Structural diversity in iron(ii) complexes of 2,6-di(pyrazol-1-yl)pyridine and 2,6-di(3-methylpyrazol-1-yl)pyridine

Article abstract of DOI:10.1039/b510370c

The syntheses, magnetochemistry and crystallography of [Fe(L ¹)₂]I_0.5[I₃]_1.5 (1), [Fe(L¹)₂][Co(C₂B₉H ₁₁)₂]₂ (2) and [Fe(L²) ₂][SbF₆]₂ (3) (L¹ = 2,6-di(pyrazol-1-yl)pyridine; L² = 2,6-di(3-methylpyrazol-1-yl) pyridine) are described. Compounds 1 and 3 are high-spin between 5-300 K. For 1, this reflects a novel variation of an angular Jahn-Teller distortion at the iron centre, which traps the molecule in its high-spin state. No such distortion is present in 3; rather, the high-spin nature of this compound may reflect ligand conformational strain caused by an intermolecular steric contact in the crystal lattice. Compound 2 exhibits a gradual high → low spin transition upon cooling with T 1/2 = 318 ± 3 K, that is only 50% complete. This reflects the presence of two distinct, equally populated iron environments in the solid. One of these unique iron centres adopts the same angular structural distortion shown by 1 and so is trapped in its high-spin state, while the other, which undergoes the spin-crossover, has a more regular coordination geometry. In contrast with 3, the solvated salts [Fe(L²)₂][BF ₄]₂·4CH₃CN and[Fe(L²) ₂][ClO₄]₂·(CH₃) ₂CO both undergo gradual thermal spin-transitions centred at 175 ± 3 K. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b510370c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Structural diversity in iron(II) complexes of 2,6-di(pyrazol-1-yl)pyridine and
2,6-di(3-methylpyrazol-1-yl)pyridine†
Je´roˆme Elha¨ık, Colin A. Kilner and Malcolm A. Halcrow*
Received 20th July 2005, Accepted 14th October 2005
First published as an Advance Article on the web 1st November 2005
DOI: 10.1039/b510370c
The syntheses, magnetochemistry and crystallography of [Fe(L1)2]I0.5[I3]1.5 (1), [Fe(L1)2][Co(C2B9H11)2]2
(2) and [Fe(L2)2][SbF6]2 (3) (L1 = 2,6-di(pyrazol-1-yl)pyridine; L2 = 2,6-di(3-methylpyrazol-1-yl)-
pyridine) are described. Compounds 1 and 3 are high-spin between 5–300 K. For 1, this reflects a novel
variation of an angular Jahn–Teller distortion at the iron centre, which traps the molecule in its
high-spin state. No such distortion is present in 3; rather, the high-spin nature of this compound may
reflect ligand conformational strain caused by an intermolecular steric contact in the crystal lattice.
Compound 2 exhibits a gradual high → low spin transition upon cooling with T 1 = 318 3 K, that is
2
only 50% complete. This reflects the presence of two distinct, equally populated iron environments in
the solid. One of these unique iron centres adopts the same angular structural distortion shown by 1
and so is trapped in its high-spin state, while the other, which undergoes the spin-crossover, has a more
regular coordination geometry. In contrast with 3, the solvated salts [Fe(L2)2][BF4]2·4CH3CN
and[Fe(L2)2][ClO4]2·(CH3)2CO both undergo gradual thermal spin-transitions centred at 175 3 K.
‘rotation’ component of the structural distortion. Undistorted
Introduction
high-spin complexes of this type exhibit a zero-field splitting
We are continuing to study the spin-transition behaviour1 of
parameter |D| ≈ 12 cm−1, while compounds with φ < 180◦
iron(II) complexes of 2,6-di(pyrazol-1-yl)pyridine2–10 and 2,6-
commonly show 3 ≤ |D| ≤ 8 cm−1.10
di(pyrazol-1-yl)pyrazine2,8,11–14 ligands. The parent compound in
this series is [Fe(L1)2]2+, whose BF4− salt undergoes an abrupt spin-
transition at 261 K.3–5,8 In contrast the ClO4−, PF6 and SbF6
−
−
salts of this material do not undergo spin-crossover, remaining
fully high-spin between 5–300 K. That reflects their adoption of
an unusual angular structural distortion, from the idealised D2d
symmetry expected with this ligand set towards C2 symmetry.4,10
This is a result of Jahn–Teller splitting of the 5E high-spin ground
state (in the D2d point group), and is promoted in compounds of
tridentate ligands with narrow bite angles like L1.4 The distortion
can be characterised by two angles: a rotation of one ligand about
the Fe ion, so that the trans-N{pyridine}–Fe–N{pyridine} angle
(φ in Scheme 1) < 180◦; and, a twist of the plane of one ligand
relative to the other about the N{pyridine}–Fe–N{pyridine} vec-
tor (i.e h < 90◦, Scheme 1). These two components usually occur
together in a distorted [Fe(L1)2]2+ salt and related compounds,
typically yielding φ ≈ 155◦ and h ≈ 60◦.9,10,14 However one example
is known that exhibits the ‘twist’ component of the distortion only,
possibly for steric reasons.10 Comparison of all these compounds
has suggested that the zero-field splitting of the high-spin species
can be used as a qualitative indicator of the presence of the
Scheme 1 The angles referred to in the discussion of the Jahn–Teller
distortion in high-spin [Fe(L1)2]2+
.
The reluctance of [Fe(L1)2]2+ salts to undergo thermal spin-
transitions contrasts with the [FeL2]2+ complex of 2,6-di(pyrazol-
3-yl)pyridine, which is a regioisomer of L1.2 A large number of hy-
drated and anhydrous salts of the latter complex dication have been
isolated, many of which undergo thermal spin-transitions exhibit-
ing novel solid state physics.15 For this reason, we have continued
to prepare new salts of [Fe(L1)2]2+, which might also show novel
spin-crossover behaviour. We report here the synthesis, and crys-
tallographic and magnetochemical characterisation of two such
compounds: [Fe(L1)2]I0.5[I3]1.5 (1) and [Fe(L1)2][Co(C2B9H11)2]2 (2).
We also describe the first complex salts of the related ligand L2,
including [Fe(L2)2][SbF6]2 (3). Taken together, 1–3 further extend
the unusual structural and magnetochemical complexity observed
in this metal/ligand system.
School of Chemistry, University of Leeds, Woodhouse Lane, Leeds, UK LS2
9JT. E-mail: M.A.Halcrow@chem.leeds.ac.uk
† Electronic supplementary information (ESI) available: Experimen-
tal details, figure and tables of bond lengths and angles from the
crystal structures of the solvate salts [Fe(L2)2][BF4]2·4CH3CN and
[Fe(L2)2][ClO4]2·(CH3)2CO at four temperatures; a plot of the variation of
Fe–N bond lengths in these compounds with temperature; and, a table of
C–H · · · I hydrogen bonding parameters in 1. See DOI: 10.1039/b510370c
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 823–830 | 823
©

View Article Online
Results and discussion
No complexes of L2 have been described before, although its
synthesis by the reaction of deprotonated 3{5}-methylpyrazole
with 0.5 equiv. of 2,6-dibromopyridine has been reported by two
groups.16,17 Both literature procedures gave a mixture of L2 and its
3,5ꢀꢀ-dimethyl regioisomer L3, which could not be separated. We
have found that, although L2 and L3 elute at almost identical rates
by flash chromatography, a low yield of pure L2 can be separated by
careful elution of the crude product mixture down a silica column
in 3 : 1 pentane/diethyl ether. The 1H NMR spectrum of pure L2
matches that previously reported for this compound.17
Fig. 1 Variable temperature susceptibility data for 1 (᭺) and 2 (♦). The
lines are the fits of these data to the expression for the zero-field splitting
of the iron centre. See text for more details.
This demonstrates that, despite its brown colouration (which is
normally exhibited by low-spin [Fe(L1)2]2+ compounds3–7,11,13,14),
1 is high-spin over this entire temperature range. The low-
temperature decrease in vMT for 1 results from zero-field splitting
5
of the E high-spin manifold.20,21 These data were modelled to
afford g = 2.109(2) and |D| = 7.6(2) cm−1. This value of |D|
implies that 1 adopts the angular structural distortion found
in several other [Fe(L1)2]2+ salts,10 which was confirmed by the
crystallographic study described below. The magnetic data for 3
(not shown in Fig. 1) are very similar to those of 1, and yielded
the parameters g = 2.140(4) and |D| = 9.9(3) cm−1. This value
of |D| is intermediate between those previously observed for
complexes in this series with regular and distorted structures.10
That 3 should remain high-spin on cooling was unexpected, since
we have crystallographically characterised two other solvated salts
of [Fe(L2)2]2+ that do undergo gradual thermal spin-transitions,
both centred at 175 3 K (ESI†).
Treatment of FeI2·4H2O with 2 equiv. of L1 in MeOH at
room temperature gives an orange solution, which yields the
orange powder [FeI2L1] following concentration in vacuo and slow
addition of Et2O. However, performing the reaction under reflux
for 2 h in air, followed by the same work-up, affords approximately
equal amounts of [FeI2L1] and a new, brown crystalline compound
[Fe(L1)2]I0.5[I3]1.5 (1). The I3 content in 1 probably arises from
−
aerobic oxidation of iodide under the more forcing reaction
conditions. These two products could be separated by decantation,
allowing 1 to be fully characterised as described below. Although
we have not obtained a crystal structure of [FeI2L1], it presumably
adopts a five-coordinate geometry analogous to that shown by
[FeCl2L4] in the crystal.18 The ES mass spectra of 1 and [FeI2L1] in
MeCN are almost identical, each containing the strong molecular
ions [Fe2I3(L1)2]+ (m/z 915), [FeI(L1)]+ (394), [FeI(L1)2]2+ (302) and
[Fe(L1)2]2+ (239). This implies that the L1 ligands in 1 are partially
displaced by the iodide ions in this solvent, which explains why
freshly crystallised 1 is always contaminated with [FeI2L1]. Similar
reactions using FeCl2·4H2O yielded [FeCl2L1]19 as the only isolable
product.
The salt [Fe(L1)2][Co(C2B9H11)2]2 (2) was obtained by reaction
of FeCl2·4H2O with 2 equiv. of L1 and excess Ag[Co(C2B9H11)2] in
refluxing CH3NO2. Diffusion of Et2O vapour into the filtered, or-
ange solution affords air-sensitive orange needles of the CH3NO2
solvate of 2. Microanalysis and 1H NMR spectroscopy both
imply that this solvent content is lost upon drying in vacuo, and
magnetochemical measurements were carried out using this dried
material. Finally, [Fe(L2)2][SbF6]2 (3) was obtained by reaction
of FeCl2·4H2O, L2 and AgSbF6 in a 1 : 2 : 2 molar ratio, under
the same conditions used for 2. Diffusion of Et2O vapour into a
solution of 3 in acetone yielded large mustard yellow, solvent-free
crystals of the compound.
In contrast, vMT for 2 decreases steadily upon cooling, from
2.7 cm3 mol−1 K at 340 K to 1.7 cm3 mol−1 K at 220 K, but
remains at 1.6–1.7 cm3 mol−1 K upon further cooling before
showing a small decrease below 50 K as before. If vMT of the
fully high-spin sample is estimated to be 3.3–3.4 cm3 mol−1 K,
this behaviour implies that approximately 50% of the iron centres
in the material undergo a gradual high → low spin transition
centred at T 1 = 318
3 K, while the remaining iron content
2
remains high-spin at all temperatures. Incomplete spin-transitions
of this type can originate through several different effects. Most
simply, they can reflect the existence of more than one type
of magnetically inequivalent iron environment in the sample.22
Alternatively, they may be a consequence of a crystallographic
phase change during spin-crossover;23 of intermolecular steric
interactions;9 of lattice strain induced by structural changes
occurring during spin-crossover;24 or, of an unusually large kinetic
barrier to the transition.25 The crystallographic investigation of 2
described below shows that the first explanation is relevant here.
Fitting the vMT vs. T data for 2 below 200 K yields parameters
g = 2.080(3) and |D| = 10.2(2) cm−1, assuming that 50% of the
material is high-spin at these temperatures. As for 3, this value
of |D| is again intermediate between the values expected if the
high-spin fraction of this compound has a distorted or undistorted
six-coordinate geometry.
Variable temperature susceptibility measurements were carried
out on 1–3. vMT for 1 remains constant at 3.3 cm3 mol−1 K between
30–300 K, before decreasing slightly at lower temperatures (Fig. 1).
The asymmetric unit of crystalline 1 contains one half-occupied
I− anion and three half-occupied I3 ions all lying on crystal-
−
lographic C2 axes; and a complex dication occupying a general
824 | Dalton Trans., 2006, 823–830
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
◦
1
Table 1 Selected bond lengths (A) and angles ( ) for [Fe(L )2]I0.5[I3]1.5 (1) and [Fe(L1)2][Co(C2B9H11)2]2·CH3NO2 (2·CH3NO2). The angles ‘φ’ and ‘h’ are
˚
defined in Scheme 1
2 at 300 K
2 at 150 K
1
Molecule A
Molecule B
Molecule A
Molecule B
Fe(1)–N(2)
Fe(1)–N(9)
Fe(1)–N(14)
Fe(1)–N(18)
Fe(1)–N(25)
Fe(1)–N(30)
2.141(4)
2.181(4)
2.185(5)
2.159(4)
2.171(5)
2.231(5)
1.981(5)
1.962(6)
2.030(4)
1.975(4)
2.031(5)
2.035(6)
2.127(5)
2.167(6)
2.158(6)
2.131(3)
2.173(4)
2.210(4)
1.917(7)
1.996(8)
1.983(7)
1.916(7)
1.958(8)
1.980(8)
2.163(8)
2.202(9)
2.172(8)
2.129(7)
2.180(7)
2.231(8)
N(2)–Fe(1)–N(9)
73.22(16)
72.45(16)
155.97(16)
129.20(16)
86.29(16)
142.24(17)
115.31(16)
99.32(16)
90.82(16)
102.44(16)
90.41(17)
102.14(17)
73.31(16)
71.62(16)
144.51(16)
80.9(3)
77.87(19)
175.7(2)
99.83(18)
104.4(2)
158.4(2)
102.4(2)
96.2(2)
74.5(3)
72.6(3)
79.6(4)
80.4(3)
178.3(3)
98.5(3)
101.6(3)
159.9(3)
101.6(3)
93.9(3)
90.0(3)
98.5(3)
90.8(3)
92.3(3)
80.2(3)
79.7(3)
160.0(3)
74.6(3)
71.7(3)
159.6(3)
124.7(3)
88.1(3)
145.3(3)
112.3(3)
100.4(3)
92.4(3)
102.2(3)
92.2(3)
94.3(3)
74.1(3)
72.8(3)
146.9(3)
N(2)–Fe(1)–N(14)
N(2)–Fe(1)–N(18) (φ)
N(2)–Fe(1)–N(25)
N(2)–Fe(1)–N(30)
N(9)–Fe(1)–N(14)
N(9)–Fe(1)–N(18)
N(9)–Fe(1)–N(25)
N(9)–Fe(1)–N(30)
N(14)–Fe(1)–N(18)
N(14)–Fe(1)–N(25)
N(14)–Fe(1)–N(30)
N(18)–Fe(1)–N(25)
N(18)–Fe(1)–N(30)
N(25)–Fe(1)–N(30)
163.11(15)
122.94(16)
90.03(16)
146.2(2)
108.27(19)
99.42(19)
93.42(18)
105.41(19)
92.2(2)
93.92(17)
73.57(15)
73.27(14)
146.72(15)
89.0(2)
99.08(17)
91.30(18)
92.5(2)
77.08(19)
78.6(2)
155.7(2)
h
89.92(5)
86.04(7)
86.12(6)
87.42(10)
87.03(9)
The cations of 1 aggregate into chains parallel to the lattice c
direction, through a long p–p interaction between pyrazole rings
1
2
N(13)–C(17) and N(8i)–C(12i) (symmetry code i: x, 2 − y, −
+
˚
z). These groups are 3.40(2) A apart, and have a dihedral angle
between them of 4.8(5)◦. The I3 ions in 1 are aligned almost
−
coparallel with each other, in rectangular channels along the
crystallographic [101] vector (Fig. 3). The channels are formed
Fig. 2 View of the complex dication in the crystal structure of
[Fe(L1)2]I0.5[I3]1.5 (1), showing the atom numbering scheme employed. All H
atoms have been omitted, and thermal ellipsoids are at the 50% probability
level.
position (Fig. 2, Table 1). The six-coordinate iron centre is clearly
˚
high-spin from its Fe–N bond◦lengths [2.141(4)–2.231(5) A] and
average L1 bite angle [72.9(3) ].3–5,8–14 However, as predicted by
the magnetic data, the complex dication deviates strongly from
the ‘ideal’ D2d-symmetric geometry expected with this ligand set.
Interestingly, only the ‘rotation’ distortion is observed with φ =
155.97(16)◦ (Scheme 1); the ‘twist’ distortion is not present since
the two L1 ligands in the molecule are perpendicular to within
experimental error [h = 89.92(5)◦]. This is the first time that the
‘rotation’ distortion has been observed in isolation in this class
of compound. Concomitant with this distortion, Fe(1) protrudes
Fig. 3 Partial packing diagram of [Fe(L1)2]I0.5[I3]1.5 (1), showing the
channels occupied by the I3 ions. The I atoms are plotted with their
−
van der Waals radii, while the complex dications are de-emphasised. For
−
clarity, only one orientation of the disordered I3 ion is shown. The view
˚
out of the least squares planes of the two ligands, by 0.406(2) A
is perpendicular to the [101] crystal vector, with the unit cell b direction
vertical.
˚
[ligand N(2)–C(17)] and 0.180(3) A [N(18)–C(33)].
This journal is The Royal Society of Chemistry 2006
Dalton Trans., 2006, 823–830 | 825
©

View Article Online
low-spin structure.3–7,11,13,14 At 300 K, five of the six Fe–N distances
have increased from their low-temperature value, by between 6
and 8 sus. That implies this crystal site now contains a mixture
of high- and low-spin molecules, albeit with the low-spin form
still predominating. In contrast, molecule B shows a C2-distorted
geometry that closely resembles that in 1. As in 1, only a significant
‘rotation’ component of the Jahn–Teller distortion is present in the
molecule [φ = 159.6(3), h = 87.03(9)◦ at 150 K]. As expected, the
Fe–N distances and ligand bite angle at both temperatures indicate
that this site is f◦ully high-spin [at 300 K, these are 2.127(5)–2.210(4)
˚
from cavities of approximate dimensions 3.4 × 9.2 A, which
−
contain two of the I3 environments, linked by constrictions of
−
˚
ca. 3.2 × 4.6 A that hold the third one. The I ions occupy pockets
in the lattice that cap one narrow side of the constrictions. All these
anions are held in place by C–H · · · I hydrogen bonds (ESI†). The
−
closest interatomic contacts between neighbouring I3 anions are
˚
4.3–4.4 A, essentially equal to the sum of the van der Waals radii of
two iodine atoms (4.3 A ). Hence there are no significant bonding
interactions between the anions in the lattice, which supports the
formulation of the structure as containing discrete I− and I3
26
˚
−
◦
centres, rather than longer polyiodide chains.27
A and 73.5(5) ; at 150 K, 2.129(7)–2.231(8) A and 73.3(6) ].
Interestingly, bonds Fe(1B)–N(2B) and Fe(1B)–N(9B) appear to
be shorter at 300 K than at 150 K, although these differences
are of marginal statistical significance at <4 sus. The structure of
molecule B is further distorted by a steric contact between C(27B)
and a cobalticarborane B–H group (not shown in Fig. 4), which
forces pyrazole ring N(24B)–C(28B)◦to bend out of the plane of
the rest of that L1 ligand, by 14.6(4) at 150 K. The open face in
molecule B created by the rotation distortion is filled by another
anion, lying above and parallel to the ligand N(2B)–C(27B) (not
shown in Fig. 4).
˚
˚
Crystals of 2·CH3NO2 are weak diffractors, particularly at room
temperature, which may reflect their needle morphology and/or
disorder caused by the mixture of spin-states they contain. Despite
this caveat, the spin-states and stereochemistries of the iron centres
at the temperatures of measurement (150 and 300 K) are unam-
biguous. The asymmetric unit of 2·CH3NO2 contains two for-
mula units, with two iron centres that are structurally very different
(Fig. 4, Table 1). Molecule A has a nearly regular D2d-symmetric
˚
geometry, with Fe–N bond lengths [1.916(7)–1.996(8) A] and aver-
age L1 bite angle [80.0(7)◦] at 150 K that are consistent with a fully
This compound is the first example of this class of iron(II)
complex to contain co-crystallised C2-distorted and undistorted
molecules, which explains its incomplete thermal spin-transition.
Molecule A, which is low-spin at 150 K but in a mixture of spin
states at 300 K, undergoes a gradual transition to its high-spin
form at higher temperatures. The bond lengths at Fe(1A) imply
that this transition is <50% complete at 300 K, consistent with the
magnetic data (Fig. 1). Molecule B, in contrast, is trapped in its
high-spin state by its angular Jahn–Teller distortion4 and remains
high-spin at all temperatures.
The complex dications in 2·CH3NO2 are well-separated from
each other. The only short intermolecular contacts between
complex molecules are: a p–p interaction between pyrazole rings
N(13A)–C(17A) and N(29B)–C(33B), which are separated by
◦
˚
3.30(4) A with a dihedral angle between them of 0.5(4) at
˚
150 K; and, a long C–H · · · p contact of 2.82 A between C(32B)–
H(32B) and the centroid of ring N(24A)–C(28A) (the van der
˚
Waals radii of an H atom and arene ring are 1.2 and 1.7 A
respectively26). Importantly, molecule A interacts with its nearest
neighbour of the same type (related by x, 1 + y, z) through
van der Waals interactions only. The lack of direct interactions
between molecules A in the lattice is consistent with the gradual,
non-cooperative spin-transition undergone by this compound.
In addition to these inter-cation interactions, there are several
1
˚
intermolecular C–H · · · H–B contacts of 2.1–2.3 A between the L
ligands and cobalticarborane anions in the lattice, and C–H · · · O
˚
contacts of 2.4–2.5 A between the ligands and solvent.
Consistent with its magnetic behaviour, the Fe–N bond lengths
◦
˚
[2.125(3)–2.214(3) A] and average ligand bite angle [73.7(2) ] in
3 at 150 K indicate that this compound has a high-spin structure
(Fig. 5, Table 2). Unlike the other high-spin compounds in this
study, however, this complex has a near-regular D2d-symmetric
geometry, with the distortion angles φ and h (Scheme 1) both being
close to their ideal values (Table 2). These parameters are very
similar to those exhibited by the high-spin forms of the solvate salts
[Fe(L2)2][BF4]2·4CH3CN and[Fe(L2)2][ClO4]2·(CH3)2CO, which do
undergo a thermal spin-transition upon cooling (ESI†). Hence
the high-spin nature of 3 cannot reflect either a structural
Fig. 4 View of the two independent complex dications in the crystal
structure of [Fe(L1)2][Co(C2B9H11)2]2·CH3NO2 (2·CH3NO2) at 150 K,
showing the atom numbering scheme employed. All H atoms have been
omitted, and thermal ellipsoids are at the 50% probability level.
826 | Dalton Trans., 2006, 823–830
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
◦
2
˚
Table 2 Selected bond lengths (A) and angles ( ) for [Fe(L )2][SbF6]2 (3).
The angles ‘φ’ and ‘h’ are defined in Scheme 1
Fe(1)–N(2)
Fe(1)–N(9)
Fe(1)–N(15)
2.125(3)
2.182(3)
2.169(3)
Fe(1)–N(20)
Fe(1)–N(27)
Fe(1)–N(33)
2.129(3)
2.183(3)
2.214(3)
N(2)–Fe(1)–N(9)
N(2)–Fe(1)–N(15)
N(2)–Fe(1)–N(20) (φ) 177.95(11) N(15)–Fe(1)–N(27)
N(2)–Fe(1)–N(27)
N(2)–Fe(1)–N(33)
N(9)–Fe(1)–N(15)
N(9)–Fe(1)–N(20)
N(9)–Fe(1)–N(27)
73.63(11) N(9)–Fe(1)–N(33)
73.94(11) N(15)–Fe(1)–N(20)
98.28(11)
106.15(11)
91.94(11)
89.87(12)
73.38(11)
73.83(11)
146.32(11)
108.67(11) N(15)–Fe(1)–N(33)
104.14(11) N(20)–Fe(1)–N(27)
147.56(11) N(20)–Fe(1)–N(33)
106.28(11) N(27)–Fe(1)–N(33)
98.03(11)
h
85.84(10)
Fig. 6 Alternative view of the complex dication in the crystal structure of
[Fe(L2)2][SbF6]2 (3), emphasising the conformational distortion in ligand
N(20)–C(37), which is vertical and to the rear. The intermolecular steric
contact that causes this distortion is also shown. All atoms have arbitrary
1
1
radii. Symmetry code ii: x, − − y, + z.
2
2
Conclusion
We have shown previously that a high-spin state can be imposed
onto an iron(II)–2,6-di(pyrazol-1-yl)pyridine complex through the
angular Jahn–Teller distortion4,9,10,14 or, when this is not present,
through intramolecular steric interactions between the two ligands
in the molecule.6,10 This study has extended our knowledge of
these phenomena in three ways. First, 1 and molecule B of 2
are the first compounds to show purely the ‘rotation’ component
of the structural distortion (i.e. φ< 180, h ≈ 90◦). We have
previously isolated one complex that undergoes only the ‘twist’
component (φ = 180, h < 90◦), using a sterically bulky 2,6-
di(pyrazol-1-yl)pyridine derivative.10 Since 1 and 2 do not contain
bulky substituents, this work confirms that the two components of
the Jahn–Teller distortion in this system can occur independently
of one another even in the absence of steric effects. Compound
2 is also important as the first material to contain co-crystallised
distorted and undistorted molecules of the same complex. This
confirms that the different forms of the same compound can co-
exist in the solid and, presumably, also in solution. The plethora of
different stereochemistries observed in salts of [Fe(L1)2]2+ simply
reflects which one crystallises preferentally from this mixture of
structures in solution.
The novel structures of 1 and 2 have allowed us to extend our
magnetostructural correlation for the angular distortion in these
compounds (Fig. 7). We previously suggested that the ‘rotation’
component was the dominant factor in the reduced zero-field
splittings |D| of 3–8 cm−1 shown by the distorted structures.10
Fig. 7 supports that contention, in that compounds showing only
the ‘rotation’ distortion exhibit reduced values of |D|, while the
example exhibiting only the ‘twist’ component does not. However,
it is also clear that the lowest values of |D| are exhibited by
those complexes adopting reduced values of both distortion angles.
Fig. 7 demonstrates that a small zero-field splitting of 3–5 cm−1
should be diagnostic for a structure with φ ≈ 155◦ and h ≈ 60◦.
Fig. 5 View of the complex dication in the crystal structure of
[Fe(L2)2][SbF6]2 (3), showing the atom numbering scheme employed. All H
atoms have been omitted, and thermal ellipsoids are at the 50% probability
level.
distortion of the iron coordination sphere, or intramolecular
steric contacts involving the ligand methyl groups (as is observed
in analogous compounds with larger pyrazole substituents10).
Rather, we tentatively attribute the fact that 3 is trapped in its high-
spin state to the very twisted conformation exhibited by ligand
N(20)–C(37), which deviates strongly from planarity (Fig. 6). The
dihedral angles between the least squares planes of pyridine ring
N(20)–C(25) and pyrazole rings N(26)–C(30) and N(32)–C(36)
are 17.2(2) and 21.8(2)◦, respectively. This distortion most likely
reflects intermolecular steric contacts between C(18ii)–H(18ii) and
1
2
˚
˚
C(35) (2.72 A) and C(36) (2.64 A) (symmetry code ii: x, −
−
1
2
y, + z). These distances are significantly shorter than the sum
of the radii of an H atom and aryl group (1.2 and 1.7 A), and
˚
are positioned to displace pyrazole ring N(32)–C(36) out of the
meridional plane of the complex (Fig. 6). We propose that its
strained conformation introduces additional rigidity into this L2
ligand, disfavouring the structural rearrangement required during
a spin-transition. For comparison the equivalent inter-heterocycle
dihedral angles in the other ligand, N(2)–C(19), are 7.01(11)
and 1.72(14)◦. There are no other noteworthy intermolecular
interactions in the crystal lattice of 3.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 823–830 | 827
©

View Article Online
1
spectrum m/z 240 [M + H]+. NMR spectra (CDCl3, 293 K): H
(250 MHz): d, ppm 8.42 (d, 2H, 2.5 Hz, Pz H5), 7.70–7.88 (m, 3H,
1
Py H3–5), 6.25 (d, 2H, 2.5 Hz, Pz H4), 2.38 (s, 6H, CH3). 13C{ H}
(63 MHz): d, ppm 151.9 (2C, Py C2/6), 149.9 (2C, Pz C3), 141.0
(1C, Py C4), 127.6 (2C, Pz C5), 108.4 and 108.0 (both 2C, Py C3/5
and Pz C4), 13.9 (2C, CH3).
Synthesis of [Fe(L1)2]I0.5[I3]1.5 (1) and [FeI2L1]
A solution of FeI2·4H2O (0.25 g, 6.6 × 10−4 mol) and L1 (0.28 g,
1.3 × 10−3 mol) in MeOH (50 cm3) was refluxed for 2 h. The
resultant brown solution was filtered, and concentrated under
reduced pressure to ca. 5 cm3. Slow diffusion of Et2O vapour
into the solution yielded large brown crystals of 1, contaminated
by [FeI2L1] as a light orange powder that could be removed by
decantation. For 1: yield 0.17 g, 57% with respect to iodine. Found
C, 23.8; H, 1.5; N, 12.8; I, 56.9%. Calcd. for C22H18FeI5N10 C,
23.7; H, 1.6; N, 12.6; I, 57.0%. Electrospray mass spectrum: m/z
915 [56Fe2127I3(L1)2]+, 435 [56Fe127I(L1)(NCMe)]+, 394 [56Fe127I(L1)]+,
339 [L1 + 127I]+, 312 [56Fe127I(L1)2(H2O)]2+, 302 [56Fe127I(L1)2]2+, 239
[56Fe(L1)2]2+, 212 [L1 + H]+. For [FeI2L1]: yield 0.11 g, 32%. Found
C, 25.3; H, 1.7; N, 13.3; I, 47.6%. Calcd. for C11H9FeI2N5 C, 25.4;
H, 1.7; N, 13.4; I, 48.7%. Electrospray mass spectrum: m/z 915
[56Fe2127I3(L1)2]+, 435 [56Fe127I(L1)(NCMe)]+, 394 [56Fe127I(L1)]+, 339
[L1 + 127I]+, 302 [56Fe127I(L1)2]2+, 239 [56Fe(L1)2]2+, 212 [L1 + H]+.
Fig. 7 Correlation of the structural distortion, as defined in Scheme 1,
with the zero-field splitting parameters shown by the high-spin iron centres
in 1–3 and the compounds in refs. 4, 10 and 14. Errors on |D| in this plot
are 0.2–0.5 cm−1
.
However, intermediate values of 7 ≤ |D| ≤ 10 cm−1 cannot be
simply assigned to a particular structural type in this system, and
should be interpreted with care.
Finally, compound 3 introduces a new way in which this class
of iron(II) complex can become trapped in a high-spin state.
Rather than showing a distorted metal coordination sphere, 3
exhibits a severe conformational distortion at one of its L2
ligands that is induced by an intermolecular steric contact.
The plethora of structures adopted by high-spin iron(II)–2,6-
di(pyrazol-1-yl)pyridine complexes has not been seen in any other
stereochemically related six-coordinate compounds, and will make
the controlled synthesis of new spin-transition materials based on
this motif a particular challenge.
Synthesis of [Fe(L1)2][Co(C2B9H11)2]2 (2)
A mixture of FeCl2·4H2O (0.12 g, 5.9 × 10−4 mol), L1 (0.25 g, 1.2 ×
10−3 mol) and Ag[Co(C2B9H11)2 (1.1 g, 2.5 × 10−3 mol) in MeNO2
(50 cm3) was refluxed for 1 h, affording an orange solution with a
white AgCl precipitate. The solution was filtered and concentrated
under reduced pressure to ca. 5 cm3. Slow diffusion of Et2O
vapour into the solution at −10◦C yielded orange needles of the
MeNO2 solvate of 2, which were collected and dried in vacuo
over P2O5. Yield 0.31 g, 46%. Found C, 31.7; H, 5.5; N, 12.6%.
Calcd. for C30H62B36Co2FeN10 C, 32.0; H, 5.6; N, 12.4%. Elec-
Experimental
Unless otherwise stated, all manipulations were carried out in
air using reagent grade solvents, except that bis(2-methoxyethyl)
ether was dried over sodium before use. 2,6-Di(pyrazol-1-
yl)pyridine (L1)16 and Ag[Co(C2B9H11)2]28 were prepared by the
literature procedures, while all other reagents were used as
supplied. Synthetic, analytical and crystallographic data for
[Fe(L2)2][BF4]2·4CH3CN and [Fe(L2)2][ClO4]2·(CH3)2CO are given
in the ESI†. The solution phase NMR, UV/vis and susceptibility
data from [Fe(L1)2]2+ (as [Fe(L1)2][BF4]2) have been reported
previously.4
trospray mass spectrum: m/z 805 [56Fe(59Co{C211B9H11}2)(L1)2]+,
635 [56Fe(59Co{C211B9H11}2)(L1)(NCMe)]+, 594 [56Fe(59Co{C2
-
11
B9H11}2)(L1)]+, 501 [56Fe(59Co{C211B9H11}2)2(L1)(NCMe)2]2+, 460
[56Fe(59Co{C211B9H11}2)2(L1)]2+, 239 [56Fe(L1)2]2+, 212 [L1 + H]+.
Synthesis of [Fe(L2)2][SbF6]2 (3)
A solution of FeCl2·4H2O (0.12 g, 5.9 × 10−4 mol), L2 (0.29 g, 1.2 ×
10−3 mol) and AgSbF6 (0.41 g, 1.2 × 10−3 mol) in MeNO2 (50 cm3)
was refluxed for 1 h. The resultant AgCl precipitrate was removed
by filtration, and the yellow solution concentrated under reduced
pressure to ca. 5 cm3. Slow diffusion of Et2O vapour into the
solution gave large mustard-yellow crystals of the product, which
were collected and dried in vacuo over P2O5. Yield 0.37 g, 62%.
Found C, 31.1; H, 2.6; N, 14.1%. Calcd. for C26H26F12FeN10Sb2
C, 31.0; H, 2.6; N, 13.9%. Electrospray mass spectrum: m/z =
Synthesis of 2,6-di(3-methylpyrazol-1-yl)pyridine (L2)
A mixture of 3{5}-methylpyrazole (3.7 g, 4.6 × 10−2 mol) and
potassium hydride (1.8 g, 4.6 × 10−2 mol) in bis(2-methoxyethyl)
ether (150 cm3) under N2 was stirred for 1 h, affording a white
suspension. Solid 2,6-dibromopyridine (5.32 g, 2.2 × 10−2 mol) was
added to the mixture, which was then stirred at 130 ◦C for 5 days.
The cooled mixture was quenched with water (300 cm3), yielding
a white precipitate of the crude product which was isolated and
dried over phosphorous pentoxide. Flash silica chromatography
(eluent 3 : 1 pentane/diethyl ether) afforded pure L2 as a white
solid. Yield 0.84 g, 16%. Mp 103–105 ◦C (lit.,17 74 ◦C, from a 4 :
1 mixture of L2 and L3). Found: C, 65.0; H, 5.5; N, 29.6%. Calcd.
for C13H13N5: C, 65.2; H, 5.5; N, 29.3%. Electron impact mass
1
267 [56Fe(L2)2]2+, 240 [L2 + H]+. H NMR spectrum (CD3CN):
d, ppm 70.7, 67.9, 37.4 (all 4H, Py H3/5 and Pz H4 and H5), 7.1
(12H, CH3), 6.2 (2H, Py H4). Peak linewidths range between 50–
70 Hz. UV/vis/NIR spectrum (MeCN, 298 K): mmax/103 cm−1
(emax/M−1 cm−1) 7.1 (8.5), 9.4 (sh), 11.4 (sh), 23.4 (sh), 25.3 (sh),
26.3 (sh), 32.3 (38800), 36.4 (23400), 40.0 (70400), 41.0 (sh).
828 | Dalton Trans., 2006, 823–830
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Table 3 Experimental details for the single crystal structure determinations in this study
1
2·CH3NO2 (1/0.75)
3
Molecular formula
Mr
C22H18FeI5N10
1112.81
Monoclinic
P2/c
15.9084(3)
13.6094(3)
15.1155(2)
104.3473(8)
3170.50(10)
4
150(2)
5.375
41942
7248
C30.75H64.25B36Co2FeN10.75O1.5
C30.75H64.25B36Co2FeN10.75O1.5
C26H26F12FeN10Sb2
1005.92
Monoclinic
P21/c
19.7007(4)
11.1095(2)
16.8557(2)
110.9156(6)
3446.02(10)
4
150(2)
2.073
51570
7885
1171.55
Orthorhombic
Pca21
37.297(2)
10.5955(6)
29.6321(17)
—
11710.1(12)
8
300(2)
0.847
146512
18205
0.039
0.044
1171.55
Orthorhombic
Pca21
37.3516(19)
10.5840(6)
29.5118(15)
—
11666.9(11)
8
150(2)
0.850
105918
22460
0.059
0.077
Crystal class
Space group
˚
a/A
˚
b/A
˚
c/A
b/◦
3
˚
V/A
Z
T/K
l(Mo-Ka)/mm−1
Measured reflections
Independent reflections
Rint
0.154
0.049
0.133
1.068
0.101
0.040
0.104
1.034
R(F)a
wR(F2)b
0.117
1.051
0.496(11)
0.214
1.078
0.49(2)
Goodness of fit
Flack parameter
—
—
1
2
a R = R[|Fo| − | Fc|]/R|Fo|. b wR = [Rw(Fo − Fc2)/RwFo ] .
2
4
−3
˚
˚
Single crystal X-ray structure determinations
Fourier peak of 1.1 e A , lying 2.5 A from two I atoms of the
same I3 half-anion.
−
Crystals of 1 were grown by slow diffusion of Et2O vapour
into a solution of the compound in methanol, while crystals of
2·CH3NO2 and 3 were similarly grown from MeNO2/Et2O. Ex-
perimental data from these structure determinations are collected
in Table 3. Diffraction data from 1 and 3 were collected on a
Nonius KappaCCD area detector diffractometer, using graphite-
Single crystal X-ray structure of [Fe(L1)2][Co(C2B9H11)2]2·
CH3NO2(1/0.75) (2·CH3NO2 (1/0.75)). The same crystal was
used for data collection at 150 and 300 K. The crystal diffracted
weakly at 300 K, and data were only collected to 2h = 47.9◦.
The crystal diffracted more strongly at 150 K, allowing data to
2h = 53.7◦ to be obtained. The precision of the refinement at
150 K is noticeably inferior to the 300 K structure, however, which
may reflect lattice strain and/or unresolved disorder induced by
the partial spin-transition undergone by the crystal upon cooling.
Although the Pca21 space group adopted by the compound is
polar, the crystal used was a racemic twin, with both hands equally
populated according to its Flack parameter (which is 0.5 within
experimental error). The asymmetric unit contains two complex
dications, four cobalticarborane anions and two nitromethane
molecules. The latter were modelled at both temperatures with
occupancies of 0.9 and 0.6. This gives a total solvent content of
1.5 molecules per asymmetric unit, or 0.75 per formula unit, in the
crystal examined.
˚
monochromated Mo-Ka radiation (k = 0.71073 A) from a sealed
tube source. Since crystals of 2·CH3NO2 diffract weakly, data
from this compound were obtained using a Bruker X8 Apex
diffractometer, with graphite-monochromated Mo-Ka radiation
generated by a rotating anode. Both diffractometers are fitted
with Oxford Cryostream low temperature devices. All structures
in this study were solved by direct methods (SHELXS9629), and
developed by full least-squares refinement on F2 (SHELXL9630).
All crystallographic figures were prepared using XSEED,31 which
incorporates POVRAY.32
Experimental details for the structure determinations of
[Fe(L2)2][BF4]2·4CH3CN and [Fe(L2)2][ClO4]2·(CH3)2CO are given
in the ESI.†
CCDC reference numbers 278897–278908.
For crystallographic data in CIF or other electronic format see
DOI: 10.1039/b510370c
At 300 K both solvent sites are disordered over two positions,
with occupancy ratios of 0.5 : 0.4 and 0.3 : 0.3 respectively. These
were modelled using the restraints C–N = 1.45(2), N–O = 1.15(2),
˚
C · · · O = 2.30(2) and O · · · O = 1.88(2) A. All non–H atoms except
Single crystal X-ray structure of [Fe(L1)2]I0.5[I3]1.5 (1). The
asymmetric unit of this compound contains one complex dication
lying on a general position; one half-occupied iodide anion lying
on the C2 axis (1, y, 0.75); two half-occupied triiodide anions lying
on C2 axes (0.5, y, 0.75) and (0.5, y, 0.25); and a disordered half-
occupied triiodide ion lying across the C2 axis (0, y, 0.25). The
disordered half-anion was modelled over two equally occupied
for the disordered solvent were refined anisotropically, and all H
atoms were placed in calculated positions and refined using a
riding model. There is one notable residual Fourier peak of 1.27 e
−3
˚
˚
A , located 1.0 A from Fe(1A).
At 150 K only one orientation of each solvent site was detected,
although discrepancies in their C–N and N–O bond lengths
imply that there may also be some unresolved disorder at these
temperatures. All non–H atoms except for the 60% occupied
solvent site were refined anisotropically, and all H atoms were
placed in calculated positions and refined using a riding model.
One carbaborane C atom became non-positive definite during
the refinement process, and was restrained so that its individual
˚
sites, with I–I bonds that were restrained to 2.92(2) A. All non-H
atoms were refined anisotropically, and all H atoms were placed in
calculated positions and refined using a riding model. Two residual
−3
˚
Fourier peaks of 1.0–1.7 e A , and the deepest Fourier hole of
−3
˚
˚
−2.3 e A , are all <1 A from an I atom. There is a third residual
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 823–830 | 829
©

View Article Online
thermal parameters approximate to isotropic behaviour during the
8 V. A. Money, J. S. Costa, S. Marce´n, G. Chastanet, J. Elha¨ık, M. A.
Halcrow, J. A. K. Howard and J.-F. Le´tard, Chem. Phys. Lett., 2004,
391, 273.
9 J. Elha¨ık, C. A. Kilner and M. A. Halcrow, CrystEngComm, 2005, 7,
151.
final least squares cycles with a SHELXL ISOR instruction. There
−3
˚
˚
are three residual Fourier peaks of 1.7–2.2 e A , each ≤1.1 A from
one of the metal atoms in the model.
10 J. Elha¨ık, D. J. Evans, C. A. Kilner and M. A. Halcrow, Dalton Trans.,
Single crystal X-ray structure of [Fe(L2)2][SbF6]2(3). No dis-
order was detected during refinement of this structure, and no
restraints were applied to it. All non-H atoms were refined
anisotropically, while H atoms were placed in calculated positions
2005, 1693.
11 J. Elha¨ık, V. A. Money, S. A. Barrett, C. A. Kilner, I. Radosavljevic
Evans and M. A. Halcrow, Dalton Trans., 2003, 2053.
12 V. A. Money, I. Radosavljevic Evans, J. Elha¨ık, M. A. Halcrow and
J. A. K. Howard, Acta Crystallogr., Sect. B, 2004, 60, 41.
13 V. A. Money, J. Elha¨ık, I. Radosavljevic Evans, M. A. Halcrow and
J. A. K. Howard, Dalton Trans., 2004, 65.
and refined using a riding model. The deepest Fourier hole of
−3
˚
˚
−1.4 e A is 0.9 A from one of the antimony atoms.
14 C. A. Kilner and M. A. Halcrow, Polyhedron, 2005,
DOI: 10.1016/j.poly.2005.06.034.
15 See e.g., T. Buchen, P. Gu¨tlich and H. A. Goodwin, Inorg. Chem., 1994,
33, 4573; T. Buchen, P. Gu¨tlich, K. H. Sugiyarto and H. A. Goodwin,
Chem. Eur. J., 1996, 2, 1134; S. Marce´n, L. Lecren, L. Capes, H. A.
Goodwin and J.-F. Le´tard, Chem. Phys. Lett., 2002, 358, 87; K. H.
Sugiyarto, W.-A. McHale, D. C. Craig, A. D. Rae, M. L. Scudder
and H. A. Goodwin, Dalton Trans., 2003, 2443; A. Bhattacharjee, V.
Ksenofontov, K. H. Sugiyarto, H. A. Goodwin and P. Gu¨tlich, Adv.
Funct. Mater., 2003, 13, 877.
16 D. L. Jameson and K. A. Goldsby, J. Org. Chem., 1990, 55, 4992.
17 X. Sun, Z. Yu, S. Wu and W.-J. Xiao, Organometallics, 2005, 24, 2959.
18 F. Calderazzo, U. Englert, C. Hu, F. Marchetti, G. Pampaloni, V.
Passarelli, A. Romano and R. Santi, Inorg. Chim. Acta, 2003, 344,
197.
19 A. R. Karam, E. L. Catari, F. Lo´pez-Linares, G. Agrifoglio, C. L.
Albano, A. D´ıaz-Barrios, T. E. Lehmann, S. V. Pekerar, L. A. Albornoz,
R. Atencio, T. Gonzalez, H. B. Ortega and P. Joskowics, Appl. Catal.,
A, 2005, 280, 165.
20 C. J. O’Connor, Prog. Inorg. Chem., 1982, 29, 203.
21 R. Bocˇa, Coord. Chem. Rev., 2004, 248, 757.
22 See e.g., S. Calogero, G. G. Lobbia, P. Cecchi, G. Valle and J. Friedl,
Polyhedron, 1994, 13, 87; J. Kusz, H. Spiering and P. Gu¨tlich, J. Appl.
Crystallogr., 2001, 34, 229 and refs. therein;; G. Psomas, N. Brefuel, F.
Dahan and J.-P. Tuchagues, Inorg. Chem., 2004, 43, 4590.
23 See e.g., D. L. Reger, C. A. Little, V. G. Young, Jr. and M. Pink, Inorg.
Chem., 2001, 40, 2870.
24 See e.g., P. Guionneau, J.-F. Letard, D. S. Yufit, D. Chasseau, G. Bravic,
A. E. Goeta, J. A. K. Howard and O. Kahn, J. Mater. Chem., 1999,
9, 985; A. F. Stassen, M. de Vos, P. J. van Koningsbruggen, F. Renz, J.
Ensling, H. Kooijman, A. L. Spek, J. G. Haasnoot, P. Gu¨tlich and J.
Reedijk, Eur. J. Inorg. Chem., 2000, 2231.
Other measurements
UV/vis spectra were obtained with a Perkin-Elmer Lambda 900
spectrophotometer, operating between 1 100–200 nm, in 1 cm
quartz cells. Electron impact and electrospray (MeCN matrix)
mass spectra were respectively performed with VG AutoSpec
and Micromass LCT TOF spectrometers. CHN microanalyses
were performed by the University of Leeds School of Chemistry
microanalytical service. NMR spectra were run on a Bruker
ARX250 spectrometer, operating at 250.1 MHz (1H) or 62.9 MHz
(13C). Magnetic susceptibility measurements were obtained using
a Quantum Design SQUID magnetometer in an applied field of
1000 G. Diamagnetic corrections were estimated from Pascal’s
constants.20 Theoretical fits of the susceptibility data to the
equation for the zero-field splitting of a high-spin d6 transition
ion20,21 were carried out using SIGMAPLOT.33
Acknowledgements
The authors thank Dr H. J. Blythe (University of Sheffield) for the
susceptibility data, and the EPSRC and the University of Leeds
for funding.
25 See e.g., M. S. Haddad, W. D. Federer, M. W. Lynch and D. N.
Hendrickson, J. Am. Chem. Soc., 1980, 102, 1468; N. Moliner, A. B.
References
1 J. A. Real, A. B. Gaspar and M. C. Mun˜oz, Dalton Trans., 2005, 2062.
2 M. A. Halcrow, Coord. Chem. Rev., 2005, DOI: 10.1016/j.ccr.
2005.03.010.
3 J. M. Holland, J. A. McAllister, Z. Lu, C. A. Kilner, M. Thornton-Pett
and M. A. Halcrow, Chem. Commun., 2001, 577.
4 J. M. Holland, J. A. McAllister, C. A. Kilner, M. Thornton-Pett, A. J.
Bridgeman and M. A. Halcrow, J. Chem. Soc., Dalton Trans., 2002,
548.
5 V. A. Money, I. Radosavljevic Evans, M. A. Halcrow, A. E. Goeta and
J. A. K. Howard, Chem. Commun., 2003, 158.
6 J. M. Holland, S. A. Barrett, C. A. Kilner and M. A. Halcrow, Inorg.
Chem. Commun., 2002, 5, 328.
Gaspar, M. C. Mun˜oz, V. Niel, J. Cano and J. A. Real, Inorg. Chem.,
2001, 40, 3986.
26 L. Pauling, The Nature of the Chemical Bond, Cornell University Press,
Ithaca NY, USA, 3rd edn, 1960, pp. 257–264.
27 P. H. Svensson and L. Kloo, Chem. Rev., 2003, 103, 1649.
28 Z. Xie, T. Jelinek, R. Bau and C. A. Reed, J. Am. Chem. Soc., 1994,
116, 1907.
29 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467.
30 G. M. Sheldrick, SHELXL97, Program for the refinement of crystal
structures, University of G
o¨ttingen, Germany, 1997.
31 L. J. Barbour, J. Supramol. Chem., 2003, 1, 189.
32 POVRAY v. 3.5, Persistence of Vision Raytracer Pty. Ltd.,
Williamstown, Victoria, Australia, 2002, http://www.povray.org.
33 SIGMAPLOT, v. 8.02, SPSS Scientific Inc., Chicago IL, 2002.
7 V. A. Money, J. Elha¨ık, M. A. Halcrow and J. A. K. Howard, Dalton
Trans., 2004, 1516.
830 | Dalton Trans., 2006, 823–830
This journal is
The Royal Society of Chemistry 2006
©