Structural, spectral and magnetic properties of carboxylato cobalt(II) complexes with heterocyclic N-donor ligands: Reconstruction of magnetic parameters from electronic spectra

Article abstract of DOI:10.1016/j.ica.2012.03.036

Heteroleptic cobalt(II) complexes with general formula of [Co(N-base) ₂(car)₂(H₂O)₂], have been synthesized and structurally characterized; the N-base stands for neutral N-donor ligands: iso-quinoline (iqu), [1]

Full text of DOI:10.1016/j.ica.2012.03.036

Inorganica Chimica Acta 388 (2012) 106–113
Contents lists available at SciVerse ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Structural, spectral and magnetic properties of carboxylato cobalt(II)
complexes with heterocyclic N-donor ligands: Reconstruction of magnetic
parameters from electronic spectra
a
c
a
b
d
a,b
J. Titiš a, , J. Hudák , J. Kozíšek , A. Krutošíková , J. Moncol’ , D. Tarabová , R. Boca
⇑
ˇ
ˇ
a Department of Chemistry (FPV), University of Ss. Cyril and Methodius, SK-917 01 Trnava, Slovakia
b Institute of Inorganic Chemistry (FCHPT), Slovak Technical University, SK-812 37 Bratislava, Slovakia
c Institute of Physical Chemistry and Chemical Physics (FCHPT), Slovak Technical University, SK-812 37 Bratislava, Slovakia
d Institute of Organic Chemistry, Catalysis and Petrochemistry (FCHPT), Slovak Technical University, SK-812 37 Bratislava, Slovakia
a r t i c l e i n f o
a b s t r a c t
Article history:
Heteroleptic cobalt(II) complexes with general formula of [Co(N-base)2(car)2(H2O)2], have been synthe-
sized and structurally characterized; the N-base stands for neutral N-donor ligands: iso-quinoline (iqu),
[1]benzofuro[3,2-c]pyridine (bzfupy), 1-(pyridin-3-yl)[1]benzofuro[3,2-c]pyridine (1-py-bzfupy) and
1-phenyl-1H-imidazole (bylim); car = acetato (ac) or benzoato (bz) ligands. The structure of [Co(iqu)2(a-
c)2(H2O)2] 1, [Co(bzfupy)2(ac)2ꢀ(H2O)2] 2, [Co(1-py-bzfupy)2(ac)2(H2O)2]ꢀH2O 3 and [Co(bylim)2(bz)2(H2O)2]
4 complexes is formed of the {CoN2O2O20} chromophore. These complexes were subjected to magneto-
chemical investigation down to 2 K (susceptibility and magnetization measurements). They show mag-
netic behavior typical for the zero-field splitting systems. The axial parameter of the zero-field
splitting adopts large values, D = 80–103 cmꢁ1. We have analyzed electronic spectra of the complexes
in detail with aim to extract the crystal-field parameters. These have been used for reconstruction of
the magnetic parameters. The calculated axial ZFS parameters are in good agreement with the
experimental values.
Received 27 January 2012
Accepted 19 March 2012
Available online 27 March 2012
Keywords:
Cobalt(II) complexes
N-heterocyclic ligands
Crystal structure
Electronic spectra
Magnetic parameters
Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction
possess a compressed tetragonal geometry and thus the D-param-
eter clearly adopts negative values.
Quinoline isosters, in which the benzene ring is replaced by fur-
an or thiophene ring (furopyridines, thienopyridines), are inten-
sively synthesized and studied [1–3]. A few from these fused
heterocycles were used as ligands in the preparation of coordina-
tion compounds with transition metals, such as Cu(II) and Ni(II).
The spectral, magnetic, thermal properties, coordination chemistry
and X-ray analysis of these compounds have been reported [4–6].
Some studies on structure–magnetism relationship have also
been reported for cobalt(II) ion [8,9]. Cobalt(II) complexes are
very attractive since its ZFS can be very large. These systems dif-
fer from the Ni(II) cases in the feature that the ground electron
term in the octahedral geometry is orbitally degenerate (Fig. 1).
4
On tetragonal compression the T1g(Oh) term splits into the terms
4A2g(D4h, ground) and 4Eg(D4h, excited); E(4Eg) ꢁ E(4A2g) = Dax. The
Recently, we outlined
a
novel type of magnetostructural
ordering of the multiplets is: 4 2g ? {1C6 < 1C7} and 4Eg ? {2C6
A <
correlations. Here, the zero-field splitting (ZFS) parameter D is cor-
related with the structural tetragonality in the series of axially
distorted Ni(II) complexes [7]. In these studies, the above heterocy-
cles have been used to prepare the thiocyanato and carboxylato
nickel(II) complexes, namely with furo[3,2-c]pyridine, 2-methylf-
uro[3,2-c]pyridine, 2,3-dimethylfuro[3,2-c]pyridine, [1]benzof-
uro[3,2-c]pyridine and iso-quinoline. In the case of the
3C6 < 2C7 < 3C7} owing to which the four members of the former
4A2g term can be considered within the spin-Hamiltonian
formalism.
The corresponding energy gap can be attributed to the
zero-field splitting: 2D = E(1C7) ꢁ E(1C6) > 0 [10]. Within this
approach, the magnetic parameters obey relationships.
ꢀ
ꢁ
gx ¼ ge ꢁ 2kA2j2x
=
Dax A2g ! Eg > ge; gz ¼ ge
ð1Þ
ð2Þ
ð3Þ
4
4
0
thiocyanato complexes, the tetaragonality of the {NiN2N4 }
ꢀ
ꢁ
D ¼ kðgz ꢁ gxÞ=2 ¼ k2A2jx2
=
Dax A2g ! Eg > 0
4
4
chromophore varied along with the sign of the ZFS parameter D.
0
h
i
The carboxylato complexes with {NiN2O2O2 } chromophore always
ꢀ
ꢁ
vTIP ¼ NAl0l2Bð2=3Þ 2A2j2x
=
Dax A2g ! Eg
4
4
⇑
where
j
is the orbital reduction factor, k is the spin–orbit splitting
Corresponding author.
E-mail addresses: jan.titis@ucm.sk, jan.titis@post.sk (J. Titiš).
parameter (for the Co(II) ion k/hc = ꢁ172 cmꢁ1) and A is the
0020-1693/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ica.2012.03.036

J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
107
3Eg(1)
3B2g
3T2g
Γ7 x 2
Δ
ax > 0
4Eg(2)
Γ7 x 2
Γ6 x 2
4T1g
10Dq
Γ
4 x 1
Γ6 x 2
Δ
ax > 0
3A2g
3B1g
Γ7 x 2
Γ6 x 2
δ
54 = −D
4A2g
δ
67 = 2D
Γ
5 x 2
Oh
D4h(compressed)
D'4
Oh
D4h(compressed)
D'4
Fig. 1. Comparison of the low-lying energy levels for high-spin Ni(II), d8 (left) and Co(II), d7 (right) ions – not to scale.
configuration interaction mixing parameter introduced by Figgis.
2. Experimental
These parameters are considered as the molecular constants charac-
teristic for the particular system. They can be retrieved by fitting
the experimental data. However, the magnetic parameters have
their origin in the electronic structure of the complex.
Several theoretical models have been proposed and a number of
reviews were written, summarizing important aspects of the ZFS.
Standardly, the ZFS can be modeled through the second-order per-
turbation theory for the spin Hamiltonian by means of the crystal-
field approach [10]. In this simple model, the ZFS is directly related
to the splitting of the crystal-field states due to the geometry of the
complex. However, further important effects arise from covalent
metal–ligand interaction. Adjustment of this purely electrostatic
model to covalency effects can be carried out via the orbital reduc-
2.1. Ligands
All chemicals were of analytical grade. Two ligands (bzfupy and
1-py-bzfupy) were prepared according to the literature recipes [15].
Detailed characterization (elemental analysis, IR and NMR spec-
tra) of the prepared ligands can be found elsewhere [16]. Here, we
will describe only the synthesis. The starting 1-benzofuran-2-carb-
aldehyde (a) was subjected to the Doebner condensation resulting
in the corresponding acid (b) (Fig. 2). The acylazide was prepared
by treatment of (b) with ethyl chloroformate and sodium azide
in one-pot reaction and then was transformed by thermal cycliza-
tion in Dowtherm with tributylamine to [1]benzofuro[3,2-c]pyri-
din-1(2H)-one (c). Refluxing of (c) with phosphorus oxychloride
gave chloro-derivative (d), which by the reduction yielded [1]ben-
zofuro[3,2-c]pyridine (e).
tion factors
j [11]. In order to forwarding beyond the spin Hamil-
tonian formalism, the orbital functions need be involved into the
basis set and the orbital angular momentum operator considered
explicitly [10]. On higher level of sophistication, it is possible to ap-
ply non-empirical methods of the quantum chemistry (correlated
ab initio methods and/or variants of the density-functional meth-
od) [12,13]. These bring quantitative data about the electronic
structure of the system under study. However, the calculated en-
ergy levels (and subsequently the ZFS) are outputs of the molecular
composition for the given geometry and usually loose transparency
in chemical terms. On the contrary, inorganic chemists are more
familiar with terminology of the Racah parameter (B and C) that,
in fact are parameters of the single-ion electron repulsion. These
can be modified via the nephelaxeutic (cloud-expanding) effect.
Moreover, electron spectra are sources about the ligand-field
strengths (F4) and the spectroscopic series is parametrized through
the metal (kM) and ligand (fL) increments: D0 = kMfL. We must also
emphasize, that using the crystal-field approach large number of
calculations can be performed with good agreement with experi-
mental data by means of minimal computing resources (with
regards to chemical complexity of the most studied systems).
Therefore, it is sometimes more helpful to investigate the zero-
field splitting and other magnetic parameters resulting from
the set of crystal-field parameters like B, C, F4(z), F4(xy), k = ꢁn/2S
The
ligand
1-(pyridin-3-yl)[1]benzofuro[3,2-c]pyridine
(1-py-bzfupy) (f) was prepared in moderate yield from
1-chloro[1]benzofuro[3,2-c]pyridine (d) and pyridine-3-boronic
acid (Py-3-B(OH)2) by Suzuki coupling reaction in the presence of
Pd(PPh3)4 catalyst in dichloromethane (Fig. 2).
Other ligands were purchased from commercial sources and
were used as received.
2.2. Preparation of the complexes
The starting materials (CoCl2ꢀ6H2O, potassium acetate and
potassium benzoate) were purchased from commercial sources.
Cobalt(II) acetate, as well as the cobalt(II) benzoate, was synthe-
sized by adding potassium acetate/benzoate to a cobalt(II) chloride
ethanol solution.
The four complexes [Co(iqu)2(ac)2(H2O)2] 1, [Co(bzfupy)2-
(ac)2(H2O)2] 2, [Co(1-py-bzfupy)2(ac)2(H2O)2]ꢀH2O 3 and [Co(by-
lim)2(bz)2(H2O)2] 4 were prepared from cobalt(II) acetate/benzoate.
Into the ethanol solution of cobalt(II) acetate/benzoate a stoichi-
ometric amount of ligand was added and stirred for half hour at
60 °C. After few days the pink solid complexes were collected.
Single crystals were grown from the ethanol solution. The compo-
sition of the obtained complexes was confirmed by elemental anal-
ysis. X-ray structure analysis was performed for all compounds.
Anal. Calc. for (1) C22H24N2CoO6: C, 56.10; H, 5.13; N, 5.94. Found:
C, 55.80; H, 5.14; N, 5.99%. Anal. Calc. for (2) C26H24N2CoO8: C,
56.63; H, 4.39; N, 5.08. Found: C, 56.62; H, 4.48; N, 5.10%. Anal.
Calc. for (3) C36H32N4CoO9: C, 59.76; H, 4.46; N, 7.74. Found: C,
58.05; H, 4.57; N, 7.57%. Anal. Calc. for (4) C34H34N4CoO6: C,
62.48; H, 5.24; N, 8.57. Found: C, 62.80; H, 5.74; N, 8.11%.
and j.
Carboxylato Co(II) complexes with N-donor ligands, in which
the equatorial positions are completed by water molecules are
widely covered in the structural database (CSD). However,
structures containing furopyridine rings absent. This paper deals
with preparation, structural and magnetic investigation of
carboxylato Co(II) complexes with iso-quinoline (iqu), [1]benzof-
uro[3,2-c]pyridine (bzfupy) and 1-(pyridin-3-yl)[1]benzofuro[3,2-
c]pyridine (1-py-bzfupy). For comparison, also the complex
[Co(bylim)2(bz)2(H2O)2] 4 was prepared and its magnetism studied
(bylim = 1-phenyl-1H-imidazole). Note that the CSD already
contains a record of the trans-bis(acetato-O)-trans-diaqua-trans-
bis(isoquinoline-N)-cobalt(II) (JAWJAT). For this compound, only
structural data were reported [14]. Remaining three prepared com-
plexes (2–4) are new.
2.3. Analytical methods and equipment
Elemental C–H–N analysis was carried out with a commercial
analyzer (Carlo Erba, 1108). Electronic spectra were measured in

108
J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
O
NH
CO2H
CH2(CO2H)2
pyridine,
1. ClCO2Et
2. NaN3
3. Δ
CHO
O
O
piperidine
O
a
b
c
POCl3
Cl
N
N
N
AcOH
3-PyB(OH)2
Pd(PPh3)4
N
O
d
O
e
O
f
Fig. 2. Preparation of the [1]benzofuro[3,2-c]pyridine (e) and the 1-(pyridin-3-yl)[1]benzofuro[3,2-c]pyridine (f).
Nujol mull (Specord 200, Analytical Jena) in the range 9000–
50000 cmꢁ1
levels – the crystal-field multiplets result from the diagonalization
^
^
^
^
.
of the effective Hamiltonian H ¼ Vee þ Vcf þ Hso. At this level of
Single-crystal X-ray diffraction experiments have been per-
formed using Gemini R CCD apparatus (Oxford Diffraction). Data
reduction and analytical absorption [17] corrections were per-
formed by CRYSALISPRO package [18]. The structures were solved by
direct methods using SIR-97 [19] or SHELXS-97 [20] and refined by
the full-matrix least-squares procedure with SHELXL-97.
sophistication the zero-field splitting is read off from the energies
of the lowest multiplets, 2D = C7
labeled according to irreducible representations of the given dou-
ble group.
ꢁ
C6. Energy levels are optionally
Density functional calculations reported in this paper have been
obtained with the ORCA electronic structure program system [24].
Unrestricted Kohn–Sham (UKS) formalism was applied for the
solution of the SCF equations without geometry optimization
(crystallographic coordinates of the complexes were used). Two
different calculations have been performed: (i) determination of
composition of the relevant molecular orbitals (MOs) for all stud-
ied complexes; (ii) calculation of the ZFS for complex 4. In the first
calculation, the local density approximation (LDA) was applied
with the Vosko–Wilk–Nusair (VWN) parameterization [25]. The
all atoms basis set was triple-f quality with one set of polarization
functions (TZVP) [26]. Tight SCF convergence criteria were applied.
The resulting UKS orbitals were transformed to natural orbitals
(UNOs) and subjected to Löwdin population analysis. For calcula-
tion of the ZFS the hybrid B3LYP exchange–correlation functional
[27] was used with combination of the TZVP basis set and the
spin–orbit mean field (SOMF) representation of the spin–orbit cou-
pling operator as described in reference manual [24].
Magnetic susceptibility and magnetization measurements were
done using a SQUID magnetometer (Quantum Design, MPMS-XL7)
between 2 and 300 K at B = 0.1 T. The magnetization data until
B = 7 T were taken at T = 2.0 K. Raw susceptibility data were cor-
rected for underlying diamagnetism using the set of Pascal con-
stants. The effective magnetic moment has been calculated as
usual: leff/l
B = 798(v0T)1/2 when SI units are employed.
3. Computational
Calculations based on the crystal-field theory were done using
the TERMS package [21]. The matrix elements of the relevant opera-
^
^
^
tors (electron repulsion Vee, crystal-field Vcf , spin–orbit Hso) were
evaluated in the basis set of free-atom terms using the full battle
of the irreducible tensor operators approach [22,23]. There is no
need to pass from the free-atom terms to the crystal-field terms
as such a process is the only unitary transformation that leaves
the eigenvalues invariant. The key role in the above calculations
plays the unit tensor operators that are expressed with the help
of the genealogic coefficients (coefficients of the fractional parent-
age) and the 6j-symbols of composition of angular momenta. Hav-
ing these reduced matrix elements of the spherical tensor
operators determined, the calculation proceeds with the help of
the Wigner–Eckart theorem where the reduced matrix element is
multiplied by a proper 3j-symbol (the integral of angular momen-
tum functions).
4. Results and discussion
4.1. Molecular and crystal structure
Crystal data for compounds 1 through 4 are collected in Table 1.
The molecular structure (ORTEP) is viewed in Supplementary
material Figs. S1–S4.
The structure of all complexes consist of [Co(N-base)2(car)2-
(H2O)2] monomeric units in which the central Co(II) atom has a
distorted octahedral configuration (Fig. 3). The coordination sphere
of the Co(II) atom is, therefore, defined by two aqua ligands (w),
two carboxylato ligands (c) and two neutral N-donor base ligands.
Two oxygen atoms from water molecules and two oxygen atoms
from carboxylato groups form the equatorial plane and two nitro-
gen atoms from the heterocycles are situated in the axial sites,
{CoN2O2O02}. The metal–ligand distance intervals are: Co–Oc
2.079–2.140, Co–Ow 2.107–2.121 and Co–N 2.135–2.188 Å (see
Table 2). Complete set of important bond distances and bond an-
gles for 1–4 can be found in Supplementary material.
In our previous works we systematically introduced a parame-
terization of the tetragonal distortion in mononuclear complexes.
The definition is simple and clear, Dstr = Rax ꢁ Req. However, in case
of the heteroleptic complexes we need be careful since we are
dealing with the heterogeneous donor set. The distortion parame-
ter for such complexes is redefined as
Having the set of spectral parameters on input (B, C, n, F4(R) and
eventually jx, jy, jz) the magnetic parameters, preferentially the
axial zero-field splitting, can be calculated in TERMS by two methods.
First, the spin Hamiltonian formalism is taken into account; here,
^
the matrix elements of the orbital angular momentum operator L
in basis of the atomic terms are transformed to the basis set of
crystal-field terms using the matrix that diagonalizes the
^
^
^
H ¼ Vee þ Vcf operator
Lcf ¼ CyLatC
ð4Þ
ð5Þ
Then the Lambda tensor (a,b = x, y, z) is constructed as
P
Kab
¼
½La ꢀ Lbꢂ=Ei ꢁ E0
i
This enters the formulae for the magnetic parameters within
the spin Hamiltonian formalism (Eqs. (1–3)). Second method
works beyond the spin Hamiltonian where the zero-field energy

J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
109
Table 1
Summary of X-ray crystallographic data.
Compound
1
2
3
4
Formula weight (g molꢁ1
Crystal system
Space group
Cell parameters
a (Å)
)
471.367
triclinic
P1
551.409
monoclinic
P121/c1
723.592
monoclinic
C12/c1
653.589
monoclinic
P121/c1
8.3670(17)
13.294(3)
19.513(4)
90
12.277(4)
7.836(6)
12.724(7)
90
27.8882(10)
17.0215(18)
7.5121(3)
90
15.9585(3)
5.5747(1)
17.0424(4)
90
b (Å)
c (Å)
a
(°)
b (°)
90
93.94(2)
90
1221.2(12)
2
95.425(3)
90
3550.0(4)
4
95.868(2)
90
1508.20(6)
4
c
(°)
89.97(3)
2170.4(8)
4
V (Å3)
Z
Density, Dcalc (g cmꢁ3
)
.345
1.500
1.350
1.661
Absorption coefficient (mm)
F(000)
0.814
908
0.757
570
0.542
1492
1.174
772
Reflections collected
Data/restrains/parameters
10355
15800/3/1117
0.0376
0.1195
831600
1985
3307
2595
2486/0/175
0.0314
0.0934
831602
4610/1/239
0.0446
0.1590
831601
3456/0/214
0.0325
0.0850
831599
R1[F2 > 2
r
(F2)]
wR2(F2)
CCDC Deposit No.
suggests a need to identify the Co–Ow bond as axially positioned.
However, in terms of the ligand-field strengths it follows: Co–
N > {Co–Ow, Co–Oc}. Therefore, the Co–N bond as axially positioned
seems to be the better choice and thus we prefer to consider small
axial distortion with large rhombic contribution for these
complexes.
The system of hydrogen bonds for the Co(II) complexes is
shown in Fig. 4. The intermolecular Ow–Hꢀ ꢀ ꢀOc hydrogen bonds
for 1 and 2 [Hꢀ ꢀ ꢀOc = 1.905 and 1.884 Å; Ow–Hꢀ ꢀ ꢀOc = 165° and
162°] are formed through aqua-ligand OH groups to uncoordinated
acetato-ligand O-atoms from neighbor-lying molecule, and they
are situated in the equatorial plane. In the crystal structure of
the complex 2, the furan ring O-atoms are not involved into the
hydrogen bonds; consequently only a two-dimensional network
is formed. Also the intramolecular Ow–Hꢀ ꢀ ꢀOc hydrogen bonds are
present in these structures [Hꢀ ꢀ ꢀOc = 1.876 and 1.842 Å; Ow–
Hꢀ ꢀ ꢀOc = 172° and 167°]. An analogous system of hydrogen bonds
can be found in compound 4. Its geometric characteristics are:
[Hꢀ ꢀ ꢀOc = 2.069 Å and Ow–Hꢀ ꢀ ꢀOc = 168°] for intermolecular con-
tacts, and [Hꢀ ꢀ ꢀOc = 1.824 Å and Ow–Hꢀ ꢀ ꢀOc = 167°] for intramolec-
ular contacts.
LN
R
H
2
CO2
O
iqu
o
C
N
H
2O
R
O2C
O
O
LN
bzfupy
N
=
R
H ,phenyl
3
C
1-py-bzfupy
LN N-donor base
=
N
N
N
bylim
N
Fig. 3. Scheme of the Co(II) monomeric units; iqu = iso-quinoline, bzfupy = [1]ben-
zofuro[3,2-c]pyridine, 1-py-bzfupy = 1-(pyridin-3-yl)[1]benzofuro[3,2-c]pyridine
and bylim = 1-phenyl-1H-imidazole.
Table 2
Selected structural parameters for the Co(II) complexes.
The registered intermolecular contacts for the complex 3 are
Complex
dz(Co–N) (Å)
dxy(Co–O) (Å)
Dstr (pm)
Estr (pm)
Ow⁄–H. . .Oc [H. . .Oc = 2.085 Å and Ow⁄–H. . .Oc = 176°] (w⁄ denotes
1
2
3
4
2.175
2.188
2.171
2.135
2.121w, 2.085c
2.107w, 2.085c
2.109w, 2.079c
2.115w, 2.140c
ꢁ2.80
ꢁ0.80
ꢁ2.30
ꢁ9.25
1.80
1.10
1.50
1.25
the
uncoordinated
water
molecules)
and
Ow–H. . .N
[H. . .N = 2.238 Å and Ow–H. . .N = 177°]. They are formed by OH
groups of the uncoordinated water molecules connected with a
free acetato-ligand O-atoms in the equatorial plane. The N-atoms
of the benzo[4,5]furo[3,2-c]pyridine fragments are connected with
the aqua-ligand OH groups, thus a complex three-dimensional
network is formed. There are also present the intramolecular
Ow–H. . .Oc hydrogen bonds as in the previous cases
[H. . .Oc = 1.879 Å and Ow–H. . .Oc = 170°].
Dstr
¼
Dz ꢁ ðDx
þ
DyÞ=2
ð6Þ
where
i
i
ꢀ
Da ¼ ðda ꢁ d Þ
ð7Þ
where a = x, y, z is a shift relative to the mean distance di for a given
bond (i = N, O). For the N- and O-donor ligands these values have
been taken from complexes containing the [Co(NH3)6]2+ and
ꢀ
4.2. Electronic spectra and crystal-field parameters
[Co(H2O)6]2+ units, respectively, giving rise to dN ¼ 2:185 Å and
The electronic spectra of 1 through 4 are essentially similar
(Fig. 5). In the range of 9000–27000 cmꢁ1 they exhibit a number
d–d transitions which are followed by an intense charge transfer
ꢀ
dO ¼ 2:085 Å [28]. For the rhombic structural distortion the rela-
ꢀ
tionship Estr = (Dx
ꢁ
Dy)/2 is valid. We also assume the validity of
the condition |Estr/Dstr| ꢃ 1/3 which is, however, sometimes difficult
to satisfy for the heteroleptic complexes. Note that for the complex
4 the above inequality is obeyed, |Estr/Dstr| = 0.14; for the remaining
band. The first visible transition is a broad band at ca 9500 cmꢁ1
The second, more intense d–d peak appears at ca 20000 cmꢁ1
.
.
Since the structure of the complexes is derived from an octahe-
dron, the interpretation of the spectra could be done in the Oh
group of symmetry in the first approximation. The estimate for
ones |Estr/Dstr| = 0.64 ꢁ 1.37. For complexes 1–3 the
D(Co–N) and
D
(Co–Oc) shifts are small while the (Co–Ow) are much larger. This
D

110
J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
Fig. 4. Coordination mode of ligands and system of hydrogen bonds in the Co(II) complexes (only asymmetric units are shown).
10000
15000
20000
25000
0.5
0.4
0.3
4
0.2
0.3
3
2
4A2g
0.2
0.4
4T1g(4P)
Δ3
4Eg
CI
0.3
4A2g(4F)
4B1g
Δ2
0.2
0.5
10Dq
4T2g(4F)
4B2g
Δ1
1
0.4
0.3
0.2
0.1
4Eg
4Eg
8Dq
4T1g(4F)
Δax
4A2g
10000
15000
20000
25000
wavenumber/cm−1
Oh
D4h
Fig. 5. Electronic d–d spectra for the studied Co(II) complexes taken in the Nujol mull; the band assignments were made from a Lorentzian analysis in D4h point group (left).
Energy levels for d7 system in near-octahedral environments (right).
the crystal-field strength for octahedral complexes is Dq(Oh) = (1/
6)F4(R), where the typical F4 value for CoII–N/O systems is ca
6500 cmꢁ1. Therefore, the first transition D1 = 8Dq lies in the NIR
region and it has not been detected by the used hardware. For
the second spin-allowed transition the estimated value is
D2 = 18Dq ꢄ 19500 cmꢁ1
.
For a more detailed analysis (in D4h symmetry), a Lorentzian
deconvolution of the spectra has been applied first. The Co(II) ion

J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
111
in tetragonal symmetry is expected to have six spin-allowed tran-
sitions, but the recorded spectra apparently have only two to three
major components. It is evident from the previous discussion that
the lowest transitions, 4A2g ? 4Eg(4T1g) and 4A2g ? 4Eg(4T2g), are lo-
cated in the region of <9000 cmꢁ1, thus we focused on the four
highest one. We fitted the spectra (without the energy level calcu-
lations) roughly with three Lorentzian primitives first, then two
minor components were added to reproduce the more suitable
shapes. The deconvoluted spectra for 1–4 are shown in Supple-
mentary material (Fig. S6). The next step was the band assignment
and determination of the crystal-field parameters. Here, we fitted
the spectra by means of the generalized crystal-field theory using
the obtained Lorentzian function characteristics (position, half-
wide, height) as fixed input parameters. The optimized set of input
parameters was {F4(xy), F4(z), B}, the Racah C parameter was con-
strained as C = 4/3B.
The optimum set of the crystal-field parameters is collected in
Table 3. There are also shown the calculated values of the lowest
transition, 4A2g ? 4Eg(4T1g) = Dax, (axial splitting parameter) which
is a measure of the crystal-field tetragonality. The band assign-
ments (see Fig. 5) have been done according of the calculated en-
ergy levels. In the used software, the levels (crystal-field terms)
are classified under the characters of the irreducible representa-
tions of the respective point group resulting from the matrix
transformation
Magnetic parameters were determined using a fitting procedure
where the energy levels results from a full-matrix diagonalization
of the spin Hamiltonian,
h ꢄ
ꢅ
ꢄ
ꢅi
ꢁ2
H ¼ ꢀh D Sz ꢁ S =3 þ E Sx ꢁ Sy þ hꢀꢁ1lBðB ꢀ g ꢀ SÞ
ð9Þ
S
2
^
2
^
2
^
2
^
^
~
~
which is a good approximation for this case since a sufficiently
strong tetragonality lifts the members of the 4Eg term outside the
thermal accessibility
(Dax > |k|). Otherwise, the Griffith/Figgis
approach should be considered. Furthermore, a molecular-field
correction is essential in reproducing the low-temperature suscep-
tibility data for T < 10 K [8]; thus the calculation has been improved
by the correction
v
vcorr
¼
þ
a
ð10Þ
1 ꢁ ðzj=NAl0l2BÞ ꢀ
v
where the parameter zj includes the isotropic exchange interaction j
with the number of nearest neighbors z; compensates uncertain-
ties in determining the underlying diamagnetism and it accounts to
the temperature-independent paramagnetism; is the net molar
magnetic susceptibility. The susceptibility data-set ( versus T at
a
v
v
B = 0.1 T) as well as the magnetization data-set (M versus B at
T = 2.0 K) have been treated simultaneously. They were used in con-
structing a common functional to be optimized by the same set of
magnetic parameters: gx, gy, D, E, zj and
a. The z-component of
the g-factor (gz) has been fixed to 2.0 (for details see Ref. [10]).
The optimum set of the magnetic parameters is listed in Table 4.
Experimental determination of the ZFS is not a trivial problem.
In any case, theoretical calculations can help in the analysis of
experimental data and verification of results. We have calculated
the magnetic parameters by crystal-field theory methods as de-
scribed in Section 3. For this purpose, a set of spectral parameters
were extracted from electronic spectra. Orbital reduction factors
Rcf ¼ CyRC
ð8Þ
where
R
is matrix representation of the symmetry operator
^
Rð
a; b;
c
Þ in jL; MLi basis and C are eigenvectors of the crystal-field
matrix.
We can conclude: (i) three spin-allowed transitions have been
identified in the measured spectra at ꢄ9500, 19000 and
21000 cmꢁ1; (ii) the transition 4A2g ? 4A2g[4T1g(4P)] is not resolved
since it is only few wavenumbers above the 4A2g ? 4Eg[4T1g(4P)]
band; (iii) at ꢄ15000 and 25000 cmꢁ1 a spin-forbidden transitions
occur, it is expected that the enhanced intensities are consequence
of the ‘‘borrowing’’ mechanism [29].
(
jx
=
jy
=
j
xy and
in real metal complexes cannot be ignored. One can approximately
set = 0.7–0.8 for general studies, however, in this case we need to
jz in D4h) are also in play since covalency effects
j
consider the reduction factors more carefully (magnetic properties
^
of complexes are quite sensitive on it). For Lz no contribution to ZFS
is found, thus only in-plane covalency (
Co(II) ion [10].
jxy) is significant for the
Conventionally
j
= 1 ꢁ (cL)2, where the cL coefficient deter-
mines the amount of ligand character that is mixed into the metal
4.3. Magnetic data and magnetic parameters
d-derived orbitals [30]. To estimate the jxy factors, DFT character-
ization of the relevant
r
d-p and
p
d-p MOs (mainly dx2
and dxy
ꢁy2
Magnetic data of the Co(II) complexes are consistent with the
zero-field splitting (Fig. 6). On cooling the effective magnetic mo-
ment stays almost constant at its room temperature leff ꢅ 5.0 lB
derived) has been performed for all studied complexes. Results
are summarized in Table 5: larger averaged O-character of the
in-plane bonds has been found for 1 and 3 (12% and 13%, respec-
tively); for 2 and 4 the O-contribution is less, 7–10%. We assumed
that these quantities are approximately equal to 100(cL)2; subse-
quently the orbital reduction factors were evaluated. Final set of
the magnetic parameters is shown in Table 6. As can bee seen
in Fig. 7, the CFTM provides somewhat better results for D-param-
eter than the SH formalism (slope of the correlation line is close
to unity for CFTM). In general, however, it can be concluded that
the calculated magnetic parameters are consistent with experi-
mental data.
For comparison, the B3LYP/TZVP coupled-perturbed (CP) meth-
od has been used for the calculation of the ZFS for complex 4. This
complex has been chosen since its negative tetragonality is the
most distinctive within the series (Dstr = ꢁ9.25 pm, |Estr/Dstr| < 1/
3). In this advanced method, the ZFS is defined as [31]
until 150–100 K when it gradually decreases to
a value of
leff ꢅ 3.5–4.0
lB at T = 2 K. The magnetization at T = 2.0 K saturates
to the value of Mmol/NAlB ꢅ 2.0–2.5.
The low-temperature susceptibility data are dominated by the
lower Kramers doublet MS = 1/2 and the limiting value is
leff ðLTÞ ¼ ½g2z =4 þ 2gx2ꢂ1=2lB irrespective of D. With gz = 2.0 and typ-
ical gx = 2.7 one gets
leff(LT) = 3.95 lB. Lower values observed at
the T ꢄ 2 K are attributed to the presence of intermolecular inter-
actions described by the molecular field correction. The high-tem-
ꢂꢀ
ꢁ
ꢃ
perature limit is leff ðHTÞ ¼ g2z þ 2gx2 ð5=4Þ 1=2lB which with the
above typical values yields 4.23 lB again irrespective of D.
Table 3
Crystal-field parameters for the studied complexes.
Complex
Dax (cmꢁ1
)
B (cmꢁ1
)
F4(z) (cmꢁ1
)
F4(xy) (cmꢁ1
)
D ¼ Dss þ DsSo;Sc þ DS;Sþ1 þ DS;Sꢁ1
ð11Þ
soc
soc
1
2
3
4
632
459
593
743
876
845
901
914
6720
6383
6646
6869
5261
5325
5273
5147
The Dss part that accounts for the spin–spin coupling contribu-
tion to the ZFS has been treated with the ‘‘UNO’’ option. This allows
the calculation of the spin–spin coupling term with a restricted

112
J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
6
5
4
3
2
1
0
6
5
4
3
2
1
0
3
2
1
3
2
1
B = 0.1 T
B = 0.1 T
T = 2.0 K
2
1
T = 2.0 K
5.0
4.5
4.0
3.5
3.0
4.4
4.2
4.0
3.8
3.6
μ
μ
μ
μ
N
N
μ
μ
0
10
20
30
40
50
0
10
20
30
40
50
0
3
0
3
0
50 100 150 200 250 300
T/K
0
1 2 3 4 5 6 7
0
50 100 150 200 250 300
T/K
0
1 2 3 4 5 6 7
B/T
B/T
6
5
4
3
2
1
0
6
5
4
3
2
1
0
B = 0.1 T
B = 0.1 T
3
4
T = 2.0 K
T = 2.0 K
2
1
0
2
1
0
4.4
4.0
3.6
4.2
4.0
3.8
3.6
3.4
3.2
μ
μ
μ
μ
N
N
μ
μ
0
10
20
30
40
50
0
10
20
30
40
50
0
50 100 150 200 250 300
T/K
0
1
2
3
4
5
6
7
0
50 100 150 200 250 300
T/K
0
1 2 3 4 5 6 7
B/T
B/T
Fig. 6. Magnetic data for the Co(II) complexes. Temperature dependence of the effective magnetic moment at B = 0.1 T and field dependence of the magnetization at T = 2.0 K.
Empty circles – experimental, lines – fitted.
Table 4
Table 6
Magnetic parameters for the Co(II) complexes extracted from SQUID data.
Reconstructed magnetic parameters (D4h).
Complex
gx
gy
D/hca
E/hca
zj/hca
ab
Complex
gxy
aTIP
DSH
D⁄
84.59
109.7
88.36
78.01
1
2
3
4
2.587
2.613
2.607
2.435
2.689
2.801
2.758
2.520
85.96
103.10
87.59
3.01
6.03
2.84
0.55
ꢁ0.01
ꢁ0.14
ꢁ0.01
ꢁ0.09
2.04
1
2
3
4
2.907
3.443
2.965
2.828
11.51
18.34
12.25
10.50
77.63
123.70
82.64
70.83
0.23
ꢁ6.49
19.43
79.99
a
In cmꢁ1
.
DSH – spin Hamiltonian formalism. D⁄ – crystal-field theory of multiplets (CFTM),
b
In 10ꢁ9 m3 molꢁ1. R-factor (%): (1) 0.49, (2) 0.54, (3) 0.83, (4) 0.68.
D
⁄ = (C7
ꢁ
C6)/2.
Table 5
Population analysis of the Co–O bonds and estimated orbital reduction factors.
140
120
100
80
1
2
3
4
a
91.2/6.6
78.6/16.5
0.88
85.4/8.3
86.1/7.4
0.93
74.8/17.0
82.4/9.2
0.87
83.3/9.1
81.1/11.9
0.90
dx2
dxy
jxy
ꢁy2
a
a
Cobalt/oxo character (%) of the MOs obtained from Löwdin population analysis.
D
SH
2D = Γ7 − Γ6
spin density obtained from the unrestricted natural orbitals. To cal-
culate the spin–orbit coupling contribution to ZFS, four types of
excitations are considered: (i) DOMO ? SOMO (b ? b), (ii)
SOMO ? VMO
DOMO ? VMO (b ?
occupied, singly occupied and virtual MOs.
(
a
?
a
a
), (iii) SOMO ? SOMO
(a ? b) and (iv)
60
60
80
100
120
140
), where the abbreviations stand for doubly
D(exp) /cm−1
The results are summarized in Table 7. As can be seen, the cal-
culated D-parameter is approximately 10-fold smaller than the
Fig. 7. Correlation between theoretical and experimental D-values for complexes
1–4 as predicted from two different crystal-field theory methods.

J. Titiš et al. / Inorganica Chimica Acta 388 (2012) 106–113
113
Table 7
Appendix A. Supplementary material
DFT-CP modeling of the ZFS for complex 4.
ZFS parameters (cmꢁ1
)
CCDC 831600, 831602, 831601 and 831599 contain the supple-
mentary crystallographic data for 1 through 4. These data can be
obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif. Supplementary
data associated with this article can be found, in the online version,
at http://dx.doi.org/10.1016/j.ica.2012.03.036.
Dsoc part
a
?
a
0.21
b ? b
? b
b ?
2.34
3.95
ꢁ0.11
a
a
Dss part
Coulomb
Exchange
D
1.53
0.17
8.07
0.09
References
[1] V. Bobošík, A. Krutošíková, U. Jordis, Monatsh. Chem. 126 (1995) 747.
[2] A. Krutošíková, R. Sleziak, Collect. Czech. Chem. Commun. 61 (1996) 1627.
|E/D|
ˇ
[3] P. Gajdoš, J. Miklovic, A. Krutošíková, Khim. Geterotsikl. Soed. 6 (2006) 825.
[4] A. Krutošíková, B. Mitasová, E. Jóna, M. Bobošíková, Chem. Pap. 55 (2001) 290.
ˇ
[5] J. Miklovic, A. Krutošíková, P. Baran, Acta Crystallogr., Sect. C 60 (2004) m227.
ˇ
ˇ
[6] V. Vrábel, L. Švorc, I. Bradiaková, J. Kozíšek, A. Krutošíková, Acta. Crystallogr.,
Sect. E 63 (2007) o4516.
experimental value. Many works shows that the small D-values of
the manganese and iron complexes (<5 cmꢁ1) are well reproduced
by the DFT-CP approach [32,33]. However, it appears that for sys-
tems with large ZFS this method has limited use.
ˇ
[7] J. Titiš, R. Boca, Inorg. Chem. 49 (2010) 3971.
[8] B. Papánková, R. Boca, L. Dlhán, I. Nemec, J. Titiš, I. Svoboda, H. Fuess, Inorg.
Chim. Acta 363 (2010) 147.
[9] C. Rajnák, J. Titiš, R. Boca, J. Moncol, Z. Padelková, Monatsh. Chem. 142 (2011)
789.
ˇ
ˇ
ˇ
ˇ
ˇ
ˇ
[10] R. Boca, Struct. Bonding 117 (2006) 1.
[11] K.W.H. Stevens, Proc. R. Soc. Lond. A219 (1953) 542.
[12] F. Neese, T. Petrenko, D. Ganyusin, G. Olbrich, Coord. Chem. Rev. 251 (2007)
288.
[13] F. Neese, Coord. Chem. Rev. 253 (2009) 526.
[14] W. Clegg, B.P. Straughan, Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 45
(1989) 1992.
[15] F. Eloy, A. Deryckere, J. Heterocycl. Chem. 8 (1971) 57.
[16] D. Tarabová, J. Titiš, N. Prónayová, A. Gatial, A. Krutošíková, Arkivoc ix (2010)
269.
[17] R.C. Clark, J.S. Reid, Acta Crystallogr., Sect. A 51 (1995) 887.
[18] Oxford DiffractionCRYSALISPRO, Oxford Diffraction Ltd., Abingdon, England, 2009.
[19] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C. Giacovazzo, A.
Guagliardi, A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr. 32
(1999) 115.
5. Conclusions
Four sixcoordinate cobalt(II) complexes of the [Co(LN)2(car)2
(H2O)2] type have been prepared and structurally characterized.
Their chromophores {CoN2O2O2 } exhibit distortions from the
0
octahedral pattern that refer to the D4h symmetry (Dstr) in first
approximation. In complexes 1–3 large rhombic contribution is
present. The analysis of magnetic data shows that the D-values
in the complexes under study adopt values between 80 and
103 cmꢁ1
.
The crystal-field theory analysis of the electronic spectra pro-
vided assignments for characteristic absorption bands in the visi-
ble region. Obtained spectral (crystal-field) parameters coupled
with the estimated orbital reduction factors were used for recon-
struction of the magnetic parameters. The results are in good
agreement with available experimental data. Parallel calculation
of the ZFS was performed with the DFT model for complex 4. Com-
parison of the individual results supports the use of crystal-field
theory methods for modeling the magnetic parameters in this class
of compounds.
[20] G.M. Sheldrick, Acta Crystallogr., Sect. A 64 (2008) 112.
ˇ
[21] R. Boca, R. Herchel, Program TERMS 04, STU, Bratislava, 2004.
ˇ
[22] R. Boca, Theoretical Foundations of Molecular Magnetism, Elsevier,
Amsterdam, 1999.
[23] B.S. Tsukerblat, Group Theory in Chemistry and Spectroscopy, Academic Press,
London, 1994.
[24] F. Neese, ORCA – An Ab initio, Density Functional and Semiempirical Program
Package, Version 2.8-20, Lehrstuhl für Theoretische Chemie, Bonn, Germany,
2010.
[25] S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200.
[26] R. Ahlrichs, K. May, Phys. Chem. Chem. Phys. 2 (2000) 943.
[27] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[28] Inorganic Crystal Structure Database, Fachinformationszentrum Karlsruhe,
Germany. <http://www.fiz-informationsdienste.de/en/DB/icsd/index.html>.
[29] B.N. Figgis, M.A. Hitchmann, Ligand Field Theory and Its Applications, Wiley,
New York, 2000.
Acknowledgments
ˇ
[30] R. Boca, Coord. Chem. Rev. 173 (1998) 167.
[31] F. Neese, E.I. Solomon, in: J.S. Miller, M. Drillon (Eds.), Magnetism: Molecules
to Materials, Wiley-VCH, Weinheim, 2002.
[32] J. Cirera, E. Ruiz, S. Alvarez, F. Neese, J. Kortus, Chem. Eur. J. 15 (2009) 4078.
[33] C. Duboc, M.N. Collomb, F. Neese, Appl. Mag. Res. 37 (2010) 229.
Slovak Grant Agencies (VEGA 1/1005/09, 1/0052/11, 1/0233/12,
APVV-0202-10 and APVV-VVCE-0004-07) are acknowledged for
the financial support.