonated pyrimidine by a shorter NH+ –O bond of 1.43 Å [21]. This structure is
consistent with a growth pattern that could lead to dissociative proton transfer within
the pyrimidineH+ (H2 O)4 cluster to form the closed shell solvated hydronium ion
H3 O+ (H2 O)3 as observed experimentally [21].
Unlike the hydration of the benzene radical cation where the C6 H+• 6 (H2 O)n clus-
ters with n up to 4 exhibit mainly “internally” solvated structures with the water
molecules forming bifurcated hydrogen bonds with the CHδ+ groups of the benzene
cation, [15] the sequential addition of water molecules onto the naphthalene cation fa-
vors external solvation where the naphthalene cation remains exterior with respect to
the water sub-cluster [19]. The calculated structures of the naphthalene•+ (H2 O)4 and
naphthalene•+ (H2 O)6 are shown in Figs. 15.9e and 15.9f, respectively. Similar struc-
tures are also obtained for naphthalene•+ (CH3 OH)4 and naphthalene•+ (CH3 OH)6
[19].
This indicates that hydrogen bonding interactions among water molecules are
more favorable than the unconventional hydrogen bonding formed between the CHδ+
sites of the naphthalene cation and the water molecules. This observation suggests
that the solvation of larger PAH cations would also favor the “externally” solvated
structures where the large organic species remain accessible at the surface of the
water nanodroplets. This could have an important implication to the reactivity of
large PAH ions in interstellar medium since the ions can interact with incoming
small molecules under UV irradiation where both ion-molecule and photochemical
processes can lead to the formation of complex organics on the surface of the ice
grains [10, 19].
The above results and discussions indicate that unlike the hydration of metal
ions, where the ion-water interaction is significantly stronger than the water-water
interaction, water-water interaction can be very similar or even stronger than ion-
water interaction in the hydration of organic ions [1]. Therefore, it appears that the
common ion hydration picture where the ion is surrounded by 4–6 water molecules
may not be valid for the hydration of organic ions. The observed trend of a large
entropy loss with the higher order hydration steps as a result of forming cyclic or
bicyclic hydrogen bonded water clusters supports this view [15, 17]. In other words,
unless the organic ion is very small (e.g. C3 H+
3 ), “internally hydrated” structures may
not form and the large organic ions remain outside the cage-like hydrogen bonded
water structures [19].
15.2.3.1 Solvation of Organic Ions by HCN Molecules
Solvation of the organic ions by HCN molecules show a tendency to form extended
HCN chains associated with the benzene and phenylacetylene radical cations and pro-
tonated pyridine. Figure 15.10 displays the calculated lowest energy structures of the
benzene•+ (HCN)n , phenylacetylene•+ (HCN)n and protonated pyridineH+ (HCN)n
clusters with n = 2 and 4. It is clear that the addition of the second HCN molecule
Fig. 15.10 Lowest energy structures of organic cation-(HCN)n clusters with n = 2 and 4. ΔE is the
calculated binding energy in kcal/mol. The calculated structures were obtained at the B3LYP/6-
311++G(d, p) level for the benzene•+ and phenylacetylene•+ radical cations, [45, 46] and at the
M06-2X/6-311++G(d, p) level for protonated pyridine [48]
is more likely to form HCN. . . HCN hydrogen bond than forming a second C-
Hδ . . . .NCH bond with the organic ion. The structures of the ions solvated by four
HCN molecules represent a competition between the assembly of a linear hydrogen

bonding HCN chain and internal solvation of the cation. Interestingly, structures with
linear hydrogen bonded chains extended from the phenylacetylene cation are only 1
kcal/mol higher in energy than the lowest energy isomer. For protonated pyridine,
the extended chain structure with four HCN molecules (Fig. 15.10j) is 1.0 kcal/mol
higher in energy than the lowest energy isomer with an extended chain of three HCN
molecules shown in Fig. 15.10i. These extended structures are good models for the
design of molecular wires formed by hydrogen bonding chains that could be utilized
for low temperature molecular devices [61–63]. For example, the lengths of these
molecular wires (the distance from the carbon atom of the phenyl ring connected
to the C≡C group to the hydrogen atom of the terminal HCN molecule in the one-
sided chain structures) can vary from 0.8 to 2.1 nm depending on the number of HCN
molecules. For the two sided chain structures where the phenyl ring can be considered
as part of the wire, the distance between the hydrogen atoms of the terminal HCN
molecules on both sides of the ring can reach 2.5 nm in the C8 H•+ 6 (HCN)4 cluster.
These wire assemblies connected by hydrogen bonds could provide unique opportu-
nities for systematic spectroscopic and structure studies of charge distributions and
separations in these systems.
It should be noted that extended linear chain structures involving HCN and acety-
lene units are common in the solid state molecular crystals of these molecules.
For example, in the solid state, both hydrogen cyanide and cyanoacetylene
(H-C≡C-C≡N) crystals consist of extended, linear hydrogen-bonded chains [64, 65].
However, in the gas phase both linear and cyclic structures are known to exist based on
several spectroscopic and theoretical studies [66–69]. Interestingly, clusters of linear
chains have been reported to exist in superfluid helium expansion [70]. In cluster ions,
the measured thermochemistry of protonated HCN clusters [H+ (HCN)n ] suggests
the formation of linear hydrogen-bonded chains extended from a strongly bound core
consisting of the proton-bound HCN dimer (HCN.H+ .NCH) [71, 72]. In the present
system, the core ion is the phenylacetylene cation which allows the formation of
HCN extended chains by hydrogen bonding to the acetylene hydrogen or to the hy-
drogen atoms of the phenyl ring. It appears that the formation of a one-sided chain by
hydrogen bonding of the HCN molecules to the acetylene hydrogen, and two-sided
chains by additional hydrogen bonding to the para hydrogen atom (with respect to
the acetylene group) of the phenyl ring represents the most favorable modes for the
growth of the hydrogen cyanide molecular wires attached to the phenylacetylene
cation.
The second group of organic ions which includes pyridine•+ , pyrimidine•+ ,
F-benzene•+ , 1,4-di-F-benzene•+ , benzonitrile•+ , and naphthalene•+ show more
tendency for internal solvation by four HCN molecules than for the formation of ex-
tended HCN chains. The calculated lowest energy structures of these solvated ions
are shown in Fig. 15.11.
The lowest energy structures of the C5 H5 N•+ (HCN)4 and C4 H4 N•+ 2 (HCN)4 clus-
ters show bifurcated structures involving multiple hydrogen bonding sites with the
ring hydrogen atoms as shown in Figs. 15.11a and 15.11b, respectively. However,
no chain structures involving four HCN molecules have been found in these clusters
Fig. 15.11 Lowest energy structures of organic radical cation-(HCN)4 clusters. ΔE is the calculated
binding energy in kcal/mol. The calculated structures were obtained at the B3LYP/6-311++G(d,
p) level for the F-benzene•+ , 1,4-di F-benzene•+ and benzonitrile•+ , [47] and at the M06-2X/6-
311++G(d, p) level for pyridine•+ and pyrimidine•+ radical cations [48]
indicating that HCN interactions with the pyridine or pyrimidine ring cations are
more favorable than the interactions within the HCN chains.
In the C6 H5 F•+ (HCN)4 cluster, the fourth HCN binds externally to the third bi-
furcated HCN with a bond length of 2.1 Å as shown in Fig. 15.11c. The measured
binding energy for this addition is 7.8 kcal/mol in good agreement with the DFT cal-
culated value of 7.4 kcal/mol. The fourth HCN in C6 H4 F•+ 2 (HCN)4 binds externally
to the second meta position on the aromatic ring as shown in Fig. 15.11d, with a bond
length of 2.2 Å. The measured binding energy of 7.7 kcal/mol agrees well with the
theoretical value of 7.3 kcal/mol. In the benzonitrile cluster ion C6 H5 CN•+ (HCN)4 ,
the fourth HCN ligand binds externally to the HCN at the ortho position, with a bond
length of 2.1 Å (Fig. 15.11e). The measured binding energy for this addition is 6.1
kcal/mol, which is lower than the DFT binding energy of the lowest energy isomer
(Fig. 15.11e) of 7.3 kcal/mol, but within the experimental uncertainty range of ± 1
kcal/mol for clustering equilibrium temperature studies.
The lowest energy structure of the naphthalene•+ (HCN)4 is very similar to that of
the benzene•+ (HCN)4 cluster where two HCN molecules are attached to the cation
ring from two opposite sides. This structure provides efficient solvation of the organic
ion. It is interesting to note that the naphthalene cation is externally solvated by four
water molecules (Fig. 15.9e) but it can be internally solvated by four HCN molecules
(Fig. 15.11f). As in all such clusters, the binding energies, with the addition of 4–6
ligands, converge to the enthalpy of vaporization of the ligand (ΔHo vap ), which is
6.0 kcal/mol for HCN at 298 K [1, 22, 48].
15.2.4 Intracluster Proton Transfer and Deprotonation Reactions
Intracluster proton transfer (IPT) coupled with cluster’s dissociation reactions have
been observed in several solvated organic radical cations R-H•+ (S)n with specific
numbers of solvent molecules (Sn ). These reactions, also known as Dissociative
Proton Transfer (DPT) reactions, generate a protonated solvent cluster Sn H+ and
an organic radical (R• ). Intracluster DPT reactions can occur within the R-H•+ (S)n
clusters if the proton transfer from the radical cation to the solvent sub-cluster (S)n be-
comes exothermic and the resulting energy is sufficient to dissociate the R• · · · H+ (S)n
cluster. The cluster size (n) at which the PT occurs can be predicted by comparing the
proton affinities of the organic radical (R• ) and the solvent sub-cluster (S)n . These
reactions have been observed in C6 H•+ + •+
6 (H2 O)n , C3 H3 (H2 O)n , C4 N2 H4 (H2 O)n , and
+
C4 N2 H5 (H2 O)n with n ≥ 4 in all cases [15, 16, 20, 21]. The IPT and DPT reactions
are represented by Eqs. (15.3) and (15.4) for the benzene•+ (H2 O)n clusters.
Intracluster Proton Transfer (IPT)
C6 H•+ •+
6 (H2 O)n → C6 H5 (H2 O)n H
+
(15.3)
Intracluster Dissociative Proton Transfer (IDPT)
C6 H•+
6 (H2 O)n → (H2 O)n H+ + C6 H5 • (15.4)
The driving force for the deprotonation of the organic ion is the formation of strong
ionic hydrogen bonds in the solvent clusters. The solvent clusters, e.g. (H2 O)n H+
can be attached to a radical carbon site through a relatively weak (H2 O)n H+ · · · C
hydrogen bond, or it can form a stronger (H2 O)n H+ · · · N bond to a basic site of the
radical such as in pyridine or pyrimidine. These are a new class of solvated clusters
where a protonated water cluster is hydrogen bonded to an organic radical. In other
words, in these clusters the deprotonated radicals are solvated externally, i.e., the
radicals remain outside, and bonded to the surface, of the solvent clusters.
The thermochemistry of the overall reaction (15.3 or 15.4) is relevant if the reaction
proceeds effectively in one step, i.e., if the cluster is assembled without stabilization,
retaining the exothermicity of all the association steps as internal energy for deproto-
nation/dissociation to form products. This may be possible in the experiments that use
neat water (solvent) vapor, where collision with every successive H2 O molecule can
lead to association with the release of the binding energy into internal energy, rather
than removing energy from the cluster. However, if a third body is present, it would
stabilize the growing cluster and no internal energy would be available for deprotona-
tion/dissociation. In fact, the third-body effect was observed in the benzene+ /water
system where the addition of He buffer gas quenched the deprotonation reaction [15].
The observation of IPT reactions in the benzene•+ /water system is consistent with
the IR dissociation spectra of the C6 H•+ 6 (H2 O)n clusters which were very similar
to the spectra of (H2 O)n H+ after n = 4, indicating that intracluster proton trans-
fer resulted in the formation of a protonated water cluster attached to the phenyl
radical [54]. However, H/D-exchange experiments under thermal ion mobility con-
ditions showed that C6 H•+ 6 (D2 O)n clusters with n = 2–8 did not exchange a proton
with D2 O to yield a C6 H•5 (D2 O)n D+ ion, although this would be expected for the
C6 H•5 (D2 O)n H+ structure [15]. It was, therefore, suggested that the intracluster pro-
ton transfer has an energy barrier that cannot be overcome in thermal systems below
280 K in the ion mobility experiments [15].
The deprotonation of organic ions such as C3 H+ •+
3 and C6 H6 requires the assem-
bly of several solvent molecules (at least 4 for H2 O) which is strongly facilitated
at lower temperatures. Therefore, the rate coefficients of these reactions showed
uniquely large negative temperature coefficients, with pseudo second-order coeffi-
cients varying as k = cT−63 and k = cT−67 - in the hydrated C3 H+ •+
3 and C6 H6 ions,
respectively [15, 16]. The large negative temperature coefficients result from a multi-
body mechanism in which five or more components need to be assembled in the
activated complex [for example, C6 H•+ ∗
6 (H2 O)4 ] for the reaction to proceed. This
means that only the fraction of the total cluster population that is in the n ≥ 4 clus-
ters is reactive. Therefore, only this fraction among all the collisions of the clusters
with H2 O molecules are reactive, and the small collision efficiency of 10−5 –10−6
reflects this effect [15, 16]. The population of the large reactive clusters increases
rapidly with decreasing temperature, and this leads to the large negative temperature
coefficient.
IPT reactions were not observed in the naphthalene •+ (H2 O)n clusters because of
the higher proton affinity (PA) of the naphthalene radical (C10 H•7 , 234 kcal/mol) [73]
as compared to that of the phenyl radical (C6 H•5 , 212 kcal/mol) [15]. This means
that more than seven or eight water molecules will be required to associate with the
naphthalene radical cation for the IPT to occur. However, IPT reactions can occur
within the C10 H•+ •+
8 (CH3 OH)n clusters if the proton transfer from the C10 H8 radical
cation to the methanol sub-cluster (CH3 OH)n becomes exothermic and the resulting
energy is sufficient to dissociate the C10 H•7 · · · H+ (CH3 OH)n cluster. Based on the
estimates of the PAs of the naphthalene radical (C10 H•7 ) [73] and the methanol clusters
[74], the PT reaction from C10 H•+ •+
8 to the methanol sub-cluster in C10 H8 (CH3 OH)n
is expected to becomes exothermic at n > 5. This is consistent with the changes in
the mass spectra of the C10 H•+ 8 (CH3 OH)n clusters at 238 K where the ion signal
of the naphthalene-containing ions disappeared and only the H+ (CH3 OH)n clusters
with n ≥ 5 were observed [19].
The calculated the structures of the clusters formed by the association of the
naphthalene radical (C10 H•7 ) with protonated methanol [H+ (CH3 OH)n ] clusters with
n = 4−6 and their binding energies are shown in Fig. 15.12.
Fig. 15.12 Structures and binding energies of the lowest energy structures of the C10 H7• .
H+ (CH3 OH)n clusters with n = 4 − 6 obtained at the B3LYP/6-311++ G(d, p) level [19]
It is clear that the binding energy of the protonated methanol cluster to the naph-
thalene radical decreases as the size of the protonated methanol cluster (n) increases.
For example, the binding energies of the H+ (CH3 OH)n clusters to the C10 H•7 radical
decrease from 6.5 to 5.4 to 4.5 for n = 4, 5 and 6, respectively according to the
B3LYP/6-311++ G(d, p) calaculations [19]. This can be explained by the decrease
in the charge-induced dipole interaction between the protonated methanol cluster
and the naphthalene radical as the effective charge becomes more delocalized on
the methanol cluster and further away from the radical. Eventually, the effect of the
charge would disappear and the binding energy would converge to that of the neutral
system. Since both the intracluster PT from the naphthalene radical cation to the
methanol sub-cluster in the C10 H•+8 (CH3 OH)n clusters becomes more exothermic as
n increases and the binding energies of the resulting C10 H•7 · · · H+ (CH3 OH)n clusters
decrease as n increases, dissociation of the larger C10 H+•
8 (CH3 OH)n clusters into the
C10 H•7 radical and protonated methanol clusters is expected to be the predominant
process at low temperatures. This is consistent with the experimental observation
mentioned above [19].
In general, IPT reactions are controlled by the relative PAs of the organic radical
and the cluster of solvent molecules and the extent of bonding interaction between
the organic radical and the protonated solvent cluster. The transfer of the proton
from the organic radical cation to the solvent molecules generates a protonated sol-
vent cluster such as (H2 O)n H+ which can be hydrogen bonded to the radical ring
carbon site [(H2 O)n H+ ···C] or at a basic nitrogen site [(H2 O)n H+ ···N]. These are
a new class of solvated clusters where a protonated solvent cluster is hydrogen bonded
to an organic radical. In other words, in these clusters the deprotonated radicals are
solvated externally, i.e., the radicals remain outside, and bonded to the surface, of the
solvent clusters. Alternatively, the protonated solvent cluster can be detached from
the deprotonated radical to generate a free protonated cluster (H2 O)n H+ and a free
organic radical (R• ). The variation of the radical/cluster bonding energy with cluster
size is of interest, and more experimental and theoretical works are needed in order
to understand the energetics and dynamics of these solvated clusters.
Acknowledgment We thank the National Science Foundation (CHE-0911146) for the support of
this work.
References
1. Meot-Ner (Mautner), M (2012) Update 1 of: strong ionic hydrogen bonds. Chem Rev
112:PR22–103; (2005) The Ionic hydrogen bond. Chem Rev 105:213–284
2. Tsuzuki S (2005) In: Wales DJ (ed) Intermolecular Forces and Clusters I, vol. 115. Springer,
Berlin, p 149
3. Cerny J, Hobza P (2007) Non-covalent interactions in biomacromolecules. Phys Chem Chem
Phys 9:5291
4. Conway BE (1981) Ionic hydration in chemistry and biophysics. Elsevier, Amsterdam
5. Sloan EDMD (2008) Clathrate hydrates of natural gases, vol 3. CRC Press, Florida
6. Nuevo M, Milam SN, Sandford SA, Elsila J, Dworkin JP (2009) Formation of uracil from the
ultraviolet photo-irradiation of pyrimidine in pure H2 O ices. Astrobiol 9:683–695
7. Menor-Salvan C, Marin-Yaseli MR (2012) Prebiotic chemistry in eutectic solutions at the
water–ice matrix. Chem Soc Rev 41:5404–5415
8. Ben-Naim A (1980) Hydrophobic interactions. Plenum, New York
9. Tanford C (1980) The hydrophobic effect: formation of micelles and biological membranes,
2nd edn. Wiley, New York
10. Gudipati RS, Allamandola LJ (2006) Unusual stability of polycyclic aromatic hydrocarbon
radical cations in amorphous water ices up to 120 K: astronomical implications. Astrophys J
638:286–292
11. Rhee YM, Lee TJ, Gudipati MS, Allamandola LJ, Head-Gordon, M (2007) Charged polycyclic
aromatic hydrocarbon clusters and the galactic extended red emission. Proc Natl Acad Sci U
S A 104:5274–5278
12. Snow TP, Bierbaum VM (2008) Ion chemistry in the interstellar medium. Annul Rev Anal
Chem 1:229
13. Herbst E, van Dishoeck EF (2009) Complex organic interstellar molecules. Annul Rev Astron
Astrophys 47:427–480
14. Ibrahim, Y, Alsharaeh E, Dias K, Meot-Ner M, El-Shall MS (2004) Stepwise hydration, and
multi-body deprotonation with steep negative temperature dependence in the Benzene∞ + –
water system. J Am Chem Soc 127:12766–12767
15. Ibrahim Y; Alsharaeh E, Meot-Ner M, El-Shall MS, Scheiner S (2005) Stepwise hydration of
ionized aromatics, energies, structures of the hydrated benzene cation, and the mechanism of
deprotonaion reactions. J Am Chem Soc 127:7053–7064
16. Mabrouki R, Ibrahim Y, Xie E, Meot-Ner M, El-Shall MS (2006) Gas phase hydration and
deprotonation of the cyclic C3 H+3 cation. Solvation with acetonitrile and comparison with the
benzene radical cation. J Phys Chem A 110:7334–7344
17. Momoh PO, El-Shall MS (2008) Gas phase hydration of organic ions. Phys Chem Chem Phys
10:4827–4834
18. Momoh PO, Hamid AM, Abrash SA, El-Shall MS (2011) Structure and hydration of the C4 H•+ 4
Ion formed by electron impact ionization of acetylene clusters. J Chem Phys 134:204315
19. Attah IK, Platt SP, Meot-Ner (Mautner) M, Aziz SG, Alyoubi AO, El-Shall MS (2014) Hydro-
gen bonding of the naphthalene radical cation to water and methanol and attachment of the
naphthalene ion to extended hydrogen bonding chains. Chem Phys Lett 613:45–53
20. Ibrahim Y, Mabrouki R, Meot-Ner M, El-Shall MS (2007) Hydrogen bonding interactions of
pyridine+ with water: stepwise solvation of distonic cations. J Phys Chem A 111:1006–1014
21. Hamid AM, Sharma P, Hilal R, Elroby S, Aziz SG, Alyoubi AO, El-Shall MS (2013) Hydration
of the pyrimidine radical cation and stepwise solvation of protonated pyrimidine with water,
methanol and acetonitrile. J Chem Phys 139:084304
22. NIST Chemistry Web Book NIST Standard Reference Database Number 69; National Institute
of Standards and Technology: Gaithersburg MD, 20899 (http://webbook.nist.gov)
23. Li QB, Jacob DJ, Yantosca RM, Heald CL, Singh HB, Koike M, Zho YJ, Sachse GW, Streets
DG (2003) A global 3-D model evaluation of the atmospheric budgets of HCN and CH3 CN:
constraints from aircraft measurements over the western pacific. J Geophys Atmos 108:8827
24. Matthews CN, Ludicky RA (1992) Hydrogen cyanide polymers on comets. Adv Space Res
12:21–32
25. Matthews CN, Minard RD (2006) Hydrogen cyanide polymers, comets and the origin of life.
Faraday Discuss 133:393–401
26. Pizzarello S, Holmes W (2009) Nitrogen-containing compounds in two CR2 meteorites: 15N
composition, molecular distribution and precursor molecules. Geochim Cosmochim Acta
73:2150–2162
27. Mumma MJ, DiSanti MA, Magee-Sauer K, Bonev BP, Villanueva GL, Kawakita H, Russo
ND, Gibb EL, Blake GA, Lyke JE, Campbell RD, Aycock J, Conrad A, Hill GM (2005) Parent
volatiles in comet 9P/Tempel 1: before and after impact. Science 310:270
28. Lahuis F, van Dishoeck EF, Boogert ACA, Pontopiddan KM, Blake GA, Dullemond CP, Evans
II NJ, Hogerheijde MR, Jørgensen JK, Kessler-Silacci JE, Knez C (2006) Hot organic molecules
toward a young low-mass star: a look at inner disk chemistry. Astrophys J Lett 636:L145
29. Haupert LJ, Enthold PG (2013) Hydration energies of aromatic ions in the gas phase. J Phys
Chem A 117:1164–1170
30. Wyttenbach T, Bowers MT (2009) Hydration of biomolecules. Chem Phys Lett 480:1–16
31. Gao B, Wyttenbach T, Bowers MT (2009) Hydration of protonated aromatic amino acids:
phenylalanine, tryptophan, and tyrosine. J Am Chem Soc 131:4695–4701
32. Liu D, Wyttenbach T, Bowers MT (2006) Hydration of mononucleotides. J Am Chem Soc
128:15155–15163
33. Kebarle P (1997) Ion thermochemistry and solvation from gas phase ion equilibria. Annu Rev
Phys Chem 28:445
34. Keesee RG, Castleman AW (1986) Thermochemical data on gasâŁłphase ionâŁłmolecule
association and clustering reactions. J Phys Chem Ref Data 15:1011
35. Larson JW, McMahon TB (1987) Hydrogen bonding in gas-phase anions. The energetics of
interaction between cyanide ion and Broensted acids determined from ion cyclotron resonance
cyanide exchange equilibria. J Am Chem Soc 109:6230
36. Gorman GS, Amster JI (1993) Gas-phase basicity measurements of dipeptides that contain
valine. J Am Chem Soc 115:5729
37. Gross DS, Schnier PD, Rodriguez-Cruz SE, Fagerquist CK, Williams ER (1996) Conformations
and folding of lysozyme ions in vacuo. Proc Natl Acad Sci U S A 93:3143
38. Deakyne CA (1997) In: Scheiner S (ed) Molecular interactions: from van der Waals to strongly
bound complexes. Wiley, New York, p 217
39. Meot-Ner (Mautner) M (2003) The proton affinity scale, and effects of ion structure and
solvation. Int J Mass Spectrom 227:525
40. Meot-Ner (Mautner) M, Lias SG (2003) Binding energies between ions and molecules, and
the thermochemistry of cluster ions. In: Linstrom PJ, Mallard WG (eds) NIST chemistry
webbook. NIST Standard Reference Database Number 69; National Institute of Standards and
Technology: Gaithersburg, MD, March 2003 (http://webbook.nist.gov)
41. Wyttenbach T, Bowers MT (2007) Intermolecular interactions in biomolecular systems
examined by mass spectrometry. Annu Rev Phys Chem 58:511
42. Kemper PR, Dupuis NF, Bowers MT (2009) A new, higher resolution, ion mobility mass
spectrometer. Int J Mass Spectrom 287:46
43. Momoh PO, El-Shall MS (2007) Stepwise hydration of ionized acetylene trimer. Further
evidence for the formation of benzene radical cation. Chem Phys Letters 436:25–29
44. Momoh PO, Xie E, Abrash SA Meot-Ner M, El-Shall MS (2008) Gas phase reactions between
acetylene radical cation and water. Energies, structures and formation mechanism of C2 H3 O+
and C2 H4 O+ ions. J Phys Chem A 112:6066–6073
45. Hamid AM, Soliman AR, El-Shall MS (2013) Stepwise association of hydrogen cyanide and
acetonitrile with the benzene radical cation: structures and binding energies of (C6 H•+
6 )(HCN)n ,
n = 1–6, and (C6 H•+6 )(CH3 CN)n , n = 1–4, clusters. J Phys Chem A 117:1069–1078
46. Hamid AM, Soliman AR, El-Shall MS (2012) Assembly of HCN hydrogen bonding chains in
the gas phase. binding energies and structures of phenylacetylene+ (HCN)n clusters, n = 1–4.
47. Attah IK, Hamid AM, Meot-Ner (Mautner) M, Aziz SG, Alyoubi AO, El-Shall MS (2013)
Substituent effects on non-covalent bonds: complexes of ionized benzene derivatives with
hydrogen cyanide. J Phys Chem A 117:10588–10597
48. Hamid AM, El-Shall MS, Hilal R, Elroby S, Aziz SG (2014) Unconventional hydrogen bonding
to organic ions in the gas phase. Stepwise association of hydrogen cyanide with the pyridine
and pyrimidine radical cations and protonated pyridine. J Chem Phys 141:054305–054315
49. Ludwig R (2001) Water: from clusters to the bulk. Angew Chem Int Ed Engl 40:1808
50. Gruenloh CJ, Carney JR, Hagemeister FC, Arrington CA, Zwier TS, Fredericks SY, Wood JT
III, Jordan KD (1998) Resonant ion-dip infrared spectroscopy of the S4 and D2d water octamers
in benzene-(water)8 and benzene2 -(water)8 . J Chem Phys 109:6601
51. Lum K, Chandler D, Weeks JD (1999) Hydrophobicity at small and large length scales. J Phys
Chem B 103:4570
52. Solca N, Dopfer O (2001) IR spectrum of the benzene–water cation: direct evidence for a
hydrogen-bonded charge–dipole complex. Chem Phys Lett 347:59–64
53. Miyazaki M, Fujii A, Ebata T, Mikami N (2001) Infrared spectroscopy of the benzene–H2 O
cluster cation: experimental study on the drastic structural change upon photoionization. Chem
Phys Lett 349:431–436
54. Miyazaki M, FujiiA, Ebata T, Mikami N (2003) Infrared spectroscopy of hydrated benzene clus-
ter cations, [C6 H6 -(H2 O)n ] + (n = 1–6): Structural changes upon photoionization and proton
transfer reactions. Phys Chem Chem Phys 5:1137–1148
55. Solca N, Dopfer O (2003) Prototype microsolvation of aromatic hydrocarbon cations by polar
ligands: ir spectra of benzene+ -Ln clusters (L) H2 O, CH3 OH). J Phys Chem A 107:4046–4055
56. Tachikawa H, Igarashi M, Ishibashi T (2001) Vibrational frequency-shifts of H2O caused by
complex formation with a molecular cation: a density functional study. Phys Chem Chem Phys
3:3052–3056
57. Tachikawa H, Igarashi M (1998) Dynamics of the ionization processes of benzene−H2O
clusters: a direct ab initio dynamics study. J Phys Chem A 102:8648–8656
58. Miyazaki M, Fujii A, Mikami N (2004) Binding energy of the benzene−water cluster cation:
an ar-mediated ir photodissociation study. J Phys Chem A 108:8269–8272
59. Liu K, Cruzan JD, Saykally RJ (1996) Water clusters. Science 271:929–933
60. Fujii A, Patwari GN, Ebata T, Mikami N (2002) Vibrational spectroscopic evidence of
unconventional hydrogen bonds. Inter J Mass Spetrom 220:289
61. Mehata MS (2008) Proton translocation and electronic relaxation along a hydrogen-bonded
molecular wire in a 6-hydroxyquinoline/acetic acid complex. J Phys Chem B 112:8383
62. Wang C, Batsanov AS, Bryce MR, Martin S, Nichols RJ, Higgins SJ, Garcia-Suarez VM,
Lambert CJ (2009) Oligoyne single molecule wires J Am Chem Soc 131:15647
63. Puigmarti-Luis J, Minoia A, Lei S, Geskin V, Li B, Lazzaroni R, De Feyter S, Amabilino
DB (2011) Self-assembly of supramolecular wires and cross-junctions and efficient electron
tunnelling across them. Chem Sci 2:1945
64. Dulmage WJ, Lipscomb WN (1951) The crystal structures of hydrogen cyanide, HCN. Acta
Crystallogr 4:330
65. Shallcross FV, Carpenter GB (1958) The crystal structure of cyanoacetylene. Acta Crystallogr
11:490
66. Jucks KW, Miller REJ (1988) Near infrared spectroscopic observation of the linear and cyclic
isomers of the hydrogen cyanide trimer. Chem Phys 88:6059
67. Ruoff RS, Emilsson T, Klots TD, Chuang C, Gutowsky HS (1988) Rotational spectrum and
structure of the linear HCN trimer. J Chem Phys 89:138
68. Karpfen A (1996) Linear and cyclic clusters of hydrogen cyanide and cyanoacetylene: a com-
parative ab initio and density functional study on cooperative hydrogen bonding. J Phys Chem
100:13474
69. Sanchez M, Provasi PF, Aucar GA, Alkorta I, Elguero J (2005) Theoretical study of HCN and
HNC neutral and charged clusters. J Phys Chem B 109:18189
70. Nauta K, Miller RE (1999) Nonequilibrium self-assembly of long chains of polar molecules
in superfluid helium. Science 283:1895
71. Meot-Ner (Mautner) M (1978) Solvation of the proton by hydrogen cyanide and acetonitrile.
Condensation of hydrogen cyanide with ions in the gas phase. J Am Chem Soc 100:4694
72. Meot-Ner (Mautner) M, Speller CV (1989) Multicomponent cluster ions. 2. Comparative
stabilities of cationic and anionic hydrogen-bonded networks. Mixed clusters of water and
hydrogen cyanide. J Phys Chem 93:3663
73. Feng WY, Lifshitz C (1996) Ion/molecule reactions of naphthalene and 1-chloronaphthalene
radical cations. Inter J Mass Spectrom Ion Proc 152:157
74. Knochenmuss R, Cheshnovsky O, Leutwyler S (1988) Proton transfer reactions in neutral
gas—phase clusters: 1-Naphthol with H2O, D2O, CH3OH, NH3 and piperidine. Chem Phys
Lett 144:317
Chapter 16
Anion-π Interactions in Supramolecular
Chemistry and Catalysis
Antonio Bauzá, Pere M. Deyà and Antonio Frontera
Abstract Non-covalent interactions play a major role in supramolecular chemistry

and biochemistry by dominating the central parts of living systems since they dictate
the functionality of many biological and host-guest systems. A good comprehen-
sion of the different non-covalent forces is necessary for the rational design of new
drugs and developing improved synthetic receptors capable to function in competi-
tive media. Interactions involving aromatic rings (or π-systems in general) are very
relevant in supramolecular chemistry, exemplified by the cation–π interaction and
its importance in protein structure and enzyme catalysis. From a traditional point of
view, the π-system is usually considered as electron rich (π-basic). The naissance of
the counterintuitive anion–π interaction –the attractive interaction between an anion
and an electron poor π-system (π-acid)– was somewhat controversially discussed
by the scientific community. However, in the last decade a great deal of theoret-
ical and experimental investigations has time-honored the anion–π interaction as
an important supramolecular bond. Herein we describe the physical nature of this
noncovalent interaction and the different strategies that can be used to modulate its
strength. Finally, selected state-of-the-art reports illustrating the rational utilization
of the anion–π interaction in supramolecular chemistry (anion receptors), biological
applications and catalysis are described in this chapter.
16.1 Introduction
Nowadays supramolecular chemistry is probably the most multidisciplinary field of

research. It is exponentially growing as indicated by the large number of articles,
reviews, and books published since the beginning of this millennium. The rapid
development of supramolecular chemistry has a profound effect on the increasing
efficiency for preparation of structures of different sizes, shapes and function-
alities. Supramolecular chemists rely on the comprehension of the non-covalent
forces, which form the basis of highly specific recognition, transport, and regulation
A. Frontera () · A. Bauzá · P. M. Deyà

Departament de Química, Universitat de les Illes Balears, 07122 Palma de Mallorca,
Baleares, Spain
e-mail: toni.frontera@uib.es
472 A. Bauzá et al.
mechanisms. The orchestration of many chemical and biological processes is often

dominated by an intricate combination of non-covalent interactions [1, 2], which are
the foundation of the life process itself, the ultimate expression of function. In general
host-guest chemistry, interactions between targeted guests and rationally designed
receptors drive the formation of assemblies of different sizes, shapes and affinities
[3–6]. The correct description of interactions between molecules is needed for the
understanding and progress of the supramolecular chemistry that usually relies on
strong, directional interactions such as hydrogen bonding and halogen bonding, and
less directional forces like ion pairing. More recently, general σ-hole interactions are
also considered as an important addition to the family of well-established directional
non-covalent interactions [7–9]. In addition, non-covalent interactions involving aro-
matic rings are enormously significant in this field [10, 11]. They play a crucial role
in chemistry and biology [12], in particular drug–receptor interactions, crystal en-
gineering, enzyme inhibition and protein folding [12]. A clear example is the role
of π–stacking interactions in DNA, RNA [14, 15] and protein-DNA aromatic inter-
actions [16]. An important and widely recognized non-covalent attractive force that
involves aromatic rings is the cation–π interaction [17, 18], of great significance
in biology [19–21] and supramolecular chemistry [22–25]. Other weak interactions
involving π-systems are also at the forefront of interdisciplinary research. For exam-
ple, CH–π [26, 27], lone pair–π [28, 29] and salt-bridge–π [30] interactions have
been recently used in several supramolecular chemistry fields, especially in crystal
engineering and protein-ligand interactions.
In the last decade, the anion–π interaction [31–33], i.e. the attractive noncovalent
force between an electron-deficient π–system and an anionic moiety [34, 35], has
also been recognized as a non-covalent bonding interaction. Its nature has been
described by a plethora of theoretical studies [36–39] in addition to an increasing
amount of experimental investigations [40–44]. Anion–π interactions are expected to
become noticeable players in fields as diverse as medicine, environmental chemistry
and biochemical processes [31, 45–47]. Moreover, their application to the design
of highly selective anion receptors and transport channels definitively confirms their
significance in the field of supramolecular chemistry [48, 49].
As a matter of fact, the design and synthesis of selective receptors designed to
bind anions is a topic of continuous interest [50–55]. The main reason is related to the
vital function played by anions, which are ubiquitous throughout biological systems
[56, 57]. In addition, some anions are increasingly recognized as problematic envi-
ronmental contaminants. For example, the overuse of anionic pollutants (phosphate
and nitrate) as fertilizers causes the disruption of aquatic ecosystems. The eutrophica-
tion caused by the fertilizing properties of these oxy-anions limits the populations of
algae and phytoplankton in rivers and lakes [58]. Other anions, such as pertechnetate
are generated from the re-processing of nuclear fuel [59] and their release to the sea
is strictly controlled or prohibited. Another harmful anion is arsenate that caused the
largest mass poisoning in human history (200 million) due to the dissolution in the
groundwater of the Bengal basin [60]. Interestingly, a new class of anion receptors
based on the anion–π interaction is emerging in the literature [31, 61]. For instance,
an interesting receptor, that combines hydrogen bonding and anion–π interaction
Fig. 16.1 a and b Host–Guest complexes reported by Schneider and coworkers. c Partial view of
the X-ray structure of thiotrithiazylium chloride
for the binding of anions with neutral π-acceptors, has been recently published by
Albrecht and collaborators [62] and its ability to trap anions both in solution and in
the solid state. Similarly, Johnson and collaborators have developed receptors for
selective nitrate binding in competitive hydrogen bonding solvents where anion–π
interactions facilitate the selectivity [63].
This chapter is not intended to be a bibliographic survey of the literature related
to anion–π interactions since several excellent reviews have been written for that
purpose [31, 35, 64–66]. Instead, we intend to emphasize the bonding relationship
between anions and π systems, at the experimental and theoretical forefront of this
noncovalent interaction. The chapter is developed under five headings. First, we
describe early publications on this topic, which have not been properly cited and
adequately placed into perspective. Second, the physical nature of the interaction is
explained for assimilation by a wide readership. Third, we highlight selected exam-
ples from the literature where molecular recognition driven by anion–π bonding is
very relevant. Next, attractive applications and wonderful examples of anion–π bond-
ing to catalysis are described. Finally, clear evidence that illustrates the increasing
interest on the study of anion–π interactions in biological systems is provided.
16.2 Early Publications
Early reports on weak attractive interactions involving negatively charged residues

and polarizable aryl groups in host–guest systems [67–69] were reported before
the publication of a series of computational studies in 2002 [36–38] supporting the
existence of attractive forces between anions and the positive electrostatic potential
on the ring edge of electron-deficient aromatic groups.
In 1993 Schneider and coworkers demonstrated attractive interactions between
simple organic host and guest molecules by means of NMR measurements [67].
Since their experiments were carried out in a highly competitive media (water) and
the aryl groups that participate in the anion–π interaction were not electron deficient
(see Fig. 16.1a and 16.1b), the calculated interactions energies were very modest,
reaching approximately 2 kJ/mol. A similar value was also obtained for the complex
between a calixarene with 4-sulfonato groups and toluene, which is not an electron
poor moiety. Thus, the attraction is only dominated by polarization effects (vide
infra). Earlier to this work, Hiraoka et al. [70] demonstrated the formation of highly
symmetric X− · · · C6 F6 anion–π complexes in the gas-phase (X = CI, Br, and I). In
addition, gas-phase clustering reactions X− @(C6 F6 )n − 1 + C6 F6 = X− @(C6 F6 )n for
X = F, C1, Br, and I were studied by means of pulsed electron-beam high-pressure
mass spectrometry. For fluoride a nucleophilic attack and covalent bond formation
was observed and, conversely, noncovalent X− · · · C6 F6 complexes were detected
for X = CI, Br, and I where the anion was located along the C6 main symmetry axis
of C6 F6 .
Another early progress in this direction [71] was reported in 1996 by Woollins
and collaborators. They named the anion–π interaction as “π-facial” [72] in their
description of the close contact between chloride and the seven-membered aromatic
[S4 N3 ]+ ring that they observe in the X-ray structure of thiotrithiazylium chloride
(see Fig. 16.1c).
16.3 Physical Nature
The physical nature of the anion–π interaction has been widely studied using high
level ab initio and Density Functional Theory (DFT) calculations. Moreover, several
partition energy schemes have been used in order to decompose the total interaction
energy into individual components [73, 74]. The general conclusion is that electro-
static forces and ion-induced polarization are the main forces that contribute to the
anion–π interaction [75–77]. The electrostatic term is explained by means of the
permanent quadrupole moment (Q) of the arene, which is the first non-zero multi-
pole moment in symmetric arenes. Moreover, in most asymmetric arenes, where the
dipole moment (μ) is non-zero, the μz component is approximately zero and, like-
wise symmetric arenes, the electrostatic attraction is basically due to the existence of
the adequate component (perpendicular to the ring plane) of the quadrupole moment
(Qzz ), which describes the charge distribution on both sides of the aromatic plane.
The Qzz of benzene is negative, but can be turned into positive by attaching electron
withdrawing substituents to the ring (see Fig. 16.2a). Similarly, the Qzz of pyridine
is negative but can be turned into a positive value by attaching a coordination metal
to the nitrogen atom [66]. Therefore, the electrostatic charge-quadrupole interac-
tion between an anion and an aromatic ring can become attractive either attaching
electron-withdrawing substituents or coordinating metal ions in case of heteroaro-
matic rings. The ion-induced polarization of the π-electron system by the anion is
significant, inducing a dipole (see Fig. 16.2b). Therefore, a polarization contribu-
tion to the total interaction energy is derived from the interaction of the anion with
the induced dipole [73–76]. An alternative explanation on the nature of anion–π
interactions involving benzene rings has been proposed by Wheeler and Houk [78],
who examined substituent effects in Cl− · · · · C6 H6 − n Xn complexes. In contrast
Fig. 16.2 a Molecular electrostatic potential of benzene and hexafluorobenzene. b Schematic

representation of the ion-induced dipole
to the intuitive view where the substituent induces changes in the aryl π system,
Wheeler and Houk propose a model where substituent effects in these systems can
be attributed mainly to direct interactions between the anion and local C–X dipoles.
A proper understanding of the physical nature (vide supra) is necessary to explain,
on one hand, the dual binding mode exhibited by arenes with negligible quadrupole
moments [79, 80]. That is, since electrostatic and polarization terms are the main
contributions to both anion–π and cation–π interactions, molecules with very small
Qzz values such as 1,3,5-trifluorobenzene (Qzz = 0.57 B) and s-triazine (Qzz = 0.90
B) are able to interact with both anions and cations because the electrostatic term is
negligible and the interaction is dominated by the polarization term, which is always
attractive. On the other hand, it is useful to rationalize the interaction of anions
with electron-rich aromatic rings, such as benzene, which are not strongly repulsive.
This is due to a compensating effect between the electrostatic (unfavorable) and ion-
induced polarization (favorable) forces that roughly cancel each other out. For this
reason, the anion–π interaction energy of benzene with chloride is very small, but
favorable [81, 82].
The electrostatic term depends on the magnitude of the Qzz and the polarization
term on the magnitude of the molecular polarizability parallel to the main symmetry
axis (denoted as α|| ) in symmetric arenes (or perpendicular to the ring plane in
asymmetric arenes, denoted as αzz ), which are intrinsic properties of the π-system.
Therefore, it is clear that in order to design an efficient anion receptor based on the
anion–π interaction, the π–binding units should have a large and positive quadrupole
moment and a large molecular polarizability. However, there is a limitation for the
Fig. 16.3 Interaction energies of pyrazine a and triazine b anion–π complexes from references
[84, 85]
first condition due to the reduced number of strong electron withdrawing groups
available for constructing the binding blocks. To have a large value of Qzz , the use of
-NO2 and -CN groups is required. However, it is synthetically complicated to attach
more than three strong electron withdrawing groups and the spacer to the aromatic
ring to build the receptor. An intelligent solution to this limitation is the utilization
of heteroaromatic rings, especially di-, tri- and tetrazines. The metal coordination to
heteroaromatic rings strongly increases the π-acidity of the ring, thus favoring the
anion–π interactions [83, 84]. It has been demonstrated that the coordination of both
pyrazine (Fig. 16.3a) and s-triazine (Fig. 16.3b) to AgI dramatically enhances their
ability to establish anion–π binding [83, 85].
The same behavior has been demonstrated [84] for s-tetrazine using both theory
and experiment. As a matter of fact, s-tetrazine coordinated to four AgI atoms is the
most powerful anion–π acceptor binding block reported to date. Theoretically, the
interaction energy of s-tetrazine with nitrate ion strengthens from − 9.6 kcal/mol to
− 62.4 kcal/mol when the arene is tetracoordinated to AgI . Experimentally, several
X-ray crystal structures of s-tetrazine μ4 –coordinated to AgI have been reported, ex-
hibiting very close contacts between the anion and the s-tetrazine ring (see Fig. 16.4)
indicating strong anion–π interactions, in agreement with theoretical predictions.
Interestingly the anion–π distance for the perchlorate anion in the X-ray structure
shown in Fig. 16.3a (2.61 Å) is the shortest reported to date.
The anion–π interaction has been also observed in the solid state of novel
hybrid inorganic–organic assemblies generated from H4 SiW12 O40 as Keggin-
type polyoxometalates (POM) and several trinuclear lanthanide clusters of type
{Na(H2 O)3 [Ln(HCAM)(H2 O)3 ]3 }4+ (Ln = La, Ce, Eu and H3 CAM = chelidamic
acid or 2,6-dicarboxy-4-hydroxypyridine) [86]. These unprecedented anion–π inter-
actions between the POM (a tetra-anion) and the coordinated aromatic ligand rings
play a crucial role in the crystal packing formation (see Fig. 16.5). This investigation
of the interaction in Keggin-type POM–based inorganic–organic frameworks also
Fig. 16.4 Fragments of the X-ray crystal structures containing μ4 -coordination of 1,2,4,5-tetrazine
from ref [84]. The relevant anion–π interactions are indicated by dashed lines (distances in Å)
Fig. 16.5 Partial view of the crystal packing of compound {Na(H2 O)3 [La(C7 H3 NO5 )(H2O)3 ]3 }
[SiW12 O40 ] published by Mirzaei et al. [86] with indication of the intermolecular anion–π
interaction established between the organic ligand and the POM (distance in Å)
includes a theoretical study devoted to analyze the effect of the ligand coordination
to the metal center that increases the π-acidity of the aromatic ring, and consequently
enhances its ability to establish anion–π interactions.
The coordination of the heteroaromatic ring to a transition metal is similar to the
effect of protonation. For instance, pyridine, diazines, etc. can be easily protonated
by simply adjusting the pH of the medium increasing the anion binding ability of
the ring (anion–π+ interactions). The geometric and energetic features of anion–π+
complexes of several aromatic cations (tropylium, quinolizinylium) and various an-
ions have been reported along with crystallographic structures [87, 88]. This field of
research has recently attracted attention and several works have appeared in the lit-
erature [89, 91]. For instance, the anion–π+ interaction participates in the formation
of a robust recognition motif in the transition metal malonate complexes using pro-
tonated 2-amino-4-picoline and 2-aminopyridine as the auxiliary ligands [89, 90].
As expected, these complexes exhibit very large (>80 kcal/mol) interaction energies
Fig. 16.6 a Partial view of the X-ray crystal structure reported by Giese et al. [62] showing the
pentafluorobenzamide, the bromide and the tetraethylammonium cation. b and c Modeled binding
motifs for the interaction of bromide with a receptor containing two pentafluorobenzamide moieties.
Distances in Å
due to strong electrostatic effects that dominate the interaction. The anion–π+ in-
teraction in protonated purine and pyrimidine bases has been recently reviewed [91]
demonstrating the importance of this interaction in biologically relevant compounds.
16.4 Anion-π Interactions in Supramolecular Chemistry
In this section of the chapter we describe very recent and especially relevant advances
in this field. The quality of the works demonstrates that there is a continuous and
increasing interest for investigating and developing novel anion-binding hosts and
transporters based on electron deficient aromatic rings.
Giese et al. [62] have studied the binding of a series of anions with neutral
π-acceptors by means of concurrent hydrogen bonding and anion-π interaction.
Interestingly, latter interaction is demonstrated both in the solid state and in
solution, and further evidenced by a computational study. The receptors are based
on pentafluorobenzamides (see Fig. 16.6), which were found appropriate systems
for studying anion-π interactions. In case of bromide, the anion–π complex was
characterized by X-ray spectroscopy and it is the first solid state structure where
anion-π interactions between an uncharged pentafluorophenyl derivative and an
anion are observed. Moreover, the investigation in solution showed differences
between electron rich and poor systems, which are explained by a cooperative effect
of N–H · · · anion and anion–π interactions and the enhanced acidity of the amide
proton by the electron-withdrawing C6 F5 -unit.
Remarkably, Watt et al. [63] have demonstrated that the selective nitrate binding
in competitive media by a tripodal urea receptor (see Fig. 16.7) is facilitated by
anion–π interactions. Using1 H–NMR titrations they show that the higher affinity
observed for nitrate over the halides for the fluorinated receptor is lost when the
fluorine atoms are absent. An anion–π interaction between the nitrate and the π-
system of the ethynyl-substituted arene is proposed as the source of this selectivity.
tBu NO2
O
N N
H H
O2N
NH X X
O NH
tBu
X
HN
tBu
HN O
X=F nitrate selectivity
X=H
NO2
Fig. 16.7 Tripodal urea receptors synthesized by Watt et al. [63]
The fluorinated receptor trends: NO− − − −

3 > Cl > Br > I , with a moderate selectivity
for nitrate. The existence of the anion–π interaction is demonstrated using an indirect
prove. That is, binding studies of the non-fluorinated receptor showed three important
issues: first, the nitrate selectivity is lost; second, the association constant of nitrate
is diminished compared to the fluorinated receptor; and, third, the binding mode
of nitrate is different than that for halides. These data support a model in which
nitrate binds to the receptors through a combination of hydrogen bonds with the urea
moieties and an anion–π-type interaction regardless of the electronic nature of the
central core, thus raising the association constant and selectivity for nitrate in the
fluorinated receptor and the opposite in the non-fluorinated. Latter receptor is more
selective for chloride since it binds the receptors via hydrogen bonding interactions,
exclusively.
Anion–π interactions have been systematically studied by Wang and Wang [92] us-
ing tetraoxacalix[2]arene[2]triazine (see Fig. 16.8), an electron-deficient and neutral
macrocyclic host. Using electrospray ionization mass spectrometry (ESI-MS), fluo-
rescence titration and X-ray crystallography, the authors demonstrate the formation of
1:1 host-guest complexes with four typical polyatomic anions of different geometries
and shapes in the gaseous phase, in solution, and in the solid state. The association
constants for the formation of anion–π complexes in acetonitrile are impressive in
some cases, ranging from 239 to 16950 M−1 , (NO− − − −
3 > BF4 > PF6 > SCN ). The
X-ray molecular structures of the complexes show that two opposed triazine rings
of the host interact with the anionic guests through cooperative anion–π and lp–π
Fig. 16.8 Crystal structures of the 1:1 host-guest complexes reported by Wang and Wang [92]
Distances in Å
Fig. 16.9 Two views in ball and stick a and space filling b of the crystal structure of the triangular
prism complexed to I−
3 reported by Stoddart and coworkers [93]
interactions. In this comprehensive study, the generality and diversity of anion–π

interaction motifs certainly provide a new dimension to molecular recognition of
anions and anion-governed self-assembly processes.
Stoddart et al. [93] have reported an interesting anion–π recognition study in
molecular triangular prisms formed by naphthalenediimide redox centers. They have
demonstrated, both experimentally and by the application of theory, through-space
orbital interactions and electron sharing phenomena in their synthesized triangu-
lar, redox-active naphthalenediimide prisms (see Fig. 16.9). The resulting electronic
communication among the naphthalenediimide units leads to an unusually large num-
ber of individually accessible redox states in the triangular prisms, opening the door
to potential applications in the field of molecular electronics. The electron deficient
cavities of the molecular prisms are ideal for studying anion–π interactions. This
ability is demonstrated by the encapsulation of linear I−3 anions inside the prismatic
cavities, causing a profound change in the packing of the prisms in the extended
solid-state architecture. The inclusion of I−
3 anions induces π–π stacking of the chi-
ral prisms into supramolecular helices, providing a unique example of anion-induced
self-assembly with potential applications as ion-channels. In addition, the chirality
endowed by the six stereogenic centers in the occupied prisms dictates the either
right- or left-handed associated with their packing in the solid state.
Fig. 16.10 Anion-templated X-ray structures of a molecular square a and a molecular pentagon b
reported by Chifotides et al. [94]
Taking advantage of the enhanced π-acidity of organic ligands upon coordination

to transition metals, Chifotides et al. [94] have reported supramolecular architectures
with 3,6-bis(2-pyridyl)-s-tetrazine cavities where anion–π interactions participate in
the remarkable stability of Fe(II) metallacycles in solution. This comprehensive in-
vestigation provides convincing evidence that anion–π interactions are the main
driving force in the formation (in high yields) of self-assembled Fe(II)-templated
metallacycles. Combining several technics, like X-ray crystallography, 1 H NMR,
solution and solid-state19 F NMR spectroscopies, cyclic voltammetry and mass spec-
troscopy, they demonstrated that the anion acts as a template occupying the π–acidic
cavities and controlling the nuclearity of the cages. That is, [BF4 ]− and [ClO4 ]−
anions template molecular squares (see Fig. 16.10a) and [SbF6 ]− , [AsF6 ]− and
[PF6 ]− anions template molecular pentagons (see Fig. 16.10b) establishing close di-
rectional F · · · C s-tetrazine contacts with the s-tetrazine rings that are up to 0.4 Å
shorter than the sum of the F · · · C van der Waals radii (3.17 Å). The number and
strength of F · · · C tetrazine contacts are maximized. They have also performed un-
precedented solid-state19 F MAS NMR studies, where the templating anions showed
downfield chemical shifts Δδ(19 F) ranging 3.5–4.0 ppm (compared to peripheral an-
ions) due to their participation in anion–π interactions. NMR, cyclic voltammetry
and mass spectroscopy studies also establish that the molecular squares and pen-
tagons remain intact in solution. This study provides unambiguous evidence that
anion–π interactions are the main driving force in the templation process leading
to the formation of Fe(II) metallacycles with π-acidic cavities. The F atoms of the
encapsulated anions are directly located over the more π-acidic s-tetrazine C atoms,
establishing six simultaneous anion–π contacts with the metallacycle edges. More
importantly, they evidenced the instrumental role of the templating anions in solution,
thus favoring and stabilizing the Fe(II) metallacycles of specific nuclearities.
Fig. 16.11 Representation of the neutral hosts designed and synthesized by Meyer’s group [95]
Meyer’s group [95] has designed and synthesized pre-organized anion traps (Cl−
and Br− ) for exploiting anion–π interactions based on 1,3-bis(pentafluorophenyl–
imino) isoindoline (Fig. 16.11a) and 3,6-di-t-butyl-1,8-bis(pentafluorophenyl)-9H–
carbazole (Fig. 16.11b) in various solvents. Both neutral receptors provide a central
N–H · · · X− hydrogen bond that directs the halide anion into a pre-organized clamp
formed by the two electron deficient arenes. Crystal structures of host–guest com-
plexes reveal that in all cases the guest is located in the cleft between the perfluorinated
flaps. In solution, association constants up to 960 M−1 have been determined de-
pending on the solvent by NMR spectroscopy. Their study also includes a detailed
computational analysis of the host–guest complexes and an energetic decomposition
of the ring–anion interactions that confirm the contribution of the anion–π inter-
actions to the stabilization of these complexes (∼50 % of the total energy). These
receptors contribute to increasing the relative low number of examples of neutral
receptors that are well pre-organized for exploiting anion–π interactions.
Ballester and his group have dedicated much effort to the experimental quantifica-
tion of anion–π interactions in solution using neutral host–guest model systems [96].
The quantification of anion–π interactions is commonly provided by computational
studies of simple models that are useful to estimate the binding energy. The scientific
community has no doubt about the existence of attractive anion-π interactions in the
gas phase and in the solid state. However, there are still few examples of attractive
anion-π interactions in solution. Ballester’s group has reported several examples of
neutral molecular receptors that bind anions in solution as a combination of anion–π
interactions and hydrogen bonding [97, 98]. The strength of the anion–π interaction
is indirectly detected as a modulation of the stronger hydrogen bonding interaction.
The dissection of the energy contribution of the anion–π interaction to the overall
binding is complex and requires the use of appropriate reference systems. Ballester
and coworkers have designed a model system based on a series of “four wall” aryl-
extended calix[4]pyrrole receptors. They contain deep aromatic cavities with fixed
walls (see Fig. 16.12a, for the nitro-derivative as example). The formation of four
concurrent H-bonds between the anion and the NH groups of the calix[4]pyrrole
Fig. 16.12 Structures of a four wall calix[4]pyrrole a and the receptor used by Ballester’s group
[97] as reference b for the estimation of chloride–π interactions. Distance in Å
scaffold fixes the halide in the aromatic cavity, above the planes of the π-systems of
the four meso-aryl substituents, as probed using 1 H NMR spectroscopy. Different
para substituents in the aromatic walls were used to tune the electronic density of
the aromatic rings and Cl− as the interacting anion. Interestingly, the magnitude
of the association constant was increased with the electron-withdrawing charac-
ter of the para-substituent in the meso-aryl groups. The difference in free energy
(ΔΔG, see Fig. 16.12) of binding between different complexes provides a direct
measurement of the relative interaction energy of the halide with the different aro-
matic systems. Using this approach, they determined a maximum contribution of
−4.4 kcal/mol for the four chloride–π interactions to the overall binding free energy
in the para-nitroaryl substituted receptor. This is likely an underestimation of the
anion–π interaction energy because the reference system used by Ballester’s group
was octamethylcalix[4]pyrrole, which provides four C–H · · · Cl interactions (see
Fig. 16.12b). Therefore, the estimated contribution of –4.4 kcal/mol likely means
that each anion–π interaction is 1.1 kcal/mol more favorable than each C–H · · · Cl−
interaction that is established in the reference complex and, consequently, it cannot
be used as an absolute estimate of the anion–π interaction in solution.
Calix[4]pyrrole based receptors featuring two additional pyrrole side arms have
been also used by Chang et al. [99] for the molecular recognition of anions. This
hexapyrroliccalix[4]pyrrole (see Fig. 16.13) has two additional pyrrole suspended
above or below the calix[4]pyrrole core and presents cis/trans isomerism. Anion
binding experiments revealed interesting differences in the binding mode depending
on the isomer. That is, whilst the trans isomer displays only hydrogen bonding
interactions, the cis isomer displays a mixed binding mode featuring a combination
of hydrogen bonding and anion–π interactions resulting in an unexpected strong
binding (see Fig. 16.13c). In fact, UV spectrophotometry and NMR titrations reveal
Fig. 16.13 X-ray structures of the trans a and cis b isomers of meso-substituted hexapyrrolic-
calix[4]pyrrolereceptors reported by Chang et al. [99] and the modeled complex c between chloride
and the cis isomer
that cis isomer displays higher affinity (105–106 M−1 ) for anions while the trans
isomer is more selective.
An interesting research has been recently published by He et al. [100] devoted
to the study of the stability of size-regulable vesicles based on anion–π interac-
tions Taking tetraoxacalix[2]arene[2]triazine as a functionalization platform (see
Fig. 16.14), He et al. synthesized a series of new amphiphilic anion receptors that
self-assemble into stable vesicles in a mixture of THF and water, with the sur-
face of the vesicles engineered by electron-deficient cavities. Strikingly, several
anions are able to influence the size of self-assembled vesicles selectively, fol-
μ μ μ
lowing the order of Fμ < ClO4 < SCNμ < BF4 < Brμ < Clμ < NO3 , as revealed by
dynamic light scattering (DLS) experiments and independently with the hydration
cost. This order of selectivity agrees with the binding strength of anions with tetraox-
acalix[2]arene[2]triazine receptor, demonstrating that the anion–π interaction most
probably competed over other possible weak interactions and is responsible for this
interesting selectivity. Furthermore, the chloride permeation process across the mem-
brane of the vesicles was also studied by He et al. using fluorescent experiments. This
investigation shows the potentiality of heteracalix aromatics as new models to con-
struct functional vesicles and gives a new dimensionality to the anion–π interaction
in aqueous medium and, potentially, in living systems.
Fig. 16.14 Schematic illustration of the vesicle (a and b) and long alkyl chain derivatives of
tetraoxacalix[2]arene[2]triazine compounds (c) reported by He et al. [100]
16.5 Catalysis
Zhao et al. reported in 2013 for the first time experimental evidence suggesting that
anion–π interactions contribute to the catalysis of the Kemp elimination reaction
(see Fig. 16.15) by π-acidic naphthalene diimides (NDI) leading to conceptually
innovative design strategies to stabilize anionic transition states [101]. Subsequent
studies [102] with modified sulfur-containing NDI catalysts confirmed the general
validity of increasing transition-state stabilization while increasing π-acidity with
regard to the Kemp elimination. Moreover, computational simulations are in excellent
agreement with experimental results, confirming that the stabilization of the anionic
transition states (but not the neutral ground states) increases with the π-acidity of
the catalysts, i.e., the existence of anion–π catalysis. The proposed catalytic cycle
is shown in Fig. 16.15 and the key point is the location of the carboxylate base on
the π-acidic surface of catalyst NDI. The initial substrate-NDI complex (NDI + S)
is likely stabilized by a combination of π–π and hydrogen bonding interactions. In
the transition state (TS), the negative charge flows over the π-acidic surface from the
carboxylate base over the carbanion of the conjugate base to the phenolate oxygen.
The TS is stabilized by the π-acidic surface of the NDI. The proton transfer from the
carboxylic acid to the phenolate in the intermediate (I) prevents product inhibition and
regenerates the catalyst NDI. Therefore, in this catalytic cycle, the NDI stabilizes the
Fig. 16.15 Catalytic cycle proposed by Zhao et al. [101] (NDI + S = catalyst–substrate complex,
I = reactive intermediate, NDI + P = catalyst–product complex)
anionic TS by means of anion–π interactions, which are proved by the acceleration

of the Kemp elimination observed experimentally. Both pioneering works [101, 102]
on catalysis with anion–π interactions are extremely important since clearly open a
new avenue and move beyond the grand principles operating in nature.
The catalytic cycle shown in Fig. 16.15 has been analyzed theoretically to examine
the most important aspects of the anion–π catalyzed Kemp elimination considering
solvent effects in the calculations [102]. Already in the early stage of the reaction
(NDI + S), the carboxylate anion is positioned so that an efficient intramolecular
anion–π interaction with the most π-acidic part of the NDI surface takes place. The
carboxylate group also anchors the benzisoxazole substrate above the NDI surface
via two C–H · · · O interactions (see Fig. 16.15), therefore favoring the π–π interac-
tion between the two parallel disposed aromatic systems. This facilitates the proton
transfer between the catalyst and the substrate, leading to the transition state TS
with the activation barrier of 15.05 kcal/mol (Fig. 16.16a). At this stage the electron
Fig. 16.16 a Free energy diagram (IEFPCM/M06-2X/def2-TZVP//B97-D/6-311G**) for the Kemp

elimination with anion–π catalyst NDI and NDICN . b Transition state (TS) for the reaction catalyzed
by catalyst NDI, negative charge transfer is highlighted in red. c Transition state (TSCN ) for the
reaction catalyzed by catalyst NDICN [101]
transfer, which occurs over several atoms from carboxylate anion to phenolate oxy-
gen, is efficiently stabilized by the π-acidic surface of NDI. The reaction progresses
toward the anionic intermediate I, while the negative charge is fully transferred to the
benzisoxazole substrate. The benzisoxazole oxygen accumulates most of the charge
and its distance from the NDI surface decreases to 2.995 Å. The conformation of this
complex once again favors anion–π interactions by placing the anionic oxygen right
above the preferential binding site of NDI. Interestingly, the Kemp elimination in the
presence of the 3,7-dicyano-substituted catalyst (NDICN , see Fig. 16.16c) follows a
similar pathway. However, the increased π-acidity enhances the TSCN transition-state
stabilization by 1.06 kcal/mol when compared to TS (Fig. 16.16a). This stabilization
enhancement of TSCN by the more π-acidic NDI surface of NDICN confirms that
anion–π interactions contribute significantly to this reaction. The comparison of cer-
tain geometric parameters during the reaction mechanism involving catalysts NDI
and NDICN also reflects enhanced anion–π implication. For instance the more pro-
nounced decrease of the distance between benzisoxazole oxygen and catalyst plane
on going from the initial complex to the TS is correlated with the higher strength of
the anion–π interaction in the NDICN .
Fig. 16.17 a Malonate covalently bonded to NDI reported by Matile’s group [103]. b DFT-
optimized geometry obtained for the enolate–π interaction
Matile’s group has taken one step further the research on anion–π catalysis [103]
by extending it to enolate chemistry. They have covalently attached a malonate moi-
ety to an NDI (see Fig. 16.17) and by means of 1 H NMR spectroscopy they have
compared the chemical shift of the acidic hydrogen atoms of free diethyl malonate
with the corresponding ones upon its attachment to the naphthalene diimide (denoted
as NDI–Mal, see Fig. 16.17a). As a result, the chemical shift of 3.38 ppm for the
acidic hydrogen atoms in free diethylmalonate changes to 1.78 ppm due to their ex-
posure to the naphthalene ring current in NDI–Mal that causes the upfield shift (see
Fig. 16.17a). This direct experimental evidence for the fixed covalent positioning of
the malonate above the π–acidic surface is very important because it assures that
any changes in acidity can be unambiguously attributed to the stabilization of the
enolate by anion–π interactions (see Fig. 16.17b). In fact, they have demonstrated
using 1 H NMR titrations that anion–π interactions stabilize the enolate by almost
two pKa units. Remarkably, the addition of these anion–π-stabilized reactive enolate
intermediates to enones and nitroolefins occurs with significant transition-state sta-
bilizations (up to 11 kJ/mol) [103]. Moreover, anionic cascade reactions that cover
aldol condensation, elimination and transesterification also accelerate on π-acidic
surfaces. These findings are very significant because enolate chemistry is funda-
mental in chemistry and biology. This research is expected to stimulate the use of
anion–π interactions in catalysis in the broadest sense.
16.6 Biologically Relevant Anion–π Interactions
Clear evidence that illustrates the growing interest in the study anion–π interactions
in biological systems is the development of software capable to search anion–π
contacts in the PDB and related biological databases. To this respect, the STAAR
(statistical analysis of aromatic rings) program [47] can identify anion–π interactions
in a large structural database of biomolecules. The program is freely available for
download through the web (http://staar.bio.utk.edu) and has been tested in a recent
version of the PDB demonstrating the high prevalence and relatively strong anion–π
energies involving side-chain/side-chain interactions in biomolecules. The program

is currently limited to phenylalanine residues and is expected to include tryptophan
and tyrosine residues in the new version. The program also provides pairwise inter-
action energies with Asp and Glu as anions. Other future project for improving the
program includes the possibility to investigate anion–π interaction in protein/ligand
complexes.
A pioneering work on the study of anion-π interaction in biomolecular recognition
was published in 2012 by Chakravarty et al. [46]. Combining theoretical and con-
vincing experimental evidences they demonstrated that the interaction between an
anion and an aromatic π system plays an important role in the formation and recog-
nition of biomolecular structures. Though less frequent than its counterpart cation–π
interactions, the examination of high-resolution structures of proteins and nucleic
acids indicated the presence anion–π interactions. Interestingly, they have been ob-
served unambiguously, occurring in protein/nucleic acid loops and often involving
conserved/coevolving sites in proteins. These findings suggest that this interaction
plays an important role in macromolecular folding and function. Two examples are
highlighted in Fig. 16.18. In the first one, the anion–π interaction is observed in
the nucleic acid backbone (tetraloop hairpin, 1m5o structure), where the guanine-
nucleotide (G75) establishes a clear anion–π interaction with the oxygen atom of
the phosphate anion. Apart from this interacting phosphate, the rest of the phosphate
anionic oxygen atoms in the nucleic acid backbone face outward. Remarkably, the
phosphate anion that faces inward is poorly solvated, facilitating the anion–π inter-
action (Fig. 16.18a, top). The nucleic acid base likely compensates for the energetic
cost of the desolvation of the anion. In many RNA hairpins, even in structures with
low resolutions the same interacting pattern is observed [46] suggesting that anion–π
interactions may be a regular feature of nucleic acid loops. The second example in-
volves an RNA-protein interaction that corresponds to a crucial biological process. It
is well-known [104] that the RRM (RNA recognition motif) of the U1A protein binds
to U1 snRNA hairpin II. The structure of the complex reveals that RRM Asp92 stacks
in parallel to the Cytosine-12 residue of the hairpin loop, establishing an anion–π
interaction. Amazingly, the same feature is observed in the structure of the RRM
of U2B” bound to hairpin IV of U2 snRNP (Fig. 16.18b). U2B” Asp92 stacks on
Guanine-12 of the hairpin loop of U2 and this arrangement is a conserved feature of
spliceosomal RNA hairpins involving a C/G12 and Asp92-carboxylates.
Frontera and coworkers [45] have studied the important role of anion–π interac-
tions in the active site of the urate oxidase enzyme (see Fig. 16.19). The inhibition
of this enzyme by cyanide ions is caused by the existence of an anion–π interaction
between the anionic inhibitor and the π-electron deficient enzymatic substrate (uric
acid). They demonstrated using high level ab initio calculations that the anion–π
interaction between the inhibitor and the substrate in a model of the active site is
energetically favorable. In addition, a favorable cooperativity effect between a π–π
stacking interaction involving the substrate and a phenylalanine residue (PHE159,
see Fig. 16.19) and the anion–π interaction also contributes to the binding energy.
This was the first example where the presence of an anion–π interaction between an
inhibitor and an enzymatic substrate was proposed to be crucial in the inhibition of
Fig. 16.18 a Anion–π interaction in nucleic acids between the base of the ith nucleotide with the
oxygen atoms of an (i + 2)th phosphate of a tetraloop hairpin. b Anion–π in RRM (split-pea green,
1a9n) of U2B (spliceosomalmachinary) in complex with U2 snRNA
Fig. 16.19 The anion–π interaction between the CN− inhibitor and the uric acid substrate (URC)
found in the X-ray structure 3bjp studied theoretically by Frontera’s group [45] is shown. Distances
in Å
an enzyme. This investigation extended the relevance of anion–π interactions to an

important field such as enzyme chemistry.
The relevance of anion–π interactions in flavoproteins has been analyzed by
Estarellas et al. [105]. An initial search in the PDB provided evidence of anion–
π interactions in a large number of proteins. Many of them are reductases that have
affinity for small negatively charged ions such as acetate, chloride and thiocyanate,
which are competitive inhibitors. Two interesting examples were described in their
investigation. In the first one (2ar8, tryptophan 7-halogenase), the anion binds the
FAD at the enzymatic center by means of an anion–π interaction and participates
in the enzymatic process [106]. This enzyme catalyzes the regioselective chlorina-
tion at the 7 position of tryptophan. The Cl− is bound near to the entrance of the
tunnel leading to the tryptophan and positioned to make a nucleophilic attack on the
flavin peroxide resulting in the formation of hydroxylated FAD and HClO that is
the real chlorination agent [106]. A close examination of the active site reveals that
the anion is located above the most electron deficient ring of the FAD establishing
an anion–π interaction. In the ternary complex (see Fig. 16.19a), the distance be-
tween the Cl− anion and the ring centroid of the pyrimidinic ring of FAD is 3.3 Å
in 2ar8. The second example is a FMN-dependent nitroreductase (1ylU) [107] that
plays a prominent role in the reduction of the antibiotics nitrofuranone and nitrofu-
rantoin to the hydroxylamine derivatives, which are the active antibacterial agents.
Inhibition studies of nitroreductase (NTR) have demonstrated that acetate displayed
competitive inhibition with respect to both the substrate and NADH [107]. More
interestingly, it has been demonstrated that acetate binds only to the oxidized form
of the enzyme. The crystal form of the enzyme complexed to the acetate and the
FMN coenzyme shows one acetate molecule in each active site (see Fig. 16.20b).
The acetate anion is bonded to the active site by means of a bifurcated hydrogen bond,
a −CH3 · · · π interaction and an anion–π interaction with the pyrimidinic ring of
FMN. The distance between the acetate anion and the ring centroid is short (3.0 Å)
indicating a strong interaction. This investigation further confirms the relevance of
anion–π interactions in the scientific field of enzyme chemistry.
Bauzá et al. [108] have studied the importance of anion-π interactions in the
mechanism of sulfide:quinone oxidoreductase. It is a flavin-dependent enzyme that
plays a physiological role in two important processes. First, it is responsible for
sulfide detoxification by oxidizing sulfide ions to elementary sulfur and the electrons
are first transferred to flavin adenine dinucleotide (FAD), which in turn passes them
to the quinone pool in the membrane [109]. Second, this enzyme plays a key role in
the sulfide-dependent respiration and anaerobic photosynthesis, deriving energy for
their growth from reduced sulfur in sulfidotrophic bacteria [110]. Two mechanisms
of action for this enzyme have been proposed [108] that involve a common anionic
intermediate that it is stabilized by a relevant anion–π interaction (INT–FAD, see
Fig. 16.21). The formation of the intermediate is facilitated by reducing the transition
state barrier, owing to an anion–π interaction that involves the π system of FAD. By
analyzing the X-ray structures of SQRs available in the Protein Data Bank (PDB)
and using DFT calculations, they have demonstrated the relevance of the anion–π
interaction in the enzymatic mechanism.
In particular, Bauzá et al. [108], have demonstrated theoretically using DFT cal-
culations (B3LYP/6-311++G**) that the energy barrier is reduced when the reaction
occurs in the present of FAD, using cyclo-L-cysteine (see Figure 16.21b) as model
of disulfide bond. In Fig. 16.22a the reaction coordinate diagram is represented
where the relative energies of the different species involved in the mechanism are
also indicated. It can be observed that the presence of FAD (in red) stabilizes all
species. However, the most important result is that the stabilization of the transi-
tion state is higher than the stabilization of the intermediates, thus supporting the
crucial role of FAD not only as an electron sink in subsequent enzymatic steps of
Fig. 16.20 3D X-ray structures of 2ar8 (a) and 1ylu (b) are shown and the anion–π interactions
observed in the FAD-Cl− (a) and FMN-acetate (b) complexes are highlighted. Distances in Å
the mechanism but also facilitating the initial addition-elimination reaction to take
place. Moreover, the presence of FAD stabilizes the reaction product placing it in an
ideal position either to attack or transfer an electron to the isoalloxazine ring. The
theoretical study also includes the computation of the NCI (Non Covalent Interac-
tion) plot to study the anion–π interactions observed in the structures retrieved from
the PDB. This visualization index based on the electron density and its derivatives
enables the identification and visualization of non-covalent interactions efficiently
[111]. The isosurfaces correspond to both favorable and unfavorable interactions,
as differentiated by the sign of the second density Hessian eigenvalue and defined
by the isosurface color. In Fig. 16.22b the representation of the NCI plot computed
around the FAD (within 4.0 Å) cofactor in 3SX6 protein is shown. Obviously, several
non-covalent regions clearly appear between the FAD and the aminoacids. For in-
stance, several green isosurfaces are found around the methyl groups of FAD, which
are characteristic of hydrophobic interactions. Interestingly the trisulfide group (rep-
resented in ball and stick) originates an extended isosurface that covers two rings of
FAD and confirms the importance of the anion–π interaction.
Finally, Bauzá et al. [112] have also studied long-range effects in anion–π in-
teractions and their crucial role in the inhibition mechanism of mycobacterium
Fig. 16.21 a Addition-elimination mechanism in the presence of FAD. b B3LYP/6-311 + G*

optimized complexes of INT-FAD and TS-FAD. Distances in Å
Fig. 16.22 a Addition-elimination mechanism in the presence of FAD. b B3LYP/6-311+G*

optimized complexes of INT-FAD and TS-FAD. Distances in Å
tuberculosis malate synthase. This enzyme together with isocitrate lyase forms the
glyoxylate shunt that is an anaplerotic bypass of the traditional Krebs cycle. It plays
a prominent role in Mycobacterium tuberculosis virulence, so it can be exploited for
the development of antitubercular therapeutics [113]. The shunt bypasses two steps
of the tricarboxylic acid cycle, allowing the incorporation of carbon, and thus, refill-
ing oxaloacetate under carbon-limiting conditions. A catalytic Mg2+ unit is located at
the bottom of the cavity, and plays a very important role. Recently, the development
of effective antituberculosis drugs based on phenyldiketo acids (PDKAs) has been
Fig. 16.23 3D X-ray structure of 3s9i is shown and the anion–π interaction observed in the active
site between aspartate-633 and the PDKA inhibitor is highlighted. Distance in Å
reported [114]. Interestingly, all the crystal structures of malate synthase–inhibitor

complexes exhibit close contact between the carboxylate of Asp633 and the face of
the aromatic ring of the inhibitor (see Fig. 16.23). Remarkably, the replacement of the
phenyl ring in PDKA by aliphatic moieties yields inactive inhibitors, suggesting that
the aromatic moiety is crucial for inhibition. However, the aromatic ring of PDKA
is not electron-deficient, and consequently, the anion–π interaction is expected to
be very weak (dominated only by polarization effects). Combining an analysis of
the recent X-ray structures of GlcB–PDKA complexes retrieved from the protein
data bank (PDB) and computational ab initio studies (RI-MP2/def2-TZVP level of
theory), Bauzá et al. [112] demonstrate the prominent role of the Mg2+ ion in the
active site, which promotes long-range enhancement of the anion–π interaction.
Specifically, in the computational analysis of the long range effects, they have
computed several theoretical models of the inhibitor and their complexes with for-
mate (as model of aspartate) to demonstrate the long range effect and to estimate
it energetically. Two possible enol forms (highlighted in green in Fig. 16.24) of
the diketo acid denoted as PDKA-taut1 and PDKA-taut2 were considered in the
energetic analysis, since it has been demonstrated experimentally that the PDKA
chemical structure is in its enol tautomer by solution-phase 1H– and 13C–NMR data
[114]. In addition, a theoretical model of the inhibitor coordinated to the Mg2+ ion
was also considered, denoted as PDKA-Mg. In this model (see Fig. 16.23c), a for-
mate ligand was added in order to keep the model neutral and to exemplify the active
site where the Mg2+ is coordinated to ASP462 (see Fig. 16.23). Interestingly, the
interaction energy of the anion–π complex of formate with PDKA-taut1 is − 4.5
kcal/mol and with PDKA-taut2 is almost negligible (− 0.2 kcal/mol) indicating that
the intramolecular hydrogen bonding is very important since it increases the electron
withdrawing ability of the carbonyl group attached to the ring. Interestingly, in the
PDKA-Mg complex, the interaction energy is large and negative (−10.6 kcal/mol),
indicating a strong binding. Therefore, the theoretical study clearly demonstrates
Fig. 16.24 Optimized formate complexes with the PDKA considering both tautomers (a and b)
and with the PDKA coordinated to Mg2+ from reference [112]. The interaction energies (ΔEBSSE )
are indicated. Distances in Å
that the presence of the Mg ion coordinated to the inhibitor enhances the anion–π
interaction. This long-range effect should be emphasized since the distance from the
Mg ion to the interacting anion is 8.3 Å (see Fig. 16.24c).
16.7 Conclusions
The purpose of this chapter has been to expose the bonding relationship between
anions and π systems by describing high quality work and demonstrating the extraor-
dinary potential of this relatively new interaction to impact the field of supramolecular
science, including catalysis and enzyme chemistry. It is now evident for the scientific
community that anion–π interactions are prominent in a wide range of systems and
should be considered as an important and general noncovalent binding force. The
potential of the anion–π interaction is not limited to the design of novel hosts and
sensors. Its important role in RNA recognition motifs, enzymatic chemistry, cata-
lysts, and crystal engineering has provided a new dimension to this interaction. The
effect of anion–π interactions should not be overlooked in chemical and biological
systems that involve anion and electron-deficient aromatic species.
Acknowledgment We are grateful to Carol Garau, Xavier Lucas, Daniel Escudero and David
Quiñonero with whom we have had the good fortune to work and their names are contained within
the pertinent references. We thank CONSOLIDER-Ingenio 2010 (project CSD2010-0065) and the
MICINN of Spain (project CTQ2011-27512 FEDER funds) for financial support. We thank the
Direcció General de Recerca, Desenvolupament Tecnològic i Innovació del Govern Balear (project
23/2011, FEDER funds) for financial support.
References
1. Schneider HJ (2009) Binding mechanisms in supramolecular complexes. Angew Chem Int

Ed 48:3924–3977
2. Schneider HJ, Yatsimirski A (2000) Principles and methods in supramolecular chemistry.
Wiley, Chichester
3. Lehn JM (1995) Supramolecular chemistry concepts and perspectives. Wiley-VCH, Weinheim
4. Vögtle F (1993) Supramolecular chemistry: an introduction. Wiley, New York
5. Beer PD, Gale PA, Smith DK (1999) Supramolecular chemistry. Oxford University Press,
Oxford
6. Steed JW, Atwood JL (2000) Supramolecular chemistry. Wiley, Chichester
7. Politzer P, Murray JS, Clark T (2010) Halogen bonding: an electrostatically driven highly
direccional noncovalent interaction. Phys Chem Chem Phys 12:7748–7757.
8. Bundhun A, Ramasami P, Murray JS, Politzer P (2013) Trends in σ-hole strenghts and
interactions of F3 MX molecules (M = C, Si, Ge and X = F, Cl, Br, I). J Mol Model
19:2739–2746
9. Bauzá A, Mooibroek TJ, Frontera A (2013) Tetrel–bonding: rediscovered supramolecular
force? Angew Chem Int Ed 52:12317–12321
and biological recognition. Angew Chem Int Ed 42:1210–1250
11. Schneider HJ, Yatsimirsky A (2008) Selectivity in supramolecular host-guest complexes.
Chem Soc Rev 37:263–277
recognition: energetics and structures. Angew Chem Int Ed 50:4808–4842
14. Burley SK, Petsko GA (1985) Aromatic-aromatic interaction: a mechanism of protein
structure stabilization. Science 229:23–28
15. Li S, Cooper VR, Thonhauser T, Lundqvist BI, Langreth DC (2009) Stacking interactions and
DNA intercalation. J Phys Chem B 113:11166–11172
16. Rutledge LR, Campbell-Verduyn LS, Wetmore SD (2007) Characterization of the stacking
interactions between DNA or RNA nucleobases and the aromatic amino acids. Chem Phys
Lett 444:67–175
17. Tsuzuki S, Yoshida M, Uchimaru T, Mikami M (2001) The origin of the cation/π interaction:
the significant importance of the induction in Li+ and Na+ complexes. J Phys Chem A
105:769–773
18. Ma JC, Dougherty DA (1997) The cation-π interaction. Chem Rev 97:1303–1324
19. Crowley PB, Golovin A (2005) Cation-π interactions in protein-protein interfaces. Proteins
Struct Funct Genet 59:231–239
20. Gallivan JP, Dougherty DA (2000) A computational study of cation-π interactions vs salt
bridges in aqueous media: implications for protein engineering. J Am Chem Soc 122:870–874
21. Lamoureux JS, Maynes JT, Glover JMN (2004) Recognition of 5 -YpG-3 sequences by
coupled stacking/hydrogen bonding interactions with amino acid residues. J Mol Biol
335:399–408
22. Frey J, Kraus T, Heitz V, Sauvage JP (2005) A catenane consisting of a large ring threaded
through both cyclic units of a handcuff-like compound. Chem Commun 49:5310–5312
23. Masu H, Sakai M, Kishikawa K, Yamamoto M, Yamaguchi K, Kohmoto S (2005) Aromatic
foldamers with iminodicarbonyl linkers: their structures and optical properties. J Org Chem
70 1423–1431
24. Bao X, Isaacsohn I, Drew AF, Smithrud DB (2007) Determining the binding and intracellular
transporting abilities of a host-[3]rotaxane. J Org Chem 72:3988–4000
25. Heemstra JM, Moore JS (2004) Helix stabilization through pyridinium-π interactions. Chem
Commun 40:1480–1481
26. Nishio M, Hirota M, Umezawa Y (1998) In the C–H/π interaction: evidence, nature,
consequences. Wiley-VCH, New York
27. Nishio M (2004) CH/π hydrogen bonds in organic and organometallic chemistry. Cryst Eng
Comm 6:130–156
28. Egli M, Sarkhel S (2007) Lone pair-aromatic interactions: to stabilize or not to stabilize. Acc
Chem Res 40:197–205
29. Mooibroek TJ Gamez P Reedijk J (2008) Lone pair–π interactions: a new supramolecular
bond? CrystEngComm 10:1501–1515
30. Mitra M, Manna P, Seth SK, DasA, Meredith J, Helliwell M, BauzáA, Choudhury SR, Frontera
A, Mukhopadhyay S (2013) Salt-bridge-π (sb-π) interactions at work: associative interac-
tions of sb-π, π-π and anion-π in Cu(II)-malonate-2-aminopyridine-hexafluoridophosphate
ternary system. CrystEngComm 15:686–696
31. Frontera A, Gamez P, Mascal M, Mooibroek TJ, Reedijk J (2011) Putting anion-π interactions
into perspective. Angew Chem Int Ed 50:9564–9583
32. Schottel BL, Chifotides HT, Dunbar KR (2008) Anion-π interactions. Chem Soc Rev 37:68–
83
33. Caltagirone C, Gale PA (2009) Anion receptor chemistry: highlights from 2007. Chem Soc
Rev 38:520–563
34. Gamez P, Mooibroek TJ, Teat SJ, Reedijk J (2007) Anion binding involving π-acidic
heteroaromatic rings. Acc Chem Res 40 435–444
35. Hay BP, Bryantsev VS (2008) Anion-arene adducts: C–H hydrogen bonding, anion-π
interaction, and carbon bonding motifs. Chem Commun 44:2417–2428
36. Quiñonero D, Garau C, Rotger C, Frontera A, Ballester P, Costa A, Deyà PM (2002) Anion-π
interactions: do they exist? Angew Chem Int Ed 41:3389–3392
37. Mascal M, Armstrong A, Bartberger MD (2002) Anion-aromatic bonding: a case for anion
recognition by π-acidic rings. J Am Chem Soc 124:6274–6276
38. Alkorta I, Rozas I, Elguero J (2002) Interaction of anions with perfluoro aromatic compounds.
J Am Chem Soc 124:8593–8598
39. Kim DY, Singh NJ, Kim KS (2008) Cyameluric acid as anion-π type receptor for ClO− 4 and
NO− 3 : π-Stacked and edge-to-face structures. J Chem Theory Comput 4:1401–1407
40. Schottel BL, Bacsa J, Dunbar KR (2005) Anion dependence of Ag(I) reactions with
3,6-bis(2-pyridyl)-1,2,4,5-tetrazine (bptz): isolation of the molecular propeller compound
[Ag2 (bptz)3 ][AsF6 ]2 . Chem Commun 41:46–47
41. Han B, Lu JJ, Kochi JK (2008) Anion recognitions via cocrystallizations with organic π-acids
in the efficient self-assembly of nanoscopic one-dimensional molecular chains. Cryst Growth
Des 8:1327–1334
42. Mascal M, Yakovlev I, Nikitin EB, Fettinger JC (2007) Fluoride-selective host based on
anion-π interactions, ion pairing, and hydrogen bonding: synthesis and fluoride-ion sandwich
complex. Angew Chem Int Ed 46:8782–8784
43. Götz RJ, Robertazzi A, Mutikainen I, Turpeinen U, Gamez P, Reedijk J (2008) Concurrent
anion-π interactions between a perchlorate ion and two π-acidic aromatic rings, namely
pentafluorophenol and 1,3,5-triazine. Chem Commun 44:3384–3386
44. Albrecht M, Müller M, Mergel O, Rissanen K, Valkonen A (2010) CH-directed anion-π
interactions in the crystals of pentafluorobenzyl-substituted ammonium and pyridinium salts.
Chem Eur J 16:5062–5069
45. Estarellas C, Frontera A, Quiñonero D, Deyà PM (2011) Relevant anion-π interactions in
biological systems: the case of urate oxidase. Angew Chem Int Ed 50:415–418
46. Chakravarty S, Sheng ZZ, Iverson B, Moore B (2012) “η6”-Type anion-π in biomolecular
recognition. FEBS Lett 586:4180–4185
47. Jenkins DD, Harris JB, Howell EF, Hinde RJ, Baudry J (2013) STAAR: Statistical Analysis
of Aromatic Rings. J Comput Chem 34:518–522
48. Dawson RE, Hennig A, Weimann DP, Emery D, Ravikumar V, Montenegro J, Takeuchi T,
Gabutti S, Mayor M, Mareda J, Schalley CA, Matile S (2010) Experimental evidence for the
functional relevance of anion-π interactions. Nat Chem 2:533–538
49. Sakai N, Mareda J, Vauthey E, Matile S (2010) Core-substituted naphthalenediimides. Chem

Commun 46:4225–4237
50. Gale PA, Caltagirone C (2014) Anion sensing by small molecules and molecular ensembles.
Chem Soc Rev. doi:10.1039/c4cs00179f
51. Caballero A, Zapata F, Beer PD (2013) Interlocked host molecules for anion recognition and
sensing. Coord Chem Rev 257:2434–2455
52. Gale PA (2011) Anion receptor chemistry. Chem Commun 47:82–86
53. Wenzel M, Hiscock JR, Gale PA (2012) Anion receptor chemistry: highlights from 2010.
Chem Soc Rev 41:480–520
54. Gale PA, Quesada R (2006) Anion coordination and anion-templated assembly: highlights
from 2002 to 2004. Coord Chem Rev 250:3219–3244
55. Bowman-James K, Bianchi A, García-Espaa E (2011) Anion coordination chemistry. Wiley-
VCH, Weinheim
56. Berg JM (1995) Zinc finger domains: from predictions to design Acc Chem Res 28:14–19
57. Ashcroft FM (2000) Ion channels and disease. Academic, San Diego
58. Moss B (1996) A land awash with nutrients-the problem of eutrophication. Chem Ind 11:407–
411
59. Asakura T, Kim SY, Morita Y, Ozawa M (2005) Reduction of pertechnetate in nitric extraction
of rare metals for future reprocessing. J Nucl Radiochem Sci 6:267–269
60. Benner S (2010) Hydrology: anthropogenic arsenic. Nat Geosci 3:5–6
61. Frontera A, Quiñonero D, Deyà PM (2011) Cation-π and anion-π interactions. WIREs
Comput Mol Sci 1:440–459
62. Giese M, Albrecht M, Krappitz T, Peters M, Gossen V, Raabe G, Valkonen A, Rissanen K
(2012) Cooperativity of H-bonding and anion-π interaction in the binding of anions with
neutral π-acceptors. Chem Commun 48:9983–9985
63. Watt MM, Zakharov LN, Haley MM, Johnson DW (2013) Selective nitrate binding in compet-
itive hydrogen bonding solvents: do anion-π interactions facilitate nitrate selectivity? Angew
Chem Int Ed 52:10275–10280
64. Gamez P, Mooibroek TJ, Teat SJ, Reedijk J (2007) Anion-π interactions involving π-acidic
rings. Acc Chem Res 40:435–444
65. Ballester P (2008) Recognition of anions. Springer Berlin
66. FronteraA (2013) Encapsulation of anions: macrocyclic receptors based on metal coordination
and anion-π interactions. Coord Chem Rev 257:1716–1727
67. Schneider HJ, Werner F, Blatter T (1993) Attractive interactions between negative charges
and polarizable aryl parts of host-guest systems. J Phys Org Chem 6:590–594
68. Schneider HJ (1991) Mechanisms of molecular recognition—investigations of organic host
guest complexes. Angew Chem Int Ed Engl 30:1417–1436
69. Hiraoka K, Mizuse S, Yamabe S (1987) High-symmetric structure of the gas-phase cluster
ions X− · · · C6 F6 (X = Cl, Br and I). J Phys Chem 91:5294–5297
70. Hiraoka K, Mizuse S, Yamabe S (1987) A determination of the stabilities and structures of
F-(C6 H6 ) and F-(C6 F6 ) clusters. J Chem Phys 86:4102–4105
71. Bauzá A, Quiñonero D, Deyà PM, Frontera A (2013) Anion-π interactions in [S4 N3 ]+ rings.
New J Chem 37:2636–2641
72. Galán-Mascarós JR, Slawin AMZ, Woollins JD, Williams DJ (1996) π-Facial interactions
between Cl− and [S4 N3 ]+ . Polyhedron 15:4603–4605
73. Kim D, Tarakeshwar P, Kim KS (2004) Theoretical investigations of anion-π interactions:
the role of anions and the nature of π systems. J Phys Chem A 108:1250–1258
74. Kim DY, Singh NJ, Lee JW, Kim KS (2008) Cyameluric acid as anion-π type receptor for
ClO− −
4 and NO3 : π-Stacked and edge-to-face structures. J Chem Theory Comput 4:1162–1169
75. Garau C, Frontera A, Quiñonero D, Ballester P, Costa A, Deyà PM (2003) A topological
analysis of the electron density in anion-π interactions. Chem Phys Chem 4:1344–1348
76. Quiñonero D, Garau C, Frontera A, Ballester P, Costa A, Deyà PM (2002) Counterintuitive
interaction of anions with benzene derivatives. Chem Phys Lett 359:486–492
77. Bauzá A, Deyà PM, Frontera A, Quiñonero D (2014) Substituent effects in cation-π inter-
actions revisited: a general approach based on intrinsic properties of the arenes. Phys Chem
Chem Phys 16:1322–1326
78. Wheeler SE, Houk KN (2010) Are anion/π interactions actually a case of simple
charge−dipole interactions? J Phys Chem A 114:8658–8664
79. Garau C, Quiñonero D, Frontera A, Ballester P, Costa A, Deyà PM (2003) Dual binding mode
of s-triazine to anions and cations. Org Lett 5:2227–2229
80. Garau C, Frontera A, Quiñonero D, Ballester P, Costa A, Deyà PM (2004) Cation-π ver-
sus anion-π interactions: energetic, charge transfer, and aromatic aspects. J Phys Chem A
108:9423–9427
81. Garau C, Quiñonero D, Frontera A, Ballester P, Costa A, Deyà PM (2003) Anion-π
interactions: must the aromatic ring be electron deficient? New J Chem 27:211–214
82. Quiñonero D, Frontera A, Garau C, Ballester P, Costa A, Deyà PM (2006) Interplay between
cation-π, anion-π and π-π interactions. Chem Phys Chem 7:2487–2491
83. Quiñonero D, Frontera A, Deyà PM (2008) High-level ab initio study of anion-π interactions
in pyridine and pyrazine rings coordinated to Ag(I). Chem Phys Chem 9:397–399
84. Gural’skiy IA, Escudero D, Frontera A, Solntsev PV, Rusanov EB, Chernega AN, Krautscheid
H, Domasevitch KV (2009) 1,2,4,5-Tetrazine: an unprecedented μ4 -coordination that
enhances ability for anion-π interactions. Dalton Trans 38:2856–2864
85. Bauzá A, Quiñonero D, Deyà PM, Frontera A (2012) Tuning of the anion-π interaction. Theor
Chem Acc 131:1219–1230
86. Mirzaei M, Eshtiagh-Hosseini H, Lotfian N, Salimi A, Bauzá A, Van Deun R, Decadt R,
Barceló-Oliver M., Frontera A (2014) Syntheses, structures, properties and DFT study of
hybrid inorganic-organic architectures constructed from trinuclear lanthanide frameworks
and Keggin-type polyoxometalates. Dalton Trans 43:1906–1916
87. Estarellas C, Frontera A, Quiñonero D, Deyà PM (2008) Theoretical and crystallographic
study of the dual σ/π binding affinity of quinolizilynium cation. J Chem Theory Comput
4:1981–1989
88. Quiñonero D, Frontera A, Escudero D, Ballester P, Costa A, Deyà PM (2007) Theoretical
study of anion-π interactions in seven membered rings. Chem Phys Chem 8:1182–1187
89. Das A, Choudhury SR, Dey B, Yalamanchili SK, Helliwell M, Gamez P, Mukhopadhyay S,
Estarellas C, Frontera A (2010) Supramolecular assembly of Mg(II) complexes directed by
asociative lone pair-π/ π-π/ π-anion-π/ π-lone pair interactions. J Phys Chem B 114:4998–
5009
90. Das A, Choudhury SR, Estarellas C, Dey B, Frontera A, Hemming J, Helliwell M, Gamez
P, Mukhopadhyay S (2011) Supramolecular assemblies involving anion-π and lone pair-π
interactions: experimental observation and theoretical analysis. Cryst Eng Comm 13:4519–
4527
91. Fiol JJ, Barceló-Oliver M, Tasada A, Frontera A, Terrón A, García-Raso A (2013) Structural
characterization, recognition patterns and theoretical calculations of longchain Nalkyl sub-
stituted purine and pyrimidine bases as ligands: on the importance of anion-π interactions.
Coord Chem Rev 257:2705–2715
92. Wang DX, Wang MX (2013) Anion–π interactions: generality, binding strength, and structure.
J Am Chem Soc 135:892–897
93. Schneebeli ST, Frasconi M, Liu Z, Wu Y, Gardner DM, Strutt NL, Cheng C, Carmieli R,
Wasielewski MR, Stoddart JR (2013) Electron sharing and anion-π recognition in molecular
triangular prisms. Angew Chem Int Ed 52:13100–13104
94. Chifotides HT, Giles ID, Dunbar KR (2013) Supramolecular architectures with π-acidic
3,6-bis(2-pyridyl)-1,2,4,5- tetrazine cavities: role of anion–π interactions in the remarkable
stability of Fe(II) metallacycles in solution. J Am Chem Soc 135:3039–3055.
95. Bretschneider A, Andrada DM, Dechert S, Meyer S, Mata RA, Meyer F (2013) Preorganized
anion traps for exploiting anion-π interactions: an experimental and computational study.
Chem Eur J 19:16988–17000
96. Ballester P (2013) Experimental quantification of anion-π interactions in solution using

neutral host-guest model systems. Acc Chem Res 46:874–884
97. Gil-Ramírez G, Escudero-Adán EC, Benet-Buchholz J, Ballester P (2008) Quantitative
evaluation of anion-π interactions in solution. Angew Chem Int Ed 47:4114–4118
98. Adriaenssens L, Gil-Ramírez G, Frontera A, Quiñonero D, Escudero-Adán EC, Ballester P
(2014) Thermodynamic characterization of halide-π interactions in solution using “two wall”
aryl extended calix[4] pyrroles as model system. J Am Chem Soc 136:3208–3218
99. Chang KC, Minami T, Koutník P, Savechenkov PY, Liu Y, Anzenbacher P (2014) Anion
binding modes in meso-substituted hexapyrrolic calix[4]pyrrole isomers. J Am Chem Soc
136:1520–1525
100. He Q, Han Y, Wang Y, Huang ZT, Wang DX (2014) Size-regulable vesicles based on anion-π
interactions. Chem Eur J 20:7486–7491
101. Zhao Y, Domoto Y, Orentas E, Beuchat C, Emery D, Mareda J, Sakai N, Matile S (2013)
Catalysis with anion-π interactions. Angew Chem Int Ed 52:9940–9943
102. Zhao Y, Beuchat C, Domoto Y, Gajewy J, Wilson A, Mareda J, Sakai N, Matile S (2014)
Anion-π catalysis. J Am Chem Soc 136:2101–2111
103. Zhao Y, Sakai N, Matile S (2014) Enolate chemistry with anion-π interactions. Nature Comm
5:3911
104. Oubridge C, Ito N, Evans PR, Teo CH, Nagai K (1994) Crystal-structure at 1.92 Angstrom
resolution of the RNA-binding domain of the U1a spliceosomal protein complexed with an
RNA hairpin. Nature 372:432–438
105. Estarellas C, Frontera A, Quiñonero D, Deyà PM (2011) Anion-π interactions in Flavopro-
teins. Chem Asian J 6:2316–2318
106. Dong C, Flecks S, Unversucht S, Haupt C, van Pee KH, Naismith JH (2005) The structure
of tryptophan 7-halogenase (PrnA) suggests a mechanism for regioselective chlorination.
Science 309:2216–2219
107. Race PR, Lovering AL, Green RM, Ossor A, White SA, Searle PF, Wrighton CJ, Hyde
EI (2005) Structural and mechanistic studies of Escherichia Coli nitroreductase with the
antibiotic nitrofurazone. Reserved binding orientations in different redox states of the enzyme
J Biol Chem 280:13256–13264
108. BauzáA, Quiñonero D, Deyà PM, FronteraA (2013) On the importance of anion-π interactions
in the mechanism of sulfide: quinone oxidoreductase. Chem Asian J 8:2708–2713
109. Cherney M, ZhangY, Solomonson M, Weiner JH, James MNG (2010) Crystal structure of sul-
fide:quinone oxidoreductase from Acidithiobacillus Ferrooxidans: insights into sulfidotrophic
respiration and detoxification. J Mol Biol 398:292–305
110. Pham VH, Yong JJ, Park SJ, Yoon DN, Chung WH, Rhee SK (2008) Molecular analysis of the
diversity of the sulfide: quinone reductase (sqr) gene in sediment environments. Microbiol
154:3112–3121
111. Johnson E, Keinan S, Mori-Sánchez P, Contreras-García J, Cohen A,Yang W (2010) Revealing
non-covalent interactions. J Am Chem Soc 132:6498–6506
112. Bauzá A, Quiñonero D, Deyà PM, Frontera A (2014) Long-range effects in anion-π interac-
tions: their crucial role in the inhibition mechanism of Mycobacterium Tuberculosis malate
synthase. Chem Eur J 20:6985–6990
113. McKinney JD, Höner zu Bentrup K, Muoz-Elías EJ, Miczak A, Chen B, Chan WT, Swenson
D, Sacchettini JC, Jacobs WR Jr, Russell DG (2000) Persistence of Mycobacterium Tubercu-
losis in macrophages and mice requires the glyoxylate shunt enzyme isocitrate lyase. Nature
406:735–738
114. Krieger IV, Freundlich JS, Gawandi VB, Roberts JP, Gawandi VB, Sun Q, Owen JL, Fraile
MT, Huss SI, Lavandera JL, Ioerger TR, Sacchettini JC (2012) Structure-guided discovery
of phenyl-diketo acids as potent inhibitors of M. Tuberculosis malate synthase. Chem Biol
19:1556–1567
Chapter 17
A Survey of DNA–Protein π–Interactions: A
Comparison of Natural Occurrences
and Structures, and Computationally Predicted
Structures and Strengths
Katie A. Wilson and Stacey D. Wetmore
Abstract Non-covalent interactions between DNA and proteins play critical roles
in cellular functions, including DNA replication and repair. To gain an appreciation
of the biomolecular components involved, several bioinformatics studies have data
mined experimental X-ray crystal structures to identify close contacts between DNA
and protein building blocks. These critical studies have revealed that DNA–protein
non-covalent interactions include π–π, C–H · · · π, O–H · · · π, N–H · · · π or lone
pair–π (X · · · π, X = O, N or S) contacts. Unfortunately, however, experimental
structural data cannot provide information about the relative strength of biologically-
relevant non-covalent interactions. Therefore, quantum mechanical calculations have
been used to determine the stability of DNA–protein π–heterodimers, as well as the
dependence of the interaction energy on changes in relative monomer orientations. In
this light, the present review summarizes work done in the literature to characterize
π–interactions between the DNA nucleobases (A, C, T and G) or deoxyribose moiety
and cyclic (His, Phe, Tyr and Trp) or acyclic (Arg, Glu and Asp) amino acid side
chains. Collectively, this body of work emphasizes the importance of DNA–protein
π–interactions for providing stability to biomolecular complexes and driving key
cellular functions.
17.1 Introduction
Non-covalent interactions are important in a wide variety of applications. For exam-

ple, Tokay geckos use non-covalent interactions to walk up walls [1–3]. As a result,
the Defense Advanced Research Project Agency in the United States of America
has used the Tokay geckos as a model to develop paddles that allow a person to
climb an eight-meter glass wall [4, 5]. Additionally, non-covalent interactions play
microscopic roles in many biological processes, including those that involve in-
teractions between DNA and proteins such as DNA repair and DNA replication.
S. D. Wetmore () · K. A. Wilson

Department of Chemistry and Biochemistry, University of Lethbridge,
4401 University Drive West, T1K 3M4 Lethbridge, Alberta, Canada
e-mail: stacey.wetmore@uleth.ca

502 K. A. Wilson and S. D. Wetmore
These non-covalent DNA–protein interactions can occur between two π–containing

systems (denoted as π–π interactions) or between one π–system and another (conju-
gated or non-conjugated) molecular component (including C–H · · · π, O–H · · · π,
N–H · · · π or lone pair–π (X · · · π, X = O, N or S) interactions).
DNA repair provides a textbook example of the wide function of DNA–protein
π–π interactions. For example, the DNA glycosylases, which are the first enzymes
involved in the base excision repair (BER) of non-bulky nucleobase damage [6], are
known to use π–π interactions between DNA nucleotide substrates and the aromatic
(or cyclic) amino acids (Phe, His, Trp and Tyr; Fig. 17.1). The best example of the
contribution of π–π interactions to BER is provided by alkyladenine DNA glyco-
sylase (AAG). Indeed, several π–interactions afforded by the AAG active site allow
this enzyme to excise a wide variety of substrates (including both cationic and neutral
nucleobases). Specifically, upon substrate binding, AAG inserts Tyr162 into the helix

3 with respect to the site of the damaged nucleobase, and this residue forms π–π
interactions that stabilize the DNA helix when the damaged base is flipped into the
enzyme active site (Fig. 17.2a) [7]. In addition, simultaneous interactions between
Tyr127, His136 and Tyr159 and the damaged nucleobase bind the substrate in the
AAG active site (Fig. 17.2b) [8]. Notably, binding of labile cationic substrates in the
AAG active site occurs due to stronger πcation –π interactions than the π–π contacts
formed with the neutral lesions. As a result, the π–π interactions between AAG and
the substrate have been shown to be catalytic for the excision of neutral lesions, but
anti-catalytic or non-catalytic for the removal of cationic lesions [8]. Although AAG
solely relies on π–π interactions to facilitate DNA repair, other DNA repair enzymes
(for example, uracil DNA glycosylase [9] and alpha-ketoglutarate-dependent dioxy-
genase [10]) use π–π interactions in conjunction with other types of non-covalent
contacts to bind and excise damaged DNA. Furthermore, other enzymes are known
to implement DNA-protein π–π interactions with π–containing (cyclic or acyclic)
amino acids (Fig. 17.1) for catalysis (for example, aldehyde dehydrogenase [11],
antibody interactions [12], tropinone reductase [13] and ketosteroid isomerase [14]).
DNA replication also exploits DNA–protein π–interactions. However, unlike the
π–π contacts between two aromatic rings used by DNA repair enzymes, all classes
of DNA polymerases invoke interactions between an amino acid π–system and de-
oxyribose. Specifically, C–H · · · π interactions between the sugar of the incoming
deoxynucleoside triphosphate (dNTP) and a conserved Glu (A-family) or Phe/Tyr
(B−, X−, Y− and RT-families) residue occurs during DNA replication (Fig. 17.2c)
[15–17]. In fact, these so-called “steric gate” interactions are responsible for the se-
lective incorporation of dNTP (DNA) over RNA nucleoside triphosphate (NTP) by
a 1,000,000 [18] to 100 [19] ratio. Nevertheless, this conserved interaction between
the deoxyribose and a π–containing amino acid was only recently acknowledged
in the literature to be more than simply a steric interaction [20, 21]. This example
illustrates that there are still many aspect of non-covalent DNA–protein contacts that
remain to be explored.
DNA–protein π–π and sugar–π interactions are ideal for binding between
biomolecules and catalysis of biological processes due to their significant strength,
yet ability to readily degrade in order to continuously afford function to the protein.
17 A Survey of DNA–Protein π–Interactions 503
Fig. 17.1 Models of

π–containing amino acids,
the natural DNA nucleobases
and DNA deoxyribose moiety
(sugar) for which
non-covalent interactions are
discussed in the present
review
Fig. 17.2 AAG active site depicting the a interaction between Y162 and T8, and b π–π interac-
tions with the (ethenoadenine) substrate, as well as c a representative sugar–π interaction between
deoxyribose and Tyr12 in a polymerase active site
Experimental methods can provide information about the abundance and structure
(NMR and X-ray crystallography), as well as catalytic contribution (mutagenesis
and kinetic studies), of DNA–protein contacts. Unfortunately, however, it is difficult
to gain information about the relative strength of discrete DNA–protein interac-
tions from experimental data. In contrast, computational chemistry is an ideal tool
for studying these non-covalent interactions since this approach can provide atomic
level details about the interactions, including the strength of intermolecular forces
acting between the monomers and the effects of geometrical variables on the inter-
action energy. As a result, quantum mechanical calculations are commonly used to
determine the binding strength (E) of DNA–protein π–interactions according to
Eq. 17.1
E = E dimer − E aa − E nt (17.1)
In this equation, E dimer stands for the electronic energy of the interaction, while E aa
and E nt stand for the electronic energy of the amino acid and deoxyribose or nucle-
obase moiety of the nucleotide, respectively. Indeed, although experimental data tells
us about the structures of π–π interactions in nature, quantum mechanical calcula-
tions are the only method that can be used to determine their energetic contribution
to DNA–protein binding, stability and function.
There are many computational methods that can be used to gain information
about DNA–protein π–interactions. Unlike hydrogen bonding, which can be accu-
rately described by many computational methodologies due to the large electrostatic
contribution, the interaction energy of π–contacts has a large dispersion compo-
nent and therefore requires methods that recover a large portion of the total electron
correlation. As a result, early studies on π–interactions implemented high-level quan-
tum mechanical techniques, which require an abundance of computer resources.
The most widely used and successful ab initio approach was MP2/6-31G*(0.25),
which replaces the standard d-exponent (0.8) with a value of 0.25 to obtain accurate
interaction energies in comparison to gold-standard CCSD(T)/CBS results at a re-
duced computational cost. Nevertheless, the accuracy of MP2/6-31G*(0.25) arises
due to a cancelation of errors [22, 23]. More recently, advances in density func-
tional theory (DFT) have produced methods that account for dispersion interactions
[24], but are much less computationally expensive, which allows for analysis of
π–interactions in biomolecules. The most commonly used DFT functionals for cal-
culating the strength of DNA–protein π–interactions include M06-2X, B3LYP-D3,
B97-B3 and ωB97-D, which have all been proven to provide energies that are com-
parable to CCSD(T)/CBS [20, 25–29]. In addition to quantum mechanical methods
that allow the structure and energy of π–contacts to be characterized, the electron
density obtained from Quantum Theory of Atoms in Molecules (QTAIM) can provide
complementary information about the nature of non-covalent interactions [30].
The goal of this review is to summarize recent work on biologically important
DNA protein π–interactions. Specifically, the occurrence, composition, structure and
strength of π–interactions between π–containing amino acids (Trp, Phe, His, Tyr,
Arg, Glu and Asp) and the natural DNA nucleobases (A, T, G and C) or deoxyribose
sugar as determined according to experimental crystal structures of DNA–protein
complexes will be examined. Additionally, the effects of geometrical parameters
that dictate relative monomer orientations and the optimal dimer arrangements that
have been determined using quantum mechanical techniques will be reviewed for
the most well studied interactions to date, including contacts between the aromatic
amino acids (Trp, Phe, His, and Tyr) and the DNA bases, and the acyclic amino acids
(Arg, Glu and Asp) and cytosine. These 7 distinct π–containing amino acids also
Fig. 17.3 Definitions of key geometrical parameters that define the relative monomer orientations
in DNA–protein π–π interactions
reveal the effects of charge (anionic, cationic and neutral), cyclic versus acyclic π–
system, and π–system size on the frequency, structure and strength of DNA–protein
π–interactions in nature, as well as on the structure and strength of the optimum
orientation of DNA–protein π–interactions. The results from purely computational
studies will be critically compared to interactions found in nature. Finally, the effects
of the biological environment, including biological backbones, solvent and additional
non-covalent interactions, will be presented, and future perspectives for the field will
be summarized.
17.2 Interactions Between Aromatic Amino Acids and

Nucleobases
As discussed in the Introduction, the presence of π–π interactions between the aro-
matic amino acids (Trp, Tyr, Phe and His) and the DNA nucleobases has been
accepted to be important in critical bioprocesses, including DNA repair. Further-
more, the function of interactions between other non-aromatic π–containing amino
acids (Arg, Glu, and Asp) and the DNA nucleobases have also been more recently
acknowledged. In general, regardless of the nature of the amino acid involved, the
structural differences between nucleobase–amino acid π–π interactions are defined
on the basis of four variables, namely R1, α, R2, ω and θ (Fig. 17.3). R1 represents
the vertical separation between the center of mass of the amino acid and the center of
mass of the nucleobase, α is the twist of the amino acid with respect to the nucleobase
(defined according to the center of mass), R2 symbolizes the horizontal displace-
ment between the center of masses of the two π–systems, and ω is the angle between
the planes of the amino acid and nucleobase rings. Notably, ω defines whether the
monomer relative orientation corresponds to a stacking (when the amino acid and
nucleobase are parallel or ω ≈ 0◦ ) or T-shaped (when the amino acid and nucleobase
are perpendicular or ω ≈ 90◦ ) binding arrangement. A final variable is also impor-
tant for defining T-shaped interactions, namely θ or the edge directed toward the
π–system. Specifically, either the edge of the amino acid can interact with the face
(π–system) of the nucleobase or the edge of the nucleobase can interact with the
face of the amino acid. The upcoming discussion will first highlight the occurrence,
structure and strength of π–interactions identified in experimental crystal structures
of DNA–protein complexes, and then describe how the calculated strength of these
important π–π interactions depends on the R1, α, R2, ω and θ variables.
17.2.1 π–Interactions in Experimental Crystal Structures
The existence of specific non-covalent interactions in biology can be verified by

searching experimental X-ray crystal structures and identifying a large number of
contacts prior to undertaking in-depth computational investigations of specific inter-
actions. In order to appreciate the persistence of DNA–protein π–π interactions in
nature, several groups have searched the protein data bank (PDB) to determine the
frequency of different types of non-covalent DNA–protein contacts, including π–π
interactions between amino acids and DNA nucleobases. The first studies of this kind
were restricted by the absence of accurate X-ray crystal structures due to limitations
in the available experimental technologies at the time [31–34]. With recent advances
in experimental methods and equipment, a large number of high-resolution crystal
structures of DNA–protein complexes became available for analysis [35–38]. How-
ever, due to differences in search criteria used and relatively undefined geometric
definitions of van der Waals interactions, these studies yielded starkly different con-
clusions. For example, one paper reports that van der Waals interactions are more
common than hydrogen bonding [35], while another study determined that hydrogen
bonding is more common than van der Waals interactions [37]. Nevertheless, these
early works agree that van der Waals interactions are common and compose more
than 30 % of all non-covalent contacts identified.
Aside from broadly determining the relative frequency of different types of
non-covalent interactions, other studies have conducted PDB searches to specif-
ically investigate π–π interactions between the aromatic amino acids and DNA
nucleobases. The first such study considered π–π interactions between adenosine

5 -triphosphate (ATP) and the aromatic amino acids, and found that, although hydro-
gen bonding occurs 2.7 times more often than π–π interactions, π–π interactions
are vital for substrate binding [39]. A more recent investigation considered the π–π
interactions between the aromatic amino acids (Trp, Tyr, Phe and His) and all four
natural nucleobases (A, T, G and C) in 141 DNA–protein complexes [40]. Unfor-
tunately, these two studies lead to different overall trends in terms of which amino
acid preferentially interacts with A. Although this discrepancy may be an artefact of
searching different experimental structures, distinctive (distance only) search crite-
ria were also used in each study. Furthermore, many other parameters besides the
separation distance contribute to the characteristics of π–π interactions. In fact, due
to the distance only search criteria, not all previously identified non-covalent inter-
actions correspond to DNA–protein π–π interactions. Indeed, closer examination
reveals some of the monomer separations that fall within the distance cut-off rep-
resent hydrogen-bonding or sugar–π interactions [20, 21]. Therefore, despite the
usefulness of previous work for verifying the presence of close contacts between
DNA and protein π–residues, none of these studies truly determined the frequency
or clarified the structure of these π–π interactions.
Although database tools have been more recently created to identify possible
DNA–protein π–interactions in experimental crystal structures [41–47] quantum
chemical calculations have also been used to gain information about the strength of
DNA–protein π–π interactions in nature. For example, Mao et al. used data mining
and MP2/6-311+G* calculations to investigate aromatic amino acid π–π contacts
with ATP in 68 crystal structures of adenylate-binding proteins [39]. The correspond-
ing interaction energies were determined to be substantial, and the work thereby
established the molecular basis for adenine recognition in proteins. Subsequently,
the Rooman group used MP2 to considered 89 interactions between His and Phe,
Tyr, Trp or A observed in X-ray structures of proteins and protein-ligand complexes
in several environments (gas phase, water, acetone, THF and CCl4 ) [48]. This analy-
sis concluded that the strength and frequent occurrence of His interactions in active
sites has implications for their importance in biological processes. In 2007, Baker and
Grant investigated 141 DNA–protein complexes for interactions between the DNA
nucleobases and Tyr, Phe, His and Trp [40]. Subsequent MP2/6-31G calculations
helped verify the importance of Phe interactions in transcription. The Tschumper
group further considered A:Phe binding by building 26 9-methyladenine–toluene
dimers from protein/ligand crystal structures and determining the binding strengths
with highly correlated methods (CCSD(T) and RI-MP2) [49]. Uniquely, MP2 opti-
mizations of the nucleobase–amino acid complexes were also implemented, which
lead to six distinctive dimers. In a follow-up study by the same group, 20 model
A:Tyr systems were fully optimized to 11 unique dimers, which were found to have
binding strengths nearly equivalent to the A:Phe systems [50]. These representa-
tive examples of combined bioinformatics and quantum chemical calculations were
critical contributions to the literature.
Despite the large number and strength of interactions between the DNA nucle-
obases and the aromatic amino acids in nature, other amino acid side chains also
have π–systems and form π–contacts with nucleobases. Indeed, small distances
between the non-aromatic π–side chains of arginine, aspartic/glutamic acid or aspar-
tate/glutamate and the DNA nucleobases have been identified in experimental X-ray
crystal structures [51–56]. More specifically, the structure of stair motifs, which are
formed upon binding of a charged amino acid side chain to two neighbouring nucle-
obases, have been assessed by critical analysis of 52 X-ray crystal structures, and
most complexes were determined to involve G and Arg [55]. This study was followed
by MP2/6-31G*(0.2) calculations on a wide variety of stair motifs [57, 58], which
verified their stability. Alternatively, while attempting to quantify the role of cation–π
interactions at the interfaces of DNA–protein complexes, Wintjens et al. identified a
large number of Arg contacts, comprising more than half of the cation–π interactions
found [54]. Furthermore, the most frequent nucleobase–amino acid pairing was de-
termine to occur between G and Arg, which was also the most stable pair according to
MP2 calculations [51, 52]. In contrast, a search of cation–π interactions in proteins
bound to ligands revealed a number of contacts between DNA nucleobases and pos-
itively charged side chains, with the most frequent pairing occurring between A and
Arg [53]. Similarly, a study of 55 cation/amino–π motifs found in X-ray structures
that involve A moieties and charged (Arg, Lys, Asn, or Gln) side chains reveals that
the Arg:A pair is the most favourable according to MP2 binding strengths and the
most commonly observed [56]. Finally, PDB data mining and molecular dynamics
simulations verify that anionic interactions, including those with Asp and Glu side
chains, mediate protein–nucleic acid interactions, albeit through an abundance of

hydrogen–bonding interactions [59]. It should be noted that the Arg, Glu and Asp
series is also of interest from a fundamental perspective since the variety of charges
these side chains adopt depending on the biological environment permits evaluation
of differences in the nature of π–π, πcation –π and πanion –π interactions.
Overall, although previous studies have been useful for verifying the existence and
determining the importance of π–interactions, differences in the crystal structures
considered, types of contacts investigated, search methods implemented and quantum
chemical calculations used for analysis makes it difficult to compare between the
studies. Therefore, in the discussion below, we will primarily summarize the most
recent work from our group that systematically considered high-resolution X-ray
crystal structures of DNA–protein complexes published before January 2014 in the
PDB [20, 21]. A unique aspect of our work is the inclusion of a visual inspection
step to definitively characterize the identified contacts between π–containing amino
acids (Phe, Trp, Tyr, His, Glu, Asp or Arg) and the natural DNA nucleobases (A, T,
G or C).
17.2.1.1 Occurrence of π–Interactions in Experimental Crystal Structures
952 DNA–protein π–π interactions were found in the 672 crystal structures searched
by our group. It has been determined that 50 % of these 672 structures contain a π–π
interaction involving Trp, Tyr, His or Phe, while 25 % of the searched PDB structures
contained a π–π interaction involving Arg, Glu or Asp. Among the 331 structures
with a π–π interaction involving Trp, Tyr, His or Phe, 61 % contain only one contact.
Similarly, of the 166 structures with a π–π interaction involving Arg, Glu or Asp,
73 % only contain one contact. These DNA–protein interactions occur in a wide
variety of protein complexes, supporting proposals that these non-covalent contacts
are important for many biological processes. However, the percentage distribution
in the types of proteins the interactions were found in is equal to that for the types of
proteins searched; therefore, π–π interactions are unlikely to preferentially occur in
a particular protein class.
17.2.1.2 Composition of π–Interactions in Experimental Crystal Structures
Among the interactions identified in nature, 75 % involve a cyclic aromatic amino

acid (Fig. 17.4a). When the amino acids involved in the contacts are considered,
the majority of interactions occur with Phe (33 %) followed by Tyr (22 %) and Arg
(21 %), while all other amino acids investigated (Trp, His, Glu and Asp) make up less
than 15 % of the contacts found. When the interacting nucleobase is considered, the
majority of contacts occur with T (30 %), while A, C, and G have a similar number
of interactions (23 %, Fig. 17.4b). However, when the composition with respect to
the nucleobase is divided according to whether the interaction occurs with a cyclic
or acyclic amino acid, T forms the most contacts with Phe, Tyr, Trp or His, but the
Fig. 17.4 Proportions of a amino acids and b nucleobases observed in nucleobase-amino acid π–π
interactions found in X-ray crystal structures of DNA–protein complexes
least with Arg, Glu or Asp, while G has the largest number of contacts with Arg, Glu
or Asp, but the least with Phe, Tyr, Trp or His. Notably, A and C have similar trends
with respect to the cyclic and acyclic amino acids, each forming about 20–30 % of
the interactions.
17.2.1.3 Structure of π–Interactions in Experimental Crystal Structures
The interactions found in nature adopt many different relative orientations of the two
residues involved. When the closest heavy atom distance between the monomers
is considered, Arg adopts the largest distances among the amino acids (up to 4.8
Å), while Asp forms the closest interactions, with separation distances < 2.8 Å.
Overall, the majority of contacts have separation distances between 3.0 and 4.0 Å,
with 45 to 90 % of the interactions for each amino acid ranging between 3.0 and 3.4
Å (Fig. 17.5). Interestingly Glu has the smallest proportion of structures between
3.0 and 3.4 Å (45 %), and the largest number of structures with a separation distance
> 4.0 Å (22 %). Notably, Arg, His and Glu are the only amino acids that never adopt
a separation distance greater than 4.0 Å. Conversely, Glu and Phe never adopt a
separation distance less than 3.0 Å. Regardless, there is no particular trend for the
separation distance with respect to the nucleobase. The range of distances adopted in
the crystal structures is indicative of possible interaction strengths, where interactions
occurring closer than the van der Waals radius of the atoms involved will be repulsive,
while interactions that are too far away will not be stabilizing.
The DNA–protein π–π interactions found in nature adopt a variety of tilt angles,
which correspond to stacked (ω = 0–20◦ ), inclined (20–70◦ ) or T-shaped (ω = 70–
90◦ ) orientations. In terms of the DNA nucleobases, all preferentially adopt the
stacked orientation. Additionally, Phe, Tyr, Trp, His and Arg are more likely to
appear in a stacked orientation (47–89 %) than an inclined or T-shaped orientation
with respect to a DNA nucleobase (less than 31 %, Fig. 17.6). Conversely, potentially
anionic amino acids (Glu and Arg), which form relatively few interactions, never
Fig. 17.5 Occupancies of different distances (Å) observed in nucleobase–amino acid π–π
interactions found in X-ray crystal structures of DNA–protein complexes
adopt a tilt angle less than 50◦ , and therefore are most commonly found in a T-
shaped orientation (56–70 %). These features thereby highlight the distinct nature of
πanion –π and πcation –π interactions [40, 60].
Each inclined and T-shaped contact was visually inspected to identify the in-
teracting edge (atoms interacting with the π–system). For Trp, His, Phe and Tyr
side chains, amino acid–edge dimers [170] are more common than nucleobase–
edge dimers [84]. Among the amino acid–edge dimers, 56 % involve a bridged
(two H atoms directed towards the π–ring) edge of Phe (Fig. 17.7a). Conversely,
the nucleobase edge–dimers are more structurally diverse, with the most prevalent
cytosine–edge occurring when two CH groups (C5 and C6) are directed towards the
π–ring (18 %, Fig. 17.7b). Conversely there are less amino acid–edge dimers (31)
than nucleobase–edge dimers (114) for the acyclic amino acids. The most common
amino acid–edge for the acyclic amino acids involves two NH groups (N and Nη)
of Arg (26 %, Fig. 17.7c), which are the most acidic hydrogens. The most common
nucleobase–edge interactions with the acyclic amino acids involve the G edge that
contains CH (C8) and N (N7) groups (16 %, Fig. 17.7d). Notably, there is larger
diversity in the nucleobase edges employed in DNA–protein T-shaped interactions,
but a preference for particular edges in the amino acid edge dimers, partly due to the
greater chemical diversity of nucleobase edges available for π–interactions.
17.2.1.4 Strength of π–Interactions in Experimental Crystal Structures
High-level quantum mechanical calculations (M06-2X/6-31G+(d,p)//MP2/6-

31G(d)) show that DNA–protein π–π interactions adopt a wide variety of interaction
Fig. 17.6 Occupancies of different binding orientations of nucleobase–amino acid π–π interactions
found in X-ray crystal structures of DNA–protein complexes
Fig. 17.7 Most common

T-shaped edges for a Phe
(3I0X; F179:A9) b C (3LDY;
Tyr35:C5) c Arg (3PX6;
Arg472:C203), and d G
(3U6Q; Arg264:G9) found in
experimental X-ray crystal
structures
energies and can provide stability to DNA–protein complexes. Specifically, contacts

with Trp, (neutral) His, Phe and Tyr most commonly have binding strengths between
− 20 and − 25 kJ mol−1 (25 %), but can have interaction energies up to −40 kJ/mol
(Fig. 17.8). While the neutral acyclic interactions are most commonly associated
with interaction energies ranging between −10 and −15 kJ mol−1 (26 %), these con-
tacts can also be as strong as the cyclic variants (−40 kJ mol−1 ). Furthermore, the
charge of the amino acid has a significant effect on the strength of the interaction.
Cationic interactions involving His or Arg most commonly range between −15 and
−30 kJ mol−1 (32 %), but can be up to −50 kJ mol−1 for His and −95 kJ mol−1 for
Arg. This is in agreement with previous literature that has noted the large strength
of cationic π–interactions [61–65]. DNA–protein anionic interactions are generally
even stronger, with the most frequent interaction energy falling from −30 to −40 kJ
mol−1 (21 %). Nevertheless, the anionic contacts led to a similar maximum binding
strength as the cationic interactions (−95 kJ mol−1 ). The significant strength of an-
ionic π–interactions compared to neutral contacts has been discussed in the literature
[64, 66]. Notably, many charged interactions calculated according to monomer ori-
entations appearing in nature are repulsive, which may simply reflect the fact that the
corresponding residues are neutral in the protein environment. In fact, the charged
His/Asp/Glu interactions that are repulsive have an interaction energy ranging from
−10 to −20 kJ mol−1 when neutral. To appreciate the importance of these DNA–
protein π–π interactions, we note that the A:T hydrogen bond, which is accepted to
be an important interaction in nature for the stabilization of DNA duplexes, has been
calculated to be approximately −70 kJ/mol at the gold-standard CCSD(T)/CBS level
of theory [67]. Thus, neutral DNA–protein π–π interactions approach the strength
of the A:T hydrogen-bonded pair, while even more impressively, the charged π–π
interactions can be more stable than the A:T dimer.
17.2.2 Effects of Geometrical Variables
To complement information gained from experimental X-ray crystal structures, com-

putational approaches have been used to study a number of isolated DNA–protein
π–π complexes. For example, in terms of the aromatic amino acids, the Tschumper
group used MP2 to fully optimize complexes of A stacked with Phe and Tyr ini-
tiated from crystal structure orientations (discussed in the Introduction) [49, 50].
Alternatively, the Rooman group compared and contrasted the stacked and T-shaped
binding orientations by scanning the A monomer of fixed geometry with respect to
Phe or His [48]. In addition to A, pyrimidine interactions with the aromatic amino
acids have been considered. For example, Cysewski investigated stacked complexes
between C or U and His, Phe, Tyr or Trp, and revealed that considerable degrees
of geometric freedom can lead to similar stacking energies [68]. The U:Phe stacked
dimer has also been critically analyzed using MP2, M05-2X and M06-2X coupled
with 6-31G** and full optimizations, as well as AIM, by Ebrahimi and co-workers
[69]. In terms of nucleobase contacts with the acyclic π–containing amino acids,
relatively little computational work has been dedicated to characterizing the dimer
potential energy surfaces, with Cysewski having considered the dimers between C
or U and Arg [68]. Aside from looking at interactions with the natural bases, contacts
Fig. 17.8 Binding strengths for π–π interactions involving the a neutral (cyclic) aromatic, b neutral
acyclic, and c charged amino acids in experimental X-ray crystal structure geometries
between non-canonical bases and the aromatic amino acids have also been studied
using computational chemistry. For example, interactions involving damaged nucle-
obases have been investigated, such as inosine paired with Phe [70] and methylated
bases with all four aromatic amino acids [71–73].
To date, our group has completed the most thorough computational studies of DNA
nucleobase interactions with the π–containing amino acids. Specifically, MP2/6-
31G*(0.25) or M06-2X/6-31+G(d, p) potential energy surface scans were used to
determine how each geometrical variable (R1, α, and R2) affects the binding energy
of the nucleobase–amino acid π–π interactions for both perfectly stacked binding
arrangements (ω = 0◦ ) and T-shaped orientations (ω = 90◦ ) at various θ values [28,
63, 74, 75]. In these scans, each variable was changed in set intervals to map the
interaction potential energy surfaces using over 1000 relative monomer orientations
for each Phe, Trp, Tyr, or His heterodimer with each canonical (DNA and RNA)
nucleobase, as well as between various protonation states of Arg, Glu, or Asp and C.
This approach identifies the minimum energy structures, and corresponding binding
strengths, as well as records the effects of different geometrical variables on the
interaction energies. As a result, the large number of data points gained from these
scans can be used to estimate the strength of interactions in experimental structures of
DNA–protein complexes by adding the deviations in variables to the global minimum
values [75].
17.2.2.1 Preferred Vertical Separation
The first geometric variable considered when mapping the potential energy surface
for the interaction energies between DNA and protein components is the vertical
separation (R1). For all amino acid and nucleobase combinations investigated, the
magnitude of the interaction energy does not change significantly with small spatial
deviations (0.1 to 0.2 Å) from the optimal R1 regardless of whether a stacked or
T-shaped orientation is considered [28, 63, 74, 75]. However, the binding energy
decreases at short R1 distances due to additional repulsion and at large R1 separations
due decreased attractive dispersion forces. Interestingly, stacked dimers typically
have a shorter prefer separation distance, but a larger dependence on R1, than T-
shaped dimers.
The preferred R1 can also be divided based on the identity of the interacting
amino acid, as well as the interaction (stacking versus T-shaped) classification. In
contrast, the identity of the nucleobase does not have a consistent effect on the
preferred vertical separation. However, amino acids with a smaller π–system (Arg,
His, Asp/Glu) prefer shorter vertical separations (3.2–3.4 Å) than amino acids with
larger, cyclic π–systems (Phe, Tyr or Trp; 3.4–3.6 Å). Interestingly, cationic His
consistently leads to shorter vertical separation distances than neutral His [63], while
there is no specific trend in the effect of a negative charge on Asp/Glu on the preferred
vertical separation distance [28].
17.2.2.2 Preferred Angle of Rotation
The angle of rotation (α) influences the interaction energy to a greater extent than
R1 and R2 for all nucleobase–amino acid dimers expect those involving cationic
His [63]. For stacking interactions, changing the angle of rotation increases the
electrostatic interactions in the system by anti-aligning the monomer dipole moment
vectors and/or decreases the steric repulsion by removing close contacts between
the monomers [28, 63, 74, 75]. Indeed, the dependence of the binding strength on
the angle of rotation increases as A < T < G < C for the nucleobases and as Phe <
Tyr < Trp < His for the neutral cyclic amino acids, which correlates with the dipole
moments of the monomers. Furthermore, the dependence on α is largest for the
anionic systems, with the dependence being up to 20, 15 and 32 kJ/mol for neutral,
cationic and anionic stacked systems, respectively.
For T-shaped DNA–protein dimers, the dependence of the interaction energy on α
is related to the strength of secondary intermolecular hydrogen bonds, the alignment
of acid/base sites of the monomers and/or steric repulsions between monomers.
Specifically, when secondary hydrogen bonding is weaker (stronger) than the dipole–
dipole stacking interaction, then the dependence on α is less (more) than for the
stacked arrangement between the same amino acid and nucleobase. For the neutral
cyclic amino acids, the dependence on α ranges between 1 and 33 kJ/mol regardless
of whether an amino acid edge or nucleobase edge is involved. Conversely, the
interaction energy changes more with α (by up to 13 kJ/mol) for cationic His T-shaped
than stacking contacts, while cationic Arg dimers exhibit the same α dependence
regardless of the binding orientation. Notably, due to greater symmetry, rotation
about α does not have as large of an effect on Arg compared to His. As for the stacking
interactions, the dependence on α is generally stronger for anionic C:Asp/Glu dimers
(up to 32 kJ/mol).
17.2.2.3 Preferred Horizontal Displacement
For all neutral systems, the original horizontal displacement aligns the center of
masses of the monomers in both the T-shaped and stacked orientations. Upon cal-
culating the preferred R2, only small changes relative to the initial position occur in
most dimers. Furthermore, for the neutral dimers, the R2 shift leads to very small
changes in the interaction energy (generally less than 2 kJ/mol). In cases where
the R2 shift affects the interaction energy, the corresponding change in the relative
monomer alignment generally reduces steric repulsion, increases the π–π overlap
or aligns the acidic/basic regions in the monomers to maximize the electrostatic
contribution, with the latter being especially true for T-shaped dimers. For charged
interactions, the horizontal displacement tends to increase the interaction energy
to a greater extent (by up to approximately 15 kJ/mol). Furthermore, for charged
dimers, the monomers shift to align the electropositive and electronegative regions
of the residues. Most importantly, R2 can affect the binding energy of heterodimers
involving cationic systems more than any other geometrical variable.
Fig. 17.9 Binding

arrangements corresponding
to the strongest DNA–protein
T-shaped edge interactions as
determined through quantum
mechanical calculations
17.2.2.4 Preferred T-Shaped Edge
The structure and interaction energy of DNA–protein T-shaped dimers exhibit a large
dependence on the interacting monomer edge. However, some overall trends in the
data are evident. For example, monomers with large dipole moments, and therefore a
large variation in the acid/base properties of the edges, show the greatest dependence
on θ. Furthermore, the strongest neutral and anionic T-shaped interactions involve the
monomer edge with the largest acidity. However, for the cationic systems, the most
stable T-shaped interactions occur with the monomer edge with the greatest basicity.
Due to greater overlap between π–systems, stronger interactions generally occur
when the monomer edge is a bridge, which involves a bond or two atoms directed
toward the π–ring, rather than single atom interactions. An exception to this rule
occurs for neutral Asp/Glu heterodimers, which have the smallest π–systems of the
amino acids considered. Notably, edges that give rise to lone pair–π interactions with
a carbonyl oxygen atom are unstable, while lone pair–π interactions with nitrogen
atoms are stable by up to −23 kJ/mol. For nucleobase–edge dimers, the glycosidic
bond is generally the preferred edge, but this is not biologically relevant for most
proteins that process DNA strands rather than nucleobase components. Therefore,
the amino group of A (N6) or C (N4), the NH groups at N1 and N2 for G, and the
NH group at N3 for T are the most preferred edges (Fig. 17.9). For amino acid–edge
dimers, the preferred edges contains a CH (Cζ) and N or NH group (N or Nδ)
(neutral) His, NH (N) or NH (N) and CH (Cζ) for (cationic) His, two CH groups
for Phe, and OH and CH group for Tyr and a NH and CH group for Trp, a OH group
for neutral Asp/Glu and the edge of the carboxylic group for anionic Asp/Glu and
two NH groups (N and Nη) for Arg (Fig. 17.9).
17.2.2.5 Energetics of Preferred Orientation
Although the above discussion provides information about the dependence of the
binding strength on the relative monomer orientation, an important remaining ques-
tion is: what is the maximum interaction energy for a particular nucleobase–amino
acid pairing? The data obtained from the potential energy surface scans gives informa-
tion about the maximum stability a particular pairing can provide to a DNA–protein
complex. The interaction strengths for the neutral cyclic amino acids range from −20
to −43 kJ/mol when in the stacked orientation, with two thirds of the contacts possess-
ing interaction energies between −25 and −35 kJ/mol. Furthermore, the T-shaped
interactions for the neutral cyclic amino acids range from −12 to −50 kJ/mol, with the
T-shaped interactions involving a nucleobase edge being stronger than interactions
involving an amino acid edge. Furthermore, the most favourable nucleobase–edge
interactions are just as strong as or stronger than the stacking interactions between
the same nucleobase–amino acid combination.
When the contacts are divided according to the interacting nucleobase, both stack-
ing and T-shaped contacts are similar in magnitude. As a function of the amino acid,
the strongest stacking and T-shaped interactions increase as Phe < < Tyr ∼ His < Trp
due to the relative magnitude of the dipole moments and π–systems, as well as the
strength of secondary intermolecular contacts between the amino acid or nucleobase
edge in the T-shaped interactions. When cationic His is considered, the stacked dimers
have binding strengths ranging from −35 to −69 kJ/mol, which is an increase of
130–223 % compared to the neutral His dimers. When cationic His adopts a T-shaped
orientation, the interaction energies range from −20 kJ/mol to −105 kJ/mol, which
is an increase of 186–332 % from neutral His dimers and larger than the increase for
stacked dimers. Notably, the increase in the interaction energy upon protonation of
His is due to increased electrostatics and is the greatest for His edge dimers.
The energetics for the acyclic amino acid π–systems with C are similar in magni-
tude to those for the cyclic systems with the same charge. In particular, the stacking
interactions are up to −26 kJ/mol, −57 kJ/mol and −48 kJ/mol for the neutral,
cationic and anionic acyclic systems, respectively. The interaction energies for these
contacts are within 4 kJ/mol of those for contacts between C and His regardless of
(neutral or cationic) charge. For the T-shaped dimers, the interactions are up to −40
kJ/mol, −91 kJ/mol and −99 kJ/mol for the neutral, cationic and anionic acyclic
systems, respectively. Therefore, the T-shaped interactions are also similar in mag-
nitude to the neutral and cationic His dimers (−33 and −92 kJ/mol, respectively). As
for the cyclic protein components, T-shaped interactions involving the acyclic amino
acids are stronger than the stacked arrangements, and the cytosine–edge dimers are
stronger than amino acid–edge counterparts (by up to 40 kJ/mol). Despite differences
between neutral and charged binding strengths, both systems adopt distinct geome-
tries in order to maximize the interaction energy. Specifically, when the preferred
geometry of the neutral Asp/Glu:C dimer (corresponding binding strength of −25.5
kJ/mol) is used to calculated the interaction energy with anionic Asp/Glu, the binding
strength decreases to only −9 kJ/mol. Similarly, when the preferred geometry for the
anionic Asp/Glu:C dimer (corresponding binding strength of −47.9 kJ/mol) is used
to calculated the binding strength for the corresponding neutral Asp/Glu interaction,
the stability decreases to only −5 kJ/mol.
Overall, analysis of DNA–protein interactions involving 7 amino acids of varying
size and charge reveals that charge is more important than the size of the π–system
for determining the stability of the contact. Additionally, these π–π interactions
are strong in the optimal binding orientations, approaching the stability provided
to DNA by the A:T base pair (−70 kJ mol−1 calculated with CCSD(T)/CBS) [67].
Therefore, when coupled with the abundance of such contacts in DNA–protein com-
plexes (discussed above), we can conclude that many types of nucleobase–amino acid
π–interactions must play an important role in biology. However, since similar inter-
action strengths are found for many monomer combinations and geometries, these
interactions likely most often contribute to binding without selectivity. Nevertheless,
examples do exist in the literature of the potential selectivity of DNA–protein π–π in-
teractions [12]. Regardless, since very stable interactions occur in many non-optimal
binding orientations, there is great flexibility in the contributions these interactions
can make to DNA–protein binding.
17.2.2.6 Effect of the Biological Environment
For all calculations discussed thus far, the amino acid and nucleobase systems were
treated with truncated models such that only the π–system of interest was considered
(without the DNA or protein backbone). However, it is possible that the biological
backbone(s) affects the structure or magnitude of the π–interaction through, for
example, polarizing the π–system. Therefore, the validity of this approximation
has been considered for all π–π (neutral) stacked nucleobase–aromatic amino acid
combinations [76, 77], while T-shaped interactions were investigated for the His:A
dimer [78], with either the DNA or protein backbone. As anticipated, the addition
of the protein or DNA backbone affects the preferred geometry of the dimer due
to steric and/or electrostatic interactions between the π–system and the backbone.
Although R1 changes very little when the preferred α for the extended system is
considered, the preferred angle of rotation changes by more than 30◦ in many dimers
upon inclusion of the DNA backbone (for example, 38 of the 105 stacked dimers
investigated exhibit larger deviations in α). Nevertheless, the α increment used in the
previous potential energy surface scans was 30◦ , and the protein backbone negligibly
affects the preferred α. Finally, although there is not a consistent trend in the effect
of the biological backbone on the horizontal displacement (R2), the binding energy
of few structures increases more than 5 kJ/mol upon accounting for the R2 shift,
which generally decreases the steric repulsion or increases the π–overlap between
the monomers, as well as allows secondary interactions between the backbone and
the π–system.
The inclusion of the biological backbone can change the strength of the interaction
between a nucleobase and amino acid. Specifically, the protein backbone strengthens
the interactions (by up to approximately 10 kJ/mol), while the DNA backbone can
strengthen or weaken the interactions (by up to approximately 15 kJ/mol). Neverthe-

less, the binding energy of some dimer pairs decreases upon including the backbone
due to changes in the relative ring orientations that move the dimer away from its
optimal geometry. However, changes in the interaction energy are not due to a mod-
ification in the π–π contact, but rather additional attractive backbone–π contacts
with the aromatic system, the presence of which was verified with QTAIM [77, 78].
These additional backbone–π interactions can increase the total DNA–protein bind-
ing energy by up to 18 kJ/mol. Furthermore, the backbone and π–π interactions are
additive [77], and therefore the energy of the π–π interactions can be accurately
calculated without the inclusion of the protein or DNA backbone. This key finding
justifies the small models used to estimate the stability of DNA–protein interactions
in nature.
In addition to the majority of previous computational studies on DNA–protein π–
π interactions employing truncated models, most calculations have been performed
in the gas phase and the results are therefore relevant to active sites of low polarity.
However, due to the presence of water in biological systems (i.e., filling cavities
in solvent exposed active sites, mediating interactions between biomolecules, as a
catalytic residue or as a product of an enzymatic reaction), solvent effects on the
magnitude of the calculated interactions energies must be considered. In this light,
select works have studied the effect of the polarity of the surrounding environment
using implicit solvation models [39, 48, 73, 74, 79]. Unfortunately, the accuracy
of this approach for monomers separated by large distances is unclear. Therefore,
to the best of our knowledge, one study in the literature to date has considered mi-
crosolvation effects by examining discrete interactions between water and (neutral
or cationic) His when in a stacked or (nucleobase or amino acid–edge) T-shaped
orientation with respect to A [80]. In this contribution, up to four water molecules
were added to His:A dimers in binding arrangements that represent (N–H · · · O,
N · · · H–O or C–H · · · O) hydrogen bonding, O–H · · · π T-shaped and O · · · π
lone pair–π interactions. The water contacts affect the (M06-2X/6-31+G(d, p)) cal-
culated strength of the neutral His:A dimer by less than 3 kJ/mol (15 %) regardless of
the number of water molecules, water binding orientation or nucleobase–amino acid
relative monomer orientations considered. This supports the use of gas–phase calcu-
lations to determine the magnitude of DNA–protein π–π interactions and suggests
that the significant strength of such π–π binding arrangements likely applies to many
biological environments (i.e., low polarity active sites and environments with solvent
or even side chain interactions with the nucleobases). However, the presence of water
drastically decreases the stability of cationic His:A dimers (by up to 15 kJ/mol or
30 %) for all (water and nucleobase–amino acid) binding arrangements considered.
This decrease arises from a reduction in the positive charge on the amino acid upon
interacting with water, which leads to an interaction energy between that of the fully
protonated and charge neutral His dimers. Nevertheless, even in the presence of
explicit water, the binding strength of the cationic His:A dimers are significant in
both stacked and T-shaped arrangements and therefore may be important for many
cellular functions.
Although studying dimers improves our fundamental understanding of DNA–

protein interactions, the effects of additional contacts on the preferred geometry and
binding strength must be determined. For example, the AAG active site discussed in
the Introduction (Fig. 17.2b) provides a prime example of DNA–protein interactions
in nature, which seldom appear as an isolated dimer. Indeed, there are three simulta-
neous nucleobase–amino acid π–π interactions in the AAG active site upon binding
ethenoadenine [7]. As a result, six test trimers composed of adenine or (cationic) 3-
methyladenine sandwiched between two (neutral and/or cationic) His residues were
considered using MP2/6-31G(0.25) and QTAIM analysis [81]. Potential energy sur-
face scans analogous to those performed for dimers determined that the presence
of an additional nucleobase–amino acid π–π contact does not affect the preferred
monomer orientation in the first interaction, which is true regardless of the charge of
the monomers involved. Furthermore, the two nucleobase–amino acid interactions
in the trimer, as well as the amino acid–amino acid 1–3 term, sum to yield the overall
trimer stabilization energy. Thus, the interaction energies reported for dimer contacts
that appear in nature, and for isolated nucleobase–amino acid dimers, likely extend
to more complicated biological environments.
17.2.3 Comparisons Between Natural Interactions and Preferred

Geometric Parameters
Comparisons can be made between the extensive studies of the structure and strength
of DNA–protein π–π interaction in the natural state [20, 21] and the optimal cal-
culated parameters [28, 63, 74, 75]. Of the variables considered in studies of the
optimal DNA–protein interactions, similarities and differences between the optimal
and the natural structure can be discussed for ω, R1 and θ. However, since the center
of mass was not added to interactions found in experimental crystal structures, slight
differences arise in the way in which these variables are measured between the two
classes of studies. Additionally, R2 and α were not recorded for the interactions
identified in experimental crystal structures. Starting with ω, studies of the optimal
interactions only considered two ω values, namely 0 (stacking) and 90◦ (T-shaped),
while interactions in nature were found to adopt all ω values between 0 and 90◦ .
Although only two orientations were investigated using scan calculations, the in-
teractions in nature that deviate significantly from ω = 0 or 90◦ have a significant
binding strength. Therefore, future work should consider the effect of rotating ω on
the interaction energy. Notably, the optimal interactions were found to be stronger
for the T-shaped interactions; however, although this trend is maintained in nature,
the stacking interactions are overall much more prevalent than T-shaped interactions.
In terms of R1, the optimal separation distance was found to be 3.2–3.6 Å and
natural interactions commonly adopt distances within this range (most commonly
being from 3–4 Å). However, these numbers are difficult to compare since the method
used to measure the distances was different for the natural and optimal interactions
(center of mass for the optimal distances versus closest heavy atom contact for the
natural interactions). Additionally, although amino acids with smaller π–systems

adopt closer optimal separation distances according to the calculations, this trend is
not observed for the natural interactions.
Finally, when the edge interacting in the T-shaped interactions is considered, it
was determined to be more favourable to have a bridged edge (a bond or two atoms)
interacting with the π–ring rather than a single proton edge, which corresponds to
the greater number of bridged edges found in nature. However, the edges that lead to
the strongest interaction energy according to the calculations are not the edges that
occur most frequently in nature. Nevertheless, it is unreasonable to expect this type
of agreement since the studies of the optimal interactions only considered perfectly
T-shaped interactions, while the edge interacting in nature is highly dependant on
the geometric constraints imposed by the protein. Nevertheless, DNA–protein π–π
interactions in nature still have significant strength and can contribute to DNA–
protein binding and other biological functions.
17.3 Interactions Between Aromatic Amino Acids and

Deoxyribose
As mentioned in the Introduction, in addition to the π–π interactions between the

DNA nucleobases and π–containing amino acids, interactions can occur between a
biological π–ring and a non-conjugated molecular component, which leads to C–
H · · · π, O–H · · · π, N–H · · · π or lone pair–π (X · · · π, X = O, N or S)
interactions. Indeed, examples of such non-covalent interactions between the aro-
matic amino acids and carbohydrates have been identified in experimental crystal
structures and the strengths of these contacts have been verified using quantum me-
chanical methods [82–85]. These so-called carbohydrate–π interactions have also
been shown to contribute to many bioprocesses including bacterial cell wall recogni-
tion, fertilization, immune response, cell–cell communication and post-translational
modifications [86–89]. Perhaps the most detailed computer modeling has been per-
formed on interactions between various pyranoses and Phe, Tyr or His, which has
shown that the amino acids interact favourably with either carbohydrate face to form
interactions that are stable by up to −50 kJ/mol [90–96].
Analogous to the carbohydrate–π interactions believed to be important in glycobi-
ology, contacts between π–containing amino acids and DNA deoxyribose have been
implicated in the mechanism of action of DNA polymerases (discussed in the In-
troduction) [15–19]. Indeed, early searches of experimental X-ray crystal structures
available in the PDB revealed close contacts between the π–containing amino acids
and the DNA deoxyribose moiety [35, 37]. Although some of these interactions likely
correspond to hydrogen bonding, including with the phosphate backbone, many rel-
ative DNA–protein orientations correspond to π–contacts (sugar–π). Similar to the
carbohydrate–π interactions, these DNA–protein sugar–π contacts can likely involve
many different sides of the sugar moiety and lead to several classifications of non-
covalent interactions, including C–H · · · π and lone pair–π contacts. In the coming
Fig. 17.10 Frequency of

particular amino acids
involved in sugar–π
interactions with the DNA
deoxyribose
sections, the frequency, structure and stability of DNA–protein sugar–π interactions

in nature will be presented, and compared to the observed nucleobase–amino acid
π–π interactions. Since the only work on sugar–π interactions in DNA–protein com-
plexes in the literature to date has been performed by our group, the discussion will
be solely based on our studies [20, 21].
17.3.1 Natural Frequency of Sugar–π Interactions
A search of 672 DNA–protein complexes available in the PDB identified 813 sugar–
π interactions between deoxyribose and the cyclic or acyclic amino acids. 31 %
of the structures searched contain a contact with the acyclic amino acids and 38 %
with the aromatic amino acids. As discussed for the nucleobase–amino acid π–π
interactions, the sugar–π contacts occur in a wide variety of proteins. Just over half
of the structures with a single sugar–π contact contained only one such contact (57
or 53 % for cyclic or acyclic amino acids, respectively). Notably, sugar–π contacts
are as common as nucleobase–amino acid π–π contacts, comprising 46 % of all
DNA–protein π–interactions. Among the structures searched, 80 % contain either a
sugar–π or π–π interaction. These statistics suggest that sugar–π and nucleobase
π–π interactions are both important components of DNA–protein binding.
17.3.2 Composition of Sugar–π Interactions
Sugar–π interactions occur with all π–containing amino acids in nature and are
almost as frequent with the cyclic (54 %) and acyclic (46 %) protein residues
(Fig. 17.10). This contrasts the nucleobase–amino acid π–π contacts, where 75 %
of interactions occur with the cyclic amino acids. Most sugar–π interactions invoke
Arg (34 %). In fact, Arg contacts comprise 73 % of all acyclic sugar–π interactions.
Nevertheless, more sugar–π interactions occur with Asp [78] than nucleobase π–π
interactions [10]. As found for the nucleobase π–π interactions, most cyclic sugar–π
interactions occur with Tyr (28 %) and Phe (17 %).
Fig. 17.11 a Numbering of the deoxyribose moiety and examples of different types of sugar–π
interactions, namely b lone pair–proton, c lone pair, d face, e bridge and f single proton contacts
17.3.3 Structure of Sugar–π Interactions
Careful consideration of the large number of sugar–π interactions identified in nature

reveals that these contacts can be classified according to the sugar atoms interacting
with the amino acid π–system. Specifically, these interactions can be categorized
as a single proton, two protons (bridged), three protons (face), a lone pair or a
lone pair and a proton (lone pair–proton) contact (Fig. 17.11). Among these clas-
sifications, the bridge and face orientations are the most common (30 % each), the
single–proton interactions are nearly as common (25 %), and the lone pair and the
lone pair–proton interactions collectively account for the remaining 15 % of inter-
actions (Fig. 17.12). Overall, each interaction adopts a range of relative monomer
orientations (Fig. 17.13). However, the most frequent interactions occur with the C5
edge of deoxyribose, with the most common sugar edge in each category including
the H4 –H5a –H5b face (20 %), the H5a –H5b bridge (13 %) and the H5a single–proton
(9 %) interactions.
The amino acid interacting with the deoxyribose affects the preferred conforma-
tion of the sugar–π interaction. Due to the larger π–system, more face interactions
Fig. 17.12 Proportions of different categories of DNA–protein sugar–π interactions with the a
(cyclic) aromatic and b acyclic amino acids
Fig. 17.13 Overlays of the most common sugar–π interactions involving the cyclic amino acids
and a a lone–pair, b H5a , c H1a H2b or d H4 H5a H5b sugar edge, and involving the acyclic amino
acids and a e lone pair, f H5a , g H5a H5b or h H4 H5a H5b sugar edge
occur with the cyclic (38 %) rather than acyclic (20 %) amino acids. Conversely,
more lone pair–π interactions occur with the acyclic (33 %) than cyclic (19 %) amino
acids, possibly due to stronger interactions with charged sites (Fig. 17.12). When
considering the individual amino acids, 59 % of Trp interactions occur with the H4 –
H5a –H5b face (Fig. 17.14a) and Asp commonly interacts with the H5a –H5b bridge
(62 %, Fig. 17.14b), which is distinct from the wide variety of interactions formed
with the other amino acids (Phe, Tyr, His, Arg and Glu).
Fig. 17.14 Examples of the

most common sugar–π
interactions found in X-ray
crystal structures of
DNA–protein complexes,
namely the a H4 H5a H5b
(4FPV; W306:T1) and b
H5a H5b (4ECS; Asp155:A9)
interactions
17.3.4 Strength of Sugar–π Interactions
Due to the range of deoxyribose–amino acid orientations adopted in nature, the

(M06-2X/6-31G(d,p)//MP2/6-31G(d)) calculated binding energies vary significantly
(Fig. 17.15). The interactions with the neutral cyclic amino acids are on average
stronger (up to −37.4 kJ mol−1 ) than interactions with the neutral acyclic amino acids
(up to −23.1 kJ mol−1 ). Additionally, the cyclic amino acids adopt a greater range of
energies, with 36 % of interactions involving the cyclic amino acids falling between
−10 and −15 kJ mol−1 and 60 % of interactions involving the acyclic amino acids
falling between −5 and −10 kJ mol−1 . The charged (cationic or anionic) interactions
are significantly stronger (up to −75 kJ mol−1 ) than the neutral sugar–π contacts. The
anionic interactions can be even stronger (up to −90 kJ mol−1 ), with most interactions
falling between −50 and −55 kJ mol−1 (19 %). Although no one particular sugar
edge leads to the strongest interactions for the neutral contacts, certain edges yield the
strongest interactions with charged amino acids. Specifically, the most stable contacts
involving the cationic amino acids occur with lone pair–π binding arrangements,
while the strongest interactions involving the anionic amino acids occur for the (H4 –
H5a –H5b ) face and (H5a –H5b ) bridge orientations. When compared to the nucleobase–
amino acid π–π interactions, the neutral deoxyribose–amino acid sugar–π contacts
are slightly weaker than the neutral nucleobase–amino acid π–π, while the charged
deoxyribose–amino acid sugar–π contacts are stronger than the charged nucleobase–
amino acid π–π interactions. Therefore, sugar–π interactions may contribute to
DNA–protein binding to a similar extent as nucleobase–amino acid π–π interactions.
17.4 Conclusions and Future Perspectives
In summary, the computational studies completed on DNA–protein π–interactions

between the amino acids and nucleobase or deoxyribose moieties to date have pro-
vided important insights into the occurrence, composition, structure, and strength
of these interactions. Data mining and computational studies based on experimen-
tal X-ray crystal structures and purely theoretical scans of important geometrical
Fig. 17.15 Binding strengths for sugar–π interactions with the a neutral aromatic, b neutral acyclic,
and c charged amino acids in experimental X-ray crystal structure geometries
parameters that dictate π–interactions between DNA and protein components have
revealed both the interaction energies and their dependence on the relative monomer
orientations. In fact, the generated database of binding energies obtained from com-
putational scans allow the strength of interactions found in new X-ray structures
to be estimated [75]. Most importantly, these collective studies have found that
both nucleobase–amino acid π–π and deoxyribose–amino acid sugar–π interactions
are common in nature and have significant strength. Therefore, both categories of
DNA–protein interactions will significantly contribute to DNA–protein binding and
function.
Unfortunately, the effects of various geometrical parameters on the stability of
π–π interactions have yet to be studied for all nucleobase–amino acid pairings.
Specifically, the acyclic amino acids have currently only been studied with C, and
therefore future work must consider pairings with G, T and A. Furthermore, al-
though the effects of some variables (R1, α and R2) on the interaction energies
have been studied, the effects of systematic variations in ω have yet to be inves-
tigated. Additionally, the interaction potential energy surfaces of the DNA–protein
sugar–π interactions should be mapped. Indeed, previous literature has investigated
the effects of R1 and R2 on the strength of the analogous carbohydrate–π contacts
[83, 97–100], and thereby revealed important information about the effects of each
geometrical variable on the binding strength. In addition to expanding our investi-
gations of known DNA–protein π–interactions, experimental structures should be
further searched to identify other classes of non-covalent π–interactions. For exam-
ple, analogous to DNA–protein π–π and sugar–π interactions, contacts between the
RNA nucleobases or ribose and the π–containing amino acids should be investigated.
Alternatively, interactions between the DNA or RNA nucleobases and other amino
acid side chains (for example, Ser and Cys) should be considered to give additional
information about the importance of C–H · · · π, O–H · · · π, N–H · · · π and/or
lone pair–π (X · · · π, X = O, N or S) interactions. Regardless of the abundance
of future work required to better understand DNA–protein π–interactions, the infor-
mation gained from computational studies thus far provides substantial insight into
these interactions.
References
1. Murthy PS (2006) Molecular handshake: recognition through weak noncovalent interactions.

J Chem Educ 83:1010
2. Autumn K, Liang YA, Hsieh ST, Zesch W, Chan WP, Kenny TW, Fearing R, Full RJ (2000)
Adhesive force of a single gecko foot-hair. Nature 405:681–685
3. Autumn K, Sitti M, Liang YA, Peattie AM, Hansen WR, Sponberg S, Kenny TW, Fearing R,
Israelachvili JN, Full RJ (2002) Evidence for van der Waals adhesion in gecko setae. Proc
4. Lee H, Lee BP, Messersmith PB. (2007) A reversible wet/dry adhesive inspired by mussels
and geckos. Nature 448:338–341
5. Yu J, Chary S, Das S, Tamelier J, Pesika NS, Turner KL, Israelachvili JN (2011) Gecko-inspired
dry adhesive for robotic applications. Adv Funct Mater 21:3010–3018
6. Schermerhorn KM, Delaney S (2014) A chemical and kinetic perspective on base excision
repair of DNA. Acc Chem Res 47:1238–1246
7. Lau AY, Wyatt MD, Glassner BJ, Samson LD, Ellenberger T (2000) Molecular basis for
discriminating between normal and damaged bases by the human alkyladenine glycosylase,
AAG. Proc Natl Acad Sci U S A 97:13573–13578
8. Rutledge LR, Wetmore SD (2011) Modeling the chemical step utilized by human alkyladenine
DNA glycosylase: a concerted mechanism AIDS in selectively excising damaged purines. J
Am Chem Soc 133:16258–16269
9. Przybylski JL, Wetmore SD (2011) A QM/QM Investigation of the hUNG2 reaction surface:
the untold tale of a catalytic residue. Biochemistry 50:4218–4227
10. Sedgwick B, Bates PA., Paik J, Jacobs SC, Lindahl T (2007) Repair of alkylated DNA: recent
advances. DNA Repair 6:429–442
11. Ahvazi B, Coulombe R, Delarge M, Vedadi M, Zhang L, Meighen E, Vrielink A (2000) Crystal
structure of the NADP+-dependent aldehyde dehydrogenase from Vibrio harveyi: structural
implications for cofactor specificity and affinity. Biochem J 349:853–861
12. An Y, Raju RK., Lu T, Wheeler SE (2014) Aromatic interactions modulate the 5 -base
selectivity of the DNA-binding autoantibody ED-10. J Phys Chem B 118:5653–5659
13. Nakajima K, Yamashita A, Akama H, Nakatsu T, Kato H, Hashimoto T, Oda Ji, Yamada Y
(1998) Crystal structures of two tropinone reductases: different reaction stereospecificities in
the same protein fold. Proc Natl Acad Sci U S A 95:4876–4881
14. Schwans JP, Sunden F, Lassila JK, Gonzalez A, Tsai Y, Herschlag D (2013) Use of anion-
aromatic interactions to position the general base in the ketosteroid isomerase active site. Proc
15. Joyce CM (1997) Choosing the right sugar: how polymerases select a nucleotide substrate.
Proc Natl Acad Sci U S A 94:1619–1622
16. DeLucia AM, Grindley NDF, Joyce CM (2003) An error-prone family Y DNA polymerase
(DinB homolog from Sulfolobus solfataricus) uses a ‘steric gate’ residue for discrimination
against ribonucleotides. Nucleic Acids Res 31:4129–4137
17. Astatke M, Ng K, Grindley ND, Joyce CM (1998) A single side chain prevents Escherichia
coli DNA polymerase I (Klenow fragment) from incorporating ribonucleotides. Proc Natl
Acad Sci U S A 95:3402–3407
18. Wang W, Wu EY., Hellinga HW, Beese LS (2012) Structural factors that determine selectivity
of a high fidelity DNA polymerase for Deoxy-, Dideoxy-, and ribonucleotides. J Biol Chem
287:28215–28226
19. Brown JA, Suo Z (2011) Unlocking the sugar “Steric Gate” of DNA polymerases.
Biochemistry 50:1135–1142
20. Wilson KA, Kellie JL, Wetmore SD (2014) DNA-protein π-interactions in nature: abundance,
structure, composition and strength of contacts between aromatic amino acids and DNA
nucleobases or deoxyribose sugar. Nucleic Acids Res 42:6726–6741
21. Wilson KA, Wells RA, Abendong MN, Anderson CB, Kung RW, Wetmore SD (2015) Land-
scape of π–π and sugar–π Contacts in DNA-protein Interactions. J. Biomol Struct Dyn. (DOI:
10.1080/07391102.2015.1013157)
22. Cerny J, Hobza P (2007) Non-covalent interactions in biomacromolecules. Phys Chem Chem
Phys 9:5291–5303
23. Sponer J, Jurecka P, Marchan I, Luque FJ, Orozco M, Hobza P (2006) Nature of base stacking:
reference quantum-chemical stacking energies in ten unique B-DNA base-pair steps. Chem-
Eur J 12:2854–2865
24. Kryachko ES, Ludea EV (2014) Density functional theory: foundations reviewed. Phys
Reports 544:123–239
25. Raju RK, Ramraj A, Hillier IH, Vincent MA, Burton NA (2009) Carbohydrate-aromatic [small
pi] interactions: a test of density functionals and the DFT-D method. Phys Chem Chem Phys
11:3411–3416
26. Hujo W, Grimme S (2012) Performance of non-local and atom-pairwise dispersion corrections
to DFT for structural parameters of molecules with noncovalent interactions. J Chem Theory
Comput 9:308–315
27. Burns LA, Vazquez-Mayagoitia A, Sumpter BG, Sherrill CD (2011) Density-functional
approaches to noncovalent interactions: a comparison of dispersion corrections (DFT-D),
exchange-hole dipole moment (XDM) theory, and specialized functionals. J Chem Phys 134
28. Wells RA, Kellie JL, Wetmore SD (2013) Significant strength of charged DNA-Protein π-π
interactions: a preliminary study of cytosine. J Phys Chem B 117:10462–10474
29. Rutledge LR, Wetmore SD (2010) The assessment of density functionals for DNA-protein
stacked and T-shaped complexes. Can J Chem-Rev Can De Chim 88:815–830
91:893–928
31. Matthews BW (1988) No code for recognition. Nature 335:294–295
32. Pabo CO, Sauer RT (1992) Transcription factors: structural families and principles of DNA
recognition. Ann Rev Biochem 61:1053–1095
33. Suzuki M (1994) A framework for the DNA-protein recognition code of the probe helix in
transcription factors: the chemical and stereochemical rules. Structure 2:317–326
34. Mandel-Gutfreund Y, Schueler O, Margalit H (1995) Comprehensive analysis of hydrogen
bonds in regulatory protein DNA-complexes: in search of common principles. J Mol Biol
253:370–382
35. Luscombe NM, Laskowski RA, Thornton JM (2001) Amino acid-base interactions: a three-
dimensional analysis of protein-DNA interactions at an atomic level. Nucleic Acids Res
29:2860–2874
36. Luscombe NM, Thornton JM (2002) Protein-DNA interactions: amino acid conservation and
the effects of mutations on binding specificity. J Mol Biol 320:991–1009
37. Lejeune D, Delsaux N, Charloteaux B, Thomas A, Brasseur R (2005) Protein-nucleic acid
recognition: statistical analysis of atomic interactions and influence of DNA structure.
Proteins: Struct, Funct, Bioinform 61:258–271
38. Sathyapriya R, Vijayabaskar M, Vishveshwara S (2008) Insights into protein-DNA interac-
tions through structure network analysis. PLoS Comput Biol 4:e1000170
39. Mao L, Wang Y, Liu Y, Hu X (2004) Molecular determinants for ATP-binding in proteins: a
data mining and quantum chemical analysis. J Mol Biol 336:787–807
40. Baker CM, Grant GH (2007) Role of aromatic amino acids in protein-nucleic acid recognition.
Biopolymers 85:456–470
41. Kirsanov DD, Zanegina ON, Aksianov EA, Spirin SA, Karyagina AS, Alexeevski AV (2013)
NPIDB: nucleic acid-protein interaction database. Nucleic Acids Res 41:D517–D523
42. Norambuena T, Melo F (2010) The protein-DNA interface database. BMC Bioinform 11:262
43. Kumar MDS, Bava KA, Gromiha MM, Prabakaran P, Kitajima K, Uedaira H, Sarai A
(2006) ProTherm and ProNIT: thermodynamic databases for proteins and protein-nucleic
acid interactions. Nucleic Acids Res 34:D204–D206
44. Contreras-Moreira B (2010) 3D-footprint: a database for the structural analysis of protein-
DNA complexes. Nucleic Acids Res 38:D91–D97
45. Lewis BA, Walia RR, Terribilini M, Ferguson J, Zheng C, Honavar V, Dobbs D (2011) PRIDB:
a protein-RNA interface database. Nucleic Acids Res 39:D277–D282
46. Kim R, Guo J-t. (2009) PDA: an automatic and comprehensive analysis program for protein-
DNA complex structures. BMC Genomics 10:S13
47. Lee S, Blundell T (2009) BIPA: a database for protein-nucleic acid interaction in 3D structures.
Bioinformatics 25:1559–1560
48. Cauet E, Rooman M, Wintjens R, Lievin J, Biot C (2005) Histidine-aromatic interactions
in proteins and protein-ligand complexes: quantum chemical study of X-ray and model
structures. J Chem Theory Comput 1:472–483
49. Copeland KL, Anderson JA, Farley AR, Cox JR, Tschumper GS (2008) Probing pheny-
lalanine/adenine π-stacking interactions in protein complexes with explicitly correlated and
CCSD(T) computations. J Phys Chem B 112:14291–14295
50. Copeland KL, Pellock SJ, Cox JR, Cafiero ML, Tschumper GS (2013) Examination of
tyrosine/adenine stacking interactions in protein complexes. J Phys Chem B 117:14001–14008
51. Gromiha MM, Santhosh C, Suwa M (2004) Influence of cation-π interactions in protein-DNA
complexes. Polymer 45:633–639
52. Gromiha MM, Santhosh C, Ahmad S (2004) Structural analysis of cation-π interactions in
DNA binding proteins. Int J Biol Macromol 34:203–211
53. Biot C, Buisine E, Kwasigroch JM, Wintjens R, Rooman M (2002) Probing the energetic and
structural role of amino acid/nucleobase cation-π interactions in protein-ligand complexes. J
Biol Chem 277:40816–40822
54. Wintjens R, Lievin J, Rooman M, Buisine E (2000) Contribution of cation-π interactions to

the stability of protein-DNA complexes. J Mol Biol 302:395–410
55. Rooman M, Lievin J, Buisine E, Wintjens R (2002) Cation-π/H-bond stair motifs at protein-
DNA interfaces. J Mol Biol 319:67–76
56. Biot C, Buisine E, Rooman M (2003) Free-energy calculations of protein-ligand cation-
π and amino-π interactions: from vacuum to proteinlike environments. J Am Chem Soc
125:13988–13994
57. Biot C, Wintjens R, Rooman M (2004) Stair motifs at protein-DNA interfaces: Nonadditivity
of H-bond, stacking, and cation-π interactions. J Am Chem Soc 126:6220–6221
58. Wintjens R, Biot C, Rooman M, Lievin J (2003) Basis set and electron correlation effects on
ab initio calculations of cation-π/H-bond stair motifs. J Phys Chem A 107:6249–6258
59. Auffinger P, Bielecki L, Westhof E (2004) Anion binding to nucleic acids. Structure 12:379–
388
60. Singh NJ, Min SK, Kim DY, Kim KS (2009) Comprehensive energy analysis for various types
of π-interaction. J Chem Theory Comput 5:515–529
61. Zacharias N, Dougherty DA (2002) Cation-π interactions in ligand recognition and catalysis.
Trends Pharmacol Sci 23:281–287
62. Dougherty DA (2012) The cation−π interaction. Acc Chem Res 46:885–893
63. Churchill CDM, Wetmore SD (2009) Noncovalent interactions involving histidine: the effect
of charge on π-π stacking and T-shaped interactions with the DNA nucleobases. J Phys Chem
B 113:16046–16058
64. Frontera A, Quinonero D, Costa A, Ballester P, Deya PM (2007) MP2 study of cooperative
effects between cation-[small pi], anion-[small pi] and [small pi]-[small pi] interactions. New
J Chem 31:556–560
65. Mahadevi AS, Sastry GN (2012) Cation−π interaction: its role and relevance in chemistry,
biology, and material science. Chem Rev 113:2100–2138
66. Schottel BL, Chifotides HT, Dunbar KR (2008) Anion-[small pi] interactions. Chem Soc Rev
37:68–83
67. Jurecka P, Sponer J, Cerny J, Hobza P (2006) Benchmark database of accurate (MP2 and
CCSD(T) complete basis set limit) interaction energies of small model complexes, DNA base
pairs, and amino acid pairs. Phys Chem Chem Phys 8:1985–1993
68. Cysewski P (2008) A post-SCF complete basis set study on the recognition patterns of uracil
and cytosine by aromatic and π-aromatic stacking interactions with amino acid residues. Phys
Chem Chem Phys 10:2636–2645
69. Ebrahimi A, Habibi-Khorassani M, Gholipour AR, Masoodi HR (2009) Interaction between
uracil nucleobase and phenylalanine amino acid: the role of sodium cation in stacking. Theor
Chem Acc 124:115–122
70. Santos LA, da Cunha EFF, Freitas MP, Ramalho TC (2014) Hydrophobic noncovalent interac-
tions of inosine-phenylalanine: a theoretical model for investigating the molecular recognition
of nucleobases. J Phys Chem A 118:5808–5817
71. Rutledge LR, Campbell-Verduyn LS, Hunter KC, Wetmore SD (2006) Characterization of
nucleobase-amino acid stacking interactions utilized by a DNA repair enzyme. J Phys Chem
B 110:19652–19663
72. Rutledge LR, Wetmore SD (2008) Remarkably strong T-shaped interactions between aromatic
amino acids and adenine: their increase upon nucleobase methylation and a comparison to
stacking. J Chem Theory Comput 4:1768–1780
73. Rutledge LR, Durst HF, Wetmore SD (2008) Computational comparison of the stacking inter-
actions between the aromatic amino acids and the natural or (cationic) methylated nucleobases.
74. Rutledge LR, Campbell-Verduyn LS, Wetmore SD (2007) Characterization of the stacking
interactions between DNA or RNA nucleobases and the aromatic amino acids. Chem Phys
Lett 444:167–175
75. Rutledge LR, Durst HF, Wetmore SD (2009) Evidence for stabilization of DNA/RNA-protein
complexes arising from nucleobase-amino acid stacking and T-shaped interactions. J Chem
Theory Comput 5:1400–1410
76. Churchill CDM, Navarro-Whyte L, Rutledge LR, Wetmore SD (2009) Effects of the biological
backbone on DNA-protein stacking interactions. Phys Chem Chem Phys 11:10657–10670
77. Churchill CDM, Rutledge LR, Wetmore SD (2010) Effects of the biological backbone on
stacking interactions at DNA-protein interfaces: the interplay between the backbone-pi and
pi-pi components. Phys Chem Chem Phys 12:14515–14526
78. Rutledge LR, Navarro-Whyte L, Peterson TL, Wetmore SD (2011) Effects of extending the
computational model on DNA-protein T-shaped interactions: the case of adenine-histidine
dimers. J Phys Chem A 115:12646–12658
79. Biot C, Buisine E, Rooman M (2003) Free-energy calculations of protein−ligand cation−π
and amino−π interactions: from vacuum to proteinlike environments. J Am Chem Soc
125:13988–13994
80. Leavens FMV, Churchill CDM, Wang S, Wetmore SD (2011) Evaluating how discrete water
molecules affect protein-DNA π-π and π+-π stacking and T-shaped interactions: the case
of histidine-adenine dimers. J Phys Chem B 115:10990–11003
81. Rutledge LR, Churchill CDM, Wetmore SD (2010) A preliminary investigation of the addi-
tivity of π − π or π + −π stacking and T-shaped interactions between natural or damaged
DNA nucleobases and histidine. J Phys Chem B 114:3355–3367
82. Spiwok V, Lipovová P, Skálová T, Buchtelová E, Hašek J, Králová B (2004) Role of
CH/π interactions in substrate binding by Escherichia coli β-galactosidase. Carbohydr Res
339:2275–2280
83. SpiwokV, Lipovová P, Skálová T,Vondráčková E, Dohnálek J, Hašek J, Králová B (2006) Mod-
elling of carbohydrate-aromatic interactions: ab initio energetics and force field performance.
J Comput-Aided Mol Des 19:887–901
84. Sujatha MS, Sasidhar YU, Balaji PV (2004) Energetics of galactose- and glucose-aromatic
amino acid interactions: Implications for binding in galactose-specific proteins. Protein Sci
13:2502–2514
85. Wimmerová M, Kozmon S, Nečasová I, Mishra SK, Komárek J, Koča J (2012) Stacking
interactions between carbohydrate and protein quantified by combination of theoretical and
experimental methods. PLoS ONE 7:e46032
86. Laughrey ZR, Kiehna SE, Riemen AJ, Waters ML (2008) Carbohydrate-pi interactions: what
are they worth? J Am Chem Soc 130:14625–14633
87. Bertozzi CR, Kiessling LL (2001) Chemical glycobiology. Science 291:2357–2364
88. Liang F-C, Chen RPY, Lin C-C, Huang K-T, Chan SI (2006) Tuning the conformation prop-
erties of a peptide by glycosylation and phosphorylation. Biochem Biophys Res Commun
342:482–488
89. Gabius H-J, Siebert H-C, André S, Jiménez-Barbero J, Rüdiger H (2004) Chemical biology
of the sugar code. ChemBioChem 5:740–764
90. Raju RK, Ramraj A, Vincent MA, Hillier IH, Burton NA (2008) Carbohydrate-protein recog-
nition probed by density functional theory and ab initio calculations including dispersive
interactions. Phys Chem Chem Phys 10:6500–6508
91. Su Z, Cocinero EJ, Stanca-Kaposta EC, Davis BG, Simons JP (2009) Carbohydrate-aromatic
interactions: a computational and IR spectroscopic investigation of the complex, methyl
alpha-L-fucopyranoside · toluene, isolated in the gas phase. Chem Phys Lett 471:17–21
92. Stanca-Kaposta EC, Gamblin DP, Screen J, Liu B, Snoek LC, Davis BG, Simons JP (2007)
Carbohydrate molecular recognition: a spectroscopic investigation of carbohydrate-aromatic
interactions. Phys Chem Chem Phys 9:4444–4451
93. Tsuzuki S, Uchimaru T, Mikami M (2011) Magnitude and nature of carbohydrate-aromatic
interactions in fucose-phenol and fucose-indole complexes: CCSD(T) level interaction energy
calculations. J Phys Chem A 115:11256–11262
94. Mohamed MNA, Watts HD, Guo J, Catchmark JM, Kubicki JD (2010) MP2, density functional
theory, and molecular mechanical calculations of C-H · · · π and hydrogen bond interactions
in a cellulose-binding module-cellulose model system. Carbohydr Res 345:1741–1751
95. Kozmon S, Matuška R, Spiwok V, Koča J (2011) Dispersion interactions of carbohydrates with
condensate aromatic moieties: theoretical study on the CH-π interaction additive properties.
96. Sujatha MS, Sasidhar YU, Balaji PV (2005) Insights into the role of the aromatic residue
in galactose-binding sites: MP2/6-311G++** study on galactose− and glucose−aromatic
residue analogue complexes. Biochemistry 44:8554–8562
97. Kumari M, Sunoj RB, Balaji PV (2012) Exploration of CH · · ·π mediated stacking interactions
in saccharide: aromatic residue complexes through conformational sampling. Carbohydrate
Res 361:133–140
98. Kozmon S, Matuška R, Spiwok V, Koča J (2011) Three-dimensional potential energy surface
of selected carbohydrates’ CH/π dispersion interactions calculated by high-level quantum
mechanical methods. Chem—A Eur J 17:5680–5690
99. Kumari M, Sunoj RB, Balaji PV (2012) Conformational mapping and energetics of saccharide-
aromatic residue interactions: implications for the discrimination of anomers and epimers and
in protein engineering. Org Biomol Chem 10:4186–4200
100. Tavagnacco L, Engström O, Schnupf U, Saboungi M-L, Himmel M, Widmalm G, Cesàro A,
Brady JW (2012) Caffeine and sugars interact in aqueous solutions: a simulation and NMR
study. J Phys Chem B 116:11701–11711

NoN Covalente Forces

Uploaded by

Copyright:

Available Formats

NoN Covalente Forces

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

NoN Covalente Forces

Uploaded by

Copyright:

Available Formats

What is the topic of discussion in the document?

What is the topic of discussion in the document?

What research areas are covered in the book series?

What research areas are covered in the book series?

Challenges and Advances in Computational

Chemistry and Physics

More information about this series at http://www.springer.com/series/6918

Challenges and Advances in Computational Chemistry and Physics

Library of Congress Control Number: 2015931919

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

1 Introduction to the Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Part I Hydrogen Bonds

2 Hydrogen Bonds Involving Sulfur: New Insights from ab Initio

3 CH· · · π Interaction in Organic Molecules . . . . . . . . . . . . . . . . . . . . . . . 47

4 The CH··O H-Bond as a Determining Factor

5 Hydrogen Bonds Involving Radical Species . . . . . . . . . . . . . . . . . . . . . . 107

6 Agostic and Hydrogen-Bonding X–H· · · M Interactions Involving a

7 What is Common for Dihydrogen Bond and H· · ·σ

Part II Other Bridging Atoms

8 The Pnicogen Bond in Review: Structures, Binding Energies,

9 Chalcogen Bonds in Protein Architecture . . . . . . . . . . . . . . . . . . . . . . . . 265

10 A Unified View of Halogen Bonding, Hydrogen Bonding and Other

11 The X-C· · ·Y Carbon Bond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

12 Interplay of Hydrogen, Halogen, Lithium and Beryllium Bonds

13 Understanding Lone Pair-π Interactions

Part III Aromatic Systems

14 Unraveling the Origin of Substituents Effects in π-Stacking

15 Noncovalent Interactions of Organic Ions with Polar Molecules

16 Anion-π Interactions in Supramolecular Chemistry and Catalysis 471

17 A Survey of DNA–Protein π–Interactions: A Comparison of Natural

Ibon Alkorta Instituto de Química Médica (IQM-CSIC), Madrid, Spain

Michio Iwaoka Department of Chemistry, School of Science, Tokai University,

Abstract The reader is introduced to the definition and diversity of noncovalent

between proton donor AH and B as an example, reaction 1 describes the association

E = E(AH · · · B) − {E(AH) + E(B)} (1.2)

1.2 Energy Dissection

positive cooperativity, or sometimes as synergistic. But the formation of a trimer

1.4 Electrostatic Potentials

improved by adding atomic dipoles, or even quadrupoles. There remains, however,

1.5 Atoms in Molecules (AIM)

1.7 Electron Density Redistributions

ρ = ρ(AH · · · B) − {ρ(AH) + ρ(B)} (1.3)

The resulting map offers a three-dimensional perspective on shifts within each

1.8 Perturbations of the Monomers

12. Jeziorski B, Moszynski R, Szalewicz K (1994) Perturbation theory approach to intermolecular

© Springer International Publishing Switzerland 2015 15

oxygen and nitrogen in specific. In fact, N–H· · · O, O–H· · · O, N–H· · · N and O–

at CCSD(T)/CBS level is − 2.85 kcal/mol. This binding energy is comparable to

Pyrrolidone carboxyl peptidase Glycogen phosphorylase b Aspartate aminotransferase

Catalase, human erythrocyte 5'-deoxy-5'-methylthioadenosine Creatine kinase

2.2 “Sulfur” as a H-Bond Acceptor

2.2.1 O–H· · · S Hydrogen Bonding

In such situations computed vibrational frequencies are very helpful in assigning

p-CR•EtOH − 217 − 422.4 746.0 − 256 0.0125 −6.33

Met- 149 His- 43

Bacteriochlorophyll Human thioredoxins Galactose oxidase

2.2.2 N–H· · · S Hydrogen Bonding