Theoretical Study of Specific Solvent Effects on the Optical and Magnetic Properties of Copper(II) Acetylacetonate.

Abstract

Specific and basicity solvent effects on the visible near-infrared electronic transitions and the electron paramagnetic resonance (EPR) parameters of the copper(II) acetylacetonate complex, Cu(acac)(2), have been investigated at the density functional theory level. The computed absorption transitions as well as the EPR parameters show a strong dependence on the direct coordination environment around the Cu(II) complex. High solvatocromic shifts are observed for 3d-3d transitions, with the highest effect observed for the d(z(2))→d(xy) transition, which is red-shifted by 6000 cm(-1) and 9000 cm(-1) in water and pyridine solvent models, respectively. Compared to the electronic g-tensors, the hyperfine coupling constants of the Cu(acac)(2) complex show a more pronounced dependence on the effect of base strength of solvent. Overall, the present methodology satisfactorily models the solvent effect on the optical and magnetic properties of the Cu(acac)(2) complex, and theory and experiment agree sufficiently well to warrant the use of the computed optical and EPR parameters to elucidate the coordination environment of the Cu(II) systems in basic solutions.

Full text

Page 1

Published:February 4, 2011
r2011 American Chemical Society
1331
dx.doi.org/10.1021/jp109826p|J. Phys. Chem. A 2011, 115, 1331–1339
ARTICLE
pubs.acs.org/JPCA
Theoretical Study of Specific Solvent Effects on the Optical and
Magnetic Properties of Copper(II) Acetylacetonate
K. J. de Almeida* and T. C. Ramalho.
Departamento de Química, Universidade Federal de Lavras, CP 3037, Lavras, MG, Brasil
Z. Rinkevicius, O. Vahtras, and H. Ågren
Department of Theoretical Chemistry, Royal Institute of Technology, SE-10691 Stockholm, Sweden
A. Cesar
Departamento de Química, Universidade Federal de Minas Gerais, Avenida Antonio Carlos, 6627, CEP-31270-901, Belo Horizonte,
Minas Gerais, Brasil
ABSTRACT: Specific and basicity solvent effects on the visible
near-infrared electronic transitions and the electron paramag-
netic resonance (EPR) parameters of the copper(II) acetylace-
tonatecomplex,Cu(acac)2,havebeeninvestigatedatthedensity
functional theory level. The computed absorption transitions as
well as the EPR parameters show a strong dependence on the
direct coordination environment around the Cu(II) complex.
High solvatocromic shifts are observed for 3d-3d transitions,
withthehighesteffectobservedforthedz2fdxytransition,which
isred-shiftedby6000cm-1and9000cm-1inwaterandpyridine
solvent models, respectively. Compared to the electronic
g-tensors, the hyperfine coupling constants of the Cu(acac)2
complex show a more pronounced dependence on the effect of base strength of solvent. Overall, the present methodology
satisfactorilymodelsthesolventeffectontheopticalandmagneticpropertiesoftheCu(acac)2complex,andtheoryandexperiment
agreesufficientlywelltowarrant theuseofthecomputedopticalandEPRparameters toelucidate thecoordinationenvironmentof
the Cu(II) systems in basic solutions.
’INTRODUCTION
Solvent effects on optical and magnetic properties of copper(II)
compoundsisamatterofconsiderableinterestasthecopper(II)ion
is second only to iron in prevalence in biological systems, where
most chemical processes take place in aqueous environment.1The
interactionwithsolventmoleculesgivesrisetosignificantchangesin
the electronic and molecular structure of copper compounds,
showing an important role in the determination of their properties
and reactivities as well as in the structure-function relationships of
copperproteinsandenzymes.Acommongoalofongoinginvestiga-
tions is to understand the dynamical behavior of molecules in
various liquid environments. While experimental and theoretical
studies of solvent effects on the chemical properties of organic
compoundsareabundant,2-6onlylittletheoreticalinformationcan
be found regarding transition metal complexes, and, indeed, in-
formation of solvent effects on the optical and magnetic properties
of copper(II) compounds are lacking. The presence of a paramag-
netic metal center is probably the limiting factor of such investiga-
tionsasitrequiresacarefulselectionofthecalculationmethodology
employed in order to ensure reliable results.
Significant solvatocromic effects has been experimentally ob-
served on the optical and magnetic spectra of bis(acetylacetonato)-
Figure 1. Molecular structure of bis(acetylacetonato)copper(II), Cu-
(acac)2.
Received:
Revised:
October 13, 2010
December 30, 2010

Page 2

1332
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
copper(II),Cu(acac)2(Figure1).7-10Inparticular,Belfordandetal.9
studied the visible and near-infrared absorption spectra of this
chelate compound in a number of different basic solvents. The
behaviorofabsorptionbandsonalterationofthesolventwasshown
to be consistent with the crystal-field splittings. In another study
carried out by Adato and Eliezer,10both the g-tensor and the
anisotropyofthenuclearhyperfineinteractiontensorsofCu(acac)2
showedaconsiderablevariationwithrespecttosolventsofdifferent
basicities.Theinterpretationofthesolventeffectinthesestudieshas
beenbasedintermsoftheinfluenceonthespectrafromtheligand-
field of the closest surrounding solvent molecules. Crystal- and
ligand-field theories have thus provided a standard qualitative
picture of the environment effect on the optical and magnetic
spectraofchelatecomplexes.7However,somequestions,suchasthe
operating mechanism of solvent molecules in Cu(II) systems and
thecoordinationnumberoftheCu(II)ioninthefirstsolvationshell
of Cu(acac)2in basic solvents, remain poorly understood. While a
structure with two solvent molecules coordinated at the axial
position of Cu(acac)2has been claimed by Belford and et al.9as a
classical model of the Jahn-Teller distortion, an alternative co-
ordination with only one axial molecule in the first solvation shell
was proposed by Ortolano and et al.8to explainthe high intensities
observed in the visible absorption bands of Cu(acac)2in basic
solvents, such as pyridine and piperidine.
On the theoretical side, the relationship between the spectral
and structural features can be successfully achieved by means of
systematiccalculationsinvolvingdifferentgeometriesofthelocal
coordination environment around the metal complex. However,
veryfewtheoreticalinvestigationshaveyetbeenreporteddealing
with solvent effects on the optical and electron paramagnetic
resonance (EPR) spectra of the copper(II) systems. Ames and
Larsen have recently reported some studies by using the density
functionaltheory (DFT)methodsfortheEPR parameters ofthe
copper(II) systems.11,12In particular, these authors published
the EPR parameters of Cu(acac)2within the tetradydrofurane
and pyridine solvent models. Their results reproduce the experi-
mentally observed trends in the parallel components of the
A-andg-tensors,providingimportant insights intothestructural
basis for the empirical trends in EPR parameters of these
systems.12Previous investigations have also been applied to
investigate the optical and EPR properties of the unsolvated
Cu(acac)2system.13-17We reported the computed visible electro-
nic transitions as well as EPR parameters for Cu(acac)2.13,14The
DFT results showed a high dependence of the four nondegenerate
3df3d excitation energies on the molecular distortion of the acac
ligand field, while spin polarization effects on the electronic g- and
A-tensors were investigated in detail for a series of square planar
copper(II) complexes, including Cu(acac)2complex.13,14In addi-
tion, the DFT methods developed by Neese for EPR parameters
have shown a good applicability for Cu(acac)2and of a class of
Cu(II) compounds in the gas-phase.15,16Saladino and Larsen have
reported the EPR parameters of the Cu(acac)2and [Cu(ox)2]2-
complexesattheDFTlevel,includingthescalarrelativisticandSOC
effects.17
In this work, we report an investigation of solvent effects on
thevisiblenear-infraredabsorptionEPRspectraofCu(acac)2in
basic solution. The main focus of this study is to assess
information about the solution environment of Cu(acac)2by
means of the computed optical and EPR parameters, where a
description of the first solvation shell around the chelate
compound is explicitly considered. The pyridine solvent was
chosen due to the experimental and computed results available
for comparative purposes, so that the accuracy of the recently
developed methods of our group could be assessed. The effect
ofsolventbasicityisalsotakenintoaccountintheworkwiththe
aim of evaluating the origin of these changes for the copper(II)
chelate systems. For this purpose, we have selected the water
due to its lower basicity, compared to that of pyridine, and also
considering the importance of the hydration process of copper-
(II) compounds.
’COMPUTATIONAL DETAILS
We have used two models for describing the first solvation
shell around Cu(acac)2(acac = 2,4-pentanedione). The geome-
try optimizations were performedbyexplicitly attaching oneand
two solvent molecules at the axial position of the chelate
complex. The cis and trans conformations of the solvent mole-
culesintheaxialpositionoftheacaccomplex(seeFigure2)were
considered in the calculations of the optical and EPR spectra.
The description of the first solvation shell by using the super-
molecule model has proved to be quite reliable for copper(II)
systems since the relatively strong and specific interactions
take place between Cu(acac)2and basic solvents.18Further-
more, the solvent effects due to long-range interactions
have been evaluated by means of PCM methodology in our
previous work,19and the obtained results showed that this effect
is very small on the Cu(II) aqua and Cu(acac)2compounds
(<5%) and can, therefore, be neglected in the present calcula-
tions.
Geometry optimization processes of the Cu(acac)2solvated
models were carried out with no symmetry constraint at the
B3LYP level in the GAMES-US program.20The standard
Gaussian 6-31G(d) basis set was used for the copper atom,21
while 6-311G** was employed for the oxygen and hydrogen
atoms.22The spin-restricted open-shell density functional linear
response (DFT-RL) formalism23was employed in the calculations
of absorption transitions. The hybrid Becke3-Lee-Yang-Parr
(B3LYP) exchange-correlation functional24,25was also used in
these calculations. The present methodology has provided reli-
ableand consistent predictions forthe excitation energies for the
Cu(II) aqua and acac complexes13,19and consequently enables
us to conduct an accurate evaluation of these properties in the
Cu(acac)2systems.
The electronic g-tensor and hyperfine coupling tensor meth-
odologies have been described in detail in our previous paper.18
The electronic g-tensors were computed using the restricted
densityfunctionalapproach,whilearestricted-unrestrictedmethod,
evaluating the Fermi contact and spin-dipolar contributions, was
used in the hyperfine coupling constant calculations. Higher order
spin-orbit contributions to the hyperfine coupling tensors were
computedusingspin-restrictedDFT-RLtheory.Intheevaluationof
this contribution, the spin-polarization effects were neglected since
theyhavebeenobservedtobeofminorimportanceinourprevious
studies.14,26All calculations of EPR parameters described above
have been performed using the modified Beck3-Lee-Yang-Parr
with38%ofexactexchange(B38LYP)functionalreparametrizedby
Solomon et al.27,28This functional is known to reproduce the
correctbalancebetweenionicandcovalentbondingcharacterinCu
d9complexes, like the Cu(II) compounds here. In the electronic
EPRcalculations,weemployedtheIGLO-IIbasissetfortheoxygen
and hydrogen atoms and an extension of uncontracting the s-type
functions in the IGLO-II basis,29,30adding two tight s-type func-
tions. For the copper ion, the CP(PPP) basis set designed by

Page 3

1333
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
Neese31for accurate calculations of the hyperfine coupling in
transition metal compounds was used. Our choice of basis set is
not compatible with the standard routines for AMFI spin-orbit
operator matrix elements implemented in DALTON 2.0, and
in order to overcome this limitation, we employed the code
provided by Schimmelpfennig,32which is capable of evaluat-
ing AMFI SO matrix elements for arbitrary basis sets. All calcula-
tions for the absorption transitions and EPR spin Hamiltonian
parameters were carried using the DALTON 2.0 quantum chem-
istry package.33
’RESULTS AND DISCUSSION
Complex Geometries. The optimized molecular structures
of the Cu(acac)2-pyridine complexes are shown in Figure 2.
The computed and experimental bond lengths and bond angles
of theCu(acac)2systems arecollectedinTable1. The structures
of Cu(acac)2-pyridine systems are characterized by strongly
coordinatedatomsofacacligandsatequatorialposition,whereas
one or two solvent molecules are weakly coordinated at axial
position, leading to the distorted square pyramidal and octahe-
Table 1. Bond Lengths (in Å) and Bond Angles (in degrees) of the Cu(acac)2Systems
Cu(acac)23(py)Cu(acac)23(py)2
Cu(acac)2
parametersa
cistransother workb
cistransother workb
C2hc
C2hd
other workb
Cu-O
C-O
C-CH
C-CH3
Cu-N(Solv)
r
r0
—Cu-O-C
—O-C-CH3
—O-C-CH
—O-Cu-O
—O0-Cu-O0
arandr0parameterscorrespondtooxygen-oxygendistancesasshowninFigure1.bTheB3LYPoptimizedparametersofAmesandLarsenfromref12.
cOur previous B3LYP results from ref 13.dExperimental results of neutron diffraction in gas phase from ref 37.
1.953
1.259
1.401
1.508
2.319
2.746
2.756
128.3
115.4
125.2
89.8
89.4
1.952
1.260
1.400
1.508
2.305
2.732
2.772
127.7
115.3
125.4
90.5
88.8
1.967 1.970
1.255
1.404
1.510
2.510
2.801
2.772
128.4
115.5
125.3
89.4
90.6
1.976
1.258
1.401
1.510
2.449
2.781
2.809
127.1
115.4
125.7
90.6
89.4
1.9811.913
1.272
1.402
1.516
1.914 ( 0.002
1.273 ( 0.002
1.402 ( 0.003
1.512 ( 0.004
1.940
2.397 2.579
125.1
115.5
124.9
92.4
87.6
124.8 ( 0.4
115.7 ( 1.0
124.4 ( 0.5
92.3 ( 0.9
Figure 2. Optimized molecular structures of the Cu(acac)2-pyridine complexes.

Page 4

1334
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
dral atomic arrangements, respectively. As shown in Table 1, our
previous results for the bond lengths and bond angles of the gas-
phase Cu(acac)2complex reproduce quite well the experimental
gas-phasedataofthiscomplex,withthemaximumerrorsof0.004
Å and 0.5 degree for bond lengths and bond angles, respectively.
The B3LYP result reported by Ames and Larsenindicate a larger
Cu-O bond length for Cu(acac)2, with a deviation of 0.03 Å.12
A comparison between the solvated and nonsolvated structures
gives us some important information about the solvent effect on
the geometrical parameters of this complex. Overall, the coordi-
nationofthe solvent molecules attheaxialposition ofCu(acac)2
give rises to significant changes in its Cu-O bond lengths and
O-Cu-O bond angles, whereas the other geometrical para-
meters of acac ligands remain nearly unchanged. The results
show that a progressive increase of 0.04 Å and 0.06 Å in the
Cu-O bond lengths is observed in the Cu(acac)23(py) and
Cu(acac)23(py)2systems, respectively. A smoother trend was,
however, verified by Ames and Larsen, in which larger absolute
values (0.06 Å in average) are computed for the Cu-O and
Cu-N bond lengths relative to the present results. These geo-
metrical differences can be taken as an indication of the use of
different atomic basis sets. The increasing behavior of Cu-O
bondlengthswiththeaxialcoordinationhasbeenexperimentally
observed in crystal structure analysis of the anhydrous and
hydratescopper(II)hexafluoroacetylacetonatecomplexesaswell
as in the aqueous hydrated copper(II) ion.12,18,34
The Cu atom in cis and trans mono-pyridine complexes are
displaced 0.2 Å and 0.15 Å, respectively, out of the plane of the
four acac O atoms toward the coordinated pyridine molecule.
Comparing the energetics of solvent complex models, the trans
conformations are more stable than cis geometries in a range
between2 and 4 cal mol-1. By analyzing the B3LYP total atomic
spin densities, which can be taken as a measure of covalent
delocalization for the single unpaired 3d-electron over the Cu-
(acac)2complex, a range of values varying from 0.28 electrons
(Cuspindensity0.70e)to0.25electrons(Cuspindensity0.73e)
were computed to the Cu(acac)2-pyridine complexes. These
quantities are equally distributed from the copper(II) ion onto
the oxygen atoms of the acac ligands in the solvated Cu(acac)
complexes. The total atomic spin density on the N atoms of
solvent molecules is not higher than 0.001e in all the complexes
investigated.Thuswecanconcludethatthecovalentcharacterof
chemical bonds of Cu(acac)2are restricted to the equatorial
bonds, whereas weak electrostatic interactions of solvent mole-
cules should take place at the axial positions of this complex.
The Visible Absorption Transitions. The computed and
experimental excitation energies and oscillator strengths of
the Cu(acac)2-pyridine complexes are listed in Table 2. Four
nondegenerateelectronictransitionswerecomputedirrespective
of whether Cu(acac)23(py) or Cu(acac)23(py)2complexes are
considered.Thisisanimportantresultsincetherestillremainsin
theliteratureadivergentviewaboutthenumberoftheelectronic
transitions in the visible absorption spectrum of Cu(acac)2, which
is not well resolved in isotropic media. Whereas Belford et al.9
used a model of three bands to interpret the absorption transitions
in different solutions, Ortolano and Funck8give their inter-
pretation in terms of a model of four bands. The DFT results of
monopyridine Cu(acac)2complexes show a better agreement
with the experimental results of Ortolano and Funck, with an
averagedeviationof335cm-1andthelargesterrorsof541cm-1
being observed for transition II of the cis conformation. The
calculated absorption positions in the Cu(acac)23(py)2complexes
are localized at lower energies than those in the Cu(acac)23(py)
complexes. This behavior becomes more pronounced for the
transition I, which shows an average shift of about 3000 cm-1
when two Cu(acac)2-pyridine models are compared.
With regard to absorption intensities, the oscillator strengths
computed show values equal to zero for all electronic transitions
oftheCu(acac)23(py)2complexes(D2hsymmetry),whereasthe
oscillator strengths of two transitions in the noncentrosymetric
Cu(acac)23(py) complexes (C2vsymmetry) are computed be-
tween 3.0 ? 10-4and 11.0 ? 10-4. These results arise from the
factthatthe3df3dtransitionsinthecopper(II)compoundsare
forbidden by Laporte selection rule (Δl = ( 1) and, conse-
quentlythe computedelectronicoscillatorstrengthsare allequal
to zero in the D2hand C2hsymmetries as well as in the two
transitions of the C2v(Cu(acac)2-(py)) structures since vibro-
nic coupling effects are not included in the present calculations.
TheB3LYPvaluesofthetransconformationagreeverywellwith
the experimental of results, indicating, therefore, that the elec-
tronic contribution should be the dominating part of the experi-
mental intensities observed for these transitions. The visible
near-infrared excitation energies in the acac pyridine complexes
are assigned to be 3df3d-based transitions localized mainly on
thecopper(II)ion.Theoptimizedmolecularorbitalsofthetrans-
Cu(acac)23(py) complex are displayed in Figure 3. The same
ordering of 3d-orbitals is obtained for both Cu(acac)2-pyridine
models. The ground-state dxy, which is well established experi-
mentally for Cu(acac)2and the related acac complex,35was
correctly accounted for by B3LYP calculations of these com-
plexes.Thedxy>dz2>dyz>dxz>dx2-y2orderisinfullagreement
with the experimental assignment of Ortolano and Funck for the
absorption spectrum of Cu(acac)2in pyridine solution.8
Table 2. Experimental and Calculated Excitation Energies (in cm-1) and Oscillator Strengths in the Cu(acac)2-Pyridine
Complexes
Cu(acac)23(py)Cu(acac)23(py)2
cistrans cis transexp.
transitionE (cm-1)fa
E (cm-1)fa
E (cm-1)fa
E (cm-1)fa
sol.b
sol.c
fc,b
I
II
III
IV
10535
13241
15031
15534
0.0
3.0
10.6
0.0
10481
13169
15216
15867
0.0
5.0
7.1
0.0
8100
12265
13546
14713
0.0
0.0
0.0
0.0
6718
11803
13946
14861
0.0
0.0
0.0
0.0
10200
12700
14900
15200
3.6
4.7
7.7
2.2
12100
14800
15100
aOscillator strength ? 10-4.bExperimental data of Cu(acac)2in the pyridine solution from ref 9.cExperimental data of Cu(acac)2in the pyridine
solution from ref 8.

Page 5

1335
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
Overall, the results discussed above show that each model of
the solvated complex gives rise to a particular spectral pattern of the
absorption spectrum, where both positions and intensities of the
visible near-infrared electronic transitions canbe used to characterize
thecoordinationenvironmentoftheCu(acac)2complexinsolution.
Theresultsoftheelectronicexcitationsindicatethesquarepyramidal
Cu(acac)2-pyridinesystemasthebestmodeltodescribetheexperi-
mental visible absorption spectrum of Cu(acac)2in pyridine.
The Electronic g- and A-Tensors. The computed and ex-
perimental EPR parameters of the Cu(acac)2-pyridine com-
plexes are tabulated in Table 3 and Table 4, respectively. As
becomes evident from the presented results in Table 3, the
calculated electronic g-tensors of all complexes have the same
orderingofindividualcomponents,i.e,gzz>gyy≈gxx,showing a
complete agreement with the experimental data.10A systematic
shift to lower values is observed for all components of g-tensors
going from the Cu(acac)23(py)2to the Cu(acac)23(py) models.
ThistrendcanbecorrelatedwiththeincreasedCu-Obondlengths
of the optimized complexes as the successive coordination of one
and two solvent molecules takes place at the axial position of
Cu(acac)2. This feature has already been discussed by Ames and
Larsen for some tetragonal Cu(II) model complexes with oxygen
Figure 3. The B3LYP molecular orbitals involved in 3d-3d electronic transitions of the monofold-coordinated Cu(acac)2-pyridine complex.
The symmetry of the d-orbitals within the C2vpoint group is shown in parentheses.
Table 3. Experimental and Calculated g-Tensors (in ppt) in the Cu(acac)2-Pyridine Complexes
Cu(acac)23(py)Cu(acac)23(py)2
cistransother worka
cistransother worka
g tensorB38LYPB38LYP BP86B3LYPB38LYPB38LYP BP86B3LYPexp.b
gxx
gyy
gzz
go
2.0839
2.0848
2.2889
2.1525
2.0835
2.0828
2.2824
2.1496
2.1176
2.1083
2.3211
2.1823
2.1132
2.1056
2.3245
2.1845
2.0859
2.0859
2.2736
2.1485
2.1272.1902.1352.201
aThe calculated absolute g)tensors of Cu(acac)2-pyridine complexes from ref 12.bExperimental results of Cu(acac)2in pyridine solution (in cm-1)
from ref 10.
Table 4. Hyperfine Coupling Constants of63Cu Dication (in MHz) in the Cu(acac)2-Pyridine Complexes
compoundmodel functionalAxx
Ayy
Azz
Aiso
AFC
Axx
dip
Ayy
dip
Azz
dip
Axx
SO
Ayy
SO
Azz
SO
Cu(acac)23(py)cis
trans
square pyramidal
B38LYP
B38LYP
BP86
B3LYP
B38LYP
B38LYP
BP86
B3LYP
73
69
76
74
-550
-557
475
554
-525
-536
469
542
489
-134
-138
-330
-332
301
301
301
301
-603
-602
105
103
111
111
373
366
other worka
Cu(acac)23(py)2
cis
trans
octahedral
100
96
104
99
-107
-114
-318
-321
305
305
305
305
-611
-610
115
112
122
121
393
385
other worka
experimentb
aThe calculated absolute A)tensors of Cu(acac)2-pyridine complexes from ref 12.bExperimental results of Cu(acac)2in pyridine solution (in cm-1)
from ref 1010.
8.848.84169

Page 6

1336
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
ligands.12That is, the shorter Cu-O bond lengths of unsolvated
Cu(acac)2, as compared to those of Cu(acac)2-pyridine system,
are related to the greater sharing of electron density between the
coppermetalcenterandtheacacligands,leadingtoadelocalization
of electron density on the Cu(II) ion, with a concomitant decrease
in the g-tensors of the former systems. It is worth noting that
this tendency is observed in the experimental data in the Peisach-
Bumberg truth tables.36
The individual g-tensor values of cis and trans structures are
slightly different from each other, whereas a larger deviation is
observed in the isotropic values of these conformations. A more
pronounceddifferenceisobservedforallcomponentsofg-tensors
ofmono-anddicoordinatedCu(acac)2-pyridinecomplexes.The
largest component, gzz, has a dominantly large local spin-orbit
contribution and is shifted 0.04 ppt from Cu(acac)23(py) to
Cu(acac)23(py)2systems, showing strong dependence with re-
spect to the number of solvent molecules coordinated at the axial
positionofCu(acac)2.Thegxxandgyycomponentsoftheg-tensor
undergoaslightlylowershiftof0.03pptinthismolecularcomplex
sequence. The best agreement with experimental results is
obtained for the calculated values of the trans-Cu(acac)23(py)
complex. The major discrepancy is observed for the gzzvalue of
this complex, which is overestimated by 0.01 ppt relative to the
experimental value. Previous DFT calculations underestimate the
g)tensors of Cu(acac)2-pyridine complexes, in about 0.15 ppt,
showing only a slight difference between g)values of square
pyramidal and octahedral models, independently of the BP86
and B3LYP exchange-correlation functionals employed. This
might be due to the fact that the conventional exchange-correla-
tionfunctionalsstillremainseverelyhandicappedintheirabilityto
describe EPR parameters of transition metal complexes. It is
important to note that for the previous calculations cited, the
unrestricted Kohn-Sham (UKS) formalism was used, whereas
the present results of the g-tensor were obtained using the
restricted Kohn-Sham (RKS) approach. A comparison with
experimental data available shows that the present B38LYP/
RKS calculations are able to describe very well the individual
components as well as the isotropic values of the g-tensors of the
Cu(acac)23(py)andCu(acac)23(py)2complexes,clearlyfavoring
the5-fold-coordinatedsystemformodelingtheabsorptionspectra
of Cu(acac)2in pyridine solution.
The computedCu(II)hyperfinecoupling tensorresultsof the
Cu(acac)2-pyridine complexes in Table 4 show a unique
ordering of principal values as that found in the electronic
g-tensor results, i.e., |Azz| > |Ayy| > |Axx|. An inspection of
Table 4 reveals a small variation, lower than 10 MHz, for the
individual and isotropic components of the cis and trans con-
formations. Comparing Cu(acac)23(py) and Cu(acac)23(py)2
complex models, however, the differences between the com-
puted hyperfine coupling values become more pronounced
around 25 MHz. A progressive increase of the absolute values
is observed in the Axxand Ayyvalues with the subsequent axial
coordination of one and two pyridine molecules in Cu(acac)2,
whereas a decrease of the Azzand Aisocomponents is found in
this sequence of complexes. The comparison between absolute
values is valid since in the experimental EPR measurements, one
can only determine the absolute values of the EPR hyperfine
parameters. It is worth noting that the absolute A)values of the
Cu(acac)2system, reported by Ames and Larsen, show the same
decreasing behavior as solvent models are considered for
the Cu(II) model with oxygen ligands.12It is interesting to note
that the B3LYP results are very close to the present results of
A-tensors, whereas a higher deviation is observed with respect to
the BP86 results. These results indicate once again the high
dependence of A-tensors in terms of the exchange-correlation
functional employed.
The contributions of the various components of the copper
hyperfine coupling tensors should be understood so that their
impact on calculation results for the hyperfine coupling becomes
clear. In conventional hyperfine coupling calculations, only the
Fermi contact and spin-dipolar contributions are evaluated. This
procedure is well suited for organic radicals, but fails severely for
transition metal compounds, where the spin-orbit contribution
becomes quite important. For Cu(acac)2-pyridine complexes,
the Fermi contact term (AFC) varies from 330 MHz Cu(acac)23
(py) to 320 MHz in the Cu(acac)23(py)2complexes. In the spin
dipolar contributions (Aii
follow a |Azz| > |Ayy| ≈ |Axx| pattern (where Azzis negative) in
both models of the pyridine acac complexes. The ordering and
sign patterns of the spin-dipolar contributions are in agreement
withthepatternobservedinthetotalhyperfinecouplingtensors.
Thefollowingorderingof|Azz|>|Axx|>|Ayy|isobservedforthe
spin-orbit contribution (Aii
indeedplaysacrucialrole forboth principal values ofA(Cu)and
Aiso(Cu). For the principal values, the spin-orbit contribution
(being opposite in sign to the Fermi contact contribution and of
similar magnitude) effectively counteracts the Fermi contact
contribution in the total hyperfine coupling tensor. Dramatic
effects of the inclusion of spin-orbit coupling are observed for
the isotropic hyperfine coupling constants, thus indicating that
the spin-orbit interaction is of major importance for a reliable
description of the hyperfine coupling constants of the Cu(acac)2-
pyridine complexes.
As shown in Table 4, a comparison with experimental results
shows that the absolute values of the isotropic Cu(II) hyperfine
coupling tensors appear to be the best EPR parameter to be
exploited to judge the coordination environment of the first
solvation shell around Cu(acac)2 in pyridine solution. The
computed values of the monopyridine Cu(acac)23(py) system
show a better qualitative agreement with the experimental data
available.Anaveragedeviationof35MHzfromtheexperimental
data is observed, likely caused by the absence of vibrational
environmental effects, which would probably improve the com-
putational results of these systems. Finally, the solvent effects on
the EPR parameters of Cu(acac)2 might be explained as a
consequence of the localization of electron density on Cu(II),
mainly caused by the increase of the Cu-O bond lengths of
Cu(acac)2due to the subsequent axial coordination of solvent
molecules in Cu(acac)2. As a consequence from this fact,
relatively higher g-tensors and smaller absolute values of Aiso
are expected for the Cu(acac)2-pyridine complexes relative to
thoseobservedinthegas-phaseCu(acac)2system.Acomparison
between experimental and computed Aisovalues of Cu(acac)23
(py) complexes gives support to our conclusions obtained from
calculations of the electronic transitions and g-tensors, which
point out the mono-pyridine complex as the most suitable
model for describing the optical and magnetic properties of
Cu(acac)2inpyridinesolution.Insummary,thegoodagreement
between calculated and experimental results of the visible near-
infrared transitions and EPR parameters shows that the present
methodology is indeed consistent for modeling the solvent
effects on the optical and magnetic properties of the Cu(acac)2
complex in pyridine solution.
dip, i = x, y, z), the Adipcomponents
SOin Table 4, i = x, y, z), which

Page 7

1337
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
Solvent Basicity Effects. In order to evaluated the basicity
solvent effects, we have selected the square pyramidal conforma-
tion of solvated complexes due to its better performance to
describe the experimental data of optical and EPR spectra of
Cu(acac)2 in pyridine. The B3LYP results for the 3d-3d
electronic transitions of Cu(acac)2in the gas-phase, water and
pyridine solvent models are displayed in Figure 4. Overall, the
results show that the coordination of solvent molecules at the
axial position of Cu(acac)2acts to gradually shift each 3df3d
electronic transition to lower energies as compared to those in
the gas phase, where no solvent molecules are coordinated at the
axial position. It is worth noting that the extension of the
computed shifts correlates with the basicity of the considered
solvents, showing lower redshift values in water than in the
pyridine solvent model. This feature might be ascribed to the
more efficient interaction of the nonligand pz(N) orbital of
pyridine, as compared to pz(O) orbital of water, with the paired
electrons in the dz2 orbital of Cu(acac)2(See Figure 3). Another
interestingresultisthatdistinctsolvatocromicshiftsarefoundfor
four 3df3d electronic transitions. The ordering of the 3d
orbitals changes from dxy< dxz< dx2-y2 < dyz< dz2 in the gas-
phase to dxy< dz2 < dxz< dyz< dx2-y2 in the solvated model
structures of water and pyridine Cu(acac)2complexes, for which
the same pattern of 3d orbital order is found. The dx2-y2fdxy
transition undergoes the smallest solvatocromic shift, remaining at
about 16500 cm-1, in agreement with the experimental referred
bandofCu(acac)2indifferentsolutions.9,10Ontheotherhand,the
more pronounced solvatocromic effect is observed for the dz2fdxy
transition, whichis red-shiftedabout6000 cm-1and 9000cm-1in
the water and pyridine complex models, respectively. The latter
result is in very good agreement with the experimental value in
pyridine(8600cm-1)reportedbyOrtolanoandFunck.8Thisresult
may be caused by the direct interactions of the pzorbitals of the
donor oxygen and nitrogen atoms of solvents with the dz2(Cu)
orbitalsasshowninFigure3.Finally,wecanseefromthisfigurethat
there is no direct interaction of the dxzand dyzorbitals with the
solvent orbitals and the analogous solvatocromic shifts computed
for the dxzfdxyand dyzfdxytransitions can be taken as indication
of this fact. It is worth noting that all above-discussed results are in
complete agreement with the experimental Cu(acac)2spectra in
solution.
The electronic g-tensors and absolute values of A-tensors of
Cu(acac)2inthegas-phaseandaquaandpyridinesolventmodels
are shown in Figures 5 and 6, respectively. Regarding the change
from gas-phase to solvent models, these figures indicate a similar
and smooth increase of the electronic g-tensors as well as in the
AxxandAyy,whereasapronounceddecreaseisverifiedinAzzand
Aiso. These trends are correlated with those observed in EPR
results as one and twopyridines coordinated at the axial position
of Cu(acac)2. The calculated Cu-O bond lengths of the square
pyramidalaquaandpyridineCu(acac)2solventmodelsare1.932
Åand1.953Å,respectively.TheCu-O(water)bondlengthsare
computed tobe 2.326 Å, while a value of 2.305 Åis found for the
Cu-N(pyridine) distance. Theseresults show an increaseof the
Cu-O bond lengths with the increase of the basicity solvent.
Thus the observed trends in EPR parameters due to the solvent
effects are related to the localization of electron density on the
transition metal center. That is, the longer the Cu-O bond
distance, the more localized electron density is found on the
Cu(II) center, leading to a increase in the g-tensors, Axxand Ayy
aswellasaconcomitantdecreaseintheabsolutevaluesofAzzand
Aisotensors. Another point relative to the solvent trends in the
EPR and optical parameters of Cu(acac)2is due to differing
copper deviations from the complex plane for the different
solvents. The half-step (θ) parameter, which corresponds to
thedistanceofthecopper(II)ionfromtheacetylacetonateligand
plane, gives us some insights in this respect. The B3LYP predic-
tion for θ in the C2hstructure of Cu(acac)2is 0.27 Å, while an
experimental value of 0.41 ( 0.07 Å has been determined for
Cu(acac)2in the gas phase. The average values in monopyridine
and monowater acac complexes decrease to 0.18 Å and 0.13 Å,
Figure 4. Solvatocromic shifts of the 3d-3d excitation energies of
Cu(acac)2in different phase models.
Figure 5. Solvatocromic shifts of the electronic g-tensors of Cu(acac)2
in different phase models.
Figure 6. Solvatocromic shifts of the absolute Cu(II) hyperfine cou-
pling constants (A) of Cu(acac)2in different phase models.

Page 8

1338
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
respectively, whereas a value equal to zero is found for all com-
plexes with two solvent molecules coordinated at axial position
of Cu(acac)2. These results suggest a correlation between the
solvent trends in the EPR and optical spectra of acac complexes
and the bonding θ parameter.
By careful inspection of the individual results, we can see that
the hyperfine constants show a more pronounced dependence
on the base strength compared to the electronic g-tensor. While
onlyaslightincreaseofallg-tensorcomponentsisobservedwith
the change of environment, a more significant variation in the
nuclear hyperfine interaction constants is found for the isotropic
and major component of A-tensors as going from the gas phase
tothesolvatedstructures.ThecomputedvaluesofAisovaryfrom
-212MHzinCu(acac)2to-183MHzinCu(acac)2-waterand
-138 MHz in the Cu(acac)2-pyridine model. The results arise
mainlyfromthefactthatthemagnitudeofthehyperfinesplitting
dependsonthemagneticmomentofthenucleusandhowtightly
the unpairedelectrons arebound toit.In the basic solvents,such
as water andpyridine,the OandN atoms actas the donor atoms
in most adducts. It is expected, therefore, that the interactions
between the donor solvent and acceptor solute, Cu(acac)2, will
essentially depend on the charge distribution on the donor
atoms. Thus, amines such as pyridine, can provide larger shifts
to lower energy in the ligand field transitions, and are character-
ized by smaller values of the isotropic hyperfine constant, while
solvents having functional oxygen donor atoms being of smaller
basicity are characterized by larger values of the isotropic
hyperfine constant. Overall, these results agree very well with
the experimental findings of Adato and Elizer which were
obtained in a wide range of different basic solvents.10
’SUMMARY
Thepresentpaperofferstheoreticalpredictionsforthesolvent
effects on the optical and magnetic spectra of Cu(II) acetylace-
tonate complex by using quantum chemistry methodologies.
DFT calculations have been performed using the supermolecule
model to describe the direct solvent interactions with the Cu-
(acac)2complex. The results indicate that coordination at the
axial position of Cu(acac)2is favored with a high basicity of the
solvent.Thevisiblenear-infraredabsorptiontransitionsaswellas
the EPR parameters show a strong dependence on the coordina-
tion environment around the copper complex. The B3LYP
results of the 3df3d excitation energies show that each model
of the solvated model complex gives rise to a particular spectral
pattern of the absorption spectrum, where both positions and
intensities of electronic transitions can be used to characterize
the coordination environment of the first solvation shell around
the Cu(acac)2complex in solution. The good comparison with
experimental results indicates that the coordination of the
solvent at axial position of Cu(acac)2should indeed be the main
solventeffectonthepositionsandintensitiesofabsorptionbands
in Cu(acac)2.
The B38LYP calculations give accurate predictions for the
individual components as well as the isotropic values of the
g-tensors of the Cu(acac)2-pyridine model complexes. On
theotherhand,thecomputedresultsofA-tensorsshowaqualita-
tive agreement with the experimental data available. The devia-
tions from theexperimentaldataare likely caused bythe absence
of vibrational environmental effects and also due to the fact that
the conventional exchange-correlation functionals still remain
severely handicapped in their ability to describe EPR parameters
oftransitionmetalcomplexes.Thecomputedexcitationenergies
and EPR parameters give support to the conclusion that the
monofold-coordinated Cu(acac)2-solvent system is the most
suitablemodelfordescribingtheopticalandmagneticproperties
of Cu(acac)2in basic solutions. Compared to the electronic
g-tensors of the Cu(acac)2complex, the hyperfine coupling
constants show a pronounced dependence on the effect of the
base strength of solvent. Overall, the present methodology used
in this investigation correctly reproduces the optical and EPR
properties of the solvated the Cu(acac)2models, providing a
satisfactory and reliable description of the solvent effects on the
optical and magnetic properties of a Cu(acac)2complex.
’REFERENCES
(1) Rorabacher, D. B. Chem. Rev. 2004, 104, 651.
(2) Orozco, M.; Luque, F. J. Chem. Rev. 2000, 100, 4187.
(3) N. S. Hush, N. S.; Reimers, J. R. Chem. Rev. 2000, 100, 775.
(4) Orozco, M.; Luque, F. J. Chem. Rev. 2001, 101, 203.
(5) Schofield, D. P.; Jordan, K. D. J. Phys. Chem. A 2007, 111,
7690.
(6) Liu, W. L.; Zheng, Z. R.; Dai, Z. F.; Liu, Z. G.; Zhu, R. B.; Wu,
W.Z.;Li,A.H.;Yang,Y.Q.;Su,W.H.J.Chem.Phys.2008,128,124501.
(7) Huheey, J. E. Inorganic Chemistry: Principles of Structure an
Reactivity, 3rd ed.; Harper & Row: New York, 1983, Chapter 9.
(8) Ortolano, T. R.; Funck, L. L. Inorg. Chem. 1968, 7, 537.
(9) Belford, R. L.; Calvin, M.; Belford, G. J. Chem. Soc. A 1957, 26,
1165.
(10) Adato, I.; Eliezer, I. J. Chem. Phys. 1971, 54, 1472.
(11) (a) Ames, W. M.; Larsen, S. C. J. Biol. Inorg. Chem. 2009, 14,
547–557. (b) Ames, W. M.; Larsen, S. C. Phys. Chem. Chem. Phys. 2009,
11, 8266–8274. (c) Ames, W. M.; Larsen, S. C. J. Phys. Chem. A 2010,
114, 589–594.
(12) Ames, W. M.; Larsen, S. C. J. Phys. Chem. A 2009, 113, 4305–
4312.
(13) deAlmeida,K.J.;Rinkevicius,Z.;Vantras,O.;Ågren,H.;Cesar,
A. Chem. Phys. Lett. 2010, 492, 14–18.
(14) Rinkevicius, Z.; de Almeida, K. J.; Oprea, C. I.; Vahtras, O;
Ågren, H.; Ruud, K. J. Chem. Phys. 2008, 129, 064109.
(15) (a)Neese,F.Int.J.QuantumChem.2001,83,104.(b)Neese,F.
J. Chem. Phys. 2001, 115, 11080. (c) Neese, F. Int. J. Quantum Chem.
2001, 83, 104.
(16) Neese, F. J. Chem. Phys. 2003, 118, 3939.
(17) Saladino, A. C.; Larsen, S. C. J. Phys. Chem. A 2003, 107, 5583.
(18) de Almeida, K. J.; Rinkevicius, Z; Ferreira, A. C.; Ågren, H.
Chem. Phys. 2007, 332, 176.
(19) de Almeida, K. J.; Murugan, N. A.; Rinkevicius, Z.; Hugosson,
H. W.; Ågren, H.; Cesar, A. Phys. Chem. Chem. Phys. 2009, 11
508–519.
(20) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.;
Gordon,M.S.;Jesen,J.H.;Koseki,S.;Matsunaga,N.;Nguyen,K.A.;Su,
S. J.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem.
1993, 14, 1347–1363.
(21) Rassolov, V.; Pople, J. A.; Ratner, M.; Windus, T. L. J. Chem.
Phys. 1998, 109, 1223–1229.
(22) Krishnan,R.;Binkley,J.S.;Seeger,R.;Pople,J.A.J.Chem.Phys.
1980, 72, 650–654.
(23) Rinkevicius, Z; Tunell, I.; Sazek, P.; Vahtras, O.; Ågren, H.
J. Chem. Phys. 2003, 119, 34–46.
(24) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(25) Lee, C; Yang, W.; Parr, R. G. Phys. Rev. 1988, 37B, 785–789.
(26) Rinkevicius, Z.; Tunell, I.; Salek, P.; Vahtras, O.; Ågren, H.
J. Chem. Phys. 2003, 119, 34–46.
(27) Solomon, E. I.; Szilagyi, R. K.; DeBeer George, S.; Basumallick,
L. Chem. Rev. 2004, 104, 419–458.

Page 9

1339
dx.doi.org/10.1021/jp109826p |J. Phys. Chem. A 2011, 115, 1331–1339
The Journal of Physical Chemistry A
ARTICLE
(28) Szilagyi, R. K.; Metz, M.; Solomon, E. I. J. Phys. Chem. A. 2002,
106, 2994–3007.
(29) Kutzelnigg, W., Fleischer, U., Schindler, M. In NMR Basic
Principles and Progress; Diehl, P, Fluck, E., G€ unther, H., Kosfeld, R.,
Eds.; Springer: Heidelberg, 1990; Vol. 23, pp 165-172.
(30) Huzinaga, S. Approximate Atomic Functions; University of
Alberta: Edmonton, AB, Canada, 1971; pp 38-39.
(31) Neese, F. Inorg. Chim. Acta 2002, 337C, 181–192.
(32) Schimmelpfennig, B. AMFI: An Atomic Spin-Orbit Mean Field
Integral Program; University of Stockholm: Sweden, 1996.
(33) DALTON,a Molecular ElectronicStructure Program, release 2.0;
see http://www.kjemi.uio.no/software/dalton/dalton.html
(34) Thomas, B. G.; Morris, M. L.; Hilderbrant, R. L. J. Mol. Struct.
1976, 35, 241.
(35) Hitchman, M. A.; Belford, R. L. Inorg. Chem. 1971, 10, 984–
988.
(36) Rinkevicius, Z.; Telyatnyk, L.; Sazek, P.; Vahtras, O.; Ågren, H.
J. Chem. Phys. 2003, 119, 34–46.
(37) Shibata, S.; Sasase, T.; Ohta, M. J. Mol. Struct. 1983, 96, 889–
894.