Synthesis, structural characterization, and magnetic properties of a copper-gadolinium complex derived from a hydroxybenzohydrazide ligand

Article abstract of DOI:10.1021/ic4027283

The reaction of hydroxybenzohydrazide with o-vanillin yields 2-hydroxy-N′-[(2-hydroxy-3-methoxyphenyl)methylidene]benzohydrazide (LH₃), a ligand that is able to give mononuclear and tetranuclear copper complexes but also to associate copper and gadolinium ions in a Cu ₂-Gd₂ heterotetranuclear complex. This synthesis is successful if the Gd ions, which are acidic in protic solvents, are introduced in a basic methanol solution of the mononuclear copper complex. In the absence of piperidine, the addition of Gd ions to a methanol solution of the mononuclear copper complex only yields a tetranuclear cubane-type copper complex. This work reports on the first structural characterization of a copper-gadolinium complex involving a benzohydrazide ligand. The resulting complex consists of two Cu-Gd pairs linked by a dihydroxo Gd-Gd bridge, in which the Cu and Gd ions are bridged by a nonsymmetric phenoxo-hydroxo bridge. The magnetostructural correlation between the ferromagnetic coupling constant and the hinge angle observed in symmetrical double-phenoxo Cu-Gd bridges remains valid for dissymmetric Cu-Gd bridges and confirms the preponderance of the structural factor over the nature of the bridge. This tetranuclear complex corresponds to two S = 4 units linked through a dihydroxo bridge introducing a weak antiferromagnetic Gd-Gd interaction and impeding the existence of a S = 8 ground state.

Full text of DOI:10.1021/ic4027283

Article
pubs.acs.org/IC
Synthesis, Structural Characterization, and Magnetic Properties of a
Copper−Gadolinium Complex Derived from a
Hydroxybenzohydrazide Ligand
Jean-Pierre Costes,*,†,‡ Carine Duhayon,†,‡ and Laure Vendier†,‡
†Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, BP 44099, F-31077 Toulouse Cedex 4, France
‡Universite
́
de Toulouse, UPS, INPT, F-31077 Toulouse Cedex 4, France
S
* Supporting Information
ABSTRACT: The reaction of hydroxybenzohydrazide with o-vanillin
yields 2-hydroxy-N′-[(2-hydroxy-3-methoxyphenyl)methylidene]-
benzohydrazide (LH3), a ligand that is able to give mononuclear and
tetranuclear copper complexes but also to associate copper and
gadolinium ions in a Cu2−Gd2 heterotetranuclear complex. This synthesis
is successful if the Gd ions, which are acidic in protic solvents, are
introduced in a basic methanol solution of the mononuclear copper
complex. In the absence of piperidine, the addition of Gd ions to a
methanol solution of the mononuclear copper complex only yields a
tetranuclear cubane-type copper complex. This work reports on the first
structural characterization of a copper−gadolinium complex involving a
benzohydrazide ligand. The resulting complex consists of two Cu−Gd
pairs linked by a dihydroxo Gd−Gd bridge, in which the Cu and Gd ions
are bridged by a nonsymmetric phenoxo−hydroxo bridge. The magnetostructural correlation between the ferromagnetic
coupling constant and the hinge angle observed in symmetrical double-phenoxo Cu−Gd bridges remains valid for dissymmetric
Cu−Gd bridges and confirms the preponderance of the structural factor over the nature of the bridge. This tetranuclear complex
corresponds to two S = 4 units linked through a dihydroxo bridge introducing a weak antiferromagnetic Gd−Gd interaction and
impeding the existence of a S = 8 ground state.
single-molecule-magnet properties.7,8 Keeping these results in
INTRODUCTION
■
mind, we decided to prepare the ligand resulting from the
reaction of hydroxybenzohydrazide with o-vanillin, (2-hydroxy-
N′-[(2-hydroxy-3-methoxyphenyl)methylidene]benzohydrazide
(LH3), in order to prepare a heterometallic 3d−4f complex.
The present work describes the characterization of the ligand
Ligands possessing N and O donor atoms are of great interest
in inorganic, biomimetic, and medicinal chemistry because they
allow syntheses of mono-, homo-, and heteropolynuclear
transition-metal complexes. Ligands associating transition-
and lanthanide-metal ions have been largely used these last
years for the syntheses of complexes possessing interesting
magnetic properties, and several reviews dedicated to that
particular research interest have appeared.1 Among the possible
ligands, it appeared to us that the potentiality of some of them
has not been completely exploited. This is the case for ligands
derived from benzohydrazide moieties. Benzohydrazide or
hydroxybenzohydrazide synthons have been reacted with
several organic reactants such as pyridinecarboxaldehyde,2 2-
acetylpyridine,3 carboxylic acids, or derivatives such as acetic
anhydride or acyl chlorides,4 or even more exotic sulfur-
containing ligands5 to yield mono- or polynuclear transition-
metal complexes. Surprisingly, literature data furnish only a few
examples making use of salicylaldehyde or o-vanillin as the
reactant,6 with structural determinations of the N′-[(2-hydroxy-
3-methoxybenzylidene]benzohydrazide ligand and of its copper
cubane complex. More recently, some structural determinations
of dinuclear dysprosium complexes prepared with diverse
benzohydrazide ligands have been described, along with their
1
(1D and 2D H and 13C NMR) and its reaction with Cu and
Gd ions. The structural determinations of the resulting
complexes do confirm the existence of mono- and tetranuclear
copper complexes and, more interesting, the isolation of a
genuine tetranuclear copper−gadolinium entity, which is the
first example of a 3d−4f complex obtained with a
benzohydrazide ligand.
EXPERIMENTAL SECTION
■
Materials. Hydrazine hydrate, phenyl salicylate, o-vanillin,
tetramethylheptanedione (Hthd), triethylamine, piperidine, Cu-
(OAc)2·2H2O, and Gd(NO3)3·6H2O (Aldrich) were used as
purchased. High-grade solvents [diethyl ether, tetrahydrofuran
(THF), dichloromethane, acetone, and methanol (MeOH)] were
used for the syntheses of ligands and complexes.
Received: November 15, 2013
Published: February 6, 2014
© 2014 American Chemical Society
2181
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187

Inorganic Chemistry
Article
1
Ligands. 2-Hydroxybenzohydrazide. Phenyl salicylate (6.4 g, 3.0
× 10−3 mol) and hydrazine hydrate (2.0 g, 4.0 × 10−3 mol) were
heated for 20 min. The addition of isopropyl alcohol (20 mL) induced
cooling and the appearance of a white precipitate, which was filtered
off, washed with a small amount of isopropyl alcohol, and dried. Yield:
2.9 g (66%). Anal. Calcd for C7H8N2O2 (152.1): C, 55.5; H, 5.3; N,
9.7. Found: C, 55.1; H, 5.1; N, 9.5. 1H NMR (400 MHz, DMSO-d6): δ
4.68 (s, l, 2 H, NH2), 6.80 (t, J = 7.8 Hz, 1 H, CH-5), 6.87 (d, J = 7.8
Hz, 1 H, CH-3), 7.32 (t, J = 7.8 Hz, 1 H, CH-4), 7.80 (d, J = 7.8 Hz, 1
H, CH-6), 10.07 (s, l, 1 H, NH),12.45 (s, l, 1 H, OH). 13C{1H} NMR
(100.63 MHz, DMSO-d6): δ 115.51 (s, ArC-1), 117.99 (s, ArC-3),
118.27 (s, ArC-5), 128.00 (s, ArC-6), 133.54 (s, ArC-4), 160.78 (s,
ArCOH), 168.14 (s, OCNH).
mode. 1D and 2D H and 13C NMR spectra were acquired at 400.16
MHz (1H) or 100.63 MHz (13C) on a Bruker Avance 400
spectrometer using (CD3)2SO as the solvent. Chemical shifts are
given in ppm versus tetramethylsilane (1H and 13C) with the
numbering specified in Scheme 1. Magnetic data were obtained with
Scheme 1. LH3 Ligand with the Numbering Scheme
Retained for NMR Data (1H and 13C)
2-Hydroxy-N′-[(2-hydroxy-3-methoxyphenyl)methylidene]-
benzohydrazide. The addition of o-vanillin (1.52 g, 1.0 × 10−3 mol)
to a stirred MeOH solution (30 mL) of 2-hydroxybenzhydrazide (1.52
g, 1.0 × 10−3 mol), followed by 30 min of heating, induced the
formation of a bulky white precipitate, which was filtered off after
cooling, washed with MeOH and diethyl ether, and dried. Yield: 2.7 g
(95%). Anal. Calcd for C15H14N2O4 (286.3): C, 62.9; H, 4.9; N, 9.8.
1
Found: C, 62.7; H, 4.8; N, 9.7. H NMR (400 MHz, DMSO-d6): δ
a Quantum Design MPMS SQUID susceptometer. Magnetic
susceptibility measurements were performed in the 2−300 K
temperature range under a 0.1 T applied magnetic field, and
diamagnetic corrections were applied by using Pascal’s constants.9
Isothermal magnetization measurements were performed up to 5 T at
2 K. The theoretical magnetic susceptibilities were computed through
exact calculations of the energy levels associated with the spin
Hamiltonian through diagonalization of the full matrix with a general
program10 and fitted by least-squares techniques11 to the sets of
experimental magnetic data for complexes 2 and 3.
3.83 (s, 3 H, CH3), 6.88 (t, J = 8 Hz, 1 H, CH-5′), 6.97 (t, J = 7.8 Hz,
1 H, CH-5), 6.99 (d, J = 7.8 Hz, 1 H, CH-3), 7.05 (d, J = 8 Hz, 1 H,
CH-4′), 7.18 (d, J = 8 Hz, CH-6′), 7.50 (t, J = 7.8 Hz, 1 H, CH-4),
7.90 (d, J = 7.8 Hz, 1 H, CH-6), 8.70 (s, 1 H, HCN), 10.88 (s, 1 H,
NH), 11.95 (s, 2 H, OH). 13C{1H} NMR (100.63 MHz, DMSO-d6): δ
56.30 (s, OCH3), 114.44, (s, ArC-4′), 116.09 (s, ArC-1′), 117.76 (s,
ArC-3), 119.35 (s, ArC-1), 119.46 (s, ArC-5′), 119.55 (s, ArC-5),
121.18 (s, ArC-6′), 129.01 (s, ArC-6), 134.42 (s, ArC-4), 147.70 (s,
ArC-3′), 148.43 (s, ArCOH′), 149.31 (s, HCN), 159.52 (s,
ArCOH), 164.99 (s, OCNH).
Crystallographic Data Collection and Structure Determi-
nation for Complexes 1−3. Crystals of 1−3 were kept in the
mother liquor until they were dipped into oil. The chosen crystals
were mounted on a Mitegen micromount and quickly cooled to 100 K
(1 and 2) or 180 K (3). The selected crystals of 1 (red brown, 0.20 ×
0.20 × 0.15 mm3), 2 (green, 0.25 × 0.03 × 0.03 mm3), and 3 (dark
purple, 0.25 × 0.17 × 0.05 mm3) were mounted on a Bruker Kappa
Apex II (1) or an Oxford-Diffraction Gemini (2 and 3) using
molybdenum (λ = 0.71073 Å, 1 and 3) or copper radiation (λ =
1.54180, 2) and equipped with an Oxford Cryosystems cooler device.
The unit cell determination and data integration were carried out using
CrysAlis RED or SAINT packages.12−14 The structures have been
solved using SUPERFLIP15 or SHELXS-9716 and refined by least-
squares procedures using the software packages CRYSTALS17 or
WinGX, version 1.63.18 Atomic scattering factors were taken from the
International Tables for X-ray Crystallography.19 All H atoms were
refined by using a riding model. When it was possible, all non-H atoms
were anisotropically refined. Drawings of molecules are performed
with the program CAMERON20 with 30% probability displacement
ellipsoids for non-H atoms.
Complexes. [LHCu(pip)]·1/2H2O (1). A mixture of H3L (0.28 g, 1.0
× 10−3 mol), Cu(Ac)2·H2O (0.20 g, 1.0 × 10−3 mol), and piperidine
(0.30 g, 3.5 × 10−3 mol) in MeOH (20 mL) was heated for 30 min,
and a green precipitate appeared. It was filtered off, washed with
MeOH, and dried. Slow diffusion of diethyl ether into a CH2Cl2
solution of the green precipitate yielded crystals suitable for X-ray
diffraction (XRD). Yield: 0.36 g (84%). Anal. Calcd for
C20H24CuN3O4.5 (442.0): C, 54.4; H, 5.5; N, 9.5. Found: C, 54.1;
H, 5.1; N, 9.5. IR (ATR): 3601w, 3369w, 3262w, 2949w, 1621s, 1598s,
1519s, 1495s, 1449 ms, 1434m, 1383m, 1363m, 1299m, 1253m, 1240s,
1214s, 1159w, 1072w, 974w, 851w, 754m, 730m, 696w, 681w cm−1.
[LHCu]4 (2). A mixture of H3L (0.28 g, 1.0 × 10−3 mol), Cu(Ac)2·
H2O (0.20 g, 1.0 × 10−3 mol), and triethylamine (0.35 g, 3.5 × 10−3
mol) in MeOH (20 mL) was heated for 30 min, and a green
precipitate appeared. It was filtered off, washed with MeOH and
diethyl ether, and dried. Yield: 0.32 g (91%). Anal. Calcd for
C15H12CuN2O4 (347.8): C, 51.8; H, 3.5; N, 8.0. Found: C, 51.4; H,
3.3; N, 7.9 . IR (ATR): 2948w, 1622m, 1601m, 1588m, 1514m, 1490s,
1450s, 1380s, 1360s, 1245s, 1218s, 1158m, 1099w, 1083w, 1070m,
970m, 924w, 790w, 750m, 736m, 699w, 681w, 650w cm−1. The
addition of Gd(NO3)3·5H2O to a MeOH solution of 1 yielded crystals
characterized by similar analytical and IR data.
Crystal data for 1: C20H24CuN3O4.5, M = 441.97, monoclinic, Pn, Z
= 8, a = 11.3862(4) Å, b = 19.1051(7) Å, c = 18.6796(7) Å, α = γ =
90°, β = 107.783(2)°, V = 3869.3(2) Å3, 91884 collected reflections,
18806 unique reflections (Rint = 0.0338), R factor = 0.028, weighted R
factor = 0.032 for 16585 contributing reflections [I > 3σ(I)]. The
crystal was twinned. Data were treated with ROTAX, which gave the
twin law between the two components (0.65:0.35).21
[LHCu(OH)Gd(thd)2)]2 (3). To a mixture of H3L (0.28 g, 1.0 × 10−3
mol) and tetramethylheptanedione (0.36 g, 2.0 × 10−3 mol) in MeOH
(30 mL) were added Cu(Ac)2·H2O (0.20 g, 1.0 × 10−3 mol) and
Gd(NO3)3·5H2O (0.44 g, 1.0 × 10−3 mol) at once and eventually
piperidine (0.30 g, 3.5 × 10−3 mol). The resulting solution was heated
for 20 min and then left to cool with stirring, thus yielding a green
precipitate that was filtered off and dried. Slow evaporation of a THF
solution of the complex yielded crystals suitable for XRD. Yield: 0.61 g
(70%). Anal. Calcd for C74H102Cu2Gd2N4O18 (1777.2): C, 50.0; H,
5.8; N, 3.1. Found: C, 49.7; H, 5.9; N, 3.1. IR (ATR): 2952m, 2865w,
1605m, 1592m, 1575m, 1551m, 1538m, 1493s, 1461m, 1395s, 1381s,
1357s, 1242m, 1218s, 1178w, 1144w, 1104w, 1065w, 965w, 869w,
794w, 756w, 736w, 629w cm−1.
Crystal data for 2: C H Cu N O , M = 1391.27, tetragonal, P4,
̅
60 48
4
8
16
Z = 2, a = 16.8827(3) Å, b = 16.8827(3) Å, c = 12.6346(2) Å, α = β =
γ = 90°, V = 3601.20(10) Å3, 27154 collected reflections, 5425 unique
reflections (Rint = 0.0382), R factor =0.074, weighted R factor = 0.085
for 4856 contributing reflections [I > 3σ(I)]. The asymmetric unit
contains two independent half-molecules. For both molecules, the HL
ligand is disordered over two positions with a 0.5:0.5 occupancy ratio.
Crystal data for 3: C74H102Cu2Gd2N4O18, M = 1777.23,
monoclinic, P21/n, Z = 4, a = 20.8730(11) Å, b = 17.923(3) Å, c =
26.0370(15) Å, β = 112.462 (5)°, V = 9001.6(17) Å3, 18371 collected
reflections, 16423 unique reflections (Rint = 0.1146), R factor = 0.0603,
weighted R factor = 0.0694 for 6940 contributing reflections [I >
2.8σ(I)].
Physical Measurements. C, H, and N elemental analyses were
carried out at the Laboratoire de Chimie de Coordination Micro-
analytical Laboratory in Toulouse, France. IR spectra were recorded
with a Perkin-Elmer Spectrum 100 FTIR spectrometer using ATR
2182
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187

Inorganic Chemistry
Article
RESULTS
■
The reaction of 2-hydroxybenzohydrazide with o-vanillin yields
the 2-hydroxy-N′-[(2-hydroxy-3-methoxyphenyl)methylidene]-
benzohydrazide ligand, which is clearly characterized by 1D and
2D 1H and 13C NMR and chemical analysis. This ligand
possesses three functions that can be deprotonated and two
different coordination sites, so that it appears as a good
candidate to give heterometallic complexes. In a first step, the
reaction with transition-metal ions has been envisaged. Because
the main coordination site involves three donor atoms, the
ligand is reacted with copper acetate in the presence of an
excess of piperidine (pip) as the deprotonating agent. Structural
determination confirms the formation of a simple mononuclear
complex, 1, with pip being able to complete the copper
coordination sphere. Replacement of pip by triethylamine gives
an insoluble complex. The IR and analytical data confirm the
absence of triethylamine, with triethylamine being unable to
make a bond with the Cu ion. Structural characterization
indicates that we are dealing with a tetranuclear complex, 2.
Surprisingly, the same complex can be prepared by the addition
of a metal ion such as Gd to the mononuclear complex 1. On
the contrary, the reaction of the ligand with a mixture of copper
acetate and gadolinium nitrate in a 1:1 ratio in the presence of
an ancillary Hthd (Hthd = tetramethylheptanedione) ligand
and a slight excess of pip yields a third complex in which the
main ligand, thd, and Cu and Gd ions are in a 1:2:1:1 ratio, as
in complex 3. Structural determinations of these three
complexes are described in the following. We can note the
prominent role of pip in these preparations, being able to enter
into the copper coordination sphere in complex 1 or to favor
formation of the heterotetranuclear complex 3, which cannot be
isolated in an acidic medium.
Structural Determinations. The asymmetric unit of 1
includes four mononuclear [LHCu(pip)] neutral molecules and
two water molecules. One of these molecules is represented in
Figure 1, with selected bonds and angles appearing in the
corresponding figure caption. The Cu ion is located in the main
tridentate coordination site of the ligand, linked to two O and
one N atoms of the ligand, with the fourth position being
occupied by the piperidine N atom. Deprotonation of the
hydrazide and of one phenol function yields a neutral copper
complex, with the Cu ion in a square-planar environment. The
2-fold deprotonated ligand is not completely planar, while pip is
in a unique chair conformation, as a consequence of the
presence of an equatorial Cu−N bond that induces an axial
position for the N−H bond. The mean planes of the main
ligand and pip are roughly perpendicular. The Cu−N and Cu−
O bonds for the four different molecules are very similar, with
the shorter bonds involving the Cu−O and Cu−N bonds of the
hydrazone part of the ligand [from 1.901(2) to 1.919(3) Å]
and the Cu−O hydrazide bond being slightly larger [1.938(2)−
1.945(2) Å] but shorter than the Cu−N piperidine bonds
[1.993(3)−2.005(3) Å]. The nondeprotonated phenol function
is involved in hydrogen bonding with the hydrazide N atom. A
water molecule is hydrogen-bonded to the piperidine N atom
and the methoxy and phenoxo O atoms of two different
molecules positioned in a roughly perpendicular arrangement.
The asymmetric unit of 2 is made of two slightly different
LHCu units. Application of symmetry operations to each
LHCu leads to four LHCu units arranged to yield a cubane
entity. Hence, the asymmetric unit contains a quarter of cubane
1 and cubane 2. The vertices of each cubane are alternately
Figure 1. View of the mononuclear molecule 1. Selected bond lengths
(Å) and angles (deg): Cu1−N1 1.913(3), Cu1−N3 1.993(3), Cu1−
O1 1.906(2), Cu1−O3 1.938(2), O1−Cu1−N3 91.8(1), O1−Cu1−
N1 93.5(1), N1−Cu1−O3 81.7(1), O3−Cu−N3 93.1(1); Cu101−
N101 1.916(3), Cu101−N103 1.991(3), Cu101−O101 1.901(2),
Cu101−O103 1.941(2), O101−Cu101−N103 91.8(1), O101−
Cu101−N101 93.6(1), N101−Cu101−O103 82.4(1), O103−
Cu101−N103 92.4(1); Cu201−N201 1.915(3), Cu201−N203
2.005(3), Cu201−O201 1.911(2), Cu201−O203 1.942(2), O201−
Cu201−N203 92.7(1), O201−Cu201−N201 93.7(1), N201−Cu201−
O203 82.8(1), O203−Cu201−N203 92.0(1); Cu301−N301 1.919(3),
Cu301−N303 1.999(3), Cu301−O301 1.915(2), Cu301−O303
1.945(2), O301−Cu301−N303 92.6(1), O301−Cu301−N301
93.4(1), N301−Cu301−O303 82.1(1), O303−Cu301−N303 92.2(1).
occupied by four phenoxo O atoms and four Cu ions in a six-
coordinate environment. The molecular structure of a cubane is
reported in Figure 2, while selected bond distances and angles
are again given in the figure caption. Each Cu center is linked to
three phenoxo O atoms, so that it can interact with the other
three Cu ions. The head-to-tail ligand strand arrangement leads
to a unique CuNO5 chromophore, implying a hydrazone N
atom, three phenoxo O atoms, and methoxy and hydrazide O
atoms coming from three different ligands. As in complex 1,
each Cu ion is coordinated to three NO2 donor atoms coming
from the same ligand, to two O atoms (phenoxo and methoxy)
from a second ligand, and to a phenoxo O atom from a third
ligand. The Cu···Cu distances are homogeneous and vary from
3.281(1) to 3.343(2) Å. These cubane structures do present
four faces characterized by Cu−O−Cu angles of 112.9(2) and
88.3(2)° (cubane 1) against 113.8(2) and 86.9(2)° (cubane 2),
while the last two faces involve two identical Cu−O−Cu angles
equal to respectively 90.4(2) and 88.6(2)°. We observe two
short Cu−O bond lengths [1.966(4) and 1.970(4) Å and
1.963(5) and 1.981(4) Å] and one long Cu−O [2.686(5) and
2.755(5) Å for cubanes 1 and 2, respectively], so that these
cubane structures can be classified as 4 + 2 (type II)
cubanes.22,23
2183
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187

Inorganic Chemistry
Article
Figure 3. View of the tetranuclear [LHCu(OH)Gd(thd)2)]2 complex
3. Selected bond lengths and distances (Å) and angles (deg): N1−Cu1
1.910(8), O9−Cu1 1.957(8), O3−Cu1 1.912(9), O1−Cu1 1.911(9),
N3−Cu2 1.941(11), O10−Cu2 1.958(9), O5−Cu2 1.906(10), O7−
Cu2 1.935(9), O2−Gd1 2.601(9), O9−Gd1 2.454(8), O1−Gd1
2.391(8), O10−Gd1 2.443(8), O11−Gd1 2.331(9), O12−Gd1
2.274(9), O13−Gd1 2.294(10), O14−Gd1 2.359(9), O9−Gd2
2.440(9), O10−Gd2 2.489(8), O5−Gd2 2.379(8), O6−Gd2
2.578(9), O15−Gd2 2.350(9), O16−Gd2 2.305(9), O17−Gd2
2.273(8), O18−Gd2 2.333(8), Cu1−Gd1 3.363(2), Cu2−Gd2
3.392(2), Gd1−Gd2 3.999(1), Cu1−O9−Gd1 99.3(4), Cu1−O9−
Gd2 117.0(4), Cu1−O1−Gd1 102.2(4), Gd2−O9−Gd1 109.6(3),
Gd1−O10−Gd2 108.4(3), Cu2−O10−Gd2 98.7(3), Cu2−O10−Gd1
116.6(4), Cu2−O5−Gd2 104.1(4).
Figure 2. View of the tetranuclear copper complex 2. Selected bond
lengths and distances (Å) and angles (deg): Cu1−N1 1.910(5), Cu1−
O1 1.970(4), Cu1−O3 1.948(4), Cu1−O1 4_757 1.966(4), Cu1−O1
3_775 2.686(5), Cu1−O2 4_757 2.298(4), Cu2−N21 1.914(5),
Cu2−O21 1.963(5), Cu2−O23 1.953(4), Cu2−O220 4_656
2.275(4), Cu2−O21 4_656 1.981(4), Cu2−O22 4_656 2.248(4),
Cu1−Cu1 4_757 3.281(1), Cu1−Cu1 2_577 3.281(1), Cu1−Cu1
3_775 3.343(2) Å, Cu2−Cu2 4_656 3.304(1), Cu2−Cu2 2_566
3.304(1), Cu2−Cu2 3_665 3.342(2), Cu1−O1−Cu1 2_577 112.9(2),
Cu1 3_775−O1−Cu1 90.4(2), Cu1 3_775−O1−Cu1 2_577 88.3(2),
Cu2 3_665−O21−Cu2 2_566 86.9(2), Cu2 3_665−O21−Cu2
88.6(2), Cu2 2_566−O21−Cu2 113.8(2).
The molecular structure of complex 3 appears in Figure 3
with again selected bond distances and angles reported in the
figure caption. It consists of two LHCu(OH)Gd(thd)2 units
assembled through two hydroxo bridges, with each μ3-OH
bridge coordinating the Cu and Gd ions of one unit to the Gd
ion of the second one. The Cu and Gd ions of one unit are
connected by a nonsymmetric double bridge involving the
phenoxo O atom of the LH ligand and the hydroxo O atoms.
The Gd ions are thus bridged by a double hydroxo bridge,
while a simple hydroxo bridge is operating between the Cu and
Gd ions belonging to different units. There is no direct link
between the Cu ions. As in complexes 1 and 2, each Cu ion is
linked to the same three donor atoms of the LH ligand (NO2),
and it completes its square-planar environment by the hydroxo
anion. The Gd ion is eight-coordinate, with two O atoms
originating from the LH ligand (methoxy and phenoxo), two
from the hydroxo anions, and four from the two thd ancillary
ligands. The Gd−O(thd) bond lengths [2.273(8)−2.359(9) Å]
are shorter than the Gd−O(phenoxo) [2.379(8)−2.391(8) Å]
and Gd−OMe [2.578(9)−2.601(9) Å] bond lengths. Three
Gd−OH [2.440(9)−2.454(8) Å] bond lengths are different
from the fourth one [2.489(8) Å], and the Gd···Gd distance is
equal to 3.999(1) Å. The Cu···Gd distances through the
phenoxo−hydroxo bridge are very similar [3.363(2)−3.392(2)
Å], while the Cu···Gd distances through the single hydroxo
bridge are larger [3.741(2)−3.753(2) Å]. The central Gd−
(OH)2−Gd core of the molecule is surrounded on one side by
the two main ligands and by two thd ligands on the other side,
while the last two thd ligands are roughly in the plane of this
core, so that the tetranuclear complexes are well isolated from
the neighboring ones, with the shorter metal distances being
larger than 10 Å.
Magnetic Properties. In agreement with structural
determination, complex 1 presents no interest from the
magnetic point of view, with the χMT product being constant
and equal to from 300 to 2 K (Figure S1 in the Supporting
Information). The magnetic behavior of complex 2 is reported
in Figure 4 in the form of thermal variation of the χMT product
(χM is the molar magnetic susceptibility of the tetranuclear
copper complex corrected for diamagnetism of the ligands).9
χMT, which is equal to 1.16 cm3 K mol−1 at 300 K decreases
Figure 4. Temperature dependence of the χMT product for complex 2
at 0.1 T applied magnetic field. The solid line corresponds to the best
data fit (see the text).
2184
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187

Inorganic Chemistry
Article
smoothly until 150 K (0.97 cm3 K mol−1) and then more
abruptly until 2 K, where it is equal to 0.04 cm3 K mol−1
(Figure 4). The χMT product at room temperature, lower than
expected for four isolated Cu ions (1.5 cm3 K mol−1 with g = 2)
decreases from 300 to 2 K, thus indicating that an
antiferromagnetic interaction through the phenoxo bridge is
active, with the nonzero value at 2 K being due to the presence
of a paramagnetic term. A qualitative analysis was first
performed with a simple isotropic Hamiltonian H =
−2J(SCu1·SCu2 + SCu2·SCu3 + SCu3·SCu4 + SCu1·SCu4). The best
fit yields an antiferromagnetic interaction parameter JCuCu
through the double phenoxo bridge equal to −36.8 cm−1,
SCu2·SGd2) − J2(SCu1·SGd2 + SCu2·SGd1) − J3(SGd1·SGd2) has been
realized with the MAGPACK program10 in order to take into
account the three possible interaction pathways evidenced by
the structural determination. The best fit leads to the parameter
values J1 = 4.8 cm−1, J2 = −0.05 cm−1, J3 = −0.013 cm−1, and g
= 2.00. The agreement factor R = ∑[(χMT)obs − (χMT)calc]2/
∑[(χMT)obs]2 = 1 × 10−4 is good. The ferromagnetic term J1
does correspond to the Cu−Gd interaction through the
nonsymmetric phenoxo−hydroxo bridge, while the Gd−Gd
interaction through the double hydroxo bridge is antiferro-
magnetic and very weak, as expected. Eventually, our
tetranuclear complex can be considered from the magnetic
point of view as an assembly of two Cu−Gd dinuclear pairs
governed by a ferromagnetic interaction. This conclusion is
confirmed by the field dependence of magnetization M at 2 K,
which is correctly fitted with these parameters (Figure 6).
with g = 2.0, a paramagnetic term of 1.7% and a correct
2
agreement factor R = ∑[(χMT)obs − (χMT)calc]2/∑[(χMT)obs
]
equal to 1 × 10−4.
In view of the structure described above, two exchange
coupling constants seem to be necessary to fit the magnetic
data, and a qualitative analysis was performed with the
following isotropic Hamiltonian H = −2J1(SCu1·SCu2 + SCu3
·
S
Cu4) − 2J2(SCu1·SCu4 + SCu2·SCu3 + SCu1·SCu3 + SCu2·SCu4).
According to the structural determination, J1 is associated with
angles close to 90° and expected to be ferromagnetic while J2,
implying a larger Cu−O−Cu angle of 114°, must be
antiferromagnetic. The best fit yields an antiferromagnetic
interaction parameter J2 through the double phenoxo bridge
equal to −35.8 cm−1, a ferromagnetic J1 parameter of 5.9 cm−1
with g = 2.0, a paramagnetic term of 1.7%, and a correct
2
agreement factor R = ∑[(χMT)obs − (χMT)calc]2/∑[(χMT)obs
]
equal to 1 × 10−4. These values are in complete agreement with
those obtained with a similar cubane complex prepared with a
Schiff base ligand resulting from the reaction of o-vanillin with
an amino acid.24
Figure 6. Field dependence of magnetization for complex 3. The solid
line corresponds to the best data fit (see the text).
Thermal variation of the χMT versus T plot of complex 3 is
similar to what was observed for previously reported copper−
gadolinium dinuclear complexes (Figure 5).25 χM is the molar
Furthermore, increasing the field up to 5 T results in a
magnetization close to the saturation value of 15.8Nβ units,
consistent with a tetranuclear complex made of two S = 4 (g =
2) units.
DISCUSSION
■
Until now, 2-hydroxybenzohydrazide has been reacted with
pyridinecarboxaldehyde2 or 2-acetylpyridine3 to yield ligands
possessing two functions that can be deprotonated, but in the
presence of 3d ions, only the NH hydrazide deprotonates.
Introducing o-vanillin instead of pyridinecarboxaldehyde
furnishes a new trianionic ligand with six potential donor
atoms distributed in two coordination sites, so that complex-
ation of 3d and 4f ions can be expected. A look at our three
structural determinations shows that the phenol function not
involved in the main coordination site never deprotonates. The
addition of gadolinium nitrate to the monomeric complex 1
yields the tetranuclear species 2, without complexation of Gd
ions. This observation can be easily explained if we remember
that positive metal ions yield acidic solutions in protic
solvents.26 Such an acidic ion protonates pip, thus leaving
free the fourth coordination site and favoring formation of the
tetranuclear complex. Replacement of pip by triethylamine
gives directly the tetranuclear complex, with triethylamine being
unable to enter in the copper coordination sphere.
Figure 5. Temperature dependence of the χMT product for complex 3
at 0.1 T applied magnetic field. The solid line corresponds to the best
data fit (see the text).
magnetic susceptibility of the tetranuclear complex, corrected
for the diamagnetism of the ligands.8 At 300 K, χMT is equal to
16.31 cm3 K mol−1, which is attributable to two CuII and two
GdIII uncoupled cations (16.5 cm3 K mol−1 with g = 2). Upon a
lowering of the temperature, χMT increases, reaching a
maximum value of 19.18 cm3 K mol−1 at 8 K, which is
followed by a slight decrease to 18.49 cm3 K mol−1 at 2 K. The
maximum value is not far from that expected (20 cm3 K mol−1)
for two pairs of ferromagnetically coupled CuII and GdIII ions. A
If benzohydrazide ligands were first used to prepare mono-
or polynuclear transition-metal complexes,2−6 it was recently
shown that picolinoylhydrazide ligands, with their supplemen-
tary pyridine N atom, were able to give dinuclear dysprosium
quantitative analysis using the Hamiltonian H = −J1(SCu1·SGd1
+
2185
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187

Inorganic Chemistry
Article
complexes.7,8 The possible preparation of homonuclear
coordinate 3d and 4f complexes encouraged us to make an
attempt to synthesize heteronuclear 3d−4f complexes, and
more precisely Cu−Gd complexes. If Gd ions are added in the
presence of an excess of pip, the solution remains basic, thus
favoring the formation of hydroxo anions and chelation of Gd
ions to the phenoxo methoxy part of the ligand. The same
complex 3 is isolated when a mixture of Cu and Gd ions is
mixed to a MeOH solution of the ligand and pip. The structural
determination of complex 3 clearly shows that the Cu ion
occupies the tridentate NO2 coordination, as in the
homonuclear copper complexes 1 and 2, while each hydroxo
anion is μ3-bridged to the two Gd ions and to one Cu ion. The
Gd ion is also linked to the phenoxo methoxy part of the ligand,
and it completes its coordination sphere with two ancillary thd
ligands and the two μ3-hydroxo anions. The nondeprotonated
phenol function is involved in hydrogen bonding with the close
hydrazide N atom.
From the magnetic point of view, each Gd ion does interact
with the other three ions (one Gd and two Cu ions), but there
is no direct link between the Cu ions. Of course, the main Cu−
Gd interaction involves the dissymmetric double phenoxo−
hydroxo bridge. Some years ago, a similar dissymmetric bridge
had been described.25 The main difference between these two
examples of dissymmetric bridges comes from the value of the
hinge angle, defined as the angle between the CuOO and
GdOO planes involving the metal ions and bridging O atoms. A
hinge angle of 0.3(1)° is associated with an interaction
parameter J value of 8.4 cm−1 in the previous case, while in
the present complex, a hinge angle of 26.1(2)° [mean value for
angles of 25.5(2) and 26.7(2)°] gives a lower J value of 4.8
cm−1. The magnetostructural correlation between the coupling
strength and hinge angle observed in symmetrical double
phenoxo Cu−Gd bridges remains valid for dissymmetric Cu−
Gd bridges and confirms the preponderance of the structural
factor over the nature of the bridge.
The magnetic coupling in dinuclear gadolinium complexes
bridged by two O atoms forming a four-membered Gd2O2 ring
has been the subject of theoretical studies.27 These complexes
are divided into two classes, according to the type of functions
involved in the bridges, carboxylate or phenolate functions that
induce differences in their geometrical parameters: unequal
Gd−O bond lengths and Gd−O−Gd angles larger than 113° in
Ln2O2 carboxylate bridges, characterized by weak ferromagnetic
interaction; nearly equal Gd−O bond lengths and Gd−O−Gd
angles lower than 113° in Ln2O2 phenolate bridges giving weak
antiferromagnetic interactions. In our complex 3, the Gd−O2−
Gd core is not planar, with a hinge angle of 17.0(2)°, but the
bond lengths [2.436(6)−2.450(7) Å], the angles at the oxygen
bridgeheads [108.2(2) and 109.7(2)°], and the Gd···Gd
distance [4.000(1) Å] do correspond to criteria favoring
antiferromagnetic interactions. Our best fit yields a J3 value of
−0.013 cm−1, in agreement with the structural parameters.
Because of the lack of literature data, it is more difficult to
speak about the third interaction pathway corresponding to the
single hydroxo bridge between the Cu ion of one dinuclear unit
and the Gd ion of the second dinuclear unit. We found an
example of μ3-hydroxo anion bridging two Dy and one Cu
ions28 that gives no information on the Cu−OH−Dy
interaction. When this interaction pathway was supposed as
secondary, the presence of a ferromagnetic Cu−Gd interaction
in the main dinuclear unit and of an antiferromagnetic Gd−Gd
interaction should result in an antiferromagnetic interaction
through the single hydroxo bridge. The J2 best fit value of
−0.05 cm−1 corresponds to what is expected and does agree
with the absence of spin frustration in that complex 3.
CONCLUSION
■
We have shown, with the help of structural determinations, that
a hydrazide ligand can yield mononuclear copper complexes in
the presence of coordinating amine ligands (pip) and
tetranuclear copper complexes in a slightly acidic medium or
in the presence of noncoordinating organic bases such as
triethylamine. Furthermore, in a basic medium, this ligand is
also able to associate Cu and Gd ions. The resulting complex
corresponds to two Cu−Gd units with a dissymmetric
phenoxo−hydroxo bridge that are associated through the Gd
ions by a dihydroxo bridge. The two μ3-hydroxo anions add a
supplementary single hydroxo bridge between the Cu and Gd
ions belonging to the different Cu−Gd units. The magnetic
study confirms that the interaction through the dissymmetric
bridge is ferromagnetic and that the magnetostructural
correlation between the coupling strength and hinge angle
observed in symmetrical double phenoxo Cu−Gd bridges
remains valid for dissymmetric Cu−Gd bridges, thus
confirming the preponderance of the structural factor over
the nature of the bridge. This tetranuclear complex furnishes a
nice example of Gd ions linked through a dihydroxo bridge
introducing a weak antiferromagnetic Gd−Gd interaction,
impeding the existence of a S = 8 ground state but consistent
with a tetranuclear complex made of two S = 4 units. This work,
reporting the first structural characterization of a copper−
gadolinium complex involving a benzohydrazide ligand,
confirms that such ligands can be used in the preparation of
heteronuclear 3d−4f complexes.
ASSOCIATED CONTENT
* Supporting Information
■
S
Crystallographic data in CIF format and magnetism of complex
1 (Figure S1). This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: jean-pierre.costes@lcc-toulouse.fr.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The work was supported by CNRS, MAGMANet (Grant
NMP3-CT-2005-515767).
■
REFERENCES
■
(1) Benelli, C.; Gatteschi, D. Chem. Rev. 2002, 102, 2369−2387.
Huang, Y.-G.; Jiang, F.-L.; Hong, M.-C. Coord. Chem. Rev. 2009, 253,
2814−2834.
(2) Zhou, P.; Zhao, Y.-G.; Bai, Y.; Pang, K.-L.; He, C. Inorg. Chim.
Acta 2007, 360, 3965−3970. Samanta, B.; Chakraborty, J.; Shit, S.;
Batten, S. R.; Jensen, P.; Masuda, J. D.; Mitra, S. Inorg. Chim. Acta
2007, 360, 2471−2484.
(3) Barbazan, P.; Carballo, R.; Vazquez-Lopez, E. M. CrystEngComm
2007, 9, 668−675.
(4) Luo, W.; Wang, X.-T.; Meng, X.-G.; Cheng, G.-Z.; Ji, Z.-P.
Polyhedron 2009, 28, 300−306. Luo, W.; Meng, X.-G.; Cheng, G.-Z.;
Ji, Z.-P. Inorg. Chim. Acta 2009, 362, 551−555. Choi, J.; Park, J.; Park,
M.; Moon, D.; Lah, M. S. Eur. J. Inorg. Chem. 2008, 5465−5470. Luo,
2186
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187

Inorganic Chemistry
Article
W.; Meng, X.-G.; Xiang, J.-F.; Duan, Y.; Cheng, G.-Z.; Ji, Z.-P. Inorg.
Chim. Acta 2008, 361, 2667−2676. Luo, W.; Wang, X.-T.; Meng, X.-
G.; Cheng, G.-Z.; Ji, Z.-P. Inorg. Chem. Commun. 2008, 11, 1044−
1047. Luo, W.; Wang, X.-T.; Cheng, G.-Z.; Song, G.; Ji, Z.-P. Inorg.
Chem. Commun. 2008, 11, 769−771. John, R. P.; Park, M.; Moon, D.;
Lee, K.; Hong, S.; Zou, Y.; Hong, C. S.; Lah, M. S. J. Am. Chem. Soc.
2007, 129, 14142−14143. Moon, D.; Lee, K.; John, R. P.; Kim, G. H.;
Suh, B. J.; Lah, M. S. Inorg. Chem. 2006, 45, 7991−7993. John, R. P.;
Lee, K.; Kim, B. J.; Suh, B. J.; Rhee, H.; Lah, M. S. Inorg. Chem. 2005,
44, 7109−7121. Lin, S.; Liu, S.-X.; Chen, Z.; Lin, B.-Z.; Gao, S. Inorg.
Chem. 2004, 43, 2222−2224. Lin, S.; Liu, S.-X.; Huang, J.-Q.; Lin, C.-
C. J. Chem. Soc., Dalton Trans. 2002, 1595−1601. Liu, S.-X.; Lin, S.;
Lin, C.-C.; Huang, J.-Q. Angew. Chem., Int. Ed. 2001, 40, 1084−1087.
(5) Beghidja, C.; Rogez, G.; Kortus, J.; Wesolek, M.; Welter, R. J. Am.
Chem. Soc. 2006, 128, 3140−3141. Beghidja, C.; Wesolek, M.; Welter,
R. Inorg. Chim. Acta 2005, 358, 3881−3888.
(28) Benelli, C.; Caneschi, A.; Gatteschi, D.; Guillou, O.; Pardi, L.
Inorg. Chem. 1990, 29, 1750−1755.
(6) Chakraborty, J.; Thakurta, S.; Pilet, G.; Luneau, D.; Mitra, S.
Polyhedron 2009, 28, 819−825. Pouralimardan, O.; Chamayou, A.-C.;
Janiak, C.; Hosseini-Monfared, H. Inorg. Chim. Acta 2007, 360, 1599−
1608. Mohan, M.; Gupta, N. S.; Gupta, M. P.; Kumar, A.; Kumar, M.;
Jha, N. K. Inorg. Chim. Acta 1988, 152, 25−36.
(7) Lin, P.-O.; Korobkov, I.; Burchell, T. J.; Murugesu, M. Dalton
Trans. 2012, 41, 13649−13655. Lin, P.-O.; Burchell, T. J.; Clerac, R.;
Murugesu, M. Angew. Chem., Int. Ed. 2008, 47, 8848−8851.
(8) Zou, L.; Zhao, L.; Chen, P.; Guo, Y.-N.; Guo, Y.-N.; Li, Y.-H.;
Tang, J. Dalton Trans. 2012, 41, 2966−2971. Guo, Y.-N.; Xu, G.-F.;
Wernsdorfer, W.; Ungur, L.; Guo, Y.; Tang, J.; Zhang, Y.-N.;
Chibotaru, L. F.; Powell, A. K. J. Am. Chem. Soc. 2011, 133, 11948−
11951. Guo, Y.-N.; Chen, X.-H.; Xue, S.; Tang, J. Inorg. Chem. 2011,
50, 9705−9713. Guo, Y.-N.; Xu, G.-F.; Gamez, P.; Zhao, L.; Lin, S.-Y.;
Tang, J.; Zhang, H.-J. J. Am. Chem. Soc. 2010, 132, 8538−8539.
(9) Pascal, P. Ann. Chim. Phys. 1910, 19, 5−70.
(10) Boudalis, A. K.; Clemente-Juan, J.-M.; Dahan, F.; Tuchagues, J.-
P. Inorg. Chem. 2004, 43, 1574−1586.
(11) James, F.; Roos, M. Comput. Phys. Commun. 1975, 10, 343−367.
(12) Xcalibur CCD system and CrysAlisPro Software system, version
1.171.35.19; Agilent Technologies U.K. Ltd.: Oxford, U.K., 2011.
(13) CrysAlis CCD and CrysAlis RED and associated programs: Oxford
Diffraction Program name(s); Oxford Diffraction Ltd.: Abingdon,
England, 2006.
(14) SAINT; Bruker AXS Inc.: Madison, WI, 2007.
(15) Superflip: Palatinus, L.; Chapuis, G. J. Appl. Crystallogr. 2007, 40,
786−790.
(16) Sheldrick, G. M. SHELX97 [includes SHELXS97, SHELXL97,
CIFTAB]: Programs for Crystal Structure Analysis, release 97-2; Institut
̈
fur Anorganische Chemie der Universitat: Gottingen, Germany, 1998.
̈
̈
̈
(17) CRYSTALS: Betteridge, P. W.; Carruthers, J. R.; Cooper, R. I.;
Prout, K.; Watkin, D. J. J. Appl. Crystallogr. 2003, 36, 1487.
(18) WINGX 1.63 Integrated System of Windows Programs for the
Solution, Refinement and Analysis of Single Crystal X-ray Diffraction
Data: Farrugia, L. J. Appl. Crystallogr. 1999, 32, 837−840.
(19) International Tables for X-ray crystallography; Kynoch Press:
Birmingham, England, 1974; Vol. IV.
(20) Watkin, D. J.; Prout, C. K.; Pearce, L. J. CAMERON; Chemical
Crystallography Laboratory: Oxford, England, 1996.
(21) Cooper, R. I.; Gould, R. O.; Parsons, S.; Watkin, D. J. J. Appl.
Crystallogr. 2002, 35, 168−174.
(22) Mergehenn, R.; Haase, W. Acta Crystallogr. 1977, B33, 1877−
1882.
(23) Tercero, J.; Ruiz, E.; Alvarez, S.; Rodriguez-Fortea, A.; Alemany,
P. J. Mater. Chem. 2006, 16, 2729−2735.
(24) Puterova-Tokarova, Z.; Mrazova, V.; Kozisek, J.; Valentova, J.;
Vranovicova, B.; Boca, R. Polyhedron 2014, 70, 52−58.
(25) Chiboub Fellah, F. Z.; Costes, J. P.; Dahan, F.; Duhayon, C.;
Novitchi, G.; Tuchagues, J. P.; Vendier, L. Inorg. Chem. 2008, 47,
6444−6451.
(26) Hawkes, S. J. J. Chem. Educ. 1996, 73, 516−517.
(27) Roy, L. E.; Hughbanks, T. J. Am. Chem. Soc. 2006, 128, 568−
575.
2187
dx.doi.org/10.1021/ic4027283 | Inorg. Chem. 2014, 53, 2181−2187