Double-CO32- centered [CoII5] wheel and modeling of its magnetic properties

Article abstract of DOI:10.1002/chem.201001418

A high-spin Co^II cluster with a rare pentagonal molecular structure and formula [Co₅(CO₃)₂(bpp) ₅]ClO₄ (1; Hbpp is 2,6-bis(phenyliminomethyl)-4- methylphenolate) has been synthesized and ch

Full text of DOI:10.1002/chem.201001418

FULL PAPER
DOI: 10.1002/chem.201001418
2ꢀ
II
Double-CO3 Centered [Co ] Wheel and Modeling of Its Magnetic
5
Properties
[
a]
[b]
[c, d]
[e]
Mrinal Sarkar, Guillem Aromꢀ,* Joan Cano,*
Valerio Bertolasi, and
[
a]
Debashis Ray*
II
II
Abstract: A high-spin Co cluster with
ordination numbers (CNs) of the Co
ions of five and six. The bulk magneti-
zation of this complicated magnetically
exchanged system has been modeled
successfully by employing a matrix di-
agonalization technique. For this, the
combination of S=3/2 ions (CN=5)
with ions exhibiting strong spin-orbit
coupling (CN=6) has been considered
and a perturbative approach to handle
the data in the whole studied range of
temperatures (2–300 K) yielding pa-
rameters of g and D (for the five-coor-
a rare pentagonal molecular structure
and formula [Co ACHTUNGTRENNUNG( CO ) ACHTUGNENNTRUG( bpp) ]ClO (1;
5 3 2 5 4
Hbpp is 2,6-bis(phenyliminomethyl)-4-
methylphenolate) has been synthesized
and characterized by single-crystal X-
ray diffraction. This topology arises
II
from fusing five [Co
A
H
U
G
R
N
U
G
dinate Co ions), of A, k, l, and D (for
2
Keywords: carbonate ligands · car-
2
ꢀ
a cyclic manner around two CO3 cen-
tral ligands, resulting in propeller-like
configuration. The irregular coordina-
tion of the carbonate ions to the metal
centers results in a combination of co-
the metals with spin-orbit coupling)
and of the exchange constants J. The
agreement with results from DFT cal-
culations, also presented here, is re-
markable.
bonates
· cobalt · coordination
chemistry · density functional calcu-
lations · magnetic properties · spin-
orbit coupling
Introduction
The self-assembly of transition-metal macrocycles and mo-
lecular spin clusters has become a frontier topic of contem-
porary research in synthetic coordination chemistry. Apart
from having aesthetically pleasing architectures (e.g.
[
a] Dr. M. Sarkar, Prof. D. Ray
Department of Chemistry, Indian Institute of Technology
Kharagpur 721 302 (India)
Fax: (91)3222-82252
E-mail: dray@chem.iitkgp.ernet.in
[1]
[2]
[3]
squares, rectangles, or pentagons ) polymetallic clusters
display unique magnetic properties at the molecular level
[
b] Dr. G. Aromꢀ
[4,5]
such as single-molecule magnet (SMM) behavior,
quan-
Department de Quꢀmica Inorgꢁnica, Facultat de Quꢀmica
Universitat de Barcelona
Diagonal, 647, 08028-Barcelona (Spain)
E-mail: guillem.aromi@qi.ub.es
[6]
tum tunneling of the magnetization, or quantum interfer-
ence. Most of the reported SMMs of 3d ions are based on
[7]
[5]
the metals Mn, Ni, and Fe. The proliferation of new exam-
ples and the investigation of their physical properties have
generally been accompanied by the development of the ap-
propriate models to describe their magnetic and spectro-
scopic behavior, and this process has taken place synergisti-
cally, bringing the state of the art of molecular nanomagnet-
ism to the current level. The case of six-coordinate cobal-
t(II) is special; large single ion anisotropy caused by a first-
order orbital contribution to the magnetic moment makes
[
c] Dr. J. Cano
Instituto de Ciencia Molecular (ICMol) and Fundaciꢂ General de la
Universitat de Valꢃncia (FGUV), Universitat de Valꢃncia
4
6980 Paterna, Valꢃncia (Spain)
E-mail: joan.cano@qi.ub.es
[
d] Dr. J. Cano
Instituciꢂ Catalana de Recerca i Estudis AvanÅats (ICREA)
Passeig Lluis Companys 23, 08010 Barcelona (Spain)
[
e] Prof. V. Bertolasi
Dipartimento di Chimica e Centro di Strutturistica Diffrattometica
Universitꢁ di Ferrara
[8]
this ion a good candidate to make SMMs and single-chain
[9]
magnets (SCMs), but it also strongly hampers the descrip-
tion of its magnetic behavior, which slows down progress in
this direction. Attempts to address this problem have been
made for quite a long time. Originally, only perfectly octahe-
Via Borsari 46, 44100 Ferrara (Italy)
Supporting information (synthesis of Hbpp, details on UV/Vis and IR
spectroscopy) for this article is available on the WWW under http://
dx.doi.org/10.1002/chem201001418.
Chem. Eur. J. 2010, 16, 13825 – 13833
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
13825

[
10]
dral systems were considered, which precluded the inclu-
sion of magnetic anisotropy. The latter has been taken into
account in recent models making use of various diagonaliza-
tion techniques or even providing approximate analytical ex-
pressions of the paramagnetic susceptibility for dinuclear
5 CoðClO
4
Þ
2
2
ꢁ 6H
2
O þ 5 Hbpp þ 9 NaOR þ 2 CO
2
!
½
Co ðCO Þ ðbppÞ ꢂClO þ 9 NaClO þ 9 ROH þ 28 H O; ð1Þ
5
3
5
4
4
2
ðR ¼ Me or HÞ
Interestingly, the use of a weaker base, such as NEt does
not facilitate the trapping of CO and consequently, the re-
action does not proceed. On the other hand, CO3 may be
added in the form of a metallic salt to the reaction mixture,
which also leads to the generation of complex 1. Thus, addi-
II
[11–13]
3
Co complexes.
More recently, some of us have pro-
2
posed a convenient model to treat the magnetic exchange
within polynuclear clusters of axially distorted six-coordi-
nate cobalt(II) ions invoking a perturbational approach,
which greatly reduces the size of the matrices to be diagon-
2ꢀ
tion of Co
ACHTUNGTNERNUNG( ClO ) to a stirred methanol solution of Hbpp
4 2
[14]
alized.
We report here the
and K CO leads to precipitation of the pentanuclear aggre-
2
3
synthesis and structure of a rare
Co ] wheel, assembled by the
gate, as described in Equation (2).
II
[
5
bridging action of the Schiff-
base ligand Hbpp (2,6-bis(phe-
nyliminomethyl)-4-methylphe-
nol, Scheme 1) and the tem-
plate effect of two carbonate
ions, with the formula [Co5-
1
0 CoðClO Þ ꢁ 6 H O þ 10 Hbpp þ 9 K CO !
4
2
2
2
3
ð2Þ
2
½Co ðCO Þ ðbppÞ ꢂClO þ 18 KClO þ 5 CO þ 65 H O
5
3
2
5
4
4
2
2
[15]
Interestingly, the yield of both processes was very similar,
Scheme 1. Structure of Hbpp.
even if the latter was carried out in both, aerobic and anae-
robic conditions. This shows that the formation of 1 consti-
tutes a significant driving force to the capture of CO from
the atmosphere if the basic conditions, necessary in this pro-
cess, are met. On the other hand, no evidence of the forma-
A
H
U
G
R
N
N
(CO )
A
H
U
G
R
N
N
(1). The
3
2
5
2
ligand Hbpp should favor the assembly of dinuclear moieties
with vacant coordination sites, which should therefore be
prone to bind to other exogenous donors, eventually facili-
tating the growth of the system into higher nuclearity spe-
cies. Surprisingly, only mononuclear complexes have been
tion of any cobalt ACHUTNGRNENUG( III) byproduct was observed. The identity
of complex 1 was consistent with the results from elemental
analysis. The complex was characterized spectroscopically,
both by UV/Vis and IR measurements (see Experimental
Section and Figure S1 and S2 in the Supporting Informa-
tion). In particular, IR stretching bands characteristic of co-
[16,17]
reported with this ligand or close derivatives,
even
though this principle has been proven with related Schiff
[18]
bases in the construction of large clusters. This potential
is exploited here, together with the capacity of the carbon-
ꢀ
1
ordinated carbonate are observed at 1552 and 1381 cm ,
ate ligand to adopt
a
large variety of coordination
to access the rare topology featured by complex
. The singularity of the bridging motifs present in this clus-
ꢀ
1
[19–24]
whereas a typical imine signal occurs at 1609 cm . The de-
tails of the molecular structure of this cluster were obtained
from single-crystal X-ray diffraction experiments (see
below).
modes,
1
ter makes it for an interesting species to study its intramo-
lecular magnetic exchange. These interactions are described
here theoretically, following a study by DFT calculations.
Most remarkably, a good simulation of the experimental
bulk magnetization data is provided by using the above-
mentioned model, where both types of metal centers found
Description of the structure: The molecular structure of 1 is
represented in Figure 1, and crystallographic data and se-
lected metric parameters are listed in Tables 1 and 2. This
compound consists of a cationic [Co
plex and a ClO4 counterion. The cation features five Co
ions disposed in a planar quasi-ideal pentagonal arrange-
ment, and connected pairwise by the phenolate donors of
+
A
H
U
G
R
N
N
(CO )
A
H
U
G
R
N
N
(bpp) ] com-
5
3
2
5
II
ꢀ
II
in 1 have been considered; high-spin five-coordinate Co
II
ions together with six-coordinate Co centers displaying
spin-orbit coupling. In this model, the exchange interactions
between the metal centers are also included. The results
from both approaches are in remarkable agreement.
ꢀ
five bpp ligands. These five bridging oxygen atoms are lo-
cated at the apices of another pentagon, approximately cir-
cumscribed around an ideal circle in which the metallic pen-
tagon would be inscribed. Besides linking two metals
ꢀ
Results and Discussion
through the phenolate, each bpp ligand forms an additional
bond to both of these metal centers through its N donor
atoms. The ensemble can thus be described as a cyclic ar-
rangement of five fused [Co2 ACHUTNRGNENUG( bpp)] moieties. To fit around
Synthesis: The Schiff-base ligand Hbpp is readily prepared
quantitatively from the reaction of 2,6-diformyl-4-methyl-
phenol with two equivalents of aniline (see Scheme S1 in
the Supporting Information), as described in the litera-
this pentagonal topology, the ligands are bent and stacked
on top of each other, mutually shifted by angles of about
708, conforming the five blades of an imaginary propeller.
This arrangement is favored by the establishment of intra-
molecular p···p interactions involving each phenyl ring and
the phenolate cycle of the bpp moiety adjacent to it. Thus,
each phenolate ring in fact forms one such interaction with
two phenyl rings, one on each side. The average inclination
[15]
ture. The reaction of Hbpp with an equimolar amount of
Co(ClO ) in the presence of a strong base (NaOMe or
ACHTUNGTRENNUNG
4
2
NaOH) in MeOH or DMF leads to either a precipitate or
crystals of 1, respectively, following CO uptake from the at-
2
mosphere, according to Equation (1).
13826
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13825 – 13833

A High-Spin Cobalt Cluster
FULL PAPER
2ꢀ
These CO3 molecules bind the metal centers (Figure 2) in
a m1,2,3 and m1,1,2,3 mode, respectively ([3.111] and [3.211] in
[25]
Harris notation).
Such coordination modes are ascribed
2
3
ꢀ
Figure 2. Mercury representation of the coordaintion modes of the CO
ligands in 1. The Co ions involved are labeled.
on the basis of what have been considered as bonding Oꢀ
2ꢀ
Co interactions between the metal centers and the CO3 li-
gands. In this complex, these are in the range of 2.033(10) to
2
.224(7) ꢅ (see Table 2 for complete list). The metal–car-
bonate contacts in complex 1, not considered as bonds, are
.420(9) ꢅ or longer. Therefore, the coordination geometries
2
of the metals in this cluster are either distorted octahedral
+
] .
Figure 1. PovRay representation of the cation of 1, [Co
5
A
H
U
G
R
N
U
G
3
)
2
A
H
U
G
R
N
N
5
with a cis-N O environment (Co1 and Co4) or distorted
2
4
Hydrogen atoms are not shown for clarity. Only non-carbon atoms and
square pyramidal with an N O environment and axial N
2
3
2ꢀ
ꢀ
CO
3
ligands are labeled. For clarity, the various bpp ligands are repre-
atoms (Co2, Co3, and Co5). The CoꢀO distances to the
sented in different gray scales.
phenoxide oxygen atoms range from 1.990(7) to 2.012(6) ꢅ.
Interestingly, the Co-O-Co angles in this [(CoO) ] ring span
5
ꢀ
of the bpp phenolate rings with respect to the idealized
a range 8.58 wide (104.29–112.768), whereas the O-Co-O
angles are much more regular (176.31–178.588). On the
other hand, the CoꢀN lengths span from 2.136(9) to
Table 1. Crystallographic data of complex [Co
1·DMF).
5
A
H
U
G
R
N
N
(CO
3
)
2
A
T
N
T
E
N
N
(bpp)
5
]ClO
4
·DMF
(
2
.214(6) ꢅ. Within the Co pentagon, the distances between
formula
M
C
107
H
85Co
5
N
10
O
11·ClO
4
·C
3
H
7
NO
adjacent metal centers are 3.148(2) to 3.310(2) ꢅ, and the
second neighbor separations are in the range 5.165(2) to
2154.05
P2 /n
space group
crystal system
a [ꢅ]
1
monoclinic
16.5226(4)
32.6191(9)
19.3264(7)
111.386(1)
9698.8(5)
4
5
.252(2) ꢅ. The crystal packing in complex 1 does not fea-
ture any obvious hydrogen bonding or p–p stacking interac-
tion network (see Figure S3 in the Supporting Information).
Regular pentagonal arrangements of 3d metals as seen in
complex 1 are virtually nonexistent, at least, with the [-(M-
b [ꢅ]
c [ꢅ]
b [8]
V [ꢅ]
Z
O) -] sequence. By contrast, the related hexagonal version is
5
T [K]
120
1.475
2ꢀ
ꢀ
3
more common, and several examples involving CO3 en-
capsulation have been reported, especially with Ni .
1
calcd [gcm
]
II [26–28]
On
F
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(000)
4436
ꢀ1
m
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(MoKa) [cm
]
9.40
the other hand, other molecular pentagonal arrangements
involving more extended bridges between metals have been
measured reflections
unique reflections
51015
13050
[3]
reported.
R
int
0.1318
obs. refl.ns [Iꢃ2s(I)]
7097
1.87–23.00
q
min–qmax/8
Magnetic properties: The synthesis of complex 1 represents
an excellent opportunity for studying the nature of the mag-
netic exchange within polynuclear complexes of high-spin
Co ions subject to spin-orbit coupling and for probing new
methodologies that are being developed to face this chal-
lenge.
hkl ranges
ꢀ18,18; ꢀ35,35; ꢀ20,21
2
R(F ) (observed reflections)
0.0874
0.2212
1283
2
wR(F ) (all reflections)
II
no. variables
goodness of fit
D1max; D1min [eꢅ
1.024
1.25; ꢀ0.67
ꢀ
3
]
DFT study of the magnetic exchange: Theoretical methods
based on density functional theory have been extensively
used for some time now to study the spin state of systems
metallic polygon is 37.78. The approximate point symmetry
[29,30]
of this assembly is thus D . This symmetry is broken by the
ranging from simple molecules
to polynuclear metal
5
2ꢀ
[31]
presence of two central CO3 ligands, capping each face of
the pentagon and mutually disposed in an approximately
staggered conformation with a C···C distance of 2.697(2) ꢅ.
clusters. Such methodology also allows one to determine
all the exchange coupling constants present in polynuclear
[32–35]
transition-metal complexes
and it has been employed
Chem. Eur. J. 2010, 16, 13825 – 13833
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13827

G. Aromꢀ, J. Cano, D. Ray et al.
here to obtain an estimate of Table 2. Selected interatomic distances [ꢅ] and angles [8] for complex [Co
5
A
H
U
G
R
N
U
G
3
)
2
A
H
U
G
R
N
N
5
]ClO
4
·DMF (1·DMF).
the magnetic interactions within
Distances
complex 1. The results from
this study have been useful for Co1ꢀN2A
the phenomenological simula-
tion of the experimental mag-
netization data (see below) and
were rationalized in light of the
ꢀ
ꢀ
ꢀ
Co1 O1A
1.967(7)
2.186(8)
1.994(7)
2.214(6)
2.189(7)
2.116(7)
Co2 O1B
Co2ꢀN2B
Co2ꢀO1C
Co2ꢀN1C
Co2ꢀO3
2.002(6)
2.138(9)
1.997(7)
2.150(8)
2.033(10)
2.732(10)
Co3 O1C
Co3ꢀN2C
Co3ꢀO1D
Co3ꢀN1D
1.990(7)
2.150(8)
1.998(7)
2.150(6)
2.454(10)
2.062(10)
Co1ꢀO1B
Co1ꢀN1B
Co1ꢀO1
Co1ꢀO6
Co3···O3
Co3ꢀO5
Co2···O6
structural features of the Co4
ꢀ
O1D
1.986(7)
2.159(6)
1.983(7)
2.157(9)
2.093(6)
2.224(7)
Co5ꢀO1A
Co5ꢀN1A
Co5ꢀO1E
Co5ꢀN1E
2.012(6)
2.154(8)
2.003(6)
2.136(9)
2.420(9)
2.187(7)
C1ꢀO1
C1ꢀO2
C1ꢀO3
C2ꢀO6
1.234(11)
1.245(11)
1.273(12)
1.217(12)
1.251(12)
1.284(12)
Co4ꢀN2D
Co4ꢀO1E
Co4ꢀN1E
Co4 O2
system. Originally, these calcu-
lations were attempted by em-
[36]
ploying the B3LYP functional
ꢀ
ꢀ
Co5···O1
C2 O5
based on all electron Gaussian Co4ꢀO4
Co5ꢀO4
C2ꢀO4
basis sets. This method howev-
Co1···Co2
er, often faces insurmountable
Co1···Co5
3.288(2)
3.214(2)
3.155(2)
3.310(2)
3.148(2)
O1A···O1B
O1B···O1C
O1C···O1D
O1D···O1E
O1E···O1A
3.969(10)
3.997(9)
3.979(10)
3.967(10)
4.014(8)
Co1···O1D
Co2···O1E
Co3···O1A
Co4···O1B
Co5···O1C
6.009(6)
6.211(7)
6.131(6)
6.058(7)
6.211(9)
difficulties when dealing with
Co2···Co3
II
Co ions that exhibit strong Co3···Co4
spin-orbit coupling, and this Co4···Co5
was the case in the current
Angles
system (see Experimental Sec-
tion). Nevertheless, the energies O1A-Co1-1B
of the various possible spin dis- O1A-Co1-1B
tributions in 1 could be deter-
mined by employing the pro-
O1A-Co1-2A
84.9(3)
177.9(3)
95.8(3)
77.1(3)
99.6(3)
97.0(3)
91.9(3)
161.0(3)
94.5(3)
83.5(3)
101.0(3)
81.0(3)
95.9(3)
163.8(3)
82.7(3)
85.1(3)
178.6(3)
96.6(3)
84.8(3)
103.8(3)
94.2(3)
90.4(3)
169.8(3)
98.6(3)
84.6(3)
95.9(3)
75.1(3)
92.2(3)
158.3(4)
82.5(3)
O1B-Co2-N2B
O1B-Co2-O1C
O1B-Co2-N1C
O1B-Co2-O3
O1B-Co2-O6
N2B-Co2-O1C
N2B-Co2-N1C
N2B-Co2-O3
N2B-Co2-O6
O1C-Co2-N1C
O1C-Co2-O3
O1C-Co2-O6
N1C-Co2-O3
N1C-Co2-O6
O3-Co2-O6
O1A-Co5-N1A
O1A-Co5-O1E
O1A-Co4-N2E
O1A-Co5-O1
O1A-Co5-O4
N1A-Co5-O1E
N1A-Co5-N2E
N1A-Co5-O4
N1A-Co5-O1
O1E-Co5-N2E
O1E-Co5-O1
O1E-Co5-O4
N2E-Co5-O4
N2E-Co5-O1
O1-Co5-O4
84.7(3)
176.3(3)
96.2(3)
102.0(3)
66.6(3)
92.9(3)
91.4(3)
167.9(4)
94.3(4)
86.6(4)
80.0(4)
110.9(4)
97.8(4)
161.2(4)
73.4(4)
85.2(3)
176.8(3)
95.7(3)
71.2(3)
107.4(3)
91.6(3)
91.2(3)
162.8(3)
97.1(3)
84.6(3)
108.9(3)
75.6(3)
98.8(3)
163.8(3)
76.8(3)
O1C-Co3-N2C
84.5(3)
177.9(3)
94.4(3)
70.4(3)
88.0(4)
97.4(3)
88.9(3)
154.1(3)
95.6(4)
84.6(3)
107.9(3)
92.8(4)
98.8(3)
175.0(4)
77.9(4)
91.1(3)
107.4(3)
110.7(3)
104.6(3)
112.8(3)
104.3(3)
126.4(7)
127.9(7)
132.2(8)
129.3(8)
124.5(7)
122.7(7)
121.9(7)
O1C-Co3-O1D
O1C-Co3-N1D
O1C-Co3-O3
O1C-Co3-O5
N2C-Co3-O1D
N2C-Co3-N1D
N2C-Co3-O3
N2C-Co3-O5
O1D-Co3-N1D
O1D-Co3-O3
O1D-Co3-O5
N1D-Co3-O3
N1D-Co3-O5
O3-Co3-O5
Co4-O4-Co5
Co1-O1A-Co5
Co1-O1B-Co2
Co2-O1C-Co3
Co3-O1D-Co4
Co4-O1E-Co5
Co1-O1-C1
O1A-Co1-O1
O1A-Co1-O6
N2A-Co1-1B
N2A-Co1-1B
[
37]
gram SIESTA
GGA
through
a
exchange-correlation N2A-Co1-O1
[38]
N2A-Co1-O6
O1B-Co1-1B
O1B-Co1-O1
O1B-Co1-O6
functional, involving only ex-
ternal electrons while substitut-
ing the atom cores by appropri-
ate pseudopotentials (see Ex- N1B-Co1-O1
perimental
Section).
This N1B-Co1-O6
O1-Co1-O6
method does not consider spin-
orbit coupling. Nevertheless,
the information this model
O1D-Co4-2D
O1D-Co4-1E
O1D-Co4-1E
gives on the coupling constants O1D-Co4-O2
O1D-Co4-O4
is very useful and may be used
N2D-Co4-1E
as a good starting point for fit-
N2D-Co4-1E
ting the experimental data (see
N2D-Co4-O2
Co1-O5-C2
Co2-O3-C1
Co3-O5-C2
Co4-O2-C1
Co4-O4-C2
Co5-O4-C2
below). Given the geometry of N2D-Co4-O4
1
, this system requires nine dif- O1E-Co4-1E
O1E-Co4-O2
ferent coupling constants for a
description of its intramolecular
O1E-Co4-O4
N1E-Co4-O2
N1E-Co4-O4
magnetic
exchange
(see
Figure 3) as represented in the O2-Co4-O4
Hamiltonian of Equation (3).
^
^ ^
^ ^
^ ^
^ ^
^ ^
5 1 5
H ¼ ꢀJ S S ꢀ J S S ꢀ J S S ꢀ J S S ꢀ J S S
1
1
3
2
2
4
3
3
5
4
1
4
ð3Þ
^
^
^ ^
^ ^
^ ^
9 2 3
ꢀ 6
J S S ꢀ J S S ꢀ J S S ꢀ J S S
1
2
7
3
4
8
4
5
Therefore, it is necessary to calculate the energy of at least
[33,39]
ten spin configurations,
chosen so that a system of nine
equations and nine unknowns (the J constants) can be pre-
pared and solved. In fact, to verify possible errors or short-
comings of the procedure, we have studied the fifteen elec-
tronic configurations that are allowed by using the broken
symmetry approach. These correspond to the following spin
Figure 3. Representation of the spin-coupling scheme used for the DFT
calculations and for modeling the magnetic data of 1. The metal number-
ing is the same as used for the crystallographic description.
13828
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13825 – 13833

A High-Spin Cobalt Cluster
FULL PAPER
i
Table 3. Magnetic exchange coupling (J), bridging ligands (L), selected structural parameters (in ꢅ and in 8), metal coordination number (CN), and the-
ꢀ
1
oretical and experimental magnetic coupling constants (in cm ) for 1.
[
a]
[a]
[a]
[b]
[c]
J
i
Sites [i,j]
[1, 3]
[2, 4]
[3, 5]
[1, 4]
dCo-Co [ꢅ]
L
d
A
H
U
G
R
N
N
(Co
i
)ꢀO [ꢅ]
d
A
T
N
T
E
N
G
f
)ꢀO [ꢅ]
Co-O-Co [8]
Co-O-C [8]
CN
J
PBE
J
exp
J
J
J
J
1
2
3
4
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
5.195
5.222
5.222
5.165
m-CO
m-CO
m-CO
m-CO
m-CO
3
3
3
3
3
2.116
2.032
2.062
2.116
2.189
2.062
2.093
2.187
2.224
2.093
–
–
–
–
–
127.9(7) 129.3(8)
132.2(8) 124.5(7)
129.3(8) 121.9(7)
126.4(7)
6, 5
5, 6
5, 5
6, 6
ꢀ5.9(5)
ꢀ4.9(5)
ꢀ3.2(5)
ꢀ1.5(5)
ꢀ3.5
ꢀ3.5
ꢀ3.5
ꢀ0.3
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
124.5(7)
1
1
–
27.9(7)
22.7(7)
J
J
J
J
J
5
6
7
8
9
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
[1, 5]
[1, 2]
[3, 4]
[4, 5]
[2, 3]
3.214
3.288
3.310
3.148
3.155
m-CO
3
2.116
1.977
2.189
1.994
2.062
1.988
2.224
1.983
1.997
2.187
2.012
2.032
2.003
2.224
1.986
2.187
2.004
1.991
–
6, 5
6, 5
5, 6
6, 5
5, 5
+0.5(5)
ꢀ1.1(5)
ꢀ6.0(5)
ꢀ0.4(5)
ꢀ0.9(5)
+0.3
ꢀ0.3
ꢀ3.5
ꢀ0.3
ꢀ0.3
m-OPh
m-CO
m-OPh
m-CO
m-OPh
m-OCO
m-OPh
m-OPh
107.3
–
110.7
–
112.8
91.0
104.3
104.6
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
3
–
–
–
–
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
3
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
2
ACHTUNGTRENNUNG
[
[
a] Labels ’i’ and ’f’ are used to specify the initial and final metal atom in the magnetic coupling. [b] Coupling constants obtained from DFT calculations.
c] Coupling constants obtained from fitting the experimental data.
states (only spin down centers indicated, using the notation
because the Co-O-Co angle (104.68) is wider and therefore,
of Figure 3): one S=15/2, five S=9/2 ([1], [2], [3], [4] and
further within the space corresponding to AF coupling.
[
[
5]) and ten S=3/2 ([1, 2], [1, 3], [1, 4], [1, 5], [2, 3], [2, 4],
2, 5], [3, 4], [3, 5] and [4, 5]). The values of the nine J con-
Bulk magnetization measurements: Variable-temperature
magnetization measurements were performed on a pow-
dered polycrystalline sample of 1 under a constant magnetic
field of 0.5 T, in the 2–300 K range. The results are repre-
sented in Figure 4 as c T versus T plots (c being the molar
i
stants resulting from these calculations are given in Table 3
see Figure 3 for notation). These results allow one to draw
(
the following conclusions: i) Regardless of the coordination
number of the metals, all the pathways involving only one
carbonate ligand (J , J , J ) lead to similar J values of
M
M
1
2
3
strength seemingly correlated to the average Co-O-C
2ꢀ
angles; ii) despite being mediated by two CO3 bridges, the
coupling J is weaker than these above, presumably because
4
it occurs through longer CoꢀO bonds and narrower Co-O-C
angles (on average); iii) the pathways leading to J , J , and
5
6
2
ꢀ
J7 are one syn,syn-CO3 and one phenoxo-like bridge. This
combination is expected to lead to a ligand counter-comple-
[40,41]
mentarity effect on the coupling strength,
by which, the
interaction of each type of ligand with the metallic magnetic
orbitals causes a decrease of the energy gaps between
SOMOs, thereby diminishing the strength of the antiferro-
[
40,41]
magnetic (AF) coupling. As observed before,
this also
Figure 4. Plots of c
The solid lines are the fit of the experimental data through a matrix diag-
onalization technique (see text).
M
5 3 2 5 4
T versus T per mole of [Co ACHTUNGTRENNGNU( CO ) AHCTUNGTNNEUGR( bpp) ]ClO (1).
causes a shift towards larger values of the M-O-M angle at
which the interaction switches from ferromagnetic (F) to
AF in phenoxo-bridged systems (qSWITCH). The calculations
on 1 indicate that this switch occurs now towards 1098;
iv) for the case of J , the metals exhibit two monoatomic m-
paramagnetic susceptibility of the compound). At room
8
3
ꢀ1
O bridges, thus precluding the effect of counter-complemen-
tarity. As it turns out, the average Co-O-Co angle featured
in this moiety (97.78) is close to qSWITCH for bis-(m-hydroxo)-
copper(II) complexes. A similar qSWITCH might be expected
here since the magnetic orbital that most efficiently propa-
temperature, c T is 11.16 cm Kmol , significantly larger
M
II
than expected for five magnetically isolated high-spin Co
ions (c T=9.375 cm Kmol if g=2), which indicates that
3
ꢀ1
M
the orbital angular momentum of the metal ions is not
quenched and therefore contributes significantly to cM.
gates the exchange (d
Therefore, a very weak coupling should be observed, in
agreement with the calculated value; v) the constant J is ex-
pected to be AF and weaker than J because there is only
x2ꢀy2
) is the same as with copper.
Upon cooling, c T decreases at an increasingly faster rate,
M
3
ꢀ1
reaching a value of 1.52 cm Kmol at 2 K, where the curve
seems to initiate a tendency to level off. This decline, which
leads to values much lower than could be expected for five
9
8
II
one m-O bridge. It is however slightly more negative than J8
non-interacting high-spin Co ions exhibiting spin-orbit cou-
Chem. Eur. J. 2010, 16, 13825 – 13833
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13829

G. Aromꢀ, J. Cano, D. Ray et al.
pling, clearly indicates the presence of AF coupling between
the metal centers, consistent with the results obtained from
DFT calculations (see above). The strong orbital contribu-
tion to the magnetic moment precludes the possibility of
using the spin-only formalism as a model to simulate the ex-
perimental data, thus rendering this problem extremely
complex. Some of us have developed procedures to inter-
and the Zeeman interaction with an external magnetic field
B.
h
i
^
^^
^
2
z
H ¼ ꢀAklLS þ D L ꢀ 1=3LðL þ 1Þ
h
i
ð5Þ
^
^
þmB ꢀAkL þ g S B
e
II
pret the magnetic behavior of polynuclear Co clusters,
[14]
This Hamiltonian does not provide for an analytical expres-
sion of cM as a function of the parameters A, k, l and D,
thus, their values must be determined through a numerical
matrix diagonalization method. This may be done by
making the analogy of this Hamiltonian with that for a hy-
pothetical system composed of two local exchange-coupled
spin moments (L=1 and S=3/2) and very different g values
taking into account the effects of spin-orbit coupling.
Complex 1 constitutes a good object of study to verify the
validity of these methods (see below).
Simulation of the magnetic data: A matrix diagonalization
method including spin-orbit coupling effects, based on a per-
turbational approach, has been recently proposed by some
of us as a tool to describe the magnetic properties of six-co-
of Ak and g , respectively. The spin triplet (represented by
e
II
[14]
L) would be seen as experiencing axial zero field splitting
ordinate high-spin Co clusters. This method is used here
to describe de magnetic behavior of 1. Since the details are
published elsewhere, only a summary is provided below.
It is known that in a purely octahedral symmetry, the
(
ZFS) quantified by D with the coupling between both cen-
ters being gauged by Akl. Under this analogy, represented
in Figure 5A, the first term of Equation (5) is the spin–spin
coupling, the second term is the energy of the ZFS, and the
third represents the Zeeman splitting of both local spins.
This procedure would allow one to obtain the temperature
dependence of cM for a mononuclear cobalt(II) complex.
Nevertheless, it can be expanded to include more metal ions
4
II
ground term F of a high-spin Co ion is split into three
4
4
4
states, T , T , and A , of which the former is the ground
1g
2g
2g
ꢀ1
state and is well separated (more than 8000 cm ) from the
other two. In addition, first-order spin-orbit coupling splits
the T1g ground state into a sextet, a quadruplet, and a
4
[42,43]
and the coupling between them.
Complex 1 exhibits five
Kramerꢆs doublet, the eigenfuctions of which can be ob-
tained by diagonalization of the matrix obtained from the
Hamiltonian in Equation (4).
II
nonequivalent Co ions, which could, for simplicity, be treat-
ed using average values of the spin-orbit coupling parame-
ters. However, the system contains metals with coordination
numbers five and six. The former centers must exhibit very
large values of D and may thus be treated as quartets (S=3/
^
^^
ð4Þ
H ¼ ꢀAklLS
2
) subject to ZFS (thus, as in the spin-only formalism). In
In this Hamiltonian, k and A are orbital reduction factors
fact, attempts to treat the problem as five equivalent metal
centers only resulted in unreasonable average values for the
various parameters. By contrast, considering three aniso-
associated, respectively, with the covalent character of the
4
metal–ligand bonds and with the interaction of the T state
1
g
4
4
4
(
T [F]) with an excited T state from a P term ( T [P]).
1g 1g 1g
II
tropic S=3/2 centers together with two Co ions with strong
Typical values of k lie within the range of 0.75 to 1, whereas
A, which provides a measure of the ligand field strength,
takes values between 3/2 (weak field limit) and 1 (strong
field limit). The parameter l is the spin-orbit constant that,
because of the covalence of the metal–ligand bond, has a
lower value than expected for
spin-orbit coupling, thus involving L and S, (Figure 5B) re-
sulted in the successful simulation of the experimental data.
Under this approach, the lowest possible value expected for
c T (8.9 cm Kmol at 0 K, for three S=3/2 spin-only cen-
ters with g=2 and two octahedral Co ions with only the
3
ꢀ1
M
II
the free ion. Most real systems
II
of Co , however, display distor-
tions from the ideal octahedral
geometry, which lead (for the
case of an axial distortion) to a
splitting of the triplet orbital
4
state T into two other states,
1g
4
4
A2g and E , separated by an
g
energy gap, D. Considering the
consequences of this distortion
and using the T and P term iso-
morphism, the Hamiltonian
given in Equation (5) can be
used to describe the system,
now involving the spin-orbit
II
Figure 5. A) Representation of the model employed for a high-spin six-coordinate Co system with spin-orbit
coupling. It is modeled as two local spins, S=1 (subject to a ZFS, D) and S=3/2 with g values of g=ꢀAk and
e
g , respectively, and coupled (with a coupling constant Akl). B) Application of the model to a cluster incorpo-
II
II
rating two six-coordinate Co ions with spin-orbit coupling and three five-coordinate Co metals (darker balls,
coupling, the axial distortion treated as spin-only centers with S=3/2).
13830
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13825 – 13833

A High-Spin Cobalt Cluster
FULL PAPER
lowest Kramer doublet populated, all uncoupled) is much
higher than observed experimentally for a wide range of
temperatures (see Figure 4), which clearly indicates the
presence of AF interactions within the complex. As men-
Conclusions
II
The particular geometric features of the [Co ACHTUNGTERNNUN(G bpp)] moiety
2
facilitates the capture of atmospheric CO (in the form of
2
II
tioned before, the coupling between Co centers may be
carbonate ligands) and the assembly of a peculiar pentago-
nal architecture with the shape of a propeller. This symme-
try forces the CO3 ligands to lie in irregular coordination
modes (unlike what is observed when carbonate templates
the formation of more common cyclic hexagonal arrange-
ments), yielding a combination of five- and six-coordinate
treated with the model proposed here. Assuming that the
exchange coupling only contains isotropic contributions op-
2ꢀ
II
erating on the real spins of the Co ions, the coupling
scheme in Figure 3 may be considered, where J represent
i
constants for the interactions between S=3/2 spins. Even in-
cluding all the approximations described above, the diago-
nalization of the resulting matrix (with the parameters A, k,
II
high-spin Co ions. Modeling the magnetic properties of
such molecules is highly challenging, and the title system
has served as a good test for the validity of a matrix diago-
nalization technique, recently developed to treat polynuclear
molecules of Co . The validity of the method is corroborat-
ed by both the success in reproducing the experimental data
l, D and J) continues to be prohibitive. However, if J is
i
smaller than l (i.e. J/l<0.1), as is usually the case, a pertur-
bative approach can be implemented, thereby simplifying
dramatically the problem. Under these conditions, each of
II
II
the Co ions with spin-orbit coupling may be considered as
and the agreement with the results obtained from DFT cal-
culations on this [Co ] cluster.
II
just an effective local spin moment with S =1/2, exhibiting
eff
5
a g factor that depends, not only on A, k, l, and D, but also
on the temperature (termed G(T)). The coupling constants
between the so defined local spins become now effective
coupling constants (J’) and their relation with the corre-
sponding real constants is J’=25J/9 (coupling between two
S =1/2) or J’=5J/3 (coupling between a S =1/2 and a S=
Experimental Section
Synthesis: The chemicals used were obtained from the following sources:
p-cresol, para-formaldehyde, hexamine, and aniline from SRL Chem
eff
eff
[14]
3/2). The value of G(T) at each temperature may be ob-
(
India) and sodium methoxide from Spectrochem (India). Cobalt per-
chlorate hexahydrate was prepared by treating cobalt(II) carbonate with
HClO (1:1) and crystallized after concentration on a water bath. 2,6-Bi-
s(phenyliminomethyl)-4-methylphenolate (Hbpp) was prepared by a lit-
tained either by using a reported empirical law or by matrix
[14]
diagonalization. The latter method is used here, since it is
more exact. The description of the magnetic exchange in
complex 1 requires nine coupling constants (Figure 3). To
avoid overparameterization, which causes the presence of
several local error minima, the various J parameters have
been grouped into three categories (J [J , J , J , J ]; J [J ,
4
[
15]
erature procedure.
All other chemicals and solvents were reagent
grade materials and were used as received without further purification.
All the reactions were performed in air at room temperature, unless oth-
erwise indicated.
A
H
U
G
R
N
U
G
5 3 2 5 4
ACHTUNGTERNNU(GN CO ) ACHTUNGTNNEUGR( bpp) ]ClO ·DMF (1·DMF): Method 1: To an orange solution
a
1
2
3
7
b
4
J , J , J ]; J [J ]) using the information obtained from DFT
6
8
9
c
5
A
H
N
T
E
N
N
4
)
2
2
calculations. Using this procedure, good simulations of the
experimental data were obtained, by using the program
followed by addition of solid NaOMe (0.22 g, 4 mmol). The stirring was
maintained for 3 h. A mustard-brown precipitate was collected by filtra-
tion, washed with cold methanol and water, and dried under vacuum
[44]
VPMAG,
for several sets of parameters. Some of these
4
over P O10. The yield of 1 was 74%. Single crystals of the complex, suita-
however, lacked any reasonable physical meaning and only
solutions involving moderate to strong metal–ligand cova-
lence, as expected for this system, were considered. The best
fit (see Figure 3) was obtained for the following parameters;
ble for X-ray analysis were obtained from DMF, by dissolving 1 (250 mg)
in this solvent (10 mL) and leaving the resulting solution unperturbed
and open to the atmosphere for five days. Elemental analysis calcd (%)
for C110H N O ClCo (2154.1): C 61.33, H 4.30, N 7.15; found: C 61.25,
92
11 16
5
ꢀ
1
ꢀ1
i) Ak=1.06, l=ꢀ126 cm , D=+1.1 cm
(six-coordinate
H 4.21, N 6.94; selected FTIR bands: (KBr): n˜ = 3447 (br), 1609 (s), 1590
(s), 1552 (vs), 1488 (m), 1430 (s), 1381 (m),1325 (m), 1195 (m), 1072 (s),
56 (vs), 693 (s), 534 cm (m); molar conductance, L
0 ohm cm mol
II
ꢀ1
Co ions), ii) g=2.049 and D=+0.7 cm (five-coordinate,
ꢀ1
7
7
M
: (DMF solution):
II
ꢀ1
ꢀ1
S=3/2 Co ions), and iii) J =ꢀ3.5 cm , J =ꢀ0.3 cm and
ꢀ1
2
ꢀ1
a
b
; UV/Vis spectra [lmax (e,)]: (DMF solution): 400
ꢀ
1
J =+0.3 cm . The small value of D resulting from this fit is
ꢀ1
ꢀ1
c
(4665), 276 nm (12495 Lmol cm ).
II
consistent with the high symmetry of the six-coordinate Co
Method 2: K CO (0.55 g, 4 mmol) was dissolved in degassed methanol
2
3
centers present in 1. In addition, the g and D parameters as-
(40 mL) by stirring for 3 h. The ligand Hbpp (0.31 g, 1 mmol) was dis-
solved in degassed methanol (20 mL) and the solution was added drop-
wise with stirring to the previous one. The resulting orange solution was
sociated with the five-coordinate ions fall perfectly within
[45]
the expected ranges. Finally, the good agreement between
the experimental data and the curve arising from this fit
must be highlighted, as well as the consistency between the
coupling constants obtained here and the values arising
from DFT calculations. These factors serve to validate the
approach and approximations used during this study, which
demonstrates that the method employed serves to obtain
physically meaningful values for various important parame-
stirred for about 10 min, and a solution of Co
4 2 2
ACHTUNGTREUNNNG( ClO ) ·6H O (0.73 g,
2
mmol) in degassed methanol (15 mL) was added to it dropwise. The re-
action mixture was stirred further for 2 h. A mustard-brown precipitate
of 1 was collected by filtration, washed with cold methanol followed by
water, and dried under vacuum over P O10. The yield was 72%.
4
Caution! Although no problems were encountered in this study, transition
metal perchlorates are potentially explosive and should be handled with
care.
Single-crystal X-ray crystallography: The crystal data of compound
II
ters governing the magnetic behavior of high-spin Co clus-
1
·DMF were collected at 120 K using a Nonius Kappa CCD diffractome-
ters.
ter with graphite-monochromated MoKa radiation. The data sets were in-
Chem. Eur. J. 2010, 16, 13825 – 13833
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13831

G. Aromꢀ, J. Cano, D. Ray et al.
[
46]
tegrated with the Denzo-SMN package and corrected for Lorentz, po-
Red EspaÇola de Supercomputaciꢂn (RES), by the Barcelona Supercom-
puting Center (BCS), and by the Supercomputing Center of the Universi-
tat de Valꢃncia (Tirant).
[
47]
larization and absorption effects (SORTAV). The structure was solved
[
48]
by direct methods (SIR97)
and refined by using full-matrix least-
squares with all non-hydrogen atoms anisotropically and hydrogens in-
cluded at calculated positions, riding on their carrier atoms. The asym-
5
metric unit is built up by a pentanuclear cation complex [(Co bpp) -
+
ꢀ
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(CO
3
)
2
] , a ClO
4
ion, and a DMF solvent molecule. All calculations
[
49]
[50]
[1] D. Mandal, V. Bertolasi, G. Aromꢀ, D. Ray, Dalton Trans. 2007,
were performed by using SHELXL-97 and PARST implemented in
[
51]
1989–1992.
the WINGX system of programs. The crystal data are given in Table 1.
Selected bond lengths and angles are given in Table 2. CCDC-736764
contains the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystallograph-
ic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[
[
2] C. L. Chen, J. Y. Zhang, C. Y. Su, Eur. J. Inorg. Chem. 2007, 2997–
010.
3
3] C. S. Campos-Fernꢇndez, B. L. Schottel, H. T. Chifotides, J. K. Bera,
J. Bacsa, J. M. Koomen, D. H. Russell, K. R. Dunbar, J. Am. Chem.
Soc. 2005, 127, 12909–12923.
Physical measurements: Microanalyses were performed by using
a
[
4] R. Sessoli, D. Gatteschi, A. Caneschi, M. A. Novak, Nature 1993,
Perkin–Elmer model 2400 microanalyzer. FT-IR spectra were recorded
on a Perkin–Elmer RX1 spectrometer. The solution electrical conductivi-
ty and electronic spectra were obtained by using a Unitech type U131C
3
65, 141–143.
[
[
5] G. Aromꢀ, E. K. Brechin, Struct. Bonding 2006, 122, 1–67.
6] A. L. Barra, L. C. Brunel, D. Gatteschi, L. Pardi, R. Sessoli, Acc.
Chem. Res. 1998, 31, 460–466.
7] L. Lecren, W. Wernsdorfer, Y. G. Li, O. Roubeau, H. Miyasaka, R.
Clerac, J. Am. Chem. Soc. 2005, 127, 11311–11317.
ꢀ
3
digital conductivity meter with a solute concentration of about 10 m and
1
a Shimadzu 1601 UV-Vis-NIR spectrophotometer, respectively. H and
[
1
3
3
C NMR spectra were recorded in CDCl with a Bruker AC 200 NMR
Spectrometer using TMS as the internal standard. Variable-temperature
magnetic susceptibility data were obtained with a Quantum Design
MPMS5 SQUID magnetometer. The field applied for temperature-de-
pendent measurements was in the range of linear dependence of M
versus H. The measured values were corrected for the experimentally de-
termined contribution of the sample holder, while Pascalꢆs constants
were utilized to estimate diamagnetic corrections to the molar paramag-
netic susceptibility.
[
[
8] M. Murrie, Chem. Soc. Rev. 2010, 39, 1986–1995.
9] Y. Z. Zheng, M. L. Tong, W. X. Zhang, X. M. Chen, Angew. Chem.
2
006, 118, 6458–6462; Angew. Chem. Int. Ed. 2006, 45, 6310–6314.
[
[
10] M. E. Lines, J. Chem. Phys. 1971, 55, 2977–2984.
11] A. V. Palii, B. S. Tsukerblat, E. Coronado, J. M. Clemente-Juan, J. J.
Borras-Almenar, Inorg. Chem. 2003, 42, 2455–2458.
12] H. Sakiyama, R. Ito, H. Kumagai, K. Inoue, M. Sakamoto, Y. Nishi-
da, M. Yamasaki, Eur. J. Inorg. Chem. 2001, 2027–2032.
[
Computational details: Calculations were performed through the Gaussi-
[13] S. M. Ostrovsky, K. Falk, J. Pelikan, D. A. Brown, Z. Tomkowicz, W.
Haase, Inorg. Chem. 2006, 45, 688–694.
[14] F. Lloret, M. Julve, J. Cano, R. Ruiz-Garcia, E. Pardo, Inorg. Chim.
Acta 2008, 361, 3432–3445.
[15] J. J. Grzybowski, F. L. Urbach, Inorg. Chem. 1980, 19, 2604–2608.
[16] W. J. Gu, Y. Z. Li, Y. Pan, S. H. Liu, Acta Crystallogr. Sect. E. 2005,
61, m614–m615.
[17] L. Q. Han, W. H. Sun, L. Y. Wang, H. G. Wang, Y. Cui, J. Chem.
Crystallogr. 2003, 33, 159–163.
[18] D. Luneau, J. M. Savariault, J. P. Tuchagues, Inorg. Chem. 1988, 27,
3912–3918.
[19] G. J. T. Cooper, G. N. Newton, D. L. Long, P. Kogerler, M. H.
Rosnes, M. Keller, L. Cronin, Inorg. Chem. 2009, 48, 1097–1104.
[20] R. P. Doyle, P. E. Kruger, B. Moubaraki, K. S. Murray, M. Nieuwen-
huyzen, Dalton Trans. 2003, 4230–4237.
[21] M. Fondo, A. M. Garcꢀa-Deibe, M. Corbella, E. Ruiz, J. Tercero, J.
Sanmartꢀn, M. R. Bermejo, Inorg. Chem. 2005, 44, 5011–5020.
[22] E. Garcꢀa-EspaÇa, P. Gavina, J. Latorre, C. Soriano, B. A. Verdejo,
J. Am. Chem. Soc. 2004, 126, 5082–5083.
[23] P. Mukherjee, M. G. B. Drew, M. Estrader, A. Ghosh, Inorg. Chem.
2008, 47, 7784–7791.
[
36]
an09 package using the B3LYP functional, the quadratic convergence
[
52]
approach and a guess function generated with the Jaguar 6.5 code. The
usual procedure involving triple-z and double-z all-electron Gaussian
basis sets for the metal and the rest of atoms, respectively, gave no satis-
[
53,54]
factory results.
Calculations engaging 4 Itanium2 1.3 GHz processors
either did not meet the convergence criterion after eight weeks, as a
result of fluctuations in the process, or led to an extremely slow conver-
gence when using a quadratic convergence algorithm. This is common
with systems containing cobalt(II) ions with strong spin-orbit coupling.
Spin distributions have been calculated by employing the program
SIESTA (Spanish Initiative for Electronic Simulations with Thousands of
[
30]
Atoms). As done previously, the GGA exchange-correlation functional
[
38]
proposed by Perdew, Burke, and Erzernhof (PBE) was used here.
Only external electrons are included in the calculations, the cores being
replaced by norm-conserving scalar relativistic pseudopotentials factor-
[
55]
ized in the Kleinman–Bylander form. These pseudopotentials are gen-
[
56]
erated following the approach proposed by Trouiller and Martins from
the ground state atomic configurations for all the atoms except for Co,
3
2
6
0
7
for which, the configuration [Ne] s 3p 4s 3d was employed. The core
radii for the s, p, d, and f components of the Co atoms are 2.38, 2.44,
[
57]
2
.26, and 2.26, respectively. The cutoff radii were 1.14 for oxygen, ni-
[24] C. Pariya, V. G. Puranik, N. R. Chaudhuri, Chem. Commun. 1997,
1307–1308.
[25] R. A. Coxall, S. G. Harris, D. K. Henderson, S. Parsons, P. A. Tasker,
R. E. P. Winpenny, J. Chem. Soc. Dalton Trans. 2000, 2349.
[26] M. L. Tong, M. Monfort, J. M. C. Juan, X. M. Chen, X. H. Bu, M.
Ohba, S. Kitagawa, Chem. Commun. 2005, 233–235.
trogen, and hydrogen atoms and 1.25 for carbon atoms, respectively. Sim-
ilar problems observed with Gaussian packages were also found here. A
search for the optimum parameters for the energy shift and for the mesh
cutoff allowed us, after considerable efforts, to obtain good converged
electronic configurations.
[
27] A. Graham, S. Meier, S. Parsons, R. E. P. Winpenny, Chem.
Commun. 2000, 811–812.
[
28] A. N. Georgopoulou, C. P. Raptopoulou, V. Psycharis, R. Balles-
teros, B. Abarca, A. K. Boudalis, Inorg. Chem. 2009, 48, 3167–3176.
Acknowledgements
[29] A. Goursot, J. P. Malrieu, D. R. Salahub, Theor. Chim. Acta 1995,
1, 225–236.
9
M.S. acknowledges the CSIR, New Delhi, India for her doctoral fellow-
ship. V.B. acknowledges Italian Ministry of University and Scientific Re-
search (MIUR, Rome). G.A. and J.C. acknowledge the Spanish Ministry
of Science and Education (Projects CTQ2007-61690, CTQ2009-06959 and
CSD2007-00010). G.A. thanks the Generalitat de Catalunya for the prize
ICREA Academia 2008. J.C. thanks the Generalitat Valenciana (Pro-
gram PROMETEO/2009/108). The authors thankfully acknowledge the
computer resources, technical expertise, and assistance provided by the
[30] J. M. Soler, E. Artacho, J. D. Gale, A. Garcꢀa, J. Junquera, P. Orde-
jon, D. Sꢇnchez-Portal, J. Phys. Condens. Matter 2002, 14, 2745–
2779.
[31] M. Castro, D. R. Salahub, Phys. Rev. B 1993, 47, 10955–10958.
[32] D. Mandal, V. Bertolasi, J. Ribas-AriÇo, G. Aromꢀ, D. Ray, Inorg.
Chem. 2008, 47, 3465–3467.
[33] E. Ruiz, P. Alemany, S. Alvarez, J. Cano, J. Am. Chem. Soc. 1997,
119, 1297–1303.
13832
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13825 – 13833

A High-Spin Cobalt Cluster
FULL PAPER
[
[
34] E. Ruiz, J. Cano, S. Alvarez, A. Caneschi, D. Gatteschi, J. Am.
Chem. Soc. 2003, 125, 6791–6794.
35] E. C. SaÇudo, T. Cauchy, E. Ruiz, R. H. Laye, O. Roubeau, S. J.
Teat, G. Aromꢀ, Inorg. Chem. 2007, 46, 9045–9047.
[46] Z. Otwinowski, Z. Minor, in Methods in Enzymology, Vol. 276, Part
A (Eds.: R. M. Sweet, C. W. Carter, J. N. Abelson, M. I. Simon),
Academic Press, London, 1977, p. 307.
[47] R. H. Blessing, Acta Crystallogr. Sect. A 1995, 51, 33.
[48] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Crystallogr. 1999, 32, 115–119.
[49] G. M. Sheldrick, SHELXL97, Program for Crystal Structure Refine-
ment, University of Gçttingen, Gçttingen, Germany, 1997.
[50] M. Nardelli, J. Appl. Crystallogr. 1995, 28, 659.
[51] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837.
[52] Jaguar 6.0, Schrçdinger, Inc., Portland, 2005.
[53] A. Schꢈfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571–
2577.
[54] A. Schꢈfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829–
5835.
[55] L. Kleinman, D. M. Bylander, Phys. Rev. Lett. 1982, 48, 1425–1428.
[56] N. Troullier, J. L. Martins, Phys. Rev. B 1991, 43, 1993–2006.
[57] S. G. Louie, S. Froyen, M. L. Cohen, Phys. Rev. B 1982, 26, 1738–
1742.
[
[
36] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
37] E. Artacho, J. D. Gale, A. Garcꢀa, J. Junquera, R. M. Martin, P. Or-
dejꢂn, D. Sꢇnchez-Portal, J. M. Soler, SIESTA 2001, 1.3.
38] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
[
[
[
[
[
[
3
865–3868.
39] E. Ruiz, J. Cano, S. Alvarez, P. Alemany, J. Am. Chem. Soc. 1998,
20, 11122–11129.
1
40] L. Gutierrez, G. Alzuet, J. A. Real, J. Cano, J. Borras, A. Castineiras,
Eur. J. Inorg. Chem. 2002, 2094–2102.
41] V. McKee, M. Zvagulis, J. V. Dagdigian, M. G. Patch, C. A. Reed, J.
Am. Chem. Soc. 1984, 106, 4765–4772.
42] G. Aromꢀ, H. Stoeckli-Evans, S. J. Teat, J. Cano, J. Ribas, J. Mater.
Chem. 2006, 16, 2635–2644.
43] M. Fondo, N. Ocampo, A. M. Garcia-Deibe, M. Corbella, M. S. El
Fallah, J. Cano, J. Sanmartin, M. R. Bermejo, Dalton Trans. 2006,
4
905–4913.
[
[
44] J. Cano, VPMAG package University of Valencia, Valencia, 2003.
45] R. Bo cˇ a, Struct. Bonding 2006, 117, 1–264.
Received: May 21, 2010
Published online: October 13, 2010
Chem. Eur. J. 2010, 16, 13825 – 13833
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13833