Organocyanide coordination chemistry: Syntheses, structural characterisations and magnetic properties of copper (II) complexes with a di-imine/pyridine ligand

Article abstract of DOI:10.1016/j.ica.2005.02.011

The reactions between the copper (II) salts [CuXL]PF₆ (L: 2,6-[1-(2,6-diisopropylphenylimino)ethyl]pyridine) (X = Cl 1, X = Br 2) and LiTCNQ, in a DMF/water mixture, or Et₃NH(TCNQ)₂, in acetone, produced the new complexes [CuXL(TCNQ)] (X = Cl 3, X = Br 4). For both compounds, crystallographic studies have clearly evidenced the existence of dimeric complexes [{CuClL}(TCNQ)]₂ owing to π-π overlap between two adjacent TCNQ^- radical anions. Compound 1 reacted with Et ₄N(C₁₀N₇) to afford the mononuclear derivative [CuClL(C₁₀N₇)] (5), while its reaction with K ₂C₁₀N₆ produced the dinuclear complex [(CuClL)₂(C₁₀N₆)] (6). The crystal structures of complexes 5 and 6 have been determined by X-ray crystallography. Magnetic studies have revealed that compound 6 displays weak antiferromagnetic interactions between the two metal centres, conversely compounds 3 and 5 exhibit purely paramagnetic behaviours.

Full text of DOI:10.1016/j.ica.2005.02.011

Inorganica Chimica Acta 358 (2005) 2513–2522
www.elsevier.com/locate/ica
Organocyanide coordination chemistry: Syntheses, structural
characterisations and magnetic properties of copper (II)
complexes with a di-imine/pyridine ligand
a
a,
a
a
Benoıt Le Gall , Franc¸oise Conan *, Nathalie Cosquer , Jean-Michel Kerbaol ,
ˆ
a
b
b
c
Jean Sala-Pala , Marek M. Kubicki , Estelle Vigier , Carlos J. Gomez-Garcia ,
d
´
Philippe Molinie
a
´ ´
Laboratoire de Chimie, Electrochimie Moleculaires et Chimie Analytique, UMR CNRS 6521, Universite de Bretagne Occidentale,
6 Avenue le Gorgeu CS 93837, 29238 Brest Cedex 3, France
´
b
`
`
´
Laboratoire de Synthese et dꢀElectrosynthese Organometalliques, UMR CNRS 5632, Universite de Bourgogne, 21000 Dijon, France
c
´
Instituto de Ciencia Molecular (ICMol), Universite de Valencia, C/Dr. Moliner 50, 46100 Burjasot (Valencia), Spain
d
´ `
Institut des Materiaux, UMR CNRS 6502, 2 rue de la Houssiniere, BP 32229, 44322 Nantes Cedex 03, France
Received 12 January 2005; accepted 6 February 2005
Available online 14 March 2005
Abstract
The reactions between the copper (II) salts [CuXL]PF6 (L: 2,6-[1-(2,6-diisopropylphenylimino)ethyl]pyridine) (X = Cl 1, X = Br
2) and LiTCNQ, in a DMF/water mixture, or Et3NH(TCNQ)2, in acetone, produced the new complexes [CuXL(TCNQ)] (X = Cl 3,
X = Br 4). For both compounds, crystallographic studies have clearly evidenced the existence of dimeric complexes
[{CuClL}(TCNQ)]2 owing to p–p overlap between two adjacent TCNQÅꢀ radical anions. Compound 1 reacted with Et4N(C10N7)
to afford the mononuclear derivative [CuClL(C10N7)] (5), while its reaction with K2C10N6 produced the dinuclear complex
[(CuClL)2(C10N6)] (6). The crystal structures of complexes 5 and 6 have been determined by X-ray crystallography. Magnetic studies
have revealed that compound 6 displays weak antiferromagnetic interactions between the two metal centres, conversely compounds
3 and 5 exhibit purely paramagnetic behaviours.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Polynitrile ligands; Copper (II); Di-imine pyridine ligand; Crystal structures
1. Introduction
studied owing to their specific properties. Actually, a
high electron affinity, associated with particular geomet-
rical and electronic features (planarity of TCNQ, near
planarity of TCNQÅꢀ, and presence of delocalised p sys-
tems), allows p–p molecular interactions resulting in
numerous combinations with transition metal com-
plexes: charge transfer salts, simple and complex TCNQ
salts involving bonded or non-bonded [(TCNQ)n]mꢀ
units with n P m [4].
Another interest deals with the highly conjugated
cyanocarbanions or azacyanocarbanions, synthesised
from the tetracyanoethylene TCNE (Scheme 1) [5]. By
reaction with inorganic fragments, these species are able
Organic polynitrile derivatives are fascinating units
since, when associated with inorganic entities, they
may lead to a wide variety of molecular arrangements
that mostly display physical properties of significant
interest in the field of magnetism and electrical conduc-
tion [1–3].
Among these compounds, the tetracyanoquinodime-
thane (TCNQ) derivatives (Scheme 1) have long been
*
Corresponding author. Tel.: +33 2 98 01 67 08; fax: +33 2 98 01 70
01.
E-mail address: francoise.conan@univ-brest.fr (F. Conan).
0020-1693/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2005.02.011

2514
B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
2-
-
compound 1 [9], LiTCNQ and Et3NH(TCNQ)2 [10],
Et4N(C10N7) and K2(C10N6) [5], were prepared as de-
scribed in the literature. Elemental analyses were per-
formed by the ‘‘Service Central dꢀAnalyses du CNRS’’,
Vernaison, France. IR spectra were obtained with a Nic-
olet Nexus spectrometer (KBr pellets).
CN
NC
CN
CN
CN
NC
CN
CN
CN
CN
CN
-
C8N5
2-
C10N6
2.1. [(CuClL)2(TCNQ)2] (3)
2-
-
CN
CN
CN
CN
Method 1: Compound
1 [CuClL]PF6 (330 mg,
NC
CN
NC
N
N
N
0.43 mmol) was dissolved in dimethylformamide
(30 mL) to afford a brown solution. An aqueous solu-
tion of LiTCNQ (91 mg, 0.43 mmol in 30 mL of water)
was then added and the mixture was stirred for 30 mn. A
green precipitate was formed, which was then filtered off
and washed with pentane (2 · 30 mL). Recrystallisation
from acetone at ꢀ20 ꢁC gave a green crystalline product.
Yield: 337 mg (30%).
Method 2: A mixture of complex 1 [CuClL]PF6
(530 mg, 0.73 mmol) and (Et3NH)2TCNQ (370 mg,
0.73 mmol) was stirred in acetone (50 mL) at room tem-
perature for 12 h. The dark solution was filtered, con-
centrated under reduced pressure and kept at ꢀ20 ꢁC
for 7 days. After filtration, green crystals were isolated.
Yield: 526 mg (80%). Anal. Calc. for C45H47ClCu-
N7 Æ2(CH3)2CO: C, 68.0; H, 6.6; N, 10.9. Found: C,
67.6; H, 6.4; N, 10.9%. IR (KBr, cmꢀ1): 2963m,
2927m and 2868m (mCH); 2187s and 2153m (mCN);
1618w and 1581m (mCN imine); 827m (dCH).
CN
CN
CN
CN
CN
2-
-
C10N7
C10N8
Scheme 1.
to generate discrete or extended polymeric structures
owing to a great number of nearby cyano groups unable
to be related to the same metal centre; nevertheless so
far only few studies have been reported [6–8].
In the course of a recent study on the reactivity of
di-imine/pyridine ligand L towards copper derivatives,
we reported that [CuXL]+ units can be bridged with
2ꢀ
anions such as the copper (I) anion Cu2Br4 to afford
polynuclear structures [9]. We became thus interested
in studying the ability of organic polynitrile anions
to bridge such inorganic ion. In this paper, we discuss
the crystalline structures and the magnetic properties of
three new copper (II) derivatives, including the 2,6-
[1-(2,6-diisopropylphenylimino)ethyl]pyridine ligand L
(Scheme 2).
2.2. [(CuBrL)2(TCNQ)2] (4)
Compound 4 was prepared in a similar manner
according to method 2, but all samples were slightly
contamined by co-crystallisation with TCNQ. Yield:
488 mg (75%). Anal. Calc. for C45H47BrCuN7 Æ0.
3(TCNQ): C, 65.5; H, 5.5; N, 12.9. Found: C, 64.7 H,
5.9; N, 13.6%. IR (KBr, cmꢀ1): 2963m, 2928m and
2868m (mCH); 2187s and 2152m (mCN); 1617w and
1581m (mCN imine); 827m (dCH).
2. Experimental
All reactions were performed in Schlenk tubes under
a dry dioxygen-free dinitrogen atmosphere. Solvents
were distilled using standard techniques and were thor-
oughly deoxygenated before use. The starting materials,
iii
iv
[CuXL]PF6
[{CuClL}2(C10N6)]
[CuClL](C10N7)
5
1 : X = Cl, 2 : X= Br
6
ii
v
i
ii
[CuXL]2(TCNQ)2
[CuXL](TCNQ)2
N
3 :X = Cl, 4: X= Br
N
N
R
R
L
R : (i-Pr)2C6H5
i : LiTCNQ (metathesis DMF/H2O)
ii : Et3NH(TCNQ)2
iii : Et4NC10N7
iv : K2C10N6 v : TCNQ
Scheme 2.

B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
2515
2.3. [CuClL(C10N7)] (5)
2.5. X-ray structures
A stoichiometric mixture of compound 1 [CuClL]PF6
(200 mg, 0.28 mmol) and (Et4N)(C10N7) (96 mg,
0.28 mmol) was refluxed in acetone (60 mL) for 24 h.
After cooling to r.t., the orange solution was filtered
off, reduced to dryness under low pressure, extracted
with methylisobutylketone (20 mL) and then left at
low temperature (ꢀ30 ꢁC) for 4 days. Light orange crys-
tals were collected. Yield: 189 mg (75%). Anal. Calc. for
C43H43ClCuN10 ÆC6H12O: C, 64.5; H, 6.2; N, 15.6.
Found: C, 64.5; H, 6.2; N, 15.2%. IR (KBr, cmꢀ1):
2966m, 2928m and 2869m (mCH); 2217m (mCN); 1588m
(mC@N).
Single crystals of 3, 5 and 6 suitable for X-ray studies
were mounted on a Nonius-Kappa CCD diffractometer.
The crystal of 5 was of rather lower quality. The unit cell
determinations and data collections were carried out
˚
with Mo Ka radiation (k = 0.71073 A) at low tempera-
ture. The measured intensities were reduced with DENZO
program [11]. The structures were solved via direct
methods with SHELXS97 and further refined with full-ma-
trix least-squares methods (SHELXL97) based on |F2| [12].
Solvent molecules were found in all three structures: 1.2
of acetone per copper in 3, 1 of methylisobutylketone
per copper atom in 5 and 3 of this last solvent per two
copper atoms in 6. All non-hydrogen atoms, except
those of a disordered over the symmetry centre acetone
molecule of occupancy 0.2 in 3, were refined with aniso-
tropic thermal parameters. The hydrogen atoms bound
to anisotropically refined atoms were included in calcu-
lated positions and refined with a riding model, while
those of highly disordered solvent molecule were ne-
glected. Crystallographic and refinement data are gath-
ered in Table 1.
2.4. [{CuClL}2(C10N6)] (6)
A mixture of compound 1 [CuClL]PF6 (242 mg,
0.33 mmol) and K2(C10N6) (94 mg, 0.33 mmol) in ace-
tone (60 mL) was kept under stirring at room tempera-
ture for 36 h. The mixture was then filtered and the
resulting solution was reduced to dryness under low
pressure, extracted with acetone or methylisobutylke-
tone (20 mL) and then left at low temperature
(ꢀ30 ꢁC) for 4 days. Yellow crystals were collected.
Yield: 330 mg (60%). Anal. Calc. for C76H86Cl2Cu2-
N12 Æ3C3H6O: C, 66.3; H, 6.8; N, 10.9. Found: C, 65.9;
H, 7.1; N, 10.1%. IR (KBr, cmꢀ1): 2963m, 2928m and
2869m (mCH); 2185vs, 2176m and 2128w (mCN); 1707
(mC@O); 1619 and 1590m (mC@N).
2.6. Magnetic measurements
ESR spectra were run on a Bruker Elexys spectrom-
eter (X-band). The solid state magnetic susceptibility
measurements were carried out on powder samples
using either a Gouy balance Johnson Matthey at room
Table 1
Crystal data and structure refinement for compounds 3, 5 and 6
3
5
6
Empirical formula
M
C
48.60H54.20ClCuN7O1.20
C49H55ClCuN10
899.02
110
O
C94H122Cl2Cu2N12O3
1666.02
120
854.58
200(2)
Temperature (K)
Crystal system, space group
ꢀ
ꢀ
triclinic, P1
9.1906(2)
13.7655(4)
triclinic, P1
monoclinic, P21/n
16.7911(4)
32.0245(8)
˚
a (A)
10.0001(4)
11.2441(5)
21.6821(11)
85.951(2)
82.782(2)
88.999(3)
2412.53(19)
2
˚
b (A)
˚
c (A)
19.1629(7)
17.0989(4)
a (ꢁ)
b (ꢁ)
c (ꢁ)
95.8700(10)
92.2760(10)
103.589(2)
92.178(2)
3
˚
V (A )
2339.04(12)
2
1.213
9187.9(4)
4
1.204
Z
Dcalc (Mg mꢀ3
)
1.238
0.554
Absorption coefficient (mmꢀ1
F(000)
)
0.567
901
0.575
3544
946
0.25 · 0.20 · 0.10
2.72–30.66
Crystal size (mmꢀ3
)
0.25 · 0.20 · 0.03
2.43–30.46
0.25 · 0.20 · 0.05
1.018–30.508
19, ꢀ34 + 29, 19
19215/11974/7929 (0.0452)
11974/0/1018
0.0592, 0.1078
0.1341, 0.1583
1.038
h range (ꢁ) for collection
hkl/ranges
ꢀ13 + 12, 19, ꢀ23 + 26
17841/11894 (0.1020)
11894/0/529
R1 = 0.0740, wR2 = 0.1327
R1 = 0.1620, wR2 = 0.1669
1.023
ꢀ14 + 13, 15, ꢀ30 + 28
16867/11571/7344 (0.404)
11571/0/560
Reflections collected/unique (Rint
Data/restraints/parameters
Final R indices [I > 2r(I)]
R indices (all data)
Goodness-of-fit on F2
˚
Largest difference peak, hole (e A
)
0.0599, 0.1787
0.1106, 0.2359
0.795
ꢀ3
)
0.766, ꢀ0.465
0.546, ꢀ0.792
1.096, ꢀ0.591

2516
B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
temperature or commercial SQUID magnetometers
from Quantum Design from 2 to 300 K. Magnetic stud-
ies for compound 6 were carried out on powder samples
at 0.1 T after field cooling, in the temperature range 2–
300 K, with a MPMS-XL SQUID magnetometer from
Quantum Design. The susceptibility was corrected for
the sample holder and the diamagnetic contributions
of all atoms Isothermal magnetisation measurements
were performed at K with magnetic fields of up to 5 T.
The susceptibilities were corrected for the intrinsic
diamagnetism of the sample container.
wave numbers found for 3 and 4 (see Section 2) clearly
attest the presence of TCNQꢀ anions in these
compounds.
The IR spectra of complexes 5 and 6 also exhibit
absorption bands in the range 2220–2120 cmꢀ1 assigned
to mCN vibrations (see Section 2). The positions of these
bands are slightly shifted with respect to thoseꢀo1f the
corresponding anion (2218 and 2202 cm
in
(Et4N)C10N7; [7] 2191, 2174 and 2121 cmꢀ1 in
K2C10N6 [8]), suggesting the coordination of the poly-
nitrile anions or at least the existence of some bonding
interactions between these anions and the inorganic
fragment.
3. Results and discussion
3.2. Crystal structures
3.1. Syntheses and general characterisation
Compounds 3 and 5 crystallise in the triclinic group
ꢀ
The reactions of the salts [CuXL]PF6 (X = Cl 1,
X = Br 2) with LiTCNQ in DMF/water mixtures at
room temperature afforded after recrystallisation in ace-
tone the new 1:1 derivatives [CuXL(TCNQ)] (X = Cl 3,
X = Br 4) in low yields. Surprisingly, similar reactions
completed with a slight excess of the TCNQ mixed-
valence salt Et3NH(TCNQ)2 in acetone at room temper-
ature provided compounds 3 and 4 in better yields and
not, as expected, the 1:2 TCNQ mixed-valence salts
[CuXL](TCNQ)2 (Scheme 2).
P1; each unit cell contains two asymmetric units consis-
tent with one copper (II) cation [CuClL]+, one nitrile en-
ꢀ
tity (TCNQ in 3, C10N7 in 5) and molecules of solvent
of crystallisation (1.2 molecule of acetone in 3, 1 mole-
cule of methylisobutylketone in 5). Compound 6 crystal-
lises in the monoclinic space group P21/n; the
asymmetric unit contains one dinuclear unit
[{CuClL}2(C10N6)] and three molecules of solvent
(methylisobutylketone).
Examination of the reactions of the azacyanocarba-
nions (Scheme 1) towards compound 1 evidenced their
reactivity, but several difficulties were encountered for
separation and purification precluding the full-charac-
terisatꢀion of new derivatives, except in the case of the
3.2.1. Copper coordination and di-imine/pyridine ligand
For the three compounds, within the [CuClL] unit,
the copper atom is surrounded by one chlorine atom
and three nitrogen atoms from the di-imine/pyridine
terdentate ligand defining an essentially square planar
geometry (Figs. 1–3). However, some distortions of the
co-ordination polyhedron, with respect to the square,
are observed (Table 2): (i) the bond angle values around
the metal atom deviate from the ideal values (N–Cu–N
from 77.86(18)ꢁ to 78.96(10)ꢁ, N–Cu–X from
98.23(13)ꢁ to 102.01(12)ꢁ), due to the usual bite of the
terdentate ligand, and (ii) the N3Cl tetra-atomic systems
are not exactly planar but exhibit slight butterfly-like
deformations. The Cu–N metal–ligand bond lengths
are usual with Cu–Nimine bond lengths (from 2.066(3)
2ꢀ
C10N7 and C10N6 anions (Scheme 2).
Hence, reaction of 1 with Et4N(C10N7) allowed the
formation of the well-defined derivative 5 for which
the elemental analysis is in good agreement with the for-
mula [CuClL(C10N7)]. On the contrary, the reaction of
K2C10N6 with compound 1 led to the new compound
6 for which the elemental analyses support a 2:1 cop-
per:organic anion ratio, and consequently the
Cu2Cl2L2(C10N6) formula.
The IR spectra of compounds 3–6 clearly exhibit the
usual absorption bands of the corresponding polynitrile
ligand and the coordinated di-imine/pyridine ligand. For
the latter, and as previously noticed for parent com-
pounds, the mCNimine position attests that the di-imine/
pyridine ligand acts according to its classical terdentate
coordination mode and not with its unusual bidentate
mode [13].
Concerning the TCNQ unit of 3 and 4, previous stud-
ies show that IR spectroscopy can be used as a diagnos-
tic tool to estimate the charge of this unit. Significant
decreases of the wave numbers are observed for the
mCN and dCH vibrations when passing from TCNQ
(2225 and 850 cmꢀ1, respectively) to TCNQꢀ (2180,
2150 and 825 cmꢀ1, respectively) [4]. The corresponding
˚
to 2.097(4) A) somewhat longer than the Cu–Npy bond
˚
lengths (from 1.930(4) to 1.943(3) A) and in good agree-
ment with those previously reported. The Cu–Cl bond
lengths distances also agree with previous literature val-
ues [9,14].
In fact, in the three compounds, a short contact be-
tween the copper centre and one nitrogen atom of a
polynitrile unit allows the square-planar copper (II) to
extend its co-ordination number up to five. The nitrogen
˚
atom occupies the apical position at 2.463(3) A from the
˚
copper atom in 3 and at 2.335(3) A in 5. These bond
lengths traduce moderate interactions with respect to
those previously observed in quite a similar com-
˚
pound [Cu2L1(TCNQ)2] (2.378(8) A) [15] or in the pseu-

B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
2517
C25
C24
C26
C30
C29
C28
N4'
C27
C7
C9
C33
C23
N5'
C41'
C34'
C22
C42'
C31
C32
C1
C40'
C39'
N3
C35'
N2'
C36'
N7
N6
C2
Cu
Cl
C38'
N1
C43'
C37'
C45'
C17
C44'
N6'
Cu'
N1'
Cl'
C16
C11
C3
N7'
C18
C5
N2
C6
C10
C4
N3'
C20
C19
C12
C15
C8
N4
C13
N5
C14
C21
Fig. 1. ORTEP representation of the asymmetric unit and atoms labelling in compound 3.
do-octahedral derivative [Cu(L2)2(TCNQ)2] (Cu–N
gand between two metallic fragments. The intradimer
˚
copper–copper distance is quite long (10.898 A).
˚
2.442(5) A); [16] these interactions are also clearly evi-
denced by the elevation of the copper atom above the
˚
N3Cl mean plane toward the apex (0.158(2) A in 3 and
At last, concerning the di-imine pyridine ligands, a
careful examination of the bond lengths, including those
of the Cpy–Cimine and even the Cimine–CH3 bonds (Table
2), clearly indicates a high p electronic delocalisation
within these ligands.
˚
0.145(2) A in 5).
In 6, it is worthy of note that the apical Cu–N bond
lengths are much shorter (Cuꢁ ꢁ ꢁN 2.197(5) and
˚
2ꢀ
2.180(5) A), indicating that in this molecular complex
the C10N6 anion acts as a strongly bonded bridging li-
3.2.2. Polynitrile unit structures
In compound 3, the TCNQ species exhibits its usual
features: the overall shape and size parameters are close
to those reported for RbTCNQ (Table 3) [17,18]. The
estimated charge q = ꢀ0.96, calculated from the Kis-
tenmacherꢀs relation, corroborates in this compound
the existence of anionic TCNQ units [19].
N7
N8
C38
C41
N10
C43
C39
C37
N6
C40
Furthermore, the TCNQ entities form isolated cen-
trosymmetric dimeric anions [(TCNQ)2]2ꢀ according to
a usual ring–ring (R–R) stack characterised by a trans-
C35
N4
C36
C42
N9
˚
versal slipping of ca. 1.2 A and an interplanar distance
C30
C21
C34
N5
˚
of 3.26 A (Fig. 1). This latter distance is slightly longer
C16
C15
C29
C4
than the distance observed for [{Fe(Cp)2}2](TCNQ)2
˚
C28
C3
˚
C14
(3.147 A) [20] or for [Cu2L1(TCNQ)2] (3.15 A) [15] but
in the same range as distances observed in the com-
C2
C5
C20
C1
C7
C6
˚
N1
Cu
Cl
pounds [M(L3)2](TCNQ)2 (M@Ni, Cu) (3.258 A) [21]
C23
C9
˚
and {[MoO(L4)3](TCNQ)}2 (3.28 A) [22]. The structure
C24
C25
C10
N3
N2
C22
of 3 can therefore be described as an arrangement of
[{CuClL}(TCNQ)]2 dimers separated, in the crystal, by
solvent molecules (acetone).
In addition, the powder X-ray diffraction patterns of
3 and 4 are similar, indicating that both compounds are
isostructural.
In compound 5, the polynitrile ligand is built on a
central nitrogen atom N(4), for which NC bond lengths
˚
(1.338(5) and 1.341(4) A) and CNC angle value
C8
C27
C13
C11
C26
C31
C12
C17
C33
C18
C19
C32
Fig. 2. ORTEP representation of the asymmetric unit and atoms
labelling in compound 5.

2518
B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
C59
C30
C53
C24
C23
C25
C26
C33
C58
C54
N6
C28
C29
C57
C52
C50
C55
C21
C40
C36
C56
C27 C22
C31
C51
C62
C20
C5
C49
C44
N3
N11
C41
N8
N5
N7
C32
C35
C60
C45
C4
C39
C38
C34
C46
N4
C3
Cl1
C42
Cl2
C61
N1
Cu2
C37
C43
Cu1
C72
C71
C48
N12
N10
C2
C47
C1
N9
C18
N2
C8
C17
C6
C63 C65
C66
C15
C14
C73
C13
C74
C70
C69
C75
C64
C19
C67
C9
C7
C16
C10
C76
C68
C12
C11
Fig. 3. ORTEP representation of the asymmetric unit and atoms labelling in compound 6.
Table 2
˚
Bond distances (A) and angles (ꢁ) in the (CuClL) units for compounds 3, 5 and 6
3
5
6
Cu co-ordination
a
b
c
1.939(2)
2.066(3)
2.1705(9)
78.96(10)
101.24(8)
1.943(3)
2.081(3)
2.1953(9)
78.48(11)
101.02(8)
1.935(4)
2.097(4)
1.930(4)
2.093(4)
2.072(2)
2.079(3)
2.087(4)
2.089(4)
2.1927(13)
78.47(16)
100.06(12)
2.1956(14)
78.77(19)
102.01(12)
a
b
78.65(10)
99.88(7)
78.43(11)
100.70(8)
78.23(16)
99.74(12)
77.86(18)
98.23(13)
L ligand
d
e
f
1.340(4)
1.485(4)
1.294(4)
1.487(4)
1.436(4)
118.1(2)
112.1(3)
1.325(4)
1.498(4)
1.290(4)
1.484(5)
1.452(4)
118.48(19)
112.4(3)
1.337(4)
1.495(6)
1.295(4)
1.496(4)
1.441(4)
119.3(2)
114.7(2)
C
1.332(4)
1.485(5)
1.285(4)
1.493(4)
1.454(4)
118.4(2)
114.3(2)
1.336(6)
1.489(7)
1.293(6)
1.492(7)
1.453(6)
119.0(3)
114.7(3)
1.336(6)
1.477(7)
1.293(6)
1.483(7)
1.452(6)
118.4(3)
113.9(3)
1.333(7)
1.490(8)
1.303(7)
1.485(8)
1.442(7)
118.9(4)
113.5(4)
1.343(7)
1.476(8)
1.284(6)
1.489(7)
1.454(7)
119.5(4)
115.3(4)
g
h
v
d
C
g
C
d
C
e
N1
d
c
e
f
a
g
a
N2
N3
Cu
c
b
h
b
h
C
C
Cl
(123.4(3)ꢁ) deeply argue in favour of a sp2 hybridisation,
and two essentially planar (CN)2C@C(CN) sub-units
(Fig. 2, Table 4) [23]. However, the C10N7 unit, which
usually is essentially planar, deviates significantly from
the planarity since a torsion angle of ca. 33ꢁ is observed
between the two wings. It is likely that this torsion sig-
nificantly reduces the repulsion between the two nitrile
groups of the central (NC)CNC(CN) unit and allows a
lowering of the central CNC angle (123.4ꢁ versus
128.3ꢁ and 129.2ꢁ for (Et4N)(C10N7) and Ag(C10N7),
respectively) [7]. This deviation from the planarity does
not preclude an important p-electronic delocalisation as
shown by the carbon–carbon bond lengths, which are all
˚
in the range 1.386(5)–1.464(6) A (Table 3).
ꢀ
Conversely in compound 5, the monodentate C10N7
ligands do not form any stack and in fact each complex
must be considered as an isolated entity. Actually, the
˚
distance between two mean-plane anions is about 4 A
long, the shortest distance between the central nitrogen
˚
atoms (N4) is 6.803 A long for anions belonging to
˚
two adjacent cells, 8.320 A long for anions within a unit
cell; these observations preclude any short contacts.

B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
2519
Table 3
Comparison of some bond lengths (A) and estimation of the charge in the TCNQ moiety
˚
Compound
Reference
a
b
c
d
e
b–c
c–d
c/(b + d) (a + c)/(b + d) qa
3
this work 1.370(5) 1.420(5) 1.417(5) 1.418(5) 1.151(5) 0.003 ꢀ0.001 0.499
0.982
0.984
0.942
ꢀ0.96
ꢀ1
RbTCNQ
TCNQ
[28]
[17]
1.373(1) 1.423(3) 1.420(1) 1.416(8) 1.153(7) 0.003
0.004 0.500
1.346(3) 1.448(4) 1.374(3) 1.441(4) 1.140(1) 0.074 ꢀ0.067 0.476
0
c
b
d
a
e
d
a
Estimation of the charge q of the TCNQ unit using Kistenmacher relation [19] q = A[c/(b + d)] + B with A = ꢀ 41.667 and B = 19.833.
Table 4
Selected bond distances (A) and bond angles (ꢁ) of the C10N7 units in compound 5, (Et4N)(C10N7) and Ag(C10N7)
ꢀ
˚
Compound 5
(Et4N)(C10N7) [7]
Ag(C10N7) [7]
a
1.338(5), 1.341(4)
1.386(5), 1.389(5)
1.416(5)–1.464(6)
1.141(5)–1.154(5)
123.4(3)
1.317(4), 1.334(4)
1.361(4), 1.382(5)
1.419(6)–1.460(5)
1.134(4)–1.137(4)
128.3(3)
1.33(1), 1.34(1)
1.36(1), 1.37(1)
1.43(2)–1.47(2)
1.13(1)–1.14(1)
129.2(9)
b
ca
da
a
b
c
d
123.6(3)
114.2(3)
122.1(3)
124.8(4)
116.2(4)
119.0(3)
124(1)
118(1)
118(1)
CN
β CN
γ
α
N
NC
CN
b
a
δ
c
CN
d
N
a
Lower to upper values.
In compound 6, as clearly described above, the
clude an intense electronic delocalisation all over the
ligand, as indicated by all CC bond lengths in the range
˚
1.408(7)–1.435(8) A (Table 5).
Finally, for the three compounds, the cyanide Csp–N
2ꢀ
C10N6 ligand, despite its six nitrile groups potentially
bridging, shows only a l2-bridging coordination mode
(Fig. 3). The central fragment (C34–C37) is essentially
˚
˚
bond lengths, falling in the range 1.140–1.162 A, corre-
planar within a maximum deviation of 0.02 A from
the corresponding mean plane and, associated with
other structural features (for each of these four C atoms,
sum of the three bond angles reaches 360ꢁ), is in agree-
ment with an sp2 hybridisation of the four central car-
spond to typical values for Csp–Nsp bonds in such poly-
nitrile compounds [20].
3.3. Magnetic measurements
2ꢀ
bon atoms. However, the C10N6 ligand is not planar
but has a propeller-shaped geometry of approximate
D3 symmetry, since the axis passing through the central
carbon atom (C35) appears as a pseudo threefold axis.
The tilt angles between the mean central plane and the
different C(CN)2 wing planes have an average value of
28.9ꢁ. This propeller-shaped geometry, previously re-
For all the compounds, EPR spectra, recorded on
powdered and on solution samples at low and room
temperatures, display the usual features for copper (II)
cations in an axial environment (Table 6). Actually,
the solution spectra show the axial pattern characteristic
of square pyramidal copper (II). The parallel signals ap-
pear as quartets owing to the hyperfine coupling with
the copper (I = 3/2) nucleus, the coupling constant val-
ues (Ai) being in the same order than those usually ob-
served for copper (II) in similar environment.
2ꢀ
ported for the C10N6 unit, decreases the CNꢁ ꢁ ꢁNC ste-
ric interactions between the nitrile groups of adjacent
C(CN)2 wings (ca. 50 kJ molꢀ1) without a strong reduc-
tion (less than 9%) of the p-overlap [24]; it does not pre-

2520
B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
Table 5
2ꢀ
˚
Selected bond lengths (A) and bond angles (ꢁ) of the C10N6 units in compound 6, Ca(C10N6)(H2O)6 and Cu(C10N6)(H2O)2
Compound 6
Ca(C10N6)(H2O)6 [24]
Cu(C10N6)(H2O)2 [8]
ac
1.416(6)–1.429(7)
1.408(7)–1.435(8)
1.140(6)–1.162(6)
a2.180(5), 2.197(5)
119.1(4)–120.7(4)
121.4(4)–123.1(5)
114.2(4)–117.0(4)
28.9
1.424
1.416, 1.426
1.15, 1.17
1.410(11)–1.425(6)
1.391(7)–1.409(8)
1.139(7)–1.141(7)
b1.977(5), 2.052(6)
119.9(3), 120.3(7)
122.3(5)–124.1(5)
113.5(5), 114.2(7)
26.9
bc
cc
Mꢁ ꢁ ꢁN
ac
120
bc
121.9, 122.6
115.5
24
cc
Tilt angle (average)
N
N
g
C
C
d
a
N
C
N
C
a
b
b
C
N
C
N
c
a
(4 + 1) coordination.
(4 + 2) coordination.
Lower to upper values.
b
c
Furthermore, for 3 and 4, a supplementary signal at
g = 2.004 is undoubtedly ascribable to the organic radi-
cal TCNQÅꢀ, traducing the dissociation of both com-
plexes in solution. Conversely, on their powdered
sample spectra, only one anisotropic signal, characteris-
tic of an axial copper (II) atom, is observed. It is worthy
to note that the narrow signal attributable to the organic
radical TCNQÅꢀ is no more detected as it is often ob-
served in the case of similar dimerised [(TCNQ)2]2ꢀ salts
[25]. This fully agrees with the magnetic measurements.
At last, considering compound 6, the room and low
temperature solution spectra exhibit similar patterns,
suggesting for this derivative a discrete dimeric molecu-
lar structure even in solution.
The solid state magnetic susceptibility measurements
were performed on compounds 3, 5 and 6 within the 2–
300 K temperature range.
The magnetic susceptibilities for compounds 3 and 5
obey a Curie law whose characteristics are given in
Table 5. For 3, the magnetic moment (2.53 lB) is in
good agreement with two independent paramagnetic
copper (II) cations bridged with a diamagnetic
[(TCNQ)2]2ꢀ dimer (S = 0) according to the structural
features, while for compound 5 the magnetic moment
(1.81 lB) corresponds to one isolated copper (II) ion.
Compound 6 shows at room temperature a vmT
product of 0.76 emu K molꢀ1 that remains constant
when decreasing temperature down to ca. 50 K. Below
this temperature, the vmT product slowly decreases to
reach a value of 0.70 emu K molꢀ1 at 2 K (Fig. 4).
The room temperature vmT value corresponds to the
expected value for two copper (II) ions:
0.75 emu K molꢀ1 for g = 2, and the smooth decrease
at low temperatures indicates the presence of very weak
antiferromagnetic interactions between the two Cu(II)
centres.
From the structural data that show a dimeric struc-
ture in this compound, we have fitted the magnetic
behaviour to the classical dimer model:
vmT ¼ ð3 ꢂ g2=8Þ ꢂ ð1=ð3 þ expðꢀ2 ꢂ J=kB ꢂ TÞÞÞ;
Table 6
Magnetic and ESR data for complexes 3, 5 and 6
C (emu3 K molꢀ1
)
J (K)
l (lB)
ga
gib Aic
b
g^
Compound
3
5
6
a
0.80
0.41
0.76
0
2.53
1.81
2.74
2.057
2.058
2.093
2.153 (129)
2.235 (130)
2.239 (127)
2.057
2.061
2.081
0
ꢀ0.267
Powder ESR spectra at 290 K.
Frozen solution samples (acetone, 140 K).
In 10ꢀ4 T.
b
c

B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
2521
4. Concluding remarks
1.0
0.9
0.8
0.7
0.6
0.5
These results clearly emphasise that new examples,
where the polynitrile anions act as good ligands bring
about magnetic interactions between metallic centres.
Actually, despite a quite long Cu–Cu distance, the dinu-
clear complex [(CuClL)2(C10N6)] exhibits weak antifer-
romagnetic interactions. This is not surprising since
strong magnetic couplings have recently been predicted
between paramagnetic metal cations through very long
bridging ligands such as dicyanoamidobenzene or ana-
˚
logues at distances as long as 25 A [27]. In the aim to
0
50
100
150
200
250
300
synthesise compounds exhibiting stronger magnetic
interactions, further works are in progress with other
di-imine/pyridine transition metal complexes.
T (K)
Fig. 4. Thermal variation of the vmT product of 6. Solid line shows the
best fit to a dimer model (see text).
where J is the intradimer exchange coupling. A very
good fit over the whole temperature range (2–300 K) is
obtained with g = 2.0054(3) and J/kB = ꢀ0.267(3) K
(ꢀ0.185(2) cmꢀ1) (solid line in Fig. 4). Note that this g
value is slightly smaller than the one obtained from
the EPR spectra, suggesting the presence of a very small
amount of diamagnetic impurity in the sample that re-
duces the molar susceptibility but does not affect the
EPR spectra. As expected, the intradimer magnetic cou-
pling through the C10N6 bridge is antiferromagnetic and
very weak with a coupling constant similar to those
obtained on other similar dimers connected through this
ligand [26].
5. Supplementary material
CCDC Nos. 256226 (for 3), 256227 (for 5) and
256228 (for 6) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained
free of charge at www.ccdcca.ac.uk/conts/retriev-
ing.html or from the Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, UK;
fax: +44-1223-336-033; e-mail: deposit@ccdc.cam.ac.uk.
Acknowledgements
A confirmation of the spin ground state and of the
presence of a weak intradimer antiferromagnetic inter-
action comes from the isothermal magnetisation at low
temperature (Fig. 5) that can be very well reproduced
with the sum of two Brillouin functions for a S = 1/2
spin ground state with a very weak antiferromagnetic
interaction (solid line in Fig. 5).
This work was supported by the CNRS (UMR 6521)
`
and Ministere de la Recherche. MR is acknowledged for
doctoral fellowship (BLG). Dr. Y. Le Mest is thanked
for fruitful discussions.
References
[1] H. Endres, in: J.S. Miller (Ed.), Extended Linear Chain Com-
pounds, 263, Plenum Press, New York, 1983.
`
[2] E. Coronado, P. Delhaes, D. Gatteschi, J.S. Miller, Molecular
2.0
1.5
1.0
0.5
0.0
Magnetism: from Molecular Assemblies to the Devices, Kluwer
Academic Publishers, Dordrecht, 1996.
[3] J.S. Miller, Inorg. Chem. 39 (2000) 4392.
[4] W. Kaim, M. Moscherosch, Coord. Chem. Rev. 129 (1994) 157.
[5] W.J. Middleton, E.L. Little, D.D. Coffman, V.A. Engelhardt, J.
Am. Chem. Soc. 80 (1958) 2795.
[6] M. Decoster, F. Conan, M.M. Kubicki, Y. Le Mest, P. Richard,
J. Sala-Pala, L. Toupet, J. Chem. Soc., Perkin Trans. II (1997)
265–271.
[7] M. Decoster, J.E. Guerchais, Y. Le Mest, J. Sala-Pala, S. Triki,
L. Toupet, Polyhedron 15 (1996) 195.
´
[8] S. Triki, J. Sala-Pala, M. Decoster, P. Molinie, L. Toupet,
0.0
0.5
1.0
H/T (T.K-1)
1.5
2.0
2.5
Angew. Chem., Int. Ed. Engl. 38 (1999) 113.
[9] B. Le Gall, F. Conan, N. Cosquer, J.M. Kerbaol, M.M. Kubicki,
E. Vigier, Y. Le Mest, J. Sala-Pala, Inorg. Chim. Acta 324 (2001)
300.
Fig. 5. Isothermal magnetisation of 6 at 2 K. Solid line shows the best
fit to the sum of two S = 1/2 Brillouin functions with a weak
antiferromagnetic interaction (see text).
[10] L.R. Melby, F.J. Harder, W.R. Hertler, W. Mahler, F.E. Benson,
W.E. Mochel, J. Am. Chem. Soc. 84 (1962) 3374.

2522
B.L. Gall et al. / Inorganica Chimica Acta 358 (2005) 2513–2522
[11] Z. Otwinowski, W. Minor, Methods Enzymol. 276 (1997) 307.
[12] G.M. Sheldrick, SHELXL97, University of Go¨ttingen, Go¨ttingen,
Germany, 1997.
[21] M.T. Azcondo, L. Ballester, S. Golhen, A. Gutierrez, L.
Ouahab, S. Yartsev, P. Delhaes, J. Mater. Chem. 9 (1999)
1237.
[13] L. Douce, A. El-Ghayoury, A. Skoulios, R. Ziessel, Chem.
Commun. (1999) 2033.
[22] M. Decoster, F. Conan, J.E. Guerchais, Y. Le Mest, J. Sala-Pala,
´
A. Leblanc, P. Molinie, E. Faulques, L. Toupet, New. J. Chem.
21 (1997) 215.
[14] B. Le Gall, F. Conan, J.M. Kerbaol, M.M. Kubicki, E. Vigier,
Y.L. Mest, J. Sala-Pala, Inorg. Chim. Acta 336 (2002) 87.
[15] P. Lacroix, O. Kahn, A. Gleizes, L. Valade, P. Cassoux, New J.
Chem. 8 (1984) 643.
[23] F.H. Allen, O. Kennard, D.G. Watson, L. Brammer, A.G. Orpen,
R. Taylor, J. Chem. Soc., Perkin Trans. II S1 (1987).
[24] D.A. Bekoe, P.K. Gantzel, N.K. Trueblood, Acta Crystallogr. 22
(1967) 657.
[16] J.P. Cornelissen, J.H.v. Diemen, L.R. Groeneveld, J.G. Haasnoot,
A.L. Spek, J. Reedijk, Inorg. Chem. 31 (1992) 198.
[17] A. Hoekstra, T. Spoelder, Acta Crystallogr. B 28 (1972) 14.
[18] S. Flandrois, D. Chasseau, Acta Crystallogr. B 33 (1977) 2744.
[19] T.J. Kistenmacher, T.J. Emge, A.N. Bloch, D.O. Cowan, Acta
Crystallogr. B 38 (1982) 1193.
[25] L. Ballester, A.M. Gil, A. Gutierrez, M.F. Perpin˜an, M.T.
Azcondo, A.E. Sanchez, U. Amador, J. Campo, F. Palacio,
Inorg. Chem. 36 (1997) 5291.
´
[26] F. Thetiot, Ph.D. Thesis, Brest, 2004.
[27] E. Ruiz, A. Rogriguez-Fortea, S. Alvarez, Inorg. Chem. 42 (2003)
4881.
[20] A.H. Reis, L.D. Preston, J.M. Williams, S.W. Peterson, G.A.
Candela, L.J. Swartzendruber, J.S. Miller, J. Am. Chem. Soc. 101
(1979) 2756.
[28] R.E. Long, R.A. Sparks, K.N. Trueblood, Acta Crystallogr. 18
(1965) 932.