Iron(II) spin transition complexes with dendritic ligands, Part I

Article abstract of DOI:10.1002/ejic.200700822

The ligands G1- and G2-oligo (benzyl ether) (PBE) dendrons and their iron(II) complexes [Fe(Gn-PBE)₃]A₂·xH₂O (with n = 1, 2 and A = triflate, tosylate) were prepared. The magnetic properties of the complexes were investigated by a SQUID magnetometer. All complexes exhibit gradual spin transition below room temperature. At very low temperatures the magnetic behaviour reflects zero-field splitting (ZFS) effects. ⁵⁷Fe-Moessbauer spectroscopy was performed to distinguish between ZFS of high spin species and spin state conversion into the low spin state. Further characterisation was carried out by thermogravimetric analysis (TGA) and FT-IR spectroscopy. Structural features have been determined by powder XRD measurements. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejic.200700822

FULL PAPER
DOI: 10.1002/ejic.200700822
Iron(II) Spin Transition Complexes with Dendritic Ligands, Part I
Prashant Sonar,[a] C. Matthias Grunert,[b] Yong-Li Wei,[b] Joachim Kusz,[b,c]
Philipp Gütlich,*[b] and A. Dieter Schlüter*[a]
Dedicated to Prof. Dieter Fenske on the occasion of his 65th birthday
Keywords: FeII spin crossover / Dendron complexes / Mössbauer spectroscopy / Magnetic susceptibility / XRD / DSC
measurements
The ligands G1- and G2-oligo (benzyl ether) (PBE) dendrons
and their iron(II) complexes [Fe(Gn-PBE)3]A2·xH2O (with n =
1, 2 and A = triflate, tosylate) were prepared. The magnetic
properties of the complexes were investigated by a SQUID
magnetometer. All complexes exhibit gradual spin transition
below room temperature. At very low temperatures the mag-
netic behaviour reflects zero-field splitting (ZFS) effects.
57Fe-Mössbauer spectroscopy was performed to distinguish
between ZFS of high spin species and spin state conversion
into the low spin state. Further characterisation was carried
out by thermogravimetric analysis (TGA) and FT-IR spec-
troscopy. Structural features have been determined by pow-
der XRD measurements.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
compound.[5–7] SCO compounds exhibiting spin state
changes in the room temperature region have attracted par-
ticular interest in recent years because of their possible use
in sensors and optical devices. Suitability of such material
for practical applications[8] depends strongly on its chemical
stability. This aspect is particularly important in the case of
iron(II) complexes because of their tendency to oxidise
more or less easily in ambient atmosphere. In recent years
researchers have been seeking for new synthetic strategies
for the preparation of SCO compounds with special empha-
sis on steering the cooperative interactions towards room
temperature switching operation on the one hand and
reaching sufficient chemical stability for long term practical
use in ambient atmosphere on the other hand.[9–11] Special
attention has thereby been paid to the role of dimensional-
ity of the lattice network and the way of interconnecting
building blocks by π-interaction and/or hydrogen bonding.
Promising approaches include also embedding a spin transi-
tion complex into polymer films, e.g. Nafion films[12] by
which the counterions for the cationic SCO complex are
supplied, or attaching the complexes directly to a polymeric
chain via ligand interaction as reported by Maeda et al.[13]
Our attention has recently been drawn to dendritic sys-
tems which could eventually be capable of complex forma-
tion with iron(II) ions via nitrogen donor atoms in 4-posi-
tion of 1,2,4-triazole molecules as “zero-generation basis”.
Many SCO compounds of iron(II) based on 1,2,4-triazole
ligands have been reported,[1] and it has turned out that
these are among the most stable ones known so far, which
in fact have already been turned into prototypes of de-
Introduction
Transition metal complexes with spin crossover (SCO)
behaviour have been studied since the last seven decades
and the field has been widely reviewed.[1] Particularly exten-
sive work has been published about iron(II) SCO complexes
with thermal high spin (HS, 5T2) ↔ low spin (LS, 1A1) tran-
sition describing the SCO behaviour as influenced by chem-
ical alteration (ligand substitution, nature of anion and
crystal solvent, metal dilution) and physical perturbation
(light, pressure, magnetic field).[2] One of the main objec-
tives has been to learn as much as possible about the
mechanism of SCO and the cooperative interactions[1–4]
that are responsible for the appearance of different types of
spin-transition curves γHS(T), the plot of the molar fraction
of the high-spin molecules as a function of temperature,
which varies from gradual to abrupt and exhibits sometimes
hysteresis and incomplete transitions.[2] Another recently
pursued objective is the attempt to combine SCO with an-
other dynamical physical phenomenon such as nonlinear
optical or liquid crystalline properties in one and the same
[a] Laboratory of Polymer Chemistry, Department of Materials,
Swiss Federal Institute of Technology, ETH Zürich,
HCI J541, 8093 Zürich, Switzerland
E-mail: dieter.schlueter@mat.ethz.ch
[b] Institut für Anorganische Chemie und Analytische Chemie,
Johannes Gutenberg Universität Mainz,
Staudingerweg 9, 55099 Mainz, Germany
Fax: 49-6131-3922990
E-mail: guetlich@uni-mainz.de
[c] University of Silesia, Institute of Physics, ul.
Uniwersytecka 4, 40007 Katowice, Poland
Eur. J. Inorg. Chem. 2008, 1613–1622
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1613

P. Sonar, C. M. Grunert, Y. Wei, J. Kusz, P. Gütlich, A. D. Schlüter
FULL PAPER
vices.[8] Hoping that iron(II) SCO compounds with triazole- using the G2 dendron 4[19] and obtained in a yield of 61%
based dendritic ligands would combine new spin transition (Scheme 2) after treating the ionic intermediate (not shown)
properties suitable for practical use with satisfactory chemi- with sodium hydroxide in a dichloromethane/water mixture
cal stability we have started the work presented here. Den- (4:1).
dritic materials are regularly branched compounds with
high molecular weight. They provide a defined number of
end groups, whose modification allows to “fine tune” their
properties. Among them dendronized polymers are one
class of materials in which dendrons are connected to every
repeating unit of the linear backbone of a polymeric
chain.[14] They have been individualized on solid substrates
and moved about with the scanning force microscopy tip.[15]
If one were to use dendronized triazoles as ligand in spin-
transition polymers, a material could be obtained which in
an ideal way combines the properties of dendronized poly-
mers (high solubility, good processability) with the physical
properties of transition metal complexes. A first step in this
direction has been accomplished by Fujigaya[16] who pre-
sented dendronized iron(II) complexes of the kind iron(II)-
tris[Gn-poly(benzyl ether)dendron] methyl sulfonic acid
(Gn, n = 0, 1, 2). These dendronized coordination polymers
turned out to be magnetically active, however, the conver-
water (4:1), NaOH, room temp., 5 h; 73%.
Scheme 1. Reagents and conditions of (i) 1H-1,2,4-triazole-1-pro-
panenitrile, acetonitrile, reflux, 12 h, 80% and (ii) dichloromethane/
sion of low to high spin states was irreversible due to an
irreversible loss of water during the first heating process
leading to the spin change. Stimulated by this experiment
we set out in a long-term project to try to find reversible
spin-transition-dendronized polymers which, when indivi-
dualized on a substrate would change colour depending
upon their state of bending. This work reports as a step
towards these goals the synthesis of first- (G1) and second-
generation (G2) Fréchet/Hawker-type dendronized triazole
complexes of the kind [Fe(G1,2-PBE)]A2·xH2O where G1-
PBE denotes 4-[3,5-bis(benzyloxy)benzyl]-4H-1,2,4-triazole
and G2-PBE represents 4-{3,5-bis[3,5-bis(benzyloxy)ben-
zyloxy]benzyl}-4H-1,2,4-triazole as ligands. Toluenesulfo-
nate (tos) and trifluorosulfonate (trif) were used as counter-
ions (A).
Results and Discussion
Synthesis of the Ligands
Alkylation of 1,2,4-triazoles leads to mixtures of 1- and
4-substituted products[17] which can be circumvented by the
elegant Horvàth method.[18] Here parent azoles are first
protected at N-1 by reacting them with, for instance, acrylo-
nitrile to give the corresponding Michael adducts. Then,
they can be selectively alkylated at the desired N-4 to fur-
nish a salt as an intermediate which is subsequently depro-
tected at N-1 by base-induced elimination. Application of
Scheme 2. Reagents and conditions of (i) 1H-1,2,4-triazole-1-pro-
panenitrile, acetonitrile, reflux, 12 h, 80% and (ii) dichloromethane/
water (4:1), NaOH, room temp., 5 h; 61%.
Synthesis of the Complexes
The complexation of iron(II) triflate or tosylate salt with
this method to triazole 3 (Scheme 1) started from the the ligands 3 or 5 in methanol and precipitation by addition
Fréchet first-generation (G1) dendron 1 whose benzylic of tetrahydrofurane (THF) yielded the following complexes:
bromide was attached to 1H-1,2,4-triazole-1-propanenitrile [Fe(G1-PBE)3.7](trif)2·1H2O,
[Fe(G1-PBE)3](tos)2·4H2O,
in refluxing acetonitrile to give intermediate 2 from which [Fe(G2-PBE)3.4](trif)2·4.3H2O and [Fe(G2-PBE)2.7](tos)2·
the G1 dendritic triazole 3 was obtained by basic treatment 2.9H2O. More details concerning the synthesis and the
in an overall yield of 73%. The purity of 3 was confirmed characterisation of the complexes (elemental analysis, neu-
1
by H NMR spectroscopy and elemental analysis. The sec- tron activation analysis, thermogravimetric analysis and in-
ond-generation (G2) triazole 5 was prepared similarly to 3 frared spectroscopy) are given in the Exp. Section.
1614
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1613–1622

Iron(II) Spin Transition Complexes with Dendritic Ligands, Part I
Differential Scanning Calorimetry
about 240 °C an additional endothermic signal is detected.
The loss of water, as observed by thermo-gravimetric In the cooling procedure the sample gives an exothermic
analysis (see Exp. Section) can also be found in measure- response at this temperature. Further cooling leads to a
ments with differential scanning calorimetry (DSC) during phase transition at about 38 °C which is similar to the pre-
the first heating process. Figure 1 (a) presents the data ob- vious sample. This phase transition is reversible and ap-
tained for [Fe(G1-PBE)3.7] (triflate)2·H2O. The graph shows pears during the following heating process at about 48 °C.
for the first heating process of the sample a complex over-
For the complexes of the second generation, G2-PBE,
lapping of endothermic signals between about 40 and 95 °C the first heating process is also irreversible as can be seen
with peak maxima at 50 °C and 80 °C, which points to a in Figure 1 (c). The DSC data show a strong endothermic
combined process of releasing water, structural rearrange- signal between about 30 °C and 100 °C with contribution
ments and eventually a partial spin state change. This will of several processes and a very pronounced peak at 48 °C
be superimposed by structural changes of the sample which in the case of [Fe(G2-PBE)3.4](triflate)2·4.3H2O and 49 °C
was prepared at low temperature and, therefore, will not in the case of [Fe(G2-PBE)2.7](tosylate)2·2.9H2O. These
have been in its energy minimum structure. Caused by the strong signals below 50 °C might be already the effect of
overlapping and not distinguishable effects only a qualita- evaporating water. Therefore, it is possible that also com-
tive analysis will be discussed. Repeated cooling and heat- plexes of G2 generation exhibit a release of water in two
ing leads to a reversible phase transition at approximately steps. The latter compound shows an additional endother-
46 °C in the cooling and 54 °C in the heating mode. Fig- mic peak at 179 °C (Figure 1, d). In this case it seems that
ure 1 (b) depicts DSC data obtained for [Fe(G1-PBE)3](tos)2· the second signal compensates for the first exothermic sig-
4H2O. The temperature behaviour is similar to the previous nal and might also be an effect of structural rearrange-
complex. In the first heating period a phase transition at ments. After the first heating procedure, the picture is sim-
about 45 °C over a range of 30 °C and the endothermic ilar to the behaviour of the tempered complexes with G1-
peak of the evaporating solvent at 100 °C can be found. At PBE ligands. The triflate complex has a phase transition at
Figure 1. Thermodynamic behaviour of a) [Fe(G1-PBE)3.7](triflate)2·1H2O, b) [Fe(G1-PBE)3](tos)2·4H2O, c) [Fe(G2-PBE)3.4](triflate)2·
4.3H2O, and d) [Fe(G2-PBE)2.7](tos)2·2.9H2O.
Eur. J. Inorg. Chem. 2008, 1613–1622
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1615

P. Sonar, C. M. Grunert, Y. Wei, J. Kusz, P. Gütlich, A. D. Schlüter
FULL PAPER
about 45 °C and the complex with tosylate counterion
Figure 3 shows the experiments performed with a fresh
shows a hysteresis with 49 °C in the cooling and 54 °C in sample of the G2 compound, [Fe(G2-PBE)2.7](tos)2·
the heating mode.
2.9H2O. The pattern between 2° and 20° shows at 300 K
only one broad reflection at 2Θ = 3.813° and a broad halo
at about 20° originating from the voluminous dendritic
branches. During the heating process to 350 K the complex
starts to crystallize and shows a strong reflection at 2Θ =
3.724° and some small reflections at 6.453°, 7,452° and
9.864°. This crystallization process is irreversible and re-
mains as such after cooling to 300 K, but the reflections
changed their position. The first strong reflection is at posi-
tion 2Θ = 3.734° and some small reflections appear at posi-
tions 6.470°, 7,473° and 9.891°. These reflections could be
indexed as (1, 0), (1, 1), (2, 0) and (2, 1) belonging to a two-
dimensional hexagonal lattice with lattice parameters a =
3.71 and 23.64 Å for 350 and 300 K, respectively. Compar-
able results were obtained by Fujigaya[16] and Seredyuk.[7]
Powder X-ray Diffraction Experiments
The result of the X-ray powder experiment with a fresh
sample of [Fe(G1-PBE)3](tos)2·4H2O measured at 300 K
shows a pronounced powder diffraction pattern in the mea-
sured range of angles 2Θ between 2° and 20° (Figure 2).
From this the unit cell can be determined as triclinic with
a = 20.445 Å, b = 16.719 Å, c = 12.164 Å and α = 85.94°,
β = 107.91° and γ = 96.031°. The volume of a unit cell
containing Z = 2 formula units is 3931.1 Å3 leading to a
density of ρ = 1.34 g/cm3. This differs from the results ob-
tained by Fujigaya[16] and Seredyuk[7] for similar iron(II)
dendritic and liquid crystalline systems who found hexago-
nal structures. The other possible interpretation of the fresh
[Fe(G1-PBE)3](tos)2·4H2O sample phase is to assume two
different two-dimensional hexagonal phases, the first with
lattice parameter a = 19.30 Å and the second with a =
22.73 Å. The experimental results do not allow to distin-
guish between both cases.
Figure 3. X-ray powder diffraction pattern of a fresh sample of
[Fe(G2- PBE)2.7](tos)2·2.9H2O measured at 300 K (dark grey), at
350 K (grey) and at 300 K after heat treatment (light grey).
Magnetic Measurements
The spin transition behaviour as a function of tempera-
ture has also been followed by magnetic susceptibility mea-
surements using a SQUID magnetometer with the following
temperature program: 4 or 200 or 250 K (for different runs)
Figure 2. X-ray powder diffraction pattern of a fresh sample of
[Fe(G1-PBE)3](tos)2·4H2O measured at 300 K (dark grey), at 350 K
(grey) and at 300 K after heat treatment (light grey).
However, when the sample is heated to 350 K the struc- Ǟ 350 Ǟ 4 Ǟ 350 K.
ture changes due to the loss of water (Figure 2). One strong
In the case of [Fe(G1-PBE)3](trif)2·H2O in (Figure 4, a)
reflection appears at 2Θ = 4.807°, a broad halo is visible at the value of χMT increases slightly from 2.9 to
about 20°, originating from the voluminous dendritic 3.1 cm3 Kmol–1 upon heating from 200 to 300 K. Between
branches, and some small reflections are at 8.334°, 9.626° 300 and 350 K, where the sample looses water, the
and 12.746°. After cooling to 300 K the positions of these χMT value increases to ca. 3.6 cm3 Kmol–1. Values between
reflections changed slightly. The first reflection appears at 3.0 and 3.5 cm3 Kmol–1 are typical of iron(II) complexes in
2Θ = 4.821° and some small reflections are at 8.356°, 9.651° the HS state. The calculation of the χMT values is based on
and 12.779°. The reflections can be indexed as (1, 0), (1, 1), the molecular mass, which was calculated using the iron
(2, 0) and (2, 1) belonging to a two-dimensional hexagonal content determined by neutron activation analysis (see Exp.
lattice with lattice parameters equal to 18.37 Å and 18.31 Å Section). With respect to the experimental error, the χMT
for 350 and 300 K, respectively. Now the structure of the value of 3.6 cm3 Kmol–1 falls into the expected range for
tempered complexes is consistent with the complexes exam- iron(II) HS species. On cooling the sample the χMT value
ined by Fujigaya[16] and Seredyuk.[7] The phase transition remains constant until 200 K, which shows that the change
upon heating is not reversible and the structure remains of χMT observed during the first heating process is not fully
stable, when the sample is cooled again to 300 K. These reversible (Figure 4, a). This is most likely due to loss of
experiments are presented in Figure 2.
water as discussed above. Between 200 and 50 K the sample
1616
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1613–1622

Iron(II) Spin Transition Complexes with Dendritic Ligands, Part I
Figure 4. Temperature-dependency behaviour followed by magnetic measurements for a) [Fe(G1-PBE)3.7](triflate)2·1H2O, b) [Fe(G1-PBE)3]-
(tos)2·4H2O, c) [Fe(G2-PBE)3.4](triflate)2·4.3H2O, and d) [Fe(G2-PBE)2.7](tos)2·2.9H2O.
undergoes a gradual spin transition and reaches a χMT [Fe(G2-PBE)3.4](triflate)2·4.3H2O, and [Fe(G2-PBE)2.7](tos)2·
value of 1.8 cm3 Kmol–1. The rather abrupt decrease of 2.9H2O are ca. 3.0 and ca. 3.5 cm3·Kmol–1, respectively.
χMT below 50 K is induced by zero-field splitting. The sec-
The change of spin state during the first heating pro-
ond heating procedure of the compound results in the same cedure of the complexes with G1-ligands as displayed in
data as measured in the cooling procedure of the respective Figure 4 (a, b) resembles the abrupt spin state change dur-
sample. The gradual spin transition below 200 K is revers- ing the release of water observed by Fujigaya[16] et al. for
ible.
their dendritic triazole complexes with different ligands, but
During the first heating procedure of [Fe(G1-PBE)3](tos)2· the effect in their studies is far more pronounced than in
4H2O (see Figure 4, b) the abrupt increase of the χMT value ours. The reason may be that Fujigaya et al. used amide
between 4 and 50 K originates from zero-field splitting. The groups to bridge the coordinating triazole with the den-
value of χMT between 50 and 200 K remains constant at dritic ligands. These amide groups are able to bind water
ca. 2.5 cm3 Kmol–1 and increases by further heating to molecules via hydrogen bonding in the close neighbour-
350 K to 3.5 cm3 Kmol–1, confirming that the compounds hood of the spin state changing iron centre. This very likely
under study are entirely in the HS state. Similar to the pre- affects the spin transition behaviour. In the present study
vious sample, the spin change is not reversible and the sam- the ligands do not contain such hydrophilic amide groups.
ple remains at ca. 3.5 cm3 Kmol–1 while cooling to 200 K. Therefore, it is assumed that the water molecules are not
Below this temperature, the sample undergoes a partial and localized in the same fashion, but rather distributed over
very gradual spin transition by ca. 0.3 cm3 Kmol–1. Below the lattice in positions with less influence on the spin transi-
50 K the magnetic curve is dominated again by the effect of tion behaviour. By the same token, it seems that in the pres-
zero-field splitting. Similar to the observation on the triflate ent study the structural rearrangements due to loss of water
complex (Figure 4, a), the magnetic behaviour in the second have less influence on the spin transition behaviour than in
heating procedure follows the behaviour observed in the Fujigaya’s study.
cooling process.
Complexes with dendritic ligands of the second genera-
tion (see parts c and d in Figure 4) do not show pronounced
irreversible effects in their magnetic behaviour. The mag-
57Fe Mössbauer Spectroscopy
netic curve below 50 K is dominated by zero-field splitting,
Mössbauer spectroscopy was performed for the tosylate
between 50 and 200 K a gradual and only partial spin tran- complexes at 80 and 4.2 K using fresh samples before heat-
sition can be observed followed by a plateau above this ing above 300 K (Figure 5). Mössbauer effect measure-
temperature. The χMT values at room temperature for ments at room temperature were not possible because of
Eur. J. Inorg. Chem. 2008, 1613–1622
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1617

P. Sonar, C. M. Grunert, Y. Wei, J. Kusz, P. Gütlich, A. D. Schlüter
FULL PAPER
Figure 5. 57Fe-Mössbauer spectra of [Fe(G1-PBE)3](tos)2·4H2O (a and b) as well as [Fe(G2-PBE)3](tos)2·2.9H2O (c and d).
the softness of the material. Responsible for this softness is slightly lower than expected from the magnetic measure-
are the flexible branches in combination with solvent mole- ments after heating to 350 K. The small deviation (ca. 4%)
cules. The high flexibility leads to an abrupt decrease of the is possibly due to the slightly higher Debye–Waller factor
Lamb–Mössbauer factor and prevents the acquisition of for the LS species compared to the HS species.[20,21] The
data of sufficient quality in the room temperature range.
57Fe-Mössbauer spectrum obtained at 4.2 K exhibits sim-
The spectrum of [Fe(G1-PBE)3](tos)2·4H2O recorded at ilar parameter values as determined at 80 K. The value of
80 K (Figure 5, a) can be separated into two doublets of γHS = 0.63 is only slightly lower than at 80 K, which is in
Lorentzian lines. One doublet with an isomer shift δ = agreement with the observed gradual spin transition and
1.19 mm/s and a quadrupole splitting ∆ = 3.45 mm/s refers proves that the strong decrease of χMT below 50 K is the
to iron(II) in the HS state. The large value of the quadru- result of zero-field splitting. All Mössbauer parameter val-
pole splitting is mainly caused by noncubic valence-electron ues are summarized in Table 1. The Mössbauer parameter
distribution to the electric field gradient and indicates that values measured for the second-generation complex [Fe(G2-
the lattice contribution to the electric field gradient is rela- PBE)3](tos)2·2.9H2O are very similar. The quadrupole split-
tively small, i.e. the inner core FeN6 apparently is not signif- ting of the HS doublet is somewhat smaller (∆ = 3.27 mm/s
icantly distorted. Obviously, the flexible and voluminous at 80 K) than observed for the G1 complex. The half width,
dentritic ligands serve as a soft medium embedding the ac- Γ, of the Lorentzian lines of the HS species of [Fe(G2-PBE)3]-
tual chromophores in relatively unperturbed fashion. A sec- (tos)2·2.9H2O is 0.46 mm/s and thus significantly higher
ond doublet with δ = 0.54 mm/s and ∆ = 0.19 mm/s can than the LS signal (about 0.3 mm/s) and observed for the
be attributed to the diamagnetic FeII LS species. The area G1 complex. The smaller quadrupole splitting as compared
fraction of the HS doublet being approximately equal to to that of the G1 system indicates that the lattice contri-
the molar fraction of the HS molecules is γHS = 0.67, which bution to the electric field gradient is somewhat larger in
Table 1. Parameter of the measured 75Fe-Mössbauer spectra.
Low-Spin State
High-Spin State
LS
HS
Complexes
T [K]
δ [mm/s]
∆EQ [mm/s] Γ [mm/s]
δ [mm/s] ∆EQ [mm/s] Γ [mm/s]
% HS area
[Fe(G1-PBE)3](tos)2·4H2O
80
4.2
80
0.54(1)
0.55(2)
0.53(1)
0.55(1)
0.19(1)
0.21(2)
0.24(2)
0.24(1)
0.28(2)
0.30(4)
0.30(3)
0.36(2)
1.19(1)
1.20(1)
1.20(1)
1.20(1)
3.45(1)
3.47(1)
3.27(3)
3.29(1)
0.30(1)
0.28(2)
0.46(2)
0.46(1)
67 (1)
63 (3)
68 (3)
70 (2)
[Fe(G2-PBE)3](tos)2·2.9H2O
4.2
1618
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1613–1622

Iron(II) Spin Transition Complexes with Dendritic Ligands, Part I
ture. Elemental analysis (C,H,N and S Analyzer) and mass spec-
troscopy (Ionspec Ultrima spectrometer) techniques were utilized
for the confirmation and purity of the ligands. The purity of the
complexes was proven by elemental analysis (C, H, N, and S) using
a Vario EL of Elementar. The iron content was determined by neu-
tron activation analysis. Samples of the dendronised complexes and
iron(II) sulfate as reference were irradiated for 6 h with thermal
neutrons in the nuclear research reactor TRIGA Mark (II) (Mainz)
at a flux of 7ϫ1011 ncm–2 s–1 affording the nuclear reaction
58Fe(n,γ)59Fe. After a delay of about two weeks for allowing short-
lived radioactive nuclides to decay, the content of 59Fe with half-
life of 44.5 d could be determined by measuring the count rate of
γ-radiation (1099 keV and 1293 keV), which is caused by the β-
decay to 59Co. The iron content was given by the ratio of the count
rates of the sample and the reference. More details are given in a
laboratory report by Grunert et al.[22] The complexation of the li-
gands Gn-PBE with their iron(II) salts could also be proven with
FT-IR spectroscopy. The experiments were carried out with a
Bruker Tensor 27 using samples prepared as KBr pellets. Thermo-
gravimetric analyses were performed by using a Q500 thermogra-
vimetry analyzer (TA Instruments, New Castle, Delaware). The
measurements were carried out in nitrogen atmosphere with a heat-
ing rate of 10 K/min and confirm the existence of included solvent.
Thermodynamic data were obtained by differential scanning calo-
rimetry (DSC) on DSC7 instrument (Perkin–Elmer, Norwalk,
Connecticut) with heating and cooling rates of 10 K/min. For
structural investigations powder X-ray diffracton were recorded at
300 K, 350 K and then again at 300 K with Cu-Kα radiation using
PANalytical X’Pert PRO diffractometer equipped with the Paar
HTK 1200. For the determination of the spin state magnetic mea-
surements were performed using a MPMSXL Quantum Design
SQUID magnetometer between 5 K and 350 K at heating and cool-
ing rates of 2 K/min. The measured data were corrected for the
magnetisation of the sample holder as well as for their own diamag-
netic contribution. Furthermore, 57Fe Mössbauer spectroscopy was
carried out at 80 K using a constant-acceleration conventional
spectrometer with a nitrogen cryostat. The source used was 57Co
in a Rh-matrix with an activity of about 10 mCi kept at room tem-
perature. For measurements at 4.2 K the samples were immersed
in helium gas in a helium cryostat. In this case the used source was
57Co in a Rh-matrix with an activity of about 5 mCi kept at 4 K.
It was impossible to record Mössbauer spectra at room temperature
because of the softness of the material causing too low resonance
effects in meaningful measuring times. The isomer shift values are
given with reference to α-iron.
the G2 system as a result of the more voluminous ligands
giving rise to more steric hindrance. The steric influence on
each chromophore is not sharply defined, but rather spread
over a certain distribution of distortions. This may be the
reason for the larger line width of the HS doublet in the
G2 system.
Conclusions
FeII triazole complexes form mostly 1D-chains using
dendritic triazoles as ligands, these chains are well encom-
passed by dendritic branches. The material properties of
these complexes are thus similar to those of the ligands
themselves. This is the reason why the preparation of the
pure complex compound with FeII/ligand ratio of 1:3 was
not possible with the present ligand material. Attempts to
fully evaporate a complexation mixture, as done by Fuji-
gaya,[16] did not afford well coordinated material. It turned
out that the best procedure for the synthesis of PBE den-
dritic complexes was to precipitate and wash them at low
temperature (approx. –80 °C). This, however, made it im-
possible to guarantee that no additional ligand precipitates
together with the complex compound. The iron concentra-
tion could be determined by neutron activation analysis,
and together with TGA and elemental analysis the correct
formula weight was calculated.
Fresh samples of the dendritic iron(II) compounds under
study lose crystal water upon heating between 40 and
100 °C with a slight increase of the χMT value. This is ac-
companied by structural changes as proven by XRD mea-
surements. The change of χMT is less pronounced in the
present study than in the report of Fujigaya,[16] probably
due to the fact that in his case the bridging NH group be-
tween the triazole and the dendritic branch offers the pos-
sibility of binding water molecules with significant influence
on the spin state of the iron centre. In our study the systems
contain bridging aliphatic groups which do not interact
with crystal water molecules, and the expected influence on
the spin state of the central iron ion is of minor importance.
All dendritic iron(II) compounds under study exhibit a par-
tial and gradual spin transition below 200 K. The spin tran-
sition is not complete and the abrupt decrease of χMT be-
low 50 K originates from zero-field-splitting, which was
proven by 57Fe-Mössbauer spectroscopy.
Synthesis of the Ligands
Compound 3 (G1-PBE): A one-neck flask was charged with 5.0 g
of 1 (13.04 mmol), 1.61 g of 1H-1,2,4-triazole-1-propanenitrile
(3.13 mmol), and 150 mL of acetonitrile. The mixture was stirred
and heated to 88 °C for 20 h under nitrogen. After cooling to room
temperature, the solvent was removed by evaporation. The residue
was dispersed in 20 mL of diethyl ether; after vigorous stirring for
1 h the ether was decanted. The washing procedure was repeated
three times. The residue was dissolved in 20 mL of aq. NaOH (2 )
and stirred overnight at room temperature for deprotection, and
after dilution with 200 mL of water, it was extracted with chloro-
form (300 mL). The organic phase was dried with MgSO4 and the
solvent was removed by evaporation. The crude product was puri-
fied by column chromatography using a CHCl3/CH3OH mixture
(95:05). The compound was dried under high vacuum to give 3.00 g
Experimental Section
Materials: All solvents used for the synthesis of the ligands were
dried before use. All reactions were carried out under nitrogen at-
mosphere with degassed dry solvents and the dendritic triazoles
were purified by column chromatography using 300–400 and 100–
200 mesh (Macherey–Nagel) silica gel, respectively. 1H-1,2,4-tri-
azole-1-propanenitrile was synthesized from 1H-1,2,4-triazole and
acrylonitrile as described in the literature.[18] All the materials nec-
essary for the syntheses of the complexes are commercially available
and were used as supplied.
1
of white G1-PBE (3) (yield 73%). H NMR (CDCl3): δ = 8.18 (s,
Methods: 1H (300 MHz) and 13C (62.5 MHz) NMR spectroscopy
was carried out on a Bruker type spectrometer at room tempera-
2 H, TrZ), 7.41–7.35 (m, 10 H, Ph), 6.63 (s, 1 H, Ph), 6.40 (s, 2 H,
Ph), 5.09–5.02 (m, 6 H, 2Ph-CH2,1 NCH2) ppm. 13C NMR
Eur. J. Inorg. Chem. 2008, 1613–1622
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1619

P. Sonar, C. M. Grunert, Y. Wei, J. Kusz, P. Gütlich, A. D. Schlüter
[Fe(G2-PBE)2.7](tos)2·2.9H2O: Yield 63.5%. C152H141FeN8O25S2
FULL PAPER
(CDCl3): δ = 160.65, 139.46, 136.31, 128.64, 128.58, 128.16, 127.44,
127.09, 106.88, 102.28, 70.27, 70.16, 49.03 ppm. MALDI-TOF (2600): calcd. C 70.09, H 5.48, N 4.36, S 2.47, Fe 2.15; found C
mass calcd. for C23H21N3O2: [M + H] + 371.44; found 371.
C23H21N3O2 (371.44): calcd. C 74.37, H 5.70, N 11.31; found C
74.20, H 5.23, N 11.17.
70.00, H 6.27, N 4.15, S 1.77, Fe 2.15.
Compound 5 (G2-PBE): A one-neck flask was charged with
2.0 g
4 (2.47 mmol), 1H-1,2,4-triazole-1-propanenitrile 0.307 g
(2.47 mmol) and 80 mL of acetonitrile. The mixture was stirred and
heated to 88 °C for 20 h. After cooling to room temperature, the
solvent was removed by evaporation. The residue was dispersed in
30 mL of diethyl ether; after vigorous stirring for 1 h the ether was
decanted. The washing procedure was repeated three times. The
residue was dissolved in 15 mL of aq. NaOH (2 ) and stirred for
overnight at room temperature for deprotection, and after dilution
with 200 mL of water, it was extracted with chloroform (300 mL).
The organic phase was dried with MgSO4 and the solvent removed
by evaporation. The crude product was purified by column
chromatography using CHCl3/CH3OH mixture in the ratio of
95:05. The compound was dried under high vacuum and gave 1.2 g
of G2-PBE 5 (yield 61%). 1H NMR (CDCl3): δ = 8.12 (s, 2 H,
Trz), 7.38–7.30 (m, 20 H, Ph), 6.61–6.50 (m, 6 H, Ph), 6.32 (s, 3 H,
Ph), 5.01 (s, 12 H, 6CH2Ph), 4.90 (s, 2 H, NCH2) ppm. 13C NMR
(CDCl3): δ = 160.12, 139.01, 138.78, 136.81, 128.53, 127.95, 127.53,
108.18, 106.56, 101.75, 70.13, 29.66 ppm. MALDI-TOF mass
calcd. for C51H45N3O6: [M + H] + 795.94; found 796. C51H45N3O6:
calcd. C 76.96, H 5.70, N 5.28; found C 76.35, H 5.43, N 5.34.
Synthesis of the Complexes: For the preparation of the complexes a
methanol solution of the iron(II) triflate and tosylate, respectively,
containing a small amount of ascorbic acid to avoid oxidation of
iron(II) to iron(III) was added dropwise into a THF solution of
the ligand with a molar ratio of iron to ligand of 1:3 at 60 °C. The
reaction mixture was stirred at this temperature for about 30 min,
then cooled down to room temperature. Part of the solvent was
removed under reduced pressure until the solution was separated
into two layers. The solution with the pink gelatine like product
was cooled with liquid nitrogen to about –50 °C to –100 °C. In that
temperature region the product became solid and inflexible and
could be collected and washed with cold methanol. The complexes
were dried at room temperature.
Elemental analysis (C, H, N and S) was carried out by conventional
combustion analysis; the iron content was determined by neutron
activation analysis. The amount of iron was used to determine the
molecular weight per formula unit of one complex (which corre-
sponds to the repeat unit of Fe triazole chains). TG analysis was
used to quantify the solvent content (see the following section).
The remainingmolecular weight was assigned to the respective li-
gand and the following values were obtained. They match the cal-
culated ones. Deviations in the ratio of iron to ligand from the
theoretical ratio of 1:3 of infinite 1D iron triazole complex chains
are due to the conditions during the synthesis and will be discussed
above.
[Fe(G1-PBE)3.7](trif)2·1H2O: Yield 40.9%. C87H80F6FeN11O14S2
(1764): calcd. C 59.91, H 4.60, N 8.90, S 3.67, Fe 3.20; found C
59.69, H 5.02, N 8.81, S 3.74, Fe 3.16.
[Fe(G1-PBE)3](tos)2·4H2O: Yield 92.6%. C83H85FeN9O16S2 (1585):
calcd. C 62.91, H 5.41, N 7.96, S 4.05, Fe 3.52; found C 63.90, H
5.40, N 7.92, S 3.91, Fe 3.39.
[Fe(G2-PBE)3.4](trif)2·4.3H2O: Yield 56.2%. C175H162F6FeN10-
O31S2 (3137.61): calcd. C 67.14, H 5.19, N 4.55, S 2.044, Fe 1.78;
found C 66.67, H 5.28, N 4.55, S 1.87, Fe 1.79.
Figure 6. Thermogravimetrical analysis of a) [Fe(G1-PBE)3.7](trif)2·
1H2O, b) [Fe(G1-PBE)3](tos)2·4H2O, c) [Fe(G2-PBE)3.4](trif)2·
4.3H2O and d) [Fe(G2-PBE)2.7](tos)2·2.9H2O.
1620
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1613–1622

Iron(II) Spin Transition Complexes with Dendritic Ligands, Part I
Characterization of the Complexes
reduction at about 80 °C, but at 150 °C a weight loss of 2.5% corre-
sponding to ca. 3.8 molecules of water in the case of [Fe(G2-PBE)3]-
(triflate)2·xH2O. [Fe(G2-PBE)3](tosylate)2·xH2O shows a gradual
release of 2% of its mass which fits to 2.9 water molecules. All four
complexes are stable up to about 260 °C, with small deviations of
a few degrees between the species. Above this temperature the com-
plexes decompose in small steps. The main decomposition takes
place very rapidly around 380 °C and 410 °C.
Thermogravimetric Analysis (TGA): TGA confirmed the existence
of included solvent, which comes from the synthesis and similar
systems are reported in the literature,[16,7] in our system we also
confirm the solvent as most likely water. The heating curves showed
two steps, one at approximately 80 °C and the other at 150 °C (Fig-
ure 6, a–d, expanded insets).
This may indicate two different positions in the lattice. The loss of
1% of mass in [Fe(G1-PBE)3](triflate)2·xH2O corresponds to about
one water molecule (Figure 6, a). The analogous tosylate com-
pound (Figure 6, b) exhibits a mass loss of 4.5% corresponding to
approximately four water molecules per formula unit. Complexes
with G2 ligands did not show the first pronounced step of mass
FT-Infrared Spectroscopy (IR): Figure 7 shows the IR spectra of
the ligands G1-PBE and G2-PBE as well as their complexes with
triflate and tosylate anions. This overview proves the complexation
of the iron salts with the ligands and gives first hints for a partial
band assignment by comparing the spectra of the free ligands and
Figure 7. IR spectra of ligands and complexes exhibit differences due to the respective generation (g) and various counterions (trif or
tos) for the complexes. Triazole bands (trz) are almost not affected in the different compounds, but often covered by other bands. Band
overlaps e.g. in the case of the counterions make assignment more difficult.
Table 2. Band assignment of selected absorption bands for the free dendritic ligands and their iron(II) complexes.
G1-PBE
G2-PBE
[Fe(G1-PBE)3]
(trif)2·H2O
[Fe(G1-PBE)3]
(tos)2·4H2O
[Fe(G2-PBE)3]
(trif)2·4.3H2O
[Fe(G2-PBE)3]
(tos)2·2.9H2O
νtrz(C–H)
νar(C–H)
3134 vw
3090 w
3064 w
3031 w
3010 vw
2946 w
2910 w
2864 w
1612 m
1598 vs
1585 m
1559 vw
1454 m
1441 m
–
3112 vw
3088 vw
3063 w
3031w
3008 vw
2925 w
2871 w
1314 vw
3092 w
3065 w
3033 w
3009 vw
2930 w
2875 w
1308 vw
3090 w
3064 w
3032 w
3010 vw
2922 w
2871 w
1310 vw
3090 w
3065 w
3032 w
3010 vw
2929 w
2872 w
1310 vw
3088 w
3063 w
3032 w
3008 vw
2923 w
2871 w
νalk(C–H)
νar(C=C)
1608 m
1595vs
1609 m
1596vs
1609 m
1596 vs
1608 m
1597vs
1608 m
1597vs
ν(N=C–N)
δ(CH2)
1560 vw
1452 s
1558 vw
1452 s
1559 vw
1452 s
1557 vw
1451 s
1559 vw
1452 s
ν(C–F)
–
1259 m
–
1255 m
–
νasym(O=S=O)
νsym(O=S=O)[a]
ν(Car–O)[b]
ν(Calk–O)[b]
δasym(O=S=O)
γ(trz)[c]
–
–
–
–
1277 m
1212 m
1123 w
1165 s
1155 s
681 m
632 w
568 m
1281 m
1214 m
1123 w
1157vs[b]
[a]
[a]
1174 vs
1155 s
–
634 w
–
1157vs[b]
1163 s
1154 s
1156vs[b]
–
680 m
632 w[c]
568 m
632 w
–
638 m[c]
636 m
[c]
[c]
δsym(O=S=O)
[a] νsym overlaps with ν(C–O) in case of A = triflate. [b] ν(Car–O) and ν(Calk–O) overlap in case of G2-PBE. [c] γ(trz) overlaps with
δsym(O=S=O) in case of A– = triflate.
Eur. J. Inorg. Chem. 2008, 1613–1622
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1621

P. Sonar, C. M. Grunert, Y. Wei, J. Kusz, P. Gütlich, A. D. Schlüter
FULL PAPER
[4]
[5]
N. Willenbacher, H. Spiering, J. Phys.: Condens. Matter 1988,
21, 1423–1439.
Y. Galyametdinov, V. Ksenofontov, A. Prosvirin, I. Ovchinni-
kov, G. Ivanova, P. Gütlich, W. Haase, Angew. Chem. Int. Ed.
2001, 40, 4269–4271.
A. B. Gaspar, V. Ksenofontov, M. Seredyuk, P. Gütlich, Coord.
Chem. Rev. 2005, 249, 2661–2676.
M. Seredyuk, A. B. Gaspar, V. Ksenofontov, S. Reiman, Y. Ga-
lyametdinov, W. Haase, E. Rentschler, P. Gütlich, Chem. Mater.
2006, 18, 2513–2519.
the ligands in the coordination sphere around the iron centre. Fur-
ther information for the band assignment was based on literature
by Socrates[23] and Fujigaya.[16]
The bands in the νC–H stretching region of around 3000 cm–1 were
assigned by intensity considerations. Whereas the C–H stretching
intensity of the triazole unit should stay constant that of the den-
dron should increase with the generation. Therefore, the bands be-
tween 2800 and 3050 cm–1 were assigned to the dendron and those
at about 3100 cm–1 to triazole.
[6]
[7]
[8]
J. F. Létard, P. Guionneau, L. Goux-Capes in Topics in Current
Chemsitry (Eds.: P. Gütlich, H. A. Goodwin), Springer-Verlag,
Berlin, Heidelberg, New York, 2004, vol. 235, 221–249.
J. A. Real, A. B. Gaspar, V. Niel, M. C. Munoz, Coord. Chem.
Rev. 2003, 236, 121–141.
A. Galet, A. B. Gaspar, M. C. Munoz, G. V. Bukin, G. Levch-
enko, J. A. Real, Adv. Mater. 2005, 17, 2949–2953.
J. A. Real, A. B. Gaspar, M. C. Munoz, Dalton Trans. 2005,
2062–2079.
A. Hauser, J. Adler, P. Gütlich, Chem. Phys. Lett. 1988, 152,
468–472.
A similar picture can be found in the region between 400 and
1700 cm–1. Bands of the triazole ring do not exhibit significant dif-
ferences between first and second generation (e.g. 1498, 695 cm–1),
while the bands of the benzyl ether show a stronger absorption for
the G2-PBE (e.g. around 1600, 1380, and 830 cm–1). Beside these
differences, the spectra of the complexes also include bands of their
respective counterions. They occur for instance at 1260, 637 and
518 cm–1 in the case of triflate and at 1213, 1011 and 568 cm–1 for
tosylate. More details can be found in Table 2. A complete band
assignment was not attempted because of overlap and general com-
plexity. Furthermore, all complexes contain a small amount of
water in the lattice giving rise to the typical broad band around
3400 cm–1 (not shown).
[9]
[10]
[11]
[12]
[13]
[14]
Y. Maeda, M. Miyamoto, Y. Takashima, H. Oshio, Inorg.
Chim. Acta 1993, 204, 231–237.
A. D. Schlüter, J. P. Rabe, Angew. Chem. Int. Ed. 2000, 39, 864–
883.
[15]
[16]
H. Frauenrath, Prog. Polym. Sci. 2005, 30, 325–384.
T. Fujigaya, D.-L. Jiang, T. Aida, J. Am. Chem. Soc. 2005, 127,
5484–5489.
M. Balasubramanian, J. G. Keay, E. F. V. Scriven, Heterocycles
1994, 37, 1951–1975.
A. Horvàth, Synthesis 1995, 9, 1183–1189.
C. Hawker, J. Frèchet, J. Am. Chem. Soc. 1990, 112, 7638–7647.
Z. Yu, C. Li, X. Z. You, H. Spiering, P. Gütlich, Y. F. Hsia,
Mater. Chem. Phys. 1997, 48, 150–155.
J. Jung, H. Spiering, Z. Yu, P. Gütlich, Hyperfine Interact.
1995, 95, 107–128.
C. M. Grunert, K. Eberhardt, P. Gütlich, P. Thörle, Annual
Report, 2005, Institut für Kernchemie, Johannes Gutenberg
University, Mainz, 2006.
Acknowledgments
We thank B. Mathiasch for recording the infrared spectra, K. Eber-
hardt and P. Thörle for the determination of the iron content by
neutron activation analysis. We cordially thank M. Colussi for the
DSC and TGA measurements and Dr. Christian Baerlocher for
the XRD measurements (both ETH Zürich). We are grateful for
financial support from the Deutsche Forschungsgemeinschaft
(DFG), Program No. 1137 (Molecular Magnetism) and the Fonds
der Chemischen Industrie.
[17]
[18]
[19]
[20]
[21]
[22]
[1] P. Gütlich, H. A. Goodwin in Spin Crossover in Topics in Cur-
rent Chemistry, Springer-Verlag, Berlin, Heidelberg, New York,
2004, vol. 233–235.
[2] P. Gütlich, A. Hauser, H. Spiering, Angew. Chem. Int. Ed. Engl.
1994, 33, 2024–2054.
[3] H. Spiering, N. Willenbacher, J. Phys.: Condens. Matter 1989,
1, 10089–10105.
[23]
G. Socrates in Infrared and Raman Characteristic Group Fre-
quencies, third ed., John Wiley & Sons, Chichester, 2001.
Received: August 5, 2007
Published Online: February 11, 2008
1622
www.eurjic.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2008, 1613–1622