View Article Online / Journal Homepage / Table of Contents for this issue

REVIEW ARTICLE

www.rsc.org/pccp | Physical Chemistry Chemical Physics

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

Density functional theory for transition metals and transition

metal chemistry
Christopher J. Cramer* and Donald G. Truhlar*
Received 8th April 2009, Accepted 20th August 2009
First published as an Advance Article on the web 21st October 2009
DOI: 10.1039/b907148b
We introduce density functional theory and review recent progress in its application to transition
metal chemistry. Topics covered include local, meta, hybrid, hybrid meta, and range-separated
functionals, band theory, software, validation tests, and applications to spin states, magnetic
exchange coupling, spectra, structure, reactivity, and catalysis, including molecules, clusters,
nanoparticles, surfaces, and solids.

1. Introduction
Density functional theory (DFT) describes the electronic states
of atoms, molecules, and materials in terms of the threedimensional electronic density of the system, which is a great
simplication over wave function theory (WFT), which involves
a 3N-dimensional antisymmetric wave function for a system
with N electrons.1 Although DFT is sometimes considered the
new kid on the block, it is now B45 years old in its modern
formulation2 (more than half as old as quantum mechanics
itself), and it has roots3,4 that are almost as ancient as the
Schrodinger equation. In practical work, DFT is almost always
applied in the form introduced by Kohn and Sham,5 including
its spin-polarized extension.6,7 The basic quantity in DFT is the
many-electron spin density, r. The spin-polarized KohnSham
formalism involves a determinant formed from a set of N
ctitious single-particle spin-orbitals corresponding to a noninteracting system of electrons with the same spin densities, ra
and rb, as the real system, where r is the sum of ra and rb, the
spin density ra is the 3-dimensional electron density of all
spin-up electrons, and rb is the same for spin-down electrons.
In the original KohnSham formalism (applicable to closedshell molecules and nonmagnetic solids), ra equals rb. (The
original KohnSham formalism may also be labeled spinrestricted KohnSham or restricted KohnSham, and the
spin-polarized version may be called spin-unrestricted or
unrestricted.) We note that spin density is a generic term for
the density associated with the subset of electrons characterized
by the same denite value of Sz, i.e., either a or b (thus one might
say spin-up density and spin-down density rather than spin
densities). In a many-electron system comprised of both spin-up
and spin-down electrons, the term spin-density is also sometimes
used to refer to the position-dependent dierence between the up
and down spin densities, or to the vector analog of this quantity.
To avoid confusion, in this article we will refer to the dierence
density as the spin polarization density.
Density functional theory (DFT) has now become the
preferred method for electronic structure theory for complex

Department of Chemistry and Supercomputing Institute,

University of Minnesota, Minneapolis, MN 55455-0431, USA.
E-mail: [email protected], [email protected]

This journal is


c

the Owner Societies 2009

chemical systems, in part because its cost scales more favorably

with system size than does the cost of correlated WFT, and yet
it competes well in accuracy except for very small systems. This
is true even in organic chemistry, but the advantages of DFT
are still greater for metals, especially transition metals. The
reason for this added advantage is static electron correlation.
It is now well appreciated that quantitatively accurate
electronic structure calculations must include electron correlation.
It is convenient to recognize two types of electron correlation,
the rst called dynamical electron correlation and the second
called static correlation, near-degeneracy correlation, or nondynamical correlation. Dynamical correlation is a short-range
eect by which electrons avoid one another to reduce electron
repulsion. It is a very general eect for all nite systems
containing two or more electrons. Accounting for dynamical
correlation by a conguration interaction wave function is
very slowly convergent and requires a very large number of
congurations. Other correlation eects, which are very system
specic and can be either medium ranged or long ranged, can
be accounted for to a large extent by mixing a small number
(sometimes two, sometimes more) of congurations that are
nearly degenerate.8,9 Such correlation eects are called
static or near-degeneracy correlation and systems exhibiting
signicant static correlation eects are often called multireference systems;10 likewise, WFT methods based on
multi-congurational zero-order states are often called multireference methods. Due to partially lled d subshells, and
nearly degenerate (n + 1)s and nd subshells, systems containing
transition metals often have a plethora of low-lying nearly
degenerate states, and near-degeneracy correlation eects on
the ground-state structure, electron distribution, and energy of
transition metal systems can be very large. Even though static
correlation can often be accounted for to a zero-order
approximation by a small number of congurations, it is often
very dicult to include correlation eects in a well balanced
way in WFT calculations on multi-reference systems.11,12
DFT, however, remains simple for such systems and is often
surprisingly accurate. This empirical fact adds to the advantage
of computational eciency in making DFT a preferred
method for transition metal chemistry.
The subject of transition-metal DFT is too large for any
single review to be complete. We selected recent papers that
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10757

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

best illustrate the promise of DFT in a number of very active

areas of transition-metal chemistry, and we also include some
older references that help to put them in perspective. We
largely exclude biological applications, which deserve and
receive1317 their own reviews.
Section 2 reviews the theory, section 3 contains some
comments on methodology, and section 4 reviews validation
studies. Section 5 reviews recent applications and section 6
contains concluding remarks.

2. Overview of DFT and functionals

2.1

Fundamental background

KohnSham spin-orbitals, cjs where s is a or b and j denotes

the other quantum numbers, are obtained by a self-consistent
eld (SCF) calculation and are formally functions of the exact
density of the system. Then
rs

occ
X

jcjs j2

where s is the spin component (a or b), the spin-orbitals are

normalized, and the sum is over occupied orbitals of a given
spin component. The electronic energy of the system is
approximated as a sum of four terms, Tn, ene, eee, and exc.
Tn is the kinetic energy of a system of noninteracting electrons
with the same spin densities as the real system; ene is the
interaction of the electron distribution with the nuclear framework; eee is the classical Coulomb energy of the spin densities
interacting with each other and with themselves; and exc, called
the exchangecorrelation energy, is everything else (everything
except Tn, ene, and eee). Therefore exc includes the interaction
correction to Tn, the correction to Vee for the fact that real
electrons do not interact with themselves, the exchange energy
(due to the indistinguishability of electrons exchanging their
space and spin variables), and the correlation energy (due to
the fact that the many-electron spin densities are not
uncorrelated products of spin-orbital densities). exc is written
as a functional, called the spin-density functional, of the spin
densities. Since the KohnSham spin-orbitals are functions of the
spin-densities, exc can depend explicitly on the spin densities and
also implicitly on them by depending on the spin-orbitals. Direct
dependence on the spin-densities can also include a dependence
on their derivatives (e.g., the magnitudes of their gradients, their
Laplacians, etc.). The HohenbergKohn theorem2 shows that the
density functional exists; but a closed-form expression for the
exact spin-density functional does not exist, and there are no
systematic routes to improving an approximate functional.
Nevertheless useful approximations have been obtained,
andby a series of ts and startsthey keep getting better.
The eective potential corresponding to exc is generated
from a functional of the spin densities. The spin-polarized
KohnSham formalism and the spin-density functional are
usually just called KohnSham theory and the density
functional, and approximations to the latter are also called
density functionals. The density functional is usually written as
the sum of an exchange part and a correlation part. One
should be careful though because the meaning of these terms is
dierent in DFT and in WFT. In particular, DFT correlation
10758 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

includes only dynamic correlation, and DFT exchange

includes not only exchange but also some static correlation,
although the latter is present in an unspecied and
uncontrolled way.18,19 In subsection 2.2 we introduce
some exchange and correlation density functionals, and in
subsections 2.4 and 2.5 we describe a greater number of them.
Transition metal chemistry often involves open-shell systems
and excited states. KohnSham theory is not general enough to
treat all open-shell systems or excited states.2022 Although
generalized KohnSham theories have been advanced to
overcome this limitation, they are not in widespread use. Thus
one encounters diculties not just due to inaccurate density
functionals but due to KohnSham theory itself. There is a rich
theoretical literature on these questions, but it contains more
than one point of view, and much practical work in DFT
involves using approximate functionals in a rough-and-ready
way (see, e.g., section 5.1). Despite such issues, KohnSham
theory, even with approximate functionals, is often the most
accurate available approach for practical work on a given
system, especially for complex systems.
One type of generalized KohnSham theory that is in
widespread use is often called hybrid DFT, and it involves
combining HartreeFock exchange, which is orbital-dependent,
with explicit functions of local spin densities and their gradients.
(Note that one can also treat HartreeFock exchange within
the ungeneralized KohnSham framework by using the
optimized eective potential (OEP) method,23,24 but we
shall not consider this formalism in detail within the present
review.) Further generalizing hybrid functionals to include
dependences on the local Laplacians of the spin densities or
on the local spin kinetic energies computed from the spinorbitals yields what are called hybrid meta functionals, and
these are the most powerful functionals now available.
Although one should distinguish between KohnSham and
generalized KohnSham methods, and making fundamental
progress at extending DFT to arbitrarily complex systems
certainly will require careful attention to the fundamentals
of the theory,25 especially for magnetic properties,26 most
practical algorithms in popular use ignore such issues and
address the theory from a computational perspective with
approximate methods and procedures being accepted or
abandoned based on their success or failure for practical
predictions, i.e., empirically. Since the present review is
application oriented, we will concentrate on such practical
aspects and put these fundamental issues aside for now. We
note that extended coverage of applications, as opposed to the
theory, is included in this review and can be found in section 5.
One should be careful not to overinterpret the KohnSham
orbitals. They correspond to a ctitious noninteracting system
with the same electron density as the correct many-body
function. They are introduced primarily to get an approximation
for the kinetic energy, which equals
T = 12ta + 12tb

(2)

where
ts

occ
2 X
h
jrcjs j2
me j

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

where h is Plancks constant divided by 2p, and me is the mass

of an electron. Since the density computed from the
KohnSham orbitals is an approximation to the exact density
(it is inexact only because we do not know the true density
functional), one-electron properties like dipole moments are
meaningful. But most properties that depend on individual
orbitals should be interpreted with care. One important
exception is the orbital energy of the HOMO. This is correctly
interpreted27 as the negative of the lowest ionization potential
(for solids that would be the work function). Nevertheless
many studies, including a large number of those cited in this
review, do employ DFT molecular orbitals to interpret the
electronic origins of chemical bonding and reactivity.
An issue of great concern in many cases is the proper treatment
of noncovalent interactions. For two ground-state closed-shell
spherical atoms (for example, Zn and Ne), the interaction
potential at long distances decreases as R6, where R is the
distance between the atoms. In the region where the electron
densities of the interacting monomers do not overlap, the longrange forces can be expressed in an asymptotic series in powers of
1/R by using WFT, a multipole expansion, and perturbation
theory, and the forces in this region of R are called dispersion
forces. The asymptotic series diverges at smaller R,28 and this is
sometimes modeled by truncating the asymptotic series and
adding damping factors to the retained terms.29,30 Semiempirical
analytic functions for potential energies are usually called
molecular mechanics (MM), and MM terms of this form are
sometimes added to DFT to improve the interaction potentials at
large R, but this is not a fully satisfactory method for several
reasons. First of all, the semiempirical damping functions are
somewhat arbitrary and must be readjusted when the underlying
DFT method is improved. Secondly, the usual functional forms
are not uniformly valid, failing, for example, for metallic nanostructures.31 A completely satisfactory solution to including longrange noncovalent interactions in DFT will probably ultimately
be based on the random phase approximation for correlation
energy,32,33 but in chemistry we are usually more interested in
noncovalent interactions at medium range, for example at the
distances typical of van der Waals molecules or of 1,3 interactions
(geminal interactions). At such distances, overlap of the change
distributions of the interacting moieties cannot be neglected (for
example, at the van der Waals minimum the gradient of the sum
of the repulsive interactions is equal in magnitude to the gradient
of the sum of the attractive interactions), and some recent density
functionals, especially those involving kinetic energy density that
were developed with special attention to medium-range exchange
and correlation energy (see section 2.2), do already seem to
provide very useful estimates of the medium-range correlation
energy in such interactions.3440 The best way to include mediumrange correlation energy in DFT is not understood in satisfactory
detail, just as it is not understood in satisfactory detail
how correlation functionals include other kinds of dynamical
correlation energy.
There are so many reviews of DFT that even a review of the
reviews would be very long, but it is still useful to point the reader
to some previous reviews. For fundamental background we
mention only a few particularly lucid expositions.1,4149 We also
mention two recent papers by Perdew and coworkers that give
useful perspective on various aspects of DFT.50,51
This journal is


c

the Owner Societies 2009

2.2 Introduction to functionals

(We consider only collinear functionals in this section, that is,
functionals for collinear DFTsee section 3.2 for a denition
of collinear DFT.)
The oldest approximation to a density functional is the
DiracSlater approximation52,53 to exchange. This must be
renormalized for use with KohnSham theory.5 This is now
usually called the local spin density approximation (LSDA)
since it depends only on spin densities (not their derivatives or
orbitals). It can be derived from the exact exchange energy of a
uniform electron gas (UEG), which is a somewhat unphysical
system in which a constant electron density is neutralized by a
constant background positive charge (rather than by discrete
nuclear charges). The UEG correlation energy can be
calculated numerically54 and t in various ways55,56 and that
leads to the LSDA for correlation, which has recently been
thoroughly reviewed.57
The next level of complexity in density functionals is to add
a dependence on the gradients of the spin densities; in
particular the functional depends on the unitless reduced
spin-density gradients, ss, which are proportional to |rrs|/r4/3
s .
(Usually, though, we just say it depends on rrs.) Such
functionals are called generalized gradient approximations
(GGAs). Popular GGAs include BP86, where B denotes
Beckes 1988 exchange functional (usually abbreviated as
B88 or just B),58 and P86 denotes Perdews 1986 correlation
functional;59 BLYP, where LYP denotes the LeeYangParr
correlation functional;60 PW91, from Perdew and Wang in
1991;61 and a functional of Perdew, Burke, and Ernzerhof
(PBE).62 The modied PerdewWang functional of Adamo
and Barone,63 called mPWPW, is very similar to PBE. Notice
that GGAs may combine an exchange functional from one
source with a correlation functional from another, or they may
both be from the same source. Thus BP86 and BLYP combine
B88 exchange with P86 or LYP correlation, respectively;
PW91 combines PW91 exchange with PW91 correlation;
PBE combines PBE exchange with PBE correlation; mPWPW
combines mPW exchange with PW91 correlation; SLYP
combines the Slater LSDA exchange with LYP correlation;
and PBELYP combines PBE exchange with LYP correlation.
Density functional theory with LSDA or GGA functionals
includes self-exchange and self-correlation, both of which are
unphysical. As a consequence such functions tend to predict
too small a HOMOLUMO gap in molecules or too small a
band gap in solids. Furthermore they tend to underestimate
the relative stability of high-spin states in molecules or of high
magnetic moments in solids. An important consequence of the
error in LSDA and GGA exchange functionals is that an
electron interacts with its own charge density; this unphysically
raises the energy of localized states and causes DFT to
produce excessively delocalized charge distributions6473 and
to incorrectly predict some materials to be metals rather than
insulators. Systems for which these errors are especially severe
are sometimes called strongly correlated systems because
delocalization is associated with the dominance of kinetic
energy terms, and localization is associated with dominance
by screened Coulomb potentials,74 and electron correlation is
important for electron localization because correlation
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10759

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

minimizes the interatomic repulsion of electrons near the same

center. The connection between covalency and an insulator
gap has been especially well studied in the metal oxide
insulators like MnO, FeO, CoO, and NiO.7592 By including
partial or full HartreeFock exchange one can decrease or
eliminate, respectively, self-exchange, and by including kinetic
energy density one can eliminate self-correlation50,79 and
decrease the sensitivity to the percentage of HartreeFock
exchange.39 HartreeFock exchange and kinetic energy
density thus reduce some of the inaccuracies of the
LSDA and GGA functionals while retaining many of the
computational advantages of GGAs as compared to WFT.
The density functionals that perform best for main-group
chemistry are not the same as those that perform best for
transition metals.9395 Whereas in solid-state chemistry local
functionals are often chosen, partly because they are easier to
apply to extended systems, in organic chemistry hybrid
functionals are the more typical choice because of their
demonstrated superior predictions of energetics; by far the
most popular such hybrid functional is B3LYP, which is a
hybrid GGA put together by Stephens et al.96 on the basis of
earlier work by Becke and others,58,60,97 especially a hybrid
GGA called B3PW91.97 The correlation functionals of
B3PW91 and B3LYP are based on PW91 and LYP,
respectively but are optimized specically for use in a hybrid
functional,97,98 whereas the most straightforward hybrid
functionals result from simply replacing a percentage
(here called X) of local density functional exchange by
HartreeFock exchange. Such functionals often have a 1
in their name to denote this one parameter (although
there are usually other parameters in the local part of the
exchangecorrelation functional). Examples include B1LYP99
(BLYP with X = 25), mPW1PW63 (mPWPW with X = 25),
PBE1PBE99,100 (PBE with X = 25, also called PBEh [but it is
not the only functional called PBEh] and here called PBE0,
which is its most common name), and MPW1K101 (mPWPW
with X = 42.8). One occasionally sees a functional, e.g.,
BH&HLYP, with X = 50. In comparison, B3LYP and
B3PW91 both have X = 20 (and each has two other new
parameters as well). Reiher et al.,102 on the basis of predicting
the relative energetics of various spin states of FeII complexes,
suggested adjusting X to 15 in B3LYP; they called the resulting
functional B3LYP*. Later, Brewer et al.103 found that X = 13
works best for a set of iron(II) and iron(III) complexes. Bredow
and Gerson81 proposed a PW91-based hybrid functional,
called HF + PWGGA81 or PW1PW,104 with X = 20, which
was optimized to the band gaps and thermodynamic and
geometric properties of MgO, CoO, and NiO.
In further related work, Radon et al.105 performed CASPT2
calculations, a method that uses second-order WFT perturbation
theory to add dynamical correlation to the complete active
space self-consistent-eld WFT method, CASSCF, which is a
multi-congurational method that includes only a small
fraction of the dynamical correlation. They assumed the
CASPT2 calculations on single molecules of bis(acetylacetonate)
cobalt(II), usually abbreviated Co(acac)2, to be of benchmark
quality, and they found a tetrahedral quartet ground state.
They then found that BP86, PBE, and TPSS (all with X = 0)
predict the ground state to be a planar doublet, whereas
10760 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

B3LYP, PBE0, and TPSSh (with X = 20, 25, and 10,

respectively) predict a tetrahedral quartet. The crystal is
known experimentally to involve planar Co(acac)2, where
the dierence from the single-molecule case is apparently
due to pp stacking interactions.
Mixing in HartreeFock exchange is not the only way to
include nonlocality. The nonlocal character of the true density
functional is more general and can be summarized by noting
that the exchangecorrelation energy density at a given point
in space depends not just on the local properties of the density
at that point but rather on the density everywhere. The
KohnSham orbitals also depend on the density everywhere.
Some methods of including nonlocality that are even older
than hybrid DFT are the weighted density approximation106111
(WDA), the self-interaction correction (SIC) method,112,113
the average-density SIC method,114,115 and the screened exchange
(sX) approximation of Bylander and Kleinman.110,111,116
The WDA is particularly interesting in that it is based on
approximating the exchangecorrelation charge density in a
way that conserves the total exchange charge and the depth of
the exchange hole.
An approximate version of the SIC method based on the
premise that most of the self-interaction error arises from
localized regions has been introduced,117,118 variously called
atomic SIC (ASIC) and pseudo-SIC. Pemmaraju et al.119
further developed the ASIC method and suggested it for use
in modeling quantum transport. Although ASIC accomplishes
some of its goals in correcting self interaction, it still suers
from the inadequacy of being built on the LSDA.120
Another approach to reducing these inaccuracies is to
invoke the so called DFT + U approximation,76,121,122 which
becomes LSDA + U or GGA + U, depending on the type of
density functional employed. The +U modication ameliorates
self-interaction by using system-dependent parameters (for
solids, the parameters should probably also depend on pressure
or molar volume and on phase, and for molecules on geometry).
The +U modication takes one out of the realm of DFT and
into a less rigorous model Hamiltonian approach, but it is
instructive in the way it corrects DFT. The method is approximate
and not uniquely dened123 and is employed in a variety of
inequivalent ways by dierent researchers, but the various
versions attempt to enforce the same physical corrections by
adding a Hubbard-like124 term to the DFT energy; this term
has the form (in the version78,85 of DFT + U employed in the
VASP program):
E DFTU  E DFT

UJX
Trrs  rs rs
2
s

where U  J (sometimes called U) is an empirical constant,

and rs is the one-electron density matrix of metal d electrons
(sometimes metal f electrons) for spin s. A key issue in
essentially all implementations is the inclusion of the
Hubbard-like term only in a basis of localized d electrons
(or f electrons). The Hubbard-like term adds a penalty for
non-idempotent density matrices in this subspace and therefore
it favors lling d orbitals that are localized on one particular
atom78 (a correlation eect), which sometimes also favors
high-spin states on each atom and therefore less covalent
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

bonding in molecules or antiferromagnetic coupling in solids.

The +U correction would be expected to be smaller for a
GGA than for LSDA, and smaller yet or not needed for a
well balanced hybrid functional; for accurate results, the +U
correction should depend on the structure.125
In some cases, for example solid MnO,87,126,127
FeO,88,126,127 and Co,88,126,127 the +U approach and the
inclusion of partial HartreeFock exchange lead to similar
results, which are much better than GGA, and one anticipates
that the methods would be even more similar if the Hubbardlike term were included for the entire electronic space rather
than just the metal d space.85
As just one example, we note that LSDA,128130 BLYP,131
PW91,129 and PBE130 predict that FeO is a metal (no band
gap), whereas LSDA with SIC,126,127 LSDA + U84 and hybrid
DFT (BLYP with X Z 10, from ref. 131) correctly predict it
to be an insulator. However, the detailed nature of the crystal
orbitals is predicted dierently by the various methods.131
A key issue in +U methods is that they approximate the
energy not only in terms of the electron density but more
generally in terms of the density matrix. Further development
of this line of approach includes LDA + DMFT132 (in which
the LSDA approximation is combined with dynamical mean-eld
theory). We will not include DFT + U or LDA + DMFT in
the rest of this review, but the reader is directed to a few
particularly relevant references for further reading.85,132149
A related approach that attempts to improve on LDA + U is
the recent Gutzwiller DFT150,151 (LDA + G).
2.3

Introduction to band theory

Since DFT saw its initial development in the physics

community but has since become an essential tool in chemistry,
there is a certain diculty in reconciling the conceptual frameworks used in the two disciplines. This is especially true for the
language developed around the important application of band
theory, and this language now occurs in many contexts in the
DFT literature.
As in many areas of chemical physics, a useful rst step is
dening terms, and that is especially important for chemists
reading the solid-state physics literature where many applications
of density functional theory to systems containing transition
metals have been reported. One key quantity associated with a
solid is its gap, but there is more than one way in which this
is dened; alternatively one can say that there is more than one
gap. First, one can dene the gap by using quasielectrons and
holes, sometimes called quasiparticles.152 The energy to
remove an electron from a system is called the ionization
potential (for metals it is called the work function), I, and it is
equal to EN1  EN where N is the number of electrons in the
material or molecule under study. Experimentally, I can be
measured by photoelectron spectroscopy (PES) [which is
also called photoemission spectroscopy and which could be
ultraviolet photoelectron spectroscopy (UPS) or X-ray photoelectron spectroscopy (XPS)] or by inverse photoemission
spectroscopy (IPES). The energy to add an electron is called
the electron anity, A, equal to EN  EN+1. One denition
of the gap is I  A.153 This is sometimes called the physical
gap, the quasiparticle gap, or the fundamental gap.
This journal is


c

the Owner Societies 2009

The gap is also sometimes associated with excitation

energies, as in optical spectroscopy. This may be called the
optical gap, and it is dened by the onset energy of absorption.
An optical excitation does not change N. In quasiparticle
language, the calculation of optical properties requires
calculating the interaction between a quasielectron (also called
an electron or a particle) and a hole (and their exchange
counterparts).154157 The optical spectrum of a solid is
the frequency-dependent imaginary part of the macroscopic
dielectric function, and the spectrum is sometimes calculated
using this equivalence.
Thirdly, the gap may be associated with the dierence
between the energy of the highest occupied molecular orbital
(HOMO) or highest occupied crystal orbital (HOCO) and the
lowest unoccupied molecular orbital (LUMO) or lowest
unoccupied crystal orbital (LUCO). This orbital-energy gap
is sometimes called the independent-particle gap, the singleparticle gap, or the one-electron gap. Strictly speaking, the
orbital-energy gap is only an approximation to the physical
quasiparticle gap or the optical gap. Thus, orbital-energy band
gaps are often 30100% below the quasiparticle band gaps in
semiconductors and insulators.158,159 The correction of the
orbital energy gap to the quasiparticle one is sometimes called
a quasiparticle shift160 or a quasiparticle correction,159 and the
correction of the physical gap to the optical one, due to
electronhole interaction,161 is sometimes called an exciton
shift or a local-eld eect. In practice there is typically some
cancellation between the quasiparticle shift and the exciton
shift because the former increases the gap, and electronhole
attraction lowers it (just as electronnucleus interactions in
atoms lower the bound states below the ionization potential).
The electronhole unit is a two-body quasiparticle, and it is
called an exciton. In semiconductors and other high-dielectric
materials, the electronhole interaction is often weak
(B0.1 eV), and the exciton is spread out over several unit cells
(this is called a MottWannier exciton). In insulators the
exciton may be localized and can be thought of as a mobile
electronically excited state of a molecule or formula unit (such
an exciton is called a Frenkel exciton); the exciton shift of a
Frenkel exciton can be very signicant, 1 eV or more.
Note that the expression band gap is itself ambiguous
since the word band is used to describe not only the
quasiparticle energies as a function of momentum hk but also
the independent-particle approximation to those curves.
The quasiparticle shifts may be approximated by the GW
approximation159,162165 for the self energy operator (also
called the mass operator or eective mass operator), where
G is a one-particle Greens function, and W is the dynamically
screened Coulomb interaction. The quasiparticle shift
calculated this way is sometimes called the GW shift. The
GW approximation is a many-body perturbative solution of
the exact equation, called the Dyson equation, for the selfenergy operator, and the calculations often begin with a DFT
calculation160,166169 (just as perturbative solutions for
molecules often begin with a HartreeFock calculation). The
non-self-consistent version of the GW method is called
G0W0.160,170 The solution to the Dyson equation may also
be approximated by uniform electron gas considerations.171
At a higher level one would solve the Dyson equation without
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10761

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

approximations. Exciton eects may be calculated by the

BetheSalpeter equation involving the two-particle Greens
function.159,163
Onida et al.159 have provided a comparison of the GW and
BetheSalpeter approaches commonly used in solid-state
physics to the TD-DFT approach commonly used for
optical spectra in chemical physics. Whereas the BetheSalpter
equation starts with the quasiparticle states corresponding to
(N  1)- and (N + 1)-electron systems, as calculated for
example by the GW approximation, TD-DFT works directly
with the N-electron systems.172 TD-DFT is based on the
response of a system to a time-dependent perturbation, and
Onida et al.159 point out that the systems response is directly
related to the N-particle excited states of an N-particle system
similar[ly] to the way the one-particle Greens function is
related to the (N + 1)- and (N  1)-particle excited states of
the same system. Both approaches are exact in principle but
subject to inevitable error due to the practical approximations,
such as approximate density functionals or the truncation173
of TD-DFT response at linear terms. The exchangecorrelation
potential of a TD-DFT calculation must provide the same
physics as self-energy corrections and electronhole interactions in the GW and BetheSalpeter calculations.159
In principle, the exchangecorrelation functional of
TD-DFT is time-dependent and depends on the entire past
history of the density. If the time-dependent potential changes
slowly, the system adjusts adiabatically, and one can approximate
the time-dependent exchangecorrelation functional by the
time-independent (local-in-time, frequency-independent)
ground-state exchangecorrelation functional. This is called
the adiabatic approximation.174 Most TD-DFT calculations
employ this approximation.
A complication with the traditional physics-literature
hierarchy described above is that hybrid functionals, which
until recently were the starting point for most chemistry
applications but very few physics ones, include some of the
interactions often included in the physics community by the
GW perturbative approach to the exchangecorrelation
functional. In particular, HartreeFock exchange is similar
to the screened exchange interaction in GW theory. Therefore
hybrid functionals can be a better starting point than local
functionals for GW calculations; this leads to smaller
quasiparticle shifts.160 Hybrid functionals can also be very
useful without GW shifts because they are less expensive than
the GW method and permit economical self-consistent
calculations of electronic eigenfunctions;83 the same can be
said for the sX approximation.175 The poor accuracy of band
gaps computed with local functionals may ultimately be
attributed to the nonanalytic dependence of the eective
potential on the density due to the derivative discontinuity
at integer numbers of electrons.153,176179
2.4

More types of functionals

Lets again start with language. Many chemists label any

functional that depends only on local properties, e.g., LSDA
or GGA, as local79,180184 and other functionals, such as
hybrid GGAs, are called nonlocal. This language focuses on
the algorithmically important consideration of whether the
10762 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

eective potential in the generalized KohnSham equations

can be applied as a multiplicative operator or requires an
integral operator (rather than focusing on the strict locality of
the functional relationship). We will follow this usage in the
present review. In the older literature, GGAs are sometimes
called nonlocal or gradient-corrected, but this usage is now
becoming uncommon. Some chemists call local functionals
pure but we deprecate this usage because it could imply to
nonspecialists that the unknown exact functional is local,
which is untrue. Some physicists separate local functionals
into strictly local, as in the LSDA, and semilocal, which is
used to describe GGAs (and meta functionals, which are
described in the next paragraph). We will not use the semilocal
language. Although we distinguished the KohnSham
equations from the generalized KohnSham equations in
section 2.1, in the rest of this article we will call them both
the KohnSham equations.
Local functionals may involve more than just spin densities
and reduced spin density gradients. The next level of complexity
is to introduce either r2ra or the magnitude of the spin kinetic
energy density. The latter is given by 12ts. For a one-electron
system, ra is equal to |c1a|2 and ta becomes 
h2|rra|2/4mera.
Thus the deviation of ta from this quantity may be used
to detect one-electron regions and eliminate spurious
self-interactions and self-correlation in such regions.79,185
Functionals that depend on the spin kinetic energy density
or r2ra are called meta functionals. The only established meta
functionals are meta GGAs, which depend on ra, rra, and
r2ra or on ra, rra, and ta. Adding HartreeFock exchange to
meta GGAs yields hybrid meta GGAs, which were already
mentioned in section 2.1.
Some early examples of meta GGAs are Becke95186 (usually
abbreviated B95) for correlation and TPSS187 for exchange
and correlation. The B95 correlation functional was originally
combined with B88 exchange, yielding BB95. Adding an
optimized amount of HartreeFock exchange to BB95 yields
B1B95 where the 1 stands for the one optimized parameter,
X. Optimizing X for TPSS yields a functional called TPSSh
where h denotes hybrid.
Notice that hybrid GGAs, meta GGAs, and hybrid meta
GGAs depend explicitly on the occupied spin orbitals as well
as on the spin densities and their derivatives. The next level of
complexity is to introduce a dependence on the unoccupied
spin orbitals. This allows one to model correlation energy in a
way analogous to how it is modeled in WFT. Functionals
depending on unoccupied orbitals are called fth-rung
functionals. Hybrid functionals depending on unoccupied
orbitals are sometimes called doubly hybrid functionals. Such
functionals have not yet been widely applied to transition
metals.
A very promising approach, also not yet applied widely to
transition metals, involves range-separated hybrid functionals
and local hybrid functionals. Range-separated functionals188196
involve separating the electronelectron interaction into a
short-range and a long-range part,188195 or even into three
parts196 (short-, middle-, and long-range) and treating the
contribution of one part of the exchange energy by the
HartreeFock method and the other part or parts by a local
exchange functional. In local hybrid functionals,197201 the
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

percentage of HartreeFock exchange that is mixed with

density functional exchange depends on the point in space.
The percentage could depend on any local function of the
point in space; for example, ta197,198 or the reduced spin
density gradient, ss.199201 A particularly successful rangeseparated functional is that of Heyd, Scuseria, and Ernzerhof
(HSE).191
2.5

Still more functionals

So far we have introduced the following functional types and

functionals:
1. local: LSDA approximation. Various approximations or
ts to UEG exchange and correlation energies. (All LSDA
functionals are very similar, and in the present review we will
not distinguish them.)
2. local: GGAs. BP86, SLYP, BLYP, PW91, PBE, PBELYP,
mPWPW, mPWLYP.
3. nonlocal: hybrid GGAs. B3PW91, B3LYP, B1LYP, PBE0,
MPW1K, B3LYP*, HSE.
4. local: meta GGAs. BB95, TPSS.
5. nonlocal: hybrid meta GGAs. B1B95, TPSSh.
In this section we introduce additional functionals of types 25.
GGAs. GGAs for exchange are dened by a single curve
corresponding to the enhancement of exchange as a function
of ss. For larger ss, addressing the GGA exchange functionals
mentioned in section 2.2, B88 has the largest enhancement,
PW91 has the lowest, and PBE and mPW are in between these
two. Other functional forms for this curve have also been
proposed, or sometimes the curve is found by multiparameter
optimization. In the former category we mention ten
particularly interesting exchange functionals. In 1996, Gill202
introduced a functional now called G96 with the goal of
nding an exchange functional that performs as well as B88
but has a simpler form. In 1998, Zhang and Yang203
introduced a revised PBE functional called revPBE, with the
goal of obtaining more accurate atomic absolute energies and
molecular atomization energies; then in 1999, Hammer
et al.204 introduced another (closely related) revision usually
called RPBE, with the goal of obtaining improved chemisorption
energies for small molecules on metal surfaces. In 2000, Vitos
et al.205 parameterized an exchange functional, called LAG,
against the local Airy gas model206 in an attempt to improve
the predictions for bulk properties of late transition metals and
semiconductors and the exchange energies of surfaces. The
local Airy gas is an edge electron gas, designed to mimic
electronic properties near an edge, such as a solid surface (that
is, a solid/vacuum interface) or the periphery of a molecule. In
2001, Handy and Cohen207 introduced an exchange functional
called OPTX, usually abbreviated O. This exchange functional
was designed to predict more accurate exchange energies of
atoms, but it does not satisfy the UEG limit. In 2004, Xu and
Goddard208 introduced an exchange functional called X.
This X exchange functional was designed to improve noncovalent interaction energies; it is very similar to mPW.
In 2005, Armiento and Mattsson209 presented a functional,
now called AM05, optimized against the jellium-surface
exchangecorrelation energy (which has recently been
claried210) and designed to give improved lattice constants,
This journal is


c

the Owner Societies 2009

bulk moduli, and vacancy formation energies. In 2006 Wu and

Cohen211 designed a new exchange functional, now called WC,
with the goal of improving lattice constants, crystal structures,
and metal surface energies. Their derivation was based on
attempting to enforce the expansion about the UEG limit
through fourth order in ss. However, due to an error in their
derivation, the claimed accuracy through fourth order does
not hold;212 their functional must therefore be judged on
empirical grounds rather than on the basis of the gradient
expansion. In 2008, Perdew et al.213 empirically modied the
PBE functional to improve the predicted lattice constants.
Their functional design strategy involved tting the jellium
surface exchangecorrelation energies, as done previously for
a GGA by Armiento and Mattsson. Also in 2008, Zhao and
Truhlar214 derived a functional, called SOGGA, that precisely
satises the gradient expansion through second order for
both exchange and correlation (none of the other functionals
discussed so far does this). The G96, OPTX, and X exchange
functionals are usually combined with the LYP correlation
functional yielding G96LYP, OLYP, and XLYP, respectively.
OPTX is also used with PBE correlation, yielding OPBE.
The LAG exchange functional is used with the LSDA for
correlation. The AM05, PBEsol, and SOGGA functionals
have corresponding correlation functionals designed to
complement them. The revPBE, RPBE, and WC functionals
are used with the PBE correlation functional.
The above functionals are not unrelated. None of them
satises all the known exact constraints on the density
functional, and they involve dierent choices of which
constraints to satisfy. Extensive discussion of the constraints
satised and not satised by these functionals, the relationships
between the functionals, and how the functionals characteristics
aect their performance has been provided by Zhang and
Yang,203 Perdew and coworkers,213,215,216 Mattsson and
coworkers,217 and Zhao and Truhlar.214 Mattsson
et al.218,219 have also provided further discussion of the
importance of surface energy. The bottom line of these
discussions is that dierent choices for the behavior of the
density functionals determine whether a functional will be
more accurate for interatomic spacings in lattices and
small-molecule bonds or for solid-state cohesive energies and
small-molecule energetics.214 The surface energy provides a
third dimension to this discussion.216

Hybrid GGAs. The very successful hybrid GGA called

B3LYP has already been mentioned in section 2.2.
Sousa et al.220 analyzed the number of occurrences of various
functional names in article titles and abstracts in the Web of
Science over the 19902006 period and found that 80% of the
references were to B3LYP, with a slightly higher percentage
over the 20022006 period. For the 19902006 period the
second most popular functional by this measure was the local
BLYP (5%), followed by the hybrid GGA called B3PW91
(4%) and the local BP86 (3%). Another hybrid GGA, B3P86,
which is similar to B3PW91 and B3LYP but with a dierent
correlation functional, was fth (2%). Hybrid density
functionals, however, are still very expensive for periodic
calculations on extended systems.221
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10763

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Other hybrid functionals that occur in this review are

X3LYP, which is like B3LYP but with the X exchange
functional replacing B88, and MPWLYP1M, which is like
mPWLYP, but with a nonzero percentage X of HartreeFock
exchange, optimized for metals.94
A number of more sophisticated hybrid GGAs have been
developed that go beyond combining separate GGAs for
exchange and correlation with an optimized value of X. These
include a sequence of multiparameter functionals from Becke,
Handy, and Tozer and their coworkers: B98,222 B97-1,223
B97-2,224 and B97-3.225 By some measures B97-3 is the most
accurate hybrid GGA for main-group chemistry, but,
like B3LYP, it is inaccurate in general for transition metal
chemistry.39

Meta GGAs and hybrid meta GGAs. Becke, Boese, Perdew,

and Savin and their collaborators and Zhao and Truhlar
developed functionals including ts in the correlation,186,226
the exchange,185,227229 or both.79,187,230,231 Most of these
density functionals can be used in either local or hybrid form.
The t-HCTH,207 BMK,208 and PW6B95205 functionals have
been especially widely tested, as have the M05 and M06
families discussed below. Prior to 2005, no available density
functional was among the better functionals for both
transition metal chemistry and main group barrier heights.
Good accuracy for both is important for studying, for
example, organometallic catalysis or organic reactions. The
former seemed to require zero or low HartreeFock exchange
(X \ 15), while the latter seemed to require high
HartreeFock exchange (X t 40). Furthermore, no available
functional predicted realistic noncovalent interactions for
weakly interacting systems. The Minnesota 2005 hybrid meta
functional,34 called M05, overcame both of these diculties,
with X = 28, by incorporating ts into exchange and correlation
in a balanced way, removing self-interaction errors in
exchange and self-correlation errors in correlation, enforcing
the UEG limit, and including a diverse set of data in the
parameterization, in particular main-group atomization
energies, ionization potentials, electron anities, barrier
heights, noncovalent interactions, and absolute atomic
energies and transition metal ionization potentials and bond
energies. The functional was parameterized against 35 data
and initially tested against 231 data. A key parameterization
strategy was to optimize X simultaneously with the other
parameters in the density functional to reduce the reliance
on compensating errors in the local component of exchange
with errors in the correlation functional. The resulting M05
functional showed, for the rst time, uniformly good results
for main-group thermochemistry, kinetics, and noncovalent
interactions and transition metal bond energies. No previous
functional was accurate for more than two of these four
categories.
In a subsequent eort, Zhao and Truhlar38 attempted to
nd the best functional with X = 0. The motivations for this
were twofold: (i) to learn more about the best functional form
and (ii) to develop a functional with lower cost for large and
extended systems. To obtain good results with X = 0 they
required a more general form for the local exchange functional.
10764 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

Great care was taken to be sure that exibility aorded by the

parameterization procedure was not too large for the size
and diversity of the chosen training sets. The resulting
functional, called M06-L, had the best overall performance
of any functional for a combination of thermochemistry,
thermochemical kinetics, metallochemical and noncovalent
interactions, bond lengths, and vibrational frequencies. The
worst performance area was barrier heights, although these
were predicted more accurately than by any other local
functional and with about the same accuracy as the popular
B3LYP functional.
In later work, Zhao and Truhlar designed a functional with
X = 100;232 this functional, called M06-HF, completely
removed one-electron self-interaction error, but it is not
recommended in general for transition metals or transitionmetal compounds because they often have multi-reference
character. Based on what was learned by designing M06-L
and M06-HF, Zhao and Truhlar redesigned and re-optimized
M05, yielding M06,39 which has X = 27, whereas M05 had
X = 28.
Zhao and Truhlar doubled the percentage of HartreeFock
exchange in M05 and M06, yielding M05-2X233 and
M06-2X,39 respectively, and they made two slightly improved
versions of M06-2X, called M08-HX and M08-SO.234
These functionals are not recommended for transition metal
chemistry, except for Zn, which Amin and Truhlar235 found to
have density-functional requirements more like a main-group
metal than a transition metal. (The Minnesota functionals
have not been tested for Cd and Hg.) We called M08-HX a
high exchange functional because there is no M08
functional whose exchange can be doubled. In this language,
low exchange would be X in the 515 range, standard
exchange would be X in the approximately 2030 range,
and high-exchange would be X in the 4060 range.
One advantage of M06-HF, as compared to most other
functionals, is that it eliminates the long-range self-interaction
error that has a disastrous eect on calculated charge transfer
excitation energies.39,232 Another way to eliminate the longrange self-interaction error is with range-separated functionals
that use HartreeFock exchange for the long-range part188,190
(unlike HSE,191 which uses it for the middle range).
An example of such a functional is CAM-B3LYP.236
Other general functionals encountered in this review are
VSXC237 (also called VS98), a meta functional, MPW1B95,238
a hybrid meta functional formed from mPW exchange and
B95 correlation, with X = 31, and OLAP3, a combination of
OPTX with the LAP3239 correlation functional.

Functionals with specic reaction parameters. Another type

of functional is one in which the parameters are adjusted not in
a general way for a broad range of systems, but rather to be as
accurate as possible for a specic reaction or a small range of
systems. Such a parametrization yields a function with specic
reaction parameters (SRP). The SRP approach was rst
developed for small-molecule reactions in the gas phase,240,241
and it has recently242 been extended to predict a potential
energy surface for H2 chemisorbing on Cu(111).
This journal is


c

the Owner Societies 2009

View Article Online

3. Methodology and software

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

3.1

Methodology: nonrelativistic aspects

Most applications of DFT in chemistry are carried out using

large electronic structure packages, and one way to understand
the available methodology is to review the capabilities of these
packages. Many of these packages are updated frequently, and
many of them have web pages where new versions and added
capabilities can be monitored. In addition in many cases the
capabilities of the packages have been summarized in journal
articles. This section includes some of those references.
One can distinguish various categories of calculations, e.g.,
molecules vs. solids, calculations with Gaussian basis sets vs.
those based on plane waves or both243,244 Gaussians and plane
waves, and all-electron calculations vs. those with eective core
potentials245265 (ECPs, also called pseudopotentials) to replace
some or all of the core electrons. The options for treating core
electrons are even more diverse when one includes methods like
the projector augmented-wave (PAW) method,266271 which is a
frozen-core all-electron method. Some programs can carry out
more than one kind of calculation; others can carry out only
one kind. Programs employing periodic boundary conditions
can include solid-state symmetry by sampling272274 many k
points in the Brillouin zone; they can be applied to crystalline
systems with nonperiodic defects or surfaces by the supercell
method,273,275,276 and they can be applied to liquids277 or
isolated molecules278,279 by evaluating only the G point in the
Brillouin zone. Several of the electronic structure packages also
include dynamics modules.280
For calculations on molecules, some programs employ or
can employ a procedure281284 called density tting or resolution
of the identity (RI) in which an auxiliary basis set is used to t
the electron density. In some cases this provides large savings
in computer time for calculations on large molecules. For
studying solids, a key procedure employed by several codes is
variable-cell-shape molecular dynamics,285,286 which can be
used to optimize structures as a function of pressure.
Another methodological element that is becoming increasingly
important is the combined quantum mechanical and molecular
mechanical (QM/MM) method, in which the electronic structure
of a primary subsystem is treated by explicit quantum
mechanics (QM) whereas the eect on the energy of the
secondary subsystem (the rest of the entire system) is included
by molecular mechanics. A key issue in such methods is how to
treat the boundary when it passes through one or more bonds.
The QM/MM method is widely used, for example in enzyme
chemistry,287 but the QM/MM boundary is not usually placed
through a bond to a transition metal. Ohnishi et al.288 have
considered the question of how to truncate the QM system at a
bond between a transition metal and a phosphorus. A longterm goal would be to develop a QM/MM method dened for
boundaries passing through any kinds of bond.289
3.2

Methodology: relativistic

Relativistic eects are non-negligible for late 3d transition

metals, important for 4d transition metals, and so important
for 5d transition metals that they must be included for even a
zero-order description.290 A fully relativistic calculation
This journal is


c

the Owner Societies 2009

involves the four-component Dirac spinor operator, but this

is seldom employed for transition metal chemistry, where, at
least until recently, at most a two-component formalism has
tended to be employed. A review from the point of view of
DFT is available.291
The reduction of the four-component formulation to a
useful two-component one can be accomplished by transformations developed by Foldy and Wouthuysen,292 Douglas
and Kroll,293 and Hess294298 and van Lenthe299 and their
coworkers, with the former work leading to the Douglas
KrollHess (DKH) Hamiltonian and the latter to the
zero-order-regular approximation299 (ZORA). Practical
two-component formulations of the DKH and ZORA
Hamiltonians suitable for use with DFT were reported by
Malkin et al.297 and Mayer et al.300 and by van Lenthe et al.301
and Wang et al.,302 respectively. The reduction to two
components eliminates the so-called small components
associated with positrons (although these are usually called
the small components, a more specic name is charge
conjugation components). In two-component calculations,
an atomic orbital is sometimes expressed as linear combination
of spin-up and spin-down orbitals associated, respectively,
with angular momentum j = l  12. A further reduction to a
one-component formulation is also possible; this yields the
spinorbit operator in its familiar form,303 as well as a
spinspin term. In the one-component formulation, the spin
direction is xed along an arbitrary axis (the z axis). This is
sometimes called the collinear approximation or the spin-free
formulation, even when the orbitals depend on the spin. In the
collinear approximation, the spin polarization density is simply
the dierence of the spin-up density and the spin-down
density. (Note that the spin polarization density is usually
called the spin density, but the term spin density is also used
to refer to the spin-up and spin-down densities in the collinear
approximation, so there is a possible source of confusion.)
In two-component calculationsin contrast to a collinear
calculation, in which every electron at every point in space is
either spin-up or spin-down along the same axisone has the
exibility for each one-electron orbital to have a spin pointing
along any axis and, furthermore, the spin associated with a
given orbital need not point in the same direction at all points
in space. A formulation with this exibility is called a noncollinear treatment.7,297,300 This is accomplished by writing
each orbital as a general spinor, that is, as a linear combination
of a spin-up orbital times a spin-up spin function and a
spin-down orbital times a spin-down spin function.
The relativistic formulation leads to a more satisfactory
denition of the spin polarization density.304 In particular the
spin polarization density is dened using the length of the spin
magnetization vector rather than z component of the spin
magnetization vector, which means that the spin polarization
vector has the desirable property of being invariant to
rotations in spin space.7 When we refer to a noncollinear
treatment, we imply this improved denition of the spin
polarization density.
In a spin-restricted DFT calculation, all spin orbitals have
the form of a product of a spatial ket and spin ket, all spin kets
have the same xed quantization axis, and every occupied
spin-up spin orbital is paired with a spin-down spin orbital
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10765

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

having the same spatial part. In a spin-polarized or

spin-unrestricted DFT calculation, the spin orbitals still
have the form of a space-spin product with a single quantization
axis of the spin parts, but now the spatial orbitals are not
doubly occupied; for example the spatial part of a spin-up
1s orbital need not be the same as the spatial part of a
corresponding 1s spin-down orbital. In a noncollinear calculation,
every orbital is a general spinor, with complex spatial parts.
(As mentioned in section 1, it is common to say restricted or
unrestricted rather than spin-restricted or spin-unrestricted.
Furthermore, the noncollinear formalism is sometimes called
the generalized spin orbital305 or general spin orbital306
description.) One can use general spinors not only in twocomponent relativistic calculations, but also in nonrelativistic
SCF calculations; for HartreeFock theory this is
called general HartreeFock theory (more general than spinunrestricted HartreeFock theory),307309 and as a generalization
of KohnSham theory it is called generalized spin density
functional theory310 or noncollinear spin density functional
theory.311 These nonrelativistic formulations are useful for
treating biradicals and certain magnetic problems.
The spin-dependent version of DFT developed by von Barth
and Hedin6 and Rajagopal and Callaway7 was originally
formulated for a general noncollinear system, but until
recently it has usually been applied only to spin-polarized
systems in the collinear approximation. The key distinction is
that a noncollinear density functional depends on the
o-diagonal elements of the spin density matrix as well as
the diagonal elements.
An illustrative example of a system requiring a noncollinear
treatment is provided by a spin frustrated Cr monolayer on a
face-centered-cubic Cu(111) substrate312 or an unsupported
Cr(111) monolayer.313,314 However, noncollinear magnetism
can also be important in molecules as small as Fe3,315317
Fe5,315319 Cr3,314,320,321 and Cr5.320322
Relativistic eects may be classied into scalar eects and
vector eects; the former are due to massvelocity and Darwin
terms in the relativistic kinetic energy, and the latter arise from
magnetic interactions involving the operators associated with
spin and orbital angular momentum.290 Spinorbit eects can
be included by treating the spinorbit terms by perturbation
theory or variationally in one-component calculations or by a
two-component calculation.290
The main eect of the scalar relativistic terms is to shrink
the s orbitals and, to a lesser extent,290,323325 the p orbitals.
This eect can be added to a nonrelativistic formulation by
using ECPs determined in relativistic calculations on atoms.
Such relativistic ECPs were recently tested for calculations on
PdCO, where the relativistic eects are scalar in nature,
and were found326 to perform excellently in reproducing
calculations327 carried out with a relativistic Hamiltonian
based on the regular approximation. When relativistic eects
are to be treated explicitly, all-electron basis sets optimized for
such calculations are required for good accuracy.328332
3.3

Software

The Gaussian program333 is based on Gaussian basis sets. It is

the most widely used program for isolated molecules, but it
10766 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

also supports periodic boundary conditions, as well as calculations

with continuum solvation. Molecular calculations with local
functionals can employ density tting. It includes TD-DFT as
well as DFT. It is very exible in the choice of basis sets,
eective core potentials, and density functionals, and it has
excellent geometry optimizers and initial guesses for the
self-consistent-eld iterations. It supports analytic Hessians334
for all functionals (even meta and hybrid meta functionals),
and it has classical dynamics capabilities and a very useful
interface to external programs. An external interface to the
335,336
POLYRATE
variational transition state theory code is available.
The Vienna Ab initio Simulation Package278,337,338 (VASP)
is the most widely used program for calculations on solids and
surfaces. (The name is confusing since most implementations
of density functional theory are not ab initio in the sense that
ab initio is used by chemists. However, many physicists and
materials scientists label all DFT methods as ab initio or rst
principles methods.) VASP employs plane waves with either
pseudopotentials or the PAW method. It includes DFT + U
and GW capabilities and has a dynamics module. VASP
has especially powerful algorithms for converging the
self-consistent-eld iterations even for systems with near
degeneracy of HOCO and LUCO.
The WIEN2k code339,340 is another well established code.
It uses the full-potential linearized augmented plane wave
(FP-LAPW) method, which uses a spherical harmonic basis
set inside of atomic spheres and a plane wave expansion in the
interstices.
In a pair of landmark papers, Kresse and coworkers used
the above three programs to demonstrate that if one converges
the results with respect to basis set, one can obtain the same
results with plane waves and Gaussians for molecules279 and in
some cases for solids.341 For molecules the Gaussian and plane
wave approaches were shown to become equivalent when the
Gaussian basis set was increased to augmented correlationconsistent polarized valence quintuple-z and the plane wave
cuto for PAW calculations was increased to 6070 Ry; the
equivalence was demonstrated for both local and hybrid
density functionals.279 For solids,193,341 reasonably good
agreement was obtained between plane wave and Gaussian
calculations with the range-separated HSE hybrid functional
for most insulators, semiconductors, and simple metals, but
not for open-shell transition metals. (Gaussian calculations
with full-range hybrids (also called global hybrids) like PBE0
are still prohibitively costly because of the slow convergence
for the long-range part of the exchange interaction.341)
For both solids341 and surfaces342 the VASP-PAW and
WIEN2k-FP-LAPW calculations agreed well.
The Amsterdam Density Functional (ADF) program,343345
somewhat uniquely, uses Slater-type orbitals rather than
Gaussians. SCF calculations are performed only for local
functionals, but post-SCF energies can be evaluated for hybrid
functionals. ADF uses density tting to reduce computational
cost, and it treats scalar and vector relativistic eects by the ZORA
method. For LSDA and GGA functionals, analytic Hessians are
available. A companion program, called BAND,346348 carries out
periodic calculations for bulk crystals, polymers, and surfaces;
various tools, such as a search program for transition states in
heterogeneous catalysis, are available.
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

The NWChem package349,350 contains both a Gaussian

module and an independent pseudopotential plane wave
module. The program is especially designed for massively
parallel calculations. Both internal (DRDY module) and
external interfaces to the POLYRATE335,336 variational transition
state theory code are available.
The General Atomic and Molecular Electronic Structure
System (GAMESS) package351,352 has a broad range of
Gaussian-based electronic structure capabilities, many of
which are parallelized. Solvent eects can be modeled by
eective fragment models. GAMESSPLUS353 allows one to
apply a well validated continuum solvation model.354
Jaguar is the electronic-structure program in the suite of
computational chemistry and drug design program oered by
Schrodinger, Inc. Jaguar carries out calculations with
Gaussians for both isolated molecules and molecules in continuum
solvent.355 An external interface to the POLYRATE335,336
variational transition state theory code is available.
MOLCAS is also a Gaussian-based electronic structure
program for studying molecules.356 Methods are included for
modeling eects of solvents, embedding in ionic solids, or
macromolecular environments.
QChem357 is a modern quantum chemistry package
available either as a stand-alone code or integrated with the
Spartan 08 package from Wavefunction, Inc. It employs
Gaussian basis sets and includes a well validated continuum
solvation model.354 QChem has been integrated with
the CHARMM molecular mechanics and classical dynamics
program.358
TURBOMOLE359 is a Gaussian-based program with a fast
RI algorithm, a TD-DFT option, and analytic Hessians.
Density tting and RI methods use an auxiliary Gaussian
basis set to describe the electron density, in addition to the
primary Gaussian basis set used to describe the orbitals.
DALTON 2.0360 is a Gaussian-based program including
linear, quadratic, and cubic361 response functions for both
singlet and triplet perturbing operators; the properties section
is especially complete.
QUICKSTEP244 is the electronic structure module of the
CP2K program. It is a pseudopotential program that uses
Gaussian basis functions to describe the orbitals and an
auxiliary plane wave basis to describe the electron density. The
CP2K package provides molecular dynamics capability.362
MOLPRO363 is an electronic structure program with very
fast post-HartreeFock capabilities, such as coupled cluster
calculations. It includes a DFT module, and DFT orbitals can
be used in coupled cluster calculations.
ORCA364 is a Gaussian-based quantum chemistry program
with special emphasis on spectroscopic properties of open-shell
molecules. It includes hybrid, meta, and hyper functionals.
The Spanish Initiative for Electronic Simulations with
Thousands of Atoms (SIESTA)365367 is a program designed
for linear-scaling calculations on materials by employing
a basis set of numerical nite-support atomic orbitals. It
employs periodic boundary conditions.
The density of Montreal package368 (deMon) carries out
DFT calculations in a Gaussian basis with density tting.
It includes TD-DFT. It was merged with the ALLCHEM
program,369,370 and the merged code is called deMon2k.371
This journal is


c

the Owner Societies 2009

DMol3 is a program that uses numerical functions on an

atom-centered grid as basis functions and can be used for
calculations in the gas phase, solvent, or solid state. Only local
functionals are supported.372374
The Octopus program375,376 carries out DFT calculations
including the inuence of time-dependent electromagnetic
elds. It can be used to calculate linear and non-linear
absorption spectra, harmonic spectra, laser-induced fragmentation,
and electronion dynamics of systems, from small clusters to
medium-sized quantum dots.
PWSCF377 (see also http://www.pwscf.org/) is the electronic
structure module of the opEn Source Package for Research in
Electronic Structure, Simulation and Optimization (Quantum
Espresso) package, which adds capabilities for classical
dynamics, geometry optimization, and transition state searches.
PWSCF is a plane wave pseudopotential code for solid-state
calculations. A linear scaling algorithm for hybrid functionals
has recently been developed for Quantum Espresso.378
The CRYSTAL electronic structure code379,380 forms Bloch
orbitals for periodic calculations as linear combinations of
Gaussians. Some examples of its usage are cited to illustrate its
capabilities.131,381,382 The tutorial article380 by the authors of
CRYSTAL is an excellent introduction to solid-state calculations
in chemistry. Whereas CRYSTAL, Quantum Espresso, and
VASP support hybrid calculations, most solid-state codes do
not support nonlocal functionals.
We also mention four other plane wave codes and examples
of their usage: ABINIT,383386 CASTEP,387391 Dacapo,271,392394
FLEUR,395 LMTART,396,397 and ParaGauss.300,398
Although we have emphasized computer programs based on
Gaussian basis functions and/or plane waves, we note that
some DFT calculations are instead accomplished by nite
dierence methods.399401
Examples of codes that can carry out noncollinear DFT
calculations are ABINIT, deMon2k, FLEUR, LMTART,
Paragauss, SIESTA, VASP, and Wien2k.

4. Validation studies
There are a large number of papers devoted to validating
density functional theory or containing a large validation
component. The benchmark data used for validation can be
either from experiment or, for simple enough systems, from
high-level WFT. (One must be cautious about WFT results,
even high-level ones, for transition-metal chemistry because if
multi-reference character is too large, WFT may not be
reliable.) In this section we briey review some recent validation
studies, beginning with those restricted to 3d metals and then
lifting this restriction, in each case summarizing in approximately
chronological order. (Additional validation studies are
included in section 5.) Most validation studies are discussed
in this section, but some solid-state validation is discussed in
section 5.3.2. Furthermore, some of the papers considered
in the application section (5.3.1) may be viewed as also
contributing to the validation eort.
Barden et al.402 considered the LSDA, BP86, BLYP, B3P86,
and B3LYP functionals applied to nine homonuclear
3d dimers (they also considered LSDA and a functional called
BHLYP that both performed very poorly and will not be
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10767

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

included in this summary). For bond energies, the mean

unsigned error (MUE) ranged from 19 kcal mol1 (BLYP)
to 30 kcal mol1 (B3P86). For bond distances the MUE
ranged from 0.020 A (BP86) to 0.053 A (B3LYP). For
vibrational frequencies the MUE ranged from 98 cm1
(BLYP) to 122 cm1 (B3P86). They characterized the
performance of DFT as surprisingly plausible, with BLYP
and BP86 rated best. Already in this study we see a trend
emerging. Whereas previous studies showed that hybrid
functionals usually perform better for main-group chemistry,
local functionals perform better for transition metal chemistry.
Diedrich et al.403 tested BP86 against the rst bond
dissociation energy of four 3d transition metal carbonyls.
The MUE was only 4 kcal mol1. First-row transition metal
monocarbonyls were also studied by Adamo and Lelj,40 who
concluded that inclusion of HF character in B3LYP was
important for the accurate calculation of vibrational frequencies
and metal-carbonyl dissociation energies.
Holthausen404 considered s/d excitation energies in 3d-series
atomic cations. A number of subtle issues such as choosing
orbital occupancies and articial mixing of 3d and 4s orbitals
were discussed. Unfortunately this study did not include scalar
relativistic eects, and it was pointed out405 that the conclusions
are quite dierent if these are included. The revised conclusion
is that local functionals are more accurate than hybrid
functionals for these intershell excitation energies. The most
accurate functionals, of 38 tested, are SLYP, PBE, BP86,
PBELYP, and PW91, with MUEs for these ve functionals
ranging from 2.8 to 3.7 kcal mol1. B3LYP had an MUE of
4.4 kcal mol1, and HCTH had an MUE of 15.1 kcal mol1.
It would be interesting to extend this study to more modern
functionals.
Furche and Perdew95 made a benchmark suite of 18 3d
transition metal reaction energies, twelve of which are bond
energies. They tested the LSDA, BP86, PBE, TPSS, B3LYP,
and TPSSh functionals. Excluding LSDA, which has an MUE
of 29 kcal mol1, the MUEs ranged from 10 kcal mol1 for
TPSSh to 12 kcal mol1 for B3LYP. In later work, this
benchmark test was extended to several more functionals,
with the following MUEs in kcal mol1: 8 for M05 and
B97-2, 9.5 for MPWLYP1M, and 11 for BLYP in the rst
extension,406 7 for M06-L, 8 for HCTH, 9 for VSXC and
OLYP, 10 for G96LYP and mPWPW, 12 for t-HCTH, and
12.5 for BB95 in the second,38 and 7 for M06, 9 for B97-3,
10 for B98, and 13 for BMK and PBE0 in the third.39
Furche and Perdew95 also examined the s/d excitation
energies of the neutral 3d-series atoms, plus Ca. They used a
procedure involving averaging non-self-consistent energies
over multiplet components. MUEs ranged from 8.3 kcal mol1
for B3LYP to 1820 kcal mol1 for PBE, BP86, and LSDA,
with TPSSh and TPSS at 14 and 17 kcal mol1, respectively.
Koster et al.370 found that GGA calculations of s/d excitation
energies of the neutral 3d atoms reproduce the experimental
sawtooth behavior as a function of atomic number, but
quantitative errors can be 15 kcal mol1 or more.
Zhao and Truhlar35 tested 18 density functionals for ZnNe,
ZnAr, ZnKr, and Zn2; these are all weakly bound complexes
with accurate binding energies in the range from 0.07 to
0.80 kcal mol1. B97-1, M05-2X, and PWB6K were the most
10768 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

accurate functionals tested for the binding energies of

Zn-rare gas dimers, and MPW1K, M05-2X, and PWB6K were
the most accurate for Zn2. Tests were also presented for
geometries.
Karttunen et al.407 tested B3LYP against WFT, in
particular, HartreeFock theory and second-order perturbation
theory (MP2) for geometry predictions on 80 hafnocenes in
the Cambridge structural database. They found MP2 is most
accurate, followed by HartreeFock theory, with B3LYP least
accurate. This somewhat surprising result was attributed to
low multi-reference character in the molecules studied.
Jensen et al.408 tested the performance of BP86, BLYP,
PBE, B3LYP, and PBE0 for 62 3d-series coordinatively
unsaturated diatomics with bonds to H, C, N, O, F, S, Cl,
and Br. Scalar relativistic eects were not included.
Root-mean-square errors (RMSEs) in bond energies range
from 13 kcal mol1 (B3LYP and PBE0) to 17 kcal mol1
(PBE). The poor performance was attributed mainly to
inaccurate s to d promotion energies. The study also included
dipole moments, and it was found that hybrid functionals lead
to more ionic (less covalent) bonding character than local
functionals; on average a PBE0 dipole moment is 0.7 D larger
than a BLYP or PBE dipole moment. The spin densities on the
metal were found to be widely dierent in several cases, with
FeC the most extreme. BP86 and PBE were found to give the
most accurate bond distances, with MUEs of B0.02 A.
B3LYP gave the most accurate ionization potentials
(RMSE = 2 kcal mol1) and BP86 and PBE0 the least
accurate (RMSE = 5.5 kcal mol1).
Goel and Masunov409 tested the BLYP, TPSS, B3LYP,
and BMK functionals for hydrides and hydride cations of
the 3d-series. The authors took special care with the SCF
process to nd broken-symmetry solutions that provide
continuous potential energy curves. BMK gave the most
accurate bond energies and TPSS the least accurate. Scalar
relativistic eects were included and led to more accurate
bond lengths. They found MUEs in ionization potentials of
3 kcal mol1 for BMK and 6 kcal mol1 for TPSS.
Legge et al.410 tested the LSDA, BLYP, BPW91, B3PW91,
and B3LYP density functionals against dissociation energies,
geometries, and vibrational frequencies of 15 diatomic
molecules containing Cu, Ag, and Au. Excluding LSDA, they
got the most accurate results with BPW91 and the least
accurate with B3LYP. They also tested vibrational frequencies
for larger molecules.
Wang and Li411 tested bond energies, geometries, and
vibrational frequencies of 17 functionals against 6 diatomics
containing Ag and 2 containing Au. Mean errors for the
various functionals are not tabulated or discussed.
Schultz et al.93 created databases of 9 bond energies and
8 bond lengths for transition metal dimers that contain only
the data for which the experiments are judged to be especially
reliable. The transition elements represented are Zr, V, Cr,
Mo, Ni, Cu, and Ag. They used these databases to test
42 dierent density functionals, including two LSDA
functionals, 12 GGAs, 13 hybrid GGAs, 7 meta GGAs, and
8 hybrid meta GGAs. The dierences between double-z
and triple-z basis sets are often quite large (greater than
10 kcal mol1), so double-z basis sets are not recommended
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

for transition metals. The bond energy MUEs (in kcal mol1)
averaged across these classes of functionals are 7 for GGAs,
8 for meta GGAs, 19 for hybrid GGAs, 20 for hybrid meta
GGAs, and 30 for LSDAs. The most accurate functional
in each of these classes (and its MUE in kcal mol1) is
respectively G96LYP and BLYP (5), TPSSKCIS (6),
B97-1 (5), TPSSh (11), and SPWL (28), and the least accurate
is respectively HCTH (12), mPWB95 (13), MPW1K (32),
BB1K (28), and SVWN3 (33). The errors are much smaller
if one limits attention to dimers with small amounts of
multi-reference character, although even here the trends are
dierent from main-group chemistry. This study shows very
clearly that the methods that perform well for transition metal
bonds are not the same ones that perform best for main-group
chemistry. This study identied a key anticorrelation: none of
the methods (e.g., MPW1K or BB1K) that did well for barrier
heights also did well for transition metal bond energies.
This was shortly before the design of the M05 functional,
which would be the rst to overcome this limitation.
In a second paper, Schultz et al.,94 again following the
guideline of using only the most reliable data for validation,
made databases of 21 metalligand bond energies and 13
metalligand bond distances. Of the 21 and 13 molecules in
these sets, respectively 17 and 9 contain transition metals.
The transition metals represented are V, Cr, Mn, Fe, Co, Rh,
Ni, Cu, and Ag. A total of 57 density functionals were tested.
A composite MUE was created based on main group atomization
energies per bond, transition metal dimer bond energies,
metalligand bond energies, atomic ionization energies
(including metals), transition metal dimer bond lengths, and
metalligand bond lengths. Based on this composite MUE the
most accurate density functionals are, in order, G96LYP
(a GGA), MPWLYP1M (a hybrid GGA), and XLYP, BLYP,
MoHLYP and mPWLYP (four more GGAs). These tests
preceded the creation of several more successful modern
functionals such as B97-3, PW6B95, M05, M06-L, and M06.
In the process of analyzing the data in the two papers just
discussed,93,94 Schultz et al. created smaller representative
databases, the rst containing a subset of four transition-metal
dimers and the second containing four metalligand
complexes. Mean errors computed with these small databases
reproduce the mean errors of the full databases remarkably
well and allow one to gauge the accuracy of new density
functionals very conveniently by comparing the results on
the small databases to those for the very large number of
functionals already tested against these databases.
The tests conducted with the databases of Furche
and Perdew and of Schultz et al. are overall reasonably
consistent.38,9395,406,412 They have been reviewed in detail by
Harvey413 in an excellent review that provides very useful
general comments about testing DFT.
de Jong and Bickelhaupt414 tested 26 density functionals for
reaction energies and saddle point energies of oxidative
addition of the CCl bond of CH3Cl to Pd atom. The best
performing functionals, excluding HCTH, and their mean
unsigned errors in kcal mol1 were found to be X3LYP
(1.4), RPBE (1.7), B3LYP (1.8), RevPBE (2.0), and BLYP
(2.9). Three dierent parametrizations of the HCTH
functional were tested with mean unsigned errors of 0.8, 1.0,
This journal is


c

the Owner Societies 2009

and 4.2. Quintal et al.415 tested 25 functionals for Pd oxidative

addition reactions, the mechanism of the Heck reaction (which
involves Pd in the 0, II, and IV oxidation states), the reduction
of acetone by hydrogenation with a Rh complex, and
other data. They found the best performance by hybrid meta
functionals in particular B1B95 and PW6B95, followed by the
hybrid PBE0, the hybrid meta TPSS25B95 (which is TPSS plus
B95 with X = 25), and the hybrid B97-1 and B97-2.
Li et al.416 tested 27 functionals (unfortunately, these were
mainly old ones25 from 2001 or earlier and one each from
2003 (TPSS) and 2004 (BMK)) against benchmark-quality
WFT data for atomization energies and clustering energies
of (MO2)n (M = Ti, Zr, and Hf) and (MO3)n (M = Cr,
Mo, and W) transition metal oxide clusters with n = 14.
Among the functionals tested, PBE0 performed best. In
another recent study that tested only older density functionals,
Mayhall et al.417 tested PW91, B3LYP, and PBE for heats of
formation of a test set of twenty small molecules (eighteen of
which have 24 atoms) containing 3d transition metals and
found mean unsigned errors of 522 kcal mol1, with B3LYP
being the best.
Schultz et al.326 obtained benchmark values and studied
bond energies, geometries, and dipole moments of PdCO and
Pd2CO with 27 density functionals. They dened a mean
percent unsigned error in three bond energies, two dipole
moments, and ve bond lengths. Both triple-z and quadruple-z
basis sets were employed. The most accurate results were
obtained with O3LYP (2%) and OLYP and PW6B95
(both 3%), followed by B97-1, B98, and MPW1K (all 4%).
The least accurate methods were BMK (18%) and PBE, TPSS,
BP86, and mPWPW (all 14%). This paper was published too
soon to test M05, M06, or M06-L. Ge et al.418 used PW91 to
study CO adsorption on Pt and Au nanoparticles. (PW91 was
not included in the study of Schultz et al.326 because it is
usually less accurate than its successors, PBE and mPWPW.)
A similar strategy for testing DFT by comparing to coupled
cluster calculations for MCO and M2CO has been applied by
Schwerdtfeger et al.,419 who applied it for M = Au as
compared to M = Pd here; their work is discussed in
section 5.3.1.9.
de Jong et al.420 tested 24 density functionals against a
CCSD(T) benchmark for the energies along the reaction path
for the oxidative addition reaction of Pd atom with methane,
in particular, the relative energies of the separated reactants,
the reaction complex, the saddle point, and the products. Their
assessment includes the LSDA, 14 GGAs, 2 hybrid GGAs,
and 7 meta functionals, but it did not include any of the six
functionals mentioned above as having performed best in the
later study of Quintal et al.415 for Pd oxidative addition
reactions. This study involved post-SCF calculations with
the BLYP density, and the VS98 and TPSSh functionals
performed best, with MUEs below 2 kcal mol1. The
least accurate functionals and their MUEs in kcal mol1
were LSDA, B16; OLAP3, B10; and PW91, PBE, and
BP86, each B6.
Ikeda et al.421 calculated coupled cluster-quality binding
energies of d6, d8, and d10 transition metals with p-conjugated
systems with up to ten carbon atoms, and they used these
results to test the B3LYP, B3PW91, BLYP, BP86,
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10769

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

BH&HLYP, mPW1PW, and LSDA functionals. They found

that LSDA overestimates the binding energies and the other
functionals signicantly (e.g., 25 kcal mol1) underestimate
them when the p-conjugated system has four or more
carbon atoms.
Li and Dixon422 used experimental electron detachment
energies for MO6 (M = Cr, Mo, and W) and M2O6
(M = Cr and W) to test the predictions of CCSD(T) and
28 functionals. They found that BP86 performed best, with a
maximum error of 0.29 eV and second largest absolute error of
0.07 eV; CCSD(T) had a maximum and second largest
absolute error of 0.31 and 0.12 eV respectively. DFT was
found to be more accurate than CCSD(T) for systems with
large multi-reference character. For heats of formation of
MO6 and M2O6, CCSD(T) was more accurate than DFT,
and among functionals tested (only three were tested for heats
of formation), B3LYP was most accurate, although the largest
absolute error was 44 kcal mol1.
S. Zhao et al.423 applied 23 density functionals to relative
energies of dierent structures, ionization potentials, bond
distances, and vibrational frequencies of neutral and ionic
clusters containing up to four Ag atoms. PBE0 was found to
be the most satisfactory functional. Functionals that tend to
the correct limit for a uniform electron gas (UEG) were judged
to be more satisfactory than those that do not.
Stevens et al.424 applied two LSDAs, seven GGAs, seven
meta GGAs, and four hybrid GGAs to bond distances,
vibrational frequencies, and dipole moments of diatomics of
Cr, Mo, and W with C, N, and O. The BP86 functional was
judged to be the most satisfactory.
Rogal et al.425 compared the LSDA, PBE, and RPBE
functionals against experiment for the heat of formation of
bulk PdO and the binding energies of O2, CO, and CO2 on
Pd(110). The mean unsigned errors in these four quantities are
0.3 eV for RPBE, 0.6 eV for PBE, and 1.9 eV for LSDA, with
9 of the 12 values being overestimates.
Song et al.426 applied one LSDA, three GGAs, and ve
hybrid GGAs to bond energies, bond lengths, and vibrational
frequencies of 20 4d-series neutral and cationic monoxides.
The main conclusions about performance were based on bond
energies of neutrals, and the local BP86, BLYP, and BPW91
functionals were deemed to have best performance.
Sears and Sherrill427,428 assessed the performance of the
BP86, BPW91, and B3LYP functionals for Sc, Ti, V, Nb, Cr,
Mo, and Mn complexes of bis(salicylaldehyde)ethylenediamine.
The accuracy for relative energies was found to be poor, which
was attributed to multi-reference character.
Ghosh et al.429 provided multi-reference WFT calculations
of the low-energy states of FeIII, CoIII, and NiIII diiminato
complexes to serve as standards against which to evaluate the
spin-state predictions of various density functionals. In particular
they tested BP86, BLYP, PW91, OLYP, OPBE, B3LYP, and
B3LYP*. The found that B3LYP performed best in these cases
but cautioned that previous work by Ghosh and Taylor430
showed opposite trends in functional performance for some
bis(salicylaldehyde)ethylenediamine compounds.
In the rst full benchmark study published after the hybrid
meta M05 functional34 became available, Zhao et al.233 tested
14 functionals against 234 data in several databases including
10770 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

the two representative metal databases of Schultz et al.93,94

For the latter, they found the best performance by BLYP
and M05 (MUE = 6 kcal mol1), and the least accurate
performance by BMK (MUE = 27 kcal mol1). M05 was
found to have the advantage of working well both for systems
with low multi-reference character and systems with high
multi-reference character and to give much more accurate
barrier heights, noncovalent interaction energies, and
main-group bond energies than any previous functional with
relatively good performance for transition metals. Zhao et al.
concluded that designing the dependence of the exchange
correlation functional to take full advantage of the
dependence on kinetic energy density and to be consistent
with a reasonably high percentage of HartreeFock exchange
was the key to this breakthrough.
Sousa et al.220 reviewed the status of DFT and assessed the
performance of many functionals for a variety of chemical
properties. The review is very thorough up to mid 2006.
Zhao and Truhlar38 carried out additional validation
studies when the local meta M06-L functional became
available. Nine GGAs, three meta GGAs, two hybrid GGAs,
and one hybrid meta GGA were tested against 22 energetic
databases plus main-group and metalligand bond lengths and
main-group vibrational frequencies. Three of the energetic
databases (all discussed above: two of Schultz et al.93,94 and
one of Furche and Perdew95) contain compounds containing
transition metals. The composite MUE for these three metal
databases ranged from 6 kcal mol1 (M06-L) to 14 kcal mol1
(t-HCTH). The best performances following M06-L were:
7 kcal mol1 for M05 and 8 kcal mol1 for G96LYP,
OLYP, and TPSS. BLYP and B3LYP had MUEs of 8.5 and
l2 kcal mol1, respectively. A broader composite including
main-group thermochemistry (atomization energies per bond
and bond energies, ionization potentials, electron anities,
proton anities, and p isomerization energies), barrier
heights, noncovalent interactions, the metal databases
mentioned above, and metal atom excitation energies yielded
the following MUEs in kcal mol1: 3 for M06-L and M05,
5 for B3LYP, 5.5 for TPSSh, 6 for eight functionals, 7 for two
others, and 7.5 for G96LYP. M06-L and M05 can be strongly
recommended.
Yet another extensive evaluation was carried out when M06
became available.39 Two GGAs, ve hybrid GGAs, two meta
GGAs, and seven hybrid meta GGAs were tested against
29 energetic databases, three bond length databases, one
vibrational frequency database, and one zero point energy
database (496 data). The M06 functional was recommended as
the best for organometallic and inorganometallic chemistry.
It also predicts accurate noncovalent interactions. A review is
available.412
Amin and Truhlar235 tested 39 density functionals against
12 bond energies, 10 bond lengths, and 8 dipole moments in
Zn coordination compounds with H, NH3, O, OH, H2O, S,
and SCH3 ligands. It was found to be important to include
scalar relativistic eects. The three best performing functionals
overall were found to be M05-2X, PW6B95, and B97-2, in that
order. A follow-up study431 involved a larger dataset that also
included binding energies and dipole moments of Zn centers in
coordination environments taken from metalloenzymes, for
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

which bond lengths can be quite dierent than in small

compounds. M05-2X continued to perform well, and B3LYP
was found to have some large errors.
Paier et al.432 published a study entitled Why does the
B3LYP hybrid functional fail for metals? This study contains
tests of the PBE, B3PW91, B3LYP, PBE0, and HSE
functionals for lattice constants, bulk moduli, and cohesive
energies of various metallic and nonmetallic systems. Transition
metals are represented by bulk Rh, Pd, Cu, and Ag, and some
aspects of the conclusions apply to all metals. In particular
they conclude that the B3LYP functional fails to describe the
transition from localized electrons (atoms) to delocalized
electrons (metals). This failure is ascribed about 2/3 to the
LYP correlation functional and about 1/3 to the B88 exchange
functional. Behavior for extended systems with large gaps
(e.g., LiF, NaF, MgO, and diamond) was quite dierent from
that for bulk metals. The authors concluded, as had many
(but not all) workers before them, that it is important to
incorporate the UEG limit.
Stroppa and Kresse433 followed up by testing the accuracy
of the BLYP, PBE, RPBE, B3LYP, HSE, AM05, and PBEsol
functionals for the adsorption site and adsorption energy of
CO on the (111) surfaces of Ru, Os, Rh, Ir, Pd, Pt, and Ag.
They found that BLYP and B3LYP are the overall best
choices for these adsorption properties among the functionals
tested, but they were simultaneously dismayed that these
functionals perform poorly for the properties of the metal
itself, as discussed in the previous paragraph.
When the RPBE functional was originally developed,204 it
was tested for adsorption energies of O, CO, and NO on
Rh(100), Ni(100), Pd(100), and Pd(111). RPBE and revPBE
gave similar results, with noticeable improvement over PBE.
In tests on molecules, Matveev et al.434 found that revPBE and
RPBE were more accurate than BP86, PW91, and PBE both
for atomization energies of small molecules and for dissociation
energies of Cr(CO)6 and Fe(CO)5. Deeth and Fey435 tested the
BLYP, PW91, PBE, revPBE, and RPBE functionals for
geometries and energies of homoleptic FeIIL6 and FeIIIL6
complexes, where L denotes a ligand. When solvation eects
were included, the best performance was obtained with RPBE.
However, the preference for low-spin states in iron complexes
was overestimated by B13 kcal mol1. More recently,
Mattsson et al.219 tested the LSDA, BLYP, PBE, RPBE,
PBE0, AM05, and HSE functionals for lattice constants and
bulk moduli of 23 solids, seven of which (W, Rh, Pd, Pt, Cu,
Ag, and Au) were transition metals. AM05, PBE0, and HSE
outperformed LSDA and PBE, which in turn outperformed
RPBE and BLYP. These trends (e.g., RPBE better than PBE
for energies but worse than PBE and AM05 for lattice
constants) can seemingly be understood from the principles
that emerged from the benchmarks of Zhao and Truhlar.214
Zhao and Truhlar tested PBE, PBEsol, SOGGA, TPSS, and
M06-L for lattice constants of 21 solids, including Rh, Pd, Cu,
Ag, and PbTiO3, cohesive energies of eight main-group solids,
six representative atomization energies of main-group molecules,
six representative barrier heights, the dissociation energy of
SF6, 20 main-group bond lengths, and the exchange energy
and total energy of the He atom. For three lattice constants
the comparisons also included LSDA, BLYP, PW91, BPW91,
This journal is


c

the Owner Societies 2009

mPWPW, and RPBE. For lattice constants, SOGGA and

M06-L were usually most accurate, followed by PBEsol, with
RPBE and BLYP least accurate. For energies, M06-L was
most accurate followed by RPBE. Again we see RPBE more
accurate than PBE for energies and less accurate for interatomic distances. This, along with several other trends, was
rationalized by examining the second-order term in the
gradient expansion of the exchange functional. The results
are also consistent with the analysis of Perdew et al.,215 and
they extend that analysis. Perdew et al.216 have made the point
that tting to the jellium surface exchangecorrelation energy
is not a possible explanation of the relative performance of
various functionals for the lattice constants of solids, and
Csonka et al.436 have recently tested selected density functionals
for lattice constants, bulk moduli, and cohesive energies of
nonmolecular solids (metals and nonmetals, including Rh, Pd,
Cu, and Ag) and provided further discussion of the tradeos
one must make in GGAs such that it may be impossible to
develop a single GGA that can accurately predict solid-state
cohesive energies, surface energies, lattice constants, and bulk
moduli. Yet another test of functionals for surface properties
has been provided recently by Haas et al.,437 who found the best
performance for lattice constants for 60 solids with PBEsol and
SOGGA, which outperformed LSDA, PBE, TPSS, WC, and
AM05. Unfortunately, neither the Csonka et al. paper436 nor
the Haas et al. paper437 considers the M06-L functional.
Ropo et al.438 tested the LSDA, PBE, LAG, AM05, and
PBEsol functionals against exchange energies of ve rare gas
atoms, atomization energies of six main group diatomic
molecules, and lattice constants and bulk moduli of eight
metals (including Fe, W, Pd, Pt, Cu, and Au), three semiconductors, and NaCl (an insulator). They suggested that
AM05 and PBEsol are superior to PBE and LAG.
Tran et al.439 compared the performance of the LSDA and
the PW91, PBE, and WC GGAs for main-group bond lengths
and noncovalent interactions, for the surface formation
energies and cohesive energies of ten bulk transition metals,
and for the lattice constants and bulk moduli of 76 solids,
including 14 bulk transition metals and 22 binary solids
containing one transition metal and one main-group element.
They found that the WC functional gives the most accurate
lattice constants (MUE of 0.03 A vs. 0.04 for PBE and 0.07 for
LSDA) but that none of the functionals tested was uniformly
better than the others for all properties. Many interesting
systematic trends were observed such as how the errors change
as one proceeds across a transition row.
Rinaldo et al.440 evaluated the errors in B3LYP thermochemical predictions for 56 transition metal atomic energies
and 71 bond energies of transition metal compounds and
found MUEs of 7.7 and 5.3 kcal mol1 respectively. They
developed an empirical scheme to correct these errors.
Buhl et al.441 tested 14 density functionals against 41 transition
metalligand bond distances in 25 5d-series molecules. They
found the best results for PBE0. They also updated previous
tests for the 3d and 4d series, and for the combined test set
comprising all three transition rows they obtained the best results
with PBE0, B3P86, and B3PW91. They found that BLYP,
VSXC, and LSDA cannot be recommended. It would be
interesting to extend these tests to M06-L and B98.
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10771

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Handzlik442 evaluated 22 density functionals against

experimental data and CCSD(T) calculations for the thermochemistry of molybdenum compounds. The best performance
was obtained with TPSS, PW91, PBE, and M05. M05 was the
only functional with nonzero HartreeFock exchange to show
good performance. The M06 and M06-L functionals were not
included in the tests.
Although the study does not include transition metals, the
work of Mosch et al.443 on the interaction of O2 with Al
clusters and Al(111) is very relevant. They found, just as
for nonmetallic reactions,444 that GGAs (for example, they
studied PBE) systematically underestimate barrier heights for
reactions of metal clusters and at metal surfaces.

5. Applications
We now consider applications of particular interest that have
appeared in the recent literature.
5.1

Spin state and magnetic properties

Spin and associated magnetic interactions in transitionmetal-containing systems are often discussed within the context
of considering the molecular system to be composed of one or
more subsystems over which unpaired spin is localized, e.g.,
the various transition-metal atoms in a cluster (for an
interesting discussion of the physical meaning of spin localization
in the context of various electronic structure theories, including
DFT, see Reiher445). It is convenient to use a spin-labeled
model Hamiltonian to describe such localized spins as if
they interact as coupled angular momenta, as proposed by
Heisenberg446 and Dirac447 and later elaborated by Van Vleck448
and Slater.449 This model has been used extensively in
conjunction with DFT to explain magnetic properties in
coordination compounds.450452
If we restrict ourselves here, for simplicity, to two interacting centers A and B, the HeisenbergDirac Hamiltonian
takes the form
H = 2JABSASB

(5)

where SX is the local spin on center X, JAB is the coupling

constant between the two centers, and the dot product
between the spins is considered only for antiparallel (antiferromagnetic, low-spin) or parallel (ferromagnetic, high-spin)
vectors. It is important to keep in mind that eqn (5) is a
shorthand description and should not be taken as an indication
that magnetic interactions are involved; they are not. Rather
the parameter JAB results from exchange integrals involving
electron repulsion. When the S2 values for the pure spin states
of the full system and the two fragments are well dened, the
eigenvalues of the operator SASB for alternative total spin
states may be determined from the relationship
S2 = (SA + SB)2 = SA2 + SB2 + 2SASB

(6)

If there is a single unpaired electron on each center A and B,

the eigenvalues of SA2 and SB2 are each 3/4; therefore hSASBi
is 3/4 for singlet coupling (for which hS2i = 0) and 1/4 for
triplet coupling (for which hS2i = 2). With such denitions of
10772 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

SASB, eigenvalues of the HeisenbergDirac Hamiltonian are

given by
ES = JABS(S + 1)

(7)

where S takes on whole or half-integer values ranging from

|SA  SB| to |SA + SB|, and where SA(SA + 1)  SB(SB + 1)
has been moved into the zero of energy.
For a single unpaired electron on each center A and B, a
linear combination of atomic orbitals (LCAO) wave function
for the high-spin triplet state is typically well represented by a
single determinant and in that respect is well suited for
treatment by DFT within the KohnSham formalism. In
another respect though, it may be less well suited than the
singlet because DFT is a ground-state theory and the triplet
state might not be the ground state. The wave function for the
low-spin singlet state, however, in the limit of weak coupling
requires at least two determinants for a balanced description
and is thus problematic for KohnSham DFT if the goal is to
compute directly the energy dierence between the two states
(the same problem holds for direct calculation of the MS = 0
triplet).453456 This discussion may be generalized to multiple
spins on each center, where the high-spin case (in which S is
equal to the highest possible value of MS) continues to be
single determinantal. As originally discussed by Ziegler et al.457
and Noodleman,458 a non-physical broken-spin-symmetry state
can be useful for predicting the proper state-energy dierence in
such instances. Cohen et al.459 have recently noted that the
broken-symmetry approach provides a formal way to circumvent
static correlation errors in existing functionals by avoiding
fractional spins in, say, stretched two-electron bonds.
The broken-symmetry state is constructed by permitting the
system to localize antiparallel, unpaired spins separately on the
two centers A and B. Such a situation is clearly non-physical,
since, for instance, an antiferromagnetically coupled singlet
would have up density on one center and down density on
the other but a true singlet has zero spin polarization density at
every position in space.452,454 In the weakly interacting limit,
however, this wave function is formally a weighted average
of pure spin states, and Noodleman458 has shown that the
coupling constant JAB may be estimated from it as460,461
HS

JAB 

E  BS E
4jSA jjSB j

where HS and BS denote high-spin and broken-symmetry,

respectively. When the high-spin state has low multi-reference
character, its KohnSham DFT energy is sometimes taken (as a
working hypothesis) as accurate (within the accuracy of the
chosen functional) and the energies of other spin states are then
computed relative to it using eqn (7) with the value of JAB
determined from a broken-symmetry calculation according to
eqn (8).
In the limit of strong coupling, Ruiz et al.462 have proposed
that it is more appropriate to consider the broken-symmetry
state to represent the true low-spin state (e.g., the singlet state
for an even number of electrons). This is equivalent to
modifying eqn (8) as
JAB 

HS
E  BS E
4jSA jjSB j 2jSB j

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

where center B has the smaller of the two spins if the spins on
A and B are not the same. In eect, eqn (9) simply expresses
the energy dierence from a DSCF calculation (that is, the
dierence in energy of separate SCF calculations on each
state) of the two spin states in the J coupling notation of the
HeisenbergDirac formalism.
Alternatively, Yamaguchi and co-workers463,464 have
suggested an approach designed to be valid over the full range
from the weak to strong coupling limits
HS

JAB  HS

E  BS E
hS2 i  BS hS 2 i

10

where the expectation values of S2 are evaluated for high-spin

and broken-symmetry KohnSham determinants, i.e., they are
treated as trial wave functions of this many-electron operator.
Since the KohnSham determinant is not the variational wave
function for the electronic Hamiltonian, use of eqn (10) is
necessarily empirical for DFT, but practical experience
suggests that it can be eective. More recently Shoji et al.465
have generalized the approach underlying eqn (10) in order to
permit its application to systems having more than two
interacting spin sites.
From a fundamental standpoint, there has been substantial
discussion in the literature with respect to the meaning of hS2i
as determined from the KohnSham determinant. Although
the KohnSham determinant has (for the unknown exact
density functional) the same density as the exact wave function
(and therefore provides a convenient parameterization of the
density), its relationship to the actual wave function is not
formally dened in DFT, and there has been lively debate with
respect to the interpretation of the expectation value of the
many-electron operator S2. In 1995, Pople et al.466 argued that
restricted open-shell KohnSham procedures were inappropriate
for DFT as they would not properly predict negative
spin densities in open-shell systems (the same can be said
of restricted open-shell HartreeFock methods), and
suggestedbased on empirical observationthat spin
contamination was not as signicant a problem in DFT as
in wave-function theories (which was rst pointed out by
Baker et al.467 and others64,468471), emphasizing, though, that
one cannot be certain that hS2iKS is necessarily a quantitative
measure of spin contamination (where h. . .iKS denotes an
expectation value over the KohnSham determinant). Wang
et al.472 pointed out that hS2i can be computed from one- and
two-particle density matrix elements, and examined hS2i as a
function of various approximations for the two-particle matrix
elements that are not strictly available from a DFT calculation.
A deeper analysis of the relationship between hS2iKS and spin
contamination was provided by Cremer and co-workers473475
based on analysis of the total and on-top pair densities for a
variety of systems having open-shell biradical character,
following an approach rst suggested by Perdew et al.21
Cremer and co-workers noted that hS2iKS is not necessarily
in good quantitative agreement with hS2i as evaluated by
other means for the KohnSham density, but that hS2iKS is
qualitatively useful in terms of its description of the degree to
which mixing of spin states is present in the KohnSham
determinant. A number of subsequent works have explored
This journal is


c

the Owner Societies 2009

the relationship between the spin properties of the

KohnSham determinant and the true wave function, in
general emphasizing that the former need not satisfy the same
restrictions as the latter.51,476479
Returning to applications, the relative utility of eqn (8)
[or its generalization eqn (10)] vs. eqn (9) for predicting magnetic
coupling has been debated in the literature, but no clear
consensus has emerged because results have been variable
depending on the chemical system studied and the functional
employed. For example, Valero et al.480 recently compared
results from eqn (8) and (9) for a number of functionals,
including those in the M06 family,39 for several binuclear
copper complexes as well as other non-metallic biradicals,
and observed that M06-L, M06, and B3LYP were more
accurate with eqn (8), while M06-2X was more accurate with
eqn (9), and PBE0 showed no preference for eqn (8) vs. eqn (9).
One of the earliest studies employing the broken-symmetry
approach was also to binuclear copper complexes, in particular
2-azido bridged compounds, where Adamo et al. observed
that predicted exchange interactions using the mPW1PW
functional were extremely sensitive to geometry.481 Ruiz and
co-workers482 compared eqn (8) to eqn (9) for a binuclear
copper complex and found that while the latter worked
better with standard functionals, the former was better
when self-interaction error was removed explicitly following
the PerdewZunger approach.112 They concluded that
self-interaction error eectively cancelled the error associated
with the non-dynamical correlation not captured by the singledeterminant KohnSham model, thereby eliminating the need
to employ eqn (8) to correct for the latter error. This interpretation was, however, challenged by Adamo et al.,483 who
asserted that in general spin symmetry must be restored
to broken-symmetry wave functions [e.g., by the mapping
associated with eqn (8)] in order to provide proper estimates
of the energy.
In a dierent study also focusing on binuclear copper
complexes, Lewin et al.484 considered singlettriplet splittings
in various supported dicopperdioxygen complexes in both
their isomeric bis(m-oxo) and m-Z2:Z2-peroxo structural forms.
These species provide a particularly interesting test set because
the instability of their singlet KohnSham orbital representations
to spin-symmetry breaking is highly dependent on the choice
of functional. Thus, with local functionals like BLYP or
mPWPW, restricted closed-shell singlets are predicted to be
stable or metastable in KohnSham calculations, while the
incorporation of HartreeFock exchange tends to lead to
instability to symmetry breaking. In this case, six dierent
functionals (BLYP, mPWPW, TPSS, TPSSh, B3LYP, and
mPW1PW) are in reasonable agreement with one another
for singlettriplet splittings of the m-Z2:Z2-peroxo isomers,
where hS2i values range from 0.00 to 0.95 depending on ligand
and functional, but only when eqn (10) is useduse of eqn (9)
leads to quite poor results for those cases having the greatest
biradicaloid character (the importance of using eqn (10) to
compute the singlet energies was also inferred from analysis
of bis(m-oxo)/m-Z2:Z2-peroxo isomerization energies484486).
Interestingly, in the bis(m-oxo) isomers, all functionals predicted
rather similar singlettriplet splittings when closed-shell
singlets were assumed, even in those cases where a
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10773

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

broken-symmetry instability existed. Similar observation

have been made for a number of monocopper complexes of
dioxygen.487,488 In all of these cases, it appears that substantial
degrees of covalency in CuO interactions, as judged from
explicitly multi-reference calculations,489 can lead to improved
predictions with restricted DFT calculations that would otherwise be inappropriate in more weakly coupled systems. In a
very recent contribution, Rivero et al. also considered
exchange coupling in binuclear copper compounds, but
examined the performance of range-separated hybrid
functionals for this task;490 they concluded that such
functionals might have utility for the study of magnetic
interactions in metal oxides and other challenging solids. A
study by Roth et al.491 on a trinuclear copper complex
supported by aminosugar Schi base ligands found B3LYP*
to give reasonable agreement with experiment for a small
doubletquartet splitting (91 cm1 vs. 24 cm1), with
some of the discrepancy being attributed to delocalization of
the spin on one of the coppers over adjacent coordinating
heteroatoms, thereby limiting the applicability of the
HeisenbergDirac model.
Another set of systems showing great sensitivity of the
ground electronic state and symmetry breaking to the choice
of functional are the intermediate-spin systems MnPc and
FePc molecules, where Pc is phthalocyanine. Marom and
Kronik492 studied these systems with B3LYP, PBE, PBE0,
and M06.
Another study comparing DFT to multi-reference
calculations has been carried out by Batista and Martin,493
who examined the binuclear ruthenium blue dimer
[(bpy)2(OH2)RuORu(OH2)(bpy)2]4+. In two prior publications,
Yang and Baik494,495 had initially reported a singlet ground
state for this species and then a triplet ground state based on
B3LYP calculations and a broken-symmetry formalism.
Batista and Martin employed CASSCF. They had diculty
converging certain of the states reported by Yang and
Baik, but noted that their CASSCF calculations predict a
singlet ground state consistent with magnetic susceptibility
measurements496 and inconsistent with the B3LYP results. As
the active space employed for the CAS wave function included
only 10 electrons in 6 orbitals and dynamical correlation
eects were not subsequently estimated, the CASSCF result
is not quantitatively accurate, but the overestimation of the
triplet stability by B3LYP is noteworthy and additional
studies to assess the inuence of HartreeFock exchange in
the functional would be interesting.
A comparison of DFT to much more highly correlated
multi-reference results has been reported by Pierloot and Vancoillie
for the singletquintet energy splittings in [Fe(L)(NHS4)],
where NHS4 is 2,2 0 -bis(2-mercaptophenylthio)diethylamine
dianion and L = NH3, N2H4, PMe3, CO, and NO+.497
On the basis of comparison to extended-basis-set CASPT2
calculations, these authors noted that DFT functionals
incorporating increased HartreeFock exchange did better in
more ionic systems, but the opposite was true for more
covalent systems (cf. the discussion of copperdioxygen
systems above). They found that OLYP oered the best
quantitative performance for computing the spin-state energy
splittings compared to other functionals that they tested, but
10774 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

noted that this functional did a poor job of predicting

geometries and that its extensibility to other systems was
suspect. Swart reported that the O exchange functional, when
combined with PBE correlation, also gave good results for the
ground spin states and geometries of a number of iron halide
ions and spin crossover compounds.498 Rong et al.499
compared various functionals for state-energy splittings in
supported iron complexes and suggested that the bias towards
low-spin states observed for local as opposed to hybrid
functionals could be attributed to a more localized frontier
orbital that was placed at higher energy in the orbital manifold
for the local functionals than for the hybrids. Carreon-Macedo
and Harvey have emphasized the signicant challenges associated
with accurately computing the very small spin-state splittings
in iron carbonyls, even with very large basis set CCSD(T)
calculations,500 where CCSD denotes coupled cluster theory
(a single-reference WFT approach) with single and double
excitations, and CCSD(T) denotes CCSD with a quasiperturbative treatment of connected triple excitations. We
note that CCSD(T) is usually reasonably reliable for singlereference systems but is of unknown validity for multi-reference
systems, as illustrated by the following example.
The diculty of evaluating the accuracy of DFT for
spin-state orderings is well illustrated by a study of NiO2. In
main-group chemistry, a high-level WFT calculation is often
used to provide a benchmark, and the most widely used
method for this purpose is CCSD(T). Song et al.501 employed
CCSD, conventional CCSD(T), and four versions of locally
renormalized CCSD(T), abbreviated as LR-CCSD(T), to
calculate the singlettriplet splitting (D, dened as energy of
singlet minus energy of triplet) in the cyclic form of NiO2. For
the same problem, they also employed other WFT methods,
namely multi-reference conguration interaction, without
(MRCI) and with (MRCI + Q) a Davidson correction,
CASSCF, and generalized Van Vleck perturbation theory
(GVVPT2, a form of multi-reference perturbation theory), as
well as two GGAs, including PBE, and two hybrid functionals,
PBE0 and TPSSh. Conventional CCSD(T), one of the four
LR-CCSD(T) methods, and GVVPT2 predicted negative D
of 7, 5, and 8 kcal mol1, respectively, whereas CCSD,
the other three LR-CCSD(T) calculations, CASSCF, MRCI,
and all four DFT calculations predicted positive D, with the
DFT predictions in the range 711 kcal mol1. Their
best estimate was that D is positive and equal to about
4 kcal mol1. The MRCI and MRCI + D calculations
predicted +12 and +13, respectively, signicantly higher than
their best estimate and also higher than DFT (of course MRCI
would become exact if enough congurations were included,
but this is impractical even for this triatomic system).
Although the argument for what is the best value of D is not
completely convincing, this study shows the danger and
uncertainties associated with using either conventional
CCSD(T) or MRCI as a benchmark for transition-metal
systems. We also note the work of Neese examining the triplet
and quintet states of [FeO(NH3)5]2+ and comparing DFT to
multi-reference WFT.502
A thorough comparison of methods including B3LYP was
undertaken by Matxain et al.503 for the scandium dimer. Based
on many prior DFT calculations and an unclear experimental
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

situation, the ground state of Sc2 came to be accepted as 5Su,

and indeed at the B3LYP level this state is predicted to be
lowest by 0.08 eV. However, large CASPT2 calculations and
diusion quantum Monte Carlo calculations both predict that
the 3Su state is 0.16 eV below the 5Su state and that the other
properties of the 3Su state, namely the vibrational frequency
and the dissociation energy, are in equally good agreement
with experiment as those predicted for the 5Su state. Sorkin
et al.479 also considered state-energy splittings in diatomic
cases, in particular Fe2, Fe2, and FeO+. They highlighted
the accuracy of predicted state-energy splittings from Kulik
et al.138 employing the PBE + U approach, although they
noted that the U potentials employed in such calculations tend
to be system-specic, and found that within the uncertainty in
experimental values, B3LYP, M05, and M06 also gave reasonable
results, but without system-dependent parameters. Sorkin
et al. interpreted all of their broken-symmetry results as
corresponding to S = MS, i.e., without resorting to spin
purication, and additionally noted the importance of permitting
spatial symmetry breaking. In particular, in the case of FeO+,
they observed a monotonic decrease in computed electronic
energies for the 4F state upon reducing the imposed symmetry
from C6v to C4v to C2v, and the predicted energy relative to
the 6S+ ground state was found to improve with reduced
symmetry.
Returning to spin-puried calculations, a particularly
interesting variation on the use of the broken-symmetry
approach to compute state-energy splitting has been reported
by Herrmann et al.504 for a study of the spin-state energetics
ranging from S = 0 to S = 4 in the [Fe2O2] core of methane
monooxygenase. In this case, the use of eqn (10) was
compared to an alternative505
HS

JAB 

E  BS E
2HS hSA  SB i  BS hSA  SB i

11

where the expectation value of the SASB operator is computed

using local projector techniques as described either by
Clark and Davidson506 or by Mayer.507 The primarily
methodological study of Herrmann et al. indicated substantial
quantitative dependence of predicted J values on the amount
of HartreeFock exchange in the functional employed
(a similar sensitivity for a dierent bridged diiron core was
observed by Hubner et al.508), and found that predictions from
eqn (11) using Clark and Davidsons SASB operator diered
from eqn (10) by 223 cm1 (about a factor of 2) for a
representative geometry, with eqn (10) being closer to the
experimental value of less than 30 cm1. Podewitz et al.509
later compared results from Clark and Davidsons approach
to that obtained using Mayers alternative for a set of
supported iron and manganese clusters; they concluded that
Mayers method provided more physically meaningful
measures of spin population on individual atoms (i.e., when
evaluating SASB), but the two approaches give nearly
identical results for SASB so that coupling constants
computed using eqn (11) are insensitive to this choice.
Another new development that has begun to see application
is the use of constrained DFT510512 (cDFT) to generate
broken-symmetry KohnSham determinants having their
This journal is


c

the Owner Societies 2009

uncoupled spins localized by the chosen constraints, as

opposed to determined only by the variational energy. As
the energy of the broken-symmetry state does not correspond
to a true spin state, including additional constraints need not
be regarded as unphysical. Thus, Diey et al.513 have used
cDFT to study the singlettriplet splittings of geminate
electron-hole pairs in the charge-transfer congurations of
zincorganic light emitting devices, and Rudra et al.514 have
adopted the approach to study spin-frustrated iron and
chromium clusters. In the latter case, Rudra et al. emphasize
that optimal results can be obtained by associating each
constrained KohnSham determinant and its energy with an
individual spin-coupling scheme in the HeisenbergDirac
Hamiltonian. Using these energies the relevant coupling
constants in multisite magnetic systems can be easily
computed from the HeisenbergDirac Hamiltonian itself
rather than by invoking a broken-symmetry projection
scheme. Schmidt et al.515 have recently described an extension
of the cDFT approach that in addition to constraining spin
localization constrains the magnitude of spin contamination
computed from the KohnSham orbitals. Schmidt et al.
observe that the additional inclusion of the spin-contamination
restraint is essential for the accurate computation of hyperne
couplings in doublet [Mn(CN)5NO]2. In other work focused
on multisite systems, Fliegl compared broken-symmetry
approaches to CASCI models in order to examine magnetic
couplings in a structural model for the tetramanganesecontaining oxygen evolving complex in photosystem II
and noted that BS DFT approaches tended to overestimate
ferromagnetic couplings between centers.516
In the context of solid-state calculations, Cioni et al. have
observed that t-dependent local functionals provide good
accuracy for the prediction of exchange coupling constants
in KNiF3 and K2NiF4 insulators by broken-symmetry
approaches.517 Avoiding hybrid functionals and the costly
computation of their HF exchange component speeds the
solid-state calculation.
The use of broken-symmetry approaches in computing
magnetic exchange is motivated by the poor performance of
standard KohnSham DFT for states that are not well
represented by single determinants, like singlet-coupled
diradicals. An alternative approach explored for cases of two
unpaired electrons is to abandon the single determinantal
formalism of the KohnSham model and adopt the restricted
ensemble-referenced KohnSham (REKS) approach of
Filatov and Shaik.518520 In this case the singlet density is
constructed from two closed-shell KohnSham determinants
diering from one another by a single orbital, i.e.,
rREKS r 2

N=21
X

jfi rj2 nr jfi rj2 ns jfi rj2

12

i1

where N is the total number of electrons and nr and ns are

fractional occupation numbers (that sum to two) for the two
non-common orbitals. The REKS energy is then taken to be
EREKS = 12[nrEKS(. . .f2r ) + nsEKS(. . .f2s )]
 s)
+f(nr,ns){EKS(. . .frfs)  12[EKS(. . .frf
 rfs)]}
+ EKS(. . .f

(13)

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10775

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

where f is a factor whose value may be chosen by analogy to

CASSCF calculations or simply empirically. Comparison of
REKS singlet energies to restricted-open-shell KohnSham
(ROKS) triplet energies then oers an alternative means to
compute 2J values in systems with 2 unpaired spins.
Moreira et al.521 compared the REKS approach with dierent
factors f to use of eqn (8) for 6 binuclear copper complexes
having J values ranging from 382 to 111 cm1. They found
that with eqn (8) the predicted J values correlated reasonably
well with experiment, but the slopes of the correlations were
4.8, 1.6, and 0.5 for the BLYP, B3LYP, and BH&HLYP
functionals, respectively [note that the slopes in this case
would be one-half of the values with eqn (8) if eqn (9) were
used instead]. Thus, there is considerable sensitivity in the
approach to the amount of HartreeFock exchange included
in the functional. With the REKS approach based on eqn (13),
Moreira et al.521 observed similar correlations and slopes of
the correlations to those obtained with eqn (8) for B3LYP
and BH&HLYP. The concluded that eqn (8) is the more
appropriate one for use in these systems when prediction from
the broken-symmetry SCF solution is undertaken (cf. the
contrasting conclusions of Valero et al.480 above with M06
results) and that improved quantitative performance of the
REKS formalism will be contingent on improved understanding
of how to avoid double counting the electron correlation
energy by designing functionals specically for use in the
REKS approach, a point which Perez-Jimenez et al.522 have
explored in more detail for various antiferromagnetic solids
and biradicals. Ukai et al.523 have also explored the application
of more general CAS-DFT approaches for computations of
magnetic properties for the complexes V(H2O)63+, V(CO)63+,
and [Cu(H2O)]2(m-CH3CO2)4.
Other alternatives to the broken-symmetry approach
include the use of so-called spin-ip DFT, where the antiferromagnetically coupled low-spin state can be rigorously
described as an excited state of a single-determinantal high-spin
reference.524 In a recent study of copperdioxygen complexes,
de la Lande et al.525 observed good agreement between
spin-ip singlettriplet splittings and either experiment or
multi-reference WFT predictions; they further emphasized that
the spin-ip formalism permits optimization of the molecular
geometry without the spin-state ambiguities associated with
broken-symmetry calculations and also that analysis of the
spin-ip excitations can be useful for identifying suitable active
spaces for subsequent multi-reference WFT calculations. Rhee
and Head-Gordon have also successfully employed spin-ip
DFT to characterize molecules containing oxidized CuPCuP
rings as singlet diradicals.526
A diculty with spin-ip DFT as originally introduced is
that, with collinear density functionals (which are used in the
overwhelming majority of applications), the spin ip can only
be introduced by HartreeFock exchange, and thus the theory
can be used only with hybrid functionals. In local collinear
functionals, the functional depends on the spin-up and
spin-down densities, the magnitudes of their gradients, and,
in meta functionals, the spin-kinetic energies or the Laplacians
of the spin densities. However, as mentioned in section 3.2,
noncollinear functionals also depend on the o-diagonal elements
of the spin density matrix, and they induce spin-ip transitions
10776 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

even in the absence of HartreeFock exchange;527 the resulting

noncollinear spin-ip TD-DFT527530 is more satisfactory; so
far, however, applications to systems containing transition
metals are in their infancy.531
Another recently described alternative to broken-symmetry
DFT involves the use of perturbation theory and the second
derivative of the KohnSham energy for a single, chosen
reference state.532 And, in somewhat older work,
Cramer et al.64,471 considered correcting broken-symmetry
DFT energies based on application of wave function
spin-projection operators to the KohnSham Slater determinant.
It will be interesting to observe how these various approaches
perform as additional applications are reported for transitionmetal containing systems.
We will not review here in detail recent developments or
applications associated with the use of DFT to compute EPR
and NMR spectral parameters for open-shell systems containing
transition metals. Rather, we refer readers to the recent review
of Neese for a thorough coverage of this subject,452 and we
note briey a few interesting reports. Teale et al.24 have
recently compared NMR chemical shifts computed from
GGA and hybrid functionals using the optimized eective
potential (OEP) formalism. Alam et al.533 predicted 19F
chemical shifts in uoropolyoxotantalates and in particular
the inuence of TaF spinorbit coupling on those shifts, and
Autschbach and Zheng534 have used two-component relativistic
DFT to rationalize the inuence of nonbonding Pt 5d orbitals
on Pt chemical shifts in supported coordination compounds
having oxidation states from II to IV. Morbec and Capelle535
have recently emphasized the connections between spinorbit
polarization terms in spin DFT and the exchange correlation
vector potential that arises in current-density functional
theory, which may serve as an interesting alternative for the
computation of magnetic spectral quantities. Hrobarik
et al.536 calculated EPR parameters for oxo-molybdenum(V)
and oxo-tungsten(V) complexes by two-component relativistic
calculations.
5.2 Electronic and vibrational spectroscopy
5.2.1 Electronic spectroscopy. The primary method for
computing transition energies to electronically excited states
within the framework of DFT is to employ a time-dependent
formalism537,538 (TD-DFT, which was introduced in section 2.3),
and the strengths and limitations of that model have been
amply reviewed elsewhere.159,232,452,539544 With respect to
transition-metal complexes, a particularly troublesome feature
of TD-DFT is the tendency of many functionals to signicantly
underestimate the energies of excited states characterized by high
degrees of non-local charge transfer, although in principle this
problem may be addressed by replacing the DFT exchange
functional by 100% HartreeFock exchange,232,545 by
using certain range-separated functionals,188,190,236 as in a recent
application to charge transfer excitation in pyridine complexes
with Ag20,546 or by going beyond rst-order response.173
Nevertheless, linear-response TD-DFT with various functionals
continues to be an especially useful one-electron model for
understanding the optical spectra of transition-metal complexes.
We highlight a few examples here.
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Salassa et al.547 have used TD-B3LYP to study the singlet

and triplet excited states of RuII polypyridine complexes and
to explain their photodissociation behaviors on the basis of the
excited-state potential energy surfaces (a number of earlier
TD-DFT studies of Ru and Os polypyridyl complexes have
been reported with the goal of rationalizing their absorption
spectra in the gas and condensed phases548554). Jackson
et al.555 have employed TD-DFT to characterize the intense
near UV absorptions in axially ligated FeIVQO complexes as
ligand to FeQO charge transfer transitions and thereby
rationalize the strong resonance enhancement of the FeQO
stretching mode upon excitation of these transitions. Hutin
et al.556 characterized the unusual mixed valence excited state
of a binuclear copper helicate with TD-B3LYP and TD-BP86
and further computed its circular dichroism in order to assign
its absolute chirality which derived from an interesting selfsorting process during crystallization. Later, Schultz et al.557
showed that this mixed valence behavior extends over all
4 copper atoms in a helicate dimer, thereby rationalizing
experimental cyclic voltammetry data and indicating a potential
means to construct a wrapped molecular copper wire. In
addition, Bar-Nahum et al.558 found TD-B98 to be especially
helpful in explaining an unusual long wavelength metal to
ligand charge-transfer transition occurring in a mixed valence
trinuclear copper disulde (Fig. 1).
A recent review provides an overview of DFT and TD-DFT
applications to electronic spectroscopy and excited-state
properties of d6 metal carbonyls, strongly phosphorescent
cyclometallated complexes, RuII photosensitizers (as used,
for example, in solar cells and light switches), and isonitrile
complexes of ReI and RuII.559
Fan et al.560 used spin-unrestricted TD-DFT with the BP86
density functional, a continuum solvation model, and empirical
corrections to calculate circular dichroism spectra of a number
of trigonal dihedral CrIII d3 complexes and compared to earlier
related work.
Perala et al.561 used TD-DFT and magnetically perturbed
TD-DFT with statistical averaging of orbital potentials562 to
calculate ultraviolet and magnetic circular dichroism spectra

Fig. 1 Experimental (light) and computed (TD-B98, 0.2 eV oset,

dark) UV/Vis spectrum for the inset trinuclear copper complex.

This journal is


c

the Owner Societies 2009

of MTAP (M = Ni, Zn) and ZnPc where TAP denotes

tetraazaporphyrin.
Another particularly useful application of TD-DFT is for
the interpretation of the optical spectra of intermediates too
reactive to be readily isolated. Thus, Kunishita et al.563
inferred the creation of a reactive CuII 2-hydroxy-2-hydroperoxypropane intermediate upon addition of hydrogen
peroxide to a solution of a supported CuI complex in acetone
based on a comparison of measured UV spectral data to
those computed at the TD-B98 level, thereby rationalizing
subsequent reactivity of this complex.
In the area of larger inorganic clusters, Stener et al.564 have
successfully reproduced and rationalized the blue shift in the
optical spectra of gold nanoparticles with decreasing cluster
size using scalar relativistic TD-DFT applied to clusters of
146, 44, and 6 gold atoms. Stener et al.565 have also examined
the optical spectroscopy of the WAu12 and MoAu12 clusters at
the TD-DFT level, noting that while spinorbit coupling
merely shifts the energies of the lower excitations in the former
compound, it gives rise to more complex splittings in many of
the excitations for the latter compound. In subsequent work,
Stener et al.566 added to their comparison TD-DFT results for
the anionic dodecahedral clusters VAu12, NbAu12, and
TaAu12.
It is important to note that medium eects on optical
spectra (solvatochromism) can be large for polar transitionmetal compounds, and the inclusion of such eects in the
TD-DFT model is typically most ecient when the surrounding
medium is modeled as a dielectric continuum. Thus, for
example, Charlot and Aukauloo552 have assessed aqueous
solvatochromic eects on the spectra of RuII polypyridine
complexes using the non-equilibrium polarized continuum
model567569 (PCM). De Angelis et al.570 have employed a
similar strategy combined with CarParrinello molecular
dynamics to simulate the optical properties of dye sensitizers
interacting with TiO2 nanoparticles, such systems being
critical components of modern dye-sensitized solar cells.
Lundqvist et al.571 used B3LYP to study the structural and
optical properties of ruthenium polypyridyl complexes on
TiO2 nanoparticles. The spectra550,572,573 and photovoltaic
parameters574 of dyes adsorbed on bulk TiO2 surfaces were
also studied with B3LYP. Abuabara et al.575 studied the
dynamics of interfacial electron transfer at catechol-sensitized
TiO2 with the PW91 functional.
In the cases of smaller inorganic molecules, various benchmark studies of DFT protocols compared to WFT methods
have been undertaken. Ramirez-Solis et al.576 have compared
DFT to CASPT2 and multi-reference averaged coupled pair
functional (ACPF) results for the lowest ligand-eld states of
AgCl2, including consideration of spinorbit eects. Considering
the B97-2, B3LYP, and PBE0 functionals, they found that
DFT predicted the rst excited 2S+
g state to be too high in
energy relative to the 2Pg ground state by about a factor of 2.
Although improved results could be obtained by increasing the
percentage of HartreeFock exchange in B3LYP to 42%, this
led to poor prediction of molecular geometries. In general, the
DFT densities were found to predict excessive delocalization
of unpaired spin density in the various states, this phenomenon
being attributable to self-interaction error.
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10777

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

In the ultra-high energy region of electronic spectroscopy,

DFT has proven to be a popular tool for the interpretation of
X-ray absorption spectra (XAS) in transition-metal containing
systems. In the most straightforward applications, the
only DFT calculation undertaken is for the ground state.
Experimental pre-edge XAS intensities D0 for ligand atoms
may be interpreted according to577
f

D0 core

!f

a2 h
hf jrjf i
3n core

14

where the core electron is typically 1s, a2 is the covalency of

the light atoms valence p orbitals mixed into the acceptor
orbital f* (expressed as a percentage), h is the number of holes
in the acceptor orbital (or orbitals in the event of degeneracy),
n is the number of absorbing atoms, and the transition dipole
moment is estimated based on empirically determined linear
relationships with measured edge XAS intensities into outer
light-atom orbitals. Following experimental near-edge
intensity measurements, eqn (14) is solved for a2 and the
covalencies so determined are compared to those computed
from atomic basis orbital contributions to computed
KohnSham virtual orbitals. Analogous procedures may be
used with metal XAS intensities. In this way, the nature of the
acceptor orbitals and the ordering of the virtual manifold may
in principle be established, oering insights into ligand and
metal oxidation states and redox behavior.
Applying this XAS analysis approach, ligand eects on the
covalency of supported FeIVQO species have been assessed,578
the varying oxidation states of S2 bridges in binuclear
copper complexes have been characterized,579 and the eects
of dierent ligands and oxidation states on the properties
of copper and nickel bis(dithiolene)580 and molybdenum
tris(dithiolene)577 complexes have been rationalized.
In terms of predicting actual near-edge XAS transition
energies, TD-DFT in principle provides a means to predict
such one-electron excitation energies in the same fashion in
which valence to valence excitations are computed. However,
the total number of all possible one-electron excitations is MN
where M is the number of occupied orbitals and N the number
of virtual orbitals, so applying TD-DFT for all one-electron
excitations is cumbersome. Instead, the TD-DFT process is
modied so that only core orbitals are included in the linear
response equations. Thus, the remaining occupied molecular
orbitals do not relax with respect to the core hole, although the
valence orbitals are all responsive since they are included in the
calculation. While this approach tends to lead to large
absolute errors, it has shown good utility in the prediction
of relative transition energies, and has been applied to assess
the covalency of metal-chloride bonds in metallocene dichlorides,
both with581 and without582 accounting for relativistic
spinorbit eects. The inclusion583 of spinorbit eects using
the two-component zeroth-order regular approximation
(ZORA) leads to signicant improvements in predicted
absolute excitation energies for heavy transition metals.
In addition, the TD-DFT for XAS approach has been used
to characterized metal ligand covalency in oxomolybdenum
compounds relevant to the sulte oxidase active site.584
A particularly interesting application of this method to study
benzenedithiolate complexes of Ni, Pd, Pt, Cu, Co, and Au
10778 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

oered insights into those factors aecting the innocence or

non-innocence of the benzodithiolenes.585 Other recent studies
have provided insights into the nature of relativistic and
solvation eects on K-edge XAS spectra for FeII and FeIII
species,586 as well as for superoxidized FeV and FeVI
complexes.587,588 Finally, in something of a DFT tour de
force, Sarangi et al.589 used both ground-state DFT and
valence and near-edge TD-DFT to compare changes in bonding,
UV spectra, and pre-edge XAS spectra between the wild-type
blue copper active site of azurin and a mutant substituting
selenocysteine for cysteine. These authors found that changes
in CuS vs. CuSe bond distances led to very similar
covalencies in the two copperchalcogen bonds, so that
spectra were minimally aected in spite of the 0.2 A variation
in bond lengths.
As an alternative method to computing electronic excitation
energies, Cheng et al.590 and Gilbert et al.591 have both
recently described an approach they call excited-state DFT
(eDFT) or the maximum overlap method (MOM), where in
each instance the KohnSham procedure is modied so that
the total energy is minimized subject to a constraint that the
nal density has maximal overlap with a target density. In this
way densities that correspond to systems with one or more
holes in low energy orbitals may in principle be constructed
and their KohnSham energies determined directly
(as opposed to using TD-DFT). In the case of a system of
non-interacting electrons, selection of a target density is
trivially accomplished by swapping one or more virtual
orbitals with occupied orbitals. By turning on the adiabatic
coupling in small steps, Cheng et al.590 found that they were
able to achieve self-consistent eDFT solutions even for
Rydberg states and Gilbert et al.591 showed the utility of MOM
for charge-transfer excited states. This work is too preliminary to
have been applied to any transition-metal containing systems, but
it has the potential to sidestep known problems with TD-DFT in
such complexes, so we mention it here.
An interesting question in the spectroscopy and optics of
metal clusters is the quenching of spinorbit coupling as a
function of geometry. This was studied using PBE for gold
clusters by Castro et al.592
One key application where it is important to predict the
correct spin-state splitting is multi-state reactivity. Yang
et al.593 used B3LYP calculations to study the gas-phase
spin-forbidden addition reactions of Y+, Zr+, Nb+, Mo+,
Pd+, and Ag+ with N2O to form MONN+, where M denotes
the metal atom; the location of the crossing scheme is sensitive
to the spin-state splitting. Additional examples are found
in the two-state reactivity examples studied by Shaik and
coworkers,594 which are discussed briey in section 5.3.1.6,
and in the two-state tripletsinglet mechanism for olen
epoxidation and oxidation reactions studied by Morokuma
and coworkers.595
5.2.2 Vibrational spectroscopy. The computation of
vibrational frequencies is nearly always undertaken in modern
practice in order to verify the nature of optimized stationary
points. The comparison of computed frequencies to values
measured by IR or Raman spectroscopy is also routine,
and we will not address such studies here, although the
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

performance of various functionals on benchmark data sets is

summarized above in section 4. One study focusing more
specically on an information provided by the vibrational
spectrum has been described by Hebben et al.596 In order to
critically assess a prior suggestion597 that homoaromatic interactions between ethylene ligands causes the geometry of
Ni(C2H4)3 to have all of its heavy atoms in a common plane,
Hebben et al.596 used BP86 calculations to assign in detail
measured IR and Raman spectra for Ni(C2H4)3, Pd(C2H4)3,
and Pt(C2H4)3. From analysis of the interfragment force
constants, they concluded that homoaromaticity plays no role
in the structures of these species: the Ni and Pd cases are well
described as DewarChattDuncanson complexes; the Pt case
includes some metallacyclopropane character. A prior study
noted the high accuracy of hybrid functionals with respect to
predicting the vibrational spectra of Ni(C2H4) and Ni(C2H4)+.598
Vibrational spectroscopy can be useful in assigning chirality
in transition metal complexes, typically from comparison of
experimental and computed vibrational circular dichroism
(VCD) or Raman optical activity (ROA) spectra.452,599601
Some recent examples in this area include prediction of VCD602
and ROA603 spectra for some chiral tris(acetylacetonato)
metal complexes.
Coupling optical and vibrational spectroscopies, resonance
Raman and o-resonance Raman spectroscopies also play key
roles in the characterization of transition-metal containing
systems.604606 Recent advances in the direct computation of
resonance Raman intensities have been reported,452,600,601,607
and applications have included examination of the rst photoexcitation step in a tetranuclear (Ru2Pd2) light-harvesting
complex608 and the determination that benzenedithiolate
ligands need not be innocent in complexes of group 8, 9, and
10 metals.609
Another interesting use of computed vibrational frequencies
is for the construction of molecular partition functions from
which equilibrium or kinetic isotope eects (EIEs and KIEs,
respectively) may be computed so as to gain improved insight
into electronic structure or mechanism. We note in particular
recent work of Popp et al.610 which reported a computed
18
O/16O KIE in good agreement with experiment for the
formation of a PdO2 adduct. Popp et al. found the ratedetermining step to involve an end-on to side-on isomerization
of the O2 unit, which was contrary to an earlier proposal611
that the measured KIE was consistent with a concerted
2-electron oxidation and direct side-on binding of the O2.
Roth and co-workers have also been quite active in comparing
experimental and DFT-derived KIEs in transition-metal
containing systems.611614 In recent work on O2 binding to
Cu614 and also in rationalizing 18O/16O KIEs in horseradish
peroxidase,615 it has been emphasized that EIEs need not
represent upper limits on KIEs, which had hitherto been a
fairly common assumption for O2-binding reactions in the
bioinorganic community.613
In another interesting study making use of the full vibrational
partition function, Brehm et al.616 used the BP86 functional to
compute the dierential vibrational entropic contribution to
temperature-dependent spin crossover in [Fe(pmea)(NCS)2]
(pmea = {bis[(2-pyridyl)methyl]-2-(2-pyridyl)ethylamine}),
providing insights into its low-spin to high-spin transition.
This journal is


c

the Owner Societies 2009

Carbonniere et al.617 studied anharmonic eects on the

vibrational spectra of d0 transition-metal tetroxides and found
B3LYP and PBE0 to oer good comparison with experiment,
while BP86, OLYP, and TPSS did less well.
5.2.3 Nonlinear optical properties. Most applications of
TD-DFT are carried out using the linear-response formalism
in which the applied electromagnetic eld is weak; this is
sucient for single-photon spectroscopy. Nonlinear optics618,619
and two-photon spectroscopy620 require high-order treatments
or a combination of TD-DFT with a sum over states, and such
treatments are beyond the scope of this review.
5.3 Structure, reactivity, and other properties
This section is divided into six subsections, focusing on
coordination complexes/clusters and organometallics, solids,
surfaces, systems having special relevance to heterogeneous
catalysis, those having special relevance to electrocatalysis and
those having special relevance to photocatalysis.
5.3.1 Coordination complexes/clusters and organometallics.
We organize this subsection, which addresses coordination
complexes/clusters and organometallics, by the transitionmetal groups in the periodic table from left to right. In some
cases this organization has the unfortunate consequence that a
given type of problem is discussed in more than one subsection,
but there is no completely satisfactory organizing principle for
such a broad survey. Of course, certain transition metals have
been the subject of considerably more theoretical attention
than others, but we make an eort to identify at least one
interesting application for each element insofar as this is
possible. In many instances, a particular modeling study
may have addressed several dierent transition metals, in
which case we have tended to discuss it within a group that
oers opportunities to compare to other, perhaps more
specically focused work. As there are a vast number of
reports from which to choose, we have inevitably applied
somewhat arbitrary selection principles in determining which
studies to highlight below. Thus, for instance, we do not
discuss full metalloenzyme studies, but we do summarize
several small-molecule model studies relevant to metalloenzyme active sites. We trust that readers will understand that
such choices were essential to maintaining a reasonable size for
the review.
Before starting to work our way across the periodic table,
we mention an especially relevant review by Ziegler and
Autschbach,621 who surveyed a large number of applications
of DFT to inorganic and organometallic reaction mechanisms
and kinetics. We also direct the reader to a somewhat older but
very concentrated thematic issue of Chem. Rev. devoted to
computational transition metal chemistry.622
5.3.1.1 Scandium, yttrium, lanthanum. We have already
discussed in section 5.1 the DFT calculations of Matxain
et al.503 associated with the computational determination of
the ground state for Sc2.
Group 3 metallocenes have been studied extensively as
catalysts for CH bond activation, and olen polymerization,
hydroamination, and hydroalkylation.623632 The primary
focus of these studies has been to elucidate the inuence of
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10779

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Fig. 2 Methane metathesis in a methyl group-3 metallocene.

varying degrees of ligand steric bulk and metal size on

reactivity. In the case of CH bond activation, transition-state
theory rate constants computed for methane metathesis
(Fig. 2) at the mPW1PW level were found to be in good
agreement with experiment as long as important contributions
from tunneling were included.623 The utility of ansa ligands
with respect to increasing the exposure, and thus reactivity, of
the metal center, has been a particular subject of study.627,632
Vastine and Hall633 have examined the transition-state
structures for the methane metathesis shown in Fig. 2, for
the illustrated scandocene, for [Cp*Ir(PMe3)(CH3)]+, and for
some additional systems; using atoms-in-molecules analysis634
of the electron density they conclude that a continuum of
mechanisms ranging from s-bond metathesis (shown in Fig. 2)
to oxidative addition/reductive elimination can be operative in
the overall metathesis process.
Cationic alkyl complexes of Sc or Y have also been found to
be eective polymerization catalysts, and Tredget et al.635 used
DFT to characterize the structural details of various such
species, noting the general presence of stabilizing agostic
interactions.
Scott et al.636 have used DFT to characterize the metalligand
binding in a novel methylidene scandium complex. In
addition, the rst examples of compounds containing a m4
bonded hydride species, the tetranuclear polyhydride clusters
[(C5Me4SiMe3)4M4H8] with M = Y and Lu, were characterized
by Luo et al.637 at the PW91 level, elucidating structural
features obscured by signicant disorder in the Lu crystal
structure.

precursor complex) but also by analysis of the atomic populations,

orbitals, and orbital interactions at stationary points.
The growth of zirconium clusters up to 15 atoms has been
studied by Sheng et al.643 using the PW91 functional. These
authors found that cluster sizes of 7, 13, and 15 were particularly
stable, and that all clusters were particularly active with
respect to the dissociative chemisorption of molecular hydrogen.
Joubert et al.644 have studied the alkane hydrogenolysis
reactivity of ZrH species supported on alumina surfaces at
the PW91 level.
Mandal et al.645 have characterized at the B3LYP level the
olen polymerization activities of two hybrid metallocenenonmetallocene polynuclear complexes, in one case a
binuclear (ZrTi) species and in the other a trinuclear (Zr2Hf)
species. They found that cationic zirconocene sites are more
energetically favorable but that cationic titanium sites are
more sterically accessible so that bimodal activity would be
expected for the polymerization of small vs. sterically hindered
alkenes. Lewin and Cramer646 have described a multilevel
quantum protocol that combines DFT for zirconocenes
with small-basis-set HartreeFock calculations or molecular
mechanics calculations using modied carbon eective core
potentials to account for changes in the electron-donating
strengths of alkyl-substituted metallocenes. Wondimagegn
et al.647,648 have use QM/MM calculations with the BP86
density functional for the QM part to suggest improved
zirconocene catalysts for olen polymerization. Morokuma
has provided an historical overview of DFT studies on the
activation of molecular nitrogen by homogeneous zirconium
catalysts.595

5.3.1.2 Titanium, zirconium, hafnium. In a recent study of

all-inorganic metallocene analogs of the early transition
metals, which have been the subject of considerable prior
theoretical attention,638640 Mercero et al.641 characterized
the electronic structures of the sandwich compounds
(cAl4)2Mn for M = Ti, V, Cr, Nb, Mo, Hf, Ta, and W,
where n is 2, 1, or 0 for metals in groups 4, 5, and 6,
respectively, and cAl4 is the 4-membered ring composed
entirely of aluminium atoms that is triply aromatic as a
dianion. Mercero et al. found that the sandwich compounds
of groups 5 and 6 were stable, maintained strong aromaticity
in the cAl42 rings, and were not prone to autodetachment of
an electron. Introduction of an alkali cation was required to
prevent such autodetachment in the case of the group 4
analogs.
Ochi et al.642 studied CH and NH bond activation at TiIV
imido complexes with B3LYP and WFT. They drew
conclusions not only from energetics (energetic stability of a

5.3.1.3 Vanadium, niobium, tantalum. We noted in section

5.2 the TD-DFT calculations of Stener et al.566 undertaken to
rationalize the optical spectra of MAu12 clusters, where M = V,
Nb, and Ta.
Owing to the importance of vanadium oxide catalysts in a
number of heterogeneous industrial processes, a substantial
amount of work has been done applying DFT to study the
reactivity of small VmOn clusters with various species. Thus, Dong
et al.649,650 have studied the gas-phase oxidation of saturated and
unsaturated hydrocarbons by such clusters at the B3LYP level,
comparing theoretical predictions to product distributions
generated in mass spectrometric experiments. In analogous work,
Ma et al.651 have examined the cyclotrimerization of acetylene
catalyzed by VO2, and Gracia et al.652 have studied the gas-phase
reaction of propyne with VO2+.
The VO2+ cation may be present in aqueous solution during
the hydrothermal synthesis of vanadophosphates. To better
understand its behavior in aqueous solution, Sadoc et al.653

10780 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

have performed CarParrinello simulations using the PBE

functional. They determined that a ve-coordinate VO2+(H2O)3
species was most stable, but that strongly acidic conditions
were required to prevent spontaneous deprotonation of this
compound.
Because of the utility of the 51V nucleus in NMR, various
studies have appeared combining DFT and NMR to gain
insights into geometric and electronic structural details, e.g., in
heptacoordinate
hydroxylamido
vanadium
picolinate
complexes654 and oxoperoxo vanadium complexes of lactic
acid.655
Michelini et al.656 used B3LYP to calculate energies along
the gas-phase reaction paths for Nb+(3P) and Nb+(5D)
reacting with ethane to produce H2, C2H4, H, or C2H6.
Validation of available functionals for the relative energies
of high-spin and low-spin insertion transition states would be
valuable for this kind of calculation.
Arteaga-Muller et al.657 have characterized the electronic
structures of supported niobium and tantalum imido
compounds in order to better rationalize their activities as
catalysts for the polymerization of methyl methacrylate. Based
on orbital analysis they emphasize the isolobal character of the
niobium and tantalum complexes compared to monocyclopentadienyl zirconium complexes.
5.3.1.4 Chromium, molybdenum, tungsten. The series of
supported transition-metal dimers Pn2 M2 where Pn* is
permethylpentalene (C8Me62) and M = V, Cr, Mn, Co, and
Ni has been prepared by Ashley et al.658 and the character of
the metal-metal bonding has been analyzed using the BP86
functional. In particular, a triple bond is observed between the
two V atoms, a double bond between the two Cr atoms, a
single bond between the two Mn atoms, and single bond
character is maintained in the cases of Co and Ni but the
coordination of the Pn* fragments to the metals is increasingly
reduced from Z5 to Z3. For the Cr case, replacing the Pn*
ligands with 1,4-triisopropylpentalene ligands leads to a
supported metal dimer having a ground state best characterized
as two antiferromagnetically coupled S = 12 Cr atoms, but the
triplet state is measured to be higher in energy by only about
0.5 kcal mol1, so that paramagnetic behavior is observed
at room temperature.659 BP86 and B3LYP calculations
incorrectly predict the triplet state to be the ground state by
23 kcal mol1, and CASPT2 calculations of the state-energy
splitting using partially relaxed DFT structures also incorrectly
predict a triplet ground state but by only 0.6 kcal mol1.
The nature of bonding between bare and supported Cr2
dimers has been extensively explored by Gagliardi and
co-workers660665 because of substantial interest in the very
high bond order between the two atoms and the signicant
challenge to many theoretical models posed by the nature of
the bonding.93,658,659,666669 In experimental practice, the very
high bond orders inferred for supported chromium dimers
compared to iron or cobalt analogs are attributable to the
tendency for the latter two metals to preferentially bond to
arenes included in the ligand structure; chromium exhibits a
much smaller tendency to engage in such bonding.664
The activity of Mo and W alkylidene complexes as catalysts
for olen polymerization has been studied by Poater et al.,670
This journal is


c

the Owner Societies 2009

who found at the B3PW91 level that turnover eciency was

related to a balance between the stability of metallacycle
intermediates in the catalytic cycle and energies required to
distort the organometallic from its resting tetrahedral geometry.
Haunschild et al.671 used B3LYP to characterize the many
possible paths for reaction of ethylene with the group 6
alkylidenes MO2(CH2)(CH3) for M = Cr, Mo, and W.
Caramori and Frenking672 have carried out energy
decomposition analysis at the DFT level to characterize the
bonding of Mo to P, N, PO, and NO ligands.
Yan et al.673 employed the BP86 functional and a
continuum model of acetonitrile solvation to study alternating
bond lengths in Mo-containing polyoxometallates, which they
attributed to pseudo JahnTeller distortions involving frontier
p bonding and antibonding orbitals.
Schenk et al.674 have used BP86 (with occasional stateenergy splittings computed using B3LYP*) to study the
nitrogenase activity of the Schrock catalyst MoHIPT
(HIPT = tris[(N-hexaisopropylterphenyl)-2-amidoethyl]amine).
The local nature of the BP86 functional was critical for
eciency with the very large HIPT ligand, which was found
to have a signicant inuence on predicted reaction energetics
compared to prior studies that employed smaller model
ligands.675682 Christian et al.683 came to similar conclusions
with respect to the importance of including the full ligand in
rationalizing dierences between the full HIPT ligand vs.
model ligands in the Schrock and related systems. Schrock684
has recently provided a very interesting comparison of theoretical
predictions and experiment for such molybdenum-based
complexes, noting that theory has proven particularly helpful
in elucidating the role of added acid in the reaction mechanism.
Stephan et al.685 have recently detailed acid eects further in
various steps relevant to catalysis. Finally, Dance686 has
summarized the implications of these and other theoretical
studies with respect to the biological nitrogenase activity of the
FeMo cofactor.
Leyssens et al.687 examined the changes in electron density
(and molecular geometry) associated with one-electron oxidations
of chromium and molybdenum phosphine and amine
complexes, and they noted that the changes are consistent
with signicant metal to phosphorus p back bonding that is
not present in the analogous amine compounds. This analysis
approximated the Fukui function as a full electron nite
dierence calculation.688

5.3.1.5 Manganese, technetium, rhenium. A key feature of

the group 7 series is the ready accessibility of their high-spin d5
states following oxidation by two electrons in the presence of a
suitably weak ligand eld. Matxain et al.689 have used
MPWB1K to study the endohedral Mn@Sn12 molecule and
its dimer in order to evaluate its potential for the construction
of a magnetic material. They predict that the MnII ion inside
the dodecahedral Sn122 cage prefers a high-spin ground state
by more than 1 eV, that it has a large ionization potential,
and that two such endohedral complexes associate, without
reorganization to a more networked solid, with a binding
energy on the order of 0.03 eV. The exchange coupling
(computed using eqn (10) above) of the dimer is computed
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10781

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

to be 5 cm1, i.e., high-spin coupled, suggesting that the

monomer should be an interesting precursor from which to
make potentially magnetic materials. In analogous work,
Matxain et al.690 considered endohedral complexes of the rst
8 third-row transition metals in ZnSn cages for n = 12 and 16.
In all cases, they found the endohedrally trapped species to be
thermodynamically stable, and negligible spin or charge transfer
was observed from the interior metal atom to the surrounding
cage, such that the high-spin state of each atom was preferred,
again suggesting that materials built from such encapsulated
species might have interesting magnetic properties.
Manganese is also of particular interest as the transitionmetal found in the active site of the oxygen evolving complex
(OEC) in photosynthesis. Yamaguchi et al.691 investigated the
electronic structures of manganeseoxo bonds using hybrid
DFT, comparing a high-valent manganeseoxo porphyrin
complex to the active MnQO bonds found in both native
OEC and articial systems. The oxygen atoms in these bonds
were found to be highly electrophilic in nature and also to
exhibit strong biradical character, not unlike the situation
observed for other high-valent metal oxos like those found
in methane monooxygenase (MMO) and cytochrome P450
(see below). Relevant to such systems, Balcells et al.692 have
emphasized that in oxomanganese porphyrins, only high spin
states placing substantial radical character on the O atom
are reactive for CH bond activation, so that trans ligands
stabilizing the singlet ground state relative to the high-spin
states decrease reactivity.
The studies of Ashley et al.658 and Balasz et al.659 on various
bimetallic pentalene sandwich compounds have been highlighted above in section 5.3.1.4. In addition, for the case of
Mn2 bis(1,4-triisopropylpentalene), a particularly interesting
phenomenon is observed: one of the two MnII atoms is lowspin (S = 12) and coordinated Z5, while the other is high-spin
(S = 5/2) and coordinated Z2.693 B3LYP calculations predict
that the state having MS = 3 is 1.1 kcal mol1 lower in
energy than that having MS = 2, consistent with magnetic
susceptibility measurements that suggest an S = 3 state lying
slightly below an S = 2 state. With respect to the binuclear
rhenium, Krapp et al.694 have performed energy decomposition
analysis at the DFT level to quantify the importance of s, p,
and d bonding in Re2Cl82.
Lundberg and Siegbahn57 tested BLYP, HCTH, B3LYP,
B3LYP*, and B98 for OH bond energies and bond lengths in
six aqua- and hydroxyl-manganese complexes. The best functional
was found to be B3LYP with an MUE of 3 kcal mol1,
and the least accurate was BLYP with an MUE of about
20 kcal mol1.
Jia et al.695 applied B3LYP to investigate interconversions
of ve-coordinate monooxo TcV and six-coordinate dioxo
TcVI amine oxime complexes, which involve both addition
reactions and proton transfer with the aid of water molecules.
Solans-Monfort et al.670 used B3PW91 to study the
structural and dynamic properties of rhenium alkylidene
complexes, which are of interest as olen polymerization
catalysts. Haunschild et al.696 used B3LYP to characterize
the many possible paths for reaction of ethylene with the
group 7 alkylidenes MO2(CH2)(CH3) for M = Re, Tc,
and Mn.
10782 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

5.3.1.6 Iron, ruthenium, osmium. Because of their tremendous

importance in biology and their possible extension to bioinspired catalytic systems, the chemistry of heme- and
non-heme-supported ironoxo species has been the subject of
extensive theoretical attention.578,697713 However, given the
number of extant reviews in this area,702,709 we will restrict our
discussion of iron here primarily to polynuclear cases.
Particularly informative studies addressing species containing a
single iron atom, however, do merit some mention. For example,
B3LYP calculations were instrumental in establishing that
bis(a-diimine) complexes of single iron atomshitherto commonly
described as high-spin Fe0 complexesare instead properly
described as antiferromagnetically coupled bis(a-diiminate)FeII species, the non-innocent ligands each having accepted
one electron from iron into a low-energy p orbital.714 In a
detailed study, the ability of various functionals to predict the
binding energy of CO, NO, and O2 to heme-supported iron
was studied by Radon and Pierloot715 using several
functionals and also the CASPT2 model. These authors found
that CASPT2 provided the best agreement with experiment
and that the OLYP functional was similarly accurate (cf. the
good performance of this functional in the prediction of state
energy splittings in supported iron compounds497 in section
5.1 above), but that BP86 predicted severe overbinding and
B3LYP and PBE0 severe underbinding. A prior study by
Strickland and Harvey716 focused on the same binding processes
(and also considered binding of H2O) and emphasized the
contribution of a spin-crossover requirement to the kinetic
barrier associated with CO binding. Mehn et al.717 characterized
the bonding of borohydride ligands to high- and low-spin iron
compounds, nding suciently strong interactions with both
H and B atoms to describe borohydride as an Z4 ligand. Shaik,
Thiel, and their respective co-workers594,718734 have been
especially active in characterizing the single- and two-state
reactivities of various supported iron compounds, particularly
for cases relevant to metalloenzymes like cytochrome P450,
heme oxygenase, horseradish peroxidase (cf. the KIE work of
Roth and Cramer615 noted in section 5.2.2), and nitric oxide
synthase, and have provided additional recent insights into
mechanistic details through DFT calculations, primarily making
use of the B3LYP functional. The most recent example734
involves triplet and quintet reactivity in nonheme oxoiron(IV),
which was modeled in B3LYP calculations.
In the case of polynuclear iron compounds, Reilly et al.735
employed PBE to study the ground state geometries and spin
multiplicities of cationic iron oxide clusters containing one or
two iron atoms and up to ve oxygen atoms. They compared
computed reaction paths for the oxidation of CO with mass
spectrometric data to assess the relative facility of CO oxidation
vs. oxygen atom replacement by CO and found good correlation
between computed barriers and observed reaction channels.
Gold/iron oxide composite nanoparticles have a number of
potential biological applications based on binding magnetic
nanoparticles to various substances, and there are many questions
about the binding strengths. Sun et al.736 showed that the PW91
functionals applied to Au6Fe13O8 clusters can explain the
dierential binding ability of various amino acids to these nanoparticles. They also studied magnetic moments, bond distances,
and gold coating energies in clusters up to size Au50Fe13O8.
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Considering biologically relevant multinuclear iron complexes,

Schwarz et al.737 employed BP86 to correlate predicted
d orbital energy splittings with those determined from magnetic
circular dichroism (MCD) for the two non-equivalent iron
atoms in the active site of toluene-4-monooxygenase (T4MO)
which is strongly similar to that of soluble methane monooxygenase (sMMO). Based on their calculations, which also
address the energetics of dierent active site geometries
formed from the two iron centers and anchored amino acid
side chains acting as ligands, they concluded that a water
molecule found to be axially bound to one of the iron centers
in several sMMO X-ray crystal structures is unlikely to remain
bound when the oxygenases are complexed with their
respective eector proteins, T4moD and MmoB; they further
conclude that complexation by these eector proteins is likely
to change the orientation of a terminal glutamate ligand on
one iron, thereby possibly facilitating formation of a Fe2O2
reactive intermediate and rationalizing the 1000-fold increase
in catalytic activity of sMMO when complexed with MmoB.
In separate work, Rinaldo et al.738 have examined the catalytic
cycle of sMMO by employing a QM/MM method in which a
QM subsystem is treated with the B3LYP functional, and they
suggest that the protein stabilizes a peroxo intermediate
having a m-Z2:Z2 coordination geometry and also adjusts
the overall thermochemistry so as to favor products over
reactants, thus emphasizing the sometime importance of
including the full protein environment in the modeling of
metalloenzymes.
An additional spectroscopic technique that can be useful for
characterizing the electron structure of iron compounds is
Mossbauer spectroscopy, and Han et al. have reported good
agreement between experimental isomer shifts for 61 dierent
iron sites and those computed at the PW91 level after applying
a linear correction.739 Han and Noodleman subsequently
evaluated the performance of a number of other functionals
for computed Mossbauer isomer shifts and found similarly
good performance from OPBE and OLYP.740,741 Based
on computed isomer shifts, geometries, pKa values, spin
populations, predicted ground states, and quadrupole splittings,
Han and Noodleman assign a cis-m-1,2 peroxo bridging structure
to intermediate state P of MMO,740 a bis(m-oxo) structure
to intermediate Q,741 and a cis to trans isomerization
pathway connecting P to Q that is coupled to specic ligand
rearrangements.740
A dierent binuclear iron site of substantial biological
interest is that found in ribonucleotide reductase, where there
has been substantial controversy with respect to the geometric
and electronic structure of intermediate X in the catalytic
cycle,699,704,742 which includes antiferromagnetically coupled
high-spin FeIII and FeIV centers. Mitic et al.743 have compared
(i) d orbital energies computed for various hypothetical
geometries at the BP86 level to splittings determined from
MCD and (ii) predicted transition energies from TD-DFT and
DSCF calculations for the same geometries to electronic
spectroscopic data; based on their calculations, they conclude
that the best model for intermediate X involves one bridging
m-oxo group and one bridging m-hydroxo group. This prediction
stands in contrast to a prior suggestion of a bis(m-oxo) core
based on a similar analysis,744 illustrating the somewhat subtle
This journal is


c

the Owner Societies 2009

spectral dierences predicted for dierent model geometries.

However, Mitic et al.743 emphasize that the use of group
theory to predict spectral intensities can oer helpful
additional information for use in resolving spectral dierences.
Turning to polyiron clusters involving sulde bridges in
place of oxo bridges, which have been extensively studied by
DFT and indeed have served as a testing and validation
ground for broken-symmetry techniques over the
years,305306,310,460,745751 Szilagyi and Winslow478 have
presented an interesting technical study of alternative
approaches to generating broken-symmetry determinants that
have an antiferromagnetic coupling of proper magnitude spin
densities on dierent centers, noting that some electronic
structure programs provide tools for doing this based on
fragment constraints, while others can be induced to converge
to the experimentally supported spin distribution when starting
from fully ionic determinants generated by enforcing block
diagonalization of the KohnSham matrix in a preceding step.
They note that spin densities in a series of Fe2S2, Fe4S4, and
MoFe3S4 clusters are relatively insensitive to choice of correlation
functional combined with B88 exchange, but are improved
relative to local exchange by inclusion of 5% HartreeFock
exchange. Such clusters are also relevant to the chemistry of
nitrogenase enzymes, and Moritz et al.752 have employed
B3LYP* (the asterisk indicates, as explained in section 2.2,
that the amount of exact exchange in the B3LYP functional is
reduced from the usual 20% to 15%) to study the binding of
N2 to Sellmann-type iron(II) complexes, noting that spin-state
changes associated with binding have thermodynamic and
kinetic consequences on this process.
In an eort to understand the widespread distribution of
Fe4S4 clusters in biology, Jensen et al.753 have carried out
BP86 calculations for MFe3S4 and M2Fe2S4 clusters where
M = Cr, Mn, Fe, Co, Ni, Cu, Zn, and Pd. They observe good
agreement for structures and spin states in cases that are
known experimentally, and note that the structures and
reduction potentials of the cubic clusters are surprisingly
insensitive to substitution of other transition metals, suggesting
that this may be an evolutionary advantage to this motif.
Jensen et al. also note that the modied clusters all have
alternative spin states that lie closer in energy to one another
than those for the all-iron case, so that spin crossover may be
more facile in the mixed-metal clusters.
Li et al.754 found that MR-CI calculations agree with
experiment that FeO2 is linear, but nine of ten tested density
functionals predict it to be bent, with only BH&HLYP giving
a linear ground state. However, BH&HLYP predicts qualitatively
incorrect results for other states.
The prediction of one-electron reduction (or oxidation)
potentials for molecules containing transition-metals has been
extensively explored in the recent past,755763 building on the
good success of DFT-based models applied to predict this
property in organic systems.764771 Recent eorts in this area
include a study by Jaque et al.762 examining the aqueous
Ru3+|Ru2+ couple with 37 density functionals and ve basis
sets and further considering the eects of one or two explicit
shells of solvent water (six or 18 water molecules, respectively)
about the bare Ru ion in addition to continuum solvation.
Jaque et al. concluded that at least two shells of explicit water
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10783

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

were necessary to approach quantitative agreement with

experiment and that an important issue is a substantial
degree of positive charge transfer from the metal ion to the
outer-shell waters in microsolvated clusters (cf. the hydrated
gold cation discussed below). For two explicit solvent shells,
the dierence in solvation energy of Ru3+ and Ru2+ varies
from 6.83 to 7.45 eV, depending on the choice of
density functional and basis set, showing that one must be
cautious about qualitative values of calculated reduction
potentials especially in articles where the sensitivity to the
functional is not explored. Kritayakornupong772,773 carried
out QM/MM calculations on both Ru2+ and Ru3+ in
aqueous solution. Although the QM region was treated by
HartreeFock theory (not DFT), the simulations provide
valuable insight into the solvated structures presented in the
solution. For example, for Ru3+ the second coordination
shell contains only ten water molecules.773 Three-body
corrections to a purely MM calculation were insucient to
reproduce the hydration structure of Ru3+.773 The structure
of the hydration shell of Fe3+(aq) was studied by
Amira et al.,774 and the hydrolysis of Fe3+ was studied by
De Abreu et al.775
The relative stability of hydrated oxo vs. hydroxy structures
has been examined by experiments and calculations for
compounds with the formula MO2H2+, where M = Fe, Co,
or Ni. For M = Ni, the (H2O)MO+ structure is favored,
whereas for M = Fe, the M(OH)2+ structure is favored; for
M = Co, the isomers are energetically similar.776
In work focusing on a large number of supported Ru
complexes, interesting because of their anticancer therapeutic
potential, Chiorescu et al.763 benchmarked various continuum
solvation models in conjunction with the B3LYP functional
for 80 reduction potentials in four solvents. Comparison with
experimental data indicated that errors were minimized when
solute cavities were constructed from Bondis set of atomic
radiithe default choices in some electronic structure
programs were found to lead to errors that were larger by
70 to 140 mV.
In work aimed at predicting aqueous one-electron reduction
potentials for the most common aqueous couple of all
of the rst-row transition metals, Moens et al.771,777 found
that the global value of the electrophilicity (a descriptor
from conceptual DFT688) was a good predictor of reduction
potentials when computed for clusters including two solvation
shells about the central metal ion. Uudsemaa and Tamm756
had previously studied this series, computing the reduction
potentials directly from DFT and continuum solvation free
energies, and had come to similar conclusions with respect to
the importance of two solvation shells.
While not computational electrochemistry per se, recent
experimental work measuring one-electron reduction
potentials of aqueous nanodrops in the gas phase778,779 is
worth noting here because of its potential to connect more
directly to the absolute half-reaction reduction potentials that
are typically computed by theory. Continued comparison of
theory to experiment in these instances is likely to be helpful in
resolving ongoing discussion on the absolute potential of the
hydrogen electrode,778,780785 to which the relative electrochemical scale is anchored.
10784 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

We turn next to ruthenium. Grubbs Ru-based catalysts for

olen metathesis are among the most economically important
catalysts invented in the past decade, especially the secondgeneration catalysts with N-heterocyclic carbene ligands like
1,3-dimesityl-4,5-dihydro-2-ylidene (H2IMes), and substantial
DFT work has been done to rationalize the inuence that
various factors may have on the activity of the supported
ruthenium complexes.670,786793 Thus, for instance, Mathew
et al.790 have used DFT to study ve dierent self-deactivation
pathways, all based on intramolecular CH bond activation,
and suggested that increased rigidity in NHC substituents
might be expected to improve catalyst stability. Benitez
et al.791 have used B3LYP and M06 to determine that
throughout the metathesis process the olen remains
bottom-bound to the (H2IMes)(Cl)2Ru catalyst and the
chlorides remain trans to one another, so that eorts to design
diastereo- and enantioselective catalysts should be undertaken
within this coordination environment. In later work Benitez
et al.794 used the M06 functional to calculate the dissociation
energy of the Ru bond to tricyclohexylphosphine (PCy3)
in a Grubbs catalyst and found excellent agreement with
experiment. They found that the M06 functional leads to
improved accuracy over B3LYP (by B0.5 kcal mol1) for
relative energies of various stable intermediates and much
improved accuracy (by B23 kcal mol1) for PCy3, and
they concluded that their calculations settle a longstanding
controversy. In very recent work, Tonner and Frenking792
have used DFT to suggest that carbodiphosphoranes might
prove to be better ligands than NHCs for the design of Grubbs
catalysts.
In a particularly comprehensive study, Zhao and Truhlar793
developed a set of benchmark relative energies for key
stationary points in the Grubbs second-generation olen
metathesis catalytic cycle using a composite model based on
CCSD(T) calculations. To make the calculations of benchmark results feasible, small model ligands were used.
They assessed spin-component-scaled MP2 approaches for
computing these relative energies, where MP2 denotes
second-order WFT perturbation theory, and they also
examined 39 dierent density functionals; they found that
the local M06-L functional combines good accuracy and, by
virtue of being local, excellent eciency with respect to
computations on reasonably large catalysts. A number of
GGA functionals incorporating the LYP correlation
functional were found to perform poorly, in part because
these functionals do a poor job of accounting for mediumrange correlation eects that are decisive for predicting the
importance of dispersion-like interactions between various
ligands in the catalytic system.789 Averaging over the entire
catalytic cycle, the mean unsigned errors in kcal mol1 were
found to be 1.2 for hybrid M06, 2.6 for hybrid MPW1B95,
3.0 for local meta M06-L and hybrid PBE0, 4.14.4 for local
PBE, hybrid M06-2X, and local meta TPSS, 5.7 for hybrid
meta t-HCTHh, 6.6 for the popular local BP86, 7.29.9 for
hybrid B98, X3LYP, and B97-3, and 11.0 for the popular
hybrid B3LYP.
When these methods were applied to real catalysts, with
their large bulky ligands, the dierences between the
predictions of the various functionals were even larger.412,789,793
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Fig. 3 Ball-and-stick depictions of Grubbs I (left) and II (right)

catalysts.

Grubbs second-generation (Grubbs II) catalysts are a hundred

to a thousand times more active than the rst-generation
(Grubbs I) Ru metathesis catalysts, and they have greater
thermal and chemical stability and better functional group
tolerance. However, the 14-electron catalyst is generated more
slowly from the 16-electron precatalyst than was the case for
the rst-generation catalysts. This can be problematic because
living polymerization can be thwarted if the ratio of the rate
constant of chain propagation to that for chain initiation is
too high.795 The dierence between the Grubbs I and Grubbs II
catalysts is the substitution of one tricyclohexylphosphine
(PCy3) in a typical biphosphine Grubbs I precatalyst,
(PCy3)2Cl2RuQCHPh, by an NHC in a typical Grubbs II
precatalyst, (PCy3)(H2IMes)Cl2RuQCHPh. These catalysts
are shown in Fig. 3.
Tsipis et al.796 noted that gas-phase calculations with
B3LYP and BP86 both predict that the PCy3 bond dissociation
energy of the Grubbs I precatalyst is larger than that of the
Grubbs II precatalyst by 1.42.3 kcal mol1, whereas the
experimental result797 is that it is smaller by 3 kcal mol1,
approximately independent of solvent. Zhao and Truhlar789,793
showed that PW91, PBE0, and TPSSh also give the wrong sign
(+0.4 to +1.1 kcal mol1) for the dierence in bond dissociation
energies but that M06-L gives 4 kcal mol1 in good
agreement with the experimental value of 3 kcal mol1.
Zhao and Truhlar then repeated the calculation with the Ru
and Cl atoms removed to nd the contribution of noncovalent
interactions of the bulky ligands. For B3LYP this noncovalent
interaction is repulsive, but for M06-L it is attractive, and in
fact it is 4.5 kcal mol1 stronger for the Grubbs I precatalyst
than for the Grubbs II precatalyst. Thus the dierence in the
bond dissociation energies is essentially completely accounted
for by noncovalent attractive interactions. This is an important
qualitative nding because the catalyst design literature is
almost entirely focused on the electron donating ability of
the ligands and their potential coordinate covalent bonding
strength, with some consideration given to steric crowding,
but there has been essentially no consideration of attractive
noncovalent interactions.
M06-L gives a bond dissociation energy of 3940 kcal mol1
for the Grubbs II precatalyst, whereas MPW1B95 gives
30 kcal mol1, and B3LYP, BP86, TPSS, PBE, TPSSh, and
PBE0 give values in the range 1423 kcal.789,793 The value
inferred from the experiment797 in toluene is 27 kcal mol1.
After the 40 kcal mol1 value was published,789 a gas-phase
experiment was published,798 yielding the same 40 kcal mol1.
This journal is


c

the Owner Societies 2009

This not only conrms the M06-L calculation (and also M06,
which gives very similar results in this case793), but it shows
that the older density functionals give much worse results for
the large, real molecules than for the smaller model catalysts;
this eect of size has also been noted in main-group
chemistry799 and is attributed to an accumulation of
medium-range correlation energy in large, complex systems,
with smaller eects in extended chains.799,800 In more recent
work, Stewart et al.801 reported that B3LYP predicts
geometries for Ru-metathesis-relevant complexes more
accurately than M06-L.
In related work, Pandian et al.802 successfully applied the
M06 functional to the Ru-catalyzed ring closing metathesis of
1,6-heptadiene. Sieert and Buhl803 studied the binding enthalpy
of a triphenylphosphine ligand in Ru(CO)Cl(PPh3)3(CHQCHPh)
with BP86, B3LYP, and the M05 and M06 families, as well as
with density functionals to which an empirical MM dispersion
is added. They found that BP86, B3LYP, and M05 do not
reproduce the experimental results, whereas all four members
of the M06 family and the empirically corrected functionals
do; M05-2X was found to be intermediate. They agreed with
the conclusion of Zhao and Truhlar789,793 that noncovalent
interactions make a very large contribution to the total
binding enthalpy. The success of new density functionals for
these noncovalent interactions allowed them to dene a
computationally ecient protocol for studying catalytic
reactions.
In other work on ruthenium, Caramori and Frenking804
have carried out energy decomposition analysis to characterize
the bonding of NO to ground- and excited-state ruthenium
tetraamine complexes. Krapp et al.805 have used BP86 and
CCSD(T) calculations to clarify details of the electronic
structures of supported iron and ruthenium complexes
coordinating naked carbon, e.g., (CO)2(PMe3)2Ru(C). And,
as noted in section 5.1, Baik and co-workers494,495 have used
B3LYP calculations to assess mechanistic details associated
with oxygen formation in water catalyzed by supported
diruthenium complexes, an area that continues to generate
intense interest in the experimental and theoretical
communities.493,806820 In particularly recent work, Bozoglian
et al.818b have shown that M06-L correctly predicts which of
two alternative pathways for water oxidation is operative
for the Ru-Hbpp catalyst; B3LYP, on the other hand, makes
the wrong prediction.495
Morokuma595 recently reviewed his DFT studies of multistep reactions of organometallic tri-ruthenium complexes.
Elschenbroich et al.821 characterized the electronic
structures of trans-Cl2-(Z1-C5H5P)4M for M = Ru and Os
and trans-Cl2-(Z1-C5H5As)4Ru, the nal case being the rst
arsenine complex of a late transition metal to have been
isolated.
Zhang et al.822,823 employed B3LYP and TD-B3LYP to
study the structures and spectroscopic properties of highly
luminescent OsII complexes.
5.3.1.7 Cobalt, rhodium, iridium. The organometallic
coenzyme B12 and related small-molecule models have been
the subject of many studies designed to probe the mechanistic
details of the CoC bond cleavage reactions critical to the
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10785

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Fig. 4

Reductive dechlorination of trichloroethylene by a cobaloxime.

function of such enzymes as 5-methyltetrahydrofolatehomocysteine methyltransferase and methylmalonyl coenzyme

A mutase.824 The more recent demonstration of the ecacy of
B12 for the reductive dechlorination of halogenated aliphatics
and olens has sparked additional interest in the design of
supported cobalt species for environmental remediation.825
For example, reaction of a pyridine-coordinated cobaloxime
with trichloroethylene leads to a cis-1,2-dichlorovinyl
organocobalt species that may undergo subsequent reductive
CoC bond cleavage (Fig. 4).826 From a mechanistic
standpoint, the nal bond-cleavage event may proceed either
homolytically to generate an organic radical and a CoI species
or heterolytically to generate an organic anion and a CoII
species. In biological systems involving B12, the Co is
supported by a corrin macrocycle and trans ligation is typically
eected by an imidazole or benzimidazole.
Pratt and van der Donk827 carried out extensive studies of
chlorinated vinylcobalamin species at the B3LYP/6-31G(d)
level and observed good correlations between computed and
experimental structural data, noting in particular that the use
of a corrin ring having all of its rim substituents removed
compared to one with 15 methyl groups (to mimic the various
side chains present in biological corrins) led to no substantial
geometric dierences. By computing one-electron reduction
potentials for dierent species and analysis of the virtual
orbitals involved, Pratt and van der Donk827 and subsequently
Birke et al.828 suggested that reductive bond cleavage may take
place via the base-o cobalamin, i.e., after equilibrium
dissociation of the trans imidazole or benzimidazole ligand.
Both Pratt and van der Donk827 and Birke et al.828 also
observed good correlations between one-electron reduction
10786 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

potentials and CoC bond lengths in the pre-reduction

complexes.
Follett et al.826 carried out similar studies at the B3LYP/
6-31+G(d) level for the system illustrated in Fig. 4. They
too analyzed one-electron reduction potentials and, more
interestingly, followed the electronic structure of reduced
cobaloximes along their bond dissociation coordinates.
By examining spin and charge distributions, they were able
to elucidate the degrees to which dierent organic ligands
(methyl, vinyl, and dichlorovinyl) tended to prefer the homolytic
vs. heterolytic pathways and the inuence of solvation upon
those preferences.
Both Pratt and van der Donk827 and Follett et al.826
conrmed observations rst reported by Jensen and Ryde829
with respect to the computed CoC bond dissociation energies
(BDEs; see Fig. 4). In particular, the B3LYP functional
consistently underestimates BDEs in both cobalamin and
cobaloxime systems by 4050 kJ mol1 when the products
include a doublet CoII species, but B3LYP provides more
accurate results when singlet CoI or CoIII products are generated.
By contrast, the local BP86 functional was found to provide
near quantitative agreement with experiment, illustrating the
degree to which inclusion of exact HartreeFock exchange
can aect predicted energy dierences between open- and
closed-shell states. Jensen et al. have discussed this point in
more detail with respect to dissociation energies of diatomic
species formed from one transition metal atom and one
non-metal atom, relating functional performance to the highor low-spin nature of the various asymptotes.408
Considerable modeling attention has also been paid to CoC
bond cleavage in the enzymatic systems 5-methyltetrahydrofolate
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

homocysteine methyltransferase and methylmalonyl coenzyme

A mutase as well. As that work has been very recently
reviewed elsewhere, we refer interested readers there.830
Morschel et al.831 have examined the inuence of
supporting ligands on cobalt-catalyzed DielsAlder reactions.
At the BP86 level, at least 4 dierent pathways are available
leading to products, with low-energy intermediates not
necessarily correlating with low-energy barriers for
subsequent steps; once all paths are considered, the BP86
predictions for regioselection in the reaction between isoprene
and phenylacetylene were found to be in good agreement with
experiment.
There is substantial interest in CO2 as a cheap C1 feedstock,
so its catalytic hydrogenation is an active area of exploration.
Huang et al.832 have used the KMLYP functional to characterize
the mechanism of hydrogenation of CO2 by a pincer
PCPrhodium complex, noting that strongly electron-donating
ligands disfavor reductive elimination of formic acid after hydrogenation (where PCP denotes phenylcyclohexylpiperidine). Hu
and Boyd833 assessed B3LYP for the reaction energy and
forward and reverse barrier heights to dissociate a CO ligand
in Cl2Rh(CO)2. The deviation of these three quantities from
CCSD(T) calculations was only 12 kcal mol1.
In an unusual example of oxygen activation, Praetorius
et al.834 have characterized a singlet square planar rhodium
complex in which the O2 ligand has not oxidized the metal
center. They suggest that prior examples of rhodium-O2
complexes with very short OO bonds that have been assigned
as superoxo species may merit reassignment in light of their
new results.
In the organic area, the mechanistic details associated
with rhodium-catalyzed asymmetric addition of boronic acids
to a,b-unsaturated 2-pyridylsulfones have been studied by
Mauleon et al.835 at the B3LYP level. In another recent report,
Fristrup et al.836 have explored the mechanism for the
rhodium-catalyzed decarbonylation of aldehydes, also at
the B3LYP level. Based on a comparison of computed to
measured H/D KIEs, migratory extrusion of CO was
established as the rate determining step for this process.
Liu et al.837 used B3LYP to study the mechanistic details of
homogeneous rhodium catalyzed [5 + 2] cycloadditions
between vinylcyclopropanes and alkynes. Orian et al.838
used BLYP calculations to study end-on side-on isomerization
of nitrile ligand half-sandwich ve-membered rhodacycles;
they analyzed the saddle points in terms of the strain energy
upon activation.
The utility of IrIII cyclometalated complexes in organic
light-emitting diodes has given rise to many TD-DFT studies
of their electronic spectral properties.839843 More relevant to
catalysis, recent studies have appeared employing DFT (usually
B3LYP) to study the mechanistic details of iridium-catalyzed
allylic etherication,844 alkene hydrosilation,845 and CH
bond activation.846 Ghosh et al.847 successfully rationalized
carboncarbon bond forming reductive elimination rates in
pincer PCPIr complexes having various degrees of steric
bulk, noting the critical importance of non-bonded interactions (as opposed to electronic eects) aecting ligand
rotations required to bring eliminating alkyl fragments into
reactive conformations.
This journal is


c

the Owner Societies 2009

Marom and Kronik848 compared KohnSham orbital

eigenvalues computed for CoPc and NiPc with the B3LYP,
PBE, PBE0, and M06 functionals to experimental ultraviolet
photoemission spectroscopy (UPS) data. They found poor
agreement for PBE and good agreement for the other three
functionals.
5.3.1.8 Nickel, palladium, platinum. Calaminici has
examined the polarizability of small nickel clusters, up to
5 atoms, using a GGA functional.849 She found that in contrast
to clusters of iron or copper, the per-atom polarizability failed to
increase with increasing cluster size, but instead showed more
complex behavior, suggesting that experiments to better
understand this situation would be of interest. Earlier work
by Arvizu and Calaminici850 established the energetics of
various isomers of Nin, Nin+, and Nin, for n up to 5, and
was useful for the development of a Ni basis set optimized for
cluster calculations.370,850 Shoji et al.851,852 have employed the
generalization465 of the Yamaguchi broken-symmetry method
[eqn (10)] with B3LYP to compute exchange interactions in
two supported Ni9 complexes and compared this approach to
the simpler Ising model, observing good agreement with
experiment for predicted ground spin states and magnetic
susceptibilities.
Conradie et al.853 observed that the BLYP, OLYP, OPBE,
and PW91 local functionals and the B3LYP and B3LYP*
hybrid functionals all predict, in gas-phase calculations on the
doublet d7 states of nickel(III) tetra(tert-butyl)porphyrin
dicyanide anion, that the ground state has the unpaired
electron in a dx2 y2 orbital, whereas the experimental ground
state has it in a dz2 orbital. They attribute the error not to
inaccurate functionals but rather to neglect of the counterion
eect.
Kozuch et al.854 studied Pd0L2 and Pd0L2Cl model
catalysts with B3LYP calculations for cross coupling reactions
with phenyl chloride as substrate. They found that the anion
lowers the barrier for the oxidative addition step, and they
explained this in terms of occupied orbital destabilization and
unoccupied orbital stabilization.
de Jong and Bickelhaupt855 used the BLYP functional to
study oxidative addition of CH, CC, CF, and CCl bonds
to model Pd coordination complexes in order to gain insight
into the competing factors of reactant strain, steric shielding,
transition-state stabilization, and anion assistance in controlling
these bond activation processes. They studied these factors not
only at the transition state but also as functions of the reaction
coordinate, and they used the KohnSham orbitals and an
energy decomposition scheme in their interpretation.
The analysis provided considerable insight into the dierences
between CC and CCl bond activations and CH alternatives.
Vines et al.856 examined palladium clusters having nuclearity
ranging from 38 to 225 by adopting a plane-wave approach
where the individual clusters occupy the centers of unit cells
having at least 1 nm of vacuum separating one cluster from
another, thus permitting the eciency of the plane-wave
approach with negligible interactions between clusters and
no symmetry constraints on individual clusters. They
determined that scalability to bulk density-of-states behavior
begins at a cluster size of about 80, with Pd140 and Pd225
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10787

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

particles showing properties very similar to bulk Pd. In a

separate study,857 Vines et al. showed that these larger nanoparticles also showed converged behavior with respect to their
tendency to self-aggregate in linear arrays, as observed in the
formation of particulate Pd nanowires.858
Betz et al.859 have explored the relative stabilities of dierent
N,S vs. N,N, vs. S,S coordination motifs in p-allyl palladium
complexes supported by asymmetric bis-thiazoline ligands and
their inuence on reactivity. Legault et al.860 employed B3LYP
to rationalize regioselectivities in homogeneous palladium
catalyzed sequential cross-couplings of polyhalogenated
heterocycles.
Continuing in the organometallic arena, the last decade has
seen substantial interest in the use of N-heterocyclic
carbenes861,862 (NHCs, already mentioned in section 5.3.1.6)
as ligands owing to their tunable steric demands and their
character as a strong s donor to the metal center. Radius and
Bickelhaupt863 have studied the nature of s vs. p bonding of
NHCs to metals in the group 10 series as a function of other
ligands using the extended transition state density decomposition
scheme of Ziegler and Rauk.864 Radius and Bickelhaupt
emphasize that the importance of p bonding, particularly with
3rd-row transition metals, should not be discounted, and the
directionality of the bonding (in terms of charge transfer) and
its magnitude are quite sensitive to the nature of the remaining
ligands, e.g., with CO ligands being better p acceptors in
this d10 series than NHCs (a similar sensitivity of metalNHC
p bonding to other ligands was noted in recent calculations on
supported f element compounds865,866). Computational
investigations into the nature of NHCs and their bonding to
all transition metals have recently been reviewed by Jacobsen
et al.867 and a review focused exclusively on the Group 10
transition metals has also been provided by Radius and
Bickelhaupt.868 Finally, Tonner et al.869 have compared the
bonding of dierent tautomers of NHCs to phosphine ligands
for complexes of group 4, 6, 8, and 11 transition metals. All of
these studies need to be re-examined in light of the importance
of attractive noncovalent interactions, as discussed in
section 5.3.1.6.
Based on computed structural data and comparison of
computed EPR spectral parameters to experiment, Harmer
et al.870 have proposed a novel nickel hydride intermediate
that plays a key role in reverse methanogenesis in the
enzyme methyl-coenzyme M reductase (MCR). In earlier
work, Yang et al.871 applied a similar protocol to identify a
related methylnickel complex.
de Jong and Bickelhaupt used the BLYP functional to study
carbonhalogen bond activation by oxidative addition at Pd
atoms414,872 and PdCl model catalysts872 and by an SN2
mechanism involving Walden inversion at C. Solvation eects
were found to change the relative favorability of the two
mechanisms. The results were explained by an activation strain
analysis.
The B3LYP functional has been used to study ligand eects
on oxidative addition,873 reductive elimination,874 and
carboncarbon bond formation875 in Pd coordination
complexes.
Adams et al.876 used B3PW91 calculations, either treating a
small model fully quantum mechanically or employing a
10788 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

QM/MM approach with the UFF force eld877 to fully

represent large phosphine ligands, to characterize the ability of
a pentanuclear Re2Pt3 complex to add up to three equivalents
of H2 at room temperature. Because of the clinical importance
of cisplatin as an anticancer drug, several recent DFT studies
have appeared focusing on the structural details of the drug or
derivatives either isolated or bound to DNA, and on the
chemical consequences of binding.878886 For example,
van der Wijst et al.887 used the BP86 functional to study the
inuence of coordination to PtII on the relative energies of
tautomers of 1-methyluracil and 1-methylthymine, nding
that such coordination causes an otherwise rare tautomer to
be favored in solution.
Datta et al.888 used the B3LYP functional to study tunneling
eects in oxidative addition of methane to a Pt complex. As in
an earlier study889 applying B3LYP to oxidative addition of
methane to a model Rh complex, they found very large
tunneling eects.
Gilbert and coworkers890 used the MPW1K functional to
study the intramolecular hydrogen bonds in cis-diamineplatinum(II) pyrophosphate complexes.
Mo and Kaxiras891 used the PBE functional to study
cyanide-transition-metal nanotubes containing Ni, Pd, or Pt.
They stated that their computational parameters ensure high
accuracy, but the accuracy of the PBE functional is not
guaranteed.
5.3.1.9 Copper, silver, gold. Studies of the coinage metals
have focused upon their activities as single-center catalysts in
supported metal complexes as well as upon their structures
and properties in clusters ranging from a few atoms to
nanoparticles; single atoms, ions, and bulk metals have also
been studied. With respect to clusters, Roldan et al.892 have
employed plane-wave LDA and PW91 calculations to study
nanoparticles of copper and silver ranging from 38 to 146 atoms
and nanoparticles of gold ranging from 38 to 225 atoms.
They observed linear convergence to bulk values of predicted
average nearest-neighbor interatomic distances with respect to
average coordination numbers as the sizes of the nanoparticles
increased. However, the reproduction of experimental bulk
values was inconsistent, with PW91 being better than LDA by
5 pm for Cu, LDA better than PW91 by 7 pm for Au, and the
experimental value for Ag falling precisely in between LDA
and PW91, diering from each by 5 pm. Trends in cohesive
energies per atom were similar, with PW91 being clearly better
for Cu, but not for the heavier coinage metals. Analysis of
density of states (DOS) plots for the three families of clusters
indicated that beyond 80 atoms, DOS parameters were
scalable to bulk values, with no special characteristics like
those associated with smaller clusters.
Calaminici et al.893 employed the same density functionals
as Roldan et al. but focused in detail on neutral and anionic
Cu9. The work was motivated by signicant disagreement
between prior theoretical studies894 and experiment895 for
the electron anity of the nonameric cluster. Employing
a cluster geometry search strategy896 at the LDA level
Calaminici et al. found six and nine minima for the neutral
and anionic nonamers, respectively. While LDA energies were
not found to be useful for identication of the lowest energy
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

structures, nor for rationalizing the measured electron anity,

energies computed at the PW91 level (for the LDA structures)
provided good agreement with both vertical and adiabatic
experimental data.
Shi et al.897 studied the mechanism of a gold-catalyzed [1,2]hydrogen shift whose rate is increased by water. A spotlight
discussion is available.898
Because of their importance in heterogeneous catalysis, the
properties of small gold clusters have been studied extensively,
and Pyykko has recently reviewed theoretical eorts in this
area.899,900 An unusual and interesting feature of gold anion
clusters compared to their Cu and Ag congeners is that various
experimental techniques901903 indicate that two-dimensional
structures are preferred over three-dimensional isomers up to a
cluster number of 12. Johansson et al.904 showed very recently
that DFT successfully predicts this 2D to 3D transition size
provided (i) one uses the meta TPSS functional, the hybrid
meta TPSSh functional, or a GGA like PBEsol that is t
(among other criteria) to yield accurate predictions of jellium
surface energies, and (ii) spinorbit contributions to cluster
energies are included. They found by contrast that BP86 and
PBE both are signicantly biased towards 2D clusters. Other
functionals that performed poorly compared to PBEsol and
TPSS include LSDA, BLYP, B3LYP, and X3LYP. It was
shown previously905 that the mean absolute relative error in
the exchangecorrelation surface energy for jellium is much
smaller (1.1%) for the TPSS and TPSSh functionals than for
PBE (5%), and Johansson et al. emphasized this property in
rationalizing why PBEsol, TPSS, and TPSSh perform better
than PBE for this transition. The conclusion about the
importance of tting the jellium surface energy is very
interesting in light of the discussion in section 2.5, and this
issue deserves further study. Olson et al.906 had previously
noted poor performance of older functionals for this same
problem, and had shown that large basis-set MP2 and
CCSD(T) calculations gave results in good agreement with
experiment. Recently it has been shown that the hybrid meta
M05 and M06 functionals and the meta M06-L functional all
predict the 2D to 3D transition correctly in going from Au11
to Au13.907 Ferrighi et al. reported a similar nding, almost
simultaneously.908 Several other authors have used DFT to
study related structural issues in slightly larger gold clusters,
from Au5 to Au20,909 Au14 to Au29,910 Au15 to Au24,911
Au20,912 Au20n+,912 Au20n (with n = 1 or 2),912 Au21 to
Au25,913,914 and Au26.915 For example, Kryachko and
Remacle, in a review,912 noted that Au20 has several planar
and 3D isomers, including a unique tetrahedral structure with
all atoms on the surface and a large HOMOLUMO gap.
Another interesting study of a gold cluster anion has been
reported by Lechtken et al.916 These authors studied Au34
and found that the TPSS functional predicts the lowest-energy
cluster to be chiral and belong to the C3 point group. Lechtken
et al. then employed BP86 structures and TD-DFT with this
functional to compute the electronic excitation spectrum of the
neutral cluster at the anion geometry, and derive therefrom a
predicted photoelectron spectrum (PES). The TD-DFT approach
is more physically appropriate than using a density of states
computation based on orbital energies, and Lechtken et al.
nd that only the chiral isomer predicted at the TPSS level to
This journal is


c

the Owner Societies 2009

be lowest in energy gives a predicted PES consistent with

experiment. This result is particularly intriguing since chiral
clusters might be expected to carry out asymmetric catalysis.
Gu et al.917 also studied Au34. They used the LSDA and
PW91 functionals and found several low-energy structures
with the form of a Au4 core surrounded by 30 uxional
outer atoms.
Li et al.918 carried out calculations on Au20 and Au20 with
the PW91 functional and found that the optimum geometry is
a pyramidal cluster that represents a small piece of bulk gold
with four (111) faces. They found an orbital energy gap of
1.8 eV.918 Further work on the anions Aun with the PBE
functional showed an evolution in shape from planar to at
cages to hollow cages to pyramidal as n increased from 13
to 20,919,920 and a transition from pyramidal to fused planar
to tubular to core/shell compact for n = 2125.914 Zubarev
and Boldyrev analyzed the bonding in the cages with the
B3PW91 functional.921
Torres et al.922 used the PBE functional to study AunO2+
for n = 48 and MAunO2+ for n = 37 where M is a dopant
atom of either Ti or Fe. The preferred geometry for adsorption
of oxygen on the doped clusters is with both O atoms on the
impurity. Prestianni et al. used B3LYP to study various
adsorption sites of O2 or CO on Aum+
with n = 1, 9, and
n
13 and m = 0, 1, and 3,923 later extending this to coadsorption
on Au13+,924 coadsorption on Au13,924,925 and separate and
coadsorption on Au9, Au13, Au9+, and Au13+.925 The latter
study included reaction paths and barriers.
Metiu and coworkers926,927 examined the binding of
propene to gold and mixed gold-silver clusters and showed
that it binds in an electron-donating mode and binds most
strongly at sites where the LUMO protrudes. They concluded
that orbital shape and orbital energy, not just low coordination
number, are important for determining the catalytic properties
of small gold clusters. They also noted927 that there are
numerous low-lying isomers of metal clusters, a subject
mentioned above.912,928,929 They used the local PW91 or
PBE functionals to study the binding of gold clusters on
TiO2 surfaces,930 CO adsorption on gold-doped ceria
surfaces,931 and CO oxidation on TiO2 doped with V, Cr,
Mo, W, and Mn.932 Supported gold atoms and clusters have
also been studied in other works.933,934 See section 5.3.3 for
further discussion of surface science.
Golightly et al.935 used B3LYP to study diphosphineprotected Au11 clusters and stressed the sensitivity of the
results to the precise nature of the ligand when the ligand
has donoracceptor abilities.
A troubling aspect of gold clusters is that even for Au3 the
most complete DFT calculations of the dissociation energy still
show discrepancies from experiment of 510 kcal mol1.936
Ending our discussion of coinage metal clusters, Crawford
et al.937 have reported a particularly interesting study of a
semiconductor cluster Ag28S26(P(O)PhOMe)12(PPh3)12. The
X-ray crystal structure of this cluster exhibits substantial
disorder in its core associated with two silver atoms that can
occupy any of six equivalent sites. DFT calculations at the
BP86 level show that as long as the two sites occupied do not
neighbor one another, essentially equivalent structures are
obtained upon optimization of any given choice. By following
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10789

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

a constrained DFT molecular dynamics trajectory, Crawford

et al. estimated the barrier for the concerted movement of
both silver atoms into new (equivalent) positions to be about
26 kJ mol1. They suggest that this estimate may also be
appropriate for cation mobility in bulk silver chalcogenides.
The largest gold cluster studied so far by DFT is the
thiolate-passivated Au102(p-MBA)44 where p-MBA is
p-mercaptobenzoic acid.373,938
Moving from clusters to supported metal compounds having
only one or two atoms, there has been extensive theoretical
activity devoted to studying the activation of molecular
oxygen by supported copper species,487,939942 owing to the
prevalence of this theme in biological and inorganic
catalysis.487,611,939,940,943946 Guell et al.947 have studied the
interaction of O2 with atomic copper and noted that a broad
range of local, hybrid, and meta-hybrid functionals all predict
the ground-state geometry of the triatomic complex to involve
end-on binding to O2, a result at odds with highly correlated
levels of electronic structure theory (CCSD(T) with a complete
basis set,947 CASPT2,948 etc.) In addition, they found that
B3LYP-like functionals required substantially dierent
amounts of exact exchange to provide satisfactory agreement
with complete basis set CCSD(T) for geometries (20%) vs.
relative electronic state energies (90%), and they attributed
this behavior to the failure of most functionals to properly
balance the covalent vs. ionic character of the dierent
geometries and states, noting that most DFT models signicantly
overestimate the rst ionization potential of Cu.
In most biological and catalytic systems, however, copper is
not found in its atomic form but is rather ionized prior to its
interaction with O2. From a formal oxidation-reduction
standpoint, the combination of O2 with a LCuI partner
(L being a generic supporting ligand environment) may occur
either without change of the copper oxidation state, or with
one or two electrons transferred so as to generate LCuIIO2()
and LCuIIIO2(2) mesomers, respectively. As O2 has a triplet
electronic ground state, weakly coupled LCuIO2 species may
also be expected to be triplets, while singlet and triplet states
may have similar energies for LCuIIO2(), since the d9 CuII

and superoxide radical anion may have separated biradical

character, while nally LCuIIIO2(2) would be expected to
have a singlet ground state given the closed-shell character of
d8 CuIII and the peroxide dianion. Accurately modeling this
spectrum of energetic possibilities is a challenge for
KohnSham DFT because of the varying degrees of dynamical
and non-dynamical electron correlation eects in the dierent
redox states, which can be eectively tuned by choice of ligand
environment L. An additional factor inuencing the electronic
structure and reactivity of LCuO2 complexes is the geometry
of O2 coordination; both end-on and side-on coordination
geometries have been observed depending on L.487
Cramer et al.949 computed the relative energies of the singlet
and triplet states for the end-on and side-on geometries of the
7 LCuO2 complexes in Fig. 5 at several dierent levels of
electronic structure theory. They observed good agreement in
predicted relative energies when comparing completely
renormalized coupled-cluster calculations950953 (CR-CC(2,3))
and multi-reference second-order perturbation theory954
(CASPT2) and based on such agreement benchmarked various
DFT protocols against the two WFT methods. They observed
that local functionals such as BLYP, mPWPW, and M06-L
provided generally good agreement with benchmark values,
but the inclusion of HartreeFock exchange in functionals like
B3LYP, mPW1PW, M06, or M06-2X led to increasingly
inaccurate results with increasing HartreeFock exchange.
Interestingly, the most accurate results from the BLYP and
mPWPW functionals were obtained when a restricted
KohnSham formalism was employed for the singlet states,
even though the KohnSham wave functions were unstable to
spin-symmetry breaking. With the M06-L functional, spin
purication of broken-symmetry energies provided slightly
improved values for singlet energies compared to restricted
KohnSham energies. For the remaining functionals, errors
were large irrespective of the protocol used to assign singlet
state energies.
By studying the ligands in Fig. 5, Cramer et al.949 were able
successfully to rationalize the observed geometries and ground
electronic states of a diverse array of experimentally

Fig. 5 Benchmark ligand systems L for comparing singlet and triplet electronic state energies for end-on and side-on LCuO2 coordination
geometries and TMG3trenCuO2+ (inset).

10790 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

characterized LCuO2 complexes, identifying the importance of

ligand stereoelectronic eects on orbital interactions between
the LCu and O2 fragments. One particular case is especially
interesting, namely TMG3trenCuO2+ (TMG3tren = 1,1,1tris{2-[N2-(1,1,3,3-tetramethylguanidino)]ethyl}amine;
see
Fig. 5 inset). Originally reported by Schatz et al.,955 the end-on
nature of the O2 coordination was assigned in part based on
good agreement between observed resonance Raman O2
stretching frequencies and those computed for a restricted
singlet state at the BP86 level. The O2 coordination geometry
was later conrmed by single-crystal X-ray crystallography;956
however, subsequent experimental work by Lanci et al.614
established the electronic ground state of TMG3trenCuO2+
to be triplet, not singlet, suggesting that the correlation in
observed (triplet) and computed (singlet) O2 stretching
frequencies was coincidental. At the mPWPW level of theory,
the triplet state is predicted to be lower in energy than the
restricted singlet by 16 kcal mol1, and it is also lower than
the broken-symmetry state by 7 kcal mol1, consistent with
the experimental ground-state determination.614 The mPWPW
level of theory also provides reasonable quantitative agreement
with the measured 18O equilibrium isotope eect (EIE) for O2
binding, with theory suggesting that solvation eects in acetone
increase the charge transfer from LCu to the O2 fragment,
compared to the gas phase, and play a signicant role in
modulating the O2 stretching frequency. De la Lande et al.957
have also considered the chemistry of the trenCuO2+ system,
applying a variable supplementary pseudopotential to the
apical nitrogen atom to tune the oxidation potential of the
supported Cu atom and examine the eects on CH bond
cleavage reaction coordinates. Gherman et al.958 have examined
the inuence of biological ligands on O2 coordination
geometries and oxidation numbers.
The reactivity of supported LCuO2 complexes has also seen
substantial study at the DFT level. Thus, Aboelella et al.959
examined the inuence of a thioether ligand on the binding of
O2 to a CuI diketiminate, using BLYP and B3LYP to
characterize the energetics and geometric details of the complex,
as well as to explain an experimentally observed equilibrium
between the monocopper complex and a bis(m-oxo) dimer (see
below for further discussion of bis(m-oxo) copper species).
Gherman et al.960 used B3LYP to compare variations in
oxygenation pathways, reduction potentials, and other
properties when comparing diketiminate supporting ligands
to anilidoimine ligands; Hill et al.961 employed the same level
of theory to explore the eects of ligand uorination; and
Heppner et al.962 examined the possible inuence of other
auxiliary ligands on Cu-diketiminate-O2 electronic structures.
(In all of the studies cited in this paragraph, the poor
performance of B3LYP for predicting singlettriplet state
energy splittings in LCuO2 complexes was suspected and
CASPT2 state energy splittings were computed instead.)
Considering a novel reaction of LCuO2 species, Hong
et al.963 have employed M06-L to characterize the reaction
path by which a supported CuI-a-ketocarboxylate-O2 complex
is transformed by decarboxylation of the ketocarboxylate into
a highly reactive CuQO species (best described as a triplet
CuIIO(/ ) moiety) which then goes on to oxygenate a ligand
phenyl ring via a concerted electrophilic attack on the ring.
This journal is


c

the Owner Societies 2009

Huber et al.488,964 expanded this initial study to explore ligand

eects on the energetics of various steps, and moreover on
computed electronic state energy dierences, singlet vs. triplet,
for all intermediates. For singlettriplet splittings M06-L was
compared to CASPT2 and RASPT2 and predictions from
the DFT and WFT levels of theory were generally within
2 kcal mol1 of one another irrespective of whether broken
symmetry energies were used or spin purication was carried
out. The starting O2 complexes were exceptions, however, with
M06-L predicting the singlet states to be substantially too high
in energy, probably owing to the large nondynamical correlation
associated with what amounts to a weakly perturbed singlet
oxygen fragment. Comba et al.965 have used B3LYP
and coupled cluster methods also to study the aromatic
hydroxylation activity of a CuQO species, in this case generated
from NO homolysis of CuTMAO complexes (TMAO =
trimethylamine N-oxide). As in the system of Hong et al.,963
concerted attack of the oxo fragment on the aromatic ring is
predicted by both B3LYP and coupled-cluster calculations to
be favored over a stepwise H-abstraction/hydroxylation path,
giving good agreement for the energetics of the reaction
pathway. The critical intermediacy of a CuQO species has
also been invoked by Yoshizawa and Shiota966 based on DFT
studies, including QM/MM analyses, of the mechanism of
methane to methanol conversion at the monocopper active site
in particulate methane monooxygenase.
Considering the nitrogen equivalent of the CuQO fragment,
namely a copper nitrene, Comba et al.967 have used B3LYP to
characterize the mechanistic details of the bispidine copper
catalyzed aziridination of olens. A spin crossover is required
to move from triplet reactants to singlet products. Bar-Nahum
et al.558 have used M06-L to determine the mechanism by
which a trinuclear copper cluster reduces N2O to N2.
The structures and reactivities of various supported
dinuclear (LCu)2O2 complexes have been the subject of
many modeling studies, in part because they pose signicant
challenges to DFT with respect to computing relative isomer
energies (Fig. 6). This challenge arises primarily from the
varying degrees of open-shell vs. closed-shell character in
singlet states depending on the oxidation states of the
coupled copper atoms; when the two copper atoms are well
described as d9 CuII the usual challenges associated with
multicongurational character in DFT arise.413,485,486,968,969
As the modeling of (LCu)2O2 systems has been very recently
reviewed,942 we will mention only one recent addition.

Fig. 6 Dierent core geometries for supported (LCu)2O2 complexes.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10791

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

In particular, Sander et al.970 have provided a detailed

characterization of the mechanistic details associated with
aromatic hydroxylation accomplished by an (LCu)2O2 complex
having a m-Z2:Z2 core.
A sometimes important reactive intermediate in the activation
of molecular oxygen by copper is the hydroperoxide CuOOH.
Osako et al.971 employed the B98 functional to establish, by
comparison of computed and experimental vibrational
frequencies and EPR parameters, that a reactive intermediate
originally presumed to be a CuII peroxide was instead a
CuIIOOH species. Ghattas et al.,972 based in part on B3LYP
calculations, have also posited the importance of a CuOOH
intermediate in the copper catalyzed conversion of 1-aminocyclopropane carboxylic acid into ethylene by hydrogen
peroxide. In acetone as solvent, the CuOOH intermediate
can react with a molecule of solvent to generate a coordinated
hydroxyhydroperoxypropane fragment that is itself capable of
aromatic ring hydroxylation;563 the mechanism for this process
was studied by Kunishita et al.973 at the mPWPW level.
Considering other ligands binding to copper, Periyasamy
et al.974 have concluded based on computed isomer energies
and EPR parameters that the binding of nitric oxide to the CuI
site in copper nitrite reductase is end-on, which is not
consistent with current structural data, suggesting that higher
resolution crystallography may be required to assess this
discrepancy.
In the area of computational electrochemistry, Holland
et al.975 have employed DFT and continuum solvent models
in the characterization of supported copper complexes that
undergo quasi-reversible one-electron reductions at biologically
accessible potentials that render them suitable as medical
imaging agents for the study of hypoxia. As noted above,
Schultz et al.557 also rationalized experimental cyclic voltammetry
data in multicopper helicates based on DFT calculations with
continuum solvation, observing wirelike behavior in a tetranuclear system. Nazmutidinov et al.976 modeled medium
eects on the multi-electron reduction of CuIICuII binuclear
complexes.
Reveles et al.977 have examined the nucleation of water
molecules around Au+, choosing a functional and basis set
combination based on good agreement with experiment for
simple water clusters. Consistent with the trend in experimental
gas-phase binding energies and earlier theoretical studies, they
predict that only 2 molecules of water can be considered to
coordinate to the Au cation; subsequent waters form a shell
that is anchored to the two ligating waters. As noted for Ru
cations in section 5.3.1.6, they found signicant charge transfer
to the outermost water molecules in increasingly large
solvation shells.
Considering supported gold compounds in homogeneous
catalysis, Cheong et al.978 employed B3LYP to establish the
mechanism of homogeneous gold-catalyzed cycloisomerization
of 1,5-allenynes to cross-conjugated trienes. Interesting, two
molecules of catalyst are required to activate the alkyne.
Soriano and Marco-Contelles979 also used B3LYP and studied
Pt- and Au-catalyzed the mechanisms of cycloisomerizations
of enynes and propargylic esters.
Schwerdtfeger et al.419 carried out accurate coupled
cluster calculations on AuCO, Au2CO, and their positive
10792 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

and negative ions. They used this data to test four density
functionals, nding that B3LYP performs better than LSDA,
BP86, and PW91. Unpublished calculations980 show that
M06-L performs about as well as B3LYP on this test.
5.3.1.10 Zinc, cadmium, mercury. Carrasco et al.981 have
used plane-wave PW91 calculations to study novel polymorphs
of ZnO formed from the combination of low-energy (MO)12
cage structures. The resulting nanoparticles would lead in the
bulk limit to ultralow density mesoporous solids with potentially
interesting properties, particularly if dopants can be
introduced into the cavities. By analysis of equation of state
data, Carrasco et al. predict that their most stable poly-cage
structure should exhibit a bulk stability falling between the
known wurtzite and rock-salt phases of ZnO, suggesting that it
should be an ideal candidate for synthetic attention.
Cadenbach et al.982 have shown the utility of alkylzinc and
monocyclopentadienylzinc fragments as one electron donors
to other transition metals, e.g., in the unusual molybdenum
complex Mo(ZnCH3)9(ZnCp*)3.
Bernasconi et al.983 used the BLYP functional to study the
dissociation of a water molecule coordinated to Zn2+ in
aqueous solution. They observed delocalization of OH charge
density over several water molecules in Zn(H2O)5(OH)+.
Zn clusters have been studied by various workers. Dai et al.984
used PW91, and Iokibe et al.985 used LSDA, PW91, and
B3LYP. Sorkin et al.431 found that B3LYP and M05-2X both
give reasonable results for the binding energies of small Zn
clusters. Iokibe et al.985 found that the bond length and
cohesive energy of the clusters depends on the extent of
4s/4p mixing, in contrast to earlier transition metals where
4s/3d mixing is a key quantity.
Botticelli et al.929 applied the PBE functional to investigate
13-atom ZnCu alloy clusters, both neural and cationic, and
found special stability for icosahedral Zn7Cu6 clusters.
However, they also found two other low-lying (within 0.2 eV
of the global minimum) isomers of Zn7Cu6, indicating that the
structure is not magic. The search for magic numbers and
low-energy isomers of metal nanoparticles is an important
eld for future activity, as theory can characterize the
rugged landscapes of such particles in a way not accessible
to experiment.928
Tachikawa et al.986 applied PW91 to calculate structures
and electronic states Znn(H2O)m with n = 132 and m = 13.
In agreement with earlier work,987 they found that the transition
from noncovalent to covalent binding in Znn occurs between
Zn3 and Zn4. This transition persists when water is bound.
Suwattanamala et al.988 carried out B3LYP and PBE0
calculations on dipropoxythiacalix[4]arenes in chloroform
solvent and examined the preferred metal coordination modes
in their Zn2+ complexes by using B3LYP.
Lu et al.989 have employed B3LYP and TD-B3LYP to
rationalize the structure and intense uorescence of a
trinuclear cadmium coordination polymer, noting that the
cadmium plays only a structural role, with the optical transitions
being well described as ligand-to-ligand charge transfer in
character. Fernandez et al.990 used a phosphinothiol ligand
designed to suppress the formation of coordination polymers
and successfully isolated a series of polynuclear Zn, Cd, and
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

Hg compounds. These authors used B3LYP calculations to

assess the nature of the bonding between the group 12 metals
and their dierent ligands. Bonding between two group 12
metals has only rarely been demonstrated, but Zhu et al.991
successfully synthesized dimers of the form ArMMAr for
Ar = tetraisopropylterphenyl and M = Zn, Cd, or Hg and
characterized the bonding at the BP86 level using energy
decomposition analysis.
With respect to Hg, Wiederhold et al.992 employed a
combination of M06-L calculations with relativistic pseudopotentials and all-electron relativistic MP2 calculations to
compute the mass-dependent and nuclear volume fractionation
factors (MDF and NVF, respectively) contributing to the
equilibrium partitioning of all Hg isotopes between dissolved
HgII species and thiol-bound HgII, where the latter is a model
for environmentally sorbed HgII. Wiederhold et al. observed
excellent agreement with experiment for HgCl2 and Hg(OH)2
and found that both MDF and NVF made signicant
contributions to the total equilibrium isotope eects.
5.3.2 Solids. There is a wide range of solids containing
transition metals. For example, transition metal oxides may be
insulators, semiconductors, conductors, or superconductors,
and they may have interesting magnetic, ferroelectric, antiferroelectric, and piezoelectric properties; they are important
minerals and constituents of rocks, and they are also important
in modern technology.993 The literature is far too large to be
fully surveyed here, but it is interesting to make some connection
to other topics discussed in this review.
The LSDA already provides a good approximation for
some, but not all, properties of some, but not all, transition
metal solids.994,995 GGAs do not necessarily improve the
results.996
Kurth et al.997 tested the accuracy of nine local functionals
for the unit cell volumes of Cu, Pd, W, Pt, and Au. The mean
unsigned errors (MUEs) are 1.0% for LSDA, 1.32.9% for
GGAs (PBE best, followed by HCTH, RPBE, and BLYP),
and 0.62.7% for meta-GGAs (PKZB best, followed by FT98,
and VS98). Simple math shows that the percentage errors in
lattice constants would be about 3 times smaller. LSDA
underestimates the volumes (overbinding) whereas the GGAs
and meta-GGAs usually overestimate them (underbinding). A
similar trend was found for Cu, Rh, Pd, Ag, Ir, Pt, and Au
with LSDA and PBE.998 For the same ve transition metals
the MUEs in bulk moduli are 19% for LSDA, 8-34% for
GGAs, and 627% for meta-GGAs.997 The TPSS meta-GGA
typically improves on PKZB for TM lattice constants but
makes TM bulk moduli worse.905
Heyd et al.999 showed that they could calculate accurate
lattice constants with the screened Coulomb hybrid HSE
functional. For example, the lattice constants of ZnS, ZnSe,
and ZnTe are overestimated by HSE by an average of 0.7%,
whereas the local PBE functional overestimates the same
lattice constants by an average of 1.4%. This is important
because HSE predicts much more accurate band structures
(discussed below) than do local functionals.
For phonon frequencies of pure TMs, LSDA usually overestimates and PBE usually underestimates, but for thermal
expansion, heat capacity, and free energy, LSDA is usually
This journal is


c

the Owner Societies 2009

more accurate than PBE.998 Suggestions have been made for

improving the accuracy of calculated thermodynamic data of
TMs and alloys by combining DFT data for temperature
dependence with more accurate results for 0 K.998 DFT has
also been applied to more complex solid materials such as
PbTiO3 perovskites10001002 and HCP iron.1003,1004
Sun et al.1005 used the LSDA, PBE, and WC functionals to
calculate the equation of state for Pt. The Helmholtz energy
was approximated as a function of lattice constants and
temperature by setting it equal to a sum of three terms: the
ground-state electronic energy (including core repulsion),
the free energy of thermally excited electronic states, and the
vibrational free energy of the lattice. Scalar relativistic eects
were included, but not spinorbit coupling. Electronic
temperature eects were nonnegligible (as has also been found
true recently for the free energies of Al nanoparticles1006).
With local functionals, the vertical HOMOLUMO orbital
energy gap is equal (to at least a good approximation) to the
lowest TD-DFT excitation energy.1007 This is also true1007 for
range-separated hybrids like HSE so that such functionals can
give more accurate band gaps with the orbital-energy gap
expression than do local hybrids.999 For example, for ZnS,
ZnSe, and ZnTe, HSE orbital energies underestimate the band
gaps by an average of 0.27 eV (714%) whereas for the same
three materials, the local PBE functional underestimates the
band gaps by an average of 1.46 eV (4158%).
Improvements can be obtained with the nonlocal weighted
density approximation (WDA; see section 2.3); for eleven
metal oxides containing transition metals (Ti, Hf, Zr, or
Cu), the average error in the LSDA gaps is 2.0 eV, whereas
the average error in the WDA calculations of Robertson et al.
is 0.5 eV.110 The error in the LSDA gap is inversely
proportional to the high-frequency dielectric constant.110,1008
The ability to calculate the gaps more accurately allows one to
study interesting problems such as the charge states of
interstitial hydrogens in oxides like TiO2, SrTiO3, PbTiO3,
ZrO2, SrZrO3, ZrSiO4, HfO2, ZnO, and CdO.110,1009
Grabowski et al.998 calculated the free-energies and phonon
dispersion curves of Rh, Ir, Pd, Pt, Ag, Au, and Cu with the
LSDA and the PBE GGA. For the electronic free energy they
used the nite-temperature DFT of Mermin.1010 For phonon
dispersion curves, the experimental results generally lie
between those calculated by LSDA and PBE, whereas for free
energies, LSDA is usually more accurate. The biggest dierences
between the two functionals were found for Ag and Au.
Theoretical studies are especially useful for studying materials
under conditions that are hard to produce experimentally, for
example, under the high-pressure conditions in the earths
mantle. Isaev et al.1011 used the PBE density functional to
calculate the electronic energy, including core repulsion, as a
function of volume at high pressure for three phases of FeH,
and they extended their calculations to free energy as a
function of pressure by adding a quasiharmonic vibrational
(i.e., phonon) contribution to the ground-state electronic
energy. Their analysis of the resulting phase diagram shows
phase transitions at 37 GPa at 300 K and 50 GPa at 1000 K.
Umemoto et al.1012 studied the spin state of Fe in ironbearing magnesium silicate perovskite, (Mg,Fe)SiO3, under
lower-mantle conditions; this spin state is a crucial parameter
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10793

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

for determining key mantle properties such as elastic and

seismic properties, the post-perovskite transition pressure,
and electrical and thermal conductivities. Unfortunately the
spin properties depend strongly on the choice of density
functional (LSDA vs. PW91129 or LSDA vs. PBE130).
Nevertheless calculations with the LSDA and PBE functionals
provide valuable insight into the eect of iron concentration,
the coordination number of iron, and the nonmetalllic band
structure at lower-mantle pressures.1012
Shi et al.175 used the PW91 and sX approximations to study
the cohesive energy, structural parameters, vibrational
frequencies, and band structure of Au2O and Au2O3. The
projected density of states was analyzed to characterize
the bonding, in particular the Au 5d and O 2p mixing and
the ionic character. Whereas PW91 predicts Au2O to be
metallic, sX predicts a band gap of 0.8 eV.
For ZnO, the local LSDA, BLYP, PBE, and TPSS
functionals predict a band gap of 0.9 eV, the local M06-L
functional predicts 1.0 eV, and the hybrid HSE functional
predicts 2.9 eV, in comparison to an experimental value of
3.4 eV.999,1013,1014 The eect of Mn cluster doping on the
optical properties of ZnO was studied with the PW1PW
density functional.104
Filippetti et al.1015 used ASIC-corrected (see section 2.2)
LSDA calculations as well as uncorrected LSDA to study the
spin-polarized band structures of the Mn-doped GaN and
GaAs semiconductors. The results conrmed the LSDA
picture for (Ga,Mn)As, but the results of the two kinds of
calculations were very dierent for (Ga,Mn)N. Additional
DFT calculations on this kind of half-metallic diluted
magnetic semiconductor were reported by several subsequent
groups.10161022 Filippetti and Fiorentini1023 applied the SIC
scheme to cuprates.
Liu and Ge1024,1025 studied hydrogen absorption in
Ti-doped and Sc-doped NaAlH4 with the PBE density functionals
in order to better understand how transition metal doping
might aect reversible hydrogen storage mechanisms.
Recent applications of linear response theory to the optical
excitations of Cu, Ag, and Au are examples of improved
approaches based on DFT.1026,1027
DFT has also been used to study spin polarization in
potential materials (such as CO1xFexS2 alloys) for spintronics
applications. An introduction to this subject is provided by
Leighton et al.1028
An area of special interest for technology is the eect of
external variables on spin-state orderings. In an atom, the
high-spin state is often preferred, whereas crystal eld splitting
can stabilize a lower spin state in coordination compounds.
As an example of the eect of an external variable, changing
the pressure can aect the competition between high-spin
(favored by electron exchange and reduced repulsion between
two electrons otherwise paired in the same orbital) and
low-spin (favored by crystal eld splitting). A recently studied
example is SrFeO2, in which experiments reveal a transition at
33 GPa from an antiferromagnetic (AFM) insulator phase, in
which the four-coordinate FeII is in a high-spin S = 2 state
to a ferromagnetic (FM) metallic phase in which the fourcoordinate FeII is in an intermediate-spin S = 1 state.1029,1030
This was studied with the PBE0 functional, but varying the
10794 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

percentage X of HartreeFock exchange from the standard

value of 25; it was found that the spin transition from
the AFM S = 2 state is to a half-metallic FM S = 1 state if
X o 20, and to a FM S = 2 state if X 4 20. With X = 15, the
calculated transition pressure is at 53 GPa.1029 The transition
was attributed to in-layer bonding interactions of Fe d orbitals
with O p orbitals.
5.3.3 Surfaces and nanoparticles. Surfaces of bulk materials
and nanoparticles, thin lms, self-assembled monolayers,
nanocrystalline coatings, and ligand-decorated materials
interfaces are very important for catalysis and for technological
applications such as molecular electronics, microelectronics,
structural components, magnetic data storage, solar devices,
and sensors. (Catalysis is mainly deferred to section 5.3.4.)
In all of these applications, DFT can be used to study
structural and morphological issues, reactivity and stability,
and optical properties such as photoabsorption. In some cases
it can also be used to calculate electrical and electron
dynamical properties such as conductance and electron
transfer or mesoscopic processes like sintering. Our review
contains only a small subset of this burgeoning eld.
Magyar et al.1031 studied the geometries and magnetic
properties of Au dimer, Au14, and Au nanoclusters of sizes
from Au28 to Au68 with B3LYP, PBE, and PBE0. They found
that the clusters are magnetized primarily on their surfaces and
that the Au14 and Au38 clusters are expanded as compared to
the bulk geometry. They attributed the magnetism of Au
nanoclusters to sd hybridization, in contrast to another
mechanism for surface magnetism in gold, in which local
magnetic moments are induced by coupling to adsorbates.1032
Wende et al.1033 used PW91 to calculate the energy of iron
porphyrin molecules as a function of their distance from a
face-centered cubic Co(100) magnetic substrate. At the
optimum distance they also computed the spin-resolved partial
densities of states in order to study the FeCo exchange
coupling.
The rest of this subsection is primarily devoted to a
sampling of the literature on adsorption; see also section 5.3.4
on heterogeneous catalysis.
Bilic et al.1034 studied physisorption of benzene on Cu, Ag,
and Au surfaces with the PW91 and PBE GGAs; they found
that the binding energy is severely underestimated due to
inaccurate treatment of correlation energy by these functionals.
In related work, Cafe et al.1035 studied monolayers of 1,10phenanthroline on Au surfaces with the PW91 functional.
Because this functional does not treat medium-range correlation
accurately the binding energy of the physisorbed horizontal
molecule was calculated to be only 2 kcal mol1; WFT
calculations were used to estimate that the actual binding
energy is much larger, 15 kcal mol1. The binding energy of
the chemisorbed vertical molecule was also underestimated by
PW91, which yielded an 8 kcal mol1 binding energy as
compared to a WFT estimate of 22 kcal mol1. Ma and
Yang1036 used B3PW91 to calculate Raman frequencies for
adsorbed p-nitroaniline on a single Au-atom model of Au
nanoparticles.
Neyman et al.10371040 calculated the adsorption energies of
all group-6 through group-11 atoms and of dimers, trimers
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

and tetramers of Cu, Ag, and Au (group 11) on MgO(001).

For Pd absorption they obtained a binding energy of 1.4 eV
with the BP86 functional and 1.1 eV with the RPBE
functional, both in good agreement with the experimental
value of 1.2  0.2 eV. The eective adsorption energy per
atom was found to be 7% less for Cu3 than for Cu2 and 22%
less for Au3 than for Au2. They also studied the optical spectra
of monomers and dimers.1041 Additional studies of these
systems were carried out by other workers.10421046 The most
recent study1046 showed that the Au adsorption site most likely
to lead to growth of a dimer (Au2) is a neutral F8 center.
Florez et al. considered the eects of adsorption on that
atomic spin states of the adsorbed metals and found that
although spin-state splittings were reduced, high-spin states
were in general not quenched upon adsorption.1047
Hinneman and Carter1048 used the PBE density functional
to study the adsorption of Y, Hf, and Pt on g-Al2O3(0001).
They found that Y and Hf form strong ionic bonds at 3-fold
hollow sites, whereas Pt bonds more weakly at an on-top site
without appreciable charge transfer.
The interaction of oxygen with Ag surfaces has been studied
for a long time, but it is still controversial, despite the
application of DFT.10491052 The structures of oxide layers
on other late transition metals have also been studied in
detail;425,1051,1053 the study of Rogal et al.425 is particularly
complete in that it uses calculated chemical potentials to
compute the phase diagram of the Pd(100) surface in
equilibrium and constrained equilibrium with an O2 and
CO gas phase. Jacob and coworkers1054 analyzed the adsorption
of oxygen on ve dierent surface faces that are involved in
the faceting of Ir(210), including the eect of aqueous electrolytes under electrochemical conditions.
Ge and coworkers used the PW91 density functional to
study the adsorption and clustering of Pt adatoms on
defect-free and defective anatase TiO2(101).1055,1056 They
found barriers of 1.0 and 1.1 eV for Pt diusion as required
for dimer formation.
Xia et al.1057 used the PW91 density functional to study
surface hydration on freshly cleaved slabs of g-manganite,
MnO(OH). They found that waterwater interactions dominate
watersurface interactions for the density functional used.
King and coworkers1058 used PW91 to explain the fact that
water dimers diuse more rapidly than water monomers on
Pd(111) by a beautiful mechanism that they dubbed the waltz
of the water dimers. The details of their interpretation in
terms of tunneling need to be checked. Another study of a
similar nature is the nding that Pd4 diuses 200 times faster
than Pd across an MgO(100) surface;1059 on this surface, Pd2
shows a walking (not quite waltzing) mechanism.1059
LSDA, PW91, PBE, and revPBE calculations predict that
CO on Pt(111) preferentially adsorbs at 3-fold hollow sites,
but the experimental result is preference for an on-top
site.1060,1061 This is consistent with a general trend for LSDA
and GGA functionals to overly favor higher-coordination
sites.1060 Hybrid B3LYP calculations, in contrast, predict the
correct preference,1061,1062 which was attributed to the
predicted HOMOLUMO gap and the position of the HOMO
and LUMO with respect to the Fermi energy.1062 The reported
contrast between the local PBE and nonlocal PBE0
This journal is


c

the Owner Societies 2009

functionals was particularly dramatic:1063 PBE was said to

favor the 3-fold site by 0.11 eV, and PBE0 was said to favor
the on-top site by 0.13 eV, with the 2-fold bridge site
intermediate for both functionals. Studies on the triatomic
molecule PdCO and the tetra-atomic molecule Pd2CO also
lead to the conclusion that only hybrid functionals can
qualitatively and quantitatively predict the nature of the
s-donation/p-back donation mechanism that is associated
with CO binding in these cases.326 Nevertheless the local
BLYP functional does correctly predict the site preference,
not only for CO/Pt(111), but also for CO/Rh(111) and
CO/Cu(111).221 The BLYP calculations also show that the
HOMOLUMO gap explanation for the site preference is not
correct.221 The PBE density functional predicts accurate CO
stretching frequencies for CO adsorbed on Pt and PtRu
surfaces.1064 The explanation for the frequency shifts in terms
of s donation and p back donation are very similar for
adsorption at surfaces and for organometallic molecules.326,1064
We will return to the important role of the p back-bonding
orbital of CO in section 5.3.4 in discussing the adsorption of
CO on Au(111).1065
Stroppa and Kresse433 applied the local BLYP and PBE
functionals and the nonlocal B3LYP and HSE functionals to
the adsorption of CO at atop and hollow sites on the (111)
surfaces of Ru, Os, Rh, Ir, Pd, Pt, and Ag. They concluded
that BLYP and B3LYP are more accurate for adsorption, with
PBE inaccurate, and HSE least accurate. This contradicts the
previous claim1063 that hybrid functionals solve the CO
adsorption puzzle; Stroppa and Kresse conclude that the
previous conclusion1063 is an artifact of the way that pseudopotentials were employed in the previous work. Stroppa and
Kresse drew general conclusions in a broad discussion that
includes not only BLYP, B3LYP, PBE, and HSE but also
LSDA, RPBE, PBE0, AM05 and PBEsol. One pays a price for
the successful chemisorption predictions of BLYP and
B3LYP, namely a poor description of the lattice constants of
transition metals and a poor description of surface energies.
They concluded that hybrid functionals are not a step forward
for metals, which agrees with the conclusion of Schultz
et al.93,94 prior to the development of M05 and M06.
Unfortunately M05 and M06 have not yet been put to most
of the tests discussed in this section.
Lo and Ziegler1066 carried out spin-polarized PBE calculations
of CO adsorption and dissociation on the (110) surface of
FeCo alloys. They then recalculated the adsorption energies
with the RPBE functional and obtained values lower by
710 kcal mol1. Furthermore, for coverages of 0.25 monolayer
and higher, the favored binding site at Fe changed from a long
bridge to an on-top site, which is consistent with previous
work1067 for Fe(110), whereas the overall favored binding site
at Co did not change. In related studies using the PW91
functional, they examined CO adsorption on Fe(310),1068
and they studied adsorption and reactions of C2 hydrocarbon
fragments on Fe(100);1069 spin polarization was included in all
calculations to account for the ferromagnetic properties of
bulk Fe.
Valcarcel et al.1070 used the PW91 density functional to
calculate the binding modes of propyne on Pd(111) and
Pt(111). Wang et al. used the PW91 density functional to
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10795

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

study the adsorption of 2-methyl-2-propanethiol on

reconstructed Au(111),1071 the adsorption of CH
p3S radicals
on unreconstructed Au(111),1072 and the 22  3 surface
reconstruction of Au(111).1073 In the last-named study they
found that the nearest-neighbor distances in the top layer are
0.090.16 A shorter than the bulk nearest-neighbor distances,
and the interlayer distance is increased by 0.050.17 A by
reconstruction. Friend and Kaxiras and coworkers1074 also
used the PW91 density functional, and they studied changes in
gold surfaces induced by sulfur adsorption. They also studied
structures and binding energies of chlorine adsorbed on Au,
again with the PW91 density functional.1075,1076
Marks et al.1077 studied the surface reconstruction and
surface energies of SrTiO3 by DFT with a limited number of
functionals, and they found TPSSh to be most accurate.
The results are analyzed in detail, which makes the paper
particularly interesting.
Mao et al.1078 calculated spin-polarized band structures
with the PBE density functional for single Mn, Fe, and Co
atoms on a single graphene sheet. Kaghazchi et al.,1054
employing the PBE density functional, evaluated interfacial
free energies of adsorbate-induced surface faces of nanofaceted Ir(210). Kuganathan and Green1079 used the LSDA
to predict the structure and conduction band of CuI in singlewalled carbon nanotubes (SWNTs). Sceats and Green1080 used
LSDA and PW91 functionals to study cobaltocene and
bis(benzene)chromium encapsulated in SWNTs. Friend and
Kaxiras and coworkers1081 used the PW91 density functional
to study MoO3 monolayers nanocrystals on Au(111) and their
selective reduction.
Carter and Yarovsky and their coworkers used the PBE and
PW91 functionals, respectively, to study a variety of adsorption,
dissociation, and diusion processes on Fe, Al/Fe, Fe/Si and
sulfur-covered-Fe surfaces10821090 and structure and magnetism
at Cr/Fe interfaces.1091 Yarovsky and coworkers1092,1093 used
the PW91 density functional to carry out spin-polarized direct
dynamics simulations of H2S dissociation at the (110) and
(100) surfaces of Fe at temperatures from 298 K to 1800 K.
The dissociation mechanism was found to change as a
function of temperature.
Iokibe et al.985 studied the adsorption and diusion of Zn
on Zn(001) with the PW91 density functional. Kurzweil and
Baer1094 used a novel DFT approach to study the surface
plasmon absorption of Au8 and Au18. Ferrando et al.1095
reviewed the application of theory to structural, optical,
magnetic, thermodynamic, and kinetic properties of mixed-metal
clusters and binary alloyed nanoparticles of transition metal
elements.
Zhu et al.1096 used hybrid GGAs to calculate quantum
well states for charge transfer excitations from Au to C60
at the C60/Au(111) interface. The experiments were later
reinterpreted by Li et al.1097 in terms of superatom molecular
orbitals1098,1099 of C60.
Geng et al.1100 and Derosa et al.1101 studied adsorption of
benzenethiolate at transition-metal (Pt, Cu, Ag, Au) surfaces.
Although there are no established rules for the precise usage
of words like nanoparticles and quantum dots, one could
classify isolated molecules in a range of ascending sizes from
complexes and clusters, dened as particles with diameters less
10796 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

than 1 nm, to nanoparticles, dened as particles with diameter

greater than 1 nm but having morphology dierent from
the bulk, up to quantum dots, often dened as having
the bulk structure but exhibiting quantum connement
(of electrons or holes) in three spatial dimensions (note
that systems with quantum connement in only two or one
dimensions are called quantum wires and quantum wells,
respectively, orespecially if smallerthey may be called,
among other possibilities, nanowires, nanotubes, nanowells,
or thin lms). However, these distinctions are uid; for
example, quantum dots are often called nanoparticles, but
nanoparticles that do not have the bulk structure are not
usually called quantum dots. Furthermore, researchers
interested in catalysis are more likely to use the nanoparticle
language, and those interested in electronic properties are
more likely to use the quantum dot language. Some recent
examples of DFT studies modeling quantum dots containing
transition metals are cited to illustrate the possibilities.11021105
5.3.4 Heterogeneous catalysis. The distinctions between
homogeneous and heterogeneous catalysis are blurred in many
applications, such as catalysis by clusters or nanoparticles,
which may be unsupported or supported on metal oxides,
zeolites, or thin lms, or catalysis by supported thin lms,
or by supported organometallic species.856,916,1054,11061122
As mentioned in section 5.3.1.9, gold atoms, gold clusters,
and gold nanoparticles have been especially widely
studied.856,1112,1113,1116,11191121,1123 Although only a fraction
of this kind of work could be included in section 5.3.1, in this
section we focus only on conventional heterogeneous catalysis,
in particular on surfaces of metals, alloys, and oxides; see
also discussions of adsorption and adsorbates on surfaces in
section 5.3.3. Gronbeck,1124 Catlow et al.,1125 Cinquini
et al.,1126 Nrskov et al.,1127,1128 and Neyman and Illas1129
recently provided introductory overviews of DFT applied to
heterogeneous catalysis, and Coquet et al.1130 provided a
critical review of the theory and simulation of heterogeneous
gold catalysis. Two recent contributions have emphasized the
nanoparticle/crystal surface distinctions for catalysis from both
experimental1131 and combined experimental/computational1132
points of view. For example, Falsig et al.1132 contrasted Au12
to Au(111). A review more focused on dynamical processes,
especially of hydrogen, has been presented by Lanzani et al.1133
In discussing dynamical processes at surfaces, especially
metal surfaces (and sometimes in discussing dynamical
processes more generally), the question arises of whether the
BornOppenheimer approximation is valid or, in contrast,
whether one must invoke electronically nonadiabatic processes
and the participation of electronic excited states.11341160 In
the surface science literature this is often called electron-hole
creation or electron-hole participation (similarly, vibrational
excitation of the surface is often called phonon creation).
There is considerable evidence that electronically excited states
play a role in some but not all processes at metal surfaces.
Kroes and coworkers1147 have oered an explanation for why
electron-hole pairs are more important for H2/Pt(111) than for
NO/Au(111). In other work discussed below, it is assumed
(usually implicitly) that only the ground-electronic-state
potential energy surface plays a role.
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

A heterogeneous catalytic process with a dramatic eect on

the course of world history is the Haber process for the
synthesis of ammonia, originally carried out on enriched Fe
and later on Ru, and this reaction is still being studied, most
recently with the PW91 and RPBE density functionals.11611163
It was found that the overall rate predicted by the RPBE
density functional is a factor of 3 to 20 too low, as compared
to experiment.1161
A familiar and practically important example of heterogeneous catalysis by transition metals is catalytic removal of
undesirable gases from automotive exhausts. Liu et al.1164
studied the selective reduction of NO by Ir under conditions
of excess oxygen such as occur in lean burns. In the absence of
an appropriate reductant, oxygen is found to poison the
catalytic process.
Although it is not a catalytic process, the chemisorption of
H2 on metal surfaces has been a very widely studied prototype
process, and much can be learned because of the possibility of
carrying out more complete quantum mechanical dynamics
calculations than for other reagents and the resulting possibility
of obtaining accurate empirical values for barrier heights.242
Recent papers on H2 dissociation over W, Ru, Pt, and Cu be
consulted as an entry point into this literature,1147,11651171 and
there is also a recent review.1172
The binding of CO on MgO(001) has proved to be a very
dicult case, but recent work has shown that the M06-2X and
M06-HF functionals can, for the rst time, provide a
simultaneously satisfactory description of adsorbate geometry,
vibrational frequency shift, and adsorption energy.1173
Although this system has no transition metals, it may provide
a route to further progress on equally challenging problems
such as the bonding and dissociation of nitrogen oxides on Nix
Mg1xO(100).1126, 11741180 In preliminary studies1180 the M06
functional shows the best performance for NO binding to
NixMg1xO(100).
Heyden et al.11811183 used B3LYP to study nitrous oxide
decomposition at iron sites in the iron-exchanged microporous
ZSM-5 zeolite. Rate parameters were determined for 167
elementary steps, and the resulting reaction mechanism
explained a variety of transient and steady-state experiments.
This work elucidated the nuclearity of the active Fe sites
during steady-state decomposition of N2O and showed that
previously invoked oxygen desorption is kinetically not
relevant for this process.
Goodrow and Bell1184 used B3LYP for methanol oxidation
on isolated vanadate ((O)3VQO3) supported on TiO2 and
showed that the dierence in rate from vanadate supported on
TiO2 is not due to the inuence of the support on the electronic
properties of vanadate, but rather is due to the higher incidence
of defects on TiO2. Introduction of an O-vacancy in the
support adjacent to vanadate lowers the apparent activation
energy of the rate limiting step (transfer of an H atom from an
adsorbed methoxy to a vanadyl O) from 23 to 16 kcal mol1.
Bonde et al.1185 used the RPBE functional to model electrochemical oxidation of nanoparticulate MoS2, WS2, CoMoS,
and CoWS with special emphasis on the edge structures.
Gokhale et al. studied the water-gas shift reaction (CO +
H2O - CO2 + H2) on Cu(111)1186 and Pt(111)394 with
the PW91 functional. They found that key mechanistic
This journal is


c

the Owner Societies 2009

determinants are the binding energies of CO, OH, and COOH

and that a mechanism involving COOH is more favorable
than direct oxidation of CO. Their results were used in a
microkinetic model to estimate turnover rates under various
conditions, and they provide an excellent example of the state
of the art in complete catalytic process modeling.
Andersson et al.1187 studied the dissociation of CO on Ni
surfaces. They used the RPBE functional with a scheme to
correct for the systematic overbinding of CO to metals with
this functional. Their study included the eect of coadsorbed
hydrogen and defects, and they included a COH species
in their mechanism; they provided a good account of the
experimental results.
Schneider and coworkers have also placed a special
emphasis on using DFT to simulate catalysis under realistic
conditions of oxidation environment, oxygen pressure,
temperature, particle size, and catalyst composition. They
attempted to quantify the conditions that lead to various
adsorbate coverages. Their work is mainly concerned with
CO and NO oxidation on Pt catalysts,11881192 and they also
used PW91 to study the inuence of oxygen coverage on
the structure, energetics, and electronics of the RuO2(110)
surface.1193
Rosch, Chen, and coworkers398,11941199 and other
researchers1200 used the PW91 and BP86 density functionals
to study the eects of crystal faces, surface composition, and
steps on a variety of decomposition steps on ZnPd, Cu, and Pd
surfaces. The recent work398 on methanol decomposition on
nanocrystallites of Pd catalysts is particularly illuminating
because it models some aspects of the nanocatalyst more
realistically than usual, in particular as a 79-atom nanocrystallite that contains nonideal specic sites and facets that
occur in real catalytic systems.1201 Unfortunately, the cluster
geometries were frozen during the calculations in an attempt
to make up for deciencies of the models and density
functional.
Li et al.1200 found that nuclear tunneling needs to be taken
into account to predict the temperature-programmed
desorption peak temperatures for b-hydride elimination of
ethyl radical on Cu. Additional studies have recently been
carried out using the PBE density functional.1202 Recent
experiments on Pd surfaces with Zn incorporation are
consistent with earlier theoretical studies11941196 but suggest
that such studies should be carried out with lower concentrations
of Zn.1203
Tang and Trout1204 used the PW91 functional to calculate
adsorption energies of SO2 and NO on surface alloys of Pt
with Ru, Ru, Ir, Ni, Pd, Cu, and Au to evaluate potential
oxidation emissions catalysts.
King and coworkers studied the dissociative chemisorption
of methane on Pt1205 and subsequent dehydrogenation steps to
produce CH2, CH, and C.391 Using PW91, they found1205
breaking bond lengths for the transition states of the CH4 CH3 + H reaction on Pt(110)-(1  2) and Pt(111) to be 1.48 A
and 1.54 A, respectively, which may be compared to earlier
calculations that yielded 1.37 A (calculated1206 by embedded
diatomics-in-molecules), 1.53 A (calculated1207 by an
embedded-cluster conguration interaction calculation), 1.59 A
(calculated1208 by PW91), 1.47 A (calculated1209,1210 by
Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10797

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

PW91), or 1.58 A (calculated1211 by PBE) on Ni(111) and

1.53 A (calculated1212 with an unspecied GGA) on Ni(100).
The calculated barriers also dier, with barriers of 17,1207
26,1208 14,1206 and 251211 kcal mol1 on Ni(111) and barriers of
121205 and 211211 kcal mol1 on Pt(111). (Liao et al.1213 and
Psofogiannakis et al.1214 also discussed the CH4 - CH3 + H
reaction at metal surfaces (Pt1213,1214 and three other
metals,1213 with PB86,1213 B3LYP,1214 and PW911214), but
they approximated the metal surfaces by small (unembedded)
clusters.) For CH2 dehydrogenation, the calculated barrier
height of King and coworkers,391 28 kcal mol1, agrees well
with an experiment that was interpreted to yield 29 kcal mol1.
Further work on CH bond activation has been reported by
Li et al.,1215 in this case including both metal surfaces and
oxides. In particular they applied the PBE functional to study
the dissociation of CH4 on Tc(0001), Ru(0001), Pd(111),
Pt(111), Cu(111), and three crystal faces of PdO. Following
Nrskov and coworkers (see next paragraph) they found a
linear correlation between the CH activation barrier and the
total chemisorption energy in the nal dissociated state, and
they discussed the heterolytic component of the dissociation
process.
Nrskov and coworkers have developed a general analysis
based on scaling properties1128 and have applied it to numerous
problems. In such an analysis one compares results for several
dierent surface compositions and tries to identify a small
number of key variables upon which the binding energy
depends linearly. A good example is provided by the adsorption
energies of hydrogen-containing compounds (CHn, NHn,
OHn, and SHn) on Ru, Rh, Ir, Ni, Pd, Pt, Cu, Ag, and Au;
here they identied n as the key variable, and they rationalized
this by an eective medium theory similar to embedded atom
theory.1216 The nding that reaction rates vary linearly with
bond energies or reaction energies is reminiscent of the
EvansPolanyiSemenov linear correlation between activation
energy and enthalpy of reaction in gas-phase kinetics.1217,1218
Another example is provided by the analysis of the selective
hydrogenation of acetylene on metallic and bimetallic
surfaces; the key variables were identied as the heats of
adsorption of acetylene and ethylene; and on this basis they
identied NiZn alloys as potential new selective catalysts
(producing ethylene in preference to ethane), and the prediction
was veried experimentally.1219 Later work based on the same
principles showed how subsurface carbon and alloying Pd
with Ag have similar eects on selective hydrogenation of
acetylene.1220 The compromise required in optimizing a
process that correlates with two or more dierent properties
is often presented in the form of a two-dimensional or threedimensional volcano plot.1221,1222
Surface-catalyzed chain growth mechanisms, that is, the
formation of CC bonds, on Re, Fe, Ru, Co, and Rh were
studied using the PBE functional.1223,1224 Surface catalyzed
hydrogenation has also been studied; for example, Saeys
et al.1225 have used BP86 to assess the mechanistic details
associated with the hydrogenation of benzene to cyclohexane
on Pt(111) surfaces.
Interpretation of the role of promoters on heterogeneous
catalysis has long been a goal of theoretical modeling.1226,1227
Zhang et al.1065 took up the question for stabilization of CO
10798 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

on Au(111); they used the PBE functional to calculate charge

densities of adsorbed species to interpret experimental
observations of promoting eects. They found depletion of
electron density from the p back-bonding orbital of CO when
coadsorbed with NO2, thereby weakening the bonding of CO
to the surface. O, S, and Cl behave analogously to NO2, and
the calculations suggest that the interaction between two
electronegative adsorbates need not be repulsive.
Pacchioni and coworkers147,148 have recently provided a
very pessimistic assessment of the ability of DFT to predict
the energetics needed for quantitative modeling of catalytic
processes in TiO2 and at TiO2 and NiO surfaces. The
functionals on which they based their conclusions are PW91,
PBE, B3PW91, B3LYP, and BH&HLYP. On the other hand,
Labat et al.1228,1229 noted that while it is disappointing that all
functionals fail to predict accurately the relative stability of
rutile and anatase, PBE0 otherwise oered a good compromise
for lattice parameters and band structure.
Porous metalorganic frameworks are becoming more
popular as hosts for catalytic reactions, and Choomwattana
et al.1230 have recently used combined QM/MM calculations
with B3LYP for the QM subsystem for the Cu-catalyzed
reaction of formaldehyde with propene. Kuc et al.1231
concluded that PBE0 greatly underestimates the interaction
energies of hydrogen molecules with the organic linkers of
Zn-based metalorganic frameworks.
Various groups have applied DFT for additional studies of
the structures and energetics of adsorption and heterogeneous
catalytic processes, and selected references are given to
illustrate recent progress.1063,1186,1222,12321250
Despite the great progress that has been made in modeling
heterogeneous catalytic processes, comparison to experimental
catalytic processes is still limited, not only by the quality of the
density functionals (and sometimes by inadequate basis sets or
noninclusion of scalar and/or vector relativistic eects), but
also by the inability to simultaneously include all aspects of
nite temperature, coverage, partial pressures of all species
present, dopants, coadsorbents, reconstruction, local
morphology, defects, and consideration of all possible pathways.
Nevertheless, much has been learned and progress is
accelerating. A key direction for making the calculations more
and more relevant to real catalysis is the increasing use of
kinetic Monte Carlo and microkinetic models in conjunction
with the DFT calculations.394,1186,12511266 Without inclusion
of DFT input and explicit treatment of diusion and
site-specic local concentration eects (spatial eects),
phenomenological models may involve eective parameters which
have only limited (if any) microscopic meaning.1255,1259
Reuter et al.1255 included dissociation, adsorption, surface
diusion, surface chemical reactions, and desorption in their
KMC model of CO oxidation on RuO2(110). Coupling to heat
transfer and uid dynamical treatments of mass transfer are in
the near future.
While the microkinetic models allow improved comparison
to experiment under a variety of realistic conditions involving
competing pathways, better experimental isolation of
preselected pathways allows more direct comparison of theory
with experiment. It is hoped that the recent experimental study
of size-preselected Pt clusters under high-temperature catalytic
This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

conditions (atmospheric pressure) presages a new round of

informative activity of this type.1267,1268 In these experiments,
size-preselected Pt8Pt10 clusters stabilized on high-surface
area supports were used to dehydrogenate propane to propene
with high eciency and good product selectivity, and B3LYP
(for clusters) and RPBE (for surfaces) calculations were
used1268 to interpret the improved catalytic properties of
clusters relative to the surface in terms of the under-coordination
of the Pt sites in the clusters.1268
Tao et al.1269 have recently investigated the performance of
selected density functionals for jellium surfaces and bulk
jellium linear response, where jellium systems serve as a model
for testing functionals.
5.3.5 Electrocatalysis. Electrochemical redox processes at
metal surfaces have also been studied by DFT, but we limit
ourselves here to presented some very recent references on
processes occurring at Pt as a starting point into the literature
of the eld.12701280 For more general remarks on DFT applied
to electrocatalysis involving transition metals, see the recent
Faraday Discussion.1281,1282
5.3.6 Photocatalysis. Doped and sensitized TiO2 are well
studied for their photocatalytic properties.12831287 Although
research on TiO2 structural, electronic, optical, photocatalytic,
photovoltaic, and hydrogen storage properties is still driven
mainly by experiment, there are many opportunities for DFT
to contribute.570575,930,932,1009,1014,1055,1056,1097,1184,12831302
Zapol and Curtis1303 reviewed the application of DFT to
adsorption and photoexcitation of organic adsorbates on TiO2
nanoparticles. Vittadini et al.1290 reviewed theoretical studies
of surface structure, stability, and reactivity of TiO2 surface
systems. Recent studies of the band structure of Mn-doped
rutile TiO21304 and of the nucleation, growth, and adsorption
of transition metal clusters on anatase TiO2 (101)1305,1306 are
of particular interest.
A very recent calculation1287 employs the local PBE
functional and the nonlocal B3LYP functional to calculate
the spin-polarized density of states for Cr/Sb codoped TiO2.
The PBE, experimental, and B3LYP band gaps are 2.6, 3.6,
and 3.9 eV, respectively, and the two calculations give dierent
predictions for the positions of the Sb and Cr states relative to
the valence and conduction bands, but the more accurate
hybrid calculations allow an interpretation of why the co-doped
system shows increased photostability.
Nakamura and Yamashita1307 used a nonequilibrium
Greens function method to calculate the reaction probability
of the photoinduced desorption/dissociation of O2 on Ag(110).
They used the PBE density functional.

6. Concluding remarks
The application of DFT to transition metal systems has
become well established, even though many studies have been
carried out with less than fully reliable density functionals.
Given the rapid pace of ongoing functional development and
the ever increasing scope of applications, the eld must be
regarded as still in its infancy. There remains considerable
room for improvement, and the future is likely to be exciting.
This journal is


c

the Owner Societies 2009

Acknowledgements
The authors are grateful to Yan Zhao for innumerable
discussions, for collaboration on density functional development
and applications, and for providing Fig. 3 and to Ilaria
Cioni, Ke Yang, and Yan Zhao for suggestions on improving
the manuscript. This work was supported in part by the
US Air Force Oce of Scientic Research under grant no.
FA9550-08-1-018 and the US National Science Foundation
under grants CHE-0610183 and CHE07-04974.

References
1 W. Kohn, A. D. Becke and R. G. Parr, J. Phys. Chem., 1996,
100, 12974.
2 P. Hohenberg and W. Kohn, Phys. Rev. B, 1964, 136, B864.
3 L. H. Thomas, Math. Proc. Cambridge Philos. Soc., 1927, 23, 542.
4 E. Fermi, Z. Phys., 1928, 48, 73.
5 W. Kohn and L. J. Sham, Phys. Rev., 1965, 140, A1133.
6 U. von Barth and L. Hedin, J. Phys. C: Solid State Phys., 1972,
5, 1629.
7 A. K. Rajagopal and J. Callaway, Phys. Rev. B: Condens. Matter
Mater. Phys., 1973, 7, 1912.
8 D. R. Hartree, W. Hartree and B. Swirles, Philos. Trans. R. Soc.
London, Ser. A, 1939, 238, 229.
9 O. Sinanoglu and D. F. Tuan, J. Chem. Phys., 1963, 38, 1740.
10 D. G. Truhlar, J. Comput. Chem., 2007, 28, 73.
11 H. J. Silverstone and O. Sinanoglu, J. Chem. Phys., 1966, 44, 1899.
12 M. W. Schmidt and M. S. Gordon, Annu. Rev. Phys. Chem.,
1998, 49, 233.
13 R. J. Deeth, Struct. Bonding, 2004, 113, 37.
14 F. Neese, JBIC, J. Biol. Inorg. Chem., 2006, 11, 702.
15 P. E. M. Siegbahn and T. Borowski, Acc. Chem. Res., 2006, 39, 729.
16 L. Noodleman and W.-G. Han, JBIC, J. Biol. Inorg. Chem.,
2006, 11, 674.
17 L. Bertini, M. Bruschi, L. de Gioia, P. Fantucci, C. Greco and
G. Zampella, Top. Curr. Chem., 2007, 268, 1.
18 V. Polo, E. Kraka and D. Cremer, Mol. Phys., 2002, 100, 1771.
19 D. Cremer, Mol. Phys., 2001, 99, 1899.
20 A. D. Becke, A. Savin and H. Stoll, Theor. Chem. Acc., 1995, 91, 147.
21 J. P. Perdew, A. Savin and K. Burke, Phys. Rev. A: At., Mol.,
Opt. Phys., 1995, 51, 4531.
22 A. Gorling, J. Chem. Phys., 2005, 123, 062203.
23 J. D. Talman and W. F. Shadwick, Phys. Rev. A: At., Mol., Opt.
Phys., 1976, 14, 36.
24 A. M. Teale, A. J. Cohen and D. J. Tozer, J. Chem. Phys., 2007,
126, 074101.
25 A. V. Arbuznikov, M. Kaupp and H. Bahrman, J. Chem. Phys.,
2006, 124, 204102.
26 A. V. Arbuznikov, J. Struct. Chem., 2007, 48, S1.
27 J. F. Janak, Phys. Rev. B: Condens. Matter Mater. Phys., 1978,
18, 7165.
28 I. G. Kaplan, Intermolecular Interactions, Wiley, Chichester,
UK, 2006, ch. 3.
29 K. T. Tang and J. P. Toennies, J. Chem. Phys., 1984, 80, 3726.
30 K. T. Tang and J. P. Toennies, Molecular Physics: An
International Journal at the Interface Between Chemistry and
Physics, 2008, 106, 1645.
31 J. F. Dobson, A. White and A. Rubio, Phys. Rev. Lett., 2006, 96,
073201.
32 J. Harl and G. Kresse, Phys. Rev. B: Condens. Matter Mater.
Phys., 2008, 77, 045136.
33 M. Rohlng and T. Bredow, Phys. Rev. Lett., 2008, 101, 266106.
34 Y. Zhao, N. E. Schultz and D. G. Truhlar, J. Chem. Phys., 2005,
123, 161103.
35 Y. Zhao and D. G. Truhlar, J. Phys. Chem. A, 2006, 110, 5121.
36 Y. Zhao and D. G. Truhlar, J. Chem. Theory Comput., 2006, 2,
1009.
37 Y. Zhao and D. G. Truhlar, J. Chem. Theory Comput., 2007, 3,
289.
38 Y. Zhao and D. G. Truhlar, J. Chem. Phys., 2006, 125, 194101.
39 Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10799

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

40 C. Adamo and F. Lelj, J. Chem. Phys., 1995, 103, 10605.

41 T. Ziegler, Chem. Rev., 1991, 91, 651.
42 J. P. Perdew, in Density Functional Theory of Molecules,
Clusters, and Solids, ed. D. E. Ellis, Kluwer, Dordrecht, 1995,
p. 47.
43 W. Kohn, Rev. Mod. Phys., 1999, 71, 1253.
44 E. J. Baerends, Theor. Chem. Acc., 2000, 103, 265.
45 J. P. Perdew and S. Kurth, in A Primer in Density Functional
Theory, ed. C. Foilhais, F. Nogueira and M. Marques,
Spinger-Verlag, Berlin, 2003, p. 1.
46 U. von Barth, Phys. Scr., 2004, t109, 9.
47 G. E. Scuseria and V. N. Staroverov, in Theory and Applications
of Computational Chemistry: The First Forty Years,
ed. C. E. Dykstra, G. Frenking, K. S. Kim and G. E. Scuseria,
Elsevier, Amsterdam, 2005, p. 669.
48 F. M. Bickelhaupt and E. J. Baerends, Rev. Comput. Chem.,
2000, 15, 1.
49 T. A. Wesolowski, in Molecular Materials with Specic Applications,
ed. W. A. Sokalski, Springer, Heidelberg, 2007, p. 153.
50 J. P. Perdew, A. Ruzsinsky, J. Tao, V. N. Staroverov, G. E. Scuseria
and G. I. Csonka, J. Chem. Phys., 2005, 123, 062201.
51 J. P. Perdew, A. Ruzsinszky, L. A. Constantin, J. Sun and
G. I. Csonka, J. Chem. Theory Comput., 2009, 5, 902.
52 P. A. M. Dirac, Math. Proc. Cambridge Philos. Soc., 1930, 26, 376.
53 J. C. Slater, Phys. Rev., 1951, 81, 385.
54 D. M. Ceperley and B. J. Alder, Phys. Rev. Lett., 1980, 45, 566.
55 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58,
1200.
56 J. Perdew and Y. Wang, Phys. Rev. B: Condens. Matter Mater.
Phys., 1992, 45, 13244.
57 J. P. Perdew, J. Tao and S. Kummel, ACS Symp. Ser., 2007, 958, 13.
58 A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098.
59 J. P. Perdew, Phys. Rev. B: Condens. Matter Mater. Phys., 1986,
33, 8822.
60 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785.
61 J. P. Perdew, in Electronic Structure of Solids 91, ed. P. Ziesche
and H. Eschrig, Akademie Verlag, Berlin, 1991, p. 11.
62 J. P. Perdew, K. Burke and M. Enzerhof, Phys. Rev. Lett., 1996,
77, 3865.
63 C. Adamo and V. Barone, J. Chem. Phys., 1998, 108, 664.
64 M. H. Lim, S. E. Worthington, F. J. Dulles and C. J. Cramer,
in Chemical Applications of Density Functional Theory,
ed. B. B. Laird, R. B. Ross and T. Ziegler, American Chemical
Society, Washington, DC, 1996, vol. 629, p. 402.
65 M. W. Nolan and G. W. Watson, J. Chem. Phys., 2006, 125,
144701.
66 T. Bally and G. N. Sastry, J. Phys. Chem. A, 1997, 101, 7923.
67 C. J. Cramer and S. E. Barrows, J. Org. Chem., 1998, 63, 5523.
68 B. Braieda, P. C. Hiberty and A. Savin, J. Phys. Chem. A, 1998,
102, 7872.
69 M. Sodupe, J. Bertran, L. Rodriguez-Santiago and
E. J. Baerends, J. Phys. Chem. A, 1999, 103, 166.
70 G. Pacchioni, F. Frigoli, D. Ricci and J. A. Weil, Phys. Rev. B:
Condens. Matter Mater. Phys., 2000, 63, 054102.
71 J. L. Gavartin, P. V. Sushko and A. L. Schluger, Phys. Rev. B:
Condens. Matter Mater. Phys., 2003, 67, 035108.
72 J. Poater, M. Sola, A. Rimola, L. Rodriguez-Santiago and
M. Sodupe, J. Phys. Chem. A, 2004, 108, 6072.
73 M. Lundberg and P. E. M. Siegbahn, J. Chem. Phys., 2005, 122,
224103.
74 F. Lechermann, A. Georges, A. Poteryaev, S. Biermann,
M. Posternak, A. Yamasaki and O. K. Andersen, Phys. Rev.
B: Condens. Matter Mater. Phys., 2006, 74, 125120.
75 K. Terakura, T. Oguchi, A. R. Williams and J. Kubler, Phys.
Rev. B: Condens. Matter Mater. Phys., 1984, 30, 4734.
76 V. I. Anisimov, J. Zaanen and O. K. Andersen, Phys. Rev. B:
Condens. Matter Mater. Phys., 1991, 44, 943.
77 V. I. Anisimov, P. Kuiper and J. Nordgren, Phys. Rev. B:
Condens. Matter Mater. Phys., 1994, 50, 8257.
78 S. L. Dudarev, G. A. Bolton and S. Y. Savrasov,
Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 1505.
79 A. D. Becke, J. Chem. Phys., 1998, 109, 2092.
80 N. M. Harrison, V. R. Saunders, R. Dovesi and W. C. Mackrodt,
Philos. Trans. R. Soc. London, Ser. A, 1998, 356, 75.

10800 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

81 T. Bredow and A. B. Gerson, Phys. Rev. B: Condens. Matter

Mater. Phys., 2000, 61, 5194.
82 O. Bengone, M. Alouani, P. Blochl and J. Hugel, Phys. Rev. B:
Condens. Matter Mater. Phys., 2000, 62, 16392.
83 J. Muscat, A. Wander and N. M. Harrison, Chem. Phys. Lett.,
2001, 342, 397.
84 S. A. Gramsch, R. E. Cohen and S. Y. Savrasov, Am. Mineral.,
2003, 88, 257.
85 A. Rohrbach, J. Hafner and G. Kresse, Phys. Rev. B: Condens.
Matter Mater. Phys., 2004, 69, 075413.
86 X.-B. Feng and N. M. Harrison, Phys. Rev. B: Condens. Matter
Mater. Phys., 2004, 69, 035114.
87 C. Franchini, V. Bayer, R. Podloucky, J. Paier and G. Kresse,
Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 045132.
88 F. Tran, P. Blaha, K. Schwarz and P. Novak, Phys. Rev. B:
Condens. Matter Mater. Phys., 2006, 74, 155108.
89 L. Wang, T. Maxisch and G. Ceder, Phys. Rev. B: Condens.
Matter Mater. Phys., 2006, 73, 195107.
90 P. Novak, J. Kunes, L. Chaput and W. E. Pickett, Phys. Status
Solidi B, 2006, 243, 563.
91 D. Kasinathan, J. Kunei, K. Koepernik, C. V. Diaconu,
R. L. Martin, I. D. Prodan, G. E. Scuseria, N. Spaldin,
L. Petit, T. C. Schulthess and W. E. Pickett, Phys. Rev. B:
Condens. Matter Mater. Phys., 2006, 74, 195110.
92 L. Giordano, G. Pacchioni, J. Goniakowski, N. Nilius, E. D. I.
Rienks and H.-J. Freund, Phys. Rev. Lett., 2008, 101, 026102.
93 N. E. Schultz, Y. Zhao and D. G. Truhlar, J. Phys. Chem. A,
2005, 109, 4388.
94 N. E. Schultz, Y. Zhao and D. G. Truhlar, J. Phys. Chem. A,
2005, 109, 11127.
95 F. Furche and J. P. Perdew, J. Chem. Phys., 2006, 124, 044103.
96 P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch,
J. Phys. Chem., 1994, 98, 11623.
97 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
98 P. J. Stephens, F. J. Devlin, C. S. Ashwar, K. L. Bak,
P. R. Taylor and M. J. Frisch, in Chemical Applications
of Density Functional Theory, ed. B. B. Laird, R. B. Ross and
T. Ziegler, American Chemical Society, Washington, DC, 1996,
vol. 629, p. 105.
99 C. Adamo and V. Barone, Chem. Phys. Lett., 1997, 274, 242.
100 M. Ernzerhof and G. E. Scuseria, J. Chem. Phys., 1999, 110, 5029.
101 B. J. Lynch, P. L. Fast, M. Harris and D. G. Truhlar, J. Phys.
Chem. A, 2000, 104, 4811.
102 M. Reiher, O. Salomon and B. A. Hess, Theor. Chem. Acc.,
2001, 107, 48.
103 G. Brewer, M. J. Olida, A. M. Schmiderkamp, C. Viragh and
P. Y. Zavalij, Dalton Trans., 2006, 5617.
104 H. Saal, T. Bredow and M. Binnewies, Phys. Chem. Chem.
Phys., 2009, 11, 3201.
105 M. Radon, M. Srebro and E. Broclawik, J. Chem. Theory
Comput., 2009, 5, 1237.
106 J. A. Alonso and L. A. Girifalco, Phys. Rev. B: Condens. Matter
Mater. Phys., 1978, 17, 3735.
107 J. P. A. Charlesworth, Phys. Rev. B: Condens. Matter
Mater. Phys., 1996, 53, 12666.
108 M. Sadd and M. P. Teter, Phys. Rev. B: Condens. Matter Mater.
Phys., 1996, 54, 13643.
109 P. P. Rushton, D. J. Tozer and S. J. Clark, Phys. Rev. B:
Condens. Matter Mater. Phys., 2002, 65, 235203.
110 J. Robertson, K. Xiong and S. J. Clark, Thin Solid Films, 2006,
496, 1.
111 J. Robertson, K. Xiong and S. J. Clark, Phys. Status Solidi B,
2006, 243, 2054.
112 J. P. Perdew and A. Zunger, Phys. Rev. B: Condens. Matter Mater.
Phys., 1981, 23, 5048.
113 A. Svane, W. M. Temmerman, Z. Szotek, J. Laegsgaard and
H. Winter, Int. J. Quantum Chem., 2000, 77, 799.
114 C. Legrand, E. Suraud and P.-G. Reinhard, J. Phys. B: At., Mol.
Opt. Phys., 2002, 35, 1115.
115 I. Cioni, H. Chermette and C. Adamo, Chem. Phys. Lett., 2003,
380, 12.
116 D. M. Bylander and L. Kleinman, Phys. Rev. B: Condens.
Matter Mater. Phys., 1990, 41, 7868.
117 D. Vogel, P. Kruger and J. Pollmann, Phys. Rev. B: Condens.
Matter Mater. Phys., 1998, 58, 3865.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

118 A. Filippetti and V. Fiorentini, Phys. Rev. B: Condens. Matter

Mater. Phys., 2005, 72, 035128.
119 C. D. Pemmaraju, T. Archer, D. Sanchez-Portal and S. Sanvito,
Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 045101.
120 A. Akande and S. Sanvito, J. Chem. Phys., 2007, 127, 034112.
121 A. I. Liechtenstein, V. I. Anisimov and J. Zaanen, Phys. Rev. B:
Condens. Matter Mater. Phys., 1995, 52, R5467.
122 V. I. Anisimov, F. Aryasetiawan and A. I. Liechtenstein,
J. Phys.: Condens. Matter, 1997, 9, 767.
123 W. E. Pickett, S. C. Erwin and E. C. Ethridge, Phys. Rev. B:
Condens. Matter Mater. Phys., 1998, 58, 1201.
124 J. Hubbard, Proc. R. Soc. London, Ser. A, 1963, 266, 238.
125 H. Hsu, K. Umemoto, M. Cococcioni and R. Wentzcovitch,
Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 125124.
126 A. Svane and O. Gunnarsson, Phys. Rev. Lett., 1990, 65, 1148.
127 Z. Szotek, W. M. Temmerman and H. Winter, Phys. Rev. B:
Condens. Matter Mater. Phys., 1993, 47, 4029.
128 D. G. Isaak, R. E. Cohen, M. J. Mehl and D. J. Singh, Phys.
Rev. B: Condens. Matter Mater. Phys., 1993, 47, 7720.
129 S. Stackhouse, J. P. Brodholt and G. D. Price, Earth Planet. Sci.
Lett., 2007, 253, 282.
130 A. Bengtson, K. Persson and D. Morgan, Earth Planet. Sci.
Lett., 2008, 265, 535.
131 M. Alfredsson, G. David Price, C. R. A. Catlow, S. C. Parker,
R. Orlando and J. P. Brodholt, Phys. Rev. B: Condens. Matter
Mater. Phys., 2004, 70, 165111.
132 G. Kotliar, S. Y. Savrasov, K. Haule, V. S. Oudovenko,
O. Parcollet and C. A. Marianetti, Rev. Mod. Phys., 2006, 78, 865.
133 P. Ravindran, R. Vidya, P. Vajeeston, A. Kjekshus and
H. Fellvag, J. Solid State Chem., 2003, 176, 338.
134 P. Zhang, W. Luo, V. H. Crespi, M. L. Cohen and S. G. Louie,
Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 70, 085108.
135 F. Zhou, M. Cococcioni, C. A. Marianetti, D. Morgan and
G. Ceder, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 70,
235121.
136 R. Pocha, D. Johrendt, B. Ni and M. M. Abd-Elmeguid, J. Am.
Chem. Soc., 2005, 127, 8732.
137 M. Cococcioni and S. de Gironcoli, Phys. Rev. B: Condens.
Matter Mater. Phys., 2005, 71, 035105.
138 H. J. Kulik, M. Cococcioni, D. A. Scherlis and N. Marzari,
Phys. Rev. Lett., 2006, 97, 103001.
139 K. Leung, S. B. Rempe, P. A. Schultz, E. M. Sproviero,
V. S. Batista, M. E. Chandross and C. J. Medforth, J. Am.
Chem. Soc., 2006, 128, 3659.
140 K. Leung and C. J. Medforth, J. Chem. Phys., 2007, 126, 024501.
141 C. Yao, W. Guan, Z. M. Su, J. D. Feng, L. K. Yan and Z. J. Wu,
Theor. Chem. Acc., 2007, 117, 115.
142 N. J. Mosey, P. Liao and E. A. Carter, J. Chem. Phys., 2008,
129, 014103.
143 H. J. Kulik and N. Marzari, J. Chem. Phys., 2008, 129, 134314.
144 S. Lany and A. Zunger, Phys. Rev. B: Condens. Matter Mater.
Phys., 2008, 78, 235104.
145 M. Cococcioni, A. Dal Corso and S. de Gironcoli, Phys. Rev. B:
Condens. Matter Mater. Phys., 2003, 67, 094106.
146 G. Rollmann, H. C. Herper and P. Entel, J. Phys. Chem. A,
2006, 110, 10799.
147 G. Pacchioni, J. Chem. Phys., 2008, 128, 182505.
148 E. Finazzi, C. Di Valentin, G. Pacchioni and A. Selloni, J. Chem.
Phys., 2008, 129, 154113.
149 F. Cinquini, L. Giordano and G. Pacchioni, Theor. Chem. Acc.,
2008, 120, 575.
150 X. Y. Deng, X. Dai and Z. Fang, Europhys. Lett., 2008, 83,
37008.
151 G.-T. Wang, X. Dai and Z. Fang, Phys. Rev. Lett., 2008, 101,
066403.
152 M. Marder, Condensed Matter Physics, John Wiley & Sons,
New York, 2000.
153 J. P. Perdew and M. Levy, Phys. Rev. Lett., 1983, 51, 1884.
154 W. Hanke and L. J. Sham, Phys. Rev. B: Condens. Matter
Mater. Phys., 1980, 21, 4656.
155 L. Brus, J. Phys. Chem., 1986, 90, 2555.
156 Y. Kayanuma, Phys. Rev. B: Condens. Matter Mater. Phys.,
1988, 38, 9797.
157 C. Delerue, M. Lannoo and G. Allan, Phys. Rev. Lett., 2000, 84,
2457.

This journal is


c

the Owner Societies 2009

158 M. S. Hybertsen and S. G. Louie, Phys. Rev. B: Condens. Matter

Mater. Phys., 1988, 37, 2733.
159 G. Onida, L. Reining and A. Rubio, Rev. Mod. Phys., 2002, 74, 601.
160 F. Fuchs, J. Furthmuller, F. Bechstedt, M. Shishkin and G. Kresse,
Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76, 115109.
161 P. H. Hahn, K. Seino, W. G. Schmidt, J. Furthmuller and
F. Bechstedt, Phys. Status Solidi B, 2005, 242, 2720.
162 L. Hedin, Phys. Rev., 1965, 139, A796.
163 S. G. Louie, in Conceptual Foundations of Materials: A Standard
Model for Ground- and Excited-State Properties, ed. S. G. Louie
and M. L. Cohen, Elsevier, Amsterdam, 2006, p. 9.
164 P. Huang and E. A. Carter, Annu. Rev. Phys. Chem., 2008, 59, 261.
165 M. Shishkin and G. Kresse, Phys. Rev. B: Condens. Matter
Mater. Phys., 2007, 75, 235102.
166 P. Rinke, A. Qteis, J. Neugebauer and M. Scheer, Phys. Status
Solidi B, 2008, 245, 929.
167 P. H. Hahn, W. G. Schmidt, K. Seino, M. Preuss, F. Bechstedt
and J. Bernholc, Phys. Rev. Lett., 2005, 94, 037404.
168 L. E. Ramos, J. Paier, G. Kresse and F. Bechstedt, Phys. Rev. B:
Condens. Matter Mater. Phys., 2008, 78, 195423.
169 S. Botti, A. Schindlmayr, R. D. Sole and L. Reisling, Rep. Prog.
Phys., 2007, 70, 357.
170 A. Marini, R. D. Sole and A. Rubio, in Time-Dependent Density
Functional Theory, ed. M. A. L. Marques, C. A. Ulrich, F. Noguiera,
K. Burke and E. K. U. Gross, Springer, Berlin, 2006, p. 161.
171 L. J. Sham and W. Kohn, Phys. Rev., 1966, 145, 561.
172 L. X. Benedict, A. Puzder, A. J. Williamson, J. C. Grossman,
G. Galli, J. E. Klepeis, J.-Y. Raty and O. Pankratov, Phys. Rev.
B: Condens. Matter Mater. Phys., 2003, 68, 085310.
173 T. Ziegler, M. Seth, M. Krykunov, J. Autschbach and F. Wang,
THEOCHEM, 2009, 914, 106.
174 E. K. U. Gross and K. Burke, in Time-Dependent Density Functional
Theory, ed. M. A. L. Marques, C. A. Ulrich, F. Noguiera, K. Burke
and E. K. U. Gross, Springer, Berlin, 2006, p. 1.
175 H. Shi, R. Asahi and C. Stamp, Phys. Rev. B: Condens. Matter
Mater. Phys., 2007, 75, 205125.
176 L. J. Sham and M. Schluter, Phys. Rev. Lett., 1983, 51, 1888.
177 R. W. Godby, M. Schluter and L. J. Sham, Phys. Rev. B:
Condens. Matter Mater. Phys., 1988, 37, 10159.
178 M. Gruning, A. Marini and A. Rubio, Phys. Rev. B: Condens.
Matter Mater. Phys., 2006, 74, 161103.
179 A. J. Cohen, P. Mori-Sanchez and W. Yang, Phys. Rev. B:
Condens. Matter Mater. Phys., 2008, 77, 115123.
180 A. D. Becke, J. Chem. Phys., 1992, 96, 2155.
181 R. van Leeuwen and E. J. Baerends, Phys. Rev. A: At., Mol.,
Opt. Phys., 1994, 49, 2421.
182 A. Dal Corso and R. Resta, Phys. Rev. B: Condens. Matter
Mater. Phys., 1994, 50, 4327.
183 A. V. Arbuznikov and M. Kaupp, Chem. Phys. Lett., 2003, 381,
495.
184 P. Mori-Sanchez, A. J. Cohen and W. Yang, J. Chem. Phys.,
2006, 125, 201102.
185 A. D. Becke and M. R. Roussel, Phys. Rev. A: At., Mol.,
Opt. Phys., 1989, 39, 3761.
186 A. D. Becke, J. Chem. Phys., 1996, 104, 1040.
187 J. Tao, J. P. Perdew, V. N. Staroverov and G. E. Scuseria,
Phys. Rev. Lett., 2003, 91, 146401.
188 A. Savin, in Recent Developments and Applications of Modern
Density Functional Theory, ed. J. Seminario, Elsevier,
New York, 1996, p. 227.
189 T. Leininger, H. Stoll, H.-J. Werner and A. Savin, Chem. Phys.
Lett., 1997, 275, 151.
190 H. Iikura, T. Tsuneda, T. Yanai and K. Hirao, J. Chem. Phys.,
2001, 115, 3540.
191 J. Heyd, G. E. Scuseria and M. Ernzerhof, J. Chem. Phys., 2003,
118, 8207.
192 J. Heyd and G. E. Scuseria, J. Chem. Phys., 2004, 120, 7274.
193 J. Heyd and G. E. Scuseria, J. Chem. Phys., 2004, 121, 1187.
194 R. Baer and D. Neuhauser, Phys. Rev. Lett., 2005, 94, 043002.
195 E. Livshitz and R. Baer, Phys. Chem. Chem. Phys., 2007, 9, 2932.
196 T. M. Henderson, A. F. Izmaylov, G. E. Scuseria and A. Savin,
J. Chem. Phys., 2007, 127, 221103.
197 J. Jaramillo, G. E. Scuseria and M. Ernzerhof, J. Chem. Phys.,
2003, 118, 1068.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10801

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

198 H. Bahmann, A. Rodenberg, A. V. Arbuznikov and M. Kaupp,

J. Chem. Phys., 2007, 126, 011103.
199 A. V. Arbuznikov and M. Kaupp, Chem. Phys. Lett., 2007, 440,
160.
200 A. V. Arbuznikov and M. Kaupp, J. Chem. Phys., 2008, 128,
214107.
201 A. V. Krukau, G. E. Scuseria, J. P. Perdew and A. Savin,
J. Chem. Phys., 2008, 129, 124103.
202 P. M. W. Gill, Mol. Phys., 1996, 89, 433.
203 Y. Zhang and W. Yang, Phys. Rev. Lett., 1998, 80, 890.
204 B. Hammer, L. B. Hansen and J. K. Nrskov, Phys. Rev. B:
Condens. Matter Mater. Phys., 1999, 59, 7413.
205 L. Vitos, B. Johansson, J. Kollar and H. L. Skriver, Phys. Rev.
B: Condens. Matter Mater. Phys., 2000, 62, 10046.
206 W. Kohn and A. E. Mattsson, Phys. Rev. Lett., 1998, 81, 3487.
207 N. C. Handy and A. J. Cohen, Mol. Phys., 2001, 99, 403.
208 X. Xu and W. A. Goddard, III, Proc. Natl. Acad. Sci. U. S. A.,
2004, 101, 2673.
209 R. Armiento and A. E. Mattsson, Phys. Rev. B: Condens. Matter
Mater. Phys., 2005, 72, 085108.
210 L. A. Constantin, J. M. Pitarke, J. F. Dobson, A. Garcia-Lekue
and J. P. Perdew, Phys. Rev. Lett., 2008, 100, 036401.
211 Z. Wu and R. E. Cohen, Phys. Rev. B: Condens. Matter Mater.
Phys., 2006, 73, 235116.
212 Y. Zhao and D. G. Truhlar, Phys. Rev. B: Condens. Matter
Mater. Phys., 2008, 78, 197101.
213 J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov,
G. E. Scuseria, L. A. Constantin, X. Zhou and K. Burke,
Phys. Rev. Lett., 2008, 100, 136406 and the erratum:
J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov,
G. E. Scuseria, L. A. Constantin, X. Zhou and K. Burke,
Phys. Rev. Lett., 2009, 102, 039902.
214 Y. Zhao and D. G. Truhlar, J. Chem. Phys., 2008, 128, 184109.
215 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1998,
80, 891.
216 J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov,
G. E. Scuseria, L. A. Constantin, X. Zhou and K. Burke, Phys.
Rev. Lett., 2008, 101, 239702.
217 A. E. Mattsson, R. Armiento and T. R. Mattsson, Phys. Rev.
Lett., 2008, 101, 239701.
218 A. E. Mattsson, R. Armiento, P. A. Schultz and T. R. Mattsson,
Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 73, 195123.
219 A. E. Mattsson, R. Armiento, J. Paier, G. Kresse, J. M. Wills
and T. R. Mattsson, J. Chem. Phys., 2008, 128, 084714.
220 S. F. Sousa, P. A. Fernandes and M. J. Ramos, J. Phys. Chem.
A, 2007, 111, 10439.
221 M. Alaei, H. Akbarzadeh, H. Gholizadeh and S. de Gironcoli,
Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 085414.
222 H. L. Schmider and A. D. Becke, J. Chem. Phys., 1998, 108, 9624.
223 F. A. Hamprecht, A. J. Cohen, D. J. Tozer and N. C. Handy,
J. Chem. Phys., 1998, 109, 6264.
224 P. J. Wilson, T. J. Bradley and D. J. Tozer, J. Chem. Phys., 2001,
115, 9233.
225 T. W. Keal and D. J. Tozer, J. Chem. Phys., 2005, 123, 121103.
226 Y. Zhao and D. G. Truhlar, J. Phys. Chem. A, 2005, 109, 5656.
227 A. D. Becke, J. Chem. Phys., 2000, 112, 4020.
228 A. D. Boese and N. C. Handy, J. Chem. Phys., 2002, 116, 9559.
229 A. D. Boese and J. M. L. Martin, J. Chem. Phys., 2004, 121, 3405.
230 J. Krieger, J. Chen, G. J. Iaate and A. Savin, in Electron
Correlations and Material Properties, ed. A. Gonis and
N. Kioussis, Plenum, New York, 1999, p. 463.
231 J. P. Perdew, S. Kurth, A. Zupan and P. Blaha, Phys. Rev. Lett.,
1999, 82, 2544.
232 Y. Zhao and D. G. Truhlar, J. Phys. Chem. A, 2006, 110, 13126.
233 Y. Zhao, N. E. Schultz and D. G. Truhlar, J. Chem. Theory
Comput., 2006, 2, 364.
234 Y. Zhao and D. G. Truhlar, J. Chem. Theory Comput., 2008, 4,
1849.
235 E. A. Amin and D. G. Truhlar, J. Chem. Theory Comput., 2008,
4, 75.
236 T. Yanai, D. P. Tew and N. C. Handy, Chem. Phys. Lett., 2004,
393, 51.
237 T. Van Voorhis and G. E. Scuseria, J. Chem. Phys., 1998, 109,
400.
238 Y. Zhao and D. G. Truhlar, J. Phys. Chem. A, 2004, 108, 6908.

10802 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

239 E. I. Proynov, S. Sirois and D. R. Salahub, Int. J. Quantum

Chem., 1997, 64, 427.
240 J. Pu and D. G. Truhlar, J. Chem. Phys., 2002, 116, 1468.
241 B. L. Kormos and C. J. Cramer, J. Phys. Org. Chem., 2002, 15,
712.
242 C. Diaz, E. Pijper, R. A. Olsen, H. F. Busnengo, D. J. Auerbach
and G. J. Kroes, preprint (April 21 2009).
243 G. Lippert, J. Hutter and M. Parrinello, Mol. Phys., 1997, 92,
477.
244 J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello,
T. Chassaing and J. Hutter, Comput. Phys. Commun., 2005, 167, 103.
245 J. C. Phillips and L. Kleinman, Phys. Rev., 1959, 116, 287.
246 M. H. Cohen and V. Heine, Phys. Rev., 1961, 122, 1821.
247 C. F. Melius, B. D. Olafson and W. A. Goddard, Chem. Phys.
Lett., 1974, 28, 457.
248 L. R. Kahn, P. Baybutt and D. G. Truhlar, J. Chem. Phys.,
1976, 65, 3826.
249 P. A. Christiansen, Y. S. Lee and K. S. Pitzer, J. Chem. Phys.,
1979, 71, 4445.
250 D. R. Hamann, M. Schluter and C. Chiang, Phys. Rev. Lett.,
1979, 43, 1494.
251 W. J. Stevens, H. Basch and M. Krauss, J. Chem. Phys., 1984,
81, 6026.
252 P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 270.
253 P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 284.
254 P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299.
255 A. M. Rappe, K. M. Rabe, E. Kaxiras and J. D. Joannopoulos,
Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 1227.
256 D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys.,
1990, 41, 7892.
257 N. Troullier and J. L. Martins, Phys. Rev. B: Condens. Matter
Mater. Phys., 1991, 43, 1993.
258 T. V. Russo, R. L. Martin and P. J. Hay, J. Phys. Chem., 1995,
99, 17085.
259 S. Goedecker, M. Teter and J. Hutter, Phys. Rev. B: Condens.
Matter Mater. Phys., 1996, 54, 1703.
260 P. Schwerdtfeger, J. R. Brown, J. K. Laerdahl and H. Stoll,
J. Chem. Phys., 2000, 113, 7110.
261 J. M. L. Martin and A. Sundermann, J. Chem. Phys., 2001, 114,
3408.
262 A. Dal Corso, Phys. Rev. B: Condens. Matter Mater. Phys.,
2001, 64, 235118.
263 M. Dolg, Theor. Chem. Acc., 2005, 114, 297.
264 E. Fromager, L. Visscher, L. Maron and C. Teichteil, J. Chem.
Phys., 2005, 123, 164105.
265 L. E. Roy, P. J. Hay and R. L. Martin, J. Chem. Theory
Comput., 2008, 4, 1029.
266 P. E. Blochl, C. J. Forst and J. Schimpl, Bull. Mater. Sci., 2003,
26, 33.
267 P. E. Blochl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994,
50, 17953.
268 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater.
Phys., 1999, 59, 1758.
269 M. Valiev, E. J. Bylaska, A. Gramada and J. H. Weare,
in Reviews of Modern Quantum Chemistry, ed. K. D. Sen,
World Scientic, Singapore, 2002, vol. 2, p. 1684.
270 M. Valiev, E. J. Bylaska and J. H. Weare, J. Chem. Phys., 2003,
119, 5955.
271 J. J. Mortensen, L. B. Hansen and K. W. Jacobsen, Phys. Rev.
B: Condens. Matter Mater. Phys., 2005, 71, 035109.
272 H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Condens. Matter
Mater. Phys., 1976, 13, 5188.
273 A. E. Mattsson, P. A. Schultz, M. P. Desjarlais, T. R. Mattsson
and K. Leung, Modell. Simul. Mater. Sci. Eng., 2005, 13, R1.
274 A. Sorouri, W. M. C. Foulkes and N. D. M. Hine, J. Chem.
Phys., 2006, 124, 064105.
275 M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias and
J. D. Joannopoulos, Rev. Mod. Phys., 1992, 64, 1045.
276 J. L. Han, L. Z. Sun, X. D. Qu, Y. P. Chen and J. X. Zhong,
Phys. B, 2009, 404, 131.
277 G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater.
Phys., 1993, 48, 13115.
278 G. Y. Sun, J. Kurti, P. Rajczy, M. Kertesz, J. Hafner and
G. Kresse, THEOCHEM, 2003, 624, 37.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

279 J. Paier, R. Hirschl, M. Marsman and G. Kresse, J. Chem. Phys.,

2005, 122, 234102.
280 J. S. Tse, Annu. Rev. Phys. Chem., 2002, 53, 249.
281 B. I. Dunlap, J. W. D. Connolly and J. R. Sabin, J. Chem. Phys.,
1979, 71, 4993.
282 O. Vahtras, J. Almlof and M. W. Feyereisen, Chem. Phys. Lett.,
1993, 213, 514.
283 M. Sierka, A. Hogekamp and R. Ahlrichs, J. Chem. Phys., 2003,
118, 9136.
284 R. Polly, H. J. Werner, F. R. Manby and P. J. Knowles,
Mol. Phys., 2004, 102, 2311.
285 R. M. Wentzcovitch, Phys. Rev. B: Condens. Matter Mater.
Phys., 1991, 44, 2358.
286 R. M. Wentzcovitch, J. L. Martins and G. D. Price, Phys. Rev.
Lett., 1993, 70, 3947.
287 H. M. Senn and W. Thiel, Angew. Chem., Int. Ed., 2009, 48, 1198.
288 Y.-y. Ohnishi, Y. Nakao, H. Sato and S. Sakaki, J. Phys. Chem.
A, 2008, 112, 1946.
289 B. Wang and D. G. Truhlar, J. Chem. Theory Comput., in press.
290 K. G. Dyall and K. Faegri, Introduction to Relativistic Quantum
Chemistry, Oxford University Press, New York, USA, 2007.
291 N. Rosch, A. Matveev, V. A. Nasluzov, K. M. Neyman,
L. Moskaleva and S. Kruger, in Relativistic Electronic Structure
Theory, Part 2. Applications, ed. P. Pyykko
and
P. Schwerdtfeger, Elsevier, Amsterdam, 2004, p. 656.
292 L. L. Foldy and S. A. Wouthuysen, Phys. Rev., 1950, 78, 29.
293 M. Douglas and N. M. Kroll, Ann. Phys., 1974, 82, 89.
294 B. A. Hess, Phys. Rev. A: At., Mol., Opt. Phys., 1986, 33, 3742.
295 G. Jansen and B. A. Hess, Phys. Rev. A: At., Mol., Opt. Phys.,
1989, 39, 6016.
296 A. Wolf, M. Reiher and B. A. Hess, J. Chem. Phys., 2002, 117, 9215.
297 I. Malkin, O. L. Malkina, V. G. Malkin and M. Kaupp,
J. Chem. Phys., 2005, 123, 244103.
298 M. Reiher, Theor. Chem. Acc., 2006, 116, 241.
299 E. van Lenthe, E. J. Baerends and J. G. Snijders, J. Chem. Phys.,
1994, 101, 9783.
300 M. Mayer, S. Kruger and N. Rosch, J. Chem. Phys., 2001, 115,
4411.
301 E. van Lenthe, J. G. Snijders and E. J. Baerends, J. Chem. Phys.,
1996, 105, 6505.
302 F. Wang, G. Hong and L. Li, Chem. Phys. Lett., 2000, 316, 318.
303 D. G. Fedorov, S. Koseki, M. W. Schmidt and M. S. Gordon,
Int. Rev. Phys. Chem., 2003, 22, 551; B. O. Roos and
P.-A. Malmqvist, Phys. Chem. Chem. Phys., 2004, 6, 2919.
304 C. van Wullen, J. Comput. Chem., 2001, 23, 779.
305 M. Shoji, K. Koizumi, R. Takeda, Y. Kitagawa, S. Yamanaka,
M. Okumura and K. Yamaguchi, Polyhedron, 2007, 26, 2335.
306 K. Yamaguchi, S. Yamanaka, M. Nishino, Y. Takano,
Y. Kitagawa, H. Nagao and Y. Yoshioka, Theor. Chem. Acc.,
1999, 102, 328.
307 S. Lunell, Chem. Phys. Lett., 1972, 13, 93.
308 K. Yamaguchi, Y. Yoshioka and T. Fueno, Chem. Phys. Lett.,
1977, 46, 360.
309 S. Hammes-Schier and H. C. Andersen, J. Chem. Phys., 1993,
99, 1901.
310 S. Yamanaka, D. Yamaki, Y. Shigeta, H. Nagao, Y. Yoshioka,
N. Suzuki and K. Yamaguchi, Int. J. Quantum Chem., 2000, 80, 664.
311 S. Yamanaka, D. Yamaki, Y. Shigeta, H. Nagao and
K. Yamaguchi, Int. J. Quantum Chem., 2001, 84, 670.
312 D. Hobbs and J. Hafner, J. Phys.: Condens. Matter, 2000, 12, 7025.
313 S. Sharma, J. K. Dewhurst, C. Ambrosch-Draxl, S. Kurth,
N. Helbig, S. Pittalis, S. Shallcross, L. Nordstrom and E. K.
U. Gross, Phys. Rev. Lett., 2007, 98, 196405.
314 D. Naveh and L. Kronik, Solid State Commun., 2009, 149, 177.
315 T. Oda, A. Pasquarello and R. Car, Phys. Rev. Lett., 1998, 80, 3622.
316 D. Hobbs, G. Kresse and J. Hafner, Phys. Rev. B: Condens.
Matter Mater. Phys., 2000, 62, 11556.
317 P. Ruiz-D az, J. Dorantes-Davila and G. M. Pastor, Eur. Phys.
J. D, 2009, 52, 175.
318 N. Fujima and T. Oda, Eur. Phys. J. D, 2003, 24, 89.
319 C. Kohler, T. Frauenheim, B. Hourahine, G. Seifert and
M. Sternberg, J. Phys. Chem. A, 2007, 111, 5622.
320 C. Kohl and G. F. Bertsch, Phys. Rev. B: Condens. Matter
Mater. Phys., 1999, 60, 4205.

This journal is


c

the Owner Societies 2009

321 J. E. Peralta, G. E. Scuseria and M. J. Frisch, Phys. Rev. B:

Condens. Matter Mater. Phys., 2007, 75, 125119.
322 S. Yamanaka, R. Takeda, T. Kawakami, K. Nakata,
T. Sakuma, T. Takada and K. Yamaguchi, J. Magn. Magn.
Mater., 2004, 272276/Supplement 1, E255.
323 K. S. Pitzer, Acc. Chem. Res., 1979, 12, 271.
324 P. Pyykko and J. P. Desclaux, Acc. Chem. Res., 1979, 12, 276.
325 P. Pyykko, Chem. Rev., 1988, 88, 563.
326 N. E. Schultz, B. F. Gherman, C. J. Cramer and D. G. Truhlar,
J. Phys. Chem. B, 2006, 110, 24030.
327 M. Filatov, Chem. Phys. Lett., 2003, 373, 131.
328 E. Van Lenthe and E. J. Baerends, J. Comput. Chem., 2003, 24,
1142.
329 B. O. Roos, R. Lindh, P.-A. Malmqvist, V. Veryazov and
P.-O. Widmark, J. Phys. Chem. A, 2004, 108, 2851.
330 R. L. A. Haiduke, L. G. M. De Macedo and A. B. F. Da Silva,
J. Comput. Chem., 2005, 26, 932.
331 B. O. Roos, R. Lindh, P.-A. Malmqvist, V. Veryazov and
P.-O. Widmark, J. Phys. Chem. A, 2005, 109, 6575.
332 K. Dyall, Theor. Chem. Acc., 2006, 115, 441.
333 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr.,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam,
S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi,
G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida,
T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li,
J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. Ochterski, P. Y. Ayala,
K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg,
V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain,
O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Cliord,
J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A.
Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe,
P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong,
C. Gonzalez and J. A. Pople, GAUSSIAN 03 (Revision D.1),
Gaussian, Inc., Wallingford, CT, 2004.
334 B. G. Johnson and M. J. Fisch, J. Chem. Phys., 1994, 100, 7429.
335 J. Zheng, M. A. Iron, B. A. Ellingson, J. C. Corchado,
Y.-Y. Chuang and D. G. Truhlar, NWChemRate 2007,
University of Minnesota, Minneapolis, 2007.
336 A. Fernandez-Ramos, B. A. Ellingson, B. C. Garrett and
D. G. Truhlar, Rev. Comput. Chem., 2007, 23, 125.
337 G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter
Mater. Phys., 1996, 54, 11169.
338 J. Hafner, J. Comput. Chem., 2008, 29, 2044.
339 P. Blaha, K. Schwarz, P. Sorantin and S. B. Trickey, Comput.
Phys. Commun., 1990, 59, 399.
340 K. Schwarz and P. Blaha, Comput. Mater. Sci., 2003, 28, 259.
341 J. Paier, M. Marsman, K. Hummer, G. Kresse, I. C. Gerber and
J. G. Angyan, J. Chem. Phys., 2006, 124, 154709.
342 A. Kiejna, G. Kresse, J. Rogal, A. De Sarkar, K. Reuter and
M. Scheer, Phys. Rev. B: Condens. Matter Mater. Phys., 2006,
73, 035404.
343 C. Fonseca Guerra, J. G. Snijders, G. te Velde and
E. J. Baerends, Theor. Chem. Acc., 1998, 99, 391.
344 G. te Velde, F. M. Bickelhaupt, E. J. Baerends, C. F. Guerra,
S. J. A. Van Gisbergen, J. G. Snijders and T. Ziegler, J. Comput.
Chem., 2001, 22, 931.
345 E. J. Baerends, J. Autschbach, J. A. Berger, A. Berces,
F. M. Bickelhaupt, C. Bo, P. L. d. Boeij, P. M. Boerrigter,
L. Cavallo, D. P. Chong, L. Deng, R. M. Dickson, D. E. Ellis,
M. van Faassen, L. Fan, T. H. Fischer, C. Fonseca Guerra,
S. J. A. van Gisbergen, A. W. Gotz, J. A. Groeneveld,
O. V. Gritsenko, M. Gruning, F. E. Harris, P. van den Hoek,
C. R. Jacob, H. Jacobsen, L. Jensen, E. S. Kadantsev, G. v. Kessel,
R. Klooster, F. Kootstra, M. V. Krykunov, E. v. Lenthe,
J. N. Louwen, D. A. McCormack, A. Michalak, J. Neugebauer,
V. P. Nicu, V. P. Osinga, S. Patchkovskii, P. H. T. Philipsen,
D. Post, C. C. Pye, W. Ravenek, J. I. Rodriguez, P. Romaniello,
P. Ros, P. R. T. Schipper, G. Schreckenbach, J. G. Snijders,
M. Sola`, M. Swart, D. Swerhone, G. t. Velde, P. Vernooijs,

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10803

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

346
347
348

349

350
351

352

353

354
355
356
357

358
359
360

361

L. Versluis, L. Visscher, O. Visser, F. Wang, T. A. Wesolowski,

E. M. van Wezenbeek, G. Wiesenekker, S. K. Wol, T. K. Woo,
A. L. Yakovlev and T. Ziegler, ADF2008, Scientic Computing
and Modeling, Amsterdam, 2008.
G. Wiesenekker and E. J. Baerends, J. Phys.: Condens. Matter,
1991, 3, 6721.
G. te Velde and E. J. Baerends, Phys. Rev. B: Condens. Matter
Mater. Phys., 1991, 44, 7888.
G. te Velde, E. J. Baerends, P. H. T. Philipsen, G. Wiesenekker,
J. A. Broeneveld, J. A. Berger, P. L. de Boeij, R. Klooster,
F. Kootstra, P. Romaniello, J. G. Snijders, E. S. Kadantsev and
T. Ziegler, BAND2008.1, Scientic Computation and Modeling,
Amsterdam, 2008.
R. A. Kendall, E. Apra, D. E. Bernholdt, E. J. Bylaska,
M. Dupuis, G. I. Fann, R. J. Harrison, J. Ju, J. A. Nichols,
J. Nieplocha, T. P. Straatsma, T. L. Windus and A. T. Wong,
Comput. Phys. Commun., 2000, 128, 260.
E. Apra`, E. J. Bylaska, D. J. Dean, A. Fortunelli, F. Gao,
P. S. Krstic, J. C. Wells and T. L. Windus, Comput. Mater. Sci.,
2003, 28, 209.
M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert,
M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunaga,
K. A. Nguyen, S. Su, T. L. Windus, M. Dupuis and
J. A. Montgomery, J. Comput. Chem., 1993, 14, 1347.
M. S. Gordon and M. W. Schmidt, in Theory and Applications
of Computational Chemistry: The First Forty Years,
ed. C. E. Dykstra, G. Frenking, K. S. Kim and G. E. Scuseria,
Elsevier, Amsterdam, 2005, p. 1167.
M. Higashi, A. V. Marenich, R. M. Olson, A. C. Chamberlin,
J. Pu, C. P. Kelly, J. D. Thompson, J. D. Xidos, J. Li, T. Zhu,
G. D. Hawkins, Y.-Y. Chuang, P. L. Fast, B. J. Lynch,
D. A. Liotard, D. Rinaldi, J. Gao, C. J. Cramer and
D. G. Truhlar, GAMESSPLUS version 2008-2, University of
Minnesota, Minneapolis, 2008.
A. V. Marenich, R. M. Olson, C. P. Kelly, C. J. Cramer and
D. G. Truhlar, J. Chem. Theory Comput., 2007, 3, 2011.
B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy,
M. N. Ringnalda, D. Sitko and B. Honig, J. Phys. Chem., 1996,
100, 11775.
G. Karlstrom, R. Lindh, P.-A. Malmqvist, B. O. Roos, U. Ryde,
V. Veryazov, P. O. Widmark, M. Cossi, B. Schimmelpfennig,
P. Neogrady and L. Seijo, Comput. Mater. Sci., 2003, 28, 222.
Y. Shao, L. F. Molnar, Y. Jung, J. Kussmann, C. Ochsenfeld,
S. T. Brown, A. T. B. Gilbert, L. V. Slipchenko, S. V. Levchenko,
D. P. ONeill, R. A. DiStasio, R. C. Lochan, T. Wang,
G. J. O. Beran, N. A. Besley, J. M. Herbert, C. Y. Lin, T.
Van Voorhis, S. H. Chien, A. Sodt, R. P. Steele, V. A. Rassolov,
P. E. Maslen, P. P. Korambath, R. D. Adamson, B. Austin,
J. Baker, E. F. C. Byrd, H. Dachsel, R. J. Doerksen, A. Dreuw,
B. D. Dunietz, A. D. Dutoi, T. R. Furlani, S. R. Gwaltney,
A. Heyden, S. Hirata, C. P. Hsu, G. Kedziora, R. Z. Khalliulin,
P. Klunzinger, A. M. Lee, M. S. Lee, W. Liang, I. Lotan, N. Nair,
B. Peters, E. I. Proynov, P. A. Pieniazek, Y. M. Rhee, J. Ritchie,
E. Rosta, C. D. Sherrill, A. C. Simmonett, J. E. Subotnik,
H. L. Woodcock, W. Zhang, A. T. Bell, A. K. Chakraborty,
D. M. Chipman, F. J. Keil, A. Warshel, W. J. Hehre,
H. F. Schaefer, J. Kong, A. I. Krylov, P. M. W. Gill and
M. Head-Gordon, Phys. Chem. Chem. Phys., 2006, 8, 3172.
H. L. Woodcock, M. Hodoscek, A. T. B. Gilbert, P. M. W. Gill,
H. F. Schaefer and B. R. Brooks, J. Comput. Chem., 2007, 28, 1485.
R. Ahlrichs, M. Bar, M. Haser, H. Horn and C. Kolmel,
Chem. Phys. Lett., 1989, 162, 165.
T. Helgaker, H. J. Aa. Jensen, P. Jrgensen, J. Olsen, K. Ruud,
H. Agren, A. A. Auer, K. L. Bak, V. Bakken, O. Christiansen,
S. Coriani, P. Dahle, E. K. Dalskov, T. Enevoldsen,
B. Fernandez, C. Hattig, K. Hald, A. Halkier, H. Heiberg,
H. Hettema, D. Jonsson, S. Kirpekar, R. Kobayashi,
H. Koch, K. V. Mikkelsen, P. Norman, M. J. Packer,
T. B. Pedersen, T. A. Ruden, P. Salek, A. Sanchez, T. Saue,
S. P. A. Sauer, B. Schimmelpfennig, K. O. Sylvester-Hvid,
P. R. Taylor and O. Vahtras, DALTON, a molecular
electronic structure program, Release 2.0, 2005, see
http://www.kjemi.uio.no/software/dalton/dalton.html.
B. Jansik, P. Salek, D. Jonsson, O. Vahtras and H. Agren,
J. Chem. Phys., 2005, 122, 054107.

10804 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

362 M. Guidon, F. Schimann, J. Hutter and J. VandeVondele,

J. Chem. Phys., 2008, 128, 214104.
363 H.-J. Werner, P. J. Knowles, R. Lindh, F. R. Manby, M. Schutz,
P. Celani, T. Korona, A. Mitrushenkov, G. Rauhut, T. B. Adler,
R. D. Amos, A. Bernhardsson, A. Berning, D. L. Cooper,
M. J. O. Deegan, A. J. Dobbyn, F. Eckert, E. Goll,
C. Hampel, G. Hetzer, T. Hrenar, G. Knizia, C. Koppl,
Y. Liu, A. W. Lloyd, R. A. Mata, A. J. May,
S. J. McNicholas, W. Meyer, M. E. Mura, A. Nickla,
P. Palmieri, K. Puger, R. Pitzer, M. Reiher, U. Schumann,
H. Stoll, A. J. Stone, R. Tarroni, T. Thorsteinsson, M. Wang
and A. Wolf, MOLPRO 2008.1, University College, Cardi,
2008.
364 F. Neese, ORCA, Universtiy of Bonn, Germany, 2009.
365 J. M. Soler, E. Artacho, J. D. Gale, A. Garc a, J. Junquera,
P. Ordejon and D. Sanchez-Portal, J. Phys.: Condens. Matter,
2002, 14, 2745.
366 D. Sanchez-Portal, P. Ordejon and E. Canadell, Struct. Bonding,
2004, 113, 103.
367 E. Artacho, E. Anglada, O. Dieguez, J. D. Gale, A. Garcia,
J. Junquera, R. M. Martin, P. Ordejon, J. M. Pruneda,
D. Sanchez-Portal and J. M. Soler, J. Phys.: Condens. Matter,
2008, 20, 064208.
368 D. R. Salahub, A. Goursot, J. Weber, A. M. Koster and A. Vela,
in Theory and Applications of Computational Chemistry: The
First Forty Years, ed. C. E. Dykstra, G. Frenking, K. S. Kim and
G. E. Scuseria, Elsevier, Amsterdam, 2005, p. 1079.
369 A. M. Koster, G. Goudtner, T. Heine and A. Vela, ALLCHEM,
Instituto Politecnico Nacional, Mexico, 2001.
370 A. M. Koster, P. Calaminici, Z. Gomez and U. Reveles,
in Reviews of Modern Quantum Chemistry, ed. K. D. Sen,
World Scientic, Singapore, 2002, p. 1439.
371 A. M. Koster, P. Calaminici, M. E. Casida, R. Flores-Moreno,
G. Geudtner, A. Goursot, T. Heine, A. Ipatov, F. Janetzko,
J. M. del Campo, S. Patchkovskii, J. U. Reveles, D. R. Salahub
and A. Vela, deMon2k, 2006 (http://www.demon-software.com/
public_html/), accessed October 12, 2009.
372 M. Ryzhkov, A. Ivanovskii and B. Delley, Theor. Chem. Acc.,
2008, 119, 313.
373 Y. Gao, N. Shao and X. C. Zeng, ACS Nano, 2008, 2, 1497.
374 X. Y. Cui, J. E. Medvedeva, B. Delley, A. J. Freeman and
C. Stamp, Phys. Rev. B: Condens. Matter Mater. Phys., 2008,
78, 245317.
375 M. A. L. Marques, A. Castro, G. F. Bertsch and A. Rubio,
Comput. Phys. Commun., 2003, 151, 60.
376 A. Castro, H. Appel, M. Oliveira, C. A. Rozzi, X. Andrade,
F. Lorenzen, M. A. L. Marques, E. K. U. Gross and A. Rubio,
Phys. Status Solidi B, 2006, 243, 2465.
377 S. Scandolo, P. Giannozzi, C. Cavazzoni, S. de Gironcoli,
A. Pasquarello and S. Baroni, Z. Kristallogr., 2005, 220, 574.
378 X. Wu, A. Selloni and R. Car, Phys. Rev. B: Condens. Matter
Mater. Phys., 2009, 79, 085102.
379 R. Dovesi, R. Orlando, C. Roetti, C. Pisani and V. R. Saunders,
Phys. Status Solidi B, 2000, 217, 63.
380 R. Dovesi, B. Civalleri, R. Orlando, C. Ruetti and
V. R. Saunders, Rev. Comput. Chem., 2005, 21, 1.
381 I. Baraille, A. Loudet, S. Lacombe, H. Cardy and C. Pisani,
THEOCHEM, 2003, 620, 291.
382 D. Munoz, N. M. Harrison and F. Illas, Phys. Rev. B: Condens.
Matter Mater. Phys., 2004, 69, 085115.
383 G. M. Rignanese, X. Gonze, G. Jun, K. Cho and A. Pasquarello,
Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 184301.
384 M. Levy and T. Pagnier, Sens. Actuators, B, 2007, 126, 204.
385 J. W. Zwanziger and M. Torrent, Appl. Magn. Reson., 2008, 33, 447.
386 A. Mellouki, L. Kalarasse, B. Bennecer and F. Kalarasse,
Comput. Mater. Sci., 2009, 44, 876.
387 Y. Imai, M. Mukaida and T. Tsunoda, Thin Solid Films, 2001,
381, 176.
388 J. W. Medlin and M. D. Allendorf, J. Phys. Chem. B, 2003, 107, 217.
389 K. Refson, S.-H. Park and G. Sposito, J. Phys. Chem. B, 2003,
107, 13376.
390 S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. J. Probert,
K. Refson and M. C. Payne, Z. Kristallogr., 2005, 220, 567.
391 M. A. Petersen, S. J. Jenkins and D. A. King, J. Phys. Chem. B,
2004, 108, 5920.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

392 B. Hinnemann and J. K. Nrskov, J. Am. Chem. Soc., 2003, 125,

1466.
393 F. Abild-Pedersen, O. Lytken, J. Engbaek, G. Nielsen,
I. Chorkendor and J. K. Nrskov, Surf. Sci., 2005, 590, 127.
394 L. C. Grabow, A. A. Gokhale, S. T. Evans, J. A. Dumesic and
M. Mavrikakis, J. Phys. Chem. C, 2008, 112, 4608.
395 A. V. Postnikov, G. Bihlmayer and S. Blugel, Comput. Mater.
Sci., 2006, 36, 91.
396 S. Y. Savrasov, Z. Kristallogr., 2005, 220, 555.
397 A. Umetani, E. Nagoshi, T. Kubodera and M. Matoba, Phys. B,
2008, 403, 1356.
398 I. V. Yudanov, A. V. Matveev, K. M. Neyman and N. Rosch,
J. Am. Chem. Soc., 2008, 130, 9342.
399 J. R. Chelikowsky, N. Troullier and Y. Saad, Phys. Rev. Lett.,
1994, 72, 1240.
400 J. L. Fattebert and J. Bernholc, Phys. Rev. B: Condens. Matter
Mater. Phys., 2000, 62, 1713.
401 J. L. Fattebert and F. Gygi, Phys. Rev. B: Condens. Matter
Mater. Phys., 2006, 73, 115124.
402 C. J. Barden, J. C. Rienstra-Kiracofe and H. F. Schaefer, III,
J. Chem. Phys., 2000, 113, 690.
403 C. Diedrich, A. Luchow and S. Grimme, J. Chem. Phys., 2005,
122, 021101.
404 M. C. Holthausen, J. Comput. Chem., 2005, 26, 1505.
405 N. E. Schultz, Y. Zhao and D. G. Truhlar, J. Comput. Chem.,
2008, 29, 185.
406 Y. Zhao and D. G. Truhlar, J. Chem. Phys., 2006, 124, 224105.
407 V. Karttunen, M. Linnolahti, T. Pakkanen, J. Maaranen and
P. Pitkanen, Theor. Chem. Acc., 2007, 118, 899.
408 K. P. Jensen, B. O. Roos and U. Ryde, J. Chem. Phys., 2007,
126, 014103.
409 S. Goel and A. E. Masunov, J. Chem. Phys., 2008, 129, 214302.
410 F. S. Legge, G. L. Nyberg and J. B. Peel, J. Phys. Chem. A, 2001,
105, 7905.
411 F. Wang and L. M. Li, J. Comput. Chem., 2004, 25, 669.
412 Y. Zhao and D. G. Truhlar, Acc. Chem. Res., 2008, 41, 157.
413 J. N. Harvey, Annu. Rep. Prog. Chem., Sect. C, 2006, 102, 203.
414 G. T. de Jong and F. M. Bickelhaupt, J. Chem. Theory Comput.,
2006, 2, 322.
415 M. M. Quintal, A. Karton, M. A. Iron, A. D. Boese and
J. M. L. Martin, J. Phys. Chem. A, 2006, 110, 709.
416 S. Li, J. M. Hennigan, D. A. Dixon and K. A. Peterson, J. Phys.
Chem. A, 2009, 113, 7861.
417 N. J. Mayhall, K. Raghavachari, P. C. Redfern and
L. A. Curtiss, J. Phys. Chem. A, 2009, 113, 5170.
418 Q. Ge, C. Song and L. Wang, Comput. Mater. Sci., 2006, 35, 247.
419 P. Schwerdtfeger, M. Lein, R. P. Krawczyk and C. R. Jacob,
J. Chem. Phys., 2008, 128, 124302.
420 G. T. de Jong, D. P. Geerke, A. Diefenbach and F. Matthias
Bickelhaupt, Chem. Phys., 2005, 313, 261.
421 A. Ikeda, Y. Nakao, H. Sato and S. Sakaki, J. Phys. Chem. A,
2007, 111, 7124.
422 S. Li and D. A. Dixon, J. Phys. Chem. A, 2007, 111, 11908.
423 S. Zhao, Z.-H. Li, W.-N. Wang, Z.-P. Liu, K.-N. Fan, Y. Xie
and H. F. Schaefer III, J. Chem. Phys., 2006, 124, 184102.
424 F. Stevens, I. Carmichael, F. Callens and M. Waroquier,
J. Phys. Chem. A, 2006, 110, 4846.
425 J. Rogal, K. Reuter and M. Scheer, Phys. Rev. B: Condens.
Matter Mater. Phys., 2007, 75, 205433.
426 P. Song, W. Guan, C. Yao, Z. Su, Z. Wu, J. Feng and L. Yan,
Theor. Chem. Acc., 2007, 117, 407.
427 J. S. Sears and C. D. Sherrill, J. Phys. Chem. A, 2008, 112, 6741.
428 J. S. Sears and C. D. Sherrill, J. Phys. Chem. A, 2008, 112, 3466.
429 A. Ghosh, E. Gonzalez, E. Tangen and B. O. Roos, J. Phys.
Chem. A, 2008, 112, 12792.
430 A. Ghosh and P. R. Taylor, Curr. Opin. Chem. Biol., 2003, 7, 113.
431 A. Sorkin, D. G. Truhlar and E. A. Amin, J. Chem. Theory
Comput., 5, 1254.
432 J. Paier, M. Marsman and G. Kresse, J. Chem. Phys., 2007, 127,
024103.
433 A. Stroppa and G. Kresse, New J. Phys., 2008, 10, 063020.
434 A. Matveev, M. Staufer, M. Mayer and N. Rosch, Int. J.
Quantum Chem., 1999, 75, 863.
435 R. J. Deeth and N. Fey, J. Comput. Chem., 2004, 25, 1840.

This journal is


c

the Owner Societies 2009

436 G. I. Csonka, J. P. Perdew, A. Ruzsinszky, P. H. T. Philipsen,

S. Lebegue, J. Paier, O. A. Vydrov and J. G. Angyan, Phys. Rev.
B: Condens. Matter Mater. Phys., 2009, 79, 155107.
437 P. Haas, F. Tran and P. Blaha, Phys. Rev. B: Condens. Matter
Mater. Phys., 2009, 79, 085104.
438 M. Ropo, K. Kokko and L. Vitos, Phys. Rev. B: Condens.
Matter Mater. Phys., 2008, 77, 195445.
439 F. Tran, R. Laskowski, P. Blaha and K. Schwarz, Phys. Rev. B:
Condens. Matter Mater. Phys., 2007, 75, 115131.
440 D. Rinaldo, L. Tian, J. N. Harvey and R. A. Friesner, J. Chem.
Phys., 2008, 129, 164108.
441 M. Buhl, C. Reimann, D. A. Pantazis, T. Bredow and F. Neese,
J. Chem. Theory Comput., 2008, 4, 1449.
442 J. Handzlik, Chem. Phys. Lett., 2009, 469, 140.
443 C. Mosch, C. Koukounas, N. Bacalis, A. Metropoulos, A. Gross
and A. Mavridis, J. Phys. Chem. C, 2008, 112, 6924.
444 J. J. Zheng, Y. Zhao and D. G. Truhlar, J. Chem. Theory
Comput., 2009, 5, 808.
445 M. Reiher, Faraday Discuss., 2007, 135, 97.
446 W. Heisenberg, Z. Phys., 1928, 49, 619.
447 P. A. M. Dirac, Proc. R. Soc. London, Ser. A, 1929, 123, 714.
448 J. H. Van Vleck, Rev. Mod. Phys., 1945, 17, 27.
449 J. C. Slater, Rev. Mod. Phys., 1953, 25, 199.
450 I. Cioni and C. A. Daul, Coord. Chem. Rev., 2003, 238239, 187.
451 J. N. Harvey, Struct. Bonding, 2004, 112, 151.
452 F. Neese, Coord. Chem. Rev., 2009, 253, 526.
453 L. Noodleman and E. R. Davidson, Chem. Phys., 1986, 109, 131.
454 F. Neese, J. Phys. Chem. Solids, 2004, 65, 781.
455 E. R. Davidson and A. E. Clark, Int. J. Quantum Chem., 2005, 103, 1.
456 S. Pittalis, S. Kurth and E. K. U. Gross, J. Chem. Phys., 2006,
125, 084105.
457 T. Ziegler, A. Rauk and E. J. Baerends, Theor. Chim. Acta, 1977,
43, 261.
458 L. Noodleman, J. Chem. Phys., 1981, 74, 5737.
459 A. J. Cohen, P. Mori-Sanchez and W. Yang, Science, 2008, 321, 792.
460 J.-M. Mouesca, J. L. Chen, L. Noodleman, D. Bashford and
D. A. Case, J. Am. Chem. Soc., 1994, 116, 11898.
461 S. Sinnecker, F. Neese, L. Noodleman and W. Lubitz, J. Am.
Chem. Soc., 2004, 126, 2613.
462 E. Ruiz, J. Cano, S. Alvarez and P. Alemany, J. Comput. Chem.,
1999, 20, 1391.
463 K. Yamaguchi, F. Jensen, A. Dorigo and K. N. Houk, Chem.
Phys. Lett., 1988, 149, 537.
464 T. Soda, Y. Kitagawa, T. Onishi, Y. Takano, Y. Shigeta,
H. Nagao, Y. Yoshioka and K. Yamaguchi, Chem. Phys. Lett.,
2000, 319, 223.
465 M. Shoji, K. Koizumi, Y. Kitagawa, T. Kawakami,
S. Yamanaka, M. Okumura and K. Yamaguchi, Chem. Phys.
Lett., 2006, 432, 343.
466 J. Pople, P. Gill and N. Handy, Int. J. Quantum Chem., 1995, 56,
303.
467 J. Baker, A. Scheiner and J. Andzelm, Chem. Phys. Lett., 1993,
216, 380.
468 J. Wang, L. A. Eriksson, R. J. Boyd, Z. Shi and B. G. Johnson,
J. Phys. Chem., 1994, 98, 1844.
469 C. J. Cramer, F. J. Dulles and D. E. Falvey, J. Am. Chem. Soc.,
1994, 116, 9787.
470 C. J. Cramer, F. J. Dulles, J. W. Storer and S. E. Worthington,
Chem. Phys. Lett., 1994, 218, 387.
471 C. J. Cramer, F. J. Dulles, D. J. Giesen and J. Almlof,
Chem. Phys. Lett., 1995, 245, 165.
472 J. Wang, A. D. Becke and J. V. H. Smith, J. Chem. Phys., 1995,
102, 3477.
473 J. Grafenstein, E. Kraka, M. Filatov and D. Cremer, Int. J. Mol.
Sci., 2002, 3, 360.
474 D. Cremer, M. Filatov, V. Polo, E. Kraka and S. Shaik, Int. J.
Mol. Sci., 2002, 3, 604.
475 J. Grafenstein and D. Cremer, Mol. Phys., 2001, 99, 981.
476 V. N. Staroverov and E. R. Davidson, Chem. Phys. Lett., 2001,
340, 142.
477 H. Carmen, Y. Lian and R. Markus, J. Comput. Chem., 2006,
27, 1223.
478 R. K. Szilagyi and M. A. Winslow, J. Comput. Chem., 2006, 27, 1385.
479 A. Sorkin, M. A. Iron and D. G. Truhlar, J. Chem. Theory
Comput., 2008, 4, 307.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10805

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

480 R. Valero, R. Costa, I. Moreira, D. G. Truhlar and F. Illas,

J. Chem. Phys., 2008, 128, 114103.
481 C. Adamo, V. Barone, A. Bencini, F. Totti and I. Cioni,
Inorg. Chem., 1999, 38, 1996.
482 E. Ruiz, S. Alvarez, J. Cano and V. Polo, J. Chem. Phys., 2005,
123, 164110.
483 C. Adamo, V. Barone, A. Bencini, R. Broer, M. Filatov,
N. M. Harrison, F. Illas, J. P. Malrieu and I. D. R. Moreira,
J. Chem. Phys., 2006, 124, 107101.
484 J. L. Lewin, D. E. Heppner and C. J. Cramer, JBIC, J. Biol.
Inorg. Chem., 2007, 12, 1221.
485 C. J. Cramer, A. Kinal, M. Wloch, P. Piecuch and L. Gagliardi,
J. Phys. Chem. A, 2006, 110, 11557.
486 C. J. Cramer, M. Wloch, P. Piecuch, C. Puzzarini and
L. Gagliardi, J. Phys. Chem. A, 2006, 110, 1991.
487 C. J. Cramer and W. B. Tolman, Acc. Chem. Res., 2007, 40, 601.
488 S. M. Huber, M. Z. Ertem, F. Aquilante, L. Gagliardi,
W. B. Tolman and C. J. Cramer, Chem.Eur. J., 2009, 15, 4886.
489 P. A. Malmqvist, K. Pierloot, A. R. Moughal Shahi, C. J. Cramer
and L. Gagliardi, J. Chem. Phys., 2008, 128, 204109.
490 P. Rivero, I. D. R. Moreira, F. Illas and G. E. Scuseria, J. Chem.
Phys., 2008, 129, 184110.
491 A. Roth, J. Becher, C. Herrmann, H. Gorls, G. Vaughan,
M. Reiher, D. Klemm and W. Plass, Inorg. Chem., 2006, 45, 10066.
492 N. Marom and L. Kronik, Appl. Phys. A: Mater. Sci. Process.,
2009, 95, 165.
493 E. R. Batista and R. L. Martin, J. Am. Chem. Soc., 2007, 129,
7224.
494 X. Yang and M. H. Baik, J. Am. Chem. Soc., 2004, 126, 13222.
495 X. Yang and M. H. Baik, J. Am. Chem. Soc., 2006, 128, 7476.
496 T. R. Weaver, T. J. Meyer, S. A. Adeyemi, G. M. Brown,
R. P. Eckberg, W. E. Hateld, E. C. Johnson, R. W. Murray
and D. Untereker, J. Am. Chem. Soc., 1975, 97, 3039.
497 K. Pierloot and S. Vancoillie, J. Chem. Phys., 2008, 128, 034104.
498 M. Swart, J. Chem. Theory Comput., 2008, 4, 2057.
499 C. Rong, S. Lian, D. Yin, B. Shen, A. Zhong, L. Bartolotti and
S. Liu, J. Chem. Phys., 2006, 125, 174102.
500 J. L. Carreon-Macedo and J. N. Harvey, Phys. Chem. Chem.
Phys., 2006, 8, 93.
501 J. Song, E. Apr, Y. G. Khait, M. R. Homann and
K. Kowalski, Chem. Phys. Lett., 2006, 428, 277.
502 F. Neese, J. Inorg. Biochem., 2006, 100, 716.
503 J. M. Matxain, E. Rezabal, X. Lopez, J. M. Ugalde and
L. Gagliardi, J. Chem. Phys., 2008, 128, 194315.
504 C. Herrmann, L. Yu and M. Reiher, J. Comput. Chem., 2006, 27,
1223.
505 C. Herrmann, M. Reiher and B. A. Hess, J. Chem. Phys., 2005,
122, 034102.
506 A. E. Clark and E. R. Davidson, J. Chem. Phys., 2001, 115, 7382.
507 I. Mayer, Chem. Phys. Lett., 2007, 440, 357.
508 O. Hubner, K. Fink and W. Klopper, Phys. Chem. Chem. Phys.,
2007, 9, 1911.
509 M. Podewitz, C. Herrmann, A. Malassa, M. Westerhausen and
M. Reiher, Chem. Phys. Lett., 2008, 451, 301.
510 P. H. Dederichs, S. Blugel, R. Zeller and H. Akai, Phys. Rev.
Lett., 1984, 53, 2512.
511 Q. Wu and T. V. Voorhis, Phys. Rev. A: At., Mol., Opt. Phys.,
2005, 72, 024502.
512 Q. Wu, C.-L. Cheng and T. Van Voorhis, J. Chem. Phys., 2007,
127, 164119.
513 S. Diey, D. Beljonne and T. Van Voorhis, J. Am. Chem. Soc.,
2008, 130, 3420.
514 I. Rudra, Q. Wu and T. Van Voorhis, Inorg. Chem., 2007, 46,
10539.
515 J. R. Schmidt, N. Shenvi and J. C. Tully, J. Chem. Phys., 2008,
129, 114110.
516 H. Fliegl, K. Fink, W. Klopper, C. E. Anson, A. K. Powell and
R. Clerac, Phys. Chem. Chem. Phys., 2009, 11, 3900.
517 I. Cioni, F. Illas and C. Adamo, J. Chem. Phys., 2004, 120, 3811.
518 M. Filatov and S. Shaik, Chem. Phys. Lett., 1999, 304, 429.
519 F. Illas, I. D. R. Moreira, J. M. Boll and M. Filatov, Theor.
Chem. Acc., 2006, 116, 587.
520 A. Kazaryan, J. Heuver and M. Filatov, J. Phys. Chem. A, 2008,
112, 12980.

10806 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

521 I. d. P. R. Moreira, R. Costa, M. Filatov and F. Illas, J. Chem.

Theory Comput., 2007, 3, 764.
522 A. J. Perez-Jimenez, J. M. Perez-Jorda, I. d. P. R. Moreira and
F. Illas, J. Comput. Chem., 2007, 28, 2559.
523 T. Ukai, K. Nakata, S. Yamanaka, T. Takada and
K. Yamaguchi, Mol. Phys., 2007, 105, 2667.
524 Y. H. Shao, M. Head-Gordon and A. I. Krylov, J. Chem. Phys.,
2003, 118, 4807.
525 A. de la Lande, V. Moliner and O. Parisel, J. Chem. Phys., 2007,
126, 035102.
526 Y. M. Rhee and M. Head-Gordon, J. Am. Chem. Soc., 2008,
130, 3878.
527 F. Wang and T. Ziegler, J. Chem. Phys., 2004, 121, 12191.
528 F. Wang and T. Ziegler, J. Chem. Phys., 2005, 122, 074109.
529 F. Wang and T. Ziegler, Int. J. Quantum Chem., 2006, 106, 2545.
530 O. Vahtras and Z. Rinkevicius, J. Chem. Phys., 2007, 126, 114101.
531 A. V. Soldatova, J. Kim, A. Rosa, G. Ricciardi, M. E. Kenney
and M. A. J. Rodgers, Inorg. Chem., 2008, 47, 4275.
532 J. E. Peralta and V. Barone, J. Chem. Phys., 2008, 129, 194107.
533 T. M. Alam, J. S. Clawson, F. o. Bonhomme, S. G. Thoma,
M. A. Rodriguez, S. Zheng and J. Autschbach, Chem. Mater.,
2008, 20, 2205.
534 J. Autschbach and S. Zheng, Magn. Reson. Chem., 2008, 46, S45.
535 J. M. Morbec and K. Capelle, Int. J. Quantum Chem., 2008, 108,
2433.
536 P. Hrobarik, O. L. Malkina, V. G. Malkin and M. Kaupp,
Chem. Phys., 2009, 356, 229.
537 E. Runge and E. K. U. Gross, Phys. Rev. Lett., 1984, 52, 997.
538 R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett., 1996, 256,
454.
539 R. Van Leeuwen, Int. J. Mod. Phys. B, 2001, 15, 1969.
540 K. Burke, M. Petersilka and E. K. U. Gross, in Recent Advances
in Density Functional Methods, ed. P. Fantucci and A. Bencini,
World Scientic, Singapore, 2002, vol. 3, p. 67.
541 R. J. Deeth, Faraday Discuss., 2003, 124, 379.
542 M. A. L. Marques and E. K. U. Gross, Annu. Rev. Phys. Chem.,
2004, 55, 427.
543 A. Dreuw and M. Head-Gordon, Chem. Rev., 2005, 105, 4009.
544 P. Elliot, F. Furche and K. Burke, Rev. Comput. Chem., 2009, 26, 91.
545 N. C. Handy, Mol. Phys., 2004, 102, 2399.
546 V. Arcisauskaite, J. Kongsted, T. Hansen and K. V. Mikkelsen,
Chem. Phys. Lett., 2009, 470, 285.
547 L. Salassa, C. Garino, G. Salassa, R. Gobetto and C. Nervi,
J. Am. Chem. Soc., 2008, 130, 9590.
548 J. F. Guillemoles, V. Barone, L. Joubert and C. Adamo, J. Phys.
Chem. A, 2002, 106, 11354.
549 P. P. Laine, I. Cioni, P. Ochsenbein, E. Amouyal, C. Adamo
and F. Bedioui, Chem.Eur. J., 2005, 11, 3711.
550 M. K. Nazeeruddin, F. De Angelis, S. Fantacci, A. Selloni,
G. Viscardi, P. Liska, S. Ito, B. Takeru and M. Gratzel, J. Am.
Chem. Soc., 2005, 127, 16835.
551 P. P. Laine, F. Loiseau, S. Campagna, I. Cioni and C. Adamo,
Inorg. Chem., 2006, 45, 5538.
552 M. F. Charlot and A. Aukauloo, J. Phys. Chem. A, 2007, 111,
11661.
553 M. Marcaccio, F. Paolucci, C. Fontanesi, G. Fioravanti and
S. Zanarini, Inorg. Chim. Acta, 2007, 360, 1154.
554 M. Abrahamsson, M. Jager, R. J. Kumar, T. Osterman,
P. Persson, H. C. Becker, O. Johansson and L. Hammarstrom,
J. Am. Chem. Soc., 2008, 130, 15533.
555 T. A. Jackson, J. U. Rohde, M. S. Seo, C. V. Sastri, R. DeHont,
A. Stubna, T. Ohta, T. Kitagawa, E. Munck, W. Nam and
L. Que, J. Am. Chem. Soc., 2008, 130, 12394.
556 M. Hutin, C. J. Cramer, L. Gagliardi, A. R. Moughal Shahi,
G. Bernardinelli, R. Cerny and J. R. Nitschke, J. Am. Chem.
Soc., 2007, 129, 8774.
557 D. Schultz, F. Biaso, A. R. M. Shahi, M. Georoy, K. Rissanen,
L. Gagliardi, C. J. Cramer and J. R. Nitschke, Chem.Eur. J.,
2008, 14, 7180.
558 I. Bar-Nahum, A. K. Gupta, S. M. Huber, M. Z. Ertem,
C. J. Cramer and W. B. Tolman, J. Am. Chem. Soc., 2009,
131, 2812.
559 A. Vlcek Jr and S. Zalis , Coord. Chem. Rev., 2007, 251, 258.
560 J. Fan, M. Seth, J. Autschbach and T. Ziegler, Inorg. Chem.,
2008, 47, 11656.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

561 G. A. Peralta, M. Seth, H. Zhekova and T. Ziegler,

Inorg. Chem., 2008, 47, 4185.
562 P. R. T. Schipper, O. V. Gritsenko, S. J. A. van Gisbergen and
E. J. Baerends, J. Chem. Phys., 2000, 112, 1344.
563 A. Kunishita, J. Teraoka, J. D. Scanlon, T. Matsumoto,
M. Suzuki, C. J. Cramer and S. Itoh, J. Am. Chem. Soc.,
2007, 129, 7248.
564 M. Stener, A. Nardelli, R. De Francesco and G. Fronzoni,
J. Phys. Chem. C, 2007, 111, 11862.
565 M. Stener, A. Nardelli and G. Fronzoni, J. Chem. Phys., 2008,
128, 134307.
566 M. Stener, A. Nardelli and G. Fronzoni, Chem. Phys. Lett.,
2008, 462, 358.
567 M. Caricato, B. Mennucci, J. Tomasi, F. Ingrosso, R. Cammi,
S. Corni and G. Scalmani, J. Chem. Phys., 2006, 124, 124520.
568 G. Scalmani, M. J. Frisch, B. Mennucci, J. Tomasi, R. Cammi
and V. Barone, J. Chem. Phys., 2006, 124, 094107.
569 B. Mennucci, Theor. Chem. Acc., 2006, 116, 31.
570 F. De Angelis, S. Fantacci and A. Sgamellotti, Theor. Chem.
Acc., 2007, 117, 1093.
571 M. J. Lundqvist, M. Nilsing, S. Lunell, B. Akermark and
P. Persson, J. Phys. Chem. B, 2006, 110, 20513.
572 P. Persson, M. J. Lundqvist, R. Ernstorfer, W. A. Goddard and
F. Willig, J. Chem. Theory Comput., 2006, 2, 441.
573 M. Nilsing, P. Persson, S. Lunell and L. Ojamae, J. Phys. Chem.
C, 2007, 111, 12116.
574 F. De Angelis, S. Fantacci, A. Selloni, M. Gratzel and
M. K. Nazeeruddin, Nano Lett., 2007, 7, 3189.
575 S. G. Abuabara, L. G. C. Rego and V. S. Batista, J. Am. Chem.
Soc., 2005, 127, 18234.
576 A. Ramirez-Solis, R. Poteau and J. P. Daudey, J. Chem. Phys.,
2006, 124, 034307.
577 A. L. Tenderholt, R. K. Szilagyi, R. H. Holm, K. O. Hodgson,
B. Hedman and E. I. Solomon, Inorg. Chem., 2008, 47, 6382.
578 A. Decker, J. U. Rohde, E. J. Klinker, S. D. Wong, L. Que and
E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 15983.
579 R. Sarangi, J. T. York, M. E. Helton, K. Fujisawa, K. D. Karlin,
W. B. Tolman, K. O. Hodgson, B. Hedman and E. I. Solomon,
J. Am. Chem. Soc., 2008, 130, 676.
580 R. Sarangi, S. DeBeer George, D. J. Rudd, R. K. Szilagyi,
X. Ribas, C. Rovira, M. Almeida, K. O. Hodgson, B. Hedman
and E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 2316.
581 M. Casarin, P. Finetti, A. Vittadini, F. Wang and T. Ziegler,
J. Phys. Chem. A, 2007, 111, 5270.
582 S. A. Kozimor, P. Yang, E. R. Batista, K. S. Boland, C. J. Burns,
C. N. Christensen, D. L. Clark, S. D. Conradson, P. J. Hay,
J. S. Lezama, R. L. Martin, D. E. Schwarz, M. P. Wilkerson and
L. E. Wolfsberg, Inorg. Chem., 2008, 47, 5365.
583 G. Fronzoni, M. Stener, P. Decleva, F. Wang, T. Ziegler, E. van
Lenthe and E. J. Baerends, Chem. Phys. Lett., 2005, 416, 56.
584 G. Fronzoni, M. Stener, A. Reduce and P. Decleva, J. Phys.
Chem. A, 2004, 108, 8467.
585 K. Ray, S. D. George, E. I. Solomon, K. Wieghardt and
F. Neese, Chem.Eur. J., 2007, 13, 2783.
586 S. DeBeer George, T. Petrenko and F. Neese, J. Phys. Chem. A,
2008, 112, 12936.
587 J. F. Berry, S. D. George and F. Neese, Phys. Chem. Chem.
Phys., 2008, 10, 4361.
588 J. F. Berry, E. Bill, E. Bothe, S. D. George, B. Mienert, F. Neese
and K. Wieghardt, Science, 2006, 312, 1937.
589 R. Sarangi, S. I. Gorelsky, L. Basumallick, H. J. Hwang,
R. C. Pratt, T. D. P. Stack, Y. Lu, K. O. Hodgson, B. Hedman
and E. I. Solomon, J. Am. Chem. Soc., 2008, 130, 3866.
590 C.-L. Cheng, Q. Wu and T. Van Voorhis, J. Chem. Phys., 2008,
129, 124112.
591 A. T. B. Gilbert, N. A. Besley and P. M. W. Gill, J. Phys. Chem.
A, 2008, 112, 13164.
592 A. Castro, M. A. L. Marques, A. H. Romero, M. J. T. Oliveira
and A. Rubio, J. Chem. Phys., 2008, 129, 144110.
593 X.-Y. Yang, Y.-C. Wang, Z.-Y. Geng and Z.-Y. Liu, Chem.
Phys. Lett., 2006, 430, 265.
594 S. Shaik, H. Hirao and D. Kumar, Acc. Chem. Res., 2007, 40,
532.
595 K. Morokuma, Bull. Chem. Soc. Jpn., 2007, 80, 2247.

This journal is


c

the Owner Societies 2009

596 N. Hebben, H.-J. Himmel, G. Eickerling, C. Herrmann,

M. Reiher, V. Herz, M. Presnitz and W. Scherer, Chem.Eur.
J., 2007, 13, 10078.
597 R. Herges and A. Papalippopoulos, Angew. Chem., Int. Ed.,
2001, 40, 4671.
598 B. D. Alexander and T. J. Dines, J. Phys. Chem. A, 2004, 108, 146.
599 V. W. Jurgensen and K. Jalkanen, Phys. Biol., 2006, 3, S63.
600 S. Sinnecker and F. Neese, Top. Curr. Chem., 2007, 268, 47.
601 C. Herrmann and M. Reiher, Top. Curr. Chem., 2007, 268, 85.
602 H. Sato, T. Taniguchi, K. Monde, S.-I. Nishimura and
A. Yamagishi, Chem. Lett., 2006, 35, 364.
603 S. Luber and M. Reiher, Chem. Phys., 2008, 346, 212.
604 B. B. Johnson and W. L. Peticolas, Annu. Rev. Phys. Chem.,
1976, 27, 465.
605 A. Warshel, Annu. Rev. Biophys. Bioeng., 1977, 6, 273.
606 R. J. H. Clark and B. Stewart, Struct. Bonding, 1979, 36, 1.
607 F. Neese, T. Petrenko, D. Ganyushin and G. Olbrich, Coord.
Chem. Rev., 2007, 251, 288.
608 C. Herrmann, J. Neugebauer, M. Presselt, U. Uhlemann,
M. Schmitt, S. Rau, J. Popp and M. Reiher, J. Phys. Chem. B,
2007, 111, 6078.
609 T. Petrenko, K. Ray, K. E. Wieghardt and F. Neese, J. Am.
Chem. Soc., 2006, 128, 4422.
610 B. V. Popp, J. E. Wendlandt, C. R. Landis and S. S. Stahl,
Angew. Chem., Int. Ed., 2007, 46, 601.
611 V. V. Smirnov, D. W. Brinkley, M. P. Lanci, K. D. Karlin and
J. P. Roth, J. Mol. Catal. A: Chem., 2006, 251, 100.
612 M. P. Lanci and J. P. Roth, J. Am. Chem. Soc., 2006, 128, 16006.
613 J. P. Roth and J. P. Klinman, in Isotope Eects in Chemistry and
Biology, ed. A. Kohen and H.-H. Limbach, CRC Press,
Boca Raton, 2006, p. 645.
614 M. P. Lanci, V. V. Smirnov, C. J. Cramer, E. V. Gauchenova,
J. Sundermeyer and J. P. Roth, J. Am. Chem. Soc., 2007, 129, 14697.
615 J. P. Roth and C. J. Cramer, J. Am. Chem. Soc., 2008, 130, 7802.
616 G. Brehm, M. Reiher, B. Le Guennic, M. Leibold, S. Schindler,
F. W. Heinemann and S. Schneider, J. Raman Spectrosc., 2006,
37, 108.
617 P. Carbonniere, I. Cioni, C. Adamo and C. Pouchan, Chem.
Phys. Lett., 2006, 429, 52.
618 B. Insuasty, C. Atienza, C. Seoane, N. Martin, J. Garin,
J. Orduna, R. Alcala and B. Villacampa, J. Org. Chem., 2004,
69, 6986.
619 L. Fang, G. Yang, Y. Qiu and Z. Su, Theor. Chem. Acc., 2008,
119, 329.
620 O. Rubio-Pons, Y. Luo and H. Agren, J. Chem. Phys., 2006,
124, 094310.
621 T. Ziegler and J. Autschbach, Chem. Rev., 2005, 105, 2695.
622 E. R. Davidson, Chem. Rev., 2000, 100, 351.
623 E. C. Sherer and C. J. Cramer, Organometallics, 2003, 22, 1682.
624 N. Barros, O. Eisenstein, L. Maron and T. D. Tilley,
Organometallics, 2006, 25, 5699.
625 N. Barros, O. Eisenstein and L. Maron, Dalton Trans., 2006, 3052.
626 N. L. Woodrum and C. J. Cramer, Organometallics, 2006, 25, 68.
627 J. L. Lewin, N. L. Woodrum and C. J. Cramer, Organometallics,
2006, 25, 5906.
628 L. Perrin, O. Eisenstein and L. Maron, New J. Chem., 2007, 31, 549.
629 P. A. Hunt, Dalton Trans., 2007, 1743.
630 V. L. Cruz, S. Martinez, J. Martinez-Salazar and J. Sancho,
Macromolecules, 2007, 40, 7413.
631 M. Zimmermann, F. Estler, E. Herdtweck, K. W. Tornroos and
R. Anwander, Organometallics, 2007, 26, 6029.
632 N. Barros, O. Eisenstein, L. Maron and T. D. Tilley,
Organometallics, 2008, 27, 2252.
633 B. A. Vastine and M. B. Hall, J. Am. Chem. Soc., 2007, 129,
12068.
634 R. F. W. Bader, Chem. Rev., 1991, 91, 893.
635 C. S. Tredget, E. Clot and P. Mountford, Organometallics, 2008,
27, 3458.
636 J. Scott, H. Fan, B. F. Wicker, A. R. Fout, M.-H. Baik and
D. J. Mindiola, J. Am. Chem. Soc., 2008, 130, 14438.
637 Y. Luo, J. Baldamus, O. Tardif and Z. Hou, Organometallics,
2005, 24, 4362.
638 M. Lein, J. Frunzke, A. Timoshkin and G. Frenking,
Chem.Eur. J., 2001, 7, 4155.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10807

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

639 E. Urnezius, W. W. Brennessel, C. J. Cramer, J. E. Ellis and

P. v. R. Schleyer, Science, 2002, 295, 832.
640 M. Lein, J. Frunzke and G. Frenking, Inorg. Chem., 2003, 42, 2504.
641 J. M. Mercero, E. Formoso, J. M. Matxain, L. A. Eriksson and
J. M. Ugalde, Chem.Eur. J., 2006, 12, 4495.
642 N. Ochi, Y. Nakao, H. Sato and S. Sakaki, J. Am. Chem. Soc.,
2007, 129, 8615.
643 X.-f. Sheng, G.-f. Zhao and L.-l. Zhi, J. Phys. Chem. C, 2008,
112, 17828.
644 J. Joubert, F. Delbecq, C. Thieuleux, M. Taouk, F. Blanc,
C. Coperet, J. Thivolle-Cazat, J.-M. Basset and P. Sautet,
Organometallics, 2007, 26, 3329.
645 S. K. Mandal, P. M. Gurubasavaraj, H. W. Roesky, G. Schwab,
D. Stalke, R. B. Oswald and V. Dolle, Inorg. Chem., 2007, 46, 10158.
646 J. L. Lewin and C. J. Cramer, J. Phys. Chem. A, 2008, 112,
12754.
647 T. Wondimagegn, D. Wang, A. Razavi and T. Ziegler, Organometallics, 2008, 27, 6434.
648 T. Wondimagegn, D. Wang, A. Razavi and T. Ziegler, Organometallics, 2009, 28, 1383.
649 F. Dong, S. Heinbuch, Y. Xie, J. J. Rocca, E. R. Bernstein,
Z.-C. Wang, K. Deng and S.-G. He, J. Am. Chem. Soc., 2008,
130, 1932.
650 F. Dong, S. Heinbuch, Y. Xie, E. R. Bernstein, J. J. Rocca,
Z.-C. Wang, X.-L. Ding and S.-G. He, J. Am. Chem. Soc., 2009,
131, 1057.
651 Y.-P. Ma, W. Xue, Z.-C. Wang, M.-F. Ge and S.-G. He, J. Phys.
Chem. A, 2008, 112, 3731.
652 L. Gracia, V. Polo, J. R. Sambrano and J. Andres, J. Phys.
Chem. A, 2008, 112, 1808.
653 A. Sadoc, S. Messaoudi, E. Furet, R. Gautier, E. Le Fur,
L. le Polles and J.-Y. Pivan, Inorg. Chem., 2007, 46, 4835.
654 K. J. Ooms, S. E. Bolte, J. J. Smee, B. Baruah, D. C. Crans and
T. Polenova, Inorg. Chem., 2007, 46, 9285.
655 L. L. G. Justino, M. L. Ramos, F. Nogueira, A. J. F. N. Sobral,
C. F. G. C. Geraldes, M. Kaupp, H. D. Burrows, C. Fiolhais
and V. M. S. Gil, Inorg. Chem., 2008, 47, 7317.
656 M. Michelini, I. Rivalta and E. Sicilia, Theor. Chem. Acc., 2008,
120, 395.
657 R. o. Arteaga-Muller, J. Sanchez-Nieves, J. Ramos, P. Royo and
M. E. G. Mosquera, Organometallics, 2008, 27, 1417.
658 A. E. Ashley, R. T. Cooper, G. G. Wildgoose, J. C. Green and
D. OHare, J. Am. Chem. Soc., 2008, 130, 15662.
659 G. Balazs, F. G. N. Cloke, L. Gagliardi, J. C. Green,
A. Harrison, P. B. Hitchcock, A. R. Moughal Shahi and
O. T. Summerscales, Organometallics, 2008, 27, 2013.
660 F. Ferrante, L. Gagliardi, B. E. Bursten and A. P. Sattelberger,
Inorg. Chem., 2005, 44, 8476.
661 M. Brynda, L. Gagliardi, P. O. Widmark, P. P. Power and
B. O. Roos, Angew. Chem., Int. Ed., 2006, 45, 3804.
662 B. O. Roos, A. C. Borin and L. Gagliardi, Angew. Chem., Int.
Ed., 2007, 46, 1469.
663 I. Infante, L. Gagliardi and G. E. Scuseria, J. Am. Chem. Soc.,
2008, 130, 7459.
664 G. La Macchia, L. Gagliardi, P. P. Power and M. Brynda,
J. Am. Chem. Soc., 2008, 130, 5104.
665 M. Brynda, L. Gagliardi and B. O. Roos, Chem. Phys. Lett.,
2009, 471, 1.
666 K. Andersson, B. O. Roos, P. A. Malmqvist and P. O. Widmark,
Chem. Phys. Lett., 1994, 230, 391.
667 C. W. Bauschlicher and H. Partridge, Chem. Phys. Lett., 1994,
231, 277.
668 H. Stoll and H. J. Werner, Mol. Phys., 1996, 88, 793.
669 K. E. Edgecombe and A. D. Becke, Chem. Phys. Lett., 1995, 244,
427.
670 A. Poater, X. Solans-Monfort, E. Clot, C. Coperet and
O. Eisenstein, J. Am. Chem. Soc., 2007, 129, 8207.
671 R. Haunschild and G. Frenking, J. Organomet. Chem., 2008,
693, 737.
672 G. F. Caramori and G. Frenking, Theor. Chem. Acc., 2008, 120, 351.
673 L. Yan, X. Lopez, J. J. Carbo, R. Sniatynsky, D. C. Duncan and
J. M. Poblet, J. Am. Chem. Soc., 2008, 130, 8223.
674 S. Schenk, B. Le Guennic, B. Kirchner and M. Reiher,
Inorg. Chem., 2008, 47, 3634.

10808 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

675 D. V. Khoroshun, D. G. Musaev and K. Morokuma,

Mol. Phys., 2002, 100, 523.
676 Z. X. Cao, Z. H. Zhou, H. L. Wan and Q. N. Zhang, Int. J.
Quantum Chem., 2005, 103, 344.
677 M. Reiher, B. Le Guennic and B. Kirchner, Inorg. Chem., 2005,
44, 9640.
678 D. C. Graham, G. J. O. Beran, M. Head-Gordon, G. Christian,
R. Stranger and B. F. Yates, J. Phys. Chem. A, 2005, 109, 6762.
679 M. Holscher and W. Leitner, Eur. J. Inorg. Chem., 2006, 4407.
680 F. Studt and F. Tuczek, J. Comput. Chem., 2006, 27, 1278.
681 D. V. Khoroshun, D. G. Musaev and K. Morokuma, J. Comput.
Chem., 2007, 28, 423.
682 A. Magistrato, A. Robertazzi and P. Carloni, J. Chem. Theory
Comput., 2007, 3, 1708.
683 G. Christian, R. Stranger and B. F. Yates, Chem.Eur. J., 2009,
15, 646.
684 R. R. Schrock, Angew. Chem., Int. Ed., 2008, 47, 5512.
685 G. C. Stephan, C. Sivasankar, F. Studt and F. Tuczek,
Chem.Eur. J., 2008, 14, 644.
686 I. Dance, Chem.Asian J., 2007, 2, 936.
687 T. Leyssens, D. Peeters, A. G. Orpen and J. N. Harvey, New J.
Chem., 2005, 29, 1424.
688 P. Geerlings, F. D. Proft and W. Langenaeker, Chem. Rev.,
2003, 103, 1793.
689 J. M. Matxain, M. Piris, E. Formoso, J. M. Mercero, X. Lopez
and J. M. Ugalde, ChemPhysChem, 2007, 8, 2096.
690 J. M. Matxain, E. Formoso, J. M. Mercero, M. Piris, X. Lopez
and J. M. Ugalde, Chem.Eur. J., 2008, 14, 8547.
691 K. Yamaguchi, S. Yamanaka, H. Isobe, M. Shoji, K. Koizumi,
Y. Kitagawa, T. Kawakami and M. Okumura, Polyhedron,
2007, 26, 2216.
692 D. Balcells, C. Raynaud, R. H. Crabtree and O. Eisenstein,
Inorg. Chem., 2008, 47, 10090.
693 G. Balazs, F. G. N. Cloke, A. Harrison, P. B. Hitchcock,
J. Green and O. T. Summerscales, Chem. Commun., 2007, 873.
694 A. Krapp, M. Lein and G. Frenking, Theor. Chem. Acc., 2008,
120, 313.
695 H.-M. Jia, D.-C. Fang, Y. Feng, J.-Y. Zhang, W.-B. Fan and
L. Zhu, Theor. Chem. Acc., 2008, 121, 271.
696 R. Haunschild and G. Frenking, J. Organomet. Chem., 2008,
693, 3627.
697 A. Ghosh, J. Almlof and L. Que, J. Phys. Chem., 1994, 98, 5576.
698 K. Yoshizawa, T. Ohta, T. Yamabe and R. Homann, J. Am.
Chem. Soc., 1997, 119, 12311.
699 E. I. Solomon, T. C. Brunold, M. I. Davis, J. N. Kemsley,
S. K. Lee, N. Lehnert, F. Neese, A. J. Skulan, Y. S. Yang and
J. Zhou, Chem. Rev., 2000, 100, 235.
700 E. I. Solomon, A. Decker and N. Lehnert, Proc. Natl. Acad. Sci.
U. S. A., 2003, 100, 3589.
701 A. Ghosh, E. Tangen, H. Ryeng and P. R. Taylor, Eur. J. Inorg.
Chem., 2004, 4555.
702 M. M. Abu-Omar, A. Loaiza and N. Hontzeas, Chem. Rev.,
2005, 105, 2227.
703 A. Decker and E. I. Solomon, Curr. Opin. Chem. Biol., 2005, 9, 152.
704 S. V. Kryatov, E. V. Rybak-Akimova and S. Schindler,
Chem. Rev., 2005, 105, 2175.
705 A. Bassan, T. Borowski, C. J. Schoeld and P. E. M. Siegbahn,
Chem.Eur. J., 2006, 12, 8835.
706 M. J. Park, J. Lee, Y. Suh, J. Kim and W. Nam, J. Am. Chem.
Soc., 2006, 128, 2630.
707 C. D. Brown, M. L. Neidig, M. B. Neibergall, J. D. Lipscomb
and E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 7427.
708 W. H. Harman and C. J. Chang, J. Am. Chem. Soc., 2007, 129,
15128.
709 W. Nam, Acc. Chem. Res., 2007, 40, 522.
710 M. L. Neidig, C. D. Brown, K. M. Light, D. G. Fujimori,
E. M. Nolan, J. C. Price, E. W. Barr, J. M. Bollinger, C. Krebs,
C. T. Walsh and E. I. Solomon, J. Am. Chem. Soc., 2007, 129,
14224.
711 M. Y. M. Pau, M. I. Davis, A. M. Orville, J. D. Lipscomb and
E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 1944.
712 L. Que, Acc. Chem. Res., 2007, 40, 493.
713 C. B. Bell, S. D. Wong, Y. M. Xiao, E. J. Klinker, A. L. Tenderholt,
M. C. Smith, J. U. Rohde, L. Que, S. P. Cramer and E. I. Solomon,
Angew. Chem., Int. Ed., 2008, 47, 9071.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

714 N. Muresan, C. C. Lu, M. Ghosh, J. C. Peters, M. Abe,

L. M. Henling, T. Weyhermoller, E. Bill and K. Wieghardt,
Inorg. Chem., 2008, 47, 4579.
715 M. Radon and K. Pierloot, J. Phys. Chem. A, 2008, 112, 11824.
716 N. Strickland and J. N. Harvey, J. Phys. Chem. B, 2007, 111,
841.
717 M. P. Mehn, S. D. Brown, T. K. Paine, W. W. Brennessel,
C. J. Cramer, J. C. Peters and L. Que, Dalton Trans., 2006, 1347.
718 S. Shaik, S. P. de Visser, F. Ogliaro, H. Schwarz and
D. Schroder, Curr. Opin. Chem. Biol., 2002, 6, 556.
719 A. Altun, S. Shaik and W. Thiel, J. Am. Chem. Soc., 2007, 129,
8978.
720 E. Derat, S. Shaik, C. Rovira, P. Vidossich and M. AlfonsoPrito, J. Am. Chem. Soc., 2007, 129, 6346.
721 C. Hazan, D. Kumar, S. P. de Visser and S. Shaik, Eur. J. Inorg.
Chem., 2007, 2966.
722 C. Li, L. Zhang, C. Zhang, H. Hirao, W. Wu and S. Shaik,
Angew. Chem., Int. Ed., 2007, 46, 8168.
723 C. V. Sastri, J. Lee, K. Oh, Y. J. Lee, J. Lee, T. A. Jackson,
K. Ray, H. Hirao, W. Shin, J. A. Halfen, J. Kim, L. Que,
S. Shaik and W. Nam, Proc. Natl. Acad. Sci. U. S. A., 2007, 104,
19181.
724 S. Shaik, H. Hirao and D. Kumar, Nat. Prod. Rep., 2007, 24,
533.
725 Y. Wang, D. Kumar, C. L. Yang, K. L. Han and S. Shaik,
J. Phys. Chem. B, 2007, 111, 7700.
726 H. Chen, H. Hirao, E. Derat, I. Schlichting and S. Shaik,
J. Phys. Chem. B, 2008, 112, 9490.
727 H. Chen, Y. Moreau, E. Derat and S. Shaik, J. Am. Chem. Soc.,
2008, 130, 1953.
728 K. B. Cho, H. Hirao, H. Chen, M. A. Carvajal, S. Cohen,
E. Derat, W. Thiel and S. Shaik, J. Phys. Chem. A, 2008, 112,
13128.
729 S. N. Dhuri, M. S. Seo, Y. M. Lee, H. Hirao, Y. Wang, W. Nam
and S. Shaik, Angew. Chem., Int. Ed., 2008, 47, 3356.
730 H. Hirao, L. Que, W. Nam and S. Shaik, Chem.Eur. J., 2008,
14, 1740.
731 S. Shaik, D. Kumar and S. P. de Visser, J. Am. Chem. Soc., 2008,
130, 10128.
732 D. Q. Wang, J. J. Zheng, S. Shaik and W. Thiel, J. Phys. Chem.
B, 2008, 112, 5126.
733 K. B. Cho, M. A. Carvajal and S. Shaik, J. Phys. Chem. B, 2009,
113, 336.
734 E. J. Klinker, S. Shaik, H. Hirao and L. Que, Angew. Chem., Int.
Ed., 2009, 48, 1291.
735 N. M. Reilly, J. U. Reveles, G. E. Johnson, J. M. del Campo,
S. N. Khanna, A. M. Koster and A. W. Castleman, J. Phys.
Chem. C, 2007, 111, 19086.
736 Q. Sun, Reddy M. Marquez, P. Jena, C. Gonzalez and Q. Wang,
J. Phys. Chem. C, 2007, 111, 4159.
737 J. K. Schwartz, P.-p. Wei, K. H. Mitchell, B. G. Fox and
E. I. Solomon, J. Am. Chem. Soc., 2008, 130, 7098.
738 D. Rinaldo, D. M. Philipp, S. J. Lippard and R. A. Friesner,
J. Am. Chem. Soc., 2007, 129, 3135.
739 W. G. Han, T. Q. Liu, T. Lovell and L. Noodleman, J. Comput.
Chem., 2006, 27, 1292.
740 W. G. Han and L. Noodleman, Inorg. Chem., 2008, 47, 2975.
741 W. G. Han and L. Noodleman, Inorg. Chim. Acta, 2008, 361,
973.
742 M. Torrent, D. G. Musaev, H. Basch and K. Morokuma,
J. Comput. Chem., 2002, 23, 59.
743 N. Mitic, M. D. Clay, L. Saleh, J. M. Bollinger and
E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 9049.
744 W.-G. Han, T. Liu, T. Lovell and L. Noodleman, Inorg. Chem.,
2006, 45, 8533.
745 L. Noodleman and D. A. Case, Adv. Inorg. Chem., 1992, 38, 423.
746 L. Noodleman, C. Y. Peng, D. A. Case and J.-M. Mouesca,
Coord. Chem. Rev., 1995, 144, 199.
747 E. Ruiz, A. Rodriguez-Fortea, J. Cano, S. Alvarez and
P. Alemany, J. Comput. Chem., 2003, 24, 982.
748 R. A. Torres, T. Lovell, L. Noodleman and D. A. Case, J. Am.
Chem. Soc., 2003, 125, 1923.
749 M. Shoji, K. Koizumi, Y. Kitagawa, S. Yamanaka,
M. Okumura and K. Yamaguchi, Int. J. Quantum Chem.,
2007, 107, 609.

This journal is


c

the Owner Societies 2009

750 S. Niu and T. Ichiye, Theor. Chem. Acc., 2007, 117, 275.
751 L. Noodleman, T. Lovell, T. Liu, F. Himo and R. A. Torres,
Curr. Opin. Chem. Biol., 2002, 6, 259.
752 G. Moritz, M. Reiher and B. A. Hess, Theor. Chem. Acc., 2005,
114, 76.
753 K. P. Jensen, B. L. Ooi and H. E. M. Christensen, J. Phys.
Chem. A, 2008, 112, 12829.
754 Z. H. Li, Y. Gong, K. N. Fan and M. F. Zhou, J. Phys. Chem.
A, 2008, 112, 13641.
755 M. H. Baik and R. A. Friesner, J. Phys. Chem. A, 2002, 106,
7407.
756 M. Uudsemaa and T. Tamm, J. Phys. Chem. A, 2003, 107, 9997.
757 J. Blumberger, L. Bernasconi, I. Tavernelli, R. Vuilleumier and
M. Sprik, J. Am. Chem. Soc., 2004, 126, 3928.
758 J. Blumberger and M. Sprik, J. Phys. Chem. B, 2004, 108, 6529.
759 J. Blumberger, Y. Tateyama and M. Sprik, Comput. Phys.
Commun., 2005, 169, 256.
760 Y. Tateyama, J. Blumberger, M. Sprik and I. Tavernelli,
J. Chem. Phys., 2005, 122, 234505.
761 J. Blumberger and M. Sprik, Theor. Chem. Acc., 2006, 115, 113.
762 P. Jaque, A. V. Marenich, C. J. Cramer and D. G. Truhlar,
J. Phys. Chem. C, 2007, 111, 5783.
763 I. Chiorescu, D. V. Deubel, V. B. Arion and B. K. Keppler,
J. Chem. Theory Comput., 2008, 4, 499.
764 P. Winget, E. J. Weber, C. J. Cramer and D. G. Truhlar, Phys.
Chem. Chem. Phys., 2000, 2, 1231.
765 E. V. Patterson, C. J. Cramer and D. G. Truhlar, J. Am. Chem.
Soc., 2001, 123, 2025.
766 C. Fontanesi, R. Benassi, R. Giovanardi, M. Marcaccio,
F. Paolucci and S. Roa, J. Mol. Struct., 2002, 612, 277.
767 A. Lewis, J. A. Bumpus, D. G. Truhlar and C. J. Cramer,
J. Chem. Ed., 2004, 81, 596; erratum, 2007, 84, 934.
768 P. Winget, C. J. Cramer and D. G. Truhlar, Theor. Chem. Acc.,
2004, 112, 217.
769 S. Bhattacharyya, M. T. Stankovich, D. G. Truhlar and
J. L. Gao, J. Phys. Chem. A, 2007, 111, 5729.
770 J. L. Hodgson, M. Namazian, S. E. Bottle and M. L. Coote,
J. Phys. Chem. A, 2007, 111, 13595.
771 J. Moens, P. Jaque, F. De Proft and P. Geerlings, J. Phys. Chem.
A, 2008, 112, 6023.
772 C. Kritayakornupong and S. Hannongbua, Chem. Phys., 2007,
332, 95.
773 C. Kritayakornupong, Chem. Phys. Lett., 2007, 441, 226.
774 S. Amira, D. Spangberg, V. Zelin, M. Probst and
K. Hermansson, J. Phys. Chem. B, 2005, 109, 14235.
775 H. A. De Abreu, L. Guimaraes and H. A. Duarte, J. Phys.
Chem. A, 2006, 110, 7713.
776 D. Schroder, S. O. Souvi and E. Alikhani, Chem. Phys. Lett.,
2009, 470, 162.
777 J. Moens, G. Roos, P. Jaque, F. D. Proft and P. Geerlings,
Chem.Eur. J., 2007, 13, 9331.
778 R. D. Leib, W. A. Donald, J. T. OBrien, M. F. Bush and
E. R. Williams, J. Am. Chem. Soc., 2007, 129, 7716.
779 W. A. Donald, R. D. Leib, J. T. OBrien, M. F. Bush and
E. R. Williams, J. Am. Chem. Soc., 2008, 130, 3371.
780 S. Trasatti, Pure Appl. Chem., 1986, 58, 955.
781 M. D. Tissandier, K. A. Cowen, W. Y. Feng, E. Gundlach,
M. H. Cohen, A. D. Earhart, J. V. Coe and T. R. Tuttle, J. Phys.
Chem. A, 1998, 102, 7787.
782 D. M. Camaioni and C. A. Schwerdtfeger, J. Phys. Chem. A,
2005, 109, 10795.
783 C. P. Kelly, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B,
2006, 110, 16066.
784 C. P. Kelly, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B,
2007, 111, 408.
785 W. R. Fawcett, Langmuir, 2008, 24, 9868.
786 A. Correa and L. Cavallo, J. Am. Chem. Soc., 2006, 128, 13352.
787 C. E. Webster, J. Am. Chem. Soc., 2007, 129, 7490.
788 B. F. Straub, Adv. Synth. Catal., 2007, 349, 204.
789 Y. Zhao and D. G. Truhlar, Org. Lett., 2007, 9, 1967.
790 J. Mathew, N. Koga and C. H. Suresh, Organometallics, 2008,
27, 4666.
791 D. Benitez, E. Tkatchouk and W. A. Goddard, Chem. Commun.,
2008, 6194.
792 R. Tonner and G. Frenking, Chem. Commun., 2008, 1584.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10809

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

793 Y. Zhao and D. G. Truhlar, J. Chem. Theory Comput., 2009, 5,

324.
794 D. Benitez, E. Tkatchouk and W. A. Goddard, Organometallics,
2009, 28, 2643.
795 C. W. Bielawski and R. H. Grubbs, Prog. Polym. Sci., 2007, 32,
1.
796 A. C. Tsipis, A. G. Orpen and J. N. Harvey, Dalton Trans., 2005,
2849.
797 M. S. Sanford, J. A. Love and R. H. Grubbs, J. Am. Chem. Soc.,
2001, 123, 6543.
798 S. Torker, D. Merki and P. Chen, J. Am. Chem. Soc., 2008, 130,
4808.
799 S. Grimme, Angew. Chem., Int. Ed., 2006, 45, 4460.
800 Y. Zhao and D. G. Truhlar, Org. Lett., 2006, 8, 5753.
801 I. C. Stewart, D. Benitez, D. J. OLeary, E. Tkatchouk,
M. W. Day, W. A. Goddard and R. H. Grubbs, J. Am. Chem.
Soc., 2009, 131, 1931.
802 S. Pandian, I. H. Hillier, M. A. Vincent, N. A. Burton,
I. W. Ashworth, D. J. Nelson, J. M. Percy and G. Rinaudo,
Chem. Phys. Lett., 2009, 476, 37.
803 N. Sieert and M. Buhl, Inorg. Chem., 2009, 48, 4622.
804 G. F. Caramori and G. Frenking, Organometallics, 2007, 26,
5815.
805 A. Krapp, K. K. Pandey and G. Frenking, J. Am. Chem. Soc.,
2007, 129, 7596.
806 F. Liu, T. Cardolaccia, B. J. Hornstein, J. R. Schoonover and
T. J. Meyer, J. Am. Chem. Soc., 2007, 129, 2446.
807 T. A. Betley, Y. Surendranath, M. V. Childress, G. E. Alliger,
R. Fu, C. C. Cummins and D. G. Nocera, Philos. Trans. R. Soc.
London, Ser. B, 2008, 363, 1293.
808 T. A. Betley, Q. Wu, T. Van Voorhis and D. G. Nocera,
Inorg. Chem., 2008, 47, 1849.
809 C. W. Cady, R. H. Crabtree and G. W. Brudvig, Coord. Chem.
Rev., 2008, 252, 444.
810 J. L. Cape and J. K. Hurst, J. Am. Chem. Soc., 2008, 130, 827.
811 J. J. Concepcion, J. W. Jurss, J. L. Templeton and T. J. Meyer,
Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 17632.
812 Y. V. Geletii, B. Botar, P. Kgerler, D. A. Hillesheim,
D. G. Musaev and C. L. Hill, Angew. Chem., Int. Ed., 2008,
47, 3896.
813 N. A. Ketterer, H. J. Fan, K. J. Blackmore, X. F. Yang,
J. W. Ziller, M. H. Baik and A. F. Heyduk, J. Am. Chem.
Soc., 2008, 130, 4364.
814 F. Liu, J. J. Concepcion, J. W. Jurss, T. Cardolaccia,
J. L. Templeton and T. J. Meyer, Inorg. Chem., 2008, 47, 1727.
815 J. T. Muckerman, D. E. Polyansky, T. Wada, K. Tanaka and
E. Fujita, Inorg. Chem., 2008, 47, 1787.
816 E. Poverenov, I. Efremenko, A. I. Frenkel, Y. Ben-David,
L. J. W. Shimon, G. Leitus, L. Konstantinovski, J. M. L.
Martin and D. Milstein, Nature, 2008, 455, 1093.
817 Y. L. Pushkar, J. Yano, K. Sauer, A. Boussac and
V. K. Yachandra, Proc. Natl. Acad. Sci. U. S. A., 2008, 105,
1879.
818 (a) I. Romero, M. Rodriguez, C. Sens, J. Mola, M. R. Kollipara,
L. Francas, E. Mas-Marza, L. Escriche and A. Llobet, Inorg.
Chem., 2008, 47, 1824; (b) F. Bozoglian, S. Romain,
M. Z. Ertem, T. K. Todorova, C. Sens, J. Mola,
M. Rodriguez, I. Romero, J. Benet-Buchholz, X. Fontrodona,
C. J. Cramer, L. Gagliardi and A. Llobet, J. Am. Chem. Soc.,
DOI: 10.1021/ja9036127.
819 W. M. C. Sameera and J. E. McGrady, Dalton Trans., 2008,
6141.
820 J. M. Saveant, Chem. Rev., 2008, 108, 2348.
821 C. Elschenbroich, J. Six, K. Harms, G. Frenking and
G. Heydenrych, Eur. J. Inorg. Chem., 2008, 3303.
822 J.-P. Zhang, X. Zhou, T. Liu, F.-Q. Bai, H.-X. Zhang and
A.-C. Tang, Theor. Chem. Acc., 2008, 121, 123.
823 J.-P. Zhang, X. Zhou, F.-Q. Bai, H.-X. Zhang and A.-C. Tang,
Theor. Chem. Acc., 2009, 122, 31.
824 Vitamin B12 and B12-Proteins, ed. B. Krautler, D. Arigoni and
B. T. Golding, Wiley-VCH, Berlin, 1998.
825 G. Glod, W. Angst, C. Holliger and R. P. Schwarzenbach,
Environ. Sci. Technol., 1997, 31, 253.
826 A. D. Follett, K. A. McNabb, A. A. Peterson, J. D. Scanlon,
C. J. Cramer and K. McNeill, Inorg. Chem., 2007, 46, 1645.

10810 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

827 D. A. Pratt and W. A. van der Donk, J. Am. Chem. Soc., 2005,
127, 384.
828 R. L. Birke, Q. Huang, T. Spataru and D. K. Gosser, J. Am.
Chem. Soc., 2006, 128, 1922.
829 K. P. Jensen and U. Ryde, J. Phys. Chem. A, 2003, 107, 7539.
830 K. P. Jensen and U. Ryde, Coord. Chem. Rev., 2009, 253, 769.
831 P. Morschel, J. Janikowski, G. Hilt and G. Frenking, J. Am.
Chem. Soc., 2008, 130, 8952.
832 K.-W. Huang, J. H. Han, C. B. Musgrave and E. Fujita,
Organometallics, 2007, 26, 508.
833 Z. Hu and R. J. Boyd, J. Chem. Phys., 2000, 113, 9393.
834 J. M. Praetorius, D. P. Allen, R. Y. Wang, J. D. Webb, F. Grein,
P. Kennepohl and C. M. Crudden, J. Am. Chem. Soc., 2008, 130,
3724.
835 P. Mauleon, I. Alonso, M. R. Rivero and J. C. Carretero, J. Org.
Chem., 2007, 72, 9924.
836 P. Fristrup, M. Kreis, A. Palmelund, P.-O. Norrby and
R. Madsen, J. Am. Chem. Soc., 2008, 130, 5206.
837 P. Liu, P. H.-Y. Cheong, Z.-X. Yu, P. A. Wender and
K. N. Houk, Angew. Chem., Int. Ed., 2008, 47, 3939.
838 L. Orian, W.-J. v. Zeist and F. M. Bickelhaupt, Organometallics,
2008, 27, 4028.
839 Q. Zhao, T. Cao, F. Li, X. Li, H. Jing, T. Yi and C. Huang,
Organometallics, 2007, 26, 2077.
840 D. Di Censo, S. Fantacci, F. De Angelis, C. Klein, N. Evans,
K. Kalyanasundaram, H. J. Bolink, M. Gratzel and
M. K. Nazeeruddin, Inorg. Chem., 2008, 47, 980.
841 T. Liu, B.-H. Xia, X. Zhou, H.-X. Zhang, Q.-J. Pan and
J.-S. Gao, Organometallics, 2007, 26, 143.
842 S. J. Lee, K.-M. Park, K. Yang and Y. Kang, Inorg. Chem.,
2009, 48, 1030.
843 T. Liu, B.-H. Xia, X. Zhou, Q.-C. Zheng, Q.-J. Pan and
H.-X. Zhang, Theor. Chem. Acc., 2008, 121, 155.
844 M. Kimura and Y. Uozumi, J. Org. Chem., 2007, 72, 707.
845 E. Calimano and T. D. Tilley, J. Am. Chem. Soc., 2008, 130,
9226.
846 D. H. Ess, S. M. Bischof, J. Oxgaard, R. A. Periana and
W. A. Goddard, Organometallics, 2008, 27, 6440.
847 R. Ghosh, T. J. Emge, K. Krogh-Jespersen and A. S. Goldman,
J. Am. Chem. Soc., 2008, 130, 11317.
848 N. Marom and L. Kronik, Appl. Phys. A: Mater. Sci. Process.,
2009, 95, 159.
849 P. Calaminici, J. Chem. Phys., 2008, 128, 164317.
850 G. L. Arvizu and P. Calaminici, J. Chem. Phys., 2007, 126,
194102.
851 M. Shoji, K. Koizumi, T. Hamamoto, Y. Kitagawa,
S. Yamanaka, M. Okumura and K. Yamaguchi, Chem. Phys.
Lett., 2006, 421, 483.
852 M. Shoji, Y. Kitagawa, T. Kawakami, S. Yamanaka,
M. Okumura and K. Yamaguchi, J. Phys. Chem. A, 2008, 112,
4020.
853 J. Conradie, T. Wondimagegn and A. Ghosh, J. Phys. Chem. B,
2008, 112, 1053.
854 S. Kozuch, C. Amatore, A. Jutand and S. Shaik, Organometallics,
2005, 24, 2319.
855 G. T. de Jong and F. M. Bickelhaupt, ChemPhysChem, 2007, 8,
1170.
856 F. Vines, F. Illas and K. M. Neyman, J. Phys. Chem. A, 2008,
112, 8911.
857 F. Vines, F. Illas and K. M. Neyman, Angew. Chem., Int. Ed.,
2007, 46, 7094.
858 Z. Shi, S. Wu and J. A. Szpunar, Nanotechnology, 2006, 17,
2161.
859 A. Betz, L. Yu, M. Reiher, A.-C. Gaumont, P.-A. Jares and
M. Gulea, J. Organomet. Chem., 2008, 693, 2499.
860 C. Y. Legault, Y. Garcia, C. A. Merlic and K. N. Houk, J. Am.
Chem. Soc., 2007, 129, 12664.
861 F. E. Hahn, Angew. Chem., Int. Ed., 2006, 45, 1348.
862 F. E. Hahn and M. C. Jahnke, Angew. Chem., Int. Ed., 2008, 47,
3122.
863 U. Radius and F. M. Bickelhaupt, Organometallics, 2008, 27,
3410.
864 T. Ziegler and A. Rauk, Inorg. Chem., 1979, 18, 1755.
865 L. Gagliardi and C. J. Cramer, Inorg. Chem., 2006, 45, 9442.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

866 H. Nakai, X. Hu, L. N. Zakharov, A. L. Rheingold and

K. Meyer, Inorg. Chem., 2004, 43, 855.
867 H. Jacobsen, A. Correa, A. Poater, C. Costabile and L. Cavallo,
Coord. Chem. Rev., 2009, 253, 687.
868 U. Radius and F. M. Bickelhaupt, Coord. Chem. Rev., 2009, 253,
678.
869 R. Tonner, G. Heydenrych and G. Frenking, Chem.Asian J.,
2007, 2, 1555.
870 J. Harmer, C. Finazzo, R. Piskorski, S. Ebner, E. C. Duin,
M. Goenrich, R. K. Thauer, M. Reiher, A. Schweiger,
D. Hinderberger and B. Jaun, J. Am. Chem. Soc., 2008, 130,
10907.
871 N. Yang, M. Reiher, M. Wang, J. Harmer and E. C. Duin,
J. Am. Chem. Soc., 2007, 129, 11028.
872 G. T. de Jong and F. M. Bickelhaupt, J. Chem. Theory Comput.,
2007, 3, 514.
873 R. Fazaeli, A. Ariafard, S. Jamshidi, E. S. Tabatabaie and
K. A. Pishro, J. Organomet. Chem., 2007, 692, 3984.
874 A. Ariafard and B. F. Yates, J. Organomet. Chem., 2009, 694,
2075.
875 V. P. Ananikov, D. G. Musaev and K. Morokuma, Eur. J.
Inorg. Chem., 2007, 5390.
876 R. D. Adams, B. Captain, C. Beddie and M. B. Hall, J. Am.
Chem. Soc., 2007, 129, 986.
877 A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard and
W. M. Ski, J. Am. Chem. Soc., 1992, 114, 10024.
878 J. Raber, C. B. Zhu and L. A. Eriksson, J. Phys. Chem. B, 2005,
109, 11006.
879 D. V. Deubel, J. Am. Chem. Soc., 2006, 128, 1654.
880 J. K. C. Lau and D. V. Deubel, J. Chem. Theory Comput., 2006,
2, 103.
881 T. Song and P. Hu, J. Chem. Phys., 2006, 125, 091101.
882 Y. Mantri, S. J. Lippard and M. H. Baik, J. Am. Chem. Soc.,
2007, 129, 5023.
883 T. Matsui, Y. Shigeta and K. Hirao, J. Phys. Chem. B, 2007,
111, 1176.
884 S. P. Oldeld, M. D. Hall and J. A. Platts, J. Med. Chem., 2007,
50, 5227.
885 S. M. Fiuza, A. M. Amado, M. P. M. Marques and L. A. E.
Batista de Carvalho, J. Phys. Chem. A, 2008, 112, 3253.
886 P. D. Dans, A. Crespo, D. o. A. Estrin and E. L. Coitino,
J. Chem. Theory Comput., 2008, 4, 740.
887 T. van der Wijst, C. Fonseca Guerra, M. Swart,
F. M. Bickelhaupt and B. Lippert, Chem.Eur. J., 2009, 15, 209.
888 A. Datta, D. A. Hrovat and W. T. Borden, J. Am. Chem. Soc.,
2008, 130, 2726.
889 J. Espinosa-Garc a, J. C. Corchado and D. G. Truhlar, J. Am.
Chem. Soc., 1997, 119, 9891.
890 R. J. Mishur, C. Zheng, T. M. Gilbert and R. N. Bose, Inorg.
Chem., 2008, 47, 7972.
891 Y. Mo and E. Kaxiras, Small, 2007, 3, 1253.
892 A. Roldan, F. Vines, F. Illas, J. Ricart and K. Neyman, Theor.
Chem. Acc., 2008, 120, 565.
893 P. Calaminici, A. M. Koster and Z. Gomez-Sandoval, J. Chem.
Theory Comput., 2007, 3, 905.
894 K. Jug, B. Zimmermann, P. Calaminici and A. M. Koster,
J. Chem. Phys., 2002, 116, 4497.
895 D. G. Leopold, J. Ho and W. C. Lineberger, J. Chem. Phys.,
1987, 86, 1715.
896 J. U. Reveles and A. M. Koster, J. Comput. Chem., 2004, 25,
1109.
897 F.-Q. Shi, X. Li, Y. Xia, L. Zhang and Z.-X. Yu, J. Am. Chem.
Soc., 2007, 129, 15503.
898 H.
Garcia,
http://pubs.acs.org/JACSbeta/jvi/issue3.html.
Accessed October 12, 2009.
899 P. Pyykko, Inorg. Chim. Acta, 2005, 358, 4113.
900 P. Pyykko, Chem. Soc. Rev., 2008, 37, 1967.
901 F. Furche, R. Ahlrichs, P. Weis, C. Jacob, S. Gilb, T. Bierweiler
and M. M. Kappes, J. Chem. Phys., 2002, 117, 6982.
902 H. Hakkinen, B. Yoon, U. Landman, X. Li, H.-J. Zhai and
L.-S. Wang, J. Phys. Chem. A, 2003, 107, 6168.
903 X. Xing, B. Yoon, U. Landman and J. H. Parks, Phys. Rev. B:
Condens. Matter Mater. Phys., 2006, 74, 165423.
904 M. P. Johansson, A. Lechtken, D. Schooss, M. M. Kappes and
F. Furche, Phys. Rev. A: At., Mol., Opt. Phys., 2008, 77, 053202.

This journal is


c

the Owner Societies 2009

905 V. N. Staroverov, G. E. Scuseria, J. Tao and J. P. Perdew,

Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 075102.
906 R. M. Olson, S. Varganov, M. S. Gordon, H. Metiu, S. Chretien,
P. Piecuch, K. Kowalski, S. A. Kucharski and M. Musial, J. Am.
Chem. Soc., 2005, 127, 1049.
907 M. Mantina, R. Valero and D. G. Truhlar, J. Chem. Phys., 2009,
131, 064706.
908 L. Ferrighi, B. Hammer and G. K. H. Madsen, J. Am. Chem.
Soc., 2009, 131, 10605.
909 L. Xiao and L. Wang, Chem. Phys. Lett., 2004, 392, 452.
910 L. Barrio, P. Liu, J. A. Rodriguez, J. M. Campos-Martin and
J. L. G. Fierro, J. Chem. Phys., 2006, 125, 164715.
911 B. Yoon, P. Koskinen, B. Huber, O. Kostko, B. v. Issendor,
H. Hakkinen, M. Moseler and U. Landman, ChemPhysChem,
2007, 8, 157.
912 E. S. Kryachko and F. Remacle, Int. J. Quantum Chem., 2007,
107, 2922.
913 M. Sterrer, T. Risse, U. M. Pozzoni, L. Giordano, M. Heyde,
H.-P. Rust, G. Pacchioni and H.-J. Freund, Phys. Rev. Lett.,
2007, 98, 096107.
914 S. Bulusu, X. Li, L.-S. Wang and X. C. Zeng, J. Phys. Chem. C,
2007, 111, 4190.
915 W. Fa and J. Dong, J. Chem. Phys., 2006, 124, 114310.
916 A. Lechtken, D. Schooss, J. R. Stairs, M. N. Blom, F. Furche,
N. Morgner, O. Kostko, B. v. Issendor and M. M. Kappes,
Angew. Chem., Int. Ed., 2007, 46, 2944.
917 X. Gu, S. Bulusu, X. Li, X. C. Zeng, J. Li, X. G. Gong and
L.-S. Wang, J. Phys. Chem. C, 2007, 111, 8228.
918 J. Li, X. Li, H.-J. Zhai and L.-S. Wang, Science, 2003, 299, 864.
919 S. Bulusu, X. Li, L.-S. Wang and X. C. Zeng, Proc. Natl. Acad.
Sci. U. S. A., 2006, 103, 8326.
920 W. Huang, S. Bulusu, R. Pal, X. C. Zeng and L.-S. Wang,
ACS Nano, 2009, 3, 1225.
921 D. Y. Zubarev and A. I. Boldyrev, J. Phys. Chem. A, 2009, 113,
866.
922 M. B. Torres, E. M. Fernandez and L. C. Balbas, J. Phys. Chem.
A, 2008, 112, 6678.
923 A. Prestianni, A. Martorana, F. Labat, I. Cioni and C. Adamo,
J. Phys. Chem. B, 2006, 110, 12240.
924 A. Prestianni, A. Martorana, I. Cioni, F. Labat and C. Adamo,
J. Phys. Chem. C, 2008, 112, 18061.
925 A. Prestianni, A. Martorana, F. Labat, I. Cioni and C.
Adamo, THEOCHEM, 2009, 903, 34.
926 S. Chretien, M. S. Gordon and H. Metiu, J. Chem. Phys., 2004,
121, 3756.
927 S. Chretien, M. S. Gordon and H. Metiu, J. Chem. Phys., 2004,
121, 9931.
928 Z. H. Li, A. W. Jasper and D. G. Truhlar, J. Am. Chem. Soc.,
2007, 129, 14899.
929 J. Botticelli, R. Fournier and M. Zhang, Theor. Chem. Acc.,
2008, 120, 583.
930 S. Chretien and H. Metiu, J. Chem. Phys., 2007, 127, 084704.
931 M. Nolan, V. S. Verdugo and H. Metiu, Surf. Sci., 2008, 602,
2734.
932 H. Y. Kim, H. M. Lee, R. G. S. Pala, V. Shapovalov and
H. Metiu, J. Phys. Chem. C, 2008, 112, 12398.
933 N. Lopez, T. V. W. Janssens, B. S. Clausen, Y. Xu,
M. Mavrikakis, T. Bligaard and J. K. Nrskov, J. Catal.,
2004, 223, 232; T. Janssens, B. Clausen, B. Hvolbk,
H. Falsig, C. Christensen, T. Bligaard and J. Nrskov,
Top. Catal., 2007, 44, 15.
934 Y. Chen, P. Hu, M.-H. Lee and H. Wang, Surf. Sci., 2008, 602,
1736.
935 A. A. Rusakov, E. Rykova, G. E. Scuseria and A. Zaitsevskii,
J. Chem. Phys., 2007, 127, 164322.
936 J. S. Golightly, L. Gao, A. W. Castleman, D. E. Bergeron,
J. W. Hudgens, R. J. Magyar and C. A. Gonzalez, J. Phys.
Chem. C, 2007, 111, 14625.
937 N. R. M. Crawford, C. Schrodt, A. Rothenberger, W. Shi and
R. Ahlrichs, Chem.Eur. J., 2008, 14, 319.
938 Y. Li, G. Galli and F. o. Gygi, ACS Nano, 2008, 2, 1896.
939 E. A. Lewis and W. B. Tolman, Chem. Rev., 2004, 104, 1047.
940 S. Itoh, Curr. Opin. Chem. Biol., 2006, 10, 115.
941 M. Rol and F. Tuczek, Angew. Chem., Int. Ed., 2008, 47, 2344.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10811

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

942 B. F. Gherman and C. J. Cramer, Coord. Chem. Rev., 2009, 253,

723.
943 T. Punniyamurthy, S. Velusamy and J. Iqbal, Chem. Rev., 2005,
105, 2329.
944 B. Isabel, M. A. Carrondo and F. L. Peter, JBIC, J. Biol. Inorg.
Chem., 2006, 11, 539.
945 J. P. Klinman, J. Biol. Chem., 2006, 281, 3013.
946 J. M. Bollinger and C. Krebs, Curr. Opin. Chem. Biol., 2007, 11,
151.
947 M. Guell, J. M. Luis, L. Rodriguez-Santiago, M. Sodupe and
M. Sola`, J. Phys. Chem. A, 2009, 113, 1308.
948 J.-y. Hasegawa, K. Pierloot and B. O. Roos, Chem. Phys. Lett.,
2001, 335, 503.
949 C. J. Cramer, J. R. Gour, A. Kinal, M. Wtoch, P. Piecuch,
A. R. Moughal Shahi and L. Gagliardi, J. Phys. Chem. A, 2008,
112, 3754.
950 P. Piecuch and M. W"och, J. Chem. Phys., 2005, 123, 224105.
951 P. Piecuch, M. W"och, J. R. Gour and A. Kinal, Chem. Phys.
Lett., 2006, 418, 467.
952 M. W"och, M. Lodriguito, P. Piecuch and J. R. Gour,
Mol. Phys., 2006, 104, 2149.
953 M. W"och, J. R. Gour and P. Piecuch, J. Phys. Chem. A, 2007,
111, 11359.
954 K. Andersson, P.-A. Malmqvist, B. O. Roos, A. J. Sadlej and
K. Wolinski, J. Phys. Chem., 1990, 94, 5483.
955 M. Schatz, V. Raab, S. P. Foxon, G. Brehm, S. Schneider,
M. Reiher, M. C. Holthausen, J. Sundermeyer and S. Schindler,
Angew. Chem., Int. Ed., 2004, 43, 4360.
956 C. Wurtele, E. Gaoutchenova, K. Harms, M. C. Holthausen,
J. Sundermeyer and S. Schindler, Angew. Chem., Int. Ed., 2006,
45, 3867.
957 A. De La Lande, H. Gerard and O. Parisel, Int. J. Quantum
Chem., 2008, 108, 1898.
958 B. F. Gherman, D. E. Heppner, W. B. Tolman and C. J. Cramer,
JBIC, J. Biol. Inorg. Chem., 2006, 11, 197.
959 N. W. Aboelella, B. F. Gherman, L. M. R. Hill, J. T. York,
N. Holm, V. G. Young, C. J. Cramer and W. B. Tolman, J. Am.
Chem. Soc., 2006, 128, 3445.
960 B. F. Gherman, W. B. Tolman and C. J. Cramer, J. Comput.
Chem., 2006, 27, 1950.
961 L. M. R. Hill, B. F. Gherman, N. W. Aboelella, C. J. Cramer
and W. B. Tolman, Dalton Trans., 2006, 4944.
962 D. E. Heppner, B. F. Gherman, W. B. Tolman and C. J. Cramer,
Dalton Trans., 2006, 4773.
963 S. Hong, S. M. Huber, L. Gagliardi, C. J. Cramer and
W. B. Tolman, J. Am. Chem. Soc., 2007, 129, 14190.
964 S. M. Huber, A. R. Moughal Shahi, F. Aquilante, C. J.
Cramer and L. Gagliardi, J. Chem. Theory Comput., DOI:
10.1021/ct900282m.
965 P. Comba, S. Knoppe, B. Martin, G. Rajaraman, C. Rolli,
B. Shapiro and T. Stork, Chem.Eur. J., 2008, 14, 344.
966 K. Yoshizawa and Y. Shiota, J. Am. Chem. Soc., 2006, 128,
9873.
967 P. Comba, C. L. Lang, C. L. de Laorden, A. Muruganantham,
G. Rajaraman, H. Wadepold and M. Zajaczkowski, Chem.Eur. J.,
2008, 14, 5313.
968 C. J. Cramer, B. A. Smith and W. B. Tolman, J. Am. Chem.
Soc., 1996, 118, 11283.
969 M. Flock and K. Pierloot, J. Phys. Chem. A, 1999, 103, 95.
970 O. Sander, A. Henss, C. Nather, C. Wurtele, M. C. Holthausen,
S. Schindler and F. Tuczek, Chem.Eur. J., 2008, 14, 9714.
971 T. Osako, S. Nagatomo, T. Kitagawa, C. J. Cramer and S. Itoh,
JBIC, J. Biol. Inorg. Chem., 2005, 10, 581.
972 W. Ghattas, M. Giorgi, Y. Mekmouche, T. Tanaka,
A. Rockenbauer, M. Reglier, Y. Hitomi and A. J. Simaan, Inorg.
Chem., 2008, 47, 4627.
973 A. Kunishita, J. D. Scanlon, H. Ishimaru, K. Honda, T. Ogura,
M. Suzuki, C. J. Cramer and S. Itoh, Inorg. Chem., 2008, 47,
8222.
974 G. Periyasamy, M. Sundararajan, I. H. Hillier, N. A. Burton
and J. J. W. McDouall, Phys. Chem. Chem. Phys., 2007, 9, 2498.
975 J. P. Holland, P. J. Barnard, D. Collison, J. R. Dilworth,
R. Edge, J. C. Green, J. M. Heslop, E. J. L. McInnes,
C. G. Salzmann and A. L. Thompson, Eur. J. Inorg. Chem.,
2008, 3549.

10812 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

976 R. R. Nazmutdinov, N. V. Roznyatovskaya, D. V. Glukhov,

I. Manyurov, V. M. Mazin, G. A. Tsirlina and M. Probst,
Inorg. Chem., 2008, 47, 6659.
977 J. U. Reveles, P. Calaminici, M. R. Beltran, A. M. Koster and
S. N. Khanna, J. Am. Chem. Soc., 2007, 129, 15565.
978 P. H.-Y. Cheong, P. Morganelli, M. R. Luzung, K. N. Houk and
F. D. Toste, J. Am. Chem. Soc., 2008, 130, 4517.
979 E. Soriano and J. Marco-Contelles, Acc. Chem. Res., 2009, 42,
1026.
980 M. Mantina and D. G. Truhlar, 2009, unpublished calculations.
981 J. Carrasco, F. Illas and S. T. Bromley, Phys. Rev. Lett., 2007,
99, 235502.
982 T. Cadenbach, T. Bollermann, C. Gemel, I. Fernandez, M.
von Hopgarten, G. Frenking and R. A. Fischer, Angew. Chem.,
Int. Ed., 2008, 47, 9150.
983 L. Bernasconi, E. J. Baerends and M. Sprik, J. Phys. Chem. B,
2006, 110, 11444.
984 Y. Dai and E. Blaisten-Barojas, J. Phys. Chem. A, 2008, 112,
11052.
985 K. Iokibe, K. Azumi and H. Tachikawa, J. Phys. Chem. C, 2007,
111, 13510.
986 H. Tachikawa, K. Iokibe, K. Azumi and H. Kawabata,
Phys. Chem. Chem. Phys., 2007, 9, 3978.
987 H. J. Flad, F. Schautz, Y. X. Wang, M. Dolg and A. Savin,
Eur. Phys. J. D, 1999, 6, 243.
988 A. Suwattanamala, A. Magalhaes and J. Gomes, Theor. Chem.
Acc., 2007, 117, 431.
989 Z. Lu, L. Wen, Z. Ni, Y. Li, H. Zhu and Q. Meng, Cryst. Growth
Des., 2007, 7, 268.
990 P. Fernandez, A. Sousa-Pedrares, J. Romero, J. A. Garc aVazquez, A. Sousa and P. Perez-Lourido, Inorg. Chem., 2008,
47, 2121.
991 Z. Zhu, M. Brynda, R. J. Wright, R. C. Fischer, W. A. Merrill,
E. Rivard, R. Wolf, J. C. Fettinger, M. M. Olmstead and
P. P. Power, J. Am. Chem. Soc., 2007, 129, 10847.
992 J. G. Wiederhold, C. J. Cramer, K. Daniel, I. Infante,
B. Bourdon and R. Kretzschmar, Geochim. Cosmochim. Acta,
2009, 73, A1438.
993 C. M. Rignanese, J. Phys.: Condens. Matter, 2005, 17, R357.
994 W. E. Pickett, Comments Solid State Phys., 1985, 12, 1.
995 K. B. Hathaway, H. J. F. Jansen and A. J. Freeman, Phys. Rev. B:
Condens. Matter Mater. Phys., 1985, 31, 7603.
996 A. Garc a, C. Elsasser, J. Zhu, S. G. Louie and M. L. Cohen,
Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46, 9829.
997 S. Kurth, J. P. Perdew and P. Blaha, Int. J. QuantumChem.,
1999, 75, 889.
998 B. Grabowski, T. Hickel and J. Neugebauer, Phys. Rev. B:
Condens. Matter Mater. Phys., 2007, 76, 024309.
999 J. Heyd, J. E. Peralta, G. E. Scuseria and R. L. Martin, J. Chem.
Phys., 2005, 123, 174101.
1000 S. Piskunov, E. Heifets, R. I. Eglitis and G. Borstel, Comput.
Mater. Sci., 2004, 29, 165.
1001 P. Baettig, C. F. Schelle, R. LeSar, U. V. Waghmare and
N. A. Spaldin, Chem. Mater., 2005, 17, 1376.
1002 S. M. Hosseini, T. Movlasooy and A. Kompany, Phys. B, 2007,
391, 316.
1003 C. M. S. Gannarelli, D. Alfe and M. J. Gillan, Phys. Earth
Planet. Interiors, 2005, 152, 67.
1004 D. F. Johnson and E. A. Carter, J. Chem. Phys., 2008, 128,
104703.
1005 T. Sun, K. Umemoto, Z. Wu, J.-C. Zheng and
R. M. Wentzcovitch, Phys. Rev. B: Condens. Matter Mater.
Phys., 2008, 78, 024304.
1006 Z. H. Li, D. Bhatt, N. E. Schultz, J. I. Siepmann and
D. G. Truhlar, J. Phys. Chem. C, 2007, 111, 16227.
1007 A. F. Izmaylov and G. E. Scuseria, J. Chem. Phys., 2008, 129,
034101; E. N. Brothers, A. F. Izmaylov, J. O. Normand,
V. Barone and G. E. Scuseria, J. Chem. Phys., 2008, 129, 11102.
1008 V. Fiorentini and A. Baldereschi, Phys. Rev. B: Condens. Matter
Mater. Phys., 1995, 51, 17196.
1009 K. Xiong, J. Robertson and S. J. Clark, J. Appl. Phys., 2007,
102, 083701.
1010 N. D. Mermin, Phys. Rev., 1965, 137, A1441.
1011 E. I. Isaev, N. V. Skorodumova, R. Ahuja, Y. K. Vekilov and
B. Johansson, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 9168.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

1012 K. Umemoto, R. M. Wentzcovitch, Y. G. Yu and R. Requist,

Earth Planet. Sci. Lett., 2008, 276, 198.
1013 J. Uddin and G. E. Scuseria, Phys. Rev. B: Condens. Matter
Mater. Phys., 2006, 74, 245115.
1014 Y. Zhao and D. G. Truhlar, J. Chem. Phys., 2009, 130, 074103.
1015 A. Filippetti, N. A. Spaldin and S. Sanvito, Chem. Phys., 2005,
309, 59.
1016 M. Peressi, A. Debernardi, S. Picozzi, F. Antoniella and
A. Continenza, Comput. Mater. Sci., 2005, 33, 125.
1017 C. Tablero, Phys. Rev. B: Condens. Matter Mater. Phys., 2005,
72, 035213.
1018 P. Boguslawski and J. Bernholc, Phys. Rev. B: Condens. Matter
Mater. Phys., 2005, 72, 115208.
1019 T. Jungwirth, J. Sinova, J. Masek, J. Kucera and
A. H. MacDonald, Rev. Mod. Phys., 2006, 78, 809.
1020 C. Tablero, J. Chem. Phys., 2007, 126, 164703.
1021 X. Y. Cui, B. Delley, A. J. Freeman and C. Stamp, Phys. Rev.
B: Condens. Matter Mater. Phys., 2007, 76, 045201.
1022 J. A. Chan, J. Z. Liu, H. Raebiger, S. Lany and A. Zunger,
Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78, 184109.
1023 A. Filippetti and V. Fiorentini, J. Magn. Magn. Mater., 2007,
310, 1648.
1024 J. Liu and Q. Ge, J. Phys. Chem. B, 2006, 110, 25863.
1025 J. Liu and Q. Ge, J. Alloys Compd., 2007, 446447, 267.
1026 P. Romaniello and P. L. de Boeij, J. Chem. Phys., 2005, 122,
164303.
1027 J. A. Berger, P. Romaniello, R. van Leeuwen and P. L. de Boeij,
Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 245117.
1028 C. Leighton, M. Manno, A. Cady, J. W. Freeland, L. Wang,
K. Umemoto, R. M. Wentzcovitch, T. Chen, C. L. Chien,
P. L. Kuhns, M. J. R. Hoch, A. P. Reyes, W. G. Moulton,
E. D. Dahlberg, J. Checkelsky and J. Eckert, J. Phys.: Condens.
Matter, 2007, 19, 315219.
1029 T. Kawakami, Y. Tsujimoto, H. Kageyama, X.-Q. Chen,
C. L. Fu, C. Tassel, A. Kitada, S. Suto, K. Hirama,
Y. Sekiya, Y. Makino, T. Okada, T. Yagi, N. Hayashi,
K. Yoshimura, S. Nasu, R. Podloucky and M. Takano,
Nat. Chem., 2009, 1, 371.
1030 M.-H. Whangbo and J. Kohler, Nat. Chem., 2009, 1, 351.
1031 R. J. Magyar, V. Mujica, M. Marquez and C. Gonzalez,
Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 144421.
1032 C. Gonzalez, Y. Simon-Manso, M. Marquez and V. Mujica,
J. Phys. Chem. B, 2006, 110, 687.
1033 H. Wende, M. Bernien, J. Luo, C. Sorg, N. Ponpandian,
J. Kurde, J. Miguel, M. Piantek, X. Xu, P. Eckhold, W. Kuch,
K. Baberschke, P. M. Panchmatia, B. Sanyal, P. M. Oppeneer
and O. Eriksson, Nat. Mater., 2007, 6, 516.
1034 A. Bilic, J. R. Reimers, N. S. Hush, R. C. Hoft and M. J. Ford,
J. Chem. Theory Comput., 2006, 2, 1093.
1035 P. F. Cafe, A. G. Larsen, W. Yang, A. Bilic, I. M. Blake,
M. J. Crossley, J. Zhang, H. Wackerbarth, J. Ulstrup and
J. R. Reimers, J. Phys. Chem. C, 2007, 111, 17285.
1036 W. Ma and Y. Fang, J. Nanopart. Res., 2006, 8, 761.
1037 K. M. Neyman, C. Inntam, V. A. Nasluzov, R. Kosarev and
N. Rosch, Appl. Phys. A: Mater. Sci. Process., 2004, 78, 823.
1038 C. Inntam, L. V. Moskaleva, K. M. Neyman, V. A. Nasluzov
and N. Rosch, Appl. Phys. A: Mater. Sci. Process., 2006, 82, 181.
1039 C. Inntam, L. V. Moskaleva, I. V. Yudanov, K. M. Neyman and
N. Rosch, Chem. Phys. Lett., 2006, 417, 515.
1040 K. M. Neyman, C. Inntam, L. V. Moskaleva and N. Rosch,
Chem.Eur. J., 2007, 13, 277.
1041 S. I. Bosko, L. V. Moskaleva, A. V. Matveev and N. Rosch,
J. Phys. Chem. A, 2007, 111, 6870.
1042 E. Florez, F. Mondragon, T. N. Truong and P. Fuentealba,
Surf. Sci., 2007, 601, 656.
1043 G. Bacaro and A. Fortunelli, New J. Phys., 2007, 9, 22.
1044 V. Bonacic-Koutecky, C. Burgel, L. Kronik, A. E. Kuznetsov
and R. Mintric, Eur. Phys. J. D, 2007, 45, 471.
1045 M. Sterrer, T. Risse, L. Giordano, M. Heyde, N. Nilius,
H.-P. Rust, G. Pacchioni and H.-J. Freund, Angew. Chem.,
Int. Ed., 2007, 46, 8703.
1046 R. Caballero, C. Quintanar, A. M. Koster, S. N. Khanna and
J. U. Reveles, J. Phys. Chem. C, 2008, 112, 14919.
1047 E. Florez, F. Mondragon, P. Fuentealba and F. Illas, Phys. Rev.
B: Condens. Matter Mater. Phys., 2008, 78, 075426.

This journal is


c

the Owner Societies 2009

1048 B. Hinnemann and E. A. Carter, J. Phys. Chem. C, 2007, 111,

7105.
1049 A. Michaelides, K. Reuter and M. Scheer, J. Vac. Sci.
Technol., A, 2005, 23, 1487.
1050 J. L. C. Faj n, M. N. l. D. S. Cordeiro and J. R. B. Gomes,
Surf. Sci., 2008, 602, 424.
1051 C. Stamp, A. Soon, S. Piccinin, H. Shi and H. Zhang, J. Phys.:
Condens. Matter, 2008, 20, 184021.
1052 B. Akdim, S. Hussain and R. Pachter, in Lecture Notes in
Computer Science (Computational ScienceICCS 2008, Part II),
Springer-Verlag, Berlin, 2008, vol. 5102, p. 353.
1053 H. Zhang, A. Soon, B. Delley and C. Stamp, Phys. Rev. B:
Condens. Matter Mater. Phys., 2008, 78, 045436.
1054 P. Kaghazchi, F. C. Simeone, K. A. Suliman, L. A. Kibler and
T. Jacob, Faraday Discuss., 2009, 140, 69.
1055 Y. Han, C.-j. Liu and Q. Ge, J. Phys. Chem. B, 2006, 110, 7463.
1056 Y. Han, C.-j. Liu and Q. Ge, J. Phys. Chem. C, 2007, 111, 16397.
1057 S. Xia, G. Pan, Z.-L. Cai, Y. Wang and J. R. Reimers, J. Phys.
Chem. C, 2007, 111, 10427.
1058 V. A. Ranea, A. Michaelides, R. Ram rez, P. L. de Andres,
J. A. Verges and D. A. King, Phys. Rev. Lett., 2004, 92, 136104.
1059 L. Xu, G. Henkelman, C. T. Campbell and H. Jonsson, Surf.
Sci., 2006, 600, 1351.
1060 P. J. Feibelman, B. Hammer, J. K. Nrskov, F. Wagner,
M. Scheer, R. Stumpf, R. Watwe and J. Dumesic, J. Phys.
Chem. B, 2001, 105, 4018.
1061 K. Doll, Surf. Sci., 2004, 573, 464.
1062 M. Neef and K. Doll, Surf. Sci., 2006, 600, 1085.
1063 Y. Wang, S. de Gironcoli, N. S. Hush and J. R. Reimers, J. Am.
Chem. Soc., 2007, 129, 10402.
1064 I. Dabo, A. Wieckowski and N. Marzari, J. Am. Chem. Soc.,
2007, 129, 11045.
1065 T. Zhang, Z.-P. Liu, S. M. Driver, S. J. Pratt, S. J. Jenkins and
D. A. King, Phys. Rev. Lett., 2005, 95, 266102.
1066 J. M. H. Lo and T. Ziegler, J. Phys. Chem. C, 2008, 112, 3679.
1067 D. E. Jiang and E. A. Carter, Surf. Sci., 2004, 570, 167.
1068 J. M. H. Lo and T. Ziegler, J. Phys. Chem. C, 2008, 112, 3692.
1069 J. M. H. Lo and T. Ziegler, J. Phys. Chem. C, 2007, 111, 13149.
1070 A. Valcarcel, A. Clotet, J. M. Ricart and F. Illas, Chem. Phys.,
2005, 309, 33.
1071 Y. Wang, N. S. Hush and J. R. Reimers, J. Phys. Chem. C, 2007,
111, 10878.
1072 Y. Wang, N. S. Hush and J. R. Reimers, J. Am. Chem. Soc.,
2007, 129, 14532.
1073 Y. Wang, N. S. Hush and J. R. Reimers, Phys. Rev. B: Condens.
Matter Mater. Phys., 2007, 75, 233416.
1074 S. Y. Quek, M. M. Biener, J. Biener, J. Bhattacharjee,
C. M. Friend, U. V. Waghmare and E. Kaxiras, J. Phys. Chem.
B, 2006, 110, 15663.
1075 W. Gao, T. A. Baker, L. Zhou, D. S. Pinnaduwage, E. Kaxiras
and C. M. Friend, J. Am. Chem. Soc., 2008, 130, 3560.
1076 T. A. Baker, C. M. Friend and E. Kaxiras, J. Chem. Phys., 2008,
129, 104702.
1077 L. D. Marks, A. N. Chiaramonti, F. Tran and P. Blaha,
Surf. Sci., 2009, 603, 2179.
1078 Y. Mao, J. Yuan and J. Zhang, J. Phys.: Condens. Matter, 2008,
20, 115209.
1079 N. Kuganathan and J. C. Green, Chem. Commun., 2008, 2432.
1080 E. L. Sceats and J. C. Green, Phys. Rev. B: Condens. Matter
Mater. Phys., 2007, 75, 245441.
1081 S. Y. Quek, M. M. Biener, J. Biener, C. M. Friend and
E. Kaxiras, Surf. Sci., 2005, 577, L71.
1082 D. E. Jiang and E. A. Carter, Surf. Sci., 2005, 583, 60.
1083 D. E. Jiang and E. A. Carter, J. Phys. Chem. B, 2005, 109, 20469.
1084 D. E. Jiang and E. A. Carter, J. Phys. Chem. B, 2006, 110, 22213.
1085 M. J. S. Spencer, I. K. Snook and I. Yarovsky, J. Phys. Chem. B,
2004, 108, 10965.
1086 M. J. S. Spencer, I. K. Snook and I. Yarovsky, J. Phys. Chem. B,
2005, 109, 9604.
1087 M. J. S. Spencer, I. K. Snook and I. Yarovsky, J. Phys.
Chem. B, 2005, 109, 10204.
1088 M. J. S. Spencer, I. K. Snook and I. Yarovsky, J. Phys. Chem. B,
2006, 110, 956.
1089 S. G. Nelson, M. J. S. Spencer, I. K. Snook and I. Yarovsky,
Surf. Sci., 2005, 590, 63.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10813

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

1090 N. Todorova, M. J. S. Spencer and I. Yarovsky, Surf. Sci., 2007,

601, 665.
1091 D. F. Johnson, D. E. Jiang and E. A. Carter, Surf. Sci., 2007,
601, 699.
1092 M. J. S. Spencer and I. Yarovsky, J. Phys. Chem. C, 2007, 111,
16372.
1093 M. J. S. Spencer, N. Todorova and I. Yarovsky, Surf. Sci., 2008,
602, 1547.
1094 Y. Kurzweil and R. Baer, Phys. Rev. B: Condens. Matter Mater.
Phys., 2006, 73, 075413.
1095 R. Ferrando, J. Jellinek and R. L. Johnston, Chem. Rev., 2008,
108, 845.
1096 X. Y. Zhu, G. Dutton, D. P. Quinn, C. D. Lindstrom,
N. E. Schultz and D. G. Truhlar, Phys. Rev. B: Condens. Matter
Mater. Phys., 2006, 74, 241401.
1097 B. Li, J. Zhao, K. Onda, K. D. Jordan, J. Yang and H. Petek,
Science, 2006, 311, 1436.
1098 M. Feng, J. Zhao and H. Petek, Science, 2008, 320, 359.
1099 Y. Pavlyukh and J. Berakdar, Chem. Phys. Lett., 2009, 468, 313.
1100 W. T. Geng, J. Nara and T. Ohno, Thin Solid Films, 2004,
464465, 379.
1101 P. A. Derosa, A. G. Zacarias and J. M. Seminario, in Reviews
of Modern Quantum Chemistry, ed. K. D. Sen, World Scientic,
Singapore, 2002, p. 1537.
1102 Z. A. Schelly, Colloids Surf., B, 2007, 56, 281.
1103 M. Walter, P. Frondelius, K. Honkala and H. Hakkinen, Phys.
Rev. Lett., 2007, 99, 096102.
1104 E. Badaeva, Y. Feng, D. R. Gamelin and X. Li, New J. Phys.,
2008, 10, 055013.
1105 H. Zeng, R. R. Vanga, D. S. Marynick and Z. A. Schelly,
J. Phys. Chem. B, 2008, 112, 14422.
1106 I. V. Yudanov, K. M. Neyman and N. Rosch, Phys. Chem.
Chem. Phys., 2004, 6, 116.
1107 K. M. Neyman, G. N. Vayssilov and N. Rosch, J. Organomet.
Chem., 2004, 689, 4384.
1108 H. Gronbeck and P. Broqvist, Phys. Rev. B: Condens. Matter
Mater. Phys., 2005, 71, 073408.
1109 L. M. Liu, B. McAllister, H. Q. Ye and P. Hu, J. Am. Chem.
Soc., 2006, 128, 4017.
1110 E. A. Ivanova Shor, V. A. Nasluzov, A. M. Shor,
G. N. Vayssilov and N. Rosch, J. Phys. Chem. C, 2007, 111,
12340.
1111 M. Yang, K. A. Jackson and J. Jellinek, J. Chem. Phys., 2006,
125, 144308.
1112 F. Cinquini, L. Giordano, G. Pacchioni, A. M. Ferrari, C. Pisani
and C. Roetti, Phys. Rev. B: Condens. Matter Mater. Phys.,
2006, 74, 165403.
1113 S. K. Brayshaw, J. C. Green, N. Hazasi and A. S. Weller, Dalton
Trans., 2007, 1781.
1114 M. Castro, Chem. Phys. Lett., 2007, 435, 322.
1115 M. Baron, D. Stacchiola, S. Ulrich, N. Nilius, S. Shaikhutdinov,
H. J. Freund, U. Martinez, L. Giordano and G. Pacchioni,
J. Phys. Chem. C, 2008, 112, 3405.
1116 G. Pacchioni, S. Sicolo, C. D. Valentin, M. Chiesa and
E. Giamello, J. Am. Chem. Soc., 2008, 130, 8690.
1117 D. Bandyopadhyay, J. Appl. Phys., 2008, 104, 084308.
1118 D. Bandyopadhyay and M. Kumar, Chem. Phys., 2008, 353, 170.
1119 S. Alayoglu, A. V. Nilekar, M. Mavrikakis and B. Eichhorn,
Nat. Mater., 2008, 7, 333.
1120 V. Simic-Milosevic, M. Heyde, N. Nilius, T. KoAa`nig,
H. P. Rust, M. Sterrer, T. Risse, H. J. Freund, L. Giordano
and G. Pacchioni, J. Am. Chem. Soc., 2008, 130, 7814.
1121 H.-J. Freund and G. Pacchioni, Chem. Soc. Rev., 2008, 37, 2224.
1122 L. C. Grabow and M. Mavrikakis, Angew. Chem., Int. Ed., 2008,
47, 7390.
1123 M. Gruber, G. Heimel, L. Romaner, J.-L. Bredas and E. Zojer,
Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 165411.
1124 H. Gronbeck, Top. Catal., 2004, 28, 59.
1125 C. R. A. Catlow, S. A. French, A. A. Sokol and J. M. Thomas,
Philos. Trans. R. Soc. London, Ser. A, 2005, 363, 913.
1126 F. Cinquini, C. Valentin, E. Finazzi, L. Giordano and
G. Pacchioni, Theor. Chem. Acc., 2007, 117, 827.
1127 J. N. Nrskov, M. Scheer and H. Toulhoat, MRS Bull., 2006,
31, 669.

10814 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

1128 J. K. Nrskov, T. Bligaard, B. Hvolbaek, F. Abild-Pedersen,

I. Chorkendor and C. H. Christensen, Chem. Soc. Rev., 2008,
37, 2163.
1129 K. M. Neyman and F. Illas, Catal. Today, 2005, 105, 2.
1130 R. Coquet, K. L. Howard and D. J. Willock, Chem. Soc. Rev.,
2008, 37, 2046.
1131 G. A. Somorjai and J. Y. Park, Angew. Chem., Int. Ed., 2008, 47,
9212.
1132 H. Falsig, B. Hvolbaek, I. S. Kristensen, T. Jiang, T. Bligaard,
C. H. Christensen and J. K. Nrskov, Angew. Chem., Int. Ed.,
2008, 47, 4835.
1133 G. Lanzani, R. Martinazzo, G. Materzanini, I. Pino and
G. Tantardini, Theor. Chem. Acc., 2007, 117, 805.
1134 B. Kasemo, E. Tornqvist, J. K. Nrskov and B. I. Lundqvist,
Surf. Sci., 1979, 89, 554.
1135 M. Persson and B. Hellsing, Phys. Rev. Lett., 1982, 49, 662.
1136 S. E. Wonchoba, W.-P. Hu and D. G. Truhlar, in Theoretical
and Computational Approaches to Interface Phenomena,
ed. H. L. Sellers and J. T. Golab, Plenum, New York, 1994, p. 1.
1137 J. T. Kindt, J. C. Tully, M. Head-Gordon and M. A. Gomez,
J. Chem. Phys., 1998, 109, 3629.
1138 H. Nienhaus, H. S. Bergh, B. Gergen, A. Majumdar,
W. H. Weinberg and E. W. McFarland, Phys. Rev. Lett.,
1999, 82, 446.
1139 J. C. Tully, Annu. Rev. Phys. Chem., 2000, 51, 153.
1140 B. Gergen, H. Nienhaus, W. H. Weinberg and
E. W. McFarland, Science, 2001, 294, 2521.
1141 H. Nienhaus, Surf. Sci. Rep., 2002, 45, 1.
1142 J. W. Gadzuk, J. Phys. Chem. B, 2002, 106, 8265.
1143 J. R. Trail, M. C. Graham, D. M. Bird, M. Persson and
S. Holloway, Phys. Rev. Lett., 2002, 88, 166802.
1144 G. Katz, R. Koslo and Y. Zeiri, J. Chem. Phys., 2004, 120,
3931.
1145 X. Z. Ji and G. A. Somorjai, J. Phys. Chem. B, 2005, 109, 22530.
1146 J. Behler, B. Delley, S. Lorenz, K. Reuter and M. Scheer, Phys.
Rev. Lett., 2005, 94, 036104.
1147 P. Nieto, E. Pijper, D. Barredo, G. Laurent, R. A. Olsen,
E.-J. Baerends, G.-J. Kroes and D. Farias, Science, 2006, 312,
86.
1148 A. C. Luntz, M. Persson and G. O. Sitz, J. Chem. Phys., 2006,
124, 091101.
1149 N. Shenvi, S. Roy, P. Parandekar and J. Tully, J. Chem. Phys.,
2006, 125, 154703.
1150 M. Lindenblatt and E. Pehlke, Phys. Rev. Lett., 2006, 97,
216101.
1151 D. Krix, R. Nunthel and H. Nienhaus, Phys. Rev. B: Condens.
Matter Mater. Phys., 2007, 75, 073410.
1152 I. Rahinov, R. Cooper, C. Yuan, X. Yang, D. J. Auerbach and
A. M. Wodtke, J. Chem. Phys., 2008, 129, 214708.
1153 N. H. Nahler, J. D. White, J. LaRue, D. J. Auerbach and
A. M. Wodtke, Science, 2008, 321, 1191.
1154 D. M. Bird, M. S. Mizielinski, M. Lindenblatt and E. Pehlke,
Surf. Sci., 2008, 602, 1212.
1155 M. Tomellini, J. Phys.: Condens. Matter, 2008, 20, 135002.
1156 J. I. Juaristi, M. Alducin, R. D. Muino, H. F. Busnengo and
A. Salin, Phys. Rev. Lett., 2008, 100, 116102.
1157 R. Cooper, I. Rahinov, C. Yuan, X. Yang, D. J. Auerbach and
A. M. Wodtke, J. Vac. Sci. Technol., A, 2009, 27, 907.
1158 S. N. Maximo and M. P. Head-Gordon, Proc. Natl. Acad. Sci.
U. S. A., 2009, 106, 11460.
1159 N. Shenvi, S. Roy and J. C. Tully, J. Chem. Phys., 2009, 130,
174107.
1160 M. Timmer and P. Kratzer, Phys. Rev. B: Condens. Matter
Mater. Phys., 2009, 79, 165407.
1161 K. Honkala, A. Hellman, I. N. Remediakis, A. Logadottir,
A. Carlsson, S. Dahl, C. H. Christensen and J. K. Nrskov,
Science, 2005, 307, 555.
1162 A. Hellman, E. J. Baerends, M. Biczysko, T. Bligaard,
C. H. Christensen, D. C. Clary, S. Dahl, R. van Harrevelt,
K. Honkala, H. Jonsson, G. J. Kroes, M. Luppi, U. Manthe,
J. K. Nrskov, R. A. Olsen, J. Rossmeisl, E. Skulason,
C. S. Tautermann, A. J. C. Varandas and J. K. Vincent,
J. Phys. Chem. B, 2006, 110, 17719.

This journal is


c

the Owner Societies 2009

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

1163 C. S. Tautermann, Y. K. Sturdy and D. C. Clary, J. Catal., 2006,

244, 199; C. S. Tautermann and D. C. Clary, Phys. Chem. Chem.
Phys., 2006, 8, 1437.
1164 Z.-P. Liu, S. J. Jenkins and D. A. King, J. Am. Chem. Soc., 2004,
126, 10746.
1165 J. K. Vincent, R. A. Olsen, G.-J. Kroes, M. Luppi and
E.-J. Baerends, J. Chem. Phys., 2005, 122, 044701.
1166 C. Crespos, H. D. Meyer, R. C. Mowrey and G. J. Kroes,
J. Chem. Phys., 2006, 124, 074706.
1167 H. L. Abbott and I. Harrison, J. Chem. Phys., 2006, 125, 024704.
1168 G. J. Kroes, E. Pijper and A. Salin, J. Chem. Phys., 2007, 127,
164722.
1169 I. M. N. Groot, H. Ueta, M. J. T. C. van der Niet, A. W. Kleyn
and L. B. F. Juurlink, J. Chem. Phys., 2007, 127, 244701.
1170 H. F. Busnengo and A. E. Martinez, J. Phys. Chem. C, 2008,
112, 5579.
1171 R. A. Olsen, D. A. McCormack, M. Luppi and E. J. Baerends,
J. Chem. Phys., 2008, 128, 194715.
1172 A. C. Luntz, Surf. Sci., 2009, 603, 1557.
1173 R. Valero, J. R. B. Gomes, D. G. Truhlar and F. Illas, J. Chem.
Phys., 2008, 129, 124710.
1174 C. Di Valentin, G. Pacchioni, T. Bredow, D. Dominguez-Ariza
and F. Illas, J. Chem. Phys., 2002, 117, 2299.
1175 D. Dominguez-Ariza, F. Illas, T. Bredow, C. D. Valentin and
G. Pacchioni, Mol. Phys., 2003, 101, 241.
1176 M. Chiesa, M. C. Paganini, E. Giamello, C. D. Valentin and
G. Pacchioni, J. Mol. Catal. A: Chem., 2003, 204205, 779.
1177 G. Pacchioni, C. D. Valentin, D. Dominguez-Ariza, F. Illas,
T. Bredow, T. Kluner and V. Staemmler, J. Phys.: Condens.
Matter, 2004, 16, S2497.
1178 A. Scagnelli, C. D. Valentin and G. Pacchioni, Surf. Sci., 2006,
600, 386.
1179 A. Rohrbach and J. Hafner, Phys. Rev. B: Condens. Matter
Mater. Phys., 2005, 71, 045405.
1180 R. Valero, J. R. B. Gomes, D. G. Truhlar and F. Illas, to be
published.
1181 N. Hansen, A. Heyden, A. T. Bell and F. J. Keil, J. Phys. Chem.
C, 2007, 111, 2092.
1182 A. Heyden, B. Peters, A. T. Bell and F. J. Keil, J. Phys. Chem. B,
2005, 109, 1857.
1183 A. Heyden, N. Hansen, A. T. Bell and F. J. Keil, J. Phys. Chem.
B, 2006, 110, 17096.
1184 A. Goodrow and A. T. Bell, J. Phys. Chem. C, 2008, 112, 13204.
1185 J. Bonde, P. G. Moses, T. F. Jaramillo, J. K. Nrskov and
I. Chorkendor, Faraday Discuss., 2009, 140, 219.
1186 A. A. Gokhale, J. A. Dumesic and M. Mavrikakis, J. Am. Chem.
Soc., 2008, 130, 1402.
1187 M. P. Andersson, F. Abild-Pedersen, I. N. Remediakis,
T. Bligaard, G. Jones, J. Engbaek, O. Lytken, S. Horch,
J. H. Nielsen, J. Sehested, J. R. Rostrup-Nielsen,
J. K. Nrskov and I. Chorkendor, J. Catal., 2008, 255, 6.
1188 Y. Xu, W. A. Shelton and W. F. Schneider, J. Phys. Chem. B,
2006, 110, 16591.
1189 R. B. Getman and W. F. Schneider, J. Phys. Chem. C, 2007, 111,
389.
1190 R. B. Getman, Y. Xu and W. F. Schneider, J. Phys. Chem. C,
2008, 112, 9559.
1191 A. D. Smeltz, R. B. Getman, W. F. Schneider and F. H. Ribeiro,
Catal. Today, 2008, 136, 84.
1192 Y. Xu, R. B. Getman, W. A. Shelton and W. F. Schneider,
Phys. Chem. Chem. Phys., 2008, 10, 6009.
1193 H. Y. Wang and W. F. Schneider, J. Chem. Phys., 2007, 127,
064706.
1194 Z.-X. Chen, K. M. Neyman, K. H. Lim and N. Rosch,
Langmuir, 2004, 20, 8068.
1195 Z.-X. Chen, K. H. Lim, K. M. Neyman and N. Rosch, J. Phys.
Chem. B, 2005, 109, 4568.
1196 K. H. Lim, Z.-X. Chen, K. M. Neyman and N. Rosch, J. Phys.
Chem. B, 2006, 110, 14890.
1197 K. H. Lim, L. V. Moskaleva and N. Rosch, ChemPhysChem,
2006, 7, 1802.
1198 Y. Cao and Z.-X. Chen, Surf. Sci., 2006, 600, 4572.
1199 Q.-L. Tang and Z.-X. Chen, J. Chem. Phys., 2007, 127, 104707.
1200 X. Li, A. J. Gellman and D. S. Sholl, J. Chem. Phys., 2007, 127,
144710.

This journal is


c

the Owner Societies 2009

1201 G.
Pacchioni,
http://pubs.acs.org/JACSbeta/issue3.html,
accessed Jan. 16, 2009.
1202 A. S. Y. Foo and K. H. Lim, Catal. Lett., 2009, 127, 113.
1203 E. Jeroro and J. M. Vohs, J. Am. Chem. Soc., 2008, 130, 10199.
1204 H. Tang and B. L. Trout, J. Phys. Chem. B, 2006, 110, 6856.
1205 A. T. Anghel, D. J. Wales, S. J. Jenkins and D. A. King,
Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 113410.
1206 S. E. Wonchoba and D. G. Truhlar, J. Phys. Chem. B, 1998, 102,
6842.
1207 H. Yang and J. L. Whitten, J. Chem. Phys., 1992, 96, 5529.
1208 P. Kratzer, B. Hammer and J. K. Nrskov, J. Chem. Phys.,
1996, 105, 5595.
1209 S. Nave and B. Jackson, J. Chem. Phys., 2007, 127, 224702.
1210 S. Nave and B. Jackson, Phys. Rev. Lett., 2007, 98, 173003.
1211 S. Nave and B. Jackson, J. Chem. Phys., 2009, 130, 054701.
1212 W. Z. Lai, D. Q. Xie and D. H. Zhang, Surf. Sci., 2005, 594, 83.
1213 M.-S. Liao, C.-T. Au and C.-F. Ng, Chem. Phys. Lett., 1997,
272, 445.
1214 G. Psofogiannakis, A. St-Amant and M. Ternan, J. Phys. Chem.
B, 2006, 110, 24593.
1215 H.-Y. Li, Y.-L. Guo, Y. Guo, G.-Z. Lu and P. Hu, J. Chem.
Phys., 2008, 128, 051101.
1216 F. Abild-Pedersen, J. Greeley, F. Studt, J. Rossmeisl,
T. R. Munter, P. G. Moses, E. Skulason, T. Bligaard and
J. K. Nrskov, Phys. Rev. Lett., 2007, 99, 016105.
1217 N. N. Semenov, Some Problems in Chemical Kinetics and
Reactivity, Princeton University Press, Princeton, NJ, 1958.
1218 M. G. Evans and M. Polanyi, Trans. Faraday Soc., 1938, 34, 11.
1219 F. Studt, F. Abild-Pedersen, T. Bligaard, R. Z. Srensen,
C. H. Christensen and J. K. Nrskov, Science, 2008, 320, 1320.
1220 F. Studt, F. Abild-Pedersen, T. Bligaard, R. Z. Srensen,
C. H. Christensen and J. K. Nrskov, Angew. Chem., Int. Ed.,
2008, 47, 9299.
1221 G. Jones, J. G. Jakobsen, S. S. Shim, J. Kleis, M. P. Andersson,
J. Rossmeisl, F. Abild-Pedersen, T. Bligaard, S. Helveg,
B. Hinnemann, J. R. Rostrup-Nielsen, I. Chorkendor,
J. Sehested and J. K. Nrskov, J. Catal., 2008, 259, 147.
1222 J. Cheng and P. Hu, J. Am. Chem. Soc., 2008, 130, 10868.
1223 J. Cheng, P. Hu, P. Ellis, S. French, G. Kelly and C. M. Lok,
J. Phys. Chem. C, 2008, 112, 6082.
1224 J. Cheng, P. Hu, P. Ellis, S. French, G. Kelly and C. M. Lok,
J. Catal., 2008, 257, 221.
1225 M. Saeys, M. F. Reyniers, M. Neurock and G. B. Marin,
J. Phys. Chem. B, 2005, 109, 2064.
1226 D. M. Bird, Faraday Discuss., 1998, 110, 335.
1227 D. G. Truhlar, Faraday Discuss., 1998, 110, 521.
1228 F. Labat, P. Baranek, C. Domain, C. Minot and C. Adamo,
J. Chem. Phys., 2007, 126, 154703.
1229 F. Labat, P. Baranek and C. Adamo, J. Chem. Theory Comput.,
2008, 4, 341.
1230 S. Choomwattana, T. Maihom, P. Khongpracha, M. Probst and
J. Limtrakul, J. Phys. Chem. C, 2008, 112, 10855.
1231 A. Kuc, T. Heine, G. Seifert and H. Duarte, Theor. Chem. Acc.,
2008, 120, 543.
1232 F. Abild-Pedersen, O. Lytken, J. Engbaek, G. Nielsen,
I. Chorkendor and J. K. Nrskov, Surf. Sci., 2005, 590, 127.
1233 J. Zhang, M. B. Vukmirovic, Y. Xu, M. Mavrikakis and
R. R. Adzic, Angew. Chem., Int. Ed., 2005, 44, 2132.
1234 S. Kandol and J. Greeley, Catal. Today, 2006, 111, 52.
1235 A. Kokalj, N. Bonini, S. de Gironcoli, C. Sbraccia, G. Fratesi
and S. Baroni, J. Am. Chem. Soc., 2006, 128, 12448.
1236 D. Torres, S. Gonzlez, K. M. Neyman and F. Illas, Chem. Phys.
Lett., 2006, 422, 412.
1237 D. Loreda, F. Delbecq, F. Vigne and P. Sautet, J. Am. Chem.
Soc., 2006, 128, 1316.
1238 C. Morin, D. Simon and P. Sautet, Surf. Sci., 2006, 600, 1339.
1239 M. C. Valero, P. Raybaud and P. Sautet, J. Phys. Chem. B, 2006,
110, 1759.
1240 A. Salameh, J. Joubert, A. Baudouin, W. Lukens, F. Delbecq,
P. Sautet, J. M. Basset and C. Coperet, Angew. Chem., Int. Ed.,
2007, 46, 3870.
1241 Z. Gu and P. B. Balbuena, J. Phys. Chem. C, 2007, 111, 17388.
1242 S. Sharifzadeh, P. Huang and E. Carter, J. Phys. Chem. C, 2008,
112, 4649.
1243 Z. Gu and P. B. Balbuena, J. Phys. Chem. C, 2008, 112, 5057.

Phys. Chem. Chem. Phys., 2009, 11, 1075710816 | 10815

Published on 21 October 2009. Downloaded by Universidade Federal de Santa Catarina (UFSC) on 11/11/2015 23:21:35.

View Article Online

1244 Y. Pluntke, L. A. Kibler and D. M. Kolb, Phys. Chem. Chem.

Phys., 2008, 10, 3684.
1245 Y. Ma and P. B. Balbuena, J. Phys. Chem. C, 2008, 112, 14520.
1246 J. R. Croy, S. Mostafa, L. Hickman, H. Heinrich and
B. R. Cuenya, Appl. Catal., A, 2008, 350, 207.
1247 F. o. Delbecq and F. Zaera, J. Am. Chem. Soc., 2008, 130,
14924.
1248 R. Callejas-Tovar and P. B. Balbuena, Surf. Sci., 2008, 602, 3531.
1249 F. Tian and A. B. Anderson, J. Phys. Chem. C, 2008, 112, 18566.
1250 Y. Gilman, P. B. Allen and M. S. Hybertsen, J. Phys. Chem. C,
2008, 112, 3314.
1251 K.-i. Aika, Y. Ogata, K. Takeishi, K. Urabe and T. Onishi,
J. Catal., 1988, 114, 200.
1252 C. G. M. Hermse, A. P. van Bavel, A. P. J. Jansen, L. A. M.
M. Barbosa, P. Sautet and R. A. van Santen, J. Phys. Chem. B,
2004, 108, 11035.
1253 S. Kandoi, J. Greeley, M. Sanchez-Castillo, S. Evans,
A. Gokhale, J. Dumesic and M. Mavrikakis, Top. Catal.,
2006, 37, 17.
1254 J. Greeley and M. Mavrikakis, J. Am. Chem. Soc., 2004, 126,
3910.
1255 K. Reuter, D. Frenkel and M. Scheer, Phys. Rev. Lett., 2004,
93, 116105.
1256 L. D. Kieken, M. Neurock and D. Mei, J. Phys. Chem. B, 2005,
109, 2234.
1257 A. B. Mhadeshwar and D. G. Vlachos, J. Phys. Chem. B, 2005,
109, 16819.
1258 P. Raybaud, Appl. Catal., A, 2007, 322, 76.
1259 B. Temel, H. Meskine, K. Reuter, M. Scheer and H. Metiu,
J. Chem. Phys., 2007, 126, 204711.
1260 O. R. Inderwildi, S. J. Jenkins and D. A. King, J. Am. Chem.
Soc., 2007, 129, 1751.
1261 H. Falsig, T. Bligaard, J. Rass-Hansen, A. Kustov,
C. Christensen and J. Nrskov, Top. Catal., 2007, 45, 117.
1262 O. R. Inderwildi, S. J. Jenkins and D. A. King, J. Am. Chem.
Soc., 2008, 130, 2213.
1263 N. V. Petrova and I. N. Yakovkin, Eur. Phys. J. B, 2008, 63, 17.
1264 J. Assmann, V. Narkhede, N. A. Breuer, M. Muhler,
A. P. Seitsonen, M. Knapp, D. Crihan, A. Farkas, G. Mellau
and H. Over, J. Phys.: Condens. Matter, 2008, 20, 184017.
1265 J. Rogal, K. Reuter and M. Scheer, Phys. Rev. B: Condens.
Matter Mater. Phys., 2008, 77, 155410.
1266 M. Rieger, J. Rogal and K. Reuter, Phys. Rev. Lett., 2008, 100,
016105.
1267 G. Pacchioni, Nat. Mater., 2009, 8, 167.
1268 S. Vajda, M. J. Pellin, J. P. Greeley, C. L. Marshall,
L. A. Curtiss, G. A. Ballentine, J. W. Elam, S. CatillonMucherie, P. C. Redfern, F. Mehmood and P. Zapol,
Nat. Mater., 2009, 8, 213.
1269 J. Tao, J. P. Perdew, L. M. Almeida, C. Fiolhais and S. Kummel,
Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 245107.
1270 J. Rossmeisl, J. K. Nrskov, C. D. Taylor, M. J. Janik and
M. Neurock, J. Phys. Chem. B, 2006, 110, 21833.
1271 V. Stamenkovic, B. S. Mun, K. J. J. Mayrhofer, P. N. Ross,
N. M. Markovic, J. Rossmeisl, J. Greeley and J. K. Nrskov,
Angew. Chem., Int. Ed., 2006, 45, 2897.
1272 E. Skulason, G. S. Karlberg, J. Rossmeisl, T. Bligaard,
J. Greeley, H. Jonsson and J. K. Nrskov, Phys. Chem. Chem.
Phys., 2007, 9, 3241.
1273 G. S. Karlberg, T. F. Jaramillo, E. Skulason, J. Rossmeisl,
T. Bligaard and J. K. Nrskov, Phys. Rev. Lett., 2007, 99,
126101.
1274 M. B. Vukmirovic, J. Zhang, K. Sasaki, A. U. Nilekar, F. Uribe,
M. Mavrikakis and R. R. Adzic, Electrochim. Acta, 2007, 52,
2257.

10816 | Phys. Chem. Chem. Phys., 2009, 11, 1075710816

1275 A. Nilekar, Y. Xu, J. Zhang, M. Vukmirovic, K. Sasaki,

R. Adzic and M. Mavrikakis, Top. Catal., 2007, 46, 276.
1276 M. Janik, C. Taylor and M. Neurock, Top. Catal., 2007, 46, 306.
1277 L. L. Wang and D. D. Johnson, J. Phys. Chem. C, 2008, 112,
8266.
1278 P. Ferrin, A. U. Nilekar, J. Greeley, M. Mavrikakis and
J. Rossmeisl, Surf. Sci., 2008, 602, 3424.
1279 M. Subhramannia and V. K. Pillai, J. Mater. Chem., 2008, 18,
5858.
1280 Y. Gohda, S. Schnur and A. Gro, Faraday Discuss., 2009, 140,
233.
1281 M. T. M. Koper, Faraday Discuss., 2009, 140, 11.
1282 D. J. Schirin, Faraday Discuss., 2009, 140, 439.
1283 F. De Angelis, A. Tilocca and A. Selloni, J. Am. Chem. Soc.,
2004, 126, 15024.
1284 C. Di Valentin, E. Finazzi, G. Pacchioni, A. Selloni, S. Livraghi,
A. M. Czoska, M. C. Paganini and E. Giamello, Chem. Mater.,
2008, 20, 3706.
1285 A. M. Czoska, S. Livraghi, M. Chiesa, E. Giamello, S. Agnoli,
G. Granozzi, E. Finazzi, C. D. Valentin and G. Pacchioni,
J. Phys. Chem. C, 2008, 112, 8951.
1286 K. Yang, Y. Dai, B. Huang and M.-H. Whangbo, Chem. Mater.,
2008, 20, 6528.
1287 C. Di Valentin, G. Pacchioni, H. Onishi and A. Kudo, Chem.
Phys. Lett., 2009, 469, 166.
1288 X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891.
1289 J. Zhao, B. Li, K. D. Jordan, J. Yang and H. Petek, Phys. Rev.
B: Condens. Matter Mater. Phys., 2006, 73, 195309.
1290 A. Vittadini, M. Casarin and A. Selloni, Theor. Chem. Acc.,
2007, 117, 663.
1291 L. K. Dash, F. Bruneval, V. Trinite, N. Vast and L. Reining,
Comput. Mater. Sci., 2007, 38, 482.
1292 G. Teobaldi, W. A. Hofer, O. Bikondoa, C. L. Pang, G. Cabailh
and G. Thornton, Chem. Phys. Lett., 2007, 437, 73.
1293 W. R. Duncan, C. F. Craig and O. V. Prezhdo, J. Am. Chem.
Soc., 2007, 129, 8528.
1294 K. Yang, Y. Dai and B. Huang, Chem. Phys. Lett., 2008, 456,
71.
1295 W. R. Duncan and O. V. Prezhdo, J. Am. Chem. Soc., 2008, 130,
9756.
1296 J. Graciani, L. J. Alvarez, J. A. Rodriguez and J. F. Sanz,
J. Phys. Chem. C, 2008, 112, 2624.
1297 P. R. McGill and H. Idriss, Surf. Sci., 2008, 602, 3688.
1298 D. P. Hagberg, J.-H. Yum, H. Lee, F. De Angelis, T. Marinado,
K. M. Karlsson, R. Humphry-Baker, L. Sun, A. Hagfeldt,
M. Gratzel and M. K. Nazeeruddin, J. Am. Chem. Soc., 2008,
130, 6259.
1299 A. Vittadini and M. Casarin, Theor. Chem. Acc., 2008, 120, 551.
1300 T. Marinado, D. P. Hagberg, M. Hedlund, T. Edvinsson,
E. M. J. Johansson, G. Boschloo, H. Rensmo, T. Brinck,
L. C. Sun and A. Hagfeldt, Phys. Chem. Chem. Phys., 2009,
11, 133.
1301 H. Perron, C. Domain, J. Roques, R. Drot, E. Simoni and
H. Catalette, Theor. Chem. Acc., 2007, 117, 565.
1302 S. Li and D. A. Dixon, J. Phys. Chem. A, 2008, 112, 6646.
1303 P. Zapol and L. A. Curtiss, J. Comput. Theor. Nanosci., 2007, 4,
222.
1304 G. Shao, J. Phys. Chem. C, 2008, 112, 18677.
1305 X.-Q. Gong, A. Selloni, O. Dulub, P. Jacobson and U. Diebold,
J. Am. Chem. Soc., 2008, 130, 370.
1306 J. Zhang, M. Zhang, Y. Han, W. Li, X. Meng and B. Zong,
J. Phys. Chem. C, 2008, 112, 19506.
1307 H. Nakamura and K. Yamashita, J. Chem. Phys., 2005, 122,
194706.

This journal is


c

the Owner Societies 2009