Pentanuclear cyanide-bridged complexes based on highly anisotropic Co II seven-coordinate building blocks: Synthesis, structure, and magnetic behavior

Article abstract of DOI:10.1021/ic201534e

Pentagonal-bipyramidal complexes [Co(DABPH)X(H₂O)]X [X = NO ₃ (1), Br (2), I (3)] were synthesized, and their magnetic behavior was investigated. Simulation of the magnetization versus temperature data revealed the complexes to be highly anisotropic (D ≈ +30 cm^-1) and the magnitude of the anisotropy to be independent of the nature of the axial ligands. The reaction of 1 with K₃[M(CN)₆] (M = Cr, Fe) produces the pentametallic clusters [{Co(DABPH)}₃{M(CN) ₆}₂(H₂O)₂] [M = Cr (4), Fe (5)]. Both clusters consist of three {Co(DABPH)} moieties separated by two {M(CN) ₆} fragments. In 4, the central and terminal Co^II ions are bound to cyanide groups cis to one another on the bridging {Cr(CN) ₆}, whereas in 5, the connections are via trans cyanide ligands, resulting in the zigzag and linear structures observed, respectively. Magnetic investigation revealed ferromagnetic intramolecular interactions; however, the ground states were poorly isolated because of the large positive local anisotropies of the Co^II ions. The effects of the local anisotropies appeared to dominate the behavior in 5, where the magnetic axes of the Co ^II ions were approximately colinear, compared to 4, where they were closer to orthogonal.

Full text of DOI:10.1021/ic201534e

Article
pubs.acs.org/IC
Pentanuclear Cyanide-Bridged Complexes Based on Highly
Anisotropic CoII Seven-Coordinate Building Blocks: Synthesis,
Structure, and Magnetic Behavior
Luke J. Batchelor,*,† Marco Sangalli,† Regis Guillot,† Nathalie Guihery,‡ Remi Maurice,‡,∥ Floriana Tuna,§
́
́
and Talal Mallah*,†
†Institut de Chimie Molec
́
ulaire et des Mater
́
iaux d’Orsay, CNRS, Universite
́
Paris-Sud 11, 91405 Orsay, France
‡Laboratoire de Chimie et Physique Quantiques, CNRS, Universite
France
́
Paul Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex 09,
§School of Chemistry, The University of Manchester, Oxford Road, Manchester M13 9PL, United Kingdom
∥
́
́
́
Departament de Quımica Fısica i Inorganica, Universitat Rovira i Virgili, Marcel.lı Domingo s/n, 43007 Tarragona, Spain
S
* Supporting Information
ABSTRACT: Pentagonal-bipyramidal complexes [Co-
(DABPH)X(H2O)]X [X = NO3 (1), Br (2), I (3)] were
synthesized, and their magnetic behavior was investigated.
Simulation of the magnetization versus temperature data revealed
the complexes to be highly anisotropic (D ≈ +30 cm−1) and the
magnitude of the anisotropy to be independent of the nature of
the axial ligands. The reaction of 1 with K3[M(CN)6] (M = Cr,
Fe) produces the pentametallic clusters [{Co(DABPH)}3{M-
(CN)6}2(H2O)2] [M = Cr (4), Fe (5)]. Both clusters consist of
three {Co(DABPH)} moieties separated by two {M(CN)6}
fragments. In 4, the central and terminal CoII ions are bound to
cyanide groups cis to one another on the bridging {Cr(CN)6}, whereas in 5, the connections are via trans cyanide ligands,
resulting in the zigzag and linear structures observed, respectively. Magnetic investigation revealed ferromagnetic intramolecular
interactions; however, the ground states were poorly isolated because of the large positive local anisotropies of the CoII ions. The
effects of the local anisotropies appeared to dominate the behavior in 5, where the magnetic axes of the CoII ions were
approximately colinear, compared to 4, where they were closer to orthogonal.
First-order SOC is present for CoII in Oh symmetry.
INTRODUCTION
■
4
However, the degeneracy of the T1g ground state is lifted via
Polynuclear complexes of first-row transition-metal ions have
the potential to behave as single molecule magnets (SMMs),1
eventually capable of storing2 and processing3 information at
the molecular level. The SMM behavior stems from a high-spin
ground state and a strong Ising-type anisotropy, which leads to
a barrier for the reorientation of the magnetization and
subsequent blocking below a given temperature. Access to
clusters with high-spin ground states is increasingly available
with our understanding of the structural motifs that promote
ferromagnetic exchange between metal centers,4,5 and the
development of clusters with large axial anisotropies, which may
lead to Ising-type anisotropy, is arguably the harder challenge in
the search for new SMMs. One of the major sources of
anisotropy in polymetallic clusters is single-ion anisotropy.6
Large axial anisotropy is the result of the simultaneous effects of
spin−orbit coupling (SOC) and axial symmetry. Thus, the use
of axial building blocks to construct high-nuclearity clusters is a
possible route to large magnetic anisotropy in molecular
complexes.
minor distortions, resulting in most cases in a 4A2g ground state
4
4
and a Eg excited state. Provided the separation between A2g
4
and Eg is large enough, only the two lowest-energy Kramers
doublets (M = ±1/2) are populated, the separation between
which is 2D.7 For six-coordinate CoII, D can be very large, up to
+83 cm−1,8 and although always positive, the correct alignment
of the axial zero-field-splitting (ZFS) tensor of each metal
center may result in molecular Ising-type anisotropies. SMM
behavior in a CoII cluster was first reported for
[Co4(hmp)4(MeOH)4Cl4] in 2002,9 and since then, energy
barriers for relaxation as high as 84 K for octametallic
phosphonate-bridged clusters10 or 96 K for a CoII dimer in
which bridging carbene ligands introduce additional spin upon
radiation have been reported.11
The ground state of seven-coordinate Co(II) complexes is
not orbitally degenerate and well separated from other excited
states. However, the axiality of the pentagonal-bipyramidal
Received: July 20, 2011
Published: November 3, 2011
© 2011 American Chemical Society
12045
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
Table 1. Details of the Crystallographic Analysis of 2, 4, and 5
2
4
5
chemical formula
C23H25N5CoI2O4
C89 H57 N27Co3 Cr2 O36
C81H103N27Co3Fe2O26
2159.39
mol wt
748.21
2361.41
crystal dimens/mm
0.20 × 0.20 × 0.03
monoclinic
0.33 × 0.08 × 0.01
monoclinic
0.20 × 0.17 × 0.03
monoclinic
C2/c
cryst syst
space group
P21/n
C2/c
a/Å
8.0749(5)
36.133(5)
33.312(2)
16.7668(9)
17.1653(10)
90.00
b/Å
14.2955(8)
32.248(5)
c/Å
22.8291(14)
10.7698(12)
α/deg
90.00
90.00
β/deg
99.069(2)
101.268(3)
95.861(2)
90.00
γ/deg
U/Å3
90.00
90.00
2602.3(3)
12307(3)
9537.5(10)
4
Z
4
4
ρcalcd
1.910
1.274
1.504
T/K
100(1)
100(1)
100(1)
2θmax/deg
34.34
23.25
31.51
data collected
12 364
66 150
15 835
unique reflns
9146
8720
11 110
no. of param
318
712
631
final R1, wR2 [I > 2σ(I)] (all data)
0.0318, 0.0622 (0.0547, 0.0668)
0.0893, 0.2378 (0.1921, 0.3088)
0.0470, 0.1167 (0.0805, 0.1321)
geometry may lead to large anisotropy. D = +25 cm−1 was
reported for an axially compressed CoII complex formed using a
heptadentate lariat ether N,N′-bis(2-aminobenzyl)-1,10-diaza-
15-crown-5.12 To date, no quantitative rationalization has been
undertaken and there have been virtually no attempts to
incorporate such complexes as building blocks for larger
clusters. The primary routes toward seven-coordinate CoII
complexes are via the use of heptadentate ligands, either cyclic
(funtionalized crown ethers)13 or acyclic (tren derivatives).14
However, these are poor candidates for large cluster develop-
ment because of the absence of labile ligands that can be
substituted to form polymetallic complexes. Complexes
possessing pentadentate equatorial ligands and labile axial
ones may be accessed using crown ethers15 or Schiff base type
compounds,16 usually synthesized initially from diacetylpyridine
(with the possibility of further functionalization).
acid hydrazone)]22 and the synthesis and magnetic properties
of similar complexes with different axial ligands. The synthesis
of polymetallic assemblies from the reaction of the seven-
coordinate CoII complex with metallocyanides is also reported,
with an accompanying magnetic investigation.
EXPERIMENTAL SECTION
■
Synthesis. Chemicals were purchased from Aldrich and used
without further purification. DABPH and [Co(DABPH)(H2O)-
(NO3)](NO3) (1) were synthesized according to literature
procedures.22 All solvents were from BDH and were used as received.
All manipulations were conducted under standard benchtop
conditions.
Synthesis of [Co(DABPH)I(H2O)]I·H2O (2). DABPH (0.150 g, 0.375
mmol) was suspended in H2O (20 mL) and the temperature raised to
55 °C. CoI2 (117 mg, 0.375 mmol) dissolved in EtOH (35 mL) was
added dropwise with stirring, forming a pale-orange solution, which
was stirred overnight at room temperature. This was then filtered and
left to slowly evaporate, yielding red crystals after 4 days (0.085 g,
32%). Elem anal. Found (calcd for CoC23H25N5O4I2): C, 36.72
(36.92); H, 3.31 (3.36); N, 9.31 (9.36). ESI+-MS (MeOH): m/z
457.09 ([Co(DABPH)]{H+})+, 585.01 ([Co(DABPH)I]+). IR (KBr,
ν/cm−1): 3461 (m), 3406 (m), 3324 (w), 3287 (w), 3178 (w), 3068
(m), 2918 (w), 1984 (w), 1642 (m), 1622 (s), 1577 (m), 1513 (s),
1481 (s), 1454 (m), 1426 (m), 1376 (m), 1310 (m), 1287 (s), 1178
(s), 1153 (m), 1128 (m), 1097 (w), 1074 (m), 1015 (w), 999 (w), 899
(w), 810 (m), 792 (w), 739 (w), 711 (s), 669 (m), 575 (w), 518 (w),
504 (w), 471 (m), 421 (m).
Synthesis of [Co(DABPH)Br(H2O)]Br·H2O (3). A reaction similar to
that for 2 but with replacement of CoI2 with CoBr2·H2O (0;088 g,
0.375 mmol) yielded brown crystals of 3 in comparable yield (0.100 g,
41%). Elem anal. Found (calcd for CoC23H25N5O4Br2): C, 42.17
(42.12); H, 3.79 (3.85); N, 10.64 (10.70). ESI+-MS (MeOH): m/z
457.09 ([Co(DABPH)]{H+})+. IR (KBr, ν/cm−1): 3453 (m), 3396
(m), 3306 (m), 3195 (w), 3063 (m), 2979 (w), 2897 (m), 2798 (w),
1625 (s), 1578 (m), 1518 (s), 1484 (m), 1455 (m), 1429 (m), 1377
(w), 1313 (m), 1295 (s), 1279, (s), 1178 (s), 1153 (m), 1129 (m),
1097 (w), 1075 (m), 1017 (w), 1000 (w), 899 (w), 812 (m), 794 (w),
743 (w), 713 (s), 672 (m), 560 (w), 518 (w), 520 (w), 466 (m), 427
(m).
Examples of seven-coordinate CoII incorporated into clusters
include the tetrametallic planar, linear array reported by Aromi
et al., consisting of a pair of seven-coordinate CoII ions, each
also bound to a six-coordinate CoII with linear tripyridyl/bis(β-
diketone) ligands; however, antiferromagnetic exchange via the
bridging O atoms results in a S = 0 ground state for the
cluster,17 as well as other bimetallic species for which magnetic
data are not reported.18 There is also one example of a three-
dimensional helical metalloframework consisting of [Cr-
(CN)6]3− units and seven-coordinate CoII that exhibits long-
range ferromagnetic order below 12 K.19 Ferromagnetic
exchange between CoII centers is promoted by bridging
[Cr(CN)6]3− because the unpaired electron density is in
orthogonal orbitals on Cr and Co, respectively, thus interacting
with orthogonal cyanide-based orbitals and resulting in
ferromagnetic exchange.20 This is the case for [Co-
(tmphen)2]3[Cr(CN)6]2, reported by Dunbar et al., displaying
a significant but unmodeled ferromagnetic exchange between
the CoII and CrIII moieties,21 and we will apply this principle to
systems including seven-coordinate CoII.
Herein we report the magnetic properties of the previously
reported seven-coordinate CoII complex [Co(DABPH)(H2O)-
(NO3)](NO3) [1; DABPH = 2,6-diacetylpyridinebis(benzoic
Synthesis of [{Co(DABPH)}3{Cr(CN)6}2(H2O)2]·15H2O (4). K3[Cr-
(CN)6] (0.022 g, 0.067 mmol) dissolved in H2O (15 mL) was added
dropwise to a hot, stirred solution of 1 (0.040 g, 0.067 mmol) in
12046
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
ligand five-coordinates the CoII ion around the equatorial plane,
leaving the axial positions free for coordination by one water
molecule and one nitrate ligand, thus completing the
pentagonal-bipyramidal coordination sphere, with charge
balance provided by the nitrate counterion.
The iodide and bromide analogues of 1 may be synthesized
by the same procedure, upon replacement of Co(NO3)2 by
CoI2 or CoBr2. The products were characterized by MS,
elemental analysis, and IR spectroscopy, and a crystal structure
of 2 was obtained (Figure 2).
MeOH/H2O (4:1, 75 mL). The clear yellow solution was stirred for 1
h at room temperature and then left to slowly evaporate. Yellow
needle-shaped crystals of 4 (0.020 g, 48%) formed after 1 week. Elem
anal. Found (calcd for Co3Cr2C81H97N27O23): C, 46.27 (46.36); H,
4.16 (4.66); N, 17.77 (17.55). IR (KBr, ν/cm−1): 3375 (s), 2127 (w),
1624 (s), 1578 (m), 1527 (s), 1489 (m), 1440 (m), 1379 (m), 1293
(s), 1180 (m), 1135 (w), 1079 (w), 1020 (w), 1002 (w), 901 (w), 811
(m), 714 (m), 455 (m).
Synthesis of [{Co(DABPH)}3{Fe(CN)6}2(H2O)2]·16H2O (5). K3[Fe-
(CN)6] (0.029 g, 0.088 mmol) dissolved in H2O (10 mL) was added
dropwise to a hot, stirred solution of 1 (0.080 g, 0.133 mmol) in
MeOH (20 mL). The mixture was stirred at 55 °C for 20 min, during
which time 5 precipitated as a brown powder (0.040 g, 42%). It was
isolated by filtration, washed with H2O (5 mL) and MeOH (10 mL),
and dried in air. The mother liquors were left to slowly evaporate,
yielding X-ray-quality block-shaped crystals after 1 month. Elem anal.
Found (calcd for Co3Fe2C81H99N27O24): C, 45.76 (45.79); H, 4.66
(4.70); N, 17.90 (17.81). IR (KBr, ν/cm−1): 3400 (s), 2138 (m), 2114
(m), 1626 (s), 1578 (m), 1532 (m), 1489 (m), 1441 (m), 1379 (w),
1294 (s), 1180 (m), 1135 (w), 1079 (w), 1020 (w), 1002 (w), 901
(w), 810 (w), 715 (m), 692 (m), 567 (w), 526 (w).
Physical Measurements. IR data were measured on KBr pellets
using a PerkinElmer FTIR spectrometer. Variable-temperature (300−2
K) magnetic data were measured on powdered samples of 1−5 in an
eicosane matrix in 1.0 and 0.1 T fields using a Quantum Design
MPMS5 SQUID magnetometer. The data were corrected for the
diamagnetic contribution of the sample holder and eicosane, and the
diamagnetism of the sample was estimated according to Pascal’s
constants. Low-temperature (2−6 K) variable-field (0−5.5 T)
measurements were carried out in the same manner. Modeling of
the magnetic susceptibility and magnetization data was performed by
matrix diagonalization methods using MAGPACK software.23
X-ray Diffraction. Single-crystal X-ray diffraction data were
measured on a Bruker APEX II CCD diffractometer. The diffraction
intensities were collected with graphite-monochromatized Mo Kα
radiation. The temperature of the crystal was maintained at the
selected value (100 K) by means of a 700 series Cryostream cooling
device to within an accuracy of ±1 K. Intensity data were corrected for
Lorenz−polarization and absorption factors. Structure solution and
refinement were performed using SHELXS97 and SHELXL97.24 The
structures were refined by direct methods. Refinement of F2 was
against all reflections with anisotropic displacement parameters for all
non-H atoms.
Figure 2. Structure of the cation in 2. H atoms, counterions, and
solvent molecules are omitted for clarity. Color code: Co, cyan; C,
black; N, blue; O, red; I, purple. Thermal ellipsoids are shown at the
50% probability level.
2 crystallizes in the monoclinic P21/c space group and
consists of a pentagonal-bipyramidal cation, in which CoII is
ligated around the equatorial plane by a pentadentate DABPH
ligand, with one iodide and one water molecule occupying the
axial positions (at distances of 2.793 and 2.097 Å from the CoII
ion). Charge balance is provided by an iodide counterion, and
there is one molecule of water in the crystal lattice. Globally,
the structure is similar to that of 1; however, the equatorial
coordination is closer to pentagonal in 2 [ligand bite angles
from 70.78° (for Npyridyl−Co−N) to 74.73° (for O−Co−O),
compared to 69.54−78.67° for the equivalent angles in 1], and
the terminal phenyl groups lie closer to the equatorial plane
(torsion angles of 16.11 and 24.07° compared to 9.89 and
34.27°), which is likely due to the difference in packing due to
the change in the counterion; there are fewer π-stacking and
hydrogen-bonding interactions than in 1, which has a
significant hydrogen-bonding network via both the coordinated
and counterionic nitrate ions and the lattice solvent water
molecules.
RESULTS
■
1 (Figure 1) was synthesized according to literature procedures
and was characterized by mass spectrometry (MS), elemental
The room temperature reaction of 1 with 1 equiv of
K3[Cr(CN)6] in MeOH/H2O yields crystalline 4 (Figure 3)
upon the slow evaporation of the reaction mixture.
4 crystallizes in the monoclinic C2/c space group and
consists of a neutral, pentametallic cluster containing three
{Co(DABPH)}2+ units bridged to one another via two
{Cr(CN)6}3− moieties in a zigzag fashion. The central seven-
coordinate CoII is coordinated equatorially by a DABPH ligand,
slightly more symmetrically than that for monometallic
complex 1 (ligand bite angles ranging from 70.47 to 77.33°,
compared to 69.54−78.67°). The axial positions are ligated by
the cyano N atoms of two crystallographically equivalent
{Cr(CN)6} units (a Co−N bond length of 2.068 Å, slightly
shorter than that reported for six-coordinate Co−NC
bonds21,25); the cyanide bridge is far from linear with respect
to the two metals (a Co−N−C angle of 154.58°). The
geometry around the CrIII atoms is virtually isotropic; the C−N
distances for the cyanide ligands are similar, ranging from 1.149
to 1.176 Å. The crystallographically equivalent, terminal CoII
Figure 1. Previously reported structure of the cation in 1. H atoms and
counterions are omitted for clarity. Color code: Co, cyan; C, black; N,
blue; O, red.
analysis, and IR spectroscopy, and a unit cell was collected to
ensure that the structure was identical with that previously
reported.22
1 crystallizes in the monoclinic P21/n space group, and the
structure was fully described previously. The neutral DABPH
12047
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
{Fe(CN)6}3− moieties in an approximately linear fashion.
Many structural features of 5 are similar to those of 4 (i.e.,
equatorial ligand bite angles in the range of 70.93−75.76° for
the central CoII, approximately isotropic environments around
the Fe ions, terminal water molecules and solvent water
molecules in the crystal lattice, and evidence of intramolecular
π−π interactions between the DABPH ligands of adjacent
clusters), with the major difference being that the bridging iron-
bound cyanides are trans to one another, inferring the overall
linearity of the structure. Neither cyanide bridge is particularly
straight in this case (Co−N−C angles are 155.94 and 153.05°).
Bond-valence-sum analysis shows the oxidation states of the Co
and Fe to be II+ and III+, respectively,26 as is also required for
charge balance.
The facile formation of these discrete {Co3Cr2} and
{Co3Fe2} units can be attributed to their neutrality. There is
no obvious capping ligand present in the reaction mixture so we
postulate that the two components polymerize (facile
substitution of labile water and nitrate ligands of {Co(DABPH}
by cyano N atoms from the anionic [M(CN)6]3−), with this
neutral combination much less soluble than any of the charged
permutations in the polar reaction media, thus crystallizing
preferentially. The global structural difference (zigzag vs linear)
may be due to the different ionic radii of CrIII and FeIII, with
FeIII being smaller with subsequently shorter distances between
the metal ion and the cyano N atoms (≈3.09 Å compared to
≈3.21 Å for CrIII) and steric hindrance subsequently preventing
the {Co(DABPH)} fragments from linking cis to one another.
The IR spectra of 4 and 5 were recorded (see the Supporting
Information). The fingerprint regions of the spectra are similar
to that reported for 1, with minor differences due to the minor
conformational changes in the ligand. CN stretches are
observed at 2127 cm−1 (4) and 2114 cm−1 (5), a frequency
attributed to terminal rather than bridging cyanides in other
reports;21 however, close inspection of the peak for 4 reveals a
shoulder toward higher frequency (as expected for bridging
CN, where the stretching frequency is increased as a result of
kinematic coupling),27 suggesting that the bridging CN stretch
is weaker and subsequently swamped by that of the terminal
CN. For 5, the full peak is observed for the bridging cyanides at
2138 cm−1. It should also be noted that there is no evidence of
CrIII−CN−CoII to CrIII−NC−CoII linkage isomerism (which
would manifest as a CN stretch at ≈2100 cm−1),21 the
occurrence of which can make crystallization of such systems
difficult, and thus CoII compounds of this type are much rarer
than their NiII analogues.
Figure 3. Structure of 4. H atoms and solvent molecules are omitted
for clarity (top), along with DABPH ligands (bottom). Color code:
Co, cyan; Cr, green; C, black; N, blue; O, red. Thermal ellipsoids are
shown at the 50% probability level.
moieties are bound to CrIII-based cyano ligands cis to those
bound to the central CoII, resulting in an anti configuration for
the zigzag pentametallic framework. The cyanide bridges are
more linear than those at the central CoII (a Co−N−C angle of
161.99°) and are slightly longer (a Co−N bond length of 2.093
Å). Again the pentagonal, equatorial planes are ligated by
DABPH (ligand bite angles ranging from 70.32 to 76.69°), with
water ligands occupying the terminal axial positions and
numerous solvent water molecules in the crystal lattice.
There is evidence of intramolecular π−π interactions between
the interwoven DABPH ligands of the terminal CoII ions of
adjacent clusters (separation ≈3.5 Å).
Similarly, the room temperature reaction of 1 with 1 equiv of
K3[Fe(CN)6] in MeOH/H2O yields crystalline 5 (Figure 4)
Although the structure of 1 is published, magnetic data have
not been reported for the complex. The measured molar
magnetic susceptibilities (χ and χT) of 1 are plotted as a
function of the temperature in Figure 5. The experiment was
performed at 0.1 T (2−150 K) and 1.0 T (5−300 K) to assess
any field dependence; none was observed in these ranges (as
was the case for all five complexes investigated), and the 1.0 T
data, for the larger temperature range, are plotted, overlaid with
the 0.1 T data from 2 to 30 K (as is the case throughout). Also
plotted is the measured magnetization as a function of the field.
χT is ca. 2.26 emu K mol−1 at 300 K (larger than expected for
the spin-only formula for a high-spin mononuclear CoII
complex, indicating a relevant orbital contribution) and remains
constant down to 100 K, below which it collapses, reaching a
value of 1.45 emu K mol−1 at 2 K, indicative of significant ZFS
in the orbitally nondegenerate S = 3/2 ground state. Similarly, at
2 K and 5.5 T, the observed magnetization value of 2.07 μB is
Figure 4. Structure of 5. H atoms and solvent molecules are omitted
for clarity. Color code: Co, cyan; Fe, brown; C, black; N, blue; O, red.
Thermal ellipsoids are shown at the 50% probability level.
upon the slow evaporation of the reaction mixture. Elemental
analysis of the precipitate corresponds to that expected from
the crystal structure, and the IR spectra of the two materials are
identical (showing both bridging and terminal CN groups at
identical frequencies).
5 crystallizes in the monoclinic C2/c space group and
consists of a neutral, pentametallic cluster containing three
{Co(DABPH)}2+ units bridged to one another via two
12048
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
Figure 5. χT vs T (left) and χ vs T (inset) for 1 and M vs H (right) collected at 2, 4, and 6 K. Data were modeled using MAGPACK (solid line); see
the text for parameters.
Figure 6. χT vs T (left) and χ vs T (inset) for 2 and M vs H (right) collected at 2, 4, and 6 K. Data were modeled using MAGPACK (solid line); see
the text for parameters.
well below the theoretical saturation for an S = 3/2 system (Msat
= 3.3 for g = 2.2), as expected when appreciable magnetic
anisotropy is inherent. The data may be simulated with
MAGPACK using Hamiltonian equation (1) and yielding the
parameters g = 2.21, D = 32.0 cm−1, and E = 0.3 cm−1 for the
best fit.
fit is performed for more than one temperature, affording us
some certainty in the sign of D from these simulations.28 Slow
relaxation of magnetization has recently been observed for a
monometallic first-row transition-metal complex (in which an
Ising-type axial anisotropy was present);29 however, given that
in our case D is positive, it is unlikely that 1 will display slow
relaxation of magnetization (as the Ms = ±1/2 states of the S =
3
/
ground state are stabilized). This anisotropic fragment
2
could, however, be used to design larger SMMs provided the
local ZFS tensors could be aligned to give rise to a molecular
Ising-type anisotropy.30 Additionally, the use of CoII as a
building block gives rise to the possibility of synthesizing
molecules with noninteger ground states (Kramers doublets).
In such cases, the Ms = ±x/2 terms do not mix, independent of
the magnitude of E, so quantum tunneling (one of the primary
routes for magnetic relaxation in SMMs) may be eliminated.
2 (Figure 6) and 3 (Supporting Information) exhibit similar
magnetic behavior, which can also be modeled using
Hamiltonian equation (1), with the parameters g = 2.18, D =
30.0 cm−1, and E = 0.3 cm−1 and g = 2.24, D = 30.0 cm−1, and
E = 0.3 cm−1, respectively. One would perhaps expect the
magnitude of the anisotropy to show greater dependence on
the nature of the ligands belonging to the coordination sphere,
as has already been reported.31−33 This may be explained by
considering the origin of the anisotropy, the interaction
between the ground and first excited states of molecule. For
the present case, the first excited state likely involves the
electronic transition from the lowest-energy degenerate xz, yz
(1)
The fitting was performed simultaneously for the four
experimental measurements (i.e., for the magnetization data at
2, 4, and 6 K and for the susceptibility data over the full
temperature range), and to reduce the variable parameters
during the fitting procedure, the g tensor was assumed to be
isotropic. The g value was fixed according to the high-
temperature χT value, with D and E then fit to the low-
temperature regime and the magnetization data. The values
obtained for g and D are consistent with those reported for
other seven-coordinate CoII complexes.12 Initial attempts to
model the data assumed purely axial ZFS (E = 0), but it was
necessary to invoke the small rhombic term to ensure a good
fit. In many cases, it is not possible to determine the sign of D
from the magnetic data alone because there will be a range of
values for D and E/D that yield equally adequate results;
however, this is less likely when D is large, E/D is small, and the
12049
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
Figure 7. χT vs T (left) and χ vs T (inset) for 4 and M vs H (right) collected at 2, 4, and 6 K. Data were modeled using MAGPACK (solid lines); see
the text for parameters.
orbitals to the xy, x2 − y2 set. Because the axial ligand primarily
interacts with the z2 orbital, it should have little effect on the
energy of this transition, and thus on the anisotropy (to which
it is inversely proportional). The minor difference could be due
to the greater π character of the interaction with the heavy
halides, which will affect the xz, yz set, although likely to a
lesser extent, potentially changing the energy of the transition.
Similar effects have been demonstrated in the case of five-
coordinate NiII complexes.34 Ab initio calculations are under-
way to rationalize the origin of the very large anisotropy found
in the seven-coordinate CoII complexes and the influence of the
axial ligands on their magnitude.35 At any rate, our experimental
results show that the seven-coordinate CoII moiety should
possess a magnetic anisotropy largely unaffected by the nature
of the axial ligands, making it a valuable building block for
introducing significant local magnetic anisotropy into poly-
nuclear species. Similar arguments may hold for other metals
(e.g., FeII), which may exhibit Ising-type anisotropies that could
again be inserted into larger systems; such work is currently
underway.
The measured molar magnetic susceptibilities (χ and χT) of
4 are plotted as a function of the temperature in Figure 7. Also
plotted is the measured magnetization as a function of the field.
Upon cooling from room temperature, χT increases from 11.81
emu K mol−1 (slightly higher than expected for three
uncoupled CoII and two CrIII ions, 10.56 emu K mol−1,
assuming gCo = 2.20 and gCr = 2.00), reaches a maximum at 4 K
(24.27 emu K mol−1), and then decreases. This behavior
implies ferromagnetic exchange between the metallic centers
(as expected for pseudolinear CrIII−CN−CoII bridges),36
although the maximum is well below that expected for an
isolated S = 15/2 ground state (≈38.5 emu K mol−1). This is
likely due to significant ZFS within the ground state and/or the
proximity of low-lying excited states. Intermolecular antiferro-
magnetic exchange may also be present and would cause similar
effects, and all of these factors would account for the collapse in
χT below 4 K.
would contribute very little to the pseudolinear cyanide bridge
along the Z axis of the CoII. As such, a large ferromagnetic
exchange coupling is expected because of interaction between
orthogonal orbitals on the Co (z2) and Cr (xz and yz).
However, because the Co−N−C angles are smaller than 180°,
the strict orthogonality criteria are not fulfilled and an
antiferromagnetic contribution to the interaction will be
present, which weakens the overall ferromagnetic exchange
coupling.
A full simulation of the data using MAGPACK was
impractical given the likelihood of overparametrization and
the incapacity of the program to model relative anisotropic
vectors (which will clearly play a defining role in determining
the observed magnetic behavior in this case, close to the weak
exchange limit); however, a reasonable fit of the susceptibility
data can be obtained to 30 K using ISOTROPIC MAGPACK
(neglecting local anisotropies), providing some insight into the
strength of the intramolecular exchange interactions. An
adequate fit was obtained with a simplistic one J model
(assuming each CrIII−CN−CoII bridge to be equal and no
other exchange interactions to be in operation), using
Hamiltonian equation (2), yielding parameters gaverage = 2.14
and J = 2.7 cm−1.
(2)
It is unlikely that the two sets of cyanide bridges actually
present the same magnitude of exchange (given the different
Co−N−C angles), but the one J model provides an adequate fit
(in agreement with other reported Cr−CN−Co exchanges)19
given the limited data that is modeled. Furthermore, this value
can at best be referred to as an “effective” exchange because the
fit is to the net result of the two competitive factors affecting
the magnitude of χT at a given temperature: (1) the
ferromagnetic exchange between metal ions and (2) the
significant local anisotropy, observed to have a large effect up
to 100 K in the monomers; thus, the absolute value of J is likely
somewhat larger. This set of parameters obviously did not
provide a good fit for the magnetization versus field data, the
low-temperature, high-field experimental regime that is
dominated by the effects of local, and subsequently molecular,
anisotropy. The magnetization value at 2 K and 5.5 T of 11.8
μB is below that expected for a S = 15/2 ground state (≈16 μB
assuming g = 2.14) and has not reached a plateau, implying an
Assuming an idealized C5v symmetry for the Co moieties
(although the CoII local geometries deviate slightly from the
idealized one), the degenerancy of the 5d orbitals is lifted to
three energy levels e1(xz,yz), e2(xy,x2−y2), and a1(z2). The z2
orbital is likely to remain the highest single occupied molecular
orbital (SOMO), and the other two SOMOs, xy and x2 − y2,
12050
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
anisotropic and/or poorly isolated ground state. We initially
considered modeling the data with a giant spin, S = 15/2;
however, this was deemed inappropriate. Even using
eigenvalues extracted from the isotropic model, considering
only the exchange, the S = 15/2 ground state is separated from
the first S = 13/2 excited state by only 3.1 cm−1, and the
introduction of significant positive local anisotropies at the CoII
sites will further reduce this separation. Thus, it is highly
unlikely that there is one state uniquely populated even in the
low-temperature, high-field regime. Future work will include a
more thorough investigation of this situation by our theoretical
collaborators.
Given the large anisotropic ground state of 4, preliminary
alternating-current measurements were performed; however,
because we believe the ZFS to be positive, it was an unlikely
candidate for SMM behavior, and the effect was not observed.
There was a slight increase in the out-of-phase susceptibility
component toward 1.8 K; however, no maximum was observed,
and furthermore there was a 2 order difference in magnitude
between the in-phase and out-of-phase components.
(below 15 K), the lowest-energy Kramers doublet (effective
spin S′ = /2) may be uniquely populated and the increase in
1
χT is assumed to be due to the ferromagnetic exchange
1
1
between the effective S′ = /2 CoII ions and the S = /2 FeIII
ions, which becomes dominant at low temperature. Similar
behavior has been reported for ferromagnetically coupled CoII
macrocycles, albeit concerning octahedral CoII species.37
The magnetization value at 2 K and 5.5 T (7.72 μB) is below
that expected for a well-isolated S = 11/2 ground state and is
indicative of substantial magnetic anisotropy. Given the
limitations in our attempts to model the data for 4 and the
range of parameters that could be governing the behavior of our
relatively limited data set, attempts were not made to model the
data at this time. The very different behaviors observed for 4
and 5 likely due to the different arrangements of the local
magnetic axes demand further investigation; attempts to
synthesize the [{Co(DABPH)}3{CoIII(CN)6}2(H2O)2] cluster
are underway because the diamagnetic bridging CoIII moieties
would negate the effects of magnetic exchange and may shed
light on the relationship between the local and molecular
anisotropies and the local geometry.
A plot of χT as a function of the temperature for 5 is shown
in Figure 8. Also plotted is the measured magnetization as a
CONCLUSION
■
In summary, seven-coordinate [Co(DABPH)X2] is a very
3
anisotropic S = /2 molecular fragment, with two labile axial
ligands and the local anisotropy largely independent of the
character of those ligands. This can be incorporated into larger
clusters and may be used to promote large molecular
anisotropies. A thorough investigation of other less common
geometries among first-row transition metals (i.e., five- or
seven-coordinate) and their use as building blocks is also
underway, particularly NiII and FeII, which should possess huge,
and not necessarily positive, local anisotropies due to near-
degeneracy effects, affording us a greater range of tools in the
search for greater anisotropy in high-spin clusters.
ASSOCIATED CONTENT
■
S
* Supporting Information
IR spectra of compounds 4 and 5 and plots of magnetic
susceptibility versus temperature and magnetization versus field
for compound 3 including MAGPACK simulation. This
material is available free of charge via the Internet at http://
pubs.acs.org. CCDC 835139, 835140, and 835141 contain the
supplementary crystallographic data for this paper. These data
can be obtained free of charge from the Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
Figure 8. χT vs T for 5 and M vs H (inset) collected at 2, 4, and 6 K.
function of the field. Upon cooling from room temperature, χT
gradually increases from 7.95 emu K mol−1 (slightly higher than
expected for three uncoupled CoII and two FeIII ions, 7.56 emu
K mol−1, assuming gCo = 2.20 and gFe = 2.00), reaches a
maximum at approximately 85 K (8.23 emu K mol−1), and then
decreases to a minimum at 15 K (7.68 emu K mol−1) before
finally increasing toward its second maximum at 4 K (8.09 emu
K mol−1). This behavior may be rationalized as follows. The
increase between 300 and 85 K is the net result of the opposing
effects of ferromagnetic exchange between metal centers and
large local anisotropy at the metal centers. The “effective”
exchange appears weaker than that in 4, which may be due to
weaker absolute exchange between metal ions, due to
differences in the bridging angles for the cyanides, or it may
be due to the greater combined effects of the local anisotropies
of the CoII ions due to the colinear arrangement of the
magnetic axes compared to the orthogonal arrangement in 4.
The decrease between 85 and 15 K reflects the greater effect of
the local anisotropies on χT at temperatures below 100 K, as
observed for the monomers 1−3. At very low temperatures
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: luke.batchelor@u-psud.fr (L.J.B.), talal.mallah@u-
psud.fr (T.M.). Tel: (+33) (0)1 69 15 47 49. Fax: (+33) (0)
1 69 15 47 54.
REFERENCES
■
(1) Sessoli, R.; Tsai, H. L.; Schake, A. R.; Wang, S.; Vincent, J. B.;
Folting, K.; Gatteschi, D.; Christou, G.; Hendrickson, D. N. J. Am.
Chem. Soc. 1993, 115, 1804.
́
(2) Aromi, G.; Brechin, E. Struct. Bonding (Berlin) 2006, 122, 1.
(3) Troiani, F.; Ghirri, A.; Affronte, M.; Carretta, S.; Santini, P.;
Amoretti, G.; Piligkos, S.; Timco, G.; Winpenny, R. E. P. Phys. Rev.
Lett. 2005, 94, 207208.
12051
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052

Inorganic Chemistry
Article
(4) Larionova, J.; Gross, M.; Pilkington, M.; Andres, H.; Stoeckli-
Evans, H.; Gudel, H. U.; Decurtins, S. Angew. Chem., Int. Ed. 2000, 39,
̈
1605.
(5) Murugesu, M.; Habrych, M.; Wernsdorfer, W.; Abboud, K. A.;
Christou, G. J. Am. Chem. Soc. 2004, 126, 4766.
(6) Barra, A.-L.; Brunel, L.-C.; Gatteschi, D.; Pardi, L.; Sessoli, R. Acc.
Chem. Res. 1998, 31, 460.
(7) Murrie, M. Chem. Soc. Rev. 2011, 39, 1986.
(8) Boca, R. Coord. Chem. Rev. 2004, 248, 757.
(9) En-Che, Y.; David, N. H.; Wolfgang, W.; Motohiro, N.; Lev, N.
Z.; Roger, D. S.; Arnold, L. R.; Marisol, L.-G.; George, C. J. Appl. Phys.
2002, 91, 7382.
(10) Langley, S. J.; Helliwell, M.; Sessoli, R.; Rosa, P.; Wernsdorfer,
W.; Winpenny, R. E. P. Chem. Commun. 2005, 5029.
(11) Yoshihara, D.; Karasawa, S.; Koga, N. J. Am. Chem. Soc. 2008,
130, 10460.
(12) Platas-Iglesias, C.; Vaiana, L.; Esteban-Gomez, D.; Avecilla, F.;
Real, J. A.; de Blas, A.; Rodriguez-Blas, T. Inorg. Chem. 2005, 44, 9704.
(13) Mathieu, F.; Weiss, R. J. Chem. Soc., Chem. Commun. 1973,
816a.
(14) Gou, S.; You, X.; Yu, K.; Lu, J. Inorg. Chem. 1993, 32, 1883.
(15) Fewings, K. R.; Junk, P. C.; Georganopoulou, D.; Prince, P. D.;
Steed, J. W. Polyhedron 2001, 20, 643.
(16) Andelkovic, K.; Ivanovic, I.; Prelesnik, B. V.; Leovac, V. M.;
Poleti, D. Polyhedron 1996, 15, 4361.
(17) Barrios, L. A.; Aguila, D.; Roubeau, O.; Murray, K. S.; Aromi, G.
Aust. J. Chem. 2009, 62, 1130.
(18) Rajput, L.; Biradha, K. Cryst. Growth Des. 2007, 7, 2376.
(19) Zhang, Y.-Z.; Sato, O. Inorg. Chem. 2009, 49, 1271.
(20) Beltran, L. M. C.; Long, J. R. Acc. Chem. Res. 2005, 38, 325.
(21) Shatruk, M.; Chambers, K. E.; Prosvirin, A. V.; Dunbar, K. R.
Inorg. Chem. 2007, 46, 5155.
(22) Giordano, T. J.; Palenik, G. J.; Palenik, R. C.; Sullivan, D. A.
Inorg. Chem. 1979, 18, 2445.
́
(23) Borras-Almenar, J. J.; Clemente-Juan, J. M.; Coronado, E.;
Tsukerblat, B. S. J. Comput. Chem. 2001, 22, 985.
(24) Sheldrick, G. M. SHELX-97; Universitat Gottingen: Gottingen,
Germany, 1997.
(25) Shiga, T.; Okawa, H.; Kitagawa, S.; Ohba, M. J. Am. Chem. Soc.
2006, 128, 16426.
(26) Brese, N. E.; O’Keeffe, M. Acta Crystallogr. 1991, B47, 192.
(27) Bignozzi, C. A.; Argazzi, R.; Schoonover, J. R.; Gordon, K. C.;
Dyer, R. B.; Scandola, F. Inorg. Chem. 1992, 31, 5260.
(28) Rogez, G. Ph.D. Thesis, Universite
2002.
́
Paris Sud XI, Orsay, France,
(29) Freedman, D. E.; Harman, W. H.; Harris, T. D.; Long, G. J.;
Chang, C. J.; Long, J. R. J. Am. Chem. Soc. 2011, 132, 1224.
(30) Rebilly, J.-N.; Mallah, T. Struct. Bonding (Berlin) 2006, 122, 103.
(31) Desrochers, P. J.; Telser, J.; Zvyagin, S. A.; Ozarowski, A.;
Krzystek, J.; Vicic, D. A. Inorg. Chem. 2006, 45, 8930.
(32) Duboc, C.; Phoeung, T.; Zein, S.; Pecaut, J.; Collomb, M. N.;
Neese, F. Inorg. Chem. 2007, 46, 4905.
(33) Long, J. R.; Karunadasa, H. I.; Arquero, K. D.; Berben, L. A.
Inorg. Chem. 2010, 49, 4738.
̀
(34) Rebilly, J.-N.; Charron, G.; Riviere, E.; Guillot, R.; Barra, A.-L.;
Serrano, M.-D.; van Slageren, J.; Mallah, T. Chem.Eur. J. 2008, 14,
1169.
(35) Ruamps, R.; Maurice, R.; Guihery, N.; Gogoi, N., ; Sutter, J.-P.;
Batchelor, L. J.; Mallah, T. Manuscript in preparation. 2011.
(36) Mallah, T.; Auberger, C.; Verdaguer, M.; Veillet, P. J. Chem. Soc.,
Chem. Commun. 1995, 61.
(37) Moragues-Canovas, M.; Talbot-Eeckelaers, C. E.; Catala, L.;
Lloret, F.; Wernsdorfer, W.; Brechin, E. K.; Mallah, T. Inorg. Chem.
2006, 45, 7038.
12052
dx.doi.org/10.1021/ic201534e|Inorg. Chem. 2011, 50, 12045−12052