Synthesis, structure, and magnetic properties of a tris[3-(2-pyridyl)-1,2,4-triazole]iron(II) spin-crossover complex

Article abstract of DOI:10.1002/1099-0682(200010)2000:10<2231::AID-EJIC2231>3.0.CO;2-B

The synthesis and characterization of tris[3-(pyridin-2-yl)-1,2,4-triazole]iron(II) bis(tetrafluoroborate), obtained from the reaction of 3-(pyridin-2-y1)-1,2,4-triazole (Hpt) and hexa-aquairon(II) tetrafluoroborate, [Fe(H₂O)6](BF₄)₂, is described, together with its crystal structures at two temperatures. X-ray crystallographic parameters are as follows: [Fe(Hpt)₃](BF₄)₂·nH₂O (n ? 2) at 250 K: orthorhombic space group Pbam, a = 15.8068(18) A, b = 17.2800(14) A, c = 21.215(2) A, V = 5794.7(10) A³, and Z = 8. [Fe(Hpt)₃](BF₄)₂·nH₂O (n ? 2) at 95 K: orthorhombic space group Pbam, a = 15.7080(12) A, b = 17.1023(16) A, c = 21.006(2) A, V = 5643.1(9) A³, and Z = 8. The Fe(II) ions are (at both temperatures) octahedrally surrounded in a mer-configuration. The complex has been subjected to both ⁵⁷Fe Mossbauer spectroscopy and magnetic susceptibility measurements. The ⁵⁷Fe Mossbauer spectrum shows the presence of two different iron(II) sites. The high-spin fraction vs. T curve of the crossover, obtained by SQUID measurements, is gradual and without hysteresis, with T 1/2 = 135 K.

Full text of DOI:10.1002/1099-0682(200010)2000:10<2231::AID-EJIC2231>3.0.CO;2-B

FULL PAPER
Synthesis, Structure, and Magnetic Properties of a Tris[3-(2-pyridyl)-1,2,4-
triazole]iron(II) Spin-Crossover Complex
[
a]
[a]
[b]
[b]
Arno F. Stassen, Matthijs de Vos, Petra J. van Koningsbruggen, Franz Renz,
[
b]
[c]
[d][°]
[a]
Jürgen Ensling, Huub Kooijman, Anthony L. Spek,
Jaap G. Haasnoot,*
[b]
[a]
Philipp Gütlich, and Jan Reedijk
th
Dedicated to Professor Heinrich Vahrenkamp on the occasion of his 60 birthday
Keywords: Spin crossover / Iron / Magnetic properties / Moessbauer spectroscopy / N ligands
˚ ˚
group Pbam, a = 15.7080(12) A, b = 17.1023(16) A, c =
The synthesis and characterization of tris[3-(pyridin-2-yl)-
,2,4-triazole]iron(II) bis(tetrafluoroborate), obtained from
the reaction of 3-(pyridin-2-yl)-1,2,4-triazole (Hpt) and hexa- both temperatures) octahedrally surrounded in a mer-config-
˚ ˚
3
II
1
21.006(2) A, V = 5643.1(9) A , and Z = 8. The Fe ions are (at
5
7
aquairon(II) tetrafluoroborate, [Fe(H
together with its crystal structures at two temperatures. X-
ray crystallographic parameters are as
follows: urements. The 57Fe Mössbauer spectrum shows the presence
Fe(Hpt) ](BF ·nH O (n ഠ 2) at 250 K: orthorhombic space
group Pbam, a = 15.8068(18) A, b = 17.2800(14) A, c =
2
O)
6
](BF
4
)
2
, is described,
uration. The complex has been subjected to both
Fe
Mössbauer spectroscopy and magnetic susceptibility meas-
[
3
4
)
2
2
of two different iron(II) sites. The high-spin fraction vs. T
curve of the crossover, obtained by SQUID measurements, is
gradual and without hysteresis, with T1/2 = 135 K.
˚
˚
˚
˚
3
2
1.215(2) A,
V
=
5794.7(10) A , and
Z
=
8.
[
Fe(Hpt) ](BF
3
4
)
2
·nH
2
O (n ഠ 2) at 95 K: orthorhombic space
Introduction
57Fe
Fe(Hpt) ](Anion) (Solvent) [Hpt ϭ 3-(pyridin-2-yl)-1,2,4-
Mössbauer
spectroscopy
studies
on
[
3
2
x
Ϫ
Ϫ
Ϫ
Ϫ
triazole (see Figure 1); Anion ϭ Cl , ClO4 , PF6 , BF4
;
Nowadays there is increasing interest in new bistable
iron(II) spin-crossover compounds.[ A variety of iron(II)
1,2]
Solvent ϭ ethanol, water] and [Fe(H3mpt) ](Anion) (Sol-
3 2
vent)x [H3Mpt ϭ 3-methyl-5-(pyridin-2-yl)-1,2,4-triazole;
compounds are known to show a transition from the high-
Ϫ
Ϫ
5
Anion ϭ ClO4 , PF6 ; Solvent ϭ water] have been re-
spin state [HS, S ϭ 2, T (O )] to the low-spin state [LS,
2
g
h
ported by Stupik et al.[
9,10]
and by Sugiyarto et al.
[31]
Cle-
1
[2Ϫ4]
S ϭ 0, A1g (O )] on cooling.
h
II
arly, these mononuclear compounds exhibit continuous Fe
spin-crossover behaviour. The Fe ion resides in a six-nitro-
Iron(II) compounds incorporating substituted 1,2,4-tria-
zole ligands have been found to undergo spin transition. In
several cases, hysteresis effects and rather high transition
II
gen environment made up of three bidentate chelating 3-
(pyridin-2-yl)-1,2,4-triazole ligands coordinating through
N4 of the 1,2,4-triazole and the pyridine N. The asymmetric
nature of the bidentate ligand may lead to the formation of
temperatures have been found.[
5Ϫ23]
Generally speaking,
II
when dealing with linear polynuclear Fe compounds of
1
,2,4-triazole ligands, the spin-crossover behaviour is ab-
[
5,7,11Ϫ16,19,21,22,24,25]
FeL units having facial or meridional geometry. Both Stu-
rupt and may show hysteresis,
but this
3
pik and Sugiyarto observed[
9,10,31]
that the spin transition
is not the case for shorter oligomers.[
26Ϫ28]
For mononu-
was affected by the presence of different solvent molecules.
This contributes to the complicated spin-crossover behavi-
our evident from the 57Fe Mössbauer spectra. They also
clear 1,2,4-triazole-containing iron(II) spin-crossover sys-
tems, the spin crossover has always been found to be grad-
ual and without any hysteresis.[
9,10,18,29,30]
observed two different iron(II) high-spin sites in the hy-
Ϫ
[
a]
drated BF
4
complex, but did not mention the ratio be-
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden
University,
[31]
tween the two sites. In addition, Sugiyarto et al. reported
comparable observations for the hydrated BF4 and ClO4
P.O. Box 9502, NL-2300 RA Leiden, The Netherlands
Fax: (internat.) ϩ31Ϫ715274671
Ϫ
Ϫ
iron(II) tris[3-(pyridin-2-yl)-1,2,4-triazole] compounds.
E-mail: haasnoot@chem.leidenuniv.nl
Institut für Anorganische Chemie und Analytische Chemie,
Universität Mainz,
[
[
[
b]
c]
Staudingerweg 9, D-55099 Mainz, Germany
Bijvoet Center for Biomolecular Research, Utrecht University,
Padualaan 8, NL-3584 CH Utrecht, The Netherlands
E-mail: a.l.spek@chem.uu.nl
Address correspondence pertaining to crystallographic studies
to this author.
°]
Figure 1. Structure of 3-(pyridin-2-yl)-1,2,4-triazole (Hpt)
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237

WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1434Ϫ1948/00/1010Ϫ2231 $ 17.50ϩ.50/0
2231

J. G. Haasnoot et al.
·nH O at 95 K and
FULL PAPER
However, these results have not yet been satisfactorily ex- Table 2. Important angles in [Fe(Hpt)
plained, and hence a further study on this interesting class
of compounds was undertaken.
](BF
)
4 2
3
2
2
50 K
Bond angle
Angle at 95 K
Angle at 250 K
In this paper, the synthesis, crystal structures at 250 K
and 95 K, and physical properties of the iron(II) Hpt coor- N(103)ϪFeϪN(303)
171.19(13)
171.04(12)
169.33(12)
80.68(12)
80.17(13)
80.33(13)
96.25(13)
92.05(12)
93.20(12)
95.08(12)
93.24(12)
91.31(12)
90.77(13)
94.41(12)
94.57(12)
164.82(14)
166.71(12)
163.95(17)
77.13(13)
77.00(15)
75.53(15)
100.75(16)
92.43(13)
93.62(15)
96.67(15)
92.74(13)
93.45(14)
91.53(15)
99.04(13)
95.42(14)
Ϫ
N(111)ϪFeϪN(211)
N(203)ϪFeϪN(311)
N(103)ϪFeϪN(111)
N(203)ϪFeϪN(211)
N(303)ϪFeϪN(311)
N(103)ϪFeϪN(203)
N(103)ϪFeϪN(211)
N(103)ϪFeϪN(311)
N(111)ϪFeϪN(203)
N(111)ϪFeϪN(303)
dination complex of the anion BF4 are described in detail.
Results and Discussion
Structure of the Complex
X-ray crystal structure analyses have been carried out at
two temperatures, i.e. at 250 K, where the complex is almost
completely high spin, and at 95 K, which is below the trans- N(111)ϪFeϪN(311)
N(203)ϪFeϪN(303)
ition temperature and hence most of the iron(II) ions parti-
N(211)ϪFeϪN(303)
cipating in the spin-crossover behaviour are in the low-spin N(211)ϪFeϪN(311)
state. The molecular structure showing the atomic num-
bering scheme is depicted in Figure 2, while relevant bond
length and bond angle information is given in Table 1 and
Table 2. At both temperatures, the space group is Pbam
Because of rotational and positional disorder of the tetra-
fluoroborate and water molecules, it was not possible to
determine any positional parameters for these molecules.
This disorder suggests the absence of hydrogen bonding.
This is surprising, because both N1 and N2 of the triazole
ring can form hydrogen bonds, either in amino or imino
with Z ϭ 8. The asymmetric unit contains one Fe(Hpt)3
moiety. Structural analysis of the complex shows that at
both temperatures the iron(II) is octahedrally surrounded
by three Hpt ligands, which bind through N2 of the pyrid-
ine ring and N4 of the triazole ring in a mer configuration.
[32]
form.
Comparison of the bond angles of the NϪFeϪN axes at
the two temperatures shows that at low temperature the
1
angles are closer to 180°, as would be expected for a A1g
octahedral structure. The NϪFeϪN angles involving the ni-
trogen atoms on different axes are also closer to those of a
regular octahedron at 95 K than at 250 K. The ligand bite
angles increase from 75.53Ϫ77.13° at 250 K to
8
0.17Ϫ80.86° at 95 K. The other nonlinear NϪFeϪN
angles decrease from 91.53Ϫ100.75° (average 95.1°) at
50 K to 90.77Ϫ96.25° (average 93.4°) at 95 K. Considering
2
that at this temperature 35% of the complexes are still in
the high-spin state, the actual angles in the low-spin com-
plex will be even closer to 90°.
The FeϪN bond length contracts by approximately
˚
0
.15 A on cooling to 95 K. These observations can be ra-
tionalized as follows. For iron(II) complexes surrounded by
neutral ligands, Equation (1) holds, where µ is the dipole
moment of the ligand and r the average metalϪligand dis-
tance.[
1]
Figure 2. Displacement ellipsoid plot drawn at a 50% probability
level of the complex cation [Fe(Hpt)
3
](BF
4
)
2
·nH
2
O at 95 K
From this equation, Equation (2) can be derived.
Table 1. Relevant bond lengths in [Fe(Hpt)
and 250 K
3
](BF
4
)
2
·nH
2
O at 95 K
˚
˚
Bond
Length at 95 K [A]
Length at 250 K [A]
FeϪN(103)
FeϪN(203)
FeϪN(303)
FeϪN(111)
FeϪN(211)
FeϪN(311)
2.003(3)
2.011(3)
2.009(3)
2.051(3)
2.041(3)
2.047(3)
2.127(4)
2.160(4)
2.154(4)
2.202(3)
2.199(4)
2.229(4)
The value for an iron(II) spin-crossover complex is norm-
ally ca. 1.74. The value obtained by applying Equation (2)
in the case of [Fe(Hpt) ](BF ) · n H O is lower than 1.74,
specifically 1.57. The difference between the calculated and
measured values arises from the fact that at both temper-
3
4 2
2
2232
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237

Properties of a Tris[3-(2-pyridyl)-1,2,4-triazole]iron(II) Spin-Crossover Complex
FULL PAPER
atures not all the iron(II) ions are in the same spin state.
From magnetic and Mössbauer measurements, it has been
concluded that at 95 K a residue of 36.1% high-spin and at
2
50 K a residue of 2.5% of low-spin iron(II) ions is present.
The actual FeϪN bond length for a pure low-spin state will
˚
be closer to 2.00 A, and using this value in Equation 2
would generate the expected dipole moment.
Because the contribution to the structure factors associ-
ated with the counter ions and solvent was taken into ac-
count using the SQUEEZE procedure as incorporated in
[
33]
PLATON,
no exact molecular formula could be deter-
mined from the crystal structure. However, considering the
number of electrons associated with the counterions and
solvent in conjunction with the elemental analysis, the mo-
lecular formula could be calculated. Assuming that only
2
ϩ
Ϫ
cations of the formula [Fe(Hpt)3] , anions BF4 , and
water were present, the molecular formula
Fe(Hpt) ](BF ) ·nH O, n ϭ 2, was derived.
[
3
4 2
2
The stability of the complex was tested by heating it for
several hours at 50 °C and by storing it in vacuo for one
day. Under neither conditions were the elemental analysis
or magnetic susceptibility affected. Thus, despite the fact
that the water molecules are seemingly not involved in
strong hydrogen bonding, they are not readily removed
from the complex.
Figure 3. 57Fe Mössbauer spectral measurements at 205, 150, 80,
and 4.2 K
57Fe Mössbauer Spectroscopic Measurements
57Fe Mössbauer spectra were recorded at temperatures
of 300 K, 205 K, 150 K, 80 K, 32 K, 10 K, and 4.2 K. A
representative selection of the spectra is shown in Figure 3.
Least-squares-fitted parameters are listed in Table 3. In all
the spectra, three doublets have been assigned. Of these,
crossover is almost complete. Further cooling to 4.2 K re-
sults only in an additional 2% increase in the intensity of
the low-spin doublet to 69%. The distribution of high-spin
and low-spin species found at the measured temperatures is
in agreement with the magnetic measurements.
II
one is characteristic of Fe in the low-spin state, while the
II
[34]
When a spin-crossover compound is cooled very rapidly,
i.e. quenched, it is possible that high-spin ions are unable
other two are characteristic of Fe in the high-spin state.
We shall refer to the high-spin doublet with the larger quad-
[
9,10]
to relax to the low-spin state, but remain thermally spin
rupole splitting as HSa, and to the other as HSb. Stupik
trapped in the metastable high-spin state.[
37Ϫ39]
This phe-
[
31]
and Sugiyarto
also detected a second high-spin site, but
nomenon is called thermal spin trapping. Mössbauer spec-
tra measured after rapid cooling, i.e. by taking a sample at
300 K and placing it in a cryostat at 4.2 K, revealed that a
fraction of the iron(II) ions, approximately 6%, is trapped
in the high-spin state (see Table 3). It is mainly the HSb site
that exhibits the spin state trapping.
The presence of a second high-spin site (HSb) indicates a
different local surrounding from the majority (HSa), arising
from the distribution of the solvent and anions.
it is not clear as to whether the site they detected was ident-
ical to HSb (these authors did not analyse the spectra in
more detail). Schreiner[29] also detected a second high-spin
site in [Fe(Hpt) ](BF ) ·H O/EtOH, to the extent of 7% at
3
4 2
2
1
00 K (31% at 300 K), which was reported to behave sim-
ilarly as that found in the present work. The area fraction
of the HSb site is only 6% at 205 K and decreases to 3% at
4
.2 K, which is close to the experimental limit. Its corres-
II
ponding Fe low-spin state doublet probably coincides with
the aforementioned major low-spin doublet. The molar ra-
tios of the molecules in the spin states can be derived from
the area fractions by taking equal MössbauerϪLamb fac-
tors for the low-spin and high-spin states.[
Magnetic Susceptibility Measurements
35,36]
Figure 4 shows the product of χT and
T for
At 300 K, a pure high-spin state doublet is present (unre- [Fe(Hpt) ](BF ) · 2 H O, where χT is the molar magnetic
3
4 2
2
solved HSa and HSb). At 205 K, a low-spin doublet of susceptibility, recorded in the temperature range 4Ϫ325 K
weak intensity is visible, associated with 8% of the ions. as a function of temperature T. The curve provides evidence
This low-spin doublet clearly increases in intensity on going for S ϭ 2 (HS) o S ϭ 0 (LS) spin-crossover behaviour. At
from 205 to 80 K (from 8% to 67%), while the doublet asso- room temperature, all iron(II) ions are in the high-spin state
3
Ϫ1
ciated with the HSa site gradually diminishes from 86% at (χT ϭ 3.52 cm Kmol ). Below 75 K, the value of χT falls
3
Ϫ1
2
05 K to 30% at 80 K. At the latter temperature, the spin to 0.72 cm Kmol . This indicates the presence of a signi-
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237
2233

J. G. Haasnoot et al.
FULL PAPER
3 4 2 2
Table 3. Least-squares fitted Mössbauer parameters for [Fe(Hpt) ](BF ) · 2 H O
[
a]
T
(
δ(LS)
∆E
Q
(LS)
Γ(LS)
A
LS
δ(HS)
∆E
Q
(HS) Γ(HS)
A
HS δ(HS)
∆E
Q
(HS) Γ(HS)
Ϫ1
A
HS
(%)
K) (mm sϪ1) (mm s
Ϫ1
)
(mm sϪ1) (%)
(mm sϪ1) (mm s
Ϫ1
)
(mm s
Ϫ1
)
(%) (mm sϪ1) (mm s
Ϫ1
)
(mm s
)
2
1
8
3
1
4
05 0.34(1)
50 0.37(1)
0.32(1)
0.33(1)
0.33(1)
0.33(1)
0.33(1)
0.33(1)
0.15(1)
0.15(1)
0.16(1)
0.16(1)
0.16(1)
0.16(1)
8
0.95(0)
0.98(1)
1.02(1)
1.03(1)
1.03(1)
1.03(1)
2.14(1)
2.36(1)
2.57(1)
2.54(1)
2.45(1)
2.47(1)
0.25(1)
0.25(1)
0.23(1)
0.22(1)
0.20(1)
0.20(1)
86
68
30
28
28
28
1.10(1)
1.10(1)
1.13(1)
1.13(1)
1.13(1)
1.13(1)
1.35(1)
1.35(1)
1.70(1)
1.70(1)
1.70(1)
1.75(1)
0.19(1)
0.19(1)
0.20(1)
0.20(1)
0.20(1)
0.20(1)
6
5
3
3
3
3
27
67
69
69
69
0
0.39(1)
0.40(1)
0.40(1)
0.40(1)
2
0
.2
After quenching from 300 K (approx. 100 K sϪ1)
4
.2
0.40(1)
0.33(1)
0.16(1)
63
1.04(1)
2.45(1)
0.20(1)
29
1.13(1)
1.80(1)
0.20(1)
7
[
a]
δ ϭ isomer shift, ∆E
Q
ϭ quadrupole splitting, Γ ϭ line width, AHS ϭ area fraction of the HS doublets. Statistical standard deviations
are given in parentheses.
ficant residue of high-spin iron(II) ions, as is also evident
from the low-temperature Mössbauer spectrum. The spin
transition is very gradual, with T1/2 ϭ 135 K. The para-
meter T1/2 is defined as the temperature in a thermally
driven spin equilibrium at which, for the iron(II) ions in-
volved in spin transition, the amounts of high-spin and low-
spin molecules are equal. The entire measurement required
5
h 45 min, corresponding to a heating rate of approxim-
Ϫ1
ately 1.1 Kmin in steps of 2.5 K.
Figure 5. Fitted high-spin molar fraction, based on Mössbauer
3 4 2 2
data, vs. T for [Fe(Hpt) ](BF ) · 2 H O (* ϭ actual Mössbauer
measurements)
In Equation (3), R is the molar gas constant, ∆H is the
enthalpy change, and ∆S is the entropy change associated
with the spin crossover. Least-squares fitting leads to ∆H ϭ
Ϫ1
Ϫ1
Ϫ1
6
.4 kJ mol , ∆S ϭ 47 J K mol , and a mean-field inter-
Ϫ1
action parameter Γ smaller than 0.1 kJ mol . The spin
crossover may be considered as being almost non-
cooperative.[
3 4 2 2
Figure 4. Plot of χT vs. T for [Fe(Hpt) ](BF ) · 2 H O
1,41]
Complexes of hexaaquairon(II) tetrafluoroborate with 3-
pyridin-2-yl)-1,2,4-triazole have been prepared in aqueous
The spin transition takes place without hysteresis. Fur-
thermore, the spin-crossover behaviour is easily detected,
since the compound is highly thermochromic. On cooling
to 77 K, its colour changes from yellow to deep purple.
Figure 5 shows the γHS vs. T curve plotted over the tem-
(
solutions at pH 7 to 14. The resulting complexes all exhibit
identical spin-crossover behaviour and also give identical
elemental analysis results. This excludes the presence of the
Ϫ
deprotonated ligand (pt ).
perature range 4Ϫ320 K. The high-spin molar fraction γ
HS
5
7
was deduced from the Fe Mössbauer data area fractions
by taking equal MössbauerϪLamb factors for low-spin
state and high-spin state doublets.[
35,36]
It may be assumed
Ligand-Field Spectroscopy
from the Mössbauer measurements that below 80 K 30% of
the high-spin sites are not involved in the spin-crossover
process. This makes it possible to simulate the spin-cross-
over curve using Equation (3), which is derived from the
regular solution model.[
The ligand-field spectrum of the present complex was
measured in the diffuse reflection mode at room temper-
5
5
ature and at liquid-nitrogen temperature. The T Ǟ E
2
Ϫ1
1
transition is seen at 850 nm (11.765 cm ), while the A Ǟ
T1 transition appears at 530 nm (18.868 cm ). These
1
40]
1
Ϫ1
values are in the regions normally observed for the corres-
ponding dϪd transitions in spin-crossover behaviour mat-
erials.[
1]
2234
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237

Properties of a Tris[3-(2-pyridyl)-1,2,4-triazole]iron(II) Spin-Crossover Complex
FULL PAPER
ordinated to the iron(II) through N2 of the triazole ring
instead of N4. This would lead to a difference in the coor-
Conclusion
In this paper, the spin-crossover complex tris(Hpt)FeII dination sphere and hence in the Mössbauer spectrum; un-
with tetrafluoroborate as anion has been reported. This fortunately, the difference that such a coordination would
complex is formed as the meridional isomer, just as the pre- cause in the structural analysis is smaller than the experi-
[
9,10]
viously
studied [Fe(Hpt) ](ClO ) ·(Solvent) . Bidentate mental error. A further possibility is that the deprotonated
3 4 2 x
Ϫ
coordination by the Hpt ligand takes place through N4 of ligand (pt ) is present. In this case, 6% of the iron(II) ions
Ϫ
the 1,2,4-triazole and the pyridine N. Details of the X-ray would be surrounded by a deprotonated pt ligand and
structure determination concerning the location of the have only one tetrafluoroborate counterion in their coor-
tetrafluoroborate and water cannot be reported owing to dination spheres. However, crystallization experiments at
high positional and rotational disorder of the counterion high pH (from 7 to 14) produced materials giving identical
and solvent. The transition temperature T1/2 has been deter- results in elemental analysis and susceptibility measure-
mined as 135 K.
ments. Therefore, this possibility can be excluded. A third
During the spin transition, the geometry of the complex possibility is that 6% of the crystals have the facial orienta-
II
does not change significantly. The average Fe ϪN bond tion. This different coordination could lead to a difference
˚
length decreases by approximately 0.15 A. The average in both the ligand field and quadrupole splitting. In the
NϪFeϪN bond angles become closer to those of an ideal mer-isomer, the three ligands are magnetically non-equiva-
octahedron on going from the high-spin to the low-spin lent, whereas in the fac-isomer, the three ligands are mag-
state.
netically equivalent as this isomer possesses a C -axis of
3
In the Mössbauer spectrum, two high-spin iron(II) sites symmetry.[45]
with different populations are observed, with the HSa site
Both the first and third options are feasible. Triazole is
clearly showing spin-crossover behaviour. Although a suit- known to be able to coordinate through N2 as well as
ably intense doublet associated with the low-spin state is through N4,[
7,21,26]
while both fac- and mer-isomers of (pyd-
observed, it is not possible to determine whether only one ridin-2-yl)-1,2,4-triazole complexes are known.[
31,45]
Both
low-spin site is present or whether the signals of two differ- the linkage and geometrical isomerism will cause a change
ent sites are overlapping. The weak intensity HSb site is at in ligand field. Whether this variation is sufficiently large to
the experimental limit of detection, hence no further con- be observed by Mössbauer spectroscopy remains uncertain.
clusions can be made concerning the spin-crossover behavi-
our at this site. On rapid cooling, a small fraction of the
high-spin ions can be thermally trapped in a metastable Experimental Section
high-spin state.
It is difficult to compare these results with previously General Remarks: Ligand-field spectra of the solid were obtained
published observations on the [Fe(Hpt)](BF ) ·2H O com- on a PerkinϪElmer Lambda 900 spectrophotometer in the
4
2
2
[
9,10,31]
2000Ϫ200 nm range using the diffuse reflectance technique, with
MgO as a reference.
plex.
These authors also noticed two types of high-
spin site, but the data were not analysed further.
The disorder of the solvent molecules and counterions
Magnetic susceptibilities were measured in the temperature range
found in the crystal structure gives an indication that nei- 5Ϫ325 K on a Quantum Design MPMS-5S SQUID operating at
ther are involved in hydrogen bonding to the triazole rings 0.1 Tesla. Data were corrected for magnetization of the sample
of the Hpt ligands. Moreover, the distance between the ni- holder and diamagnetic contributions, which were estimated using
Ϫ4
3
Ϫ1 [46]
Pascal constants (χ
D
ϭ Ϫ2.85·10 cm mol ).
trogens of the triazole ring and the electron density of the
counterions and water molecules is too great for hydrogen
bonding. The water molecules are nevertheless strongly
57Fe
Mössbauer spectra of the polycrystalline compound
](BF ·2H O were recorded using a conventional spec-
trometer with the sample placed in a helium cryostat allowing vari-
[Fe(Hpt)
3
4
)
2
2
bound to the complex and cannot be easily removed.
II
Approximately 30% of the Fe ions do not participate in ation of the temperature between 4 K and 300 K. As the source,
5
7
the thermal spin transition. It is possible that the distribu-
Co in rhodium at 293 K was used. The data were evaluated using
tion of water molecules in the crystal lattice is responsible the programs MOSFUN[ and EFFI.
35]
[36]
The high-spin molar
fraction was estimated from the area fractions by taking equal
MössbauerϪLamb factors for low-spin and high-spin states.
for
this.
In
the
almost
identical
system
[
42]
[
Fe(Hpt) ](BF ) (SO ) ·3H O, the spin transition does
3 4 1.2 4 0.4 2
not have an associated high-spin residue at low temperature.
Preparation of the Compounds: All chemicals were obtained com-
It is known that in a few systems the solvent molecules are mercially in adequately pure states. The ligand Hpt was synthesized
essential for the spin transition.[
43,44]
It seems likely that in from 2-cyanopyridine, hydrazine monohydrate, and formic acid ac-
the present system, due to a non-uniform distribution of cording to the method described in literature.[
water molecules in the lattice, some Fe ions remain in the
high-spin state.
32,47]
7 6 4
Hpt: C H N
II
(
146.2): calcd. C 57.53, H 4.14, N 38.33; found C 56.68, H 4.65,
N 38.63.
A small proportion (6%) of the iron(II) ions (HSb) ex-
perience a different coordination than that described in this
[
Fe(Hpt)
3
](BF
4
)
2
·2H
O)
2
O was prepared under argon atmosphere by
](BF (0.34 g, 1 mmol) and Hpt (0.44 g,
dissolving [Fe(H
2
6
4 2
)
article. There are several possibilities for the different coor- 3 mmol) in hot, deoxygenated water (20 mL), to yield a clear brown
dination of HSb. One explanation is that the ligand is co- solution. On standing in a closed flask at room temperature for
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237
2235

J. G. Haasnoot et al.
FULL PAPER
Table 4. Crystallographic data for [Fe(Hpt)
Compound
3
4 2 2
](BF ) ·nH O at 95 K and 250 K
T ϭ 95 K
T ϭ 250 K
Crystal data
[
a]
Formula
C
21
703.94
H
18FeN12·2BF
4
·2H
2
O
21 4 2
C H18FeN12·2BF ·2H O
[
a]
Molecular weight
Crystal system
Space group
703.94
Orthorhombic
Pbam (no. 55)
15.7080(12)
17.1023(16)
21.006(2)
5643.1(9)
1.6572(2)*
Orthorhombic
Pbam (no. 55)
15.8068(18)
17.2800(14)
21.215(2)
5794.7(10)
1.6138(2)*
1.68
˚
a [A]
˚
b [A]
˚
c [A]
˚
3
V [A ]
Ϫ3 [a]
D
D
calcd [g cm
]
Ϫ3
obsd [g cm
]
Z (mononuclear units)
8
8
[
a]
F(000)
2848
0.66
2848
Ϫ1
][a]
µ [mm [Mo-K
α
0.64
Crystal colour
red-brown
yellow-brown
Data collection
θ
min, θmax [deg.]
1.6, 25.25
4
1.6, 25.25
2
X-ray exposure [h]
Data set
Ϫ18:18, Ϫ20:20, Ϫ25:25
Ϫ18:14, Ϫ20:16, Ϫ25:17
Total data
44707
5273
0.0907
0.0395
22091
5337
0.0541
0.0552
Total unique data
R
R
int
σ
Refinement
No. of refined params.
307
307
[
b]
Final R1
0.0681 [4433 I Ͼ 2σ(I)]
0.0740 [3766 I Ͼ 2σ(I)]
[
c]
Final wR2
0.1587
1.140
0.2039
1.041
Goodness of Fit
Ϫ1[d]
2
2
2
2
2
2
w
σ (F ) ϩ (0.462P) ϩ 11.59P
Ͻ0.001, Ͻ0.001
σ (F ) ϩ (0.00755P) ϩ 13.51P
Ͻ0.001, Ͻ0.001
(
∆/σ)av, (∆/σ)max
Min. and max.
Ϫ0.36, 0.36
Ϫ0.35, 0.33
residual density [e AϪ3]
˚
[
[
a]
d]
[b]
|. Ϫ [c] wR2 ϭ {Σ[w(F
2
2)2]/Σ[w(F
2)2]}1/2. Ϫ
Including disordered anion and solvent contributions. Ϫ R1 ϭ Σ | |F
o
| Ϫ |F
c
| |/Σ |F
o
o
Ϫ F
c
o
2
2
o c
P ϭ [Max(F ,0) ϩ 2F ]/3.
a few days, clear yellow block-shaped crystals were deposited. Ϫ
Fe(Hpt) ](BF · 2 H O (703.9): calcd. C 35.83, H 3.15, N 23.88, also containing disordered solvent molecules, probably water. The
F 21.59; found C 35.93, H 3.32, N 23.9, F 19.6.
counterions were found to be disordered and located in a region
[
3
4
)
2
2
contribution to the structure factors associated with counterions
and solvent was taken into account using the SQUEEZE procedure
Crystal Structure Determination of [Fe(Hpt) ](BF
3
4
)
2
: A single crys-
[33]
as incorporated in PLATON.
A total of 840 e were found in a
tal of dimensions 0.10 ϫ 0.15 ϫ 0.30 mm was mounted on a Linde-
mann glass capillary and placed in the cold nitrogen stream of the
diffractometer. Data were collected on an EnrafϪNonius
3
˚
volume of 1800 A , equally distributed between two symmetry-re-
lated cavities. Neutral atom scattering factors and anomalous dis-
persion corrections were taken from the International Tables for
KappaCCD diffractometer on a rotating anode (Mo-K
α
radiation,
[50]
Crystallography. Geometrical calculations and the generation of
˚
graphite monochromator, λ ϭ 0.71073 A, Lp correction, no ab-
sorption correction). Two data sets were recorded from this crystal,
one at T ϭ 95 K and one at T ϭ 250 K. Structure determination
revealed that the only geometrical differences between these struc-
tures were those related to the spin transition. Since the data for
the 250 K structure were rather poor due to crystal deterioration,
an additional data set was collected on a second crystal of dimen-
sions 0.20 ϫ 0.35 ϫ 0.35 mm. Geometric data pertaining to the
two structures measured at 250 K proved to be equal within experi-
mental error. The data for the 250 K structure reported here refer
to the second data set. Pertinent data relating to the structure deter-
minations are collected in Table 4. The structures were solved by
[33]
illustrations were performed with PLATON.
Crystallographic data (excluding structure factors) for the struc-
tures reported in this paper have been deposited with the Cam-
bridge Crystallographic Data Centre as supplementary publication
nos. CCDC-140484/140485. Copies of the data can be obtained
free of charge on application to the CCDC, 12 Union Road, Cam-
bridge CB2 1EZ, U.K. [Fax: (internat.) ϩ44 (0)1223/336033; E-
mail: deposit@ccdc.cam.ac.uk].
Acknowledgments
[
48]
2
direct methods using SHELXS-86.
Refinement on F was per-
Financial support by the European Union, allowing regular ex-
change of preliminary results with several European colleagues
within the TOSS network under contract ERB-FMRX-CT980199,
is gratefully acknowledged. Additional support by the ESF pro-
gram Molecular Magnets is also kindly acknowledged. The work
described in this paper has been supported by the Leiden Univer-
sity Study Group, WFMO, and has been performed under the aus-
formed with SHELXL-97.[ The hydrogen atoms were included
in the refinement in calculated positions riding on their carrier
atoms. The non-hydrogen atoms were refined with anisotropic ther-
mal parameters. The hydrogen atoms were refined with a fixed iso-
tropic displacement parameter related to the value of the equivalent
49]
isotropic displacement parameter of their carrier atoms. The BF
4
2236
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237

Properties of a Tris[3-(2-pyridyl)-1,2,4-triazole]iron(II) Spin-Crossover Complex
FULL PAPER
[
[
22]
23]
C. Cantin, H. Daubric, J. Kliava, Y. Servant, L. Sommier, O.
Kahn, J. Phys. Condens. Matt. 1998, 31, 7057Ϫ7064.
C. Janiak, T. G. Scharmann, J. C. Green, R. P. G. Parkin, M.
J. Kolm, E. Riedel, W. Mickler, J. Elguero, R. M. Claramunt,
D. Sanz, Chem. Eur. J. 1996, 8, 992Ϫ1000.
pices of the Graduate Research School, HRSMC, a joint activity
of Leiden University and the two Universities in Amsterdam. This
work was also partly financially supported (A.L.S.) by the Council
for the Chemical Sciences of the Netherlands Organization for Sci-
entific Research (CW-NWO). Finally, financial support from the
Materialwissenschaftliches Forschungszentrum der Universität
Mainz and the Fonds der Chemischen Industrie is also gratefully
acknowledged.
[
[
24]
25]
K. H. Sugiyarto, D. C. Graig, A. D. Rae, H. A. Goodwin, Aust.
J. Chem. 1994, 869Ϫ890.
E. Codjovi, L. Sommier, O. Kahn, C. Jay, New J. Chem. 1996,
5
, 503Ϫ505.
[26]
J. J. A. Kolnaar, G. van Dijk, H. Kooijman, A. L. Spek, V. G.
Ksenofontov, P. Gütlich, J. G. Haasnoot, J. Reedijk, Inorg.
Chem. 1997, 11, 2433Ϫ2440.
[
[
1]
2]
P. Gütlich, A. Hauser, H. Spiering, Angew. Chem. Int. Ed. Engl.
994, 2024Ϫ2054.
E. König, G. Ritter, S. K. Kulshreshtha, Chem. Rev. 1985,
19Ϫ234.
[
[
27]
28]
J. J. A. Kolnaar, Ph.D. Thesis, Leiden University, 1998.
C. Janiak, T. G. Scharmann, T. Brauniger, J. Holubova, M.
Nadvornik, Z. Anorg. Allg. Chem. 1998, 5, 769Ϫ774.
A. Schreiner, Ph.D. thesis, University of Mainz 1995.
P. J. Kunkeler, P. J. van Koningsbruggen, J. P. Cornelissen, A.
N. van der Horst, A. M. van der Kraan, A. L. Spek, J. G.
Haasnoot, J. Reedijk, J. Am. Chem. Soc. 1996, 9, 2190Ϫ2197.
K. H. Sugiyarto, D. C. Craig, A. D. Rae, H. A. Goodwin, Aust.
J. Chem. 1995, 35Ϫ54.
R. Hage, Ph.D. Thesis, Leiden University, 1991.
A. L. Spek, Acta Crystallogr. 1990, C34.
P. Gütlich, R. Link, A. Trautwein, Mössbauer Spectroscopy and
Transition Metal Chemistry, Springer, Berlin, 1978.
E. W. Müller, Ph.D. Thesis, University of Mainz, Germany,
1982.
1
2
[
[
29]
30]
[
[
3]
4]
O. Kahn, C. J. Martinez, Science 1998, 5347, 44Ϫ48.
P. Gütlich, in Structure and Bonding (Eds.: M. J. Clarke, P.
Hemmerich, J. B. Goodenough, J. A. Ibers, C. K. Jørgensen, J.
B. Neilands, D. Reiden, R. Weiss, R. J. P. Williams), Springer,
Berlin, 1981, p. 83Ϫ196.
[
31]
[
[
[
[
[
5]
6]
7]
8]
9]
W. Vreugdenhil, J. H. van Diemen, R. A. G. de Graaff, J. G.
[
[
[
32]
33]
34]
Haasnoot, J. Reedijk, Polyhedron 1990, 24, 2971Ϫ2979.
J. G. Vos, R. A. G. de Graaff, J. G. Haasnoot, A. M. van der
Kraan, P. de Vaal, J. Reedijk, Inorg. Chem. 1994, 2905Ϫ2910.
W. Vreugdenhil, J. G. Haasnoot, O. Kahn, P. Thu e´ ry, J. Reed-
ijk, J. Am. Chem. Soc. 1987, 5272Ϫ5273.
G. Vos, R. A. le F eˆ bre, R. A. G. de Graaff, J. G. Haasnoot, J.
Reedijk, J. Am. Chem. Soc. 1983, 1682Ϫ1693.
[
[
[
35]
36]
37]
H. Spiering, EFFI Program for Mössbauer Spectra Fitting, Uni-
versity of Mainz, Germany, 1999.
H. A. Goodwin, K. H. Sugiyarto, Chem. Phys. Lett. 1987, 5,
P. Stupik, J. H. Zhang, M. Kwiecien, W. M. Reiff, J. G. Haas-
noot, R. Hage, J. Reedijk, Hyperfine Interactions 1986,
4
70Ϫ474.
725Ϫ727.
[
[
38]
39]
[
[
[
10]
11]
12]
A. Hauser, Chem. Phys. Lett. 1986, 6, 543Ϫ548.
P. Stupik, W. M. Reiff, R. Hage, J. Jacobs, J. G. Haasnoot, J.
Reedijk, Hyperfine Interactions 1988, 343Ϫ346.
A. Ozarowski, Y. Shunzhong, B. R. McGarvey, A. Mislankar,
J. E. Drake, Inorg. Chem. 1991, 3167Ϫ3174.
W. Vreugdenhil, S. Gorter, J. G. Haasnoot, J. Reedijk, Polyhed-
ron 1985, 10, 1769Ϫ1775.
G. Ritter, E. Konig, W. Irler, H. A. Goodwin, Inorg. Chem.
1
978, 2, 225.
[
[
[
40]
41]
42]
C. P. Schlichter, H. G. Drickamer, J. Chem. Phys. 1972, 2142.
P. Gütlich, A. Hauser, H. Spiering, Angew. Chem. 1994, 2109.
O. Roubeau, A. F. Stassen, M. d. Vos, J. G. Haasnoot, J. Reed-
ijk, to be published.
R. Claude, J. Zarembowitch, M. Philoche-Levisalles, F. d’Y-
voire, New J. Chem. 1991, 625Ϫ641.
R. Hernandez-Molina, A. Mederos, S. Dominguez, P. Gili, C.
Ruiz-Perez, A. Castineiras, X. Solans, F. Lloret, J. A. Real, In-
org. Chem. 1998, 20, 5102Ϫ5108.
R. Hage, J. G. Haasnoot, J. Reedijk, J. G. Vos, Inorg. Chim.
Acta 1986, 73Ϫ76.
I. M. Kolthoff, P. J. Elving, Treatise on Analytical Chemistry,
Vol. 4, New York, 1963.
M. Uda, Y. Hisazumi, K. Sato, S. Kubota, Chem Pharm. Bull.
[
[
13]
14]
O. Kahn, J. Krober, C. Jay, Adv. Mater. 1992, 11, 718Ϫ728.
C. Jay, F. Groli e` re, O. Kahn, J. Kröber, Mol. Cryst. Liq. Cryst.
[
43]
1993, 255Ϫ262.
[44]
[
15]
J. Kröber, J. Audi e` re, R. Claude, E. Codjovi, O. Kahn, J. G.
Haasnoot, F. Groli e` re, C. Jay, A. Bousseksou, J. Linar e` s, F.
Varret, A. Gonthier-Vassal, Chem. Mater. 1994, 1404Ϫ1412.
J. Kröber, E. Codjovi, O. Kahn, F. Groli e` re, C. Jay, J. Am.
Chem. Soc. 1993, 9810Ϫ9811.
[45]
[46]
[47]
[48]
[49]
[50]
[
[
[
[
16]
17]
18]
19]
K. H. Sugiyarto, H. A. Goodwin, Aust. J. Chem. 1994,
263Ϫ277.
K. H. Sugiyarto, D. C. Craig, A. D. Rae, H. A. Goodwin, Aust.
J. Chem. 1993, 1269Ϫ1290.
1
976, 12, 3103Ϫ3108.
G. M. Sheldrick, SHELXS-86: Program for crystal structure
determination, University of Göttingen, Germany, 1986.
G. M. Sheldrick, SHELXL-97Ϫ2: Program for crystal structure
refinement, University of Göttingen, Germany, 1996.
A. J. C. Wilson, International Tables for Crystallography, Vol-
ume C, Kluwer Academic Publishers, Dordrecht, The
Netherlands, 1992.
P. J. van Koningsbruggen, Y. Garcia, E. Codjovi, R. Lapouy-
ade, O. Kahn, L. Fourn e` s, L. Rabardel, J. Mater. Chem. 1997,
10, 2069Ϫ2075.
[
[
20]
21]
Y. Garcia, P. J. van Koningsbruggen, G. Bravic, P. Guionneau,
A. Chasseau, G. L. Cascarano, J. Moscovici, K. Lambert, A.
Michalowicz, O. Kahn, Inorg. Chem. 1997, 6357Ϫ6365.
Y. Garcia, P. J. van Koningsbruggen, E. Codjovi, R. Lapouy-
Received February 18, 2000
ade, O. Kahn, L. Rabardel, J. Mater. Chem. 1997, 6, 857Ϫ858.
[I00060]
Eur. J. Inorg. Chem. 2000, 2231Ϫ2237
2237