The impact of anion-modulated structural variations on the magnetic coupling in trinuclear heterometallic CuII-CoII complexes derived from a salen-type Schiff base ligand

Article abstract of DOI:10.1002/ejic.201402151

Three new trinuclear heterometallic [(Cu^IIL)₂Co ^IIX₂] complexes [H₂L = N,N′- bis(salicylidene)-1,3-propanediamine and X = thiocyanate (1), benzoate (2), or azide (3)] have been synthesized by reacting the metalloligand [CuL] with Co(ClO₄)₂·6H₂O and the NH ₄⁺ or Na⁺ salt of the corresponding anion in methanol. Structural characterization reveals that the central Co^II ion is connected to two terminal metalloligands through μ_1,1- diphenoxido bridges in all three complexes. However, two monodentate thiocyanato ions, which are mutually cis coordinated to the Co atom in 1, generate a "bent" structure, whereas the trans-coordinated syn-syn bridging benzoato (1κO:2κO′) and the end-on bridging azido (μ_1,1) coligands in 2 and 3, respectively, produce linear structures. The changes in the number and nature of the bridges with a shortening of the distances between the metal centers leads to a concomitant decrease of the average Cu^II-O-Co^II bridging angle from 99.3(2) to 97.1(4) and 91.5(1)° for 1, 2 and 3, respectively. Variable-temperature magnetic susceptibility measurements show the presence of a dominant antiferromagnetic coupling between the Cu-Co pairs in all three complexes. However, a steady decrease of the magnitude of the exchange coupling constant (J_Cu-Co) is observed from -33.4 (for 1) to -11.4 (for 2) and -2.15 cm^-1 (for 3). This trend suggests that larger Cu-O-Co angles are associated with stronger antiferromagnetic coupling. Anion-mediated geometrical variations generate three heterometallic complexes with distinct trinuclear Cu₂Co cores as a result of different diphenoxido bridging angles between the Cu^II and Co^II ions. Magnetic studies reveal decreased antiferromagnetic exchange interactions associated with a decrease of this bridging angle. Copyright

Full text of DOI:10.1002/ejic.201402151

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
FULL PAPER
DOI:10.1002/ejic.201402151
The Impact of Anion-Modulated Structural Variations on
the Magnetic Coupling in Trinuclear Heterometallic
CuII–CoII Complexes Derived from a Salen-Type Schiff
Base Ligand
Soumavo Ghosh,[a] Guillem Aromí,*[b] Patrick Gamez,[b,c] and
Ashutosh Ghosh*[a]
Keywords: Metalloligands / Copper / Cobalt / Bridging ligands / Schiff bases / Magnetic properties / Magnetostructural
correlations
Three new trinuclear heterometallic [(CuIIL)2CoIIX2] com- tures. The changes in the number and nature of the bridges
plexes [H2L
=
N,NЈ-bis(salicylidene)-1,3-propanediamine
with a shortening of the distances between the metal centers
leads to a concomitant decrease of the average CuII–O–CoII
bridging angle from 99.3(2) to 97.1(4) and 91.5(1)° for 1, 2 and
3, respectively. Variable-temperature magnetic susceptibility
measurements show the presence of a dominant antiferro-
magnetic coupling between the Cu–Co pairs in all three com-
plexes. However, a steady decrease of the magnitude of the
exchange coupling constant (JCu-Co) is observed from –33.4
(for 1) to –11.4 (for 2) and –2.15 cm–1 (for 3). This trend sug-
gests that larger Cu–O–Co angles are associated with
stronger antiferromagnetic coupling.
and X = thiocyanate (1), benzoate (2), or azide (3)] have been
synthesized by reacting the metalloligand [CuL] with Co-
(ClO4)2·6H2O and the NH4 or Na+ salt of the corresponding
+
anion in methanol. Structural characterization reveals that
the central CoII ion is connected to two terminal metallo-
ligands through μ1,1-diphenoxido bridges in all three com-
plexes. However, two monodentate thiocyanato ions, which
are mutually cis coordinated to the Co atom in 1, generate
a “bent” structure, whereas the trans-coordinated syn–syn
bridging benzoato (1κO:2κOЈ) and the end-on bridging azido
(μ1,1) coligands in 2 and 3, respectively, produce linear struc-
sensing, and various spintronic devices.[3] It should be noted
that magnetostructural correlations have been drawn mostly
between isotropic spins, and several groups have success-
fully determined the critical impact of the bridging angle
on the nature and magnitude of exchange interactions.[1,4]
However, such attempts are rather scarce in complexes with
heterospins, although the generation of large ground spin
states with slow magnetic relaxation dynamics in hetero-
metallic transition-metal complexes has become of para-
mount importance in the field of single-molecule magnets
(SMMs).[5]
Introduction
For decades, the combination of coordination chemistry
and molecular magnetism has greatly inspired scientists to
investigate the magnetic properties of isolated molecules, es-
pecially mixed-valence or heterometallic compounds that
show low-temperature ferrimagnetic coupling.[1] One of the
primary goals of these studies is to explore the magne-
tostructural correlations of the exchange interactions be-
tween multiple nonequivalent spin-carrying centers, which
may be ferro- or antiferromagnetically coupled.[2] Such cor-
relations can help to direct future synthetic strategies for
the design of magnetic materials for efficient information
storage, quantum information processing, solvatomagnetic
Recently, we have been investigating trinuclear [(ML)2-
MЈXn] complexes (H2L is a N2O2-donor salen-type Schiff
base ligand, X– is an anionic coligand, and M and MЈ are
two different paramagnetic transition-metal ions).[6] These
complexes showed promising evidence of linear–bent con-
formational flexibility, which led to antiferro- or ferromag-
netic switching of exchange interactions.[7] We further pro-
[a] Department of Chemistry, University College of Science,
University of Calcutta,
92, A. P. C. Road, Kolkata 700009, India
E-mail: ghosh_59@yahoo.com
ceeded to tune the bridging angle in heterometallic CuII
-
2
[b] Departament de Química Inorgànica, Universitat de Barcelona,
Martí i Franquès 1-11, 08028 Barcelona, Spain
E-mail: guillem.aromi@qi.ub.es
MnII complexes by using various coligands (X–) to deter-
mine their influence on the exchange coupling and the pos-
sible ferromagnetic to antiferromagnetic crossover angle.[8]
In the present study, we continue our quest to obtain a cor-
relation between the bridging angle and magnetic coupling
http://www.gmmf-ub.com/
[c] Institució Catalana de Recerca i Estudis Avançats (ICREA),
Passeig Lluís Companys 23, 08010 Barcelona, Spain
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201402151.
Eur. J. Inorg. Chem. 0000, 0–0
1
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
for trinuclear CuII2CoII complexes. Such systems are ex- of the spin-carrying centers and the bridging angle on the
pected to have intricate magnetic properties as a result of a nature of the magnetic coupling in CuII/CoII complexes.
considerable orbital and spin–orbit coupling contribution This anion-dependent variation of the oxido bridging angle
to the effective magnetic moment arising from the intrinsic for any heterometallic spin-coupled systems has rarely been
orbital angular momentum in the octahedral ground state observed.[8]
of high-spin CoII ions [4T1g(F)].[1,5b,9] Moreover, analogous
Cu/Co/MЈЈ (MЈЈ = CdII and PbII) spin-coupled systems
have shown strong geometrical dependence of exchange
Results and Discussion
coupling.[10] The anionic coligands have been chosen in
such a way that a wide variation in the phenoxido bridging
Syntheses and Spectral Characterizations of the Complexes
angle can be achieved by exploiting the differences in their
mode of ligation.[8] A large number of structurally and mag-
Previously, we have synthesized several trinuclear com-
netically characterized heterometallic CuII/CoII molecular plexes with [CuL] as a metalloligand and diamagnetic NaI,
clusters constructed by diphenoxido bridges have been re- PbII, ZnII, CdII, and HgII or paramagnetic NiII, MnII, and
ported; however, they are mostly derived from acyclic or TbIII central metal ions.[6,8a,12] In the present study, we have
cyclic Schiff base ligands.[11] In those complexes, the diphe- prepared three trinuclear complexes by reactions between
noxido-bridged dimetallic CuII/CoII molecular clusters are [CuL], Co(ClO4)2·6H2O, and the anionic coligands thio-
–
rather rigid and, hence, do not allow a significant anion- cyanate (SCN–), benzoate (PhCOO–), and azide (N3 ) in
dependent variation of the CuII–O–CoII angles.
methanol (Scheme 1). As observed previously with related
Herein, we report the synthesis, characterization, crystal compounds, the pseudohalide SCN– leads to an angular tri-
structures, and magnetic properties of three analogous nuclear complex (1), whereas the benzoate and azide ions
[(CuIIL)2CoIIX2] complexes [H2L = N,NЈ-bis(salicylidene)- yield linear trinuclear complexes (2 and 3, respectively).
–
1,3-propanediamine and X = SCN–, OBz–, N3 ]. The con-
In addition to elemental analysis, all complexes have
formational flexibility of the trinuclear [(CuIIL)2CoII]2+ co- been primarily characterized by IR spectroscopy. The Schiff
ordination cluster permits different modes of ligation of the base moiety shows strong and sharp IR absorption bands
anions, which results in significant modulation of the CuII– at 1619, 1624, and 1613 cm–1 for 1, 2 and 3, respectively.
O–CoII angles and the relative positions of the CuII–CoII– The precursor metalloligand [CuL] is neutral and, therefore,
CuII atoms. These subtle structural changes give rise to has no counteranion. The anionic coligands in these three
clear differences in the magnitude of the antiferromagnetic complexes show characteristic IR absorption bands at 2044,
–
exchange interactions in these complexes and, thus, em- 1561, and 2062 cm–1 for SCN–, PhCOO–, and N3 , respec-
phasize the relevance of the distinct geometrical positions tively.[13]
Scheme 1. Formation of the precursor metalloligand and 1, 2, and 3.
Eur. J. Inorg. Chem. 0000, 0–0
2
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
All three complexes and the free metalloligand have been significantly longer than the Co–NNCS bond lengths
characterized by UV/Vis/NIR spectroscopy in methanol as [2.028(3) and 2.068(2) Å], whereas the axial Co–O bond
well as in the solid state. In solution, all compounds display lengths [2.102(2) and 2.085(2) Å] are intermediate; all of
very similar UV/Vis spectra with intense bands at ca. 595 them are typical of high-spin CoII ions.[16] The equatorial
(CuII d–d transition; Figure S1, upper right, Supporting In- N–Co–N bond angle involving the two cis-thiocyanato
formation), 360 (ligand-to-metal charge-transfer transition groups [101.27(9)°] and those involving the oxygen atoms of
of [CuL]), and 274 nm (intraligand π–π* transition of each chelating [CuL] ligand [70.17(7) and 68.90(7)°] deviate
[CuL]), which are similar to those of the free metalloligand significantly from the ideal value (90°). In addition, the ax-
(Figure S1, upper left). The solid-state electronic spectra ial O–Co–O bond angle [163.13(7)°] is also far from that
exhibit the expected bathochromic as well as hypochromic of an ideal octahedron (180°). All these metric parameters
shifts of the d–d transition bands of the metalloligand, as characterize a highly distorted octahedral CoII environ-
a result of its complexation with the CoII center (Figure S1, ment, which is very similar to those found in related struc-
bottom left). Hence, the corresponding band at 595 nm for tures.[7,16]
the free [CuL] moiety is observed at 619 (1), 617 (2), and
Each of the two terminal copper atoms of the trinuclear
609 nm (3). The metalloligand [CuL] has no characteristic (CuIIL)2CoII core is bonded to four donor atoms [O(1),
absorption maxima in the region 800–1400 nm, but the O(2), N(1), N(2) for Cu(1) and O(3), O(4), N(3), N(4) for
octahedral high-spin d7 CoII ions are expected to exhibit Cu(2)] of one tetradentate L2– ligand to generate a square-
absorption bands in the NIR region that correspond to the planar geometry. The Cu–O and Cu–N bond lengths are in
spin-allowed d–d transitions. Accordingly, the bands at the ranges 1.921(2)–1.937(2) and 1.955(2)–1.973(2) Å,
1310, 1290, and 1281 nm are assignable to the characteristic respectively. This geometry is also confirmed by the so-
[4T1g(P)ǟ4T1g(F)] transitions of the octahedral CoII centers called τ4 index.[17] This value measures the distortion be-
in 1, 2, and 3, respectively (Figure S1, bottom right).[14]
tween a perfect tetrahedron (τ4 = 1) and a perfect square
plane (τ4 = 0) and can be calculated with the formula τ4 =
[360°– (α + β)]/141°; α and β (in °) are the two largest angles
around the metal center in the complex. The τ4 values of
0.137 and 0.164 for Cu(1) and Cu(2), respectively, confirm
their slightly distorted square-planar geometries. The root-
mean-square (r.m.s.) deviations from the mean plane pass-
ing through the four atoms coordinated to the copper ions
are 0.027 and 0.107 Å for Cu(1) and Cu(2), respectively,
and the metal atoms are located 0.017(1) [Cu(1)] and
0.013(1) Å [Cu(2)] away from this plane. The metal···metal
separations and the angle between the metal centers in
the angular (CuL)2Co unit are Cu(1)···Co(1) 3.122(1) Å,
Cu(2)···Co(1) 3.091(1) Å, Cu(1)···Cu(2) 4.403(1) Å, and
Cu(1)···Co(1)···Cu(2) 90.24(1)°.
Description of the Structures
The X-ray crystal structure of 1 consists of a “bent”[15]
trinuclear asymmetric unit, viz., [(CuL)2Co(NCS)2] (Fig-
ure 1). Selected coordination bonds and angles are listed in
Table 1.
In contrast to 1, both 2 and 3 consist of centrosymmetric
linear trinuclear units[15] of formulae [(CuL)2Co(O2CPh)2]
and [(CuL)2Co(N3)2], respectively. The corresponding
molecular structures are depicted in Figure 2, and selected
bond lengths and angles are listed in Table 1.
The structures of both complexes consist of a hexaco-
ordinate central cobalt atom in an octahedral geometry to-
gether with two pentacoordinate square-planar copper
atoms. The phenoxido oxygen atoms O(1) and O(2) and
their symmetry related ones (O1a and O2a) from each
[CuL] metalloligand form the equatorial plane of the octa-
hedral coordination sphere of the CoII centers with bond
Figure 1. Representation of the molecular structure of 1 with par-
tial atom-numbering scheme. Atom color scheme: grey C, blue N,
red O, yellow S, green Cu, pink Co. Hydrogen atoms are omitted
for clarity.
The structure contains a hexacoordinate central cobalt lengths of 2.084(6)–2.163(2) Å for the two complexes. The
atom in a CoN2O4 octahedral environment together with axial positions are occupied by oxygen atoms O(3) of the
two tetracoordinate square-planar copper atoms from two syn–syn bridging benzoate ligand (1κO:2κOЈ) and the nitro-
[CuL] metalloligands. The equatorial plane of the octahe- gen atom N(3) of the end-on bridging azide ligand (μ1,1
)
dral coordination sphere comprises two cis-coordinated with bond lengths of 2.065(8) and 2.075(2) Å for 2 and 3,
thiocyanato N atoms [N(5) and N(6)] and two phenoxido respectively. The bond lengths of the ligating atoms to the
O atoms [O(1) and O(4)] from two different [CuL] ligands. CoII center are comparable for the CoO6 and the CoO4N2
The remaining two phenoxido O atoms [O(2) and O(3)] of octahedra in 2 and 3, respectively, and show weak Jahn–
the two [CuL] moieties occupy the axial positions. The Teller distortion, which is a typical feature for high-spin
equatorial Co–O bond lengths [2.164(2) and 2.236(2) Å] are CoII ions. Owing to the presence of centers of inversion in
Eur. J. Inorg. Chem. 0000, 0–0
3
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
Table 1. Selected bond lengths [Å] and angles [°] for 1–3.
1
2
3
Cu(1)–O(1)
Cu(1)–O(2)
Cu(1)–N(1)
Cu(1)–N(2)
Cu(2)–O(3)
Cu(2)–O(4)
Cu(2)–N(3)
Cu(2)–N(4)
Co(1)–O(1)
Co(1)–O(2)
Co(1)–O(3)
Co(1)–O(4)
Co(1)–N(5)
Co(1)–N(6)
Co(1)–Cu(1)
Co(1)–Cu(2)
Cu(1)–Cu(2)
1.926(2)
1.928(2)
1.960(2)
1.970(2)
1.921(2)
1.937(2)
1.973(2)
1.955(2)
2.164(2)
2.102(2)
2.085(2)
2.236(2)
2.028(3)
2.068(2)
3.122(1)
3.091(1)
4.403(1)
Cu(1)–O(1)
Cu(1)–O(2)
Cu(1)–N(1)
1.951(7)
1.969(7)
1.989(9)
1.958(9)
2.218(8)
1.941(2)
1.926(2)
1.937(2)
1.968(2)
2.572(3)
Cu(1)–N(2)
Cu(1)–O(4)*/N(3a)#[a]
Co(1)–O(1)
Co(1)–O(2)
Co(1)–O(3)*/N(3)#
2.084(6)
2.110(6)
2.065(8)
2.163(2)
2.102(2)
2.075(2)
Co(1)–Cu(1)
3.041(1)
6.082(2)
2.916(1)
5.831(2)
Cu(1)–Cu(1a)/Cu(1b)
O(1)–Cu(1)–O(2)
O(1)–Cu(1)–N(1)
O(1)–Cu(1)–N(2)
O(2)–Cu(1)–N(1)
O(2)–Cu(1)–N(2)
N(1)–Cu(1)–N(2)
O(3)–Cu(2)–O(4)
O(3)–Cu(2)–N(3)
O(3)–Cu(2)–N(4)
O(4)–Cu(2)–N(3)
O(4)–Cu(2)–N(4)
N(3)–Cu(2)–N(4)
O(1)–Co(1)–O(2)
O(1)–Co(1)–O(3)
O(1)–Co(1)–O(4)
O(1)–Co(1)–N(5)
O(1)–Co(1)–N(6)
O(2)–Co(1)–O(3)
O(2)–Co(1)–O(4)
O(2)–Co(1)–N(5)
O(2)–Co(1)–N(6)
O(3)–Co(1)–O(4)
O(3)–Co(1)–N(5)
O(3)–Co(1)–N(6)
O(4)–Co(1)–N(5)
O(4)–Co(1)–N(6)
N(5)–Co(1)–N(6)
Cu(1)–Co(1)–Cu(2)
79.05(7)
91.96(8)
170.13(8)
170.61(8)
91.11(8)
97.91(9)
78.76(7)
91.35(8)
168.59(9)
168.30(8)
92.70(8)
97.90(9)
70.17(7)
95.62(6)
76.61(7)
96.04(8)
157.54(8)
163.13(7)
98.14(7)
95.61(8)
93.82(8)
68.90(7)
94.90(8)
97.02(8)
161.06(8)
90.77(8)
101.27(9)
90.24(1)
O(1)–Cu(1)–O(2)
O(1)–Cu(1)–N(1)
O(1)–Cu(1)–N(2)
O(1)–Cu(1)–O(4)*/N(3a)#
O(2)–Cu(1)–N(1)
O(2)–Cu(1)–N(2)
O(2)–Cu(1)–O(4)*/N(3a)#
N(1)–Cu(1)–N(2)
N(1)–Cu(1)–O(4)*/N(3a)#
N(2)–Cu(1)–O(4)*/N(3a)#
81.4(3)
89.8(3)
163.3(3)
98.1(3)
170.5(3)
90.7(3)
94.0(3)
96.9(3)
90.8(3)
97.2(3)
80.89(8)
92.36(8)
160.36(8)
76.65(8)
165.95(8)
92.24(9)
73.45(8)
97.93(9)
93.06(8)
119.18(8)
O(1)–Co(1)–O(2)
O(1)–Co(1)–O(3)*/N(3)#
O(1)–Co(1)–O(1a)
75.1(3)
89.4(3)
180.00
104.9(3)
90.6(3)
90.2(3)
180.00
89.8(3)
180.00
72.03(7)
95.80(9)
180.00
107.97(7)
84.20(9)
98.06(9)
180.00
O(1)–Co(1)–O(2a)
O(1)–Co(1)–O(3b)*/N(3a)#
O(2)–Co(1)–O(3)*/N(3)#
O(2)–Co(1)–O(2a)
O(2)–Co(1)–O(3b)*/N(3a)#
O(3)*/N(3)#–Co(1)–O(3b)*/N(3a)#
81.94(9)
180.00
Cu(1)–Co(1)–Cu(1a)/Cu(1b)
180
180
[a] * and # are for 2 and 3, respectively. Symmetry codes: a = –x, –y, –z and b = –x + 1/2, –y + 1/2, –z.
Figure 2. Representations of the molecular structures of 2 (left) and 3 (right) with partial atom-numbering schemes; a = –x + 1/2, –y +
1/2, –z and –x, –y, –z for 2 and 3, respectively. Atom color scheme: grey C, blue N, red O, green Cu, Pink Co. Hydrogen atoms are
omitted for clarity.
Eur. J. Inorg. Chem. 0000, 0–0
4
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
both 2 and 3, all of the trans angles are ideal (180°), but tion further. Consequently, the phenoxido bridging angle
the cis angles in the equatorial plane [75.1(3) and 104.9(3)° decreases to an average value of 91.5(1)°.
for 2 and 72.03(7) and 107.97(7)° for 3] and the cis angles
with the axially coordinated atoms [89.4(3)–90.6(3)° for 2
Table 2. List of μ1,1 bridging angles for 1–3.
μ1,1 Bridging angle
1
2
3
and 81.94(9)–98.06(9)° for 3] deviate considerably from the
perfect value (90°). The range of values also indicates that
the distortion is higher in 3 than it is in 2. The basal planes
of each of the two terminal copper atoms of the trinuclear
(CuIIL)2CoII cores of 2 and 3 contain four donor atoms of
the Schiff base, namely, two imine N atoms [N(1) and N(2)]
with Cu–N bond lengths in the range 1.937(2)–1.989(9) Å
and two phenoxido O atoms [O(1) and O(2)] with Cu–O
bond lengths in the range 1.926(2)–1.969(7) Å. The bridg-
ing benzoate and the azide ions occupy the axial positions
of the square pyramids with Cu(1)–O(4) and Cu(1)–N(3)
distances of 2.218(8) and 2.572(3) Å, respectively. Thus, the
axial interactions of the bridging coligands with the CuII
centers are much weakened in 3 compared to those in 2.
Moreover, the angles between the axially and equatorially
coordinated atoms vary within a narrow range of 90.8(3)–
98.1(3)° in 2, but those in 3 cover a wide range [73.45(8)–
119.18(8)°]. The so-called Addison parameter (τ5) amounts
to 0.120 and 0.093 for 2 and 3, respectively, and, hence,
confirms the slightly distorted square-pyramidal geometry
for the copper(II) ions (τ5 is 0 for a perfect square pyramid,
Cu(1)–O(1)–Co(1)
Cu(1)–O(2)–Co(1)
Cu(2)–O(3)–Co(1)
Cu(2)–O(4)–Co(1)
Cu(1a)–N(3)–Co(1)
99.37(8)
101.47(8)
100.94(8)
95.34(8)
–
97.8(3)
96.4(3)
–
–
–
90.39(7)
92.63(8)
–
–
76.90(8)
Scheme 2. The average phenoxido bridging angles of the Cu(μ1,1
O)2Co units in 1–3 (from left to right, respectively).
-
Magnetic Properties
The study of the magnetic properties of heterometallic
whereas it is 1 for a trigonal bipyramid).[18] The r.m.s. devia- complexes comprising exchange-coupled copper and high-
tion of the four basally coordinated atoms from the mean spin octahedral cobalt(II) involves difficulties derived from
plane passing through them is much less (0.087 Å) in 2 that the nonquenched orbital angular momentum of the latter
it is in 3 (0.241 Å), and the metal atoms are 0.169(1) and ions. This has been done in a number of different ways. One
0.056(1) Å away from this plane toward the axially coordi- approach, for instance, is to consider only a lower range of
nated oxygen atoms. The Co···Cu distances in 2 and 3 temperatures, at which the lowest J multiplet of the CoII
[3.041(1) and 2.916(1) Å] are slightly shorter than those in 4T1 term is mainly populated (J = 1/2) and to treat it as an
1, but the Cu···Cu separations [6.082(2) and 5.831(2) Å] are effective S = 1/2 and use a spin-only model for the whole
much longer than that of 1 as a consequence of the linear problem.[19] Sometimes, the lowest two J multiplets have
structure with triply bridged metal centers. Another signifi- been included as effective S = 1/2 and S = 3/2 spin moments
cant structural difference arising from the different bridges with the splitting at zero field between both states modeled
in 2 and 3 is the relative position of the terminal CuII ions as the result of magnetoanisotropy.[10,20] Ideally, however,
with respect to the equatorial CoO4 plane. The perpendicu- the model should include the effect of the orbital angular
lar distances of the copper atoms from the said plane are momentum of the CoII ion (L = 1), as it couples with the
ca. 0.743 and 1.102 Å in 2 and 3, respectively. Although the electronic spin (S = 3/2), and also the effect of its interac-
heterometallic coordination clusters of both 2 and 3 com- tion with an external magnetic field. Unfortunately, the
prise one Co octahedron and two Cu square pyramids, they problem does not allow for an analytical solution to the
are joined together through two monoatomic bridges in an wave equation from the resulting Hamiltonian;[21] therefore,
edge-sharing manner in 2, whereas they share common to simulate the magnetic properties in these systems, it is
faces through three monoatomic bridges in 3 (Figure S2).
necessary to employ matrix diagonalization procedures (see
As all three complexes possess a μ1,1-diphenoxido- below).
bridged (CuIIL)2CoII unit, they can be compared through
Molar paramagnetic susceptibility (χM) measurements in
their respective bridging angles (Table 2). For 1, the trinu- the 2–300 K temperature range were performed on pow-
clear unit is formed only by diphenoxido bridges, which dered polycrystalline samples of 1, 2, and 3 under a con-
connect the central CoII ion (bound to two terminal thio- stant magnetic field of 0.5 T. The results are shown in Fig-
cyanato coligands) to two terminal CuII ions, and the bridg- ure 3 as χMT versus T plots. To simulate these data, the
ing angles range from 95.34(8) to 101.47(8)° with an average coupling between the orbital and the spin magnetic mo-
value of 99.3(2)° (Scheme 2). For 2, the trinuclear unit in- ments (L = 1 and S = 3/2) of the octahedral high-spin CoII
cludes an additional μ1,3-carboxylato bridge between the ion was taken into consideration. To this, the exchange be-
terminal CuII centers and the central CoII ion, which brings tween the magnetic moments of the CoII and CuII ions was
them closer, and there is a concomitant decrease of the also added by following the treatment of Lines,[22] which
average phenoxido bridging angles to 97.1(4)°. In 3, the ad- considers only the coupling between true spins (S = 3/2 and
ditional μ1,1-azido bridge shortens the CuII···CoII separa- 1/2 for CoII and CuII ions, respectively). In addition, the
Eur. J. Inorg. Chem. 0000, 0–0
5
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
effect of the magnetic field on the various magnetic mo- dition to the phenoxide bridges, syn–syn carboxylato or
ments resulting from these interactions (the Zeeman effect) end-on azido bridging ligands, respectively, which are
was included. Thus, the Hamiltonian employed is that in known to be active as magnetic couplers.[25] Nevertheless,
Equation (1).
as these ligands are bound to the CuII ions through the long
axial positions, their interaction with the dx2–y2 magnetic or-
bitals of the CuII ions and, thus, their contribution to the
magnetic coupling is expected to be negligible. In the pres-
ent system, a trend seems to emerge that suggests that the
larger Cu–O–Co angles (involving the phenolate bridges)
are conducive to stronger antiferromagnetic couplings (Fig-
ure S3). This observation is reminiscent of the many pre-
viously established correlations in which [Cu2] complexes
with wider Cu–O–Cu angles involving equatorial bonding
positions cause stronger antiferromagnetic interactions.[26]
We have similarly examined a possible relationship between
the Co–O or the Cu–O bond lengths (Figures S4 and S5)
and the extent of the magnetic coupling. The bond lengths
to the Co center seem to be unrelated to the coupling,
whereas longer average Cu–O bond lengths are ac-
companied by an increase of the coupling. This observation
seems contradictory if the coupling is understood in part as
a measure of the overlap between the magnetic orbitals of
the metal ions and the orbitals from the ligands that facili-
tating the superexchange, unless the slight elongation is
more detrimental to potential ferromagnetic pathways than
antiferromagnetic ones, as was previously suggested for a
related system.[19] It could also be that this correlation is
accidental and that the true parameter related to the cou-
pling is indeed the bonding angle. In view of these observa-
tions, it was deemed appropriate to examine in more detail
all the previously reported compounds that feature Co–Cu
heterometallic fragments with both metals exhibiting two
monoatomic oxygen bridges. The Cambridge Structural
(1)
Figure 3. Plots of χMT vs. T for 1, 2, and 3. The solid lines are fits
to the experimental data (see text for details).
ˆ
In this Hamiltonian, L[and ]1 and S1 are the orbital and
spin angular moments of the CoII ion, and S2 and S3 are
the spin operators of each copper ion. Likewise, g1, g2, and
g3 are the isotropic gyromagnetic ratios of the CoII and
both CuII centers, respectively. The terms J, λ, and σ corre-
spond to the exchange coupling constant between Co and
Cu, the spin-orbit coupling constant of Co, and a combined
ˆ
ˆ
orbitalreductionparameterofthismetal,respectively;[21]μBand
Ǟ
B have the usual meanings. The program PHI[23] was em- Database (February 2014 update) yields 56 hits involving
ployed for the diagonalization of the matrix arising from this moiety. We summarize the systems for which magnetic
this Hamiltonian and to obtain the parameters that best fit data have been modeled in Table 3, together with the mag-
the experimental data. To avoid the overparameterization nitude of the coupling constants reported as well as their
of the problem, the values of g1 and g2 = g3 were fixed to most relevant structural parameters. Analysis of these data
reasonable values of 2.0 and 2.2, respectively. Additionally, reveals an approximate correlation between the average Cu–
the data could only be fitted satisfactorily if the contri- O–Co angle and J (Figure 4). This correlation is exception-
bution of a small amount of a CoII paramagnetic impurity ally good if one considers the following aspects: (1) the
was considered (amounting to molar fractions of 0.1, 0.1, magnetic data have been simulated by using a large variety
and 0.05 in 1, 2, and 3, respectively). The best solutions of different models (see above) for the various systems,
were obtained for J, λ, and σ values of (in the 1/2/3 format) (2) the coordination geometries and even the coordination
–33.4/–11.4/–2.15 cm–1, –180/–170/–165 cm–1, and –1.30/ numbers of the CoII ions change significantly from com-
–1.16/–0.99, respectively, and the ensuing simulations are pound to compound and, thus, they have different degrees
represented as solid lines in Figure 3. Of these results, the of influence of the orbital angular momentum on the mag-
most salient feature is the gradation in the strength of the netic properties, (3) other parameters, which are known to
magnetic exchange between the CuII and CoII ions from 1 influence the magnetic coupling, such as the planarity of
to 3. Thus, it is tempting to search for a correlation between the [Cu(μ-O)2Co] core, the M–O bond angles, and the C–
the magnitude of this coupling and some structural param- O···O angles, vary from complex to complex. Taking these
eter. In this context, a very common occurrence is a corre- and other aspects into consideration, it can be concluded
lation between the M–X–M angle (“X” is a monoatomic that the bridging Cu–O–Co angle must indeed be consid-
bridge) and the exchange coupling constant.[24] However, ered as a very relevant parameter in defining the extent of
such correlations can only be performed when the remain- the magnetic exchange. It is very likely that if the data for
ing structural parameters or the chemical nature of the sys- all these compounds were treated by the same model, a
tems compared remain constant. In this case, the systems much better correlation would be obtained. A possible rela-
may not be directly comparable as 2 and 3 exhibit, in ad- tionship between the average Cu–O and Co–O bond lengths
Eur. J. Inorg. Chem. 0000, 0–0
6
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
Table 3. List of all complexes from the literature containing the Co(μ-O)2Cu moiety for which the magnetic data have been analyzed,
together with the J constants determined for them and relevant metric parameters.
Complex
Cu–O–Co [°][a]
Cu–O [Å][a]
Co–O [Å][a]
J [cm–1][b]
Ref.
[c]
[9c]
[CuCo3(MeDea)3Cl3(CH3OH)0.55(H2O)0.45
[Cu2Co(NCS)3(N(CH3)2C2H4O)3]
[Cu2Co(OH)I2(N(CH3)2C2H4O)3]2
]
93.74
101.04
99.145
98.54
96.7
100.34
95.3
102.6
100.07
100.11
98.48
98.99
104.77
97.74
101.14
2.171
1.941
1.958
1.909
1.892
1.947
1.99
1.884
1.904
1.902
1.935
1.912
1.955
1.921
1.955
2.130
2.095
2.128
2.082
2.122
2.088
2.046
2.124
2.041
2.103
2.016
2.004
2.021
2.097
1.993
–2.38
–20.05
–24.35
–12.5
–16.3
–19
+1.6
–8.0
–15
–20.2
–26.2
–51
–50.6
–8.0
[27]
[27]
[d]
[11i]
[19]
[Cu2Co2(L1)2(H2O)2]n
[CuCo(prp2en)(hfa)2][e]
[f]
[10]
[Cu2CoPbCl4(L2)4]2
[10]
[Cu2CoCd2Cl6(L2)4(MeHO)2][f]
[g]
[20]
[(CuL3)2(CuCoL3(H2O)3)](ClO4)2
[11a]
[11k]
[28]
[CuCo(prp2pr)(hfa)2][e,h]
[e,i]
[CuCo(L4)(hfa)]2
[Cu2Co(L5)2(H2O)][j]
[CuCo(salen)Cl2][k]
[CuCo(L6)(NCS)2][l]
[29]
[11d]
[11c]
[11c]
[m]
[CoCu(L7)](ClO4)2
[n]
[CoCu(L8)(MeCN)(PrOH)](ClO4)2
–38
[a] Average values. [b] The Hamiltonian convention employed is H = –2JS1S2. [c] H2MeDea = N-methyldiethanolamine. [d] H4L1 = 2-
hydroxy-3-[(E)-({2-[(2-hydroxybenzoyl)amino]ethyl}imino)methyl]benzoic acid. [e] H2prp2en = N,NЈ-ethylenebis(2-hydroxypropiophen-
oneimine); hfa– = hexafluoroacetylacetonato. [f] HL2 = 2-(dimethylamino)ethanol. [g] H2L3 = N,NЈ-ethylenebis(3-ethoxysalicylaldimine).
[h] H2L4 = N,NЈ-propylenebis(2-hydroxypropiophenoneimine). [i] H3L4 = 1-(2-hydroxybenzamido)-2-[(2-hydroxy-3-methoxybenzylidene)
amino]ethane. [j] H3L5 = N-3-carboxylsalicylidene-NЈ-salicylaldehyde-1,2-diaminoethane. [k] H2salen = N,NЈ-ethylenebis(salicyliden-
imine). [l] H2L6 = N,NЈ-propylenebis(3-methoxysalicylideneimine). [m] H2L7 = two 2-[(methylamino)methyl]-6-(iminomethyl)-4-
bromophenol units linked at both ends by a –(CH2)2– chain. [n] H2L8 = two 2-[(methylamino)methyl]-6-(iminomethyl)-4-bromophenol
units linked by a –(CH2)2– chain at one end and a –(CH2)3– chain at the other.
with J was also investigated. For the former parameter, and Cu(μ1,1-O)2Co(μ1,1-O)2Cu core, depending upon the coor-
contrary to what the same analysis for only 1, 2, and 3 dination modes of the anionic coligands. The average Cu–
suggested (see above), almost no correlation was observed O–Co angle of 91.5(1)° in 3 appears to be the lowest re-
(Figure S6). As for the dependence of J with the average ported to date for a diphenoxido-bridged CuII–CoII com-
Co–O bond lengths, an inverse relationship could be in- pound. All of these complexes are overall antiferromagnetic
ferred, although of a much lower quality (Figure S7, Sup- and lead to final ground states with effective spins of 1/2
porting Information).
at low temperature; hence, the crossover angle is yet to be
achieved. However, a trend is clearly observed as the anti-
ferromagnetic interaction, which has the highest value for 1
with the largest bridging angle, decreases as the phenoxido
bridging angle decreases. Thus, the present study validates
the strategy of using metalloligands to synthesize hetero-
metallic spin-coupled systems with a predetermined nu-
clearity in an easy and effective manner. In addition, the
inherent flexibility of the basic structural unit allows the
modulation of the bridging angles through the judicious
choice of anion, which make these groups of complexes ex-
tremely valuable for investigations of the influence of the
bridging angle on magnetic exchange interactions between
heterometallic centers.
Figure 4. Plot of –J vs. Cu–O–Co angle (α) for 1–3 and the com-
pounds from Table 3. The red line is a linear regression, which pro-
duced a best fit for –J = –286 + 3.1α.
Experimental Section
Conclusions
Starting Materials: Reagent grade salicylaldehyde and 1,3-propane-
diamine were purchased from Lancaster and used as received. The
reagents and solvents used were of commercially available reagent
quality, unless otherwise stated.
To assess the magnetic coupling in phenoxido-bridged
CuII–CoII entities and to find the antiferromagnetic-to-fer-
romagnetic crossover angle, three trinuclear Cu–Co com-
plexes have been prepared by utilizing a copper metalloli-
gand derived from a N2O2-donor salen-type Schiff base and
different anionic coligands. The complexes obtained exhibit
linear or bent arrangements of the metal centers associated
Caution! Perchlorate and azide salts of metal complexes with or-
ganic ligands are potentially explosive. Only a small amount of
material should be prepared and it should be handled with care.
Schiff Base Ligands H2L and the Metalloligand [CuL]: The Schiff
with a wide range of phenoxido bridging angles within the base ligand H2L was prepared by following reported methods.[30]
Eur. J. Inorg. Chem. 0000, 0–0
7
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
Briefly, 1,3-propanediamine (0.42 mL, 5 mmol) was mixed with sal-
icylaldehyde (1.0 mL, 10 mmol) in methanol (20 mL). The resulting
solution was heated to reflux for ca. 2 h and then cooled to room
temperature. The yellow methanolic solution was used directly for
complex formation. To a methanolic solution (20 mL) of Cu(ClO4)2·
6H2O (1.852 g, 5 mmol), a methanolic solution of H2L (5 mmol,
20 mL) and triethylamine (1.4 mL, 10 mmol) were added to pre-
pare the precursor metalloligand [CuL], as reported previously.[30]
the Patterson method by using SHELXS 97. Subsequent difference
Fourier syntheses and least-squares refinements revealed the posi-
tions of the remaining non-hydrogen atoms, which were refined
with independent anisotropic displacement parameters. However,
for 1, the disordered carbon atom C(26) was refined isotropically.
Hydrogen atoms were placed in idealized positions, and their dis-
placement parameters were fixed to be 1.2 times those of the at-
tached non-hydrogen atom. Successful convergence was indicated
by maximum shift/errors of 0.001 for the last cycle of the least-
squares refinement. Owing to the intrinsic nature of the crystal, the
data for 2 were not ideal. Absorption corrections were applied by
using the SADABS program.[31] All calculations were performed
by using SHELXS 97,[32] SHELXL 97,[33] PLATON 99,[34] OR-
TEP-32,[35] and WinGX system version1.64.[36] The data collection
and structure refinement parameters and crystallographic data for
the three complexes are given in Table 4.
Complexes [(CuL)2Co(NCS)2] (1), [(CuL)2Co(O2CPh)2] (2), and
[(CuL)2Co(N3)2] (3): To a methanolic solution (20 mL) of the pre-
cursor complex [CuL] (0.718 g, 2 mmol), a solution of Co(ClO4)2·
6H2O (0.367 g, 1 mmol) in methanol (2 mL) was added, followed
by an aqueous solution of ammonium thiocyanate (0.159 g,
2.0 mmol, 2 mL), sodium benzoate (0.288 g, 2.0 mmol, 4 mL), or
sodium azide (0.130 g, 2.0 mmol, 2 mL) for 1, 2, and 3, respectively.
In all cases, green microcrystalline products started to separate
rapidly. The stirring was continued for 1 h, and then the solid com-
pound was collected by filtration, washed with methanol, and dried
in a vacuum desiccator containing anhydrous CaCl2. The filtrate
was allowed to stand overnight at room temperature, and single
crystals of X-ray quality appeared at the bottom of the vessel of
each solution. The crystals were washed with a methanol–water
mixture, dried in a desiccator containing anhydrous CaCl2, and
subsequently characterized by elemental analysis, spectroscopic
methods, and X-ray diffraction.
Table 4. Parameters for data collection and structure refinement for
1–3.
1
2
3
Formula
C
36H32N6O4-
C48H42N4O8-
Cu2Co
988.89
monoclinic
C2/c
25.577(5)
10.524(5)
17.594(5)
96.308(5)
4707(3)
4
C34H32N10O4-
Cu2Co
830.73
monoclinic
P21/c
9.366(5)
11.135(5)
16.009(5)
102.556(5)
1629.7(12)
2
S2Cu2Co
862.81
monoclinic
P21/c
12.267(5)
10.089(5)
28.578(5)
101.597(5)
3465(2)
4
Formula weight
Space group
Crystal system
a /Å
b /Å
c /Å
Complex 1: Yield 0.561 g, 65%. C36H32CoCu2N6O4S2 (862.81):
calcd. C 50.11, H 3.74, N 9.74; found C 50.34, H 3.87, N 9.54.
UV/Vis: λmax (MeOH, absorbance) = 596, 360, 274 nm; λmax (solid,
β /°
V /Å3
reflectance) = 1310, 867, 619, 356 nm. IR: ν = 2044 [ν(N=C=S–)],
˜
Z
1619 [ν(C=N)] cm–1.
Dcalcd. /gcm–3
μ /mm–1
F(000)
1.654
1.861
1756
1.395
1.299
2028
1.693
1.854
846
Complex 2: Yield 0.613 g, 62%. C48H42CoCu2N4O8 (988.89): calcd.
C 58.30, H 4.28, N 5.67; found C 58.46, H 4.12, N 5.85. UV/Vis:
λmax (MeOH, absorbance) = 598, 360, 274 nm; λmax (solid, reflec-
Rint
0.0267
1.5–25.3
23596
0.1425
1.6–25.1
21680
0.0269
2.3–25.3
11159
tance) = 1290, 617, 356 nm. IR: ν = 1598, 1561 [νs+as(COO–)],
θ range /°
Total reflections
Unique reflections
˜
[1624 ν(C=N)] cm–1.
6237
4200
2918
Data with IϾ2σ(I) 5328
2858
2402
Complex 3: Yield 0.573 g, 69%. C34H32CoCu2N10O4 (830.73):
calcd. C 49.16, H 3.88, N 16.86; found C 49.25, H 4.05, N 16.98.
R1 on IϾ2σ(I)
wR2 [IϾ2σ(I)]
GOF on F2
0.0290
0.0912
0.2581
1.138
0.0290
0.0723
1.025
0.0710
1.020
293
UV/Vis: λmax (MeOH, absorbance) = 593, 360, 274 nm; λmax (solid,
–
reflectance) = 1281, 609, 356 nm. IR: ν = 2062 [ν(N )], 1613
˜
3
Temperature /K
293
293
[ν(C=N)] cm–1.
Physical Measurements: Elemental analyses (C, H, and N) were
performed by using a Perkin–Elmer 2400 series II CHN analyzer.
IR spectra (4000–500 cm–1) of samples as KBr pellets were re-
corded by using a Perkin–Elmer RXI FTIR spectrophotometer. All
solutions were prepared with spectroscopic grade solvents. Elec-
tronic spectra were recorded with samples in methanol (800–
200 nm) in a 1 cm optical glass cuvette as well as in the solid state
(1400–300 nm) with a Hitachi U-3501 spectrophotometer. The
variable-temperature magnetic susceptibility data of crystalline
samples of 1–3 were collected with a Quantum Design supercon-
ducting quantum interference device (SQUID) magnetometer at
the Serveis Cientificotècnics of the Universitat de Barcelona. Pas-
cal’s constants were used to estimate diamagnetic corrections to the
molar paramagnetic susceptibility, and a correction was applied for
the sample holder.
CCDC-988253 (for 1), -988254 (for 2), and -988255 (for 3) contain
the supplementary crystallographic data for this paper. These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Supporting Information (see footnote on the first page of this arti-
cle): Electronic absorption and reflectance spectra of the com-
plexes, schematic differences between 2 and 3, plot of –J vs. Cu–
O–Co angle, average Cu–O and Co–O bond lengths in 1–3, plot
of –J vs. average Cu–O and Co–O bond lengths in the complexes
listed in Table 3.
Acknowledgments
A. G. thanks the Department of Science and Technology (DST),
New Delhi for financial support (grant number SR/S1/IC/0034/
2012) as well as for the DST-FIST funded single-crystal X-ray dif-
fractometer facility at the Department of Chemistry, University of
Calcutta. S. G. is thankful to the University Grants Commission
(UGC), New Delhi for a Senior Research Fellowship. G. A.
acknowledges the Spanish Ministerio de Economía y Competitivi-
dad (MINECO) (project number CTQ2012-32247) and the Catalan
Crystallographic Data Collection and Refinement: Suitable single
crystals of each of the three complexes were mounted in a Bruker
AXS SMART APEX II diffractometer equipped with a graphite
monochromator and Mo-Kα (λ = 0.71073 Å) radiation. The crys-
tals were positioned 60 mm from the CCD. 360 frames were mea-
sured with a counting time of 5 s. The structures were solved by
Eur. J. Inorg. Chem. 0000, 0–0
8
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
6645; n) C. Fraser, R. Ostrander, A. L. Rheingold, C. White,
B. Bosnich, Inorg. Chem. 1994, 33, 324–337.
Government for the ICREA Academia 2013 prize for excellence in
research. P. G. acknowledges the Spanish ICREA (Institució Catal-
ana de Recerca i Estudis Avançats) and the MINECO (project
number CTQ2011-27929-C02-01). Prof. Michael G. B. Drew,
School of Chemistry, University of Reading, UK is thanked for
solving the disorder in 1.
[12] a) S. Biswas, S. Naiya, M. G. B. Drew, C. Estarellas, A. Fron-
tera, A. Ghosh, Inorg. Chim. Acta 2011, 366, 219–226; b) S.
Biswas, A. Ghosh, Indian J. Chem. Sect. A: Inorg. Bio-Inorg.,
Phys., Theor. Anal. Chem. 2011, 50, 1356–1365; c) S. Biswas, A.
Ghosh, Polyhedron 2013, 65, 322–331; d) S. Biswas, A. Ghosh,
Polyhedron 2011, 30, 676–681; e) L. K. Das, R. M. Kadam, A.
Bauzá, A. Frontera, A. Ghosh, Inorg. Chem. 2012, 51, 12407–
12418; f) S. Biswas, C. Diaz, A. Ghosh, Polyhedron 2013, 51,
96–102; g) S. Ghosh, Y. Ida, T. Ishida, A. Ghosh, Cryst. Growth
Des. 2014, 14, 2588Ϫ2598.
[13] K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, Wiley, New York, 4th ed., 1986.
[14] a) E. Koenig, Struct. Bonding (Berlin) 1972, 9, 175–212; b)
Y. S. Dou, J. Chem. Educ. 1990, 67, 134–135.
[15] L. K. Das, S.-W. Park, S. J. Cho, A. Ghosh, Dalton Trans. 2012,
41, 11009–11017.
[16] J. Vallejo, I. Castro, R. Ruiz-García, J. Cano, M. Julve, F. Llo-
ret, G. De Munno, W. Wernsdorfer, E. Pardo, J. Am. Chem.
Soc. 2012, 134, 15704–15707.
[17] L. Yang, D. R. Powell, R. P. Houser, Dalton Trans. 2007, 955–
964.
[18] A. W. Addison, T. N. Rao, J. Reedijk, J. van Rijn, G. C. Ver-
schoor, J. Chem. Soc., Dalton Trans. 1984, 1349–1356.
[19] C. J. O’connor, D. P. Freyberg, E. Sinn, Inorg. Chem. 1979, 18,
1077–1088.
[1] O. Kahn, Molecular Magnetism, Wiley-VCH, Weinheim, Ger-
many, 1993.
[2] a) M. Andruh, Chem. Commun. 2007, 2565–2577; b) O. Kahn,
Adv. Inorg. Chem. 1995, 43, 179–259; c) K. S. Murray, Adv.
Inorg. Chem. 1995, 43, 261–358.
[3] a) O. Kahn, P. Tola, G. Galy, H. Coudanne, J. Am. Chem. Soc.
1978, 100, 3931–3933; b) A. Bencini, C. Benelli, A. Caneschi,
R. L. Carlin, A. Dei, D. Gatteschi, J. Am. Chem. Soc. 1985,
107, 8128–8136; c) Y. Journaux, O. Kahn, I. Morgenstern-Bad-
arau, J. Galy, J. Jaud, A. Bencini, D. Gatteschi, J. Am. Chem.
Soc. 1985, 107, 6305–6312; d) J. M. Zadrozny, D. J. Xiao, M.
Atanasov, G. J. Long, F. Grandjean, F. Neese, J. R. Long, Nat.
Chem. 2013, 5, 577–581; e) M. N. Leuenberger, D. Loss, Nature
2001, 410, 789–793.
[4] K. J. Cho, D. W. Ryu, K. S. Lim, W. R. Lee, J. W. Lee, E. K.
Koh, C. S. Hong, Dalton Trans. 2013, 42, 5796–5804.
[5] a) D. Gatteschi, R. Sessoli, J. Villain, Molecular Nanomagnets,
Oxford University Press, New York, 2006; b) E. Pardo, R.
Ruiz-García, F. Lloret, J. Faus, M. Julve, Y. Journaux, F.
Delgado, C. Ruiz-Pérez, Adv. Mater. 2004, 16, 1597–1600.
[6] a) S. J. Gruber, C. M. Harris, E. Sinn, Inorg. Nucl. Chem. Lett.
1967, 3, 495–499; b) O. Atakol, L. Tatar, M. A. Akay, D. Ülkü,
Anal. Sci. 1999, 15, 101–102; c) S. Biswas, S. Naiya, C. J.
Gómez-García, A. Ghosh, Dalton Trans. 2012, 41, 462–473; d)
P. Seth, L. K. Das, M. G. B. Drew, A. Ghosh, Eur. J. Inorg.
Chem. 2012, 2232–2242.
[20] M. Nayak, R. Koner, H.-H. Lin, U. Flörke, H.-H. Wei, S. Mo-
hanta, Inorg. Chem. 2006, 45, 10764–10773.
[21] F. Lloret, M. Julve, J. Cano, R. Ruiz-García, E. Pardo, Inorg.
Chim. Acta 2008, 361, 3432–3445.
[22] M. E. Lines, J. Chem. Phys. 1971, 55, 2977–2984.
[23] N. F. Chilton, R. P. Anderson, L. D. Turner, A. Soncini, K. S.
Murray, J. Comput. Chem. 2013, 34, 1164–1175.
[7] S. Ghosh, S. Mukherjee, P. Seth, P. S. Mukherjee, A. Ghosh,
Dalton Trans. 2013, 42, 13554–13564.
[8] a) P. Seth, S. Ghosh, A. Figuerola, A. Ghosh, Dalton Trans.
2014, 43, 990–998; b) A. K. Ghosh, A. Bauzá, V. Bertolasi, A.
Frontera, D. Ray, Polyhedron 2013, 53, 32–39.
[24] a) T. K. Karmakar, B. K. Ghosh, A. Usman, H.-K. Fun, E.
Rivière, T. Mallah, G. Aromí, S. K. Chandra, Inorg. Chem.
2005, 44, 2391–2399; b) V. H. Crawford, H. W. Richardson,
J. R. Wasson, D. J. Hodgson, W. E. Hatfield, Inorg. Chem.
1976, 15, 2107–2110; c) M. A. Halcrow, J.-S. Sun, J. C. Huff-
man, G. Christou, Inorg. Chem. 1995, 34, 4167–4177.
[25] a) G. Aromí, S. M. J. Aubin, M. A. Bolcar, G. Christou, H. J.
Eppley, K. Folting, D. N. Hendrickson, J. C. Huffman, R. C.
Squire, H.-L. Tsai, S. Wang, M. W. Wemple, Polyhedron 1998,
17, 3005–3020; b) A. Escuer, G. Aromí, Eur. J. Inorg. Chem.
2006, 4721–4736.
[26] a) S. S. Tandon, S. D. Bunge, D. Motry, J. S. Costa, G. Aromí,
J. Reedijk, L. K. Thompson, Inorg. Chem. 2009, 48, 4873–4881;
b) L. K. Thompson, S. K. Mandal, S. S. Tandon, J. N. Bridson,
M. K. Park, Inorg. Chem. 1996, 35, 3117–3125.
[27] V. G. Makhankova, O. Y. Vassilyeva, V. N. Kokozay, B. W.
Skelton, L. Sorace, D. Gatteschi, J. Chem. Soc., Dalton Trans.
2002, 4253–4259.
[28] R.-J. Tao, C.-Z. Mei, S.-Q. Zang, Q.-L. Wang, J.-Y. Niu, D.-Z.
Liao, Inorg. Chim. Acta 2004, 357, 1985–1990.
[29] A. Gutiérrez, M. F. Perpiñán, A. E. Sánchez, M. C. Torralba,
M. R. Torres, Inorg. Chim. Acta 2010, 363, 1837–1842.
[30] S. Ghosh, S. Biswas, A. Bauzá, M. Barceló-Oliver, A. Frontera,
A. Ghosh, Inorg. Chem. 2013, 52, 7508–7523.
[31] SAINT, v. 6.02 and SADABS, v. 2.03, Bruker AXS, Inc., Madi-
son WI, 2002.
[32] G. M. Sheldrick, SHELXS 97, Program for Structure Solution,
University of Göttingen, Germany, 1997.
[9] a) O. Kahn, R. Claude, H. Coudanne, J. Chem. Soc., Chem.
Commun. 1978, 1012–1013; b) G. E. Kostakis, S. P. Perlepes,
V. A. Blatov, D. M. Proserpio, A. K. Powell, Coord. Chem. Rev.
2012, 256, 1246–1278; c) D. S. Nesterov, V. N. Kokozay, J. Jezi-
erska, O. V. Pavlyuk, R. Boca, A. J. L. Pombeiro, Inorg. Chem.
ˇ
2011, 50, 4401–4411.
[10] D. S. Nesterov, V. N. Kokozay, B. W. Skelton, J. Jezierska, A.
Ozarowski, Dalton Trans. 2007, 558–564.
[11] a) G. A. Brewer, E. Sinn, Inorg. Chem. 1987, 26, 1529–1535; b)
T. Aono, H. Wadda, M. Yonemura, H. Furutachi, M. Ohba,
H. Okawa, J. Chem. Soc., Dalton Trans. 1997, 3029–3034; c)
M. Yonemura, K. Arimura, K. Inoue, N. Usuki, M. Ohba,
H. Okawa, Inorg. Chem. 2002, 41, 582–589; d) J.-P. Costes, R.
Gheorghe, M. Andruh, S. Shova, J.-M. C. Juan, New J. Chem.
2006, 30, 572–576; e) D. G. Branzea, A. M. Madalan, S. Ciat-
tini, N. Avarvari, A. Caneschi, M. Andruh, New J. Chem. 2010,
34, 2479–2490; f) A. Hori, Y. Mitsuka, M. Ohba, H. Okawa,
Inorg. Chim. Acta 2002, 337, 113–121; g) Y. Zhang, S.-Z.
Zhang, J.-D. Li, J. Dang, R.-J. Tao, Chin. J. Struct. Chem. 2010,
29, 1321–1326; h) R.-J. Tao, F.-A. Li, S.-Q. Zang, Y.-X. Cheng,
Q.-L. Wang, J.-Y. Niu, D.-Z. Liao, J. Coord. Chem. 2006, 59,
901–919; i) R.-J. Tao, F.-A. Li, S.-Q. Zang, Y.-X. Cheng, Q.-L.
Wang, J.-Y. Niu, D.-Z. Liao, Polyhedron 2006, 25, 2153–2159;
j) M. Mikuriya, H. Okawa, S. Kida, I. Ueda, Bull. Chem. Soc.
Jpn. 1978, 51, 2920–2923; k) S. Osa, Y. Sunatsuki, Y. Yamam-
oto, M. Nakamura, T. Shimamoto, N. Matsumoto, N. Re, In-
org. Chem. 2003, 42, 5507–5512; l) Y. Nakamura, M. Yone-
mura, K. Arimura, N. Usuki, M. Ohba, H. Okawa, Inorg.
Chem. 2001, 40, 3739–3744; m) N. B. O’Bryan, T. O. Maier,
I. C. Paul, R. S. Drago, J. Am. Chem. Soc. 1973, 95, 6640–
[33] G. M. Sheldrick, SHELXL 97, Program for Crystal Structure
Refinement, University of Göttingen, Germany, 1997.
[34] A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7–13.
[35] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565.
[36] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837–838.
Received: March 7, 2014
Published Online: ᭿
Eur. J. Inorg. Chem. 0000, 0–0
9
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Job/Unit: I42151
/KAP1
Date: 16-06-14 11:53:06
Pages: 10
www.eurjic.org
FULL PAPER
Magnetic Properties
S. Ghosh, G. Aromí,* P. Gamez,
A. Ghosh* ....................................... 1–10
Anion-mediated geometrical variations
generate three heterometallic complexes
with distinct trinuclear Cu2Co cores as a
result of different diphenoxido bridging
angles between the CuII and CoII ions.
Magnetic studies reveal decreased antifer-
romagnetic exchange interactions associ-
ated with a decrease of this bridging angle.
The Impact of Anion-Modulated Struc-
tural Variations on the Magnetic Coupling
in Trinuclear Heterometallic CuII–CoII
Complexes Derived from
Schiff Base Ligand
a Salen-Type
Keywords: Metalloligands / Copper / Co-
balt / Bridging ligands / Schiff bases / Mag-
netic properties / Magnetostructural corre-
lations
Eur. J. Inorg. Chem. 0000, 0–0
10
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim