The Zn polymorphic analogues of the [Fe(PM-PEA)2(NCS)2] spin-transition compound

Article abstract of DOI:10.1016/j.ica.2008.03.011

We have studied the peculiarities of complex [Zn(PM-PEA)₂(NCS)₂], which presents large void cavities. As powder it does not show solvent inclusion, while as single crystals we have evidenced the existence of various polymorphs, with

Full text of DOI:10.1016/j.ica.2008.03.011

Inorganica Chimica Acta 361 (2008) 3519–3524
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
The Zn polymorphic analogues of the [Fe(PM–PEA)2(NCS)2]
spin-transition compound
*
Frédéric Le Gac, Philippe Guionneau, Jean-Francßois Létard, Patrick Rosa
ICMCB, CNRS, Université Bordeaux 1, 87, Avenue du Dr. A. Schweitzer, 33608 Pessac Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
We have studied the peculiarities of complex [Zn(PM–PEA)2(NCS)2], which presents large void cavities. As
powder it does not show solvent inclusion, while as single crystals we have evidenced the existence of
various polymorphs, with and without solvent inclusion. These results are preliminary results for the
comprehension of the behaviour of mixed [FexZn1ꢀx(PM–PEA)2(NCS)2], where Zn has been used to
decrease the cooperative interactions between Fe atoms and check the effects on the wide hysteresis
reported for this compound.
Received 18 February 2008
Received in revised form 27 February 2008
Accepted 3 March 2008
Available online 10 March 2008
Dedicated to Dante Gatteschi.
Ó 2008 Elsevier B.V. All rights reserved.
Keywords:
X-ray crystal structures
Polymorphism
Spin crossover
1. Introduction
tal packing effects on magnetic properties, through solid dilution,
in order to see effects not seen with pure compounds (see e.g.,
In light of the evergoing interest in new materials for photonics
or information storage, the study of spin-crossover compounds has
undergone in recent years a strong acceleration. This is due to re-
cent results that are very promising, either with pulsed laser irra-
diation which has shown the possibility of photoconversion within
the hysteresis [1], that is at higher temperatures than previously
seen with the light-induced excited spin-state trapping (LIESST) ef-
fect, or the creativity of molecular chemists in combining the li-
gand-driven light-induced spin crossover (LD-LISC) effect with
the remarkable properties of the diarylethene framework devel-
oped by Irie [2], or also the studies of cascading effects performed
with ultrafast techniques [3]. But potential applications still need
to conform to practical considerations, and one of the strongest
is the temperature range of the application. For information stor-
age applications, bistability, that is the presence of a hysteresis,
must not only be centred at or near room temperature, but its
width should be of no less than 80 K. But the presence of hysteretic
behaviour depends mainly upon packing effects in the crystalline
media, the so-called cooperativity being assured through hydrogen
bonds, p-stacking or Van der Waals contacts, which are not usually
easy to understand, and even less to control.
the seminal work of Hauser et al. on diluted LS species) [4] or to
have structural data necessary to distinguish between thermal ef-
fects and spin crossover [5–7].
The so-called PM–L family of ligands has been extensively stud-
ied in our laboratory over the last twelve years, with much informa-
tion from a variety of techniques. Out of the [Fe(PM–L)2(NCS)2]
complexes synthesized so far, two ligands have shown to be of par-
ticular interest (see Scheme 1). (1) PM–BiA has been shown to crys-
tallize in two different phases; phase I in the orthorhombic system
shows a particularly abrupt transition with a small hysteresis
(T1/2; = 168 K and T1/2" = 173 K) which is not accompanied by a crys-
tallographic phase transition, and phase II in the monoclinic system
shows a more gradual spin transition (T1/2 = 198 K), with both
phases showing interesting photomagnetic behaviour [8]. (2)
PM–PEA has been shown to give rise to a very large hysteresis of
37–60 K, which is concomitant with a quite severe crystallographic
phase transition from monoclinic to orthorhombic, but shows only a
weak LIESST effect at low temperatures [9].
Of the possible doping cations mentioned earlier, Zn(II) is the
most interesting one since it is diamagnetic, and it has an ionic ra-
dius at 0.74 Å which is intermediary between Fe(II) in the low-spin
state (0.61 Å) and in the high-spin state (0.78 Å) [10], which should
greatly facilitate its insertion in the crystalline matrix. In the
course of an ongoing study on the effects of solid dilution of
spin-crossover complexes with non-active metals, and in line with
previous results with Co(II) [5], we have undertaken recently to
In this sense, the use of other cations (diamagnetic Zn2+ or para-
magnetic Co2+, Mn2+, Ni2+, Cd2+) is a precious tool to study the crys-
* Corresponding author. Fax: +33 540002649.
E-mail address: rosa@icmcb-bordeaux.cnrs.fr (P. Rosa).
0020-1693/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2008.03.011

3520
F. Le Gac et al. / Inorganica Chimica Acta 361 (2008) 3519–3524
C
42H28N6S2Zn (746.23 g molꢀ1): C, 67.60; H, 3.78; N, 11.26; S,
8.59; Zn, 8.76. Found: C, 66.98; H, 3.75; N, 11.11; S, 8.52 (8.42
for ICP); Zn, 8.89%.
Thermogravimetric analysis was performed under argon on a
Setaram TAG2400 (symmetrical thermobalance coupled with dif-
ferential thermoanalysis), coupled with a mass spectrometer Ther-
mostar BALZERS.
Scheme 1.
2.2. X-ray powder diffraction and crystal structure determination
dilute the complex of PM–PEA with Zn(II) in order to check the
effects on the cooperativity of the spin transition and the photo-
magnetic behaviour [11]. We will present here corresponding crys-
tallographic results showing the existence of various polymorphic
forms of the [Zn(PM–PEA)2(NCS)2] complex.
Powder X-ray diffraction data were recorded using a PANalyti-
cal X’Pert MPD diffractometer with Bragg–Brentano geometry, Cu
Ka radiation and a backscattering graphite 370 monochromator.
Yellow well-formed single crystals can be obtained in a few
days by layering a freshly prepared solution of Zn(NCS)2 in metha-
nol over a solution of PM–PEA in either dichloromethane or a mix-
ture of methanol and THF, in a test tube.
2. Experimental
Single crystals were mounted on a Bruker-Nonius j-CCD dif-
fractometer, Mo Ka radiation (0.71073 Å). Data collection was per-
formed at either 293 K or 140 K, using mixed / and x scans. The
structural determination by direct methods and the refinement
of atomic parameters based on full-matrix least squares on F2 were
performed using the SHELX-97 [12] programs within the WINGX pack-
age [13]. Refinement details are given in Table 1. Hydrogen atoms
were geometrically positioned.
2.1. Materials
Synthesis was performed with usual techniques without any
special care to exclude oxygen or moisture. Nevertheless, solvents
were used freshly distilled, methanol over Mg and CH2Cl2 over
CaH2. PM–PEA = N-(20-pyridylmethylene)-4-(phenylethynyl)ani-
line was prepared as described earlier [9a]. The Zn compound
was synthesized according to the same procedure reported for
the parent Fe compound [9a]. Zn(H2O)7 ꢁ SO4 (58 mg, 0.2 mmol)
and potassium thiocyanate (39 mg, 0.4 mmol) were dissolved in
methanol (10 mL). The resulting solution was stirred for 30 min.
It was then filtered slowly over a solution of PM–PEA (113 mg,
0.4 mmol) in dichloromethane (3 mL). The yellow precipitate that
forms very quickly is filtered, rinsed with diethylether, and left
to dry on the bench. Yield: 130 mg, 87%.
3. Results and discussion
The synthesis of [Zn(PM–PEA)2(NCS)2] is straightforward. The
yellow powder obtained in excellent yield was confirmed by ele-
mental analysis to have the expected composition. The X-ray pow-
der diffraction spectrum as reported in Fig. 1 shows that the solid is
crystalline. Enlarged poorly resolved peaks point towards a precip-
itation that was probably too quick. Comparison with the powder
diffraction pattern calculated from the crystal structure of
[Fe(PM–PEA)2(NCS)2], at room temperature in the monoclinic
C, H, N and S elemental analysis was performed with a Thermo-
Fischer Flash EA1112 microanalyzer. Zn and S elemental analysis
was performed with
a Varian ES-720 ICP-OES. Calc. for
Table 1
Crystal and experimental data for [M(PM–L)2(NCS)2] ꢁ X complexes, with (M = Zn, L = PEA, X = Ø, MeOH, MeOH ꢁ H2O, this work), (M = Fe, L=PEA) [9a], (M = Co, L = PEA, X = MeOH)
[5], (M = Fe, L=Bia) [18b]
[M(PM–L)2(NCS)2] ꢁ X
L:
PEA
BiA
Fe
M:
Zn
Fe
Co
X:
MeOH
Phase I
Ø
MeOH H2O
Ø
MeOH
Phase I
Ø
Phase:
Phase II
Phase I
Phase II
Formula
Crystallization solvent
T (K)
ZnC43H32N6S2O
CH2Cl2/MeOH
140
ZnC42H28N6S2
CH2Cl2/MeOH
293
ZnC43H34N6S2O2
CH2Cl2/MeOH
293
FeC42H28N6S2
MeOH
293
CoC43H32N6S2O
MeOH
293
FeC38H28N6S2
CH2Cl2/MeOH
293
140
Crystal dimensions (mm)
Crystal description
System
Space group
Z
0.1 ꢂ 0.1 ꢂ 0.1
yellow prism
monoclinic
P21/c
0.2 ꢂ 0.1 ꢂ 0.1
yellow prism
monoclinic
P21/c
0.3 ꢂ 0.2 ꢂ 0.1
yellow prism
monoclinic
P21/c
monoclinic
P21/c
4
monoclinic
P21/c
monoclinic
P21/c
monoclinic
P21/c
4
4
4
4
4
a (Å)
b (Å)
c (Å)
15.721(3)
14.417(5)
16.855(5)
93.93(2)
3811.3(2)
0.796
17.062(8)
14.824(8)
17.156(9)
117.71(3)
3841.5(36)
0.785
16.920(5)
14.824(5)
16.939(4)
117.64(2)
3763.7(21)
0.802
17.326(7)
14.711(6)
17.817(5)
117.59(2)
4024.8(28)
0.750
15.637(1)
14.566(8)
16.821(1)
92.95(4)
3826.2(1)
15.599(5)
14.727(5)
16.870(5)
93.65(5)
3867.6(2)
17.570(5)
12.602(5)
17.358(5)
115.68(1)
3464(2)
b (°)
V (Å3)
l (mmꢀ1
)
Independent reflections
8675
0.035
9092
0.050
8955
0.035
9133
0.041
Rint(F2)
Observed (I > 2r(I))
Parameters/restraints
Robs/Rall
7390
490/5
0.038/0.049
0.102
5005
460/0
0.050/0.111
0.144
7212
460/0
0.036/0.051
0.090
5585
487/0
0.073/0.118
0.274
wR2 all(F2)
Goodness-of-fit on Fall
qmin/qmax
1.068
1.028
1.093
1.034
ꢀ0.522/0.411
ꢀ0.406/0.295
ꢀ0.422/0.273
ꢀ0.502/1.328
Cell void (Å3) (ratio in % of cell volume)
0
245.9(6.4)
235.0(6.2)
0
228.7(6)
0
33.3(1)

F. Le Gac et al. / Inorganica Chimica Acta 361 (2008) 3519–3524
3521
solvent in the solids we are synthesizing, and elemental analysis
by itself is not enough to rule out the presence of some traces of
solvent.1 We completed the check of the eventual presence of sol-
vent inclusions in the Zn complex powder with a thermogravimetric
analysis. Loss of mass is negligible up to 125 °C, from 125 to 250 °C it
amounts to only 0.5%, excluding thus any systematic inclusion of
water, methanol or dichloromethane in the voids of the lattice. This
is quite surprising given the large void space available. It can be
noted that the situation is quite different for the PM–BiA ligand,
the other ligand giving hysteretic behaviour, where a more efficient
packing leaves no space for solvent inclusion.
Single crystals were obtained by diffusion of a methanol solu-
tion of Zn(NCS)2 in a solution of the ligand in either dichlorometh-
ane or methanol/THF (the ligand is not very soluble in methanol
alone). With methanol/THF as solvent, the small needle-like yellow
crystals that formed after a few days proved upon the resolution of
the structure to be isomorphous with the iron complex structure,
but this structure will be commented in more detail in a forthcom-
ing publication [11].
With dichloromethane, huge perfect yellow elongated prisms
grew in a few days. Out of those crystals, not less than three differ-
ent crystalline forms of [Zn(PM–PEA)2(NCS)2] could be identified
and characterized, two with solvent inclusion and one without,
see Table 1.
All the three forms crystallize in the monoclinic setting, and the
structures could be solved in the same space group than the Fe(II)
compound, P21/c. But only one form proves to be isomorphic to
Fe(II), while the two others show a very different monoclinic angle
b around 118°. The analogy with the phase II of PM–BiA that we
mentioned earlier is immediate, see Table 1, and we will call those
new forms in the same way. By extension, the Fe(II) structure and
related ones will be labelled as phase I.
On the one hand, the methanol solvated phase I polymorph is
essentially identical to the non-solvated Fe(II) complex. In fact
the methanol occupies the void space previously identified, with-
out disturbing the network of close contacts (hydrogen bonds, C–
Sꢀ contacts, p-stacking, etc.) that have been previously described.
[5,8,18] The available OH enters in close hydrogen-bonding with
a neighbouring Sꢀ of another complex thiocyanate, without dis-
turbing the other contacts of this thiocyanate.
Fig. 1. Comparison of experimental X-ray powder diffraction pattern for Zn(PM–
PEA)2(NCS)2] (in black) and the pattern of the various phases calculated from the
single crystal atomic positions of (a) [Fe(PM–PEA)2(NCS)2] at 293 K (dark grey) and
[Zn(PM–PEA)2(NCS)2] ꢁ MeOH phase I at 293 K (light grey); (b) [Zn(PM–PEA)2(NCS)2]
phase II at 293 K (dark grey) and [Zn(PM–PEA)2(NCS)2] ꢁ MeOH ꢁ H2O phase II at
293 K (light grey). 2h was restricted to 5–24° where differences are more apparent,
and spectra were all normalized.
On the other hand, the two other phase II polymorphs, the one
solvated with methanol and water molecules included in the cell,
the other unsolvated, differ quite markedly from the original Fe(II)
complex. In Fig. 3, the molecular structures of those three com-
plexes superposed are represented.2 The differences are striking
both for the NCS arms and for the aromatic arms. This time, the dif-
ference is marked also between the solvated and the non-solvated
form of the Zn complex (Fig. 3b). Here, the solvent does not simply
fill the void space that can be identified (62 Å3 at 293 K to 59 Å3 at
140 K) in the non-solvated species (Fig. 4). Nevertheless, the metal
intramolecular environment is similar for all the three complexes,
as can be seen in Table 2. The only significative difference regards
the metal–imine bonds which are slightly longer in the case of Zn(II).
This is not unexpected, since Zn(II) cannot stabilize the ligands by
metal-to-ligand charge transfer, due to its too high third ionization
potential, and imines with their low-lying p* system will be particu-
larly affected, which explains the observed elongation. Angles be-
tween the two thiocyanates and the pyridylimines are also very
similar. For the non-solvated Zn complex, the Zn–NCS angles are
161° and 163°, while for the solvated complex these angles are
165.4° and 166.8°. The difference, visible also in Fig. 3b, is mainly
system, shows an excellent correspondence. The Zn compound is
isomorphic with the Fe compound as was expected. This is for
molecular compounds the one principal condition to have solid
solutions of one complex into a matrix of another complex, the
other conditions of size, valency or electronegativity [14] were al-
ready fulfilled with the choice of Zn(II). This result is in line with
our previously reported study of Co(II) complexes [5], where the
Co(II) complex of PM–PEA was shown to be isomorphic with the
Fe(II) compound, but where it was also shown that it could co-
crystallize with a molecule of methanol.
And indeed, if one calculates void spaces existing within the
crystalline cell [15], starting from the structures of [Fe(PM–
PEA)2(NCS)2] and [Co(PM–PEA)2(NCS)2] at room temperature, then
one finds a void space of 57 Å3 (60 Å3) near each complex that is
available for solvent inclusion, see Fig. 2. This void is large enough
to contain at least one water molecule, or even methanol. Interest-
ingly, with the change of symmetry occurring with the spin transi-
tion, the void found in the orthorhombic phase decreases down to
46 Å3 per complex. That means that the effective inclusion of a
molecule of solvent would have to interfere with the spin crossover
of the complex since at low temperature there would not be en-
ough space anymore for the solvent. And this is of course without
even considering possible hydrogen-bonding or other contacts
which are known to modulate spin crossover behaviour [16,17].
It is thus important to check carefully the presence or absence of
1
In fact calculated values for the Zn complex with 0.4H2O or 0.4MeOH give a better
fit to the experimental values.
2
Superposition was performed with Accelrys DS Visualizer 2.0. The coordination
sphere ZnN6 was used as framework for the superposition.

3522
F. Le Gac et al. / Inorganica Chimica Acta 361 (2008) 3519–3524
Fig. 2. Void spaces in the cells of (a) [Fe(PM–PEA)2(NCS)2], monoclinic phase at 293 K, viewed in perspective along axis a; (b) [Fe(PM–PEA)2(NCS)2], orthorhombic phase at
140 K, viewed in perspective along axis b; (c) [Co(PM–PEA)2(NCS)2], monoclinic phase at 293 K, viewed in perspective along axis a.
Fig. 3. (a) Superposition of the structures at 293 K of the phase I Fe complex (dark grey) and the phase II Zn complex (light grey) in its non-solvated (left figure) and solvated
form (right figure). (b) Superposition of the structures at 293 K of the water–methanol solvated (light grey) and the non-solvated (dark grey) Zn phase II complexes.
Fig. 4. Left: representation of cavities in the unit cell of non-solvated phase II Zn complex viewed in perspective along axis b; Right: ORTEP representation of the unit cell of
the solvated phase II Zn complex viewed in perspective along axis b, solvent atoms are represented in space-filling mode. Hydrogens were omitted for clarity.

F. Le Gac et al. / Inorganica Chimica Acta 361 (2008) 3519–3524
3523
Table 2
which are disposed antiparallel along a. But while for the phase I
Fe(II) compound the intrasheet contacts were insured by a mixture
of Cꢁ ꢁ ꢁC and Sꢁ ꢁ ꢁH contacts, for the non-solvated Zn complex the
intrasheet contacts are limited to one interaction between an aro-
matic C–H of the pyridine ring of PM–PEA and one of the alkyne
carbons. For the solvated complex these contacts change, and the
thiocyanates are involved once again. Intersheet contacts for their
part are mainly assured by Sꢀꢁ ꢁ ꢁH–C contacts, for both the non-sol-
vated and the solvated Zn species, as for the phase I Fe compound.
And it is here that lies the main difference between former species:
for the solvated Zn complex, the included methanol is involved in
H-bonding with one of the thiocyanates, which does not establish
any other close contact, and it can be expected that the intrasheet
interaction is thus weakened. This can be verified through the
lengths of the intermolecular contacts: for phase I Fe(II), the short-
est contact involving a sulphur puts a neighbouring carbon at only
3.46 Å, while for phase II Zn complexes, the shortest contact is at
more than 3.6 Å. We can then confidently expect that if we ever
find the Fe(II) analogue of this new phase II, the spin transition will
be more gradual, similarly to what happens with PM–BiA. Until we
find this compound, a more detailed analysis of the packing of the
Zn species is not necessary.
MN6 polyhedron geometry in [M(PM–PEA)2(NCS)2] ꢁ X complexes
[Fe(PM–
PEA)2(NCS)2]
Phase I, 293 K
[Zn(PM–
PEA)2(NCS)2]
Phase II, 293 K
[Zn(PM–
PEA)2(NCS)2]MeOH ꢁ H2O
Phase II, 293 K
Bond lengths (Å)
M–NCS
M–NCS
M–Npy
M–Npy
M–Nimine
M–Nimine
Angles (°)
SCN–M–NCS
SCN–M–Npy
2.055
2.056
2.164
2.167
2.246
2.270
2.050
2.052
2.134
2.155
2.276
2.335
2.052
2.062
2.136
2.145
2.300
2.351
96.0
92.0, 93.4, 99.4,
101.0
98.0
93.6, 94.0, 95.7,
101.6
92.4, 94.0
75.0, 73.6
96.8
93.7, 94.3, 95.7, 99.2
SCN–M–Nimine 93.1, 94.5
PM–PEA bite
angle
92.4, 93.8
74.5, 74.5
73.8, 74.2
Standard deviations are less than 0.005 Å and 0.2°.
due to the contact with the methanol molecule for the solvated spe-
cies. For the iron complex, these angles are 162.3° and 167.4°. But as
can be observed in Fig. 3a, while values for these angles do not seem
to differ, the orientations of the NCS are significantly different.
The packing of the Zn complexes can be described in the same
terms than the Fe(II) parent compound [18a], see Fig. 5. The com-
plexes are organized in sheets parallel to the (ab) plane. These
sheets are constituted of lines along b made of parallel complexes,
A last point concerns the comparison of the calculated powder
X-ray diffraction pattern. We have reported the corresponding
simulations for phase II in Fig. 1. It can be seen, even though the
crystallinity of our powder sample is not very high, that the new
phase seems to be absent or in negligible amount from the synthe-
sized powder (see in particular peaks for 2h ꢃ 18–24°).
Fig. 5. Crystal packing of [Zn(PM–PEA)2(NCS)2] phase II, viewed along axis c (Upper figure) and axis b (lower figure).

3524
F. Le Gac et al. / Inorganica Chimica Acta 361 (2008) 3519–3524
[2] K. Sénéchal-David, GdR ‘‘Magnétisme et Commutation Moléculaires” launch
4. Conclusions
meeting, December, 2007, oral communication; see: M. Morimoto, M. Irie,
Chem. Commun. (2005) 3895.
We have synthesized and characterized complexes of [Zn(PM–
PEA)2(NCS)2] ꢁ X (X = MeOH, H2O). We have evidenced the pecu-
liarity of these complexes which present large void cavities but
as powder do not show solvent inclusion. Crystalline material
was fully characterized by X-ray diffraction and evidenced the
existence of various polymorphs, with and without solvent inclu-
sion. These results are important for the comprehension of the
behaviour of the tricky [Fe(PM–PEA)2(NCS)2], which remains so
far one of the spin-crossover compounds with the largest hystere-
sis observed, and, as a general matter of fact, they underline how
carefully powder and single crystals experimental data must be
handled when interpreting physical properties.
[3] M. Buron-Le Cointe, E. Collet, L. Guerin, M.-H. Lemée-Cailleau, H. Cailleau, M.
Wulff, T. Luty, S. Koshihara, K. Tanaka, J. Lumin. 112 (2005) 235.
[4] (a) A. Hauser, P. Guetlich, H. Spiering, Inorg. Chem. 25 (1986) 4245;
(b) W. Wang, I.Y. Chan, S. Schenker, A. Hauser, J. Chem. Phys. 106 (1997) 3817;
(c) J. Jeftic, A. Hauser, J. Phys. Chem. B 101 (1997) 10262;
(d) S. Schenker, A. Hauser, W. Wang, I.Y. Chan, Chem. Phys. Lett. 297 (1998) 281;
(e) S. Schenker, A. Hauser, W. Wang, I.Y. Chan, J. Chem. Phys. 109 (1998) 9870.
[5] P. Guionneau, M. Marchivie, G. Bravic, J.-F. Létard, D. Chasseau, J. Mater. Chem.
12 (2002) 2546.
[6] C. Baldé, C. Desplanches, A. Wattiaux, P. Guionneau, P. Gütlich, J.-F. Létard,
Chem. Commun. submitted for publication.
[7] C. Baldé, C. Desplanches, P. Gütlich, J.-F. Létard, Inorg. Chim. Acta 361 (2008)
3529.
[8] J.-F. Letard, G. Chastanet, O. Nguyen, S. Marcen, M. Marchivie, P. Guionneau, D.
Chasseau, P. Gütlich, Monatsh. Chem. 134 (2003) 165.
[9] (a) J.-F. Létard, P. Guionneau, E. Codjovi, O. Lavastre, G. Bravic, D. Chasseau, O.
Kahn, J. Am. Chem. Soc. 119 (1997) 10861;
Acknowledgements
(b) H. Daubric, C. Cantin, C. Thomas, J. Kliava, J.-F. Létard, O. Kahn, Chem. Phys.
244 (1999) 75.
[10] R.D. Shannon, Acta Crystallogr., Sect. A 32 (1976) 751.
[11] P. Rosa, F. Le Gac, Y. Bouhadja, P. Guionneau, J.-F. Létard, J. Sánchez Costa, C.M.
Grunert, P. Gutlich, in preparation.
[12] G.M. Sheldrick, Programs for Crystal Structure Analysis (Release 97-2), Institüt
für Anorganische Chemie der Universität, Tammanstrasse 4, D-3400
Göttingen, Germany, 1998.
The authors wish to thank CNRS and Région Aquitaine for fund-
ing. L. Etienne from ICMCB is gratefully acknowledged for ICP-OES
measurements.
[13] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837.
Appendix A. Supplementary material
[14] By extension of the Hume-Rothery rules used in metallurgy, see: W. Hume-
Rothery, R.W. Smallman, C.W. Haworth, The Structure of Metals and Alloys,
Springer, Berlin, 1969.
[15] A.L. Spek, PLATON, A Multipurpose Crystallographic Tool, Utrecht University,
Utrecht, The Netherlands, 2007. The minimum radius for void drawing was
fixed at 1.5 Å.
[16] A.M. Greenaway, C.J. O’Connor, A. Schrock, E. Sinn, Inorg. Chem. 18 (1979)
2692.
[17] B.A. Katz, C.E. Strouse, J. Am. Chem. Soc. 101 (1979) 6214.
[18] (a) P. Guionneau, J.-F. Létard, D.S. Yufit, D. Chasseau, G. Bravic, A.E. Goeta, J.A.K.
Howard, O. Kahn, J. Mater. Chem. 9 (1999) 985;
CCDC 678016, 678017, 678018 and 678301 contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif. Supplementary
data associated with this article can be found, in the online version,
at doi:10.1016/j.ica.2008.03.011.
(b) M. Marchivie, P. Guionneau, J.-F. Létard, D. Chasseau, Acta Crystallogr., Sect.
B 59 (2003) 479;
References
(c) M. Marchivie, P. Guionneau, J.-F. Létard, D. Chasseau, J.A.K. Howard, J. Phys.
Chem. Solids 65 (2004) 17;
(d) P. Guionneau, M. Marchivie, G. Bravic, J.-F. Létard, D. Chasseau, Top. Curr.
Chem. 234 (2004) 97.
[1] (a) E. Freysz, S. Montant, S. Létard, J.-F. Létard, Chem. Phys. Lett. 394 (2004)
318;
(b) S. Bonhommeau, G. Molnar, A. Galet, A. Zwick, J.-A. Real, J.J. McGarvey, A.
Bousseksou, Angew. Chem., Int. Ed. 44 (2005) 4069.