Structure and properties of novel M(dmit)2 salts with the Me3N+-TEMPO cation radical (ME3N+-TEMPO = N,N,N-trimethyl(1-oxyl-2,2,6,6-tetramethylpiperidin-4-yl)ammonium)

Article abstract of DOI:10.1039/b008424g

Salts of the Me₃N⁺-TEMPO cation radical (Me₃N⁺-TEMPO = N,N,N-trimethyl(1-oxyl-2,2,6,6-tetramethylpiperidin-4-yl)ammonium) with I^- and PF₆^- are prepared. Their cr8ystal structure and magnetic properties are reported (antiferromagnetically coupled dimers for (Me₃N-TEMPO)I and antiferromagnetically coupled chains for (Me₃N-TEMPO)PF₆). The Me₃N⁺-TEMPO magnetic cation radical is associated with M(dmit)₂ (M = Ni, Pd; dmit^2- = 1,3-dithiole-2-thioxo-4,5-dithiolato) to afford the divalent (Me₃N-TEMPO)₂[M(dmit)₂] (M = Ni, Pd) salts and the monovalent (Me₃N-TEMPO)[Ni(dmit)₂] salt. The crystal structures of (Me₃N-TEMPO)₂[M(dmit)₂] (M = Ni, Pd) and (Me₃N-TEMPO)[Ni(dmit)₂] are determined. In the (Me₃N-TEMPO)₂[M(dmit)₂] salts, a unique, non planar, chair conformation of the M(dmit)₂ molecule is observed. These salts display conventional Curie behaviour. The (Me₃N-TEMPO)[Ni(dmit)₂] salt exhibits a single-crystal room-temperature conductivity of 4 × 10^-3 S cm^-1 which is rather high for an integral oxidation state salt.

Full text of DOI:10.1039/b008424g

View Article Online / Journal Homepage / Table of Contents for this issue
Structure and properties of novel M(dmit)2 salts with the Me3Nz-
TEMPO cation radical (Me3Nz-TEMPO~N,N,N-trimethyl(1-
oxyl-2,2,6,6-tetramethylpiperidin-4-yl)ammonium){
Shuji Aonuma,*a{ HeÂleÁne Casellas,a Christophe Faulmann,a BeÂneÂdicte Garreau de Bonneval,a
Isabelle Malfant,*a Patrick Cassoux,*a Pascal G. Lacroix,b Yuko Hosokoshic and
Katsuya Inouec
aEquipe PreÂcurseurs MoleÂculaires et MateÂriaux, LCC-CNRS, 205, Route de Narbonne,
31077 Toulouse, Cedex 04, France. E-mail: aonuma@lcc-toulouse.fr,
cassoux@lcc-toulouse.fr
bEquipe MateÂriaux MoleÂculaires et BiomimeÂtiques, LCC-CNRS, 205, Route de Narbonne,
31077 Toulouse, Cedex 04, France
cInstitute for Molecular Science, Myodaiji, Okazaki, 444±8585, Japan
Received 9th August 2000, Accepted 24th October 2000
First published as an Advance Article on the web 12th December 2000
Salts of the Me3Nz-TEMPO cation radical (Me3Nz-TEMPO~N,N,N-trimethyl(1-oxyl-2,2,6,6-
2
tetramethylpiperidin-4-yl)ammonium) with I2 and PF6 are prepared. Their crystal structure and magnetic
properties are reported (antiferromagnetically coupled dimers for (Me3N-TEMPO)I and antiferromagnetically
coupled chains for (Me3N-TEMPO)PF6). The Me3Nz-TEMPO magnetic cation radical is associated with
M(dmit)2 (M~Ni, Pd; dmit22 ~1,3-dithiole-2-thioxo-4,5-dithiolato) to afford the divalent (Me3N-
TEMPO)2[M(dmit)2] (M~Ni, Pd) salts and the monovalent (Me3N-TEMPO)[Ni(dmit)2] salt. The crystal
structures of (Me3N-TEMPO)2[M(dmit)2] (M~Ni, Pd) and (Me3N-TEMPO)[Ni(dmit)2] are determined. In the
(Me3N-TEMPO)2[M(dmit)2] salts, a unique, non planar, chair conformation of the M(dmit)2 molecule is
observed. These salts display conventional Curie behaviour. The (Me3N-TEMPO)[Ni(dmit)2] salt exhibits a
single-crystal room-temperature conductivity of 461023 S cm21, which is rather high for an integral oxidation
state salt.
temperature or applying pressure. Third, the molecular
Introduction
building blocks may have various degrees of freedom (e.g.
conformational, rotational) which may result in a variety of
structures, and hence, of physical properties. Fourth, and
especially for planar systems, the frequently encountered low
dimensionality of the structures often results in unusual
physical properties such as spin density wave, charge density
wave or spin±Peierls instabilities for the conducting properties,
and spin ladder behaviour (quantum spin system) for the
magnetic properties. Due to the rather weak intermolecular
interactions, the tight-binding view of the electronic band
structure is often well suited for understanding the mechanism
of conducting properties. On the other hand, magnetism of
organic radicals (in the narrow sense), where magnetic spins are
isotropic because they originate from s and/or p electrons, can
be rationalised in terms of the ideal Heisenberg spin system.
Co-existence or interplay of conductivity and magnetism is of
great current interest in molecule-based material science.1 Two
different types of mechanism for this interplay can be
considered: (1) magnetic coupling via an indirect exchange
interaction through the conduction electrons;2,3 (2) interaction
of conduction electrons through magnetically ordered spins.4
The most interesting results recently obtained along this line
concern the characterisation of a paramagnetic superconduc-
tor, b@-(BEDT-TTF)4(H3O)[Fe(C2O4)3]?(C6H5CN) (BEDT-
TTF~bis(ethylenedithio)tetrathiafulvalene),5 the ®rst obser-
vation of a magnetic ®eld-induced restoration of a highly
conducting state in the l-(BETS)2FeCl4 phase (BETS~bis-
(ethylenedithio)tetraselenafulvalene),6 and the preparation of
the ®rst antiferromagnetic superconductor, k-(BETS)2FeBr4.7
For such purposes, some of the advantages in using
molecule-based materials may be recollected as follows:
First, their crystal lattice is constructed with molecules that
can be extensively and minutely modi®ed by synthetic methods.
Second, in most cases, molecules are connected by van der
Waals interactions which are much weaker than ionic, covalent
or metallic bonds. This means that a molecule-based crystal is
soft, and that large effects can be expected by changing the
M(dmit)2
(dmit22~1,3-dithiole-2-thioxo-4,5-dithiolato;
Chart 1) is a planar acceptor molecule known as a precursor
to charge-transfer salts exhibiting metal-like and even super-
conducting behaviour.8,9 In other respects, after the discovery
in 1985 of the ®rst charge-transfer salt-based bulk ferromagnet,
[Fe(C5Me5)2](TCNE) (TCNE~tetracyanoethylene),10 a great
research effort has been devoted to the study of charge-transfer
salts involving planar p acceptor anion radicals associated with
paramagnetic cations,9,11 including the ®rst dmit-based bulk
ferromagnet, [Mn(C5Me5)2][Ni(dmit)2].12
{The crystal structure of (Me3N-TEMPO)2[M(dmit)2] (Fig. S1) and the
molar magnetic susceptibility of (Me3N-TEMPO)PF6 (Fig. S2) are
available as supplementary data. For direct electronic access see http://
www.rsc.org/suppdata/jm/b0/b008424g/
{Visiting research fellow from The Institute for Solid State Physics, The
University of Tokyo, Kashiwa, Chiba 277-8581, Japan.
The ®rst purely organic (i.e., without magnetic d spins) bulk
ferromagnet,13 the nitronyl aminoxyl radical crystal phase, b-p-
NPNN (p-NPNN~p-nitrophenyl nitronyl nitroxide14), was
reported in 1991.15 Other purely organic ferromagnets, such as
diazaadamantanedioxyl16 and several TEMPO derivatives
DOI: 10.1039/b008424g
J. Mater. Chem., 2001, 11, 337±345
This journal is # The Royal Society of Chemistry 2001
337

View Article Online
Scheme 1
reaction of equimolar (Me3N-TEMPO)I with Na2[M(dmit)2]
(M~Ni, Pd) (Method B).
Metathesis between Na[Ni(dmit)2] and (Me3N-TEMPO)I
(Method C in Scheme 3) or iodine oxidation of the divalent
(Me3N-TEMPO)2[Ni(dmit)2] salt (Method D) afforded the
monovalent (Me3N-TEMPO)[Ni(dmit)2], which is readily
soluble in acetone or in acetonitrile.
In some cases, Method C simultaneously produced further
oxidised species assumed to be (Me3N-TEMPO)[Ni(dmit)2]2.
This prompted us to study the redox behaviour of Me3Nz-
TEMPO to evaluate its potential oxidation ability. The cyclic
voltammogram of (Me3N-TEMPO)PF6 measured in acetoni-
trile shows one pair of reversible waves. The half wave potential
values at 0.92 V or 0.90 V vs. SCE were obtained by cyclic
voltammetry (CV) or square wave voltammetry (SQW),
respectively. With this last method, the half width of the
peak corresponds to the transfer of one electron. In the case of
(Me3N-TEMPO)I, the redox potential was also ca. 0.9 V vs.
SCE, with the corresponding wave overlapping with the I2/I?
redox wave (0.86 V vs. SCE). Thus, given the oxidation
potential observed for Me3Nz-TEMPO, this radical cannot be
considered as responsible for a possible further oxidation of
(Me3N-TEMPO)[Ni(dmit)2].
Chart 1
(TEMPO~2,2,6,6-tetramethylpiperidin-1-oxyl),17 have been
reported. Charge-transfer salts based on organic radicals
have also been extensively studied.18 One of the most
spectacular recent results obtained by using charge-transfer
salts of organic radicals is the observation of a spin ladder
behaviour in a nitronyl aminoxyl-pyridinium (EPYNN) salt of
Ni(dmit)2.19 Charge-transfer salts of tetrathiafulvalene (TTF)
donors bearing nitroxyl20 radicals have been also reported.21,22
For promoting a possible interaction between conducting
electrons and magnetic spins in a hybrid molecule-based
system, the conducting and magnetic components should be
geometrically located close to each other. Therefore, less
sterically hindered and compact organic radicals should be
more suitable for use as counter cations. For example, alkyl
ammonium cations (e.g. Me3Nz-TEMPO, N,N,N-trimethyl(1-
oxyl-2,2,6,6-tetramethylpiperidin-4-yl)ammonium; see Chart
1) should be preferred to pyridinium cations such as
EPYNN. Nitronyl or imino aminoxyl cations with small
ammonio groups (e.g. Me3Nz-NN, Me3Nz-IN) may be even
more attractive because of their small size and also because of
the delocalisation of the radical spin in the nitronyl aminoxyl,
which could possibly enhance intermolecular interaction.
In this paper, we report on the crystal structure and magnetic
properties of (Me3N-TEMPO)X salts (X~I, PF6), which can
be readily prepared, are very stable, and can be associated with
M(dmit)2 radical anions. We also describe the preparation,
crystal structure, and magnetic properties of (Me3N-
TEMPO)x[M(dmit)2] (M~Ni, x~2, 1; M~Pd, x~2).
Crystal structures
(Me3N-TEMPO)I. Fig. 1 shows the molecular structure,
together with the atomic labelling scheme, for the two
independent Me3Nz-TEMPO cations and the two I2 anions.
The structure is formed by groups of four cations (A, B, A', B'),
in which the four oxygen atoms form a parallelogram (Fig. 2).
¼
The O O distances are 4.578(6) between A and B' and
Results and discussions
Scheme 2
Syntheses
(Me3N-TEMPO)I was prepared by reductive amination of
oxo-TEMPO to Me2N-TEMPO, followed by methylation with
MeI. (Me3N-TEMPO)PF6 was prepared from (Me3N-
TEMPO)I by anion exchange with KPF6 (Scheme 1).
The (Me3N-TEMPO)2[Ni(dmit)2] and (Me3N-TEMPO)2-
[Pd(dmit)2] divalent salts were obtained by a metathesis
reaction of an excess of (Me3N-TEMPO)I with the corre-
sponding (n-Bu4N)2[M(dmit)2] (M~Ni, Pd) (Method A in
Scheme 2), which afforded single crystals suitable for X-ray
structural analysis.
TEMPO)2[M(dmit)2] was also obtained by a similar metathesis
A
crystalline powder of (Me3N-
Scheme 3
338
J. Mater. Chem., 2001, 11, 337±345

View Article Online
Fig. 1 Molecular structure and atom labelling scheme for (Me3N-
TEMPO)I.
Fig. 4 Projection of the structure of (Me3N-TEMPO)PF6 showing two
adjacent chains along the a axis (anions are omitted for clarity). Dotted
¼
¼
Ê
lines represent the O O distances of 5.7540(7) A and the N
O
Ê
distances of 4.827(3) A.
between their mean planes (except H and Me) is 89.96(3)³) to
each other. Six other groups of cations for which the oxygen
planes form an angle of about 60³ with the central plane
(Fig. 2b) surround each A, B, A', or B' group along the [110]
and [011] directions.
(Me3N-TEMPO)PF6. The molecular structure, together
with the atom-labelling scheme, is shown in Fig. 3. The
asymmetric unit contains one half Me3Nz-TEMPO unit and
2
one disordered PF6 unit, both lying on a mirror plane.
Consequently, the unit cell contains four Me3Nz-TEMPO
cations and four PF62 anions. The structure consists of chains
of cationic units, parallel to the a direction, with the anions
lying in between (Fig. 4). Each chain (for example, ¼A±B±C±
D¼ in Fig. 4) is built on Me3Nz-TEMPO units perpendicular
to each other (angle between their mean planes (except H and
Me) is 90.41(4)³). Along the chains, the Me3Nz-TEMPO units
¼
¼
are connected through N O and O O interactions with
Ê
interatomic distances of 4.827(3) and 5.7540(7) A, respectively.
The interatomic distances between atoms belonging to adjacent
Ê
chains are much longer (w7.7 A), indicating the absence of any
interchain interactions.
Fig. 2 (a) Projection of the structure of (Me3N-TEMPO)I showing the
parallelogram formed by the four cations. Dotted lines represent the
(Me3N-TEMPO)2[M(dmit)2] (M~Ni, Pd). The molecular
structure of these two compounds, together with the atom-
labelling scheme is shown in Fig. 5. These two compounds are
isostructural. The unit cell contains two [M(dmit)2]22 and four
¼
Ê
O
O distances of 4.578(6) (AB' and A'B) and 5.921(6) A (AB and
A'B'). (b) Projection of the structure of (Me3N-TEMPO)I along the a
and c axes showing the arrangement of the adjacent groups of oxygen
atoms (only oxygen atoms are shown for clarity).
¼
Ê
5.921(6) A between A and B. The corresponding N
O
Ê
Ê
distances are 3.789(6) A and 5.024(6) A, respectively
(Fig. 2a). This may indicate that the interaction between A
and B' may be stronger than that between A and B, leading to a
possible dimeric arrangement (see the section below on the
magnetic properties of this compound). Within each group of
four cations, the Me3Nz-TEMPO cations (A and B, A' and B,
A and B', A' and B'; see Fig. 2)are perpendicular (angle
Fig. 3 Molecular structure and atom labelling scheme for (Me3N-
TEMPO)PF6.
Fig. 5 (a) Molecular structure and atom labelling scheme for (Me3N-
TEMPO)2[M(dmit)2] (M~Ni, Pd) and (b) its side view.
J. Mater. Chem., 2001, 11, 337±345
339

View Article Online
Me3Nz-TEMPO units. The structure consists of sheets of
parallel [M(dmit)2]22 units lying in the ac plane and alternating
with sheets of cations along the b direction (Fig. S1). The two
shortest distances between oxygen atoms are 7.655(2) and
7.651(3) A. As seen on the side view of the [M(dmit)2]22 shown
Ê
in Fig. 5b, the molecule is not planar: the two dmit planes form
an angle of 164.02(5)³ (Ni) and 165.49(7)³ (Pd) with the central
MS4 planar core. To our knowledge, such a chair conformation
has never been observed in any other mononuclear dmit
complex-based compound. Nevertheless, such a bent confor-
mation has been observed in dinuclear compounds containing
22
the dmit ligand together with the tto ligand (tto~C2S4
,
tetrathiooxalato), namely (R4N)2{tto[Ni(dmit)]2} (R~alkyl
group).23 In these complexes, the central tto fragment forms
an angle of 178 to 155.5³ with the external Ni(dmit) moieties. In
(Me3N-TEMPO)2[M(dmit)2], the interplanar distance between
Ê
the two dmit units is 0.868(5) A (Ni) and 0.845(4) A (Pd).
Ê
Fig. 7 Projection of the structure of (Me3N-TEMPO)[Ni(dmit)2] in the
bc plane, showing the chains of [Ni(dmit)2]2 dimers along the b
direction.
(Me3N-TEMPO)[Ni(dmit)2]. The
molecular
structure,
together with the atom-labelling scheme, is shown in Fig. 6.
The unit cell contains four [Ni(dmit)2]2 units and four Me3Nz-
TEMPO cations. The structure consists of layers of
[Ni(dmit)2]2 units lying in the ab plane and separated by
sheets of cations along the c direction (Fig. 7). The layers are
built on [Ni(dmit)2]2 units connected to each other through
Magnetism
(Me3N-TEMPO)I and (Me3N-TEMPO)PF6. The tempera-
ture dependence of the molar magnetic susceptibility (x)
measured for a crystalline sample of (Me3N-TEMPO)I is
shown in Fig. 8. Within the 100±300 K temperature range,
observed data are well described by the Curie±Weiss law with a
Curie constant, C, of 0.366 emu K mol21 and a Weiss constant,
h, of 26.3 K. The Curie constant corresponds to 98% of the
normal value of one isolated S~1/2 spin. The magnetic
susceptibility exhibits a maximum at ca. 12 K. Below this
temperature, x decreases down to ca. 5 K. Upon further
cooling, x increases again. As previously indicated, some
features in the crystal structure of (Me3N-TEMPO)I suggest a
possible dimeric arrangement. Should this compound be purely
dimeric, then x should decrease down to zero for TA0.
Therefore, the behaviour below 10 K is clearly indicative of the
in¯uence of Curie impurities due to lattice defects. From the xT
value extrapolated to T~0, this Curie-like component is
determined as C~0.06 emu K mol21, which corresponds to
16% of spins. For crystals of good quality, this value is usually
lower, but, in the present case, the larger value of 16% may be
attributed to considerable crystal defects in the polycrystalline
sample studied.
¼
¼
Ê
two short S S contacts (S1 S6, 3.516(8) A), thus forming
Ê
dimers with an interplanar distance of 3.81 A. Along the b
direction, each dimer is connected to adjacent dimers through
¼
¼
¼
several S S contacts (S4 S4: 3.519(9), S4 S5: 3.677(8),
¼
¼
¼
Ê
S4 S9: 3.362(6), S4 S10: 3.623(7) and S5 S9: 3.573(7) A),
thus forming chains of dimers along [010]. These chains are not
isolated as they are connected to adjacent chains along the a
¼
¼
Ê
direction through S1 S1 (3.614(10) A) and S3 S7
Ê
(3.533(7) A) contacts.
Single crystal conductivity of (Me3N-TEMPO)[Ni(dmit)2] is
461023 S cm21 at room temperature, and shows semiconduc-
tive temperature dependence (Ea~0.12 eV). The rather high
conductivity for a monovalent Ni(dmit)2 salt can be attributed
to the many short contacts between Ni(dmit)2 molecules.
In the structure of the ®ve compounds described above, the
piperidine ring of the Me3Nz-TEMPO cation adopts a chair
conformation. Consequently, the six ring atoms form a very
distorted plane. A useful gauge for estimating the planarity of
the C2NO moiety is given by the parameter a, which is de®ned24
as the angle between the N±O bond and the CNC plane
(Table 1). The C±N±C±C (t1, t'1), N±C±C±C (t2, t'2) and C±
C±C±C (t3, t'3) torsion angles are between those observed for
mono and di-substituted rings. The N±O distances and CNC
angles are within the range observed in other related
compounds.25
Generally, the spin density of TEMPO derivatives is
mainly localized on the aminoxyl (NO) moiety.26 ROHF/
INDO calculations carried out on both (Me3N-TEMPO)I
and (Me3N-TEMPO)PF6 con®rm that the aminoxyl (NO)
moiety bears at least 95% of the spin density (Fig. 9).
Therefore, the magnetic properties of (Me3N-TEMPO)I
(and (Me3N-
TEMPO)PF6 as well) must essentially arise from intermo-
lecular interactions involving the NO moieties. In (Me3N-
TEMPO)I, the strongest interaction between NO moieties of
¼
¼
two neighbouring radicals, A B' on the one hand and A
on the other hand (see Fig. 2a) are evidenced by the short
B
¼
Ê Ê
O O distances (4.578(6) A and 5.921(6) A respectively; see
Fig. 2a) or the even shorter corresponding N O distances
¼
Ê
Ê
(3.789(6) A and 5.024(6) A, respectively). If the interaction
between radicals A and B could be considered as negligible
compared to that between A and B', then (Me3N-TEMPO)I
could be considered as an essentially dimeric system.
Therefore, with this assumption, the temperature depen-
dence of the magnetic susceptibility can readily be ascribed
to an antiferromagnetic coupling between S~1/2 dimers,
according to eqn. (1) including the corresponding term for
Curie impurities:27
Fig. 6 Molecular structure and atom labelling scheme for (Me3N-
TEMPO)[Ni(dmit)2].
340
J. Mater. Chem., 2001, 11, 337±345

View Article Online
Table 1 Comparison of some geometrical featuresa of (Me3N-TEMPO)I, (Me3N-TEMPO)PF6, (Me3N-TEMPO)2[Ni(dmit)2], (Me3N-
TEMPO)2[Pd(dmit)2] and (Me3N-TEMPO)[Ni(dmit)2]
t1/³
t
t'2/³
12/³
t
t'3/³
13/³
Ê
N±O/A
Compound
a/³
t'1/³
C±N±C/³
124.1(4)
123.6(4)
(Me3N-TEMPO)I
14.47
18.73
33.6(7)
247.3(6)
48.1(6)
48.0(6)
246.0(6)
48.1(3)
245.8(5)
46.9(5)
246.3(5)
48.2(6)
252(2)
50(3)
65.5(6)
1.275(5)
1.264(5)
234.5(7)
235.6(7)
34.4(7)
38.2(3)
33.9(5)
234.5(5)
34.0(6)
235.0(6)
40(3)
264.7(6)
263.5(6)
62.3(6)
62.7(3)
61.3(5)
261.9(5)
63.0(6)
263.5(6)
70(2)
(Me3N-TEMPO)PF6
(Me3N-TEMPO)2[Ni(dmit)2]
20.0
19.82
1.282(4)
1.272(4)
123.4(3)
124.9(3)
(Me3N-TEMPO)2[Pd(dmit)2]
(Me3N-TEMPO)[Ni(dmit)2]
19.19
14.44
1.273(5)
1.32(2)
124.4(4)
126(2)
239(3)
270(2)
aa is the angle between the N±O bond and the C±N±C plane. t1, t'1, t2, t'2, t3 and t'3 are endocyclic torsion angles in the piperidine ring.
!
X
ꢀ
ꢁ
n{1
2Nb2g2
1
NB2g2
kT
.
Si Siz1
H~{J
(2)
i~l
Á
x~(1{r)
|
zr
(1)
{J
kT
kT
3z exp
where the summation runs over the n sites of the chain. When n
tends to in®nity, the results can be ®tted by the numerical
expression eqn. (3):30
where N is Avogadro's number, g is the Zeeman splitting
parameter (®xed in this case at 2.00), b is the Bohr magneton, k
is the Boltzman constant, J is the singlet±triplet energy gap, and
r is the proportion of non coupled impurity. The least-square
best ®t is found for J~217.3 cm21 and r~0.145.28 The
agreement is satisfactory as illustrated in Fig. 8. The dis-
crepancy factor de®ned as R(x)~[S (xobs2xcalc)2/S(xobs)2] is
equal to 5.261025 in the temperature range of 5±300 K.
In the case of (Me3N-TEMPO)PF6, the interaction between
the NO moieties of neighbouring radicals (see Fig. 4) extends
over the entire crystal. The structure clearly reveals a magnetic
chain of equally spaced spin S~1/2. The temperature
dependence of the product of the magnetic susceptibility on
the temperature is shown in Fig. S2. At room temperature, xT
is equal to 0.328 emu K mol21, which corresponds to 88% of
the expected value. The spin Hamiltonian best suited to
describe the isotropic interaction between nearest neighbours is
give in eqn. (2)29
!
Nb2g2
0:25z0:074975xz0:075235x2
1:0z0:9931xz0:172135x2z0:757825x3
xT~
|
(3)
k
where x~|J|/kT. Least squares ®tting of the experimental data
leads to J~20.17 cm21 with a discrepancy factor R(x) of
9.661026
.
Since the nature of the magnetic exchange pathway is so
similar in both materials, it may appear surprising that their
magnetic response is so different. Indeed, both structures reveal
that the cations are placed close together and the angle between
their mean planes is close to 90³ (see Fig. 2 and Fig. 4). It can be
inferred from these structural features that direct overlap of the
NO orbitals is the sole pathway responsible for the magnetic
exchange in these salts. Therefore, it can be assumed that, to a
large extent, the main relevant structural parameter that
¼
correlates with the J value is the shortest N O distance (D)
between the two centres bearing the spin density. Therefore,
although the Coffman expression31 is usually used for
Fig. 9 Picture of the SOMO (Singly Occupied Molecular Orbital)
calculated by ROHF/INDO, and views of intermolecular arrangement
in (a) (Me3N-TEMPO)I and (b) (Me3N-TEMPO)PF6. Numbers in
round brackets are calculated spin densities on each atom in NO
moiety.
Fig. 8 Experimental (à) and calculated (solid line) temperature
dependence of the molar magnetic susceptibility of (Me3N-TEMPO)I.
J. Mater. Chem., 2001, 11, 337±345
341

View Article Online
intramolecular interaction, the J and D values may be
tentatively analysed in the present case, with eqn. (4):
of the xT value below 100 K. Air oxidation of (Me3N-
TEMPO)2[Pd(dmit)2] in the presence of acetic acid afforded a
black powder sample. Elemental analyses of this sample are
consistent with a (Me3N-TEMPO)[Pd(dmit)2]4 formulation.
J~A exp ({eD)
(4)
The room temperature powder conductivity is 161022 S cm21
.
where A is an appropriate exchange constant and e is an
exponential exchange coef®cient. When going from
At this stage of the work, it is clear that we have still a long
way to go before obtaining compounds in this series exhibiting
interplay of conductivity and magnetism. It should be however
noted that, among all molecule-based material, such an
interplay has been clearly established in only one series of
compounds, i.e., those based on the BETS donor.6
Ê
D~3.789(6) to 4.827(3) A, the present magnetic data indicate
that J drops from 217.3 to 20.17 cm21, which is consistent
with an e parameter equal to 4.4. This value lies in the typical
3.0±5.5 range previously reported for direct intramolecular
overlaps of magnetic orbitals.32
For ensuring interaction between conduction electrons (here,
those of the M(dmit)2) and localised spins, both the conducting
and magnetic components should be geometrically located
close to each other. Therefore the little sterically hindered
nitronyl and imino aminoxyl (see Chart 1) cation radicals that
are smaller and, in addition, delocalised, may be even more
ef®cient than TEMPO-based radicals. Preparation of these
ammonium cations is in progress.
The situation encountered in (Me3N-TEMPO)I and (Me3N-
TEMPO)PF6 contrasts with the recent report on the ferro-
magnetic coupling in various TEMPO-based molecules.17b,25e,f
It should be noted that (Me3N-TEMPO)I and (Me3N-
¼
TEMPO)PF6 exhibit short NO NO distances with the NO
moieties lying orthogonal to each other. This leads to non-zero
overlaps, which favour antiferromagnetic interactions. By
contrast, as stressed by Nogami et al.17b,25e,f TEMPO-based
ferromagnetically coupled materials involves NO moieties with
¼
Ê
intermolecular NO NO distances larger than 5.5 A, with the
corresponding orbitals being nearly parallel to each other.
Experimental
General
(Me3N-TEMPO)2[M(dmit)2] (M~Ni, Pd). In both the
isostructural (Me3N-TEMPO)2[Ni(dmit)2] and (Me3N-
TEMPO)2[Pd(dmit)2] divalent salts, the room temperature xT
value is very close to the value expected for two S~1/2 spin
systems. Both salts show Curie behaviour without any striking
magnetic interactions. This agrees with the absence of short
direct distances between NO moieties in their crystal structure.
oxo-TEMPO was prepared from 2,2,6,6-tetramethyl-4-piper-
idone or its hydrochloride.33 (n-Bu4N)2[Ni(dmit)2], (n-Bu4N)-
[Ni(dmit)2] and (n-Bu4N)2[Pd(dmit)2] were prepared following
previously reported methods.34 Other starting materials used in
this work were used without further puri®cation unless
otherwise noted. New compounds were characterised from
single-crystal structure by X-ray analysis, or elemental analysis
data are given.
Conclusions
(Me3N-TEMPO)I
In this work, the great versatility of both the TEMPO radical
and M(dmit)2-based anion systems has been established again.
Indeed, both Me3N-TEMPO and M(dmit)2 (M~Ni, Pd) can
be associated in 2 : 1, 1 : 1, and even 1 : 2 non integral oxidation
state salts. We have also shown that a minor chemical
modi®cation, as for example that between (Me3N-TEMPO)I
and (Me3N-TEMPO)PF6, may lead to a strikingly different
structure and solid state properties, i.e., antiferromagnetically
coupled dimers for (Me3N-TEMPO)I, antiferromagnetic
Me2N-TEMPO was prepared by reductive amination35 of oxo-
TEMPO (13.8 g, 81.0 mmol). Crude Me2N-TEMPO dissolved
in ethanol (100 cm3) was reacted with methyl iodide (11.0 g,
77.2 mmol). This afforded the (Me3N-TEMPO)I ammonium
salt as pale orange microcrystals (10.8 g, 31.7 mmol) in a 40%
yield.36 Single crystals were obtained by slow evaporation of
acetonitrile solutions.
(Me3N-TEMPO)PF6
chains for (Me3N-TEMPO)PF6. Though
a number of
ferromagnetic TEMPO-based compounds were recently repor-
Filtered solutions of (Me3N-TEMPO)I (3.67 g, 10.8 mmol) in
water (100 cm3) and KPF6 (27.0 g, 147 mmol) in water
(200 cm3) were mixed together. Pale orange microcrystals
immediately formed, which were ®ltered off and washed with
ethanol and ether to afford 3.46 g (9.63 mmol) of (Me3N-
TEMPO)PF6 in a 89% yield. Single crystals were obtained by
slow evaporation of acetonitrile solutions.
e,f
ted,17b,25 no ferromagnetic coupling has been observed in the
Me3Nz-TEMPO-based compounds studied in this work. This
may be related to the non-zero overlap of the orbitals of the
NO moieties in these compounds.
The resolution of the structure of the 2 : 1 (Me3N-
TEMPO)2[M(dmit)2] (M~Ni, Pd) salts reveals a unique
feature, i.e., the chair conformation of the [M(dmit)2] moieties
that has never been previously observed in any known
(Me3N-TEMPO)2[Ni(dmit)2]
[M(dmit)2]-based
compound.
Nevertheless,
(Me3N-
Method A. Filtered solutions of (Me3N-TEMPO)I (1.00 g,
2.93 mmol) in acetonitrile (200 cm3) and (n-Bu4N)2[Ni(dmit)2]
(0.642 g, 0.664 mmol) in acetone (180 cm3) were mixed
together, and left for a week. Filtration and washing with
acetonitrile and acetone afforded deep dark violet blocks
TEMPO)2[M(dmit)2] just displays Curie behaviour, indicating
no magnetic interaction between the NO moieties.
The structure of the derived 1 : 1 (Me3N-TEMPO)[Ni(dmit)2]
salt consist of chains of [Ni(dmit)2]2 dimers connected through
¼
¼
short S S contacts. These S S interactions may explain the
value of the room-temperature single-crystal conductivity,
461023 S cm21, which is rather high for a 1 : 1 integral
oxidation state salt.
(0.381 g,
0.433 mmol;
typical
dimensions
were
0.360.260.2 mm3) in a 65% yield.
Preparation of further oxidised salts is in progress. As
previously mentioned, an insoluble dark green powder has been
obtained as a side product in the preparation of (Me3N-
TEMPO)[Ni(dmit)2] by method C (see Scheme 3). Elemental
Method B. Na2(dmit) was generated in situ from dmit-
(COPh)2 (0.541 g, 1.33 mmol) and CH3ONa (0.439 g,
8.13 mmol) in methanol (10 cm3). To the resulting mixture,
solutions of NiCl2?6H2O (0.166 g, 0.697 mmol) in methanol
(20 cm3) and (Me3N-TEMPO)I (0.386 g, 1.13 mmol) in 95%
methanol (65 cm3) were successively added. After stirring over-
night, a dark green crystalline powder was ®ltered off,
redissolved in acetone±acetonitrile, and reprecipitated with
analyses of this sample are consistent with
a (Me3N-
TEMPO)[Ni(dmit)2]2 formulation. This sample has a room-
temperature powder conductivity of 161022 S cm21, and
exhibits a ferromagnetic behaviour evidenced by an increase
342
J. Mater. Chem., 2001, 11, 337±345

View Article Online
Table 2 Crystallographic data for (Me3N-TEMPO)I, (Me3N-TEMPO)PF6, (Me3N-TEMPO)2[Ni(dmit)2], (Me3N-TEMPO)2[Pd(dmit)2] and
(Me3N-TEMPO)[Ni(dmit)2]
Compound
Chemical formula
Formula weight
T/K
C
24H52I2N4O2
682.50
293(2)
C12H26F6N2OP
359.32
160(2)
C30H52N4O2NiS10
880.06
293(2)
C30H52N4O2PdS10
927.76
293(2)
C18H26N2ONiS10
665.72
293(2)
Crystal system
Space group
monoclinic
P21/n
orthorhombic
Pnma
11.400(1)
10.852(1)
13.354(2)
Ð
orthorhombic
Pbca
14.521(2)
15.982(3)
18.150(5)
Ð
orthorhombic
Pbca
14.646(3)
16.012(4)
18.126(5)
Ð
monoclinic
P21/c
Ê
a/A
14.437(4)
8.372(1)
11.998(2)
28.592(4)
91.25(2)
2871.2(7)
1.540
Ê
b/A
13.763(4)
17.278(4)
112.45(2)
3173(2)
1.429
Ê
c/A
b/³
V/A
Ê 3
1652.1(3)
1.445
4212(1)
1.388
4250(2)
1.450
r
calcd/g cm23
Z
4
2.006
9478/9161
0.0717
0.0327
4
0.229
15600/1676
0.0534
0.0453
4
0.989
7420/1957
0.0645
0.0271
4
0.960
5550/5527
0.0113
0.0401
4
1.419
27237/5644
0.2928
0.0715
m(Mo-Ka)/mm21
Re¯ections collected/unique
Rint
R(Fo)a
Rw(Fo)b
0.1191
0.1130
0.0525
0.1048
0.1551
2 1/2
] .
aR(Fo)~S(||Fo|2|Fc||)/S|Fo|. bRw(Fo)~[Sw(|Fo|2|Fc|)2/SwFo
propan-2-ol to afford dark purple microcrystals (0.378 g,
0.429 mmol) in a 65% yield. Found: C, 41.07; H, 5.52; N,
6.31%; (C12H26N2O)2(NiC6S10) requires C, 40.94; H, 5.96; N,
6.37%.
formed during this reaction. The ®ltrate was reprecipitated by
adding ether (100 cm3) to give crude (Me3N-TEMPO)[Ni-
(dmit)2]. Recrystallisation from acetone yielded dark green
needles and micro-crystals (0.139 g, 0.209 mmol) in a 75%
yield.
Found:
C,
31.14;
H,
2.86;
N,
4.47%;
(C12H26N2O)(PdC6S10) requires C, 32.47; H, 3.94; N, 4.21%.
(Me3Nz-TEMPO)2[Pd(dmit)2]
Method A. Filtered solutions of (Me3N-TEMPO)I
(0.494 g, 1.45 mmol) in acetonitrile (60 cm3) and (n-
Bu4N)2[Pd(dmit)2] (0.323 g, 0.328 mmol) in acetone
(50 cm3) were mixed together, and left at room temperature
for one day. Filtration and washing with acetonitrile and
acetone afforded deep violet blocks (0.142 g, 0.153 mmol,
typical dimensions 0.360.260.1 mm3) in a 47% yield. A
second crop (0.108 g) is obtained after concentration of the
resulting solution.
Electrochemical studies
Cyclic voltammetry data (homemade potentiostat interfaced
with a PC computer):37 Au working microelectrode (125 mm),
acetonitrile solution, 0.1 mol dm23 (n-Bu4N)ClO4 as support-
ing electrolyte (dried by melting and pumping under vacuum
immediately before use), Pt wire as auxiliary electrode, SCE as
reference electrode. Square wave voltammetry38 data were
recorded on a PAR273.
Method B. Na2(dmit) was generated in situ from dmit-
(COPh)2 (0.429 g, 1.09 mmol) and CH3ONa (0.325 g,
6.02 mmol) in methanol (10 cm3). To the resulting mixture,
solutions of K2PdCl4 (0.190 g, 0.582 mmol) in 50% aq.
X-Ray diffraction
For (Me3N-TEMPO)I and (Me3N-TEMPO)2[M(dmit)2]
(M~Ni, Pd), data collection was carried out on a Enraf-
Nonius CAD4 diffractometer with graphite-monochromated
methanol
(20 cm3)
and
(Me3N-TEMPO)I
(0.368 g,
1.08 mmol) in 95% methanol (100 cm3) were successively
added. After stirring overnight, a dark red-violet powder was
®ltered off, redissolved in acetone±acetonitrile, and reprecipi-
tated with propan-2-ol to afford dark purple microcrystals
(0.219 g, 0.236 mmol) in a 44% yield. Found: C, 38.66; H, 5.27;
N, 5.87%; (C12H26N2O)2(PdC6S10) requires C, 38.83; H, 5.65;
N, 6.04%.
Ê
Mo-Ka radiation (l~0.71073 A) by using the CAD4-Express
package.39 For every compound, the intensity of three
re¯ections was monitored throughout the data collection,
and no signi®cant decay was observed. Accurate unit cell
parameters were obtained by least-squares re®nements on the
basis of the setting angles of 25 re¯ections. Empirical
absorption corrections based on psi-scans were applied for
every compound.40
(Me3N-TEMPO)[Ni(dmit)2]
For (Me3N-TEMPO)PF6 and (Me3N-TEMPO)[Ni(dmit)2],
data collection and cell re®nement were performed on a IPDS
Stoe diffractometer, by using the IPDS software41 at 160 and
293 K, respectively. All calculations were carried out using the
WinGX package (version 1.62),42 with SIR9243 or
SHELXS8644 for the structure solution and SHELXL9745
for the subsequent re®nements. Drawings were obtained with
CAMERON.46
All non-H atoms were re®ned anisotropically, except in
(Me3N-TEMPO)[Ni(dmit)2], for which C, N and O were
re®ned isotropically. Positions of H atoms were calculated,
with an isotropic displacement parameter 20% higher than that
of the parent atom. Crystallographic data for every compound
are summarised in Table 2. CCDC reference number 1145/260.
See http://www.rsc.org/suppdata/jm/b0/b008424g/ for crystal-
lographic ®les in .cif format.
Method C. Na2[Ni(dmit)2] was prepared in situ from
Na2(dmit) (0.300 g, 1.23 mmol) in acetone (30 cm3) and
NiCl2?6H2O (0.161 g, 6.76 mmol) in MeOH (7 cm3). Iodine
(0.038 g, 0.150 mmol) in acetone (30 cm3) and (Me3N-
TEMPO)I (0.234 g, 0.682 mmol) in MeOH (30 cm3) were
successively added. The solution was concentrated to half
volume and placed in a freezer. Addition of i-PrOH gave a
mixture of a brown powder and black shiny needles. These
needles were characterised as the monovalent (Me3N-TEM-
PO)[Ni(dmit)2] salt by X-ray analysis.
Method D. A solution of iodine (0.0363 g, 0.143 mmol) in
acetone (30 cm3) was added to
a solution of (Me3N-
TEMPO)2[Ni(dmit)2] (0.245 g, 0.278 mmol) in acetone
(1 dm3). After 2 days, the reaction mixture was concentrated
down to 25 cm3 and ®ltered to remove (Me3N-TEMPO)I
J. Mater. Chem., 2001, 11, 337±345
343

View Article Online
12 C. Faulmann, A. E. Pullen, E. Riviere, Y. Journaux, L. Retailleau
and P. Cassoux, Synth. Met., 1999, 103, 2296.
Conductivity measurements
Temperature dependent single-crystal conductivity measure-
ments were carried out following the standard four-probe
technique. Electrical contacts were obtained by gluing four
gold wires to the crystal with Emetron M8001 gold paint.
13 For a review, see M. Kinoshita, Jpn. J. Appl. Phys., 1994, 33, 5718.
14 IUPAC recommended name is 2-(p-nitrophenyl)-4,4,5,5-tetra-
methyl-4,5-dihydro-1H-imidazol-1-oxyl 3-oxide.
15 M. Tamura, Y. Nakazawa, D. Shiomi, K. Nozawa, Y. Hosokoshi,
M. Ishikawa, M. Takahashi and M. Kinoshita, Chem. Phys. Lett.,
1991, 186, 401; Y. Nakazawa, M. Tamura, N. Shirakawa,
D. Shiomi, M. Takahashi, M. Kinoshita and M. Ishikawa,
Phys. Rev. B: Condens. Matter, 1992, 46, 8906.
Magnetic measurements
Static magnetic measurements were carried out using a
Quantum Design MPMS 5S SQUID magnetometer with the
temperature range of 1.8±300 K and 0±5 T for randomly
oriented crystals. The samples were contained in a thin plastic
wrap (New Krewrap1, Kureha Chemicals Industry Co., Ltd.).
The diamagnetic contribution of the sample holder was
corrected experimentally. The diamagnetic contributions
(xdia) of the constituent atoms were experimentally determined.
16 R. Chiarelli, M. A. Novak, A. Rassat and J. L. Tholence, Nature
(London), 1993, 363, 147.
17 (a) K. Togashi, R. Imachi, K. Tomioka, H. Tsuboi, T. Ishida,
T. Nogami, N. Takeda and M. Ishikawa, Bull. Chem. Soc. Jpn.,
1996, 69, 2821; (b) T. Nogami, T. Ishida, M. Yasui, F. Iwasaki,
N. Takeda, M. Ishikawa, T. Kawakami and K. Yamaguchi, Bull.
Chem. Soc. Jpn., 1996, 69, 1841.
18 See for example: Y. Mazaki, K. Awaga and K. Kobayashi,
J. Chem. Soc., Chem Commun., 1992, 1661; S. Nakatsuji, A. Takai,
K. Nishikawa, Y. Morimoto, N. Yasuoka, K. Suzuki, T. Enoki
and H. Anzai, Chem. Commun., 1997, 275; S. Nakatsuji, A. Takai,
K. Nishikawa, Y. Morimoto, N. Yasuoka, K. Suzuki, T. Enoki
and H. Anzai, J. Mater. Chem., 1999, 9, 1747.
x
dia/1026 emu mol21
;
(Me3N-TEMPO)I, 2185; (Me3N-
TEMPO)PF6 2190; (n-Bu4N)2[Ni(dmit)2], 2580; (n-
Bu4N)2[Pd(dmit)2], 2539; (n-Bu4N)Br, 2203. From these
values and Pascal's constants for I2 and Br2, determined as
Me3Nz-TEMPO, ±133; [Ni(dmit)2]22, 2246; [Pd(dmit)2]2±,
2205.
19 H. Imai, T. Inabe, T. Otsuka, T. Okuno and K. Awaga, Phys. Rev.
B, 1996, 54, R6838; H. Imai, T. Otsuka, T. Naito, K. Awaga and
T. Inabe, J. Am. Chem. Soc., 1999, 121, 8098.
20 The IUPAC recommended name for the NO radical moiety is
aminoxyl. The terms nitroxide and nitroxyl are also conventionally
used.
Semiempirical computations
The INDO (intermediate neglect of differential overlap)47
method was employed for the calculation of the molecular
orbitals of Me3Nz-TEMPO. Calculations were performed
using ZINDO with the INDO/1 Hamiltonian in the CAChe
work system.48 As Me3Nz-TEMPO is paramagnetic, the
restricted open-shell Hartree±Fock (ROHF) formalism was
adopted. Metrical parameters used for the calculation were
taken from the present crystal structures.
21 (a) T. Sugimoto, S. Yamaga, M. Nakai, M. Tsuji, H. Nakatsuji
and N. Hosoito, Chem. Lett., 1993, 1817; (b) R. Kumai,
M. Matsushita, A. Izuoka and T. Sugawara, J. Am. Chem. Soc.,
1994, 116, 4523; (c) S. Nakatsuji, N. Akashi, K. Suzuki, T. Enoki,
N. Kinishita and H. Anzai, Mol. Cryst. Liq. Cryst., 1995, 268, 153;
(d) S. Nakatsuji, A. Hirai, J. Yamada, K. Suzuki, T. Enoki and
H. Anzai, Mol. Cryst. Liq. Cryst., 1997, 306, 409; (e) H. Fujiwara
and H. Kobayashi, Synth. Met., 1999, 102, 1740.
22 For a review, see S. Nakatsuji and H. Anzai, J. Mater. Chem.,
1997, 7, 2161.
23 A. E. Pullen, R.-M. Olk, S. Zeltner, E. Hoyer, K. A. Abboud and
J. R. Reynolds, Inorg. Chem., 1997, 36, 958; A. E. Pullen,
S. Zeltner, R.-M. Olk, E. Hoyer, K. A. Abboud and J. R. Reynolds,
Inorg. Chem., 1997, 36, 4163.
24 J. Lajzerowicz-Bonneteau in Spin labeling: Theory and Applica-
tions, ed. L. J. Berliner, Academic Press, New York, 1976, pp. 239±
249.
25 (a) M. Le Bars-Combe, Acta Crystallogr. Sect. B, 1982, 38, 2749;
(b) F. A. Cotton and T. R. Felthouse, Inorg. Chem., 1982, 21, 2667;
(c) M. Cygler, Can. J. Chem., 1982, 60, 2392; (d) F. Lanfranc de
Panthou, D. Luneau, J. Laugier and P. Rey, J. Am. Chem. Soc.,
1993, 115, 9095; (e) F. Iwasaki, J. H. Yoshikawa, H. Yamamoto,
K. Takada, E. Kan-Nari, M. Yasui, T. Ishida and T. Nogami,
Acta Crystallogr., Sect. B, 1999, 55, 231; (f) F. Iwasaki,
J. H. Yoshikawa, H. Yamamoto, E. Kan-Nari, K. Takada,
M. Yasui, T. Ishida and T. Nogami, Acta Crystallogr., Sect. B,
1999, 55, 1057.
26 D. Bordeaux, J. X. Boucherle, B. Delley, B. Gillon, E. Ressouche
and J. Schweizer in Magnetic Molecular Materials (NATO ASI
Series. Series E, Applied Science - vol. 198), eds. D. Gatteschi,
O. Kahn, J. S. Miller and F. Palacio, Kluwer Academic Publishers,
Dordrecht, 1991, pp. 371±383.
27 B. Bleaney and K. D. Bowers, Proc. R. Soc. London, Ser. A, 1952,
A214, 451.
28 This r value of 0.145 is in good agreement with the experimental
estimation of 16% of impurities.
29 O. Kahn, Molecular Magnetism, VCH Publishers, New York,
1993.
30 W. E. Estes, D. P. Gavel, W. E. Hat®eld and D. J. Hodgson, Inorg.
Chem., 1978, 17, 1415.
31 R. E. Coffman and G. R. Buettner, J. Phys. Chem., 1979, 83, 2387.
32 See for example: (a) J. Kondo, Solid State Phys., 1969, 23, 183;
(b) A. Niemann, U. Bossek, K. Wiegardt, C. Butzlaff,
A. X. Trautwein and B. Nuber, Angew. Chem., Int. Ed. Engl.,
1992, 31, 311.
33 S. Banerjee and G. K. Trivedl, Tetrahedron, 1992, 48, 9939.
34 G. Steimecke, H.-J. Sieler, R. Kirmse and E. Hoyer, Phosphorus
Sulfur Relat. Elem., 1979, 7, 49.
Acknowledgements
Printed circuits for conductivity measurements were kindly
supplied by the Motorola Company. We thank D. de
Montauzon for assistance in electrochemical studies. This
research was supported in part (YH and KI) by a Grant-in-Aid
for Scienti®c Research on Priority Area (B) of Molecular
Conductors and Magnets (Area No. 730/11224214) from the
Ministry of Education, Science, Sports and Culture, Japan.
References
1
H. Kobayashi, A. Kobayashi and P. Cassoux, Chem. Soc. Rev.,
2000, 29, 325.
2
P. Day, M. Kurmoo, T. Mallah, I. R. Marsden, R. H. Friend,
F. L. Platt, W. Hayes, D. Chasseau, J. Gaultier, G. Bravic and
L. Ducasse, J. Am. Chem. Soc., 1992, 114, 10722.
3
E. Coronado, L. R. Falvello, J. R. GalaÂn-MascaroÂs, C. GimeÂnez-
Saiz, C. J. GoÂmez-GarcÂõa, V. N. Lauhkin, A. PeÂrez-BenÂõtez,
C. Rovira and J. Veciana, Adv. Mater., 1997, 9, 984.
For a review, see R. Kato, Bull. Chem. Soc. Jpn., 2000, 73, 515.
M. Kurmoo, A. W. Graham, P. Day, S. J. Coles, M. B. Hursthouse,
J. L. Cau®eld, J. Singleton, P. L. Pratt, W. Hayes, L. Ducasse and
P. Guionneau, J. Am. Chem. Soc., 1995, 117, 12209.
L. Brossard, R. Clerac, C. Coulon, M. Tokumoto, T. Ziman,
D. K. Petrov, V. N. Laukhin, M. J. Naughton, A. Audouard,
F. Goze, A. Kobayashi, H. Kobayashi and P. Cassoux, Eur. Phys.
J. B, 1998, 1, 439.
4
5
6
7
8
9
E. Ojima, H. Fujiwara, K. Kato, H. Kobayashi, H. Tanaka,
A. Kobayashi, M. Tokumoto and P. Cassoux, J. Am. Chem. Soc.,
1999, 121, 5581.
P. Cassoux and L. Valade, in Inorganic Materials, eds. D. W. Bruce
and D. O'Hare, 2nd edn., John Wiley & Sons, New York, 1996,
pp. 1±64, and references therein.
P. Cassoux and J. Miller in Chemistry of Advanced Materials: An
Overview, eds. L. V. Interrante and M. J. Hampden-Smith, VCH
Publishers, New York, 1997, pp. 19±72, and references therein.
35 Organic Synthesis, John Wiley & Sons, New York, 1998, Coll.
vol. 6, pp. 499±501.
36 This compound can be also prepared from MeNH±TEMPO and
10 J. S. Miller, A. J. Epstein and W. M. Reiff, Mol. Cryst. Liq. Cryst.,
1985, 120, 27.
11 J. S. Miller and A. J. Epstein, Chem. Eng. News, 1995, 73, 30.
344
J. Mater. Chem., 2001, 11, 337±345

View Article Online
MeI: E. J. Rauckman, G. M. Rosen and W. W. Hord, Org. Prep.
Proced. Int., 1977, 9, 53.
43 A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi,
J. Appl. Crystallogr., 1993, 26, 343.
37 P. Cassoux, R. Dartiguepeyron, P.-L. Fabre and D. de Mon-
tauzon, Actual. Chim., 1985, 79; P. Cassoux, R. Dartiguepeyron,
P.-L. Fabre and D. de Montauzon, Electrochim. Acta, 1985, 11,
1485.
44 G. M. Sheldrick, Program for Crystal Structure Solution,
Germany, 1986.
45 G. M. Sheldrick, Programs for Crystal Structure Analysis (Release
97-2), GoÈttingen, Germany, 1998.
38 L. Ramaley and M. S. Krause Jr, Anal. Chem., 1969, 41, 1362.
39 CAD4 Express Software. Enraf-Nonius, Delft, The Netherlands,
1994.
46 D. M. Watkin, L. Pearce and C. K. Prout, CAMERON Ð A
Molecular Graphics Package, Chemical Crystallography Labora-
tory, University of Oxford, 1993.
40 A. C. T. North, D. C. Phillips and F. S. Mathews, Acta.
Crystallogr., Sect. A, 1968, 24, 351.
41 IPDS Software, version 2.86, Stoe & Cie, Darmstadt, Germany,
1996.
42 WINGX Main Reference, L. J. Farrugia, J. Appl. Crystallogr.,
1999, 32, 837.
47 (a) M. Zerner, G. Loew, R. Kirchner and U. Mueller-Westerhoff,
J. Am. Chem. Soc., 1980, 102, 589; (b) W. P. Anderson, D. Edwards
and M. C. Zerner, Inorg. Chem., 1986, 25, 2728.
48 The MO calculations were performed using CAChe version 3.8,
CAChe Scienti®c Inc., 1995.
J. Mater. Chem., 2001, 11, 337±345
345