Synthesis of 2,6-Di(pyrazol-1-yl)pyrazine derivatives and the spin-state behavior of their iron(II) complexes

Article abstract of DOI:10.1002/ejic.201201100

Chlorination of 2,6-bis(pyrazol-1-yl)pyrazine (bppz) with Na- ClO in acetic acid afforded 2,6-bis(4-chloropyrazol-1-yl)pyrazine (L²Cl). 2,6-Bis(4-bromopyrazol-1-yl)pyrazine (L²Br), 2,6-bis(4-iodopyrazol-1- yl)pyrazine (L²I), 2,6-bis(4-methylpyrazol- 1-yl)pyrazine (L ²Me), and 2,6-bis(4-nitropyrazol-1- yl)pyrazine (L²NO ₂) were also prepared by reactions of the preformed 4-substituted pyrazoles with 2,6-dichloropyrazine. The reduction of L²NO ₂ with iron powder gave 2,6-bis(4- aminopyrazol-1-yl)pyrazine (L ²NH₂) and L²I was converted into 2,6-bis[4-(phenylethynyl)pyrazol-1-yl]pyrazine (L²CCPh) by a Sonogashira coupling reaction. The salts [Fe(L²Me)₂]X ₂ (X- = BF₄ - and ClO₄ -) underwent thermal spin-crossover abruptly at around 200 K in one and two steps, respectively. The [Fe(L²Me)₂]X₂ salts exhibited dif- ferent light-induced excited spin-state trapping (LIESST) behavior; the BF₄ - salt behaves classically [T(LIESST) = 93 K], but the ClO₄ - salt undergoes a multistep LIESST relaxation. In contrast, solid [Fe(L ²Cl)₂][BF₄]2 adopts a fixed 2:1 high/ low-spin-state population that does not change with temperature below 300 K, whereas [Fe(L²Br)₂][BF₄]2 and [Fe(L ²I)₂][BF₄]2 form low-spin solvated crystals that are transformed into high-spin powders on drying. The pyrazinyl group in the L²R ligands slightly stabilizes the low-spin state of the complexes, as determined by solution-phase magnetic measurements. The crystal structure of [Fe(L²CCPh)(OH₂)z]- [BF₄]2 contains a disordered mixture of six- (z = 3) and sevencoordinate (z = 4) iron centers.

Full text of DOI:10.1002/ejic.201201100

FULL PAPER
DOI:10.1002/ejic.201201100
Synthesis of 2,6-Di(pyrazol-1-yl)pyrazine Derivatives and
the Spin-State Behavior of Their Iron(II) Complexes
CLUSTER
ISSUE
Rufeida Mohammed,[a] Guillaume Chastanet,[b] Floriana Tuna,[c]
Tamsin L. Malkin,[a] Simon A. Barrett,[a] Colin A. Kilner,[a]
Jean-François Létard,*[b] and Malcolm A. Halcrow*[a]
Keywords: Spin crossover / Magnetic properties / Iron / N ligands
Chlorination of 2,6-bis(pyrazol-1-yl)pyrazine (bppz) with Na-
ClO in acetic acid afforded 2,6-bis(4-chloropyrazol-1-yl)pyr-
azine (L2Cl). 2,6-Bis(4-bromopyrazol-1-yl)pyrazine (L2Br),
2,6-bis(4-iodopyrazol-1-yl)pyrazine (L2I), 2,6-bis(4-methyl-
pyrazol-1-yl)pyrazine (L2Me), and 2,6-bis(4-nitropyrazol-1-
yl)pyrazine (L2NO2) were also prepared by reactions of the
preformed 4-substituted pyrazoles with 2,6-dichloropyrazine.
The reduction of L2NO2 with iron powder gave 2,6-bis(4-
aminopyrazol-1-yl)pyrazine (L2NH2) and L2I was con-
verted into 2,6-bis[4-(phenylethynyl)pyrazol-1-yl]pyrazine
(L2CCPh) by a Sonogashira coupling reaction. The salts
ferent light-induced excited spin-state trapping (LIESST) be-
–
havior; the BF4 salt behaves classically [T(LIESST) = 93 K],
–
but the ClO4 salt undergoes a multistep LIESST relaxation.
In contrast, solid [Fe(L2Cl)2][BF4]2 adopts a fixed 2:1 high/
low-spin-state population that does not change with tem-
perature below 300 K, whereas [Fe(L2Br)2][BF4]2 and
[Fe(L2I)2][BF4]2 form low-spin solvated crystals that are trans-
formed into high-spin powders on drying. The pyrazinyl
group in the L2R ligands slightly stabilizes the low-spin state
of the complexes, as determined by solution-phase magnetic
measurements. The crystal structure of [Fe(L2CCPh)(OH2)z]-
[BF4]2 contains a disordered mixture of six- (z = 3) and seven-
coordinate (z = 4) iron centers.
–
[Fe(L2Me)2]X2 (X– = BF4 and ClO4–) underwent thermal
spin-crossover abruptly at around 200 K in one and two
steps, respectively. The [Fe(L2Me)2]X2 salts exhibited dif-
idine ring allows a [Fe(bpp)2]2+ center to be functionalized
without significantly perturbing the iron center. The good
availability of crystallographic data from this series of com-
plexes has allowed structure–function relationships to be
elucidated that control their spin-crossover under both
thermodynamic[5] and kinetic (light-induced excited spin-
state trapping, LIESST) conditions.[6] In addition, their ease
of functionalization has allowed multimetallic complexes,[7]
coordination polymers,[8] multifunctional spin-crossover
molecules,[9] and surface assemblies[10] to be constructed by
using [Fe(bpp)2]2+ components.
An alternative way of modifying the bpp ligand is to in-
corporate additional heteroatoms into its constituent
rings.[11–13] One such derivative is 2,6-di(pyrazol-1-yl)pyr-
azine (bppz), which was first prepared by Hosseini and co-
workers.[11] Several salts of [Fe(bppz)2]2+[14,15] and
[Fe(bppz*)2]2+[14,16] are spin-crossover active. However,
these materials are generally not isostructural with their an-
alogues from the [Fe(bpp)2]2+ series, and there is no clear
relationship between the temperatures of spin-crossover in
[Fe(bpp)2]2+ and [Fe(bppz)2]2+ derivatives in the solid
state.[3] With a view to clarifying these observations, we
have now extended the number of known [Fe(bppz)2]2+ de-
rivatives by derivatizing the bppz ligand at the pyrazole 4-
position, yielding a new series of ligands L2R.
Introduction
Salts of [Fe(bpp)2]2+ [bpp = 2,6-di(pyrazol-1-yl)pyridine]
and its derivatives commonly undergo thermal spin transi-
tions[1,2] at accessible temperatures, often close to room
temperature.[3] They have been particularly useful for spin-
crossover research because the bpp ligand framework can
be derivatized at any position of its heterocyclic rings.[3,4]
Substituents at the pyrazole 3-position in [Fe(bpp)2]2+ have
a strong steric and inductive influence on the coordinated
metal ion, whereas substituents at the pyrazole 4-position
(i.e., L1R) allow the ligand field to be modulated without
any steric consequences. Conversely, substitution of the pyr-
[a] School of Chemistry, University of Leeds,
Woodhouse Lane, Leeds LS2 9JT, UK
Fax: +44-113-343-6565
E-mail: m.a.halcrow@leeds.ac.uk
Homepage: www.chem.leeds.ac.uk/People/Halcrow.html
[b] CNRS, Université de Bordeaux, ICMCB,
87 avenue du Dr. A. Schweitzer, 33608 Pessac, France
Fax: +33-5-40002649
E-mail: letard@icmcb-bordeaux.cnrs.fr
Homepage: www.icmcb-bordeaux.cnrs.fr/groupes/groupe6.html
[c] School of Chemistry and Photon Science Institute, University
of Manchester,
Oxford Road, Manchester M13 9PL, UK
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201201100.
Eur. J. Inorg. Chem. 2013, 819–831
819
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 1. View of the crystal structure of L2Cl, showing the atom
numbering scheme. Thermal ellipsoids are drawn at the 50% prob-
ability level. Symmetry code: (i) 1 – y, 1 – x, ½ – z.
Results and Discussion
colorized over a period of hours and the free LR ligands
precipitated; the single crystals of L2Cl described above
were obtained by this route. Similarly, the iron complexes
of L2NO2 and L2NH2 could not be purified from uncom-
plexed ligand, even in the presence of triethyl orthoformate,
and reactions of Fe[BF4]2 and 2 equiv. L2CCPh yielded the
1:1 complex [Fe(L2CCPh)(OH2)x][BF4]2 (x = 3 or 4; see be-
low). We previously noted that [Fe(bppz)2]2+ is more sensi-
The pyrazole rings of bpp can be halogenated in good
yield under electrophilic conditions.[17,18] Similar reactions
with bppz were less successful, however, perhaps because
the more electron-deficient pyrazine ring in bppz deacti-
vates its pyrazole groups to electrophilic attack. Chlorina-
tion of bppz to give L2Cl in moderate yield was achieved
by using NaOCl in acetic acid. However, Br2/acetic acid and
I2/[NH4]2[Ce(NO3)6], which readily halogenate bpp,[17]
yielded only unchanged starting material when reacted with
bppz.
tive to protic solvents than [Fe(bpp)2]2+ [14]
It is likely that
.
the electron-withdrawing substituents in L2Cl, L2Br, L2I,
L2NO2, and L2CCPh further reduce the basicity of the pyr-
azole donors to the extent that they are easily displaced
from the iron center. Further evidence for the solution la-
bility of [Fe(L2R)2]2+ was provided by their electrospray
mass spectra, which only exhibit peaks from the free L2R
ligands with no detectable iron-containing fragments.
The ligands L2Me, L2Br, L2I, and L2NO2 were obtained
by the alternative approach of treating the appropriate de-
protonated 4-substituted pyrazoles with 0.5 equiv. 2,6-
dichloropyrazine in warm dmf.[11,12,14] These products are
less soluble than the corresponding L1R derivatives,[6,17]
and only L2Me was sufficiently soluble to be purified by
column chromatography. The other L2R ligands, including
L2Cl, were purified by washing the yellow crude materials
with dmf or ethyl acetate until all the colored impurities
had leached out of the white insoluble product. The synthe-
sis of L2NO2 is notable because L1NO2 could not be pre-
pared by this method, because the more forcing conditions
required (3–5 d at 130 °C) led to product decomposition.[19]
Reduction of L2NO2 with iron powder afforded L2NH2
with good NMR purity. Finally, coupling of L2I with
2 equiv. of phenylacetylene under Sonogashira conditions
yielded L2CCPh. No Heck reaction took place between L2I
and styrene with Pd(OAc)2/PPh3 or [Pd(PPh3)4] as catalyst,
however.
–
Attempts to prepare the ClO4 salts of these complexes led
to only [Fe(L2Me)2][ClO4]2, [Fe(L2Br)2][ClO4]2, and
[Fe(L2I)2][ClO4]2 being obtained in analytical purity. This
may again reflect the poor stability of the complexes with
electron-withdrawing pyrazole substituents in the presence
of the more nucleophilic perchlorate anion.[22]
Variable-temperature magnetic susceptibility data
–
–
showed that the BF4 and ClO4 salts of [Fe(L2Me)2]2+ are
high-spin complexes at room temperature and exhibit spin-
crossover upon cooling (Figure 2). For [Fe(L2Me)2][BF4]2,
The identity of L2Cl was confirmed by a crystal structure
determination, which showed it to adopt the expected con-
formation with near-coplanar and transoid pyrazole and
pyrazine rings (Figure 1).[8,11,17,20,21] The molecules in the
lattice associate into alternating canted sheets parallel to
(001) through a combination of electrostatic π–π interac-
tions and van der Waals contacts.
Treatment of Fe[BF4]2·6H2O with 2 equiv. L2R (R = Me
or a halogen) in nitromethane at reflux afforded [FeL2]-
[BF4]2 (L = L2Me, L2Cl, L2Br, and L2I) as orange or brown
solids following the usual work-up. The halogenated ligand
complexes are moisture-sensitive in solution and could only
be crystallized in the presence of the drying agent triethyl
Figure 2. Variable-temperature magnetic susceptibility data for
[Fe(L2Me)2][BF4]2 (᭹) and [Fe(L2Me)2][ClO4]2 (᭛). The data for
orthoformate. If this reagent was omitted, the solutions de- both salts were measured in both cooling and warming modes.
Eur. J. Inorg. Chem. 2013, 819–831
820
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
the transition occurs abruptly at T = 242 K with a 3 K with [Fe(L1Me)2][BF4]2.[6] The cations in the crystal pack
½
hysteresis loop. The temperature of the transition was con- into four-fold layers through interdigitation of the pyrazolyl
firmed by a DSC measurement with a warming temperature substituents (see the Supporting Information).[23] Although
ramp, which exhibited a first-order endotherm at 245 K this crystal packing can afford intermolecular π–π inter-
with ΔH = 23.0 kJmol–1 and ΔS = 94 Jmol–1 K–1. The form actions between these groups,[23] that is not the case in
of this transition and its thermodynamic parameters are [Fe(L2Me)2][BF4]2, whose overlapping pyrazolyl groups are
–
similar to those found for [Fe(bppz)2]X2 (X– = BF4 and separated by 3.80(1) Å. This reflects the steric influence of
– [14]
ClO4 )
and several complexes from the [Fe(bpp)2]2+ the L2Me methyl substituents. The adoption of this mode
series.[5,17] The spin transition in [Fe(L2Me)2][ClO4]2 is more of crystal packing by [Fe(L2Me)2][BF4]2 is consistent with
complicated, occurring in two steps with a discontinuity the abrupt form of its spin transition in the susceptibility
near 50% conversion (Figure 2). The midpoint tempera- data (Figure 2).[5] At 150 K, the same crystal had under-
tures of the two steps are T = 171 and 207 K, with the gone a crystallographic phase change and diffracted more
½
higher temperature step exhibiting a 3 K hysteresis loop as weakly, so the resultant dataset could not be solved.
before. A DSC analysis of this transition was not possible
Powder X-ray diffraction analysis of a bulk sample of
because its lower temperature is outside the range of our [Fe(L2Me)2][BF4]2 showed it to be phase pure and iso-
instrument.
structural with the single-crystal phase of the compound
A single-crystal X-ray analysis of [Fe(L2Me)2][BF4]2 was (Figure 4). Variable-temperature measurements showed
achieved at 300 K in its high-spin state (Figure 3, Table 1). some changes in peak intensity between 250 and 220 K,
¯
The compound adopts the space group P421c and is iso- most notably the disappearance of a moderate intensity dif-
structural with high-spin [Fe(L1Me)2][ClO4]2,[6] [Fe(L1Br)2]- fraction peak near 2θ = 22.5°. This is consistent with the
[BF4]2, and one polymorph of [Fe(L1Cl)2][BF4]2,[17] but not spin transition at 242 K in the susceptibility data (Figure 2).
These changes were fully reversed upon rewarming to room
temperature. The powder pattern of [Fe(L2Me)2][BF4]2 re-
sembles those of α-[Fe(L1Cl)2][BF4]2 and [Fe(L1Br)2][BF4]2
in the high-spin state, when they are all isostructural with
¯
the P421c space group, but not those of the low-spin-state
complexes.[17] Hence, the phase change undergone by
[Fe(L2Me)2][BF4]2 during spin-crossover must be different
to those of the [Fe(L1R)2][BF4]2 compounds, which adopt
the P21 space group when in the low-spin state.
Powder X-ray diffraction analysis of [Fe(L2Me)2][ClO4]2
showed that the chlorate salt is less crystalline, but iso-
–
structural with the BF4 salt at 290 K (Figure 4 and the
Supporting Information). The powder pattern was un-
changed at 220 K, but showed progressive changes on fur-
ther cooling to 200, 190, and 140 K (Figure 4). The changes
in the powder pattern are similar to those observed for the
–
BF4 salt, but occur over a wider temperature range. This
is again consistent with the spin-crossover shown by these
two compounds (Figure 2). Therefore the different forms of
the spin transitions in the salts of [Fe(L2Me)2]2+ do not re-
flect large differences in their structural chemistry.
Figure 3. View of the dication complex in the crystal structure of
[Fe(L2Me)2][BF4]2 at 300 K showing the atom numbering scheme
employed. Thermal ellipsoids are drawn at the 50% probability
level and H atoms have been omitted for clarity. Symmetry codes:
(ii) –1 + y, 1 – x, –z; (iii) –x, 2 – y, z; (iv) 1 – y, 1 + x, –z.
The behavior of [Fe(L2Me)2][ClO4]2 is reminiscent of
–
–
[Fe(bppz*)2]X2 (X– = BF4 and ClO4 ).[14] These adopt a
Table 1. Selected bond lengths [Å] and angles [°] for the crystal structures of [Fe(L2R)2][BF4]2 in this work. α, Σ, and Θ are indices
showing the spin state of the complex,[3,24] and θ and φ are measures of the angular Jahn–Teller distortion sometimes shown by these
iron centers in their high-spin state (see the text for details).[25,26] Typical values for these parameters in [Fe(bpp)2]2+ and [Fe(bppz)2]2+
derivatives are given in ref.[3]
L2R = L2Me
(high-spin)
L2R = L2Cl
(Molecule 1)
L2R = L2Cl
(Half-molecule 2)
L2R = L2Br
L2R = L2I
Fe–N(pyrazinyl)
Fe–N(pyrazolyl)
α
2.150(3)
2.218(2)
73.19(6)
153.7(2)
475
180
90
2.148(2), 2.165(2)
2.210(2)–2.226(3)
73.0(2)
156.0(3)
483
171.98(9)
89.23(2)
1.913(2)
2.000(3), 2.000(3)
80.0(1)
86.7(4)
285
178.48(15)
89.56(2)
1.900(3), 1.903(3)
1.981(3)–1.987(3)
80.0(3)
87.0(4)
285
178.19(11)
88.33(4)
1.922(3), 1.922(3)
1.998(3)–2.012(3)
79.8(2)
88.9(4)
291
176.85(11)
88.34(3)
Σ
Θ
φ
θ
Eur. J. Inorg. Chem. 2013, 819–831
821
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 5. (a) Temperature dependence of χMT for [Fe(L2Me)2]-
[BF4]2. (᭺) Data recorded in the cooling and warming modes with-
out irradiation, (Δ) data recorded with irradiation at 514 nm at
10 K, and (᭿) T(LIESST) measurement, data recorded in the
warming mode with the laser turned off after irradiation for 1 h.
The inset shows the derivative of the dχMT/dT vs. T curve, the
minimum of which corresponds to the T(LIESST) value. (b) Relax-
ation kinetics at various temperatures from 77.5 to 90 K. The solid
lines are simulations, see the text.
Figure 4. Variable-temperature powder X-ray diffraction patterns
for [Fe(L2Me)2]X2 [X– = BF4 (left), ClO4 (right)].
–
–
closely related crystal lattice type, in a different space group,
but also with expanded intermolecular distances within the
four-fold layers. Two-step spin transitions also occur in
these compounds, with the discontinuity corresponding to iron(II) HS state.[30] The maximum χMT values for the two
a change in X– anion disorder at 50% conversion.[16] No- compounds, reached at around 30 K, correspond to a pho-
tably, the discontinuity in [Fe(bppz*)2]X2 is more pro- toconversion efficiency of almost 90%.
nounced in the salt containing the larger perchlorate anion,
For [Fe(L2Me)2][BF4]2, the HSǞLS relaxation becomes
efficient above 50 K, and χMT decreases rapidly above
as in [Fe(L2Me)2][ClO4]2 (Figure 2).
The photomagnetic properties of [Fe(L2Me)2][BF4]2 and 80 K. The limiting temperature T(LIESST), above which
[Fe(L2Me)2][ClO4]2 were investigated by using thin layers the light-induced magnetic high-spin information is erased,
of polycrystalline samples. For both compounds, the most was determined to be 93 K from the minimum point of the
efficient wavelength to induce the LIESST effect[27] was dχMT/dT vs. T plot (Figure 5, a, inset).[29] The curve is a
found to be 514 nm, which leads to a strong increase in classical shape for mononuclear complexes with only one
the magnetic signal at 10 K (Figure 5, a and Figure 6). The crystallographic site. The LIESST relaxation kinetics were
T(LIESST) curves were then recorded for each compound systematically investigated at temperatures between 75 K
(Figure 5, a and Figure 6) to determine the stability of the and the highest temperatures accessible with our SQUID
photoinduced HS states.[28] In this procedure, the irradia- apparatus, which are close to the T(LIESST) value (Fig-
tion was maintained until the signal was saturated, then the ure 5, b). The relaxation curves strongly deviate from a sin-
light was switched off and the temperature was slowly in- gle exponential and can be modeled by using a sigmoidal
creased at 0.3 Kmin–1.[29] The increase in χMT from 10 to law describing the self-accelerating behavior predicted for
30 K is a typical signature for zero-field splitting of the strong cooperative systems. This cooperativity arises from
Eur. J. Inorg. Chem. 2013, 819–831
822
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
which confirms that the experimental kinetic parameters
used in this simulation and the fitting procedures them-
selves are appropriate.
In contrast, the T(LIESST) curve of [Fe(L2Me)2][ClO4]2
is clearly nonclassical because at least two steps are ob-
served (Figure 6). The shape of this T(LIESST) curve re-
flects the two-step character of the thermal spin-crossover
curve (Figure 6). The dχMT/dT vs. T plot reveals a first
minimum at 47 K and two minima, close together, at 92
and 98 K. The minimum at 47 K is a broad “peak” and
suggests a noncooperative relaxation, whereas the double
“peak” at 92–98 K is sharper, which indicates the presence
of cooperative interactions in the relaxation process. The
large temperature range between the two processes means
that their relaxation regimes can be treated separately. We
confirmed this by recording the relaxation kinetics at 60 K,
that is, in the plateau region separating the two processes
(Figure 7). As expected, the magnetic signal remains con-
Figure 6. Temperature dependence of χMT for [Fe(L2Me)2]-
[ClO4]2. (᭺) Data recorded in the cooling and warming modes
without irradiation, (Δ) data recorded with irradiation at 514 nm
at 10 K, and (᭿) T(LIESST) measurement, data recorded in the
warming mode with the laser turned off after irradiation for 1 h. stant for up to 8 h at 60 K, which confirms that the low-
The inset shows the derivative of the dχMT/dT vs. T curve, the
minimum of which corresponds to the T(LIESST) value. The full
perature relaxation step has not commenced at this tem-
pale- and dark-gray lines are simulations, see the text.
temperature relaxation step is complete and the high-tem-
perature. This temperature was therefore used to delimit the
T(LIESST) curve into two areas, one relating to the low-
the propagation of the large difference in metal–ligand
temperature process (Step 1) and the other to the high-tem-
bond lengths between the HS and LS states through the
perature process (Step 2; Figure 7). From the kinetic data
crystal lattice, resulting in elastic interactions caused by the
at 60 K, the baseline relating to Step 1 was recorded by
change in internal pressure inside the solid as the spin tran-
decreasing the temperature to 10 K (Figure 7) and was used
as a reference to convert the χMT values into fractions of
HS centers, γHS. The relaxation kinetics from Steps 1 and 2
sition proceeds.[31] Thus, the height of the activation barrier
to LIESST relaxation changes as a function of γHS (the
fraction of spin centers in the sample that are high spin at
are reported in Figure 8a and b, respectively.
a given temperature), and the relaxation rate k*HL(T, γHS
)
depends exponentially on both γHS and T [Equations (1)
and (2) in which a(T) (= Ea*/kBT) is the acceleration factor
at a given temperature].
ѨγHS
= –k*HLγHS
(1)
Ѩt
k*HL(T, γHS) = kHL(T)exp[a(T)(1 – γHS)]
(2)
The curves fitted with Ea* = 245 cm–1 are shown as solid
lines in Figure 5 (b). The apparent activation energy, Ea
(1670 cm–1), and the apparent pre-exponential factor, kϱ
(4.9ϫ107 s–1), of the activated region were calculated from
the straight-line fit given by the Arrhenius plot, lnkHL(T)
vs. 1/T.
The validity of these kinetic parameters was tested by
using them to reproduce the experimental T(LIESST)
curve. The procedure models both the time and temperature
dependence of the relaxation, and combines the quantum
mechanical tunneling and the thermally activated regions,
in accord with Equation (3).[28,32]
Figure 7. (᭿) T(LIESST) measurement for compound [Fe-
(L2Me)2][ClO4]2 with various kinetics (ٌ) recorded for Steps 1 and
2. (ᮀ) T(LIESST) measurement for Step 1 only and the related
baseline.
The relaxation kinetics for Step 1 are clearly of ex-
ponential shape. The data were fitted by using a stretched
exponential model that introduces a distribution, σ, of the
activation energy, Ea. The curves fitted with σ = 30 cm–1 are
shown as solid lines in Figure 8 (a). The apparent activation
kHL(T) = k0 + kϱexp(–Ea/kBT)
(3)
The rate constant, k0, characterizes the relaxation in the energy, Ea (550 cm–1), and the apparent pre-exponential
quantum tunneling region and is estimated to be an upper factor, kϱ (2.3ϫ104 s–1), were calculated from the straight-
limit from the last complete kinetic data recorded at low line fit given by the Arrhenius plot, lnkHL(T) vs. 1/T.
temperature. The agreement between the calculated and ex-
The situation with Step 2 is more complicated because
perimental T(LIESST) curves is excellent (Figure 5, a), two relaxation processes occur simultaneously. One of these
Eur. J. Inorg. Chem. 2013, 819–831
823
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[ClO4]2. The simulation includes 15% of the stretched ex-
ponential for Step 1 and 85% for Step 2, which corresponds
to the proportions determined from the T(LIESST) curve
(Figure 6). The first calculation (Figure 6, pale-gray line)
was performed by using the dynamic parameters deduced
for Step 1 (stretched exponential) and the mean values from
the two approaches for Step 2 described above (Ea
=
1620 cm–1, kϱ = 5.0ϫ107 s–1, and Ea* = 90 cm–1). The
agreement with experiment is relatively good in that the
two-step character is well reproduced and the positions of
the two T(LIESST) values are also well simulated. The
main divergence occurs in fact from the description of the
two steps occurring in the 90–100 K region. For this, we
performed a second calculation taking into account three
contributions. Step 1 was treated as before by using the de-
rived kinetic parameters and a 15% weighting on the
T(LIESST) curve. Step 2 [representing 85% of the
T(LIESST) curve] was subdivided into two independent
contributions of equal weight; 42.5% with the parameters
extracted from the stretched exponential approach and
42.5% using the parameters obtained from the sigmoidal
law. The result of the simulation is presented as the dark-
gray curve in Figure 6. This gave a much better agreement
with the description of the steps in the experimental
T(LIESST) curve because the three-step character is well
reproduced.
Be that as it may, however, whatever the aesthetics of the
T(LIESST) simulations, the only relevant conclusion from
this study is that the two simultaneous relaxation processes
Figure 8. Relaxation kinetics at various temperatures: (a) Step 1
from 42.5 to 49 K and (b) Step 2 from 60 to 97.5 K. The solid lines
are simulations, see the text.
in Step 2 have similar dynamic parameters with Ea
≈
1600 cm–1 and kϱ ≈ 5.0ϫ107 s–1. They are distinguished by
the presence of cooperativity in one relaxation (Ea*
is cooperative and can be described by a sigmoidal law, ≈ 180 cm–1) and a distribution of activation energies in the
whereas the other is noncooperative and follows an ex- other (σ = 25 cm–1).
ponential shape, probably stretched. The simultaneous na-
Single-crystal structures were achieved from solvated
ture of the relaxations leads to very difficult simulations crystals of the BF4– salts of all the halogenated ligand com-
with many parameters. Hence we were only able to describe plexes. The formation of solvates by these compounds con-
the global relaxation rate of Step 2, which we performed trasts with [Fe(L1R)2][BF4]2 (R = Cl and Br), which adopt
by using two approaches. First, the relaxation curves were solvent-free phases under the same crystallization condi-
simulated by a stretched exponential law. This does not pro- tions (see above).[5,17] Crystals of [Fe(L2Cl)2][BF4]2·
vide information on the cooperative character of the kinet- 3.33CH3NO2·(C2H5)2O contain substantial anion and sol-
ics, but usually gives quite relevant values for the relaxation vent disorder, including channels of disordered solvent, but
rates because a stretched exponential equates to a simulta- a good refinement of this material was achieved at 100 K.
neous multiple-relaxation law. The simulations are pre- Its asymmetric unit contains 1.5 formula units with a half-
sented in Figure 8 (b) and the dynamic parameters ex- molecule of the complex spanning a crystallographic C2
tracted from the lnkHL(T) vs. 1/T plot are Ea = 1640 cm–1, axis. From its metric parameters, the whole molecule is
kϱ = 7.5ϫ107 s–1, and σ = 25 cm–1. The second approach clearly high spin at this temperature, whereas the half-mole-
consisted of trying to simulate the cooperative character of cule is low spin (Table 1).
the curve by using a sigmoidal law (see the Supporting In-
The magnetic susceptibility measurements on dried, sol-
formation). The sample behavior is not reproduced on long vent-free [Fe(L2Cl)2][BF4]2 are consistent with the crystal
timescales by this method, but nonetheless, the dynamic pa- structure in showing an almost constant value of χMT =
rameters obtained are close to those obtained from the 2.4–2.6 cm3 Kmol–1 between 50–300 K (plotted in the Sup-
stretched exponential simulation (Ea = 1600 cm–1, kϱ
=
porting Information; χMT decreases upon further cooling
2.0ϫ107 s–1, and Ea* = 180 cm–1). They are also compar- owing to zero-field splitting of the high-spin iron center[27]).
able to the values obtained for [Fe(L2Me)2][BF4]2 (see This is close to the predicted value for a sample comprising
above), which lends them additional credit.
two-thirds high-spin and one-third low-spin iron(II) centers
The validity of these results was strengthened by simula- (ca. 2.3 cm3 Kmol–1),[33] which shows that the spin-state
tion of the multistep T(LIESST) curve of [Fe(L2Me)2]- population of the compound does not change significantly
Eur. J. Inorg. Chem. 2013, 819–831
824
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
upon drying. Moreover, little or none of the high-spin con- 330 K to 2.3 cm3 Kmol–1 at 30 K indicates a very gradual
tent of the sample undergoes spin-crossover upon cooling. partial spin-crossover in approximately 20% of the iron cen-
High-spin complexes of this type can exhibit an angular ters in the sample over this temperature range. Thus,
Jahn–Teller distortion, which inhibits spin-crossover in the [Fe(L2Br)2][BF4]2·nCH3NO2 transforms fully from a low-
solid state.[3,25,26,34] This is manifested through a reduction spin crystalline solvate (n = 3) into a high-spin powder (n
in the trans-N(pyridyl)–Fe–N(pyridyl) angle (φ) from its = 1), but this change in spin state is incomplete upon drying
ideal value of 180° and/or a twisting of the two tridentate [Fe(L2I)2][BF4]2·nCH3NO2. Solid [Fe(L2Br)2][ClO4]2 and
ligands away from the perpendicular (θ Ͻ 90°, θ is the dihe- [Fe(L2I)2][ClO4]2 are also high spin at room temperature
dral angle between the least-squares planes of the two li- and exhibit very gradual, partial spin-crossover on cooling,
gands; Figure 9). The angle φ in molecule 1 of [Fe(L2Cl)2]- but single crystals of these salts were not obtained.
[BF4]2 [171.98(9)°, Table 1] implies a weak distortion of this
The influence of ligand substituents on spin-crossover in
type, being close to the threshold value below which spin- the absence of lattice effects can be probed by solution-
crossover is not observed (Figure 9).[3] An alternative expla- phase measurements.[36,37] The variable-temperature Evans
nation, that spin-crossover is prevented by intermolecular method was used to study the spin-state behavior of
steric contacts in the crystal lattice,[34,35] seems less likely. [Fe(L2R)2][BF4]2 (R = Me, Cl, and Br) and [Fe(bppz)2]-
The only noteworthy intermolecular interactions involving [BF4]2 (Figure 10).[38] Data for [Fe(bppz)2][BF4]2 were ob-
molecule 1 are C–H···F contacts of 2.3–2.5 Å between the tained in (CD3)2CO to maximize the temperature range of
C–H of the pyrazolyl 3-position and disordered anions in the study, but the other compounds were measured in
the lattice. Comparable cation···anion contacts are also CD3NO2 for reasons of solubility. The use of these dif-
present in [Fe(L2Me)2][BF4]2 and do not prevent spin-cross- ferent, weakly interacting solvents is unlikely to influence
over in that material.
the results of the study.[37] Solutions of [Fe(bppz)2][BF4]2
and [Fe(L2Me)2][BF4]2 exhibited the expected gradual spin-
state equilibria; their T values were estimated to be 268
½
and 291 K, respectively (Figure 10). These values are both
around 20 K higher than those of the corresponding com-
plexes of the [Fe(L1R)]2+ series,[25,39] and the higher T
½
value for [Fe(L2Me)2][BF4]2 reflects the electron-donating
character of its methyl substituents. The enthalpies of the
spin-crossover derived from these analyses are 20.2 kJmol–1
for [Fe(bppz)2][BF4]2 and 25.3 kJmol–1 for [Fe(L2Me)2]-
[BF4]2. These are typical values for this class of complex
and show that the compounds do not undergo significant
solvolysis under these conditions.[40]
Figure 9. Plot of the Jahn–Teller distortion indices θ vs. φ for pub-
lished complexes of the [Fe(bpp)2]2+ and [Fe(bppz)2]2+ series[3]
showing compounds that remain high spin on cooling (᭹), spin-
crossover complexes in their high-spin state (᭛), and low-spin com-
plexes (᭡). The gray arrow indicates the high-spin molecule in the
solvate crystal structure of [Fe(L2Cl)2][BF4]2.
In contrast, [Fe(L2Br)2]2+ and [Fe(L2I)2]2+ form tris-ni-
tromethane solvate crystals that are isostructural with each
other but not with the [Fe(L2Cl)2][BF4]2 solvate. Although
there are small differences in their Fe–N distances, both
compounds are clearly in a low-spin state at the tempera-
ture of measurement, 150 K (Table 1). The molecules show
only minor deviations from their ideal D2d symmetry and
associate into layers in the crystal that are related to the
terpyridine embrace lattice structure.[23]
Figure 10. Solution-phase magnetic susceptibility data for
[Fe(bppz)2][BF4]2 in (CD3)2CO (᭺) and [Fe(L2Me)2][BF4]2 in
CD3NO2 (᭹).
Drying these dark-brown crystalline materials in vacuo
resulted in their decomposition into yellow powders. These
powders are mostly solvent-free, although solid [Fe(L2Br)2]-
In contrast, data derived from the Evans method for
[BF4]2 may contain some residual nitromethane by elemen- [Fe(L2Cl)2][BF4]2 and [Fe(L2Br)2][BF4]2 in CD3NO2 show
tal microanalysis. Variable-temperature magnetic measure- χMT values that are almost constant between 253–323 K,
ments showed that [Fe(L2Br)2][BF4]2 is fully high spin at and that lie between the values expected for a high-spin and
1
330 K and below (plotted in the Supporting Information). low-spin complex. In the latter case, H NMR peaks from
Although [Fe(L2I)2][BF4]2 is also predominantly high spin the free L2Br ligand also became resolved at higher tem-
at 330 K, a steady decrease in χMT from 3.0 cm3 Kmol–1 at peratures. This indicates substantial decomposition of the
Eur. J. Inorg. Chem. 2013, 819–831
825
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
complexes in this weakly associating solvent to the extent for more details). The apparent loss of a L2CCPh ligand
that insufficient intact [Fe(L2R)2]2+ centers are present for during the crystallization process again emphasizes the
their spin transitions to be measured. This is consistent with solution lability of the [Fe(L2R)2]2+ complexes.
the above observations regarding the solution stability of
these complexes, but contrasts with [Fe(L1R)2]2+ (R = Cl
Conclusions
and Br), which are stable in solution under the same condi-
tions.[17] The electron-withdrawing halide substituents on
the L2Cl and L2Br ligands reduce the basicity of the pyr-
azole groups,[41] which makes their complexes more reactive
in solution for two reasons. First, the resultant weaker li-
gand field increases the high-spin population of the com-
plexes at a particular temperature. Secondly, the Fe–N
bonds in the complex are more kinetically labile, and so
more reactive towards solvent.
Several bppz derivatives substituted at the pyrazolyl C-
4 positions (L2R) have been prepared. A wider range of
derivatives could be accessed than for the corresponding
2,6-di(pyrazol-1-yl)pyridine ligand series (L1R) because the
standard L2R synthesis employs milder conditions. How-
ever, the [Fe(L2R)2]2+ complexes were harder to isolate than
[Fe(L1R)2]2+ because the L2R ligands dissociate more easily
from the iron center in solution and are less soluble. This
led to hydrolysis of the complexes and ligand precipitation
to the extent that [Fe(L2NO2)2]2+, [Fe(L2NH2)2]2+, and
[Fe(L2CCPh)2]2+ could not be obtained in pure form. The
solution lability of [Fe(L2R)2]2+ was also evident from their
NMR behavior when R is an electron-withdrawing halo
substituent.
Attempts to crystallize [Fe(L2CCPh)2][BF4]2 from un-
dried MeNO2/Et2O yielded yellow prisms. The asymmetric
unit contains half a [Fe(L2CCPh)(OH2)2][BF4]2·(C2H5)2O
moiety spanning a crystallographic C2 axis together with
three closely spaced Fourier peaks 2.1–2.2 Å from the iron
atom that refined reasonably as partial oxygen sites (Fig-
ure 11). Therefore this complex was interpreted as being a
random co-crystal of seven-coordinate [Fe(L2CCPh)-
(OH2)4]2+ and six-coordinate [Fe(L2CCPh)(OH2)3]2+ in a
2:3 occupancy ratio, which is consistent with the microanal-
ysis of the dried material (see the Supporting Information
Comparison of the [Fe(L1R)2]X2 and [Fe(L2R)2]X2 salts
in the solid state is difficult because these pairs of com-
plexes are rarely isostructural for the same “R” group and
anion. Thus, for example, [Fe(L2Cl)2][BF4]2 and [Fe(L2Br)2]-
[BF4]2 form solvated crystals that are not spin-crossover
active, which contrasts with the spin transitions undergone
by solvent-free, crystalline [Fe(L1Cl)2][BF4]2 and [Fe-
(L1Br)2][BF4]2.[5,17] However, solution data imply that the
pyrazinyl groups in [Fe(L2R)2]2+ (R = H or Me) thermody-
namically stabilize the low-spin state of the complex to a
small extent compared with the pyridyl donors in [Fe-
(L1R)2]2+. This is consistent with other studies of spin-
crossover in pyrazine-containing complexes.[42] This result
is counter-intuitive because the lower basicity of the pyr-
azine ring should lead to weaker NǞFe dative bonds. Pre-
sumably, therefore, stronger FeǞL2R back-bonding to the
more electron-deficient pyrazine ring has a more significant
effect on the d-orbital splitting of the iron center.
Although they are isostructural in both spin states, as
determined by powder X-ray diffraction analysis, the salts
of [Fe(L2Me)2][BF4]2 and [Fe(L2Me)2][ClO4]2 show very dif-
ferent thermal and light-induced spin-transition behavior.
Thermal spin-crossover in [Fe(L2Me)2][BF4]2 occurs in one
step at T = 242 K, which is associated with a T(LIESST)
½
value of 93 K. This is broadly consistent with the previously
reported relationship between T and T(LIESST) for
½
[Fe(L1R)2]2+ and [Fe(L2R)2]2+ complexes [Equation (4), T0
= 150 K].[6,17,43]
T(LIESST) = T0 – 0.3T
(4)
½
In contrast, thermal spin-crossover in [Fe(L2Me)2]-
[ClO4]2 occurs in two steps at 171 and 207 K, which is
largely reflected by the stepped T(LIESST) curve. The re-
laxation behavior of the photoinduced phase was found to
be even more complicated and was tentatively simulated. A
first attempt to relate these data to Equation (4) is to corre-
Figure 11. View of the crystal structure of the complex molecule of
[Fe(L2CCPh)(OH2)4]x[Fe(L2CCPh)(OH2)3]1–x[BF4]2·(C2H5)2O (x ≈
0.4). Thermal ellipsoids are drawn at the 50% probability level, and
carbon-bound H atoms have been omitted for clarity. The complex
is seven-coordinate when O(20A) and O(20Avi) are both occupied,
and six-coordinate with either O(20B), O(20C), or O(20Cvi) occu-
pied. Symmetry code: (vi) –x, y, 3/2 – z.
late the highest thermal T (207 K) with the lowest
½
Eur. J. Inorg. Chem. 2013, 819–831
826
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
trospray mass spectra (ESI MS) were obtained with a Waters
ZQ4000 spectrometer from MeCN feed solutions. All mass peaks
have the correct isotopic distributions for the proposed assign-
ments. Powder X-ray diffraction analyses were performed with a
Bruker D8 Advance A25 diffractometer using Cu-Kα radiation
(λ = 1.5418 Å).
T(LIESST) value (47 K), and the second step of the ther-
mal spin transition at 171 K with the T(LIESST) value near
95 K. Notably, the abrupt nature of the first step of the spin
transition does not match the gradual shape of the low-
temperature step of the T(LIESST) curve. Nonetheless,
structural rearrangements occurring during the cooling
and/or photoexcitation processes can give rise to such ob-
servations. For example, this has been observed in [FeL2]-
(ClO4)2 {L = 2-[3-(2Ј-pyridyl)pyrazol-1-ylmethyl]pyridine},
in which competition between cooperative interactions and
structural disorder leads to an abrupt thermal spin-cross-
over and a gradual T(LIESST) curve.[44] As described
above, the two-step thermal spin-crossover in [Fe(L2Me)2]-
[ClO4]2 could be controlled by changes to perchlorate anion
disorder during the transition,[14,16] or it could reflect the
lower crystallinity of the perchlorate salt, which will lead to
a more heterogeneous distribution of [Fe(L2Me)2]2+ sites in
the lattice. Both these hypotheses could explain the LIESST
relaxation behavior of the compound, showing multiple re-
laxation pathways with similar activation energies but dif-
ferent cooperativities. Unfortunately, the absence of crystal-
lographic data for [Fe(L2Me)2][ClO4]2 prevents us from vali-
dating these suggestions.
Magnetic susceptibility measurements were performed by using a
Quantum Design SQUID magnetometer in an applied field of 1000
or 5000 G. A diamagnetic correction for the sample was estimated
from Pascal’s constants;[33] a diamagnetic correction for the sample
holder was also used. Photomagnetic measurements were per-
formed by using a Spectrum Physics Series 2025 Kr+ laser (λ =
514 nm) coupled via an optical fibre to the cavity of a MPMS-55
Quantum Design SQUID magnetometer. The optical power at the
sample surface was adjusted to 5 mWcm–2, and it was verified that
this resulted in no change in magnetic response due to heating of
the sample. Photomagnetic samples consisted of a thin layer of
compound, the weight of which was determined by comparison of
the thermal spin-crossover curve with that of a more accurately
weighed sample of the same material. Susceptibility measurements
in solution were obtained by Evans method using a Bruker
DRX500 spectrometer operating at 500.13 MHz.[38] A diamagnetic
correction for the sample[33] and a correction for the variation of
the density of the solvent with temperature[47] were applied to these
data.
A second approach to understanding [Fe(L2Me)2]-
Materials and Methods: Unless otherwise stated, all reactions were
carried out in air and in non-pre-dried AR-grade solvents. 2,6-Di-
(pyrazol-1-yl)pyrazine (bppz) was synthesized by the literature
method[11] and all other reagents were used as commercially sup-
plied.
[ClO4]2 by using Equation (4) is to consider that the T
½
value of 207 K is correlated with the T(LIESST) value of
95 K, and that the T step at 171 K with the T(LIESST)
½
value of 47 K. In this case the cooperative (and reversely,
the gradual) thermal spin-transition curve is associated with
the cooperative (gradual) T(LIESST) regime. Such behavior
can be linked to changes in the coordination sphere of two
different crystallographic sites, which has been, for example,
observed for [Fe(NCS)2(PM-BiA)2] [PM-BiA = N-(2Ј-pyr-
idylmethylene)-4-aminobiphenyl], which possesses two dis-
tinct isomorphs.[32,45]
Synthesis of 2,6-Bis(4-chloropyrazol-1-yl)pyrazine (L2Cl): A solu-
tion of bppz (1.1 g, 5.2 mmol) in glacial acetic acid (70 cm3) was
added to a 5% aqueous solution of NaOCl (40 cm3) diluted in
100 cm3 of water. The yellow mixture was stirred at room tempera-
ture for 2 h and further heated for 3 h at 70 °C. The mixture was
poured onto ice–water (200 cm3) and the resultant precipitate ex-
tracted into CHCl3 (200 cm3, then 3ϫ 30 cm3). The combined or-
ganic layers were washed with saturated aqueous K2CO3 (80 cm3)
and water (80 cm3), and then dried with MgSO4. The solvent was
removed under reduced pressure to yield a pale-yellow solid, which
was washed with dimethylformamide and diethyl ether to afford a
white powder, yield 0.67 g, 46%, m.p. 146–147 °C. 1H NMR
[300 MHz, (CD)3SO]: δ = 8.02 (s, 2 H, Pz 3-H), 9.01 and 9.04 (both
s, 2 H, Pz 5-H and Pyz 3,5-H) ppm. MS (ESI): m/z = 281.9
[HLCl]+. C10H6Cl2N6 (281.11): calcd. C 42.7, H 2.15, N 30.0; found
C 43.1, H 2.10, N 30.0.
Whichever explanation is correct, if we relate the result-
ant T(LIESST)/T correlations to the T(LIESST) data-
½
base,[28] the stability of the photoinduced high-spin state of
[Fe(L2Me)2][ClO4]2 does not lie on the same T0 line as other
complexes of this type.[6,17,43] This is only the second com-
pound from the [Fe(bpp)2]2+ series that deviates from the
T(LIESST) vs. T relationship [Equation (4)] to be iden-
½
tified.[46] This point is particularly interesting because it
may allow identification of the factors required to increase
the stability of a photoinduced high-spin state at room tem-
perature.
2,6-Bis(4-bromopyrazol-1-yl)pyrazine (L2Br): A solution of 4-bro-
mopyrazole (5.00 g, 34 mmol) in dry dmf (100 cm3) was added to
NaH (0.90 g, 37 mmol) under N2, and the resultant suspension was
stirred at 50 °C for 20 min. Solid 2,6-dichloropyrazine (2.43 g,
16 mmol) was then added in one portion and the mixture was
stirred at 90 °C for a further 16 h. After cooling, a large excess of
cold water was added to the mixture to yield a yellow solid. The
solid was collected and washed with ethyl acetate to yield a white
Experimental Section
1
powder, yield 1.95 g, 31%, m.p. 140–142 °C. H NMR [300 MHz,
Instrumentation: Elemental microanalyses were performed by the
University of Leeds School of Chemistry microanalytical service.
IR spectra were recorded as Nujol mulls pressed between NaCl
windows between 600–4,000 cm–1 by using a Nicolet Avatar 360
(CD)3SO]: δ = 7.74 (s, 2 H, Pz 3-H), 8.50 (s, 2 H, Pz 5-H), 9.17 (s, 2
H, Pyz 3,5-H) ppm. MS (ESI): m/z = 392.9 [NaLBr]+. C10H6Br2N6
(370.01): calcd. C 32.5, H 1.63, N 22.7% found C 32.2, H 1.55, N
22.3.
2,6-Bis(4-iodopyrazol-1-yl)pyrazine (L2I): Method as for L2Br, but
by using 4-iodopyrazole (6.60 g, 34 mmol). The crude yellow solid
was washed with ethyl acetate to afford the pure product as a white
1
spectrophotometer. H NMR spectra were recorded with a Bruker
DPX300 spectrometer operating at 300.2 MHz. UV/Vis/NIR spec-
tra were recorded with a Perkin–Elmer Lambda900 spectropho-
tometer in 1 cm quartz solution cells between 200–3000 nm. Elec-
Eur. J. Inorg. Chem. 2013, 819–831
827
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
1
powder, yield 1.67 g, 21%, m.p. 140–142 °C. H NMR [300 MHz, (ca. 3 h). The cooled solution was concentrated in vacuo to around
(CD)3SO]: δ = 8.01 (s, 2 H, Pz 3-H), 9.13 (s, 2 H, Pz 5-H), 9.25 (s,
5 cm3. Slow diffusion of diethyl ether vapor into the filtered solu-
2 H, Pyz 3,5-H) ppm. MS (ESI): m/z = 486.9 [NaLI]+. C10H6I2N6 tion afforded orange crystals of the product. The other complex
(464.01): calcd. C 25.9, H 1.30, N 18.1; found C 25.6, H 1.20, N
17.7.
salts were prepared using appropriate amounts of the relevant L2R
ligand and/or Fe[ClO4]2·6H2O, as required. Yields after recrystalli-
zation ranged from 42–77%. Microanalytical data for the com-
plexes are listed below.
2,6-Bis(4-methylpyrazol-1-yl)pyrazine (L2Me): Method as for L2Br,
but by using 4-methylpyrazole (2.79 g, 34 mmol). The resultant yel-
low solid was purified by flash silica column chromatography in
dichloromethane/ethyl acetate/hexane (1:1:2) to give LMe as a white
[Fe(L2Me)2][BF4]2: C24H24B2F8FeN12 (709.99): calcd. C 40.6, H
3.41, N 23.7; found C 40.7, H 3.35, N 23.7.
1
powder, yield 1.89 g, 46%, m.p. 147–148 °C. H NMR [300 MHz,
[Fe(L2Me)2][ClO4]2: C24H24Cl2FeN12O8 (735.28): calcd. C 39.2, H
3.29, N 22.9; found C 39.4, H 3.25, N 23.0.
(CD)3SO]: δ = 2.18 (s, 6 H, CH3), 7.61 (s, 2 H, Pz 3-H), 8.24 (s, 2
H, Pz 5-H), 9.08 (s, 2 H, Pyz 3,5-H) ppm. MS (ESI): m/z = 263.1
[NaLMe]+. C12H12N6 (240.27): calcd. C 60.0, H 5.03, N 35.0; found
C 60.0, H 5.00, N 35.2.
[Fe(L2Cl)2][BF4]2: C20H12B2Cl4F8FeN12 (791.67): calcd. C 30.3, H
1.53, N 21.2; found C 30.2, H 1.72, N 21.2.
2,6-Bis(4-nitropyrazol-1-yl)pyrazine (L2NO2): Method as for L2Br,
but by using 4-nitropyrazole (3.85 g, 34 mmol). This yielded a yel-
low powder, which was purified by washing with ethyl acetate to
give a white solid, yield 1.72 g, 33%, m.p. 151–153 °C. 1H NMR
[300 MHz, (CD)3SO]: δ = 8.84 (s, 2 H, Pz 3-H), 9.35 (s, 2 H, Pyz
3,5-H), 10.49 (s, 2 H, Pz 5-H) ppm. MS (ESI): m/z = 302.3
[LNO2]+. C10H6N8O4 (302.23): calcd. C 39.7, H 2.00, N 37.1; found
C 39.8, H 1.95, N 37.1.
[Fe(L2Br)2][BF4]2·CH3NO2:
(1030.51): calcd. C 24.4, H 1.47, N 17.7; found C 24.3, H 1.75, N
17.5.
C20H12B2Br4F8FeN12·CH3NO2
[Fe(L2Br)2][ClO4]2: C20H12Br4Cl2FeN12O8 (994.76): calcd. C 24.1,
H 1.22, N 16.9; found C 23.9, H 1.50, N 16.8.
[Fe(L2I)2][BF4]2: C20H12B2F8FeI4N12 (1157.47): calcd. C 20.8, H
1.05, N 14.5; found C 20.8, H, 1.04, N 14.5.
2,6-Bis(4-aminopyrazol-1-yl)pyrazine (L2NH2): A mixture of L2NO2
(0.66 g, 2.2 mmol), iron powder (0.73 g, 3.7 mmol), and ammonium
chloride (1.17 g, 21.9 mmol) was suspended in a ethanol/water (2:1,
20 cm3) under N2. The mixture was heated at reflux for 30 min.
The black mixture was filtered whilst hot and the yellow filtrate
concentrated to around 5 cm3 under reduced pressure. Storage at
0 °C for 1 h afforded a yellow precipitate, which was collected, re-
peatedly washed with water, and dried. Although there were no
organic impurities in the product by 1H NMR, its poor microanaly-
sis may indicate a minor inorganic salt contaminant that we were
unable to remove, yield 0.20 g, 56%. 1H NMR [300 MHz,
(CD)3SO]: δ = 5.60 (br. s, 4 H, NH2), 7.71 and 7.89 (both s, 2 H,
Pz 5-H and 5-H), 9.01 (s, 2 H, Pyz 3,5-H) ppm. MS (ESI): m/z =
265.5 [NaLNH2]+. C10H10N8 (242.24): calcd. C 49.6, H 4.16, N 46.3;
found C 48.0, H 4.10, N, 43.2.
2,6-Bis[4-(phenylethynyl)pyrazol-1-yl]pyrazine Hemihydrate (L2CCPh·
½H2O): This ligand was prepared under N2. L2I (0.36 g,
0.78 mmol) was dissolved in fresh triethylamine (10 cm3) and diox-
ane (2 cm3), and the mixture was stirred for 15 min. [Pd(PPh3)2Cl2]
(55 mg, 0.078 mmol), PPh3 (41 mg, 0.156 mmol), and CuI (20 mg,
0.1 mmol) were then added, followed by phenylacetylene (0.35 g,
3.5 mmol). The resultant mixture was heated at 80 °C for 30 min
until it became a dark orange. It was then cooled to room tempera-
ture and left to stir for 1 h during which a light-orange product
precipitated. The mixture was neutralized in an ice-bath with dilute
HCl and the product was extracted with CH2Cl2 (3ϫ 20 cm3). The
orange organic phase was collected, washed with a saturated solu-
tion of NH4Cl, and dried with MgSO4. The solvent was removed
under reduced pressure to give a dark-orange solid, which was
washed with ethanol (50 cm3) to give a white powder, yield 0.26 g,
78%, m.p. 126–128 °C. 1H NMR [300 MHz, (CD)3SO]: δ = 7.47
(m, 3 H, Ph 3,4,5-H), 7.56 (m, 2 H, Ph 2,6-H), 8.38 (s, 2 H, Pz 3-
H), 8.50 (s, 2 H, Pz 5-H), 9.56 (s, 2 H, Pyz 3,5-H) ppm. MS (ESI):
m/z = 413.1 [HLCCPh]+. C26H16N6·½H2O (421.46): calcd. C 74.1,
H 4.07, N 19.9; found C 73.7, H 3.80, N 19.5.
[Fe(L2I)2][ClO4]2: C20H12Cl2FeI4N12O8 (1182.76): calcd. C 20.3, H
1.02, N 14.2; found C 20.3, H 1.35, N 14.6.
[Fe(L2CCPh)(OH2)4]x[Fe(L2CCPh)(OH2)3]1–x[BF4]2 (x
≈
0.5):
Complexation of Fe[BF4]2·6H2O (0.14 g, 0.42 mmol) with L2CCPh
(0.34 g, 0.83 mmol, 2 equiv.) following the above procedure gave
this 1:1 metal/ligand product as a yellow solid, yield 0.20 g, 45%.
(C26H24B2F8FeN6O4)0.5(C26H22B2F8FeN6O3)0.5 (704.96): calcd. C
44.3, H 3.29, N 11.9; found 44.8, H 3.80, N 11.3.
Single-Crystal Structure Analyses: Single crystals of the complexes
were all obtained by slow diffusion of diethyl ether vapor into ni-
tromethane solutions of the compounds. All diffraction data were
collected with a Bruker X8 Apex diffractometer using graphite-
monochromated Mo-Kα radiation (λ = 0.71073 Å) generated by a
rotating anode. The structure was solved by direct methods
(SHELXS-97)[48] and then developed by least-squares refinement
on F2 (SHELXL-97).[48] Crystallographic figures were prepared by
using XSEED,[49] which incorporates POVRAY.[50] Experimental
data obtained from the structure determinations are presented in
Table 2.
Structure Refinement of L2Cl: The asymmetric unit contains half a
molecule with N(1) and N(4) lying on the crystallographic C2 axis
[x, 1 – x, ¼]. No restraints were applied in the final model. All non-
H atoms were refined anisotropically. All H atoms were located in
the Fourier map and allowed to refine freely with a common Uiso
thermal parameter of 0.027(3) Å2. The refined C–H distances were
in the range of 0.92(2)–1.01(2) Å.
Structure Refinement of [Fe(L2Me)2][BF4]2: The compound was
originally solved in the space group P21 and then transformed into
P421c by using the ADSYMM routine in PLATON.[51] The crystal
¯
was refined as a racemic twin. The asymmetric unit contains a
quarter of the complex dication, with Fe(1) occupying the crystal-
lographic S4 site [0, 1, 1] and N(2) and N(5) lying on the C2 axis
–
[0, 1, z], and half a BF4 anion that is disordered about the C2
Synthesis of the Complexes: The same basic method, described here
for [Fe(L2Me)2][BF4]2, was used for the synthesis of all the com-
plexes. A solution of L2Me (0.20 g, 0.83 mmol) and Fe[BF4]2·6H2O
(0.14 g, 0.42 mmol) in nitromethane (15 cm3) and triethyl orthofor-
mate (3 drops) was heated at reflux until all the solid had dissolved
axis [0, ½, z]. Two equally occupied partial environments for this
disordered half-anion were refined subject to the refined restraints
B–F 1.40(2) Å and F···F 2.29(2) Å. All non-H atoms except the
disordered anion were refined anisotropically and all H atoms were
placed in calculated positions and refined by using a riding model.
Eur. J. Inorg. Chem. 2013, 819–831
828
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 2. Experimental data for the crystal structure determinations in this work.
L2Cl
[Fe(L2Me)2][BF4]2
[Fe(L2Cl)2][BF4]2·3.33CH3NO2·(C2H5)2O
Formula
C10H6Cl2N6
281.11
tetragonal
P41212
5.1884(3)
–
41.708(3)
–
1122.77(13)
150(2)
4
C24H24B2F8FeN12
710.02
C27.33H32B2Cl4F8FeN15.33O7.67
Mr [gmol–1]
Crystal system
Space group
a [Å]
1069.29
monoclinic
C2/c
40.025(4)
17.0702(17)
19.958(2)
101.775(4)
13349(2)
100(2)
12
tetragonal
¯
P421c
9.7297(14)
–
17.5425(16)
–
1660.7(4)
300(2)
2
b [Å]
c [Å]
β [°]
V [Å3]
T [K]
Z
Dcalcd. [gcm–3]
1.663
0.567
1.420
0.535
1.596
0.675
μ [mm–1]
Min./max. transmission
θmax [°]
Measured reflections
Unique reflections
Reflections [Fo Ͼ 4σ(Fo)]
Rint
0.770/0.902
30.61
27746
1744
1727
0.043
0.635/0.900
27.60
20713
1926
1561
0.061
0.699/0.851
28.50
144862
16845
13010
0.064
Parameters
93
0.031
0.076
128
0.039
0.112
983
0.058
0.200
[a]
R1 [Fo Ͼ 4σ(Fo)]
[b]
wR2 [all data]
Gof
1.286
0.32/–0.25
0.01(9)
1.062
0.25/–0.24
0.46(3)
1.074
1.86/–0.77
–
Δρmax./Δρmin. [eÅ–3]
Flack parameter
[Fe(L2Br)2][BF4]2·3CH3NO2 [Fe(L2I)2][BF4]2·3CH3NO2
[Fe(L2CCPh)(OH2)4]x [Fe(L2CCPh)(OH2)3]1–x
[BF4]2 (x = 0.4)
-
Formula
C23H21B2Br4F8FeN15O6
1152.66
monoclinic
P21/c
17.678(4)
17.144(3)
14.180(3)
113.26(3)
3948.2(14)
150(2)
C23H21B2I4F8FeN15O6
1340.62
monoclinic
P21/c
18.320(3)
17.445(3)
14.502(2)
112.866(6)
4270.6(11)
150(2)
C30H32.8B2F8FeN6O4.4
777.29
monoclinic
C2/c
18.9558(15)
13.6788(12)
17.928(2)
120.731(3)
3995.9(7)
150(2)
Mr [gmol–1]
Crystal system
Space group
a [Å]
b [Å]
c [Å]
β [°]
V [Å3]
T [K]
Z
4
4
4
Dcalcd. [gcm–3]
1.939
4.527
2.085
3.335
1.292
0.455
μ [mm–1]
Min./max. transmission
θmax [°]
Measured reflections
Unique reflections
Reflections [Fo Ͼ 4σ(Fo)]
Rint
0.644/1.029
27.53
75061
9056
6445
0.574/0.979
29.02
166820
11348
9191
0.075
0.763/0.972
28.46
53701
5056
4110
0.077
0.050
Parameters
553
0.036
0.084
518
0.033
0.080
257
0.086
0.283
[a]
R1 [Fo Ͼ 4σ(Fo)]
[b]
wR2 [all data]
Gof
1.008
0.77/–0.57
–
1.031
1.26/–1.21
–
1.096
1.82/–0.56
–
Δρmax./Δρmin. [eÅ–3]
Flack parameter
2
2 2
½
[a] R1 = Σ||Fo| – |Fc||/Σ|Fo|. [b] wR2 = {Σ[w(Fo – Fc ) ]/Σ[w(Fo2)2]} .
Structure Refinement of [Fe(L2Cl)2][BF4]2·3.33CH3NO2·(C2H5)2O: located overlying the two half-occupied nitromethane sites. Finally,
The asymmetric unit contains 1.5 formula units with one half-mo-
lecule of the complex spanning the crystallographic C2 axis [0, y,
¼]. All three of the unique BF4– ions are disordered and were mod-
eled over two or three sites by using the refined restraints B–F
1.40(2) Å and F···F 2.29(2) Å. Five nitromethane solvent sites were
also identified. Three of these are wholly occupied and the other
two are half-occupied. A half-molecule of diethyl ether was also
a badly disordered region near the inversion center at [0, ½, ½]
forms channels parallel to the unit cell c axis of approximate di-
mensions 5.0ϫ3.3 Å (measured from the ordered solvent mole-
cules lining the edges of the channels). A SQUEEZE analysis[51]
found four voids of 445 Å3 each per unit cell, containing a total of
602 electrons. That equates to 151 electrons per void, or 75 elec-
trons per asymmetric unit, which could correspond to one molecule
Eur. J. Inorg. Chem. 2013, 819–831
829
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
of nitromethane (32 electrons) plus one molecule of diethyl ether
(42 electrons). This formula was used for the density and F(000)
calculations.
anisotropically and all carbon-bound H atoms were placed in cal-
culated positions and refined by using a riding model. The water
H atoms bound to O(19) were located in the Fourier map, and this
water molecule was refined as a rigid group in the final least-
squares cycles. The H atoms bound to O(20A)–O(20C) were not
located, but are included in the density calculation.
CCDC-899454 {for [Fe(L2Cl)2][BF4]2·3.33CH3NO2·(C2H5)2O},
-899455 {for [Fe(L2Me)2][BF4]2}, -899456 {for [Fe(L2I)2][BF4]2·
3CH3NO2}, -899457 {for [Fe(L2Br)2][BF4]2·3CH3NO2}, -899458
{for [Fe(L2CCPh)(OH2)4]x[Fe(L2CCPh)(OH2)3]1–x[BF4]2·(C2H5)2-
O}, and -899459 (for L2Cl) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
The original dataset, rather than that obtained by SQUEEZE, was
used for the final refinement cycles. The more intense Fourier peaks
in the disordered channels were refined as 0.2-occupied C atoms,
but no attempt was made to assign these to individual partial sol-
vent sites. All wholly occupied non-H atoms and the partial anion
sites with an occupancy Ն0.5 were refined anisotropically, whereas
H atoms were placed in calculated positions and refined by using
a riding model. There were 14 residual Fourier peaks of 1.0–
2.0 eÅ–3 in the final model, most of which also lie within the unas-
signed region of the disordered solvent.
Structure Refinement of [Fe(L2Br)2][BF4]2·3CH3NO2: Both BF4
–
Supporting Information (see footnote on the first page of this arti-
cle): Additional crystallographic figures, magnetic susceptibility
and LIESST relaxation data, and powder diffraction simulations.
ions are disordered over two sites. The refined occupancy ratio for
the disordered orientations of anion B(38)–F(42) was 0.5:0.5,
whereas that for anion B(43)–F(47) was 0.72:0.28. The refined re-
straints B–F = 1.38(2) and F···F = 2.25(2) Å were applied to the
anions. All wholly non-H atoms with an occupancy Ͼ0.5 were re-
fined anisotropically and all H atoms were placed in calculated
positions and refined by using a riding model.
Acknowledgments
The authors thank Dr. H. J. Blythe (University of Sheffield) for
some of the magnetic susceptibility data and A. Kazlauciunas
(University of Leeds) for the DSC measurement. The Aquitaine
Region is acknowledged for the development of the International
Centre of Photomagnetism in Aquitain (ICPA) platform. This
work was funded by the Engineering and Physical Science Research
Council.
Structure Refinement of [Fe(L2I)2][BF4]2·3CH3NO2: The two BF4
–
ions are disordered over two sites. The sites of disorder for anion
B(38)–F(42) were refined with a common wholly occupied B atom,
B(38), with a refined occupancy ratio of 0.8:0.2. The refined occu-
pancies of the disordered orientations of the other anion B(43)–
F(47) were 0.6:0.4. The refined restraints B–F 1.40(2) and F···F
2.29(2) Å were applied to the anions. One of the three nitromethane
molecules C(52)–O(55) is also disordered over two orientations
with a 0.60:0.40 occupancy ratio. This was modeled by using the
fixed restraints C–N 1.47(2), N–O 1.21(2), O···O 2.10(2), and C···O
2.30(2) Å. All wholly occupied non-H atoms were refined aniso-
tropically and all H atoms were placed in calculated positions and
refined by using a riding model. There are five residual Fourier
peaks of 1.0–1.3 eÅ–3 within the disordered anion B(38)–F(42),
which may represent another, minor site of disorder for that resi-
due. The deepest Fourier hole of –1.3 eÅ–3 is also located in the
same anion.
[1] P. Gutlich, H. A. Goodwin (Eds.), Spin-Crossover in Transition-
Metal Compounds I–III, in: Topics in Current Chemistry, vol.
233–235, Springer, Berlin, 2004.
[2] a) P. Gamez, J. S. Costa, M. Quesada, G. Aromí, Dalton Trans.
ˇ
2009, 7845–7853; b) I. Salitrosˇ, N. T. Madhu, R. Bocˇa, J. Pav-
lik, M. Ruben, Monatsh. Chem. 2009, 140, 695–733; c) A.
Bousseksou, G. Molnár, L. Salmon, W. Nicolazzi, Chem. Soc.
Rev. 2011, 40, 3313–3335; d) M. A. Halcrow, Chem. Soc. Rev.
2011, 40, 4119–4142.
[3] M. A. Halcrow, Coord. Chem. Rev. 2009, 253, 2493–2514.
[4] M. A. Halcrow, Coord. Chem. Rev. 2005, 249, 2880–2908.
[5] R. Pritchard, C. A. Kilner, M. A. Halcrow, Chem. Commun.
2007, 577–579.
[6] C. Carbonera, J. S. Costa, V. A. Money, J. Elhaïk, J. A. K.
Howard, M. A. Halcrow, J.-F. Létard, Dalton Trans. 2006,
3058–3066.
Structure Refinement of [Fe(L2CCPh)(OH2)4]x[Fe(L2CCPh)-
(OH2)3]1–x[BF4]2·(C2H5)2O: The asymmetric unit contains half a
complex dication, with Fe(1), N(2), N(5), and O(20B) lying on the
–
crystallographic C2 axis [0, y, ¾], and one BF4 anion and a half-
occupied, disordered diethyl ether molecule occupying general lat-
tice sites. In addition to the L2CCPh ligand and the wholly occu-
pied water ligand O(19), there are three weaker Fourier peaks close
to the C2 axis. These are at appropriate distances from Fe(1) to be
considered partial water ligands and they were modeled as such, as
O(20A) (occupancy 0.4), O(20B) (lying on the C2 axis, occupancy
0.1), and O(20C) (occupancy 0.2). O(20A) and its symmetry equiv-
alent O(20Ai) [symmetry code: (i) –x, y, 3/2 – z] are 2.76 Å from
each other, so these sites can be simultaneously occupied. The prox-
imity of O(20B) and O(20C) to O(20A) and to each other means
that only one of those sites can be occupied at any one time. We
interpret this as implying that the crystal contains a mixture of
seven-coordinate [Fe(L2CCPh)(OH2)4]2+ when O(20A) and
O(20Ai) are both occupied and six-coordinate [Fe(L2CCPh)-
(OH2)3]2+ when either O(20B), O(20C), or O(20Ci) is occupied.
[7] a) M. Nihei, L. Han, H. Oshio, J. Am. Chem. Soc. 2007, 129,
5312–5313; b) C. A. Tovee, C. A. Kilner, S. A. Barrett, J. A.
Thomas, M. A. Halcrow, Eur. J. Inorg. Chem. 2010, 1007–1012.
[8] a) C. Rajadurai, O. Fuhr, R. Kruk, M. Ghafari, H. Hahn, M.
Ruben, Chem. Commun. 2007, 2636–2638; b) J. Elhaïk, C. M.
Pask, C. A. Kilner, M. A. Halcrow, Tetrahedron 2007, 63, 291–
298.
[9] a) M. Nihei, N. Takahashi, H. Nishikawa, H. Oshio, Dalton
Trans. 2011, 40, 2154–2156; b) Y. Hasegawa, K. Takahashi, S.
Kume, H. Nishihara, Chem. Commun. 2011, 47, 6846–6848; c)
R. González-Prieto, B. Fleury, F. Schramm, G. Zoppellaro, R.
Chandrasekar, O. Fuhr, S. Lebedkin, M. Kappes, M. Ruben,
Dalton Trans. 2011, 40, 7564–7570.
[10] M. S. Alam, M. Stocker, K. Gieb, P. Müller, M. Haryono, K.
Student, A. Grohmann, Angew. Chem. 2010, 122, 1178; Angew.
Chem. Int. Ed. 2010, 49, 1159–1163.
[11] M. Loï, M. W. Hosseini, A. Jouaiti, A. De Cian, J. Fischer, Eur.
J. Inorg. Chem. 1999, 1981–1985.
[12] J. C. Rodriguez-Ubis, R. Sedano, G. Barroso, O. Juanes, E.
Brunet, Helv. Chim. Acta 1997, 80, 86–96; corrigendum: J. C.
Rodriguez-Ubis, R. Sedano, G. Barroso, O. Juanes, E. Brunet,
Helv. Chim. Acta 1997, 80, 621.
–
The BF4 ion is crystallographically ordered, but the diethyl ether
molecule is disordered over two sites labeled “A” (occupancy 0.3)
and “B” (0.2). The fixed restraints C–C 1.52(2), C–O 1.43(2), 1,3-
C···O 2.42(2), and 1,3-C···C = 2.34(2) Å were applied to this resi-
due. All wholly occupied non-H atoms and O(20A) were refined
Eur. J. Inorg. Chem. 2013, 819–831
830
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[13] See, for example: a) K. H. Sugiyarto, D. C. Craig, A. D. Rae,
H. A. Goodwin, Aust. J. Chem. 1993, 46, 1269–1290; b) M.
Ostermeier, M.-A. Berlin, R. M. Meudtner, S. Demeshko, F.
Meyer, C. Limberg, S. Hecht, Chem. Eur. J. 2010, 16, 10202–
10213; c) N. Chandrasekhar, R. Chandrasekar, Dalton Trans.
2010, 39, 9872–9878.
[14] J. Elhaïk, V. A. Money, S. A. Barrett, C. A. Kilner, I. R. Evans,
M. A. Halcrow, Dalton Trans. 2003, 2053–2060.
[15] V. A. Money, I. R. Evans, J. Elhaïk, M. A. Halcrow, J. A. K.
Howard, Acta Crystallogr., Sect. B 2004, 60, 41–45.
[16] V. A. Money, J. Elhaïk, I. R. Evans, M. A. Halcrow, J. A. K.
Howard, Dalton Trans. 2004, 65–69.
[17] R. Pritchard, H. Lazar, S. A. Barrett, C. A. Kilner, S. Asthana,
C. Carbonera, J.-F. Létard, M. A. Halcrow, Dalton Trans. 2009,
6656–6666.
[18] a) G. Zoppellaro, M. Baumgarten, Eur. J. Org. Chem. 2005,
2888–2892; corrigendum: G. Zoppellaro, M. Baumgarten, Eur.
J. Org. Chem. 2005, 4201; b) S. Basak, P. Hui, R. Chandrase-
kar, Synthesis 2009, 4042–4048.
[19] R. Pritchard, Ph. D. Thesis, University of Leeds, UK, 2008.
[20] a) C. A. Bessel, R. F. See, D. L. Jameson, M. R. Churchill, K. J.
Takeuchi, J. Chem. Soc., Dalton Trans. 1992, 3223–3228; b)
M. A. Halcrow, C. A. Kilner, M. Thornton-Pett, Acta Crys-
tallogr., Sect. C 2000, 56, 213–214; c) F. Calderazzo, U. Englert,
C. Hu, F. Marchetti, G. Pampaloni, V. Passarelli, A. Romano,
4432–4441; b) J.-F. Létard, L. Capes, G. Chastanet, N. Moliner,
S. Létard, J.-A. Real, O. Kahn, Chem. Phys. Lett. 1999, 313,
115–120; c) S. Marcén, L. Lecren, L. Capes, H. A. Goodwin,
J.-F. Létard, Chem. Phys. Lett. 2002, 358, 87–95.
[30] O. Kahn, Molecular Magnetism, VCH Verlagsgesellschaft,
Weinheim, Germany, 1993.
[31] a) A. Hauser, Coord. Chem. Rev. 1991, 111, 275–290; b) A.
Hauser, Comments Inorg. Chem. 1995, 17, 17–40; c) A. Hauser,
Top. Curr. Chem. 2004, 234, 155–198; d) A. Hauser, C. Enas-
chescu, M. L. Daku, A. Vargas, N. Amstutz, Coord. Chem. Rev.
2006, 250, 1642–1652.
[32] J.-F. Létard, G. Chastanet, O. Nguyen, S. Marcèn, M.
Marchivie, P. Guionneau, D. Chasseau, P. Gütlich, Monatsh.
Chem. 2003, 134, 165–182.
[33] C. J. O’Connor, Prog. Inorg. Chem. 1982, 29, 203–283.
[34] J. Elhaïk, C. A. Kilner, M. A. Halcrow, Dalton Trans. 2006,
823–830.
[35] See, for example: a) G. S. Matouzenko, A. Bousseksou, S. Lec-
ocq, P. J. van Koningsbruggen, M. Perrin, O. Kahn, A. Collet,
Inorg. Chem. 1997, 36, 5869–5879; b) H. Z. Lazar, T. Forestier,
S. A. Barrett, C. A. Kilner, J.-F. Létard, M. A. Halcrow, Dalton
Trans. 2007, 4276–4285; c) R. Pritchard, S. A. Barrett, C. A.
Kilner, M. A. Halcrow, Dalton Trans. 2008, 3159–3168; corri-
gendum: R. Pritchard, S. A. Barrett, C. A. Kilner, M. A.
Halcrow, Dalton Trans. 2009, 10622.
R. Santi, Inorg. Chim. Acta 2003, 344, 197–206; d) G. Zoppel- [36] a) J. P. Jesson, S. Trofimenko, D. R. Eaton, J. Am. Chem. Soc.
laro, A. Geies, V. Enkelmann, M. Baumgarten, Eur. J. Org.
Chem. 2004, 2367–2374; e) N. Chandrasekhar, R. Chandrase-
kar, Chem. Commun. 2010, 46, 2915–2917; f) S. Basak, R.
Chandrasekar, Adv. Funct. Mater. 2011, 21, 667–673.
[21] Twinned single crystals of L2NO2 were also obtained. A pre-
liminary refinement showed that the molecule adopts the same
conformation as L2Cl (see the Supporting Information). This
structure has not been deposited with the CCDC. Unit cell
data for L2NO2: C10H6N8O4, Mr 302.23, monoclinic, P21/c, a
= 5.3564(6), b = 10.1292(12), c = 23.201(3) Å, β = 96.463(6)°,
V = 1250.8(2) Å3, Z = 4.
[22] R. Lungwitz, S. Spange, New J. Chem. 2008, 32, 392–394.
[23] a) M. L. Scudder, H. A. Goodwin, I. G. Dance, New J. Chem.
1999, 23, 695–705; b) J. McMurtrie, I. Dance, CrystEngComm
2005, 7, 216–229; c) J. McMurtrie, I. Dance, CrystEngComm
2010, 12, 2700–2710.
1967, 89, 3158–3164; b) M. A. Hoselton, L. J. Wilson, R. S.
Drago, J. Am. Chem. Soc. 1975, 97, 1722–1729; c) M. F. Twee-
dle, L. J. Wilson, J. Am. Chem. Soc. 1976, 98, 4824–4834; d)
W. Linert, M. Konecny, F. Renz, J. Chem. Soc., Dalton Trans.
1994, 1523–1531.
[37] a) H. L. Chum, J. A. Vanin, M. I. D. Holanda, Inorg. Chem.
1982, 21, 1146–1152; b) T. Zhu, C. H. Su, D. Schaeper, B. K.
Lemke, L. J. Wilson, K. M. Kadish, Inorg. Chem. 1984, 23,
4345–4349; c) S. A. Barrett, C. A. Kilner, M. A. Halcrow, Dal-
ton Trans. 2011, 40, 12021–12024.
[38] a) D. F. Evans, J. Chem. Soc. 1959, 2003–2005; b) E. M. Schu-
bert, J. Chem. Educ. 1992, 69, 62.
[39] R. Pritchard, C. A. Kilner, S. A. Barrett, M. A. Halcrow, Inorg.
Chim. Acta 2009, 362, 4365–4371.
[40] H. Toftlund, J. J. McGarvey, Top. Curr. Chem. 2004, 233, 151–
166.
[24] a) J. K. McCusker, A. L. Rheingold, D. N. Hendrickson, Inorg.
Chem. 1996, 35, 2100–2112; b) P. Guionneau, M. Marchivie,
G. Bravic, J.-F. Létard, D. Chasseau, Top. Curr. Chem. 2004,
234, 97–128.
[41] J. Elguero, E. Gonzalez, R. Jacquier, Bull. Soc. Chim. Fr. 1968,
5009–5017.
[42] R. J. Smithson, C. A. Kilner, A. R. Brough, M. A. Halcrow,
Polyhedron 2003, 22, 725–733.
[25] J. M. Holland, J. A. McAllister, C. A. Kilner, M. Thornton-
Pett, A. J. Bridgeman, M. A. Halcrow, J. Chem. Soc., Dalton
Trans. 2002, 548–554.
[26] a) J. Elhaïk, D. J. Evans, C. A. Kilner, M. A. Halcrow, Dalton
Trans. 2005, 1693–1700; b) C. A. Kilner, M. A. Halcrow, Poly-
hedron 2006, 25, 235–240; c) M. Haryono, F. W. Heinemann,
K. Petukhov, K. Gieb, P. Müller, A. Grohmann, Eur. J. Inorg.
Chem. 2009, 2136–2143.
[43] V. A. Money, J. S. Costa, S. Marcén, G. Chastanet, J. Elhaïk,
M. A. Halcrow, J. A. K. Howard, J.-F. Létard, Chem. Phys.
Lett. 2004, 391, 273–277.
[44] V. Mishra, R. Mukherjee, J. Linares, C. Baldé, C. Desplanches,
J.-F. Létard, E. Collet, L. Toupet, M. Castro, F. Varret, Inorg.
Chem. 2008, 47, 7577–7587.
[45] M. Marchivie, P. Guionneau, J.-F. Létard, D. Chasseau, Acta
Crystallogr., Sect. B 2005, 61, 25–28.
[27] S. Decurtins, P. Gütlich, C. P. Köhler, H. Spiering, A. Hauser,
Chem. Phys. Lett. 1984, 105, 1–4.
[28] a) J.-F. Létard, P. Guionneau, O. Nguyen, J. S. Costa, S.
Marcén, G. Chastanet, M. Marchivie, L. Capes, Chem. Eur. J.
2005, 11, 4582–4589; b) J.-F. Létard, J. Mater. Chem. 2006,
16, 2550–2559; c) J.-F. Létard, G. Chastanet, P. Guionneau, C.
[46] V. A. Money, C. Carbonera, J. Elhaïk, M. A. Halcrow, J. A. K.
Howard, J.-F. Létard, Chem. Eur. J. 2007, 13, 5503–5514.
[47] a) J. C. Philip, H. B. Oakley, J. Chem. Soc. Trans. 1924, 125,
1189–1195; b) W. A. Felsing, S. A. Durban, J. Am. Chem. Soc.
1926, 48, 2885–2893.
[48] G. M. Sheldrick, Acta Crystallogr., Sect. A 2008, 64, 112–122.
Desplanches, Optimising the stability of trapped metastable spin [49] L. J. Barbour, J. Supramol. Chem. 2001, 1, 189–191.
states, in: Spin-crossover materials - properties and applications
(Ed.: M. A. Halcrow), Wiley, Chichester, UK, 2013, in the
press.
[50] POVRAY, v.3.5, Persistence of Vision Raytracer Pty. Ltd., Wil-
liamstown, Victoria, Australia, 2002 (http://www.povray.org).
[51] A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7–13.
Received: September 18, 2012
[29] a) J.-F. Létard, P. Guionneau, L. Rabardel, J. A. K. Howard,
A. E. Goeta, D. Chasseau, O. Kahn, Inorg. Chem. 1998, 37,
Published Online: December 4, 2013
Eur. J. Inorg. Chem. 2013, 819–831
831
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim