Liquid crystalline octahedral iron(III) complexes with 1,4,7-tris[3,4-bis-(decyloxy)benzyl]-1,4,7-triazacyclononane: Thermal characterization and Mossbauer investigations of bridging and redox behavior

Article abstract of DOI:10.1002/(SICI)1521-3765(199801)4:1<93::AID-CHEM93>3.0.CO;2-X

The complexation of iron(III) chloride with 1,4,7-tris[3,4-bis(decyloxy)-benzyl]-1,4,7-triazacyclononane leads to a liquid crystalline product, which forms a second mesophase after thermal treatment at around 70°C above the first clearing temperature. Pol

Full text of DOI:10.1002/(SICI)1521-3765(199801)4:1<93::AID-CHEM93>3.0.CO;2-X

FULL PAPER
Liquid Crystalline Octahedral Iron(iii) Complexes with 1,4,7-Tris[3,4-bis-
(decyloxy)benzyl]-1,4,7-triazacyclononane: Thermal Characterization and
Mössbauer Investigations of Bridging and Redox Behavior
G. Henning Walf, Rüdiger. Benda, F. Jochen Litterst,* Uwe Stebani, Steven Schmidt, and
Günter Lattermann*
Abstract: The complexation of iron(iii) chloride with 1,4,7-tris[3,4-bis(decyloxy)-
benzyl]-1,4,7-triazacyclononane leads to a liquid crystalline product, which forms a
second mesophase after thermal treatment at around 708C above the first clearing
temperature. Polarizing microscopy, DSC, and UV observations reveal a trans-
formation, which was further investigated by Mössbauer spectroscopy. In different
temperature ranges, various octahedral, binuclear iron(iii) complexes are present
along with a monomeric species. In the presence of water and oxygen, the iron(iii)
center participates in a redox equilibrium.
Keywords: bridging
iron ´ liquid crystals ´ Moessbauer
spectroscopy ´ redox chemistry
ligands
´
Introduction
The possibility of combining the properties of liquid crystals
and metal complexes has led to intensive research on metal-
lomesogens.[1±5] The great majority of them exhibit linear or
square planar geometry at the metal center. However,
because of the dependence of the physical properties (e.g.
of Mössbauer spectroscopic investigations, which revealed
their octahedral structure, their bridging behavior, and a
redox equilibrium of the iron centers.
magnetism, color) on the geometry of the related ligand field,
it is highly desirable to vary the geometry of the complex. In
1992 an octahedral liquid crystalline nickel(ii) complex was
reported.[6] Details of a further series of octahedral metal
complexes, containing various ligands and forming smectic or
columnar mesophases, have been published recently.[7±14] In
the past, a limited number of low molecular[15±25,8] and
polymeric[26±28,21] liquid crystalline iron(iii) complexes have
been investigated with regard to their magnetic properties.
Besides ESR, Mössbauer spectroscopic investigations have
also been performed.[18,19,23] In this paper, we describe the
mesomorphism of iron(iii) complexes containing the ligand
1,4,7-tris[3,4-bis(decyloxy)benzyl]-1,4,7-triazacyclononane
(L), their thermal transformations, and the results
Experimental Section
Materials: Tetrahydrofuran (THF) and dioxane were dried and purified
under inert gas by refluxing with potassium and subsequent distillation.
Inert gas for chemical reactions (N2 or Ar) was purified and dried in
columns filled with a molecular sieve and potassium on aluminum oxide.
3,4-Bis(decyloxy)benzoyl chloride was obtained from methyl 3,4-dihydrox-
ybenzoate (Lancaster) by a reported procedure.[29]
Instruments: FTIR: Bio Rad/Digilab FTS40. UV/VIS: Hitachi U-3000. 1H
NMR: Bruker AC250 (250 MHz). GC: Waters model 510 with UV
detector (Waters 440, 254 nm), RI detector (Waters 410); eluent THF;
calibration with PS standards; column combination: PL-gel 600 Â 7.5, pore
width 100, 500 Š; particle size: 5 mm. EI/MS: Finnigan, MAT8500;
MAT112 S Varian. Elemental analysis: Mikroanalytisches Labor Beetz
Kronach; Fe: Mikroanalytisches Labor Pascher Remagen. Polarizing
microscopy: Leitz Labolux 12-Pol, hot stage Mettler FP82, Mettler
FP80; camera Wild MPS45/51S. DSC: Perkin-Elmer DSC7; standard
heating rate: 10 Kmin 1. TGA: Netzsch STA409C.
[*] F. J. Litterst, G. H. Walf, R. Benda
Institut für Metallphysik, Technische Universität Braunschweig
D-38106 Braunschweig (Germany)
Fax: Int. code ‡ 49531391-5129
e-mail: j.litterst@tu-bs.de
G. Lattermann, U. Stebani, S. Schmidt
Makromolekulare Chemie I, Universität Bayreuth
D-95440 Bayreuth (Germany)
Mössbauer absorption experiments: 57Co-in-Rh source at RT, standard
Mössbauer cryostat; spectra were recorded between 4 K and 250 K;
calculation of the spectra by least-square fits. Isomer shifts are referred to
metallic iron at RT, the values for linewidth are not calculated in thin
Fax: Int. code ‡ 4992155-3206
e-mail: guenter.lattermann@uni-bayreuth.de
Chem. Eur. J. 1998, 4, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0401-0093 $ 17.50+.25/0
93

F. J. Litterst, G. Llattermann et al.
FULL PAPER
absorber approximation but by solving the transmission integral as
described previously.[30,31] Thermal treatments were performed in a stand-
ard oven under defined atmospheric environments (air, O2, or Ar). The
Mössbauer spectra of thermally treated samples were obtained after the
sample had cooled to RT, been removed from the oven, and been
immediately loaded into the He atmosphere of the Mössbauer cryostat.
1,4,7-Tris[3,4-bis(decyloxy)benzyl]-1,4,7-triazacyclononane: Under inert
gas, 1,4,7-triazacyclononane ([9]aneN3, 2 mmol, Fluka), 3,4-bis(decyloxy)-
benzoyl chloride (6.1 mmol), and 4-dimethylaminopyridine (DMAP)
(6.1 mmol) were dissolved in dry dioxane (100 mL). The reaction mixture
was stirred under reflux for 12 hours. After cooling to room temperature,
DMAP ´ HCl was filtered off, the solvent evaporated, and the slightly
yellow, oily residue purified by chromatography [alumina N, activity
grade I (ICN)] with hexane/ethyl acetate 1:1, and subsequent recrystalliza-
tion (methanol/THF 5:1). Yield: 50% as a white waxy solid, freeze-dried
from benzene solution.
The corresponding amine ligand L was obtained by reduction of the
reaction product as described earlier[6] and freeze-dried from benzene
solution. The purity of the products was checked by 1H NMR, IR, GPC, EI-
MS, and elemental analysis.
[a]
L ´ FeCl3 (Samples B and C): Anhydrous FeCl3 (1.43 mmol) in dry THF
(70 mL) was added dropwise at 408C to a solution of L (1.43 mmol) in dry
THF (70 mL). The mixture was stirred at 408C for 12 hours, and the solvent
evaporated. The residue was dissolved in hexane, insoluble particles were
filtered off, and the solvent evaporated again. The residue was freeze-dried
from benzene solution (sample C). Sample B was recrystallized from
isooctane over a period of several weeks in the refrigerator. Yield: 96%.
Sample B: IR (KBr): nĈ 2955, 2924, 2855, 1605, 1588, 1516, 1467, 1429,
1391, 1267, 1144, 1016 cm 1; C87H153Cl3FeN3O6 (1499.32 gmol 1): calcd: C
69.89, H 10.28, N 2.80, Cl 7.09, Fe 3.72, O 6.40; found: a) no thermal
treatment, freeze-dried: C 68.99, H 10.24, N 3.08, Cl 7.17, Fe 3.76, O 6.76;
b) annealed for 30 minutes at 1308C (H2O atmosphere), freeze-dried: C
68.66, H 10.24, N 2.91, Cl 7.11, Fe 3.74, O 7.34; c) annealed for 2 days at
1308C (H2O atmosphere), freeze-dried: C 67.73, H 9.50, N 2.82, Cl 6.95, Fe
3.79, O 9.21.
L ´ FeCl3 ´ 6H2O[a] (Sample A): The same synthetic procedure as for [L ´
FeCl3], but with FeCl3 ´ 6H2O. Recrystallization from diethyl ether/ethanol
1:1 over a period of several weeks in the refrigerator. Yield: 94%. IR
(KBr): nĈ 2955, 2924, 2855, 1605, 1588, 1516, 1467, 1429, 1391, 1267, 1144,
1016, 667 cm 1; C87H153Cl3FeN3O6 (1499.32 gmol 1): calcd: C 69.89, H
10.28, N 2.80, Cl 7.09, Fe 3.72, O 6.40; found: C 69.22, H 10.36, N 3.00, Cl
6.18, Fe 3.98, O 7.26.
Figure 1. Optical textures of complex L ´ FeCl3: top) M1, at 538C ( Â 120);
bottom) M2, at 958C ( Â 120).
[a] Empirical formulae only are used here, since we do not know the precise
structural formulae of the complexes.
Results
Thermal behavior: after melting at 488C, the two complexes
of L with anhydrous FeCl3 (L ´ FeCl3, samples B and C)
formed a viscous phase, M1, which exhibited slight double
refraction under the polarizing microscope (Figure 1, top). A
transition to the isotropic phase was observed at 598C. DSC
measurements of a second heat treatment (Figure 2) revealed,
besides the clearing temperature, only a glass transition
around 258C. Some small black particles still existed in the
isotropic phase, but they disappeared at 130 ± 1408C. Surpris-
ingly, after heating to 130 ± 1408C, a new mesophase was
observed under the polarizing microscope on cooling to 988C
(Figure 1, bottom). On subsequent heating, this new meso-
phase, M2, exhibited a clearing temperature of 103.58C.
In principle, the complex of L with FeCl3 ´ 6H2O (L ´ FeCl3 ´
6H2O) exhibited the same behavior with somewhat different
transition temperatures. Above the melting point (478C), M1
cleared at 568C. When heating for the first time to a maximum
temperature of 1408C (Figure 3a), DSC measurements
Figure 2. DSC thermogram of complex L ´ FeCl3; Tmax ˆ 708C, heating
rate: 2 Kmin 1: a) first heating; b) second and further heatings; c) cool-
ing.
Figure 3. DSC thermogram of the complex L ´ FeCl3 ´ 6H2O on first
heating up to 1408C (heating rate: 10 Kmin 1): a) first heating; b) cooling.
94
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0401-0094 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 1

Liquid Crystalline Iron Complexes
93±99
showed a significant change in the heat flow above 958C. The
clearing temperature of M2 on heating for the first time is
about 1258C. On cooling from higher temperatures in the
isotropic phase (Figure 3b), M2 reappears at 1308C. There
was no transition at ꢀ 958C, no transition to M1, and no
recrystallization; however, a glass transition was detected. On
further heating, besides the glass transition at 208C, only the
clearing peak of M2, stabilized at 1368C, was observed. The
transition temperatures of both complexes are summarized in
Table 1.
reprecipitated from isooctane over a period of several weeks
in the refrigerator, sample C was used directly after synthesis.
The spectra given in Figure 4 show superpositions of two
quadrupole doublets (I and II) for all three samples A, B, and
Table 1. Transition temperatures, transition enthalpies (in brackets) and DCp values of the
complexes L ´ FeCl3 and L ´ FeCl3 ´ 6H2O.[a]
Tg
DCp
K !M1
M1 !M2
M2 !I
1
[8C] [kJmol 1 K
]
[8C (kJmol 1)] [8C (kJmol 1)] [8C (kJmol 1)]
[L ´ FeCl3]
[L ´ FeCl3 ´ 6H2O] 20
25
0.23
0.18
53.0 (11.3)[b]
47.0 (10.0)[b]
60.5 (1.1)[c]
56.0 (0.8)[c]
103.5 (1.7)[d]
136.0 (2.6)[d]
[a] Tg ˆ glass transition temperature; K ˆ crystalline phase; M ˆ mesophase; I ˆ isotropic
phase. [b] Only on first heating. [c] Only at a maximum heating temperature of Tmax ˆ 708C.
[d] Only on further heating, after a first maximum heating temperature of Tmax ˆ 1408C.
The change of the heat flow around 958C does not
correspond to the onset of the pyrolytic decomposition of
the complexes at 1708C, which was revealed by thermogravi-
metrical measurements (TGA). Furthermore, a dissociation
(without weight loss) to the free ligand and the salt cannot
explain the transition around 958C because of the existence of
the mesophase M2, which in consequence must then be caused
by another process: a transformation of the original complex,
for example. This transformation is also indicated by UV
spectroscopy. In the samples not heated above 708C, a new
band at 330 nm appeared for both complexes after annealing
for 30 minutes at 1308C.
X-ray measurements[32] of M1 of both complexes (L ´ FeCl3 ´
Figure 4. a ± c) Mössbauer spectra of samples L ´ FeCl3 ´ 6H2O (A), L ´
FeCl3 (B and C) in the initial state (i.e. without thermal treatment) at
different temperatures; d) Mössbauer spectrum of sample C (L ´ FeCl3)
after heating under Ar to 1308C.
6H2O and L ´ FeCl3) revealed only one reflection at 29.7 Š
and a halo at 4.4 Š for the fluid alkoxy chains. This indicates
the existence, but not the type, of the mesophase. In M2, in
addition to the first-order reflection at 33.7 Š and the halo at
4.4 Š, a second-order reflection at 16.9 Š was observed. This
is consistent with either a columnar or a lamellar phase.
The fact that the observed transformation is not due to
decomposition leads to the assumption of a modification in
the coordination sphere of the complexes.
A precise determination of the complex structure by X-ray
structural analysis is not possible, because of the very poor
recrystallization behavior of the complexes before the ther-
mal transformation (common for liquid crystals) and the total
disappearance of the crystalline phase after the thermal
transformation, even for months.
C, along with a further signal in the middle (IIIA, B, C). In all
the samples, each of the doublets I and II exhibits the same
hyperfine parameters, and the areas of the two doublets are in
a ratio of 1:1. The values of the quadrupole splittings are DI ˆ
1
1.58 Æ 0.01 and DII ˆ 1.26 Æ 0.02 mms . They are nearly
temperature-independent, slightly decreasing at higher tem-
1
peratures (dDIIIA/dTˆ (2.5 Æ 0.3)  10 4 mms 1 K ). The
temperature dependence of the isomer shifts dI ˆ 0.37 Æ 0.01
1
and dII ˆ 0.45 Æ 0.02 mms is negligible. The parameters of
linewidth for doublet I and II could be kept at the same value
and the doubletsꢁ relative areas remained as 1:1 between 4 K
and 200 K when fitting all the spectra (within a limit for
deviation of 1%).
Mössbauer spectroscopy:
These characteristics are typical for octahedrally coordi-
nated, binuclear high-spin iron(iii) complexes, for example
haemerythrin derivatives[33±35] as well as for nonmesogenic
iron(iii) complexes with 1,4,7-triazacyclononane ligands.[36,37]
The two kinds of iron sites, present in a 1:1 ratio, have two
slightly different environments. Therefore, this species can be
regarded as an asymmetric dimer. The spectra differ in shape
The initial state of the samples: In order to shed light on the
situation, three samples were investigated by Mössbauer
spectroscopy: Sample A corresponds to the complex L ´
FeCl3 ´ 6H2O, reprecipitated from a 1:1 mixture of ethanol/
diethylether over a period of several weeks in the refrigerator.
Both samples B and C correspond to L ´ FeCl3; sample B was
Chem. Eur. J. 1998, 4, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0401-0095 $ 17.50+.25/0
95

F. J. Litterst, G. Llattermann et al.
FULL PAPER
Table 2. Mössbauer hyperfine parameters. Isomer shifts are relative to
metallic iron, quadrupole splittings and isomer shifts are at 70 K.
and intensity with respect to the third signal in the middle.
Futhermore, signal IIIA of sample A (L ´ FeCl3 ´ 6H2O) shows
a broadened doublet. The quadrupole splitting (DIIIA ˆ 0.7 Æ
1
1
1
d [mms
]
D [mms
]
G [mms
]
Bhf [T]
1
0.03 mms ) and the weak temperature dependence of the
Doublet I
Doublet II
IIIA
IIIC
Fe(ii)
0.37 Æ 0.01
0.45 Æ 0.02
0.45 Æ 0.01
0.45 Æ 0.02
1.04 Æ 0.03
1.00 Æ 0.06
0.37 Æ 0.03
1.58 Æ 0.01
1.26 Æ 0.02
0.70 Æ 0.03
0.151 Æ 0.002
0.151 Æ 0.002
0.35 Æ 0.05
±
±
50
±
±
±
isomer shift (dIIIA ˆ 0.45 Æ 0.01 mms 1; ddIIIA/dT as for dou-
blets I and II) are typical for monomeric high-spin iron(iii)
species.[33] At higher temperatures, the recoil-free fraction of
the g-radiation (f-factor) is found to decrease much more
slowly than for all the other signals. Therefore, we were able
to obtain spectra consisting only of the signal IIIA at temper-
atures as high as 250 K (Figure 5 c).
0.1 < D < 0.25 0.3 < G < 0.5
2.90 Æ 0.01
Fe(ii)
Doublet IV
2.40 Æ 0.01
1.30 Æ 0.05
0.25 Æ 0.01
±
atmosphere. This treatment transformed equal parts of the
intensity of line IIIC into the doublets I and II (see Fig-
ure 4d). Therefore, IIIC could be interpreted as a metastable
molecular modification, which is, despite a very small quadru-
pole splitting, closely related to the binuclear or dimeric
structure corresponding to doublets I and II (asymmetric
dimer), but with a high local symmetry. This interpretation as
a symmetric dimer is supported by the temperature depend-
ence of the f-factor, which is exactly the same for IIIC as for
doublets I and II, but quite different from IIIA, which
corresponds to the monomeric iron(iii) structure (discussed
in detail below). Apparently, the symmetrical dimer is a
metastable modification first formed in the water-free com-
plex at low temperatures and then rearranged irreversibly to
the thermodynamically more stable asymmetric dimer (in the
spectra seen as doublet I and II) by heating above 1308C
under an atmosphere of inert gas.
The f-factor is proportional to the negative exponent of the
mean square displacement of the Mössbauer nucleus. In the
harmonic approximation[39,40] for the atomic vibrations it is
possible to derive a Debye temperature from the temperature
dependence of f: a value of qD,dimer ˆ 65 K was determined
from signals I, II, and IIIC, and qD,monomer ˆ 86 K from IIIA.
We performed calculations on the f-factors of these two
different molecular configurations, using the model of a chain
of harmonic oscillators for the intramolecular oscillations and
the Debye model for the intermolecular oscillations. We
obtained a relative change in the Debye temperatures for
monomeric molecules in comparison to dimeric ones, result-
p
3
ing in qD,monomer
ˆ
2qD,dimer for intermolecular oscillations
(bigger molecules/longer chains have more phonon modes, i.e.
they are softer) and qD,monomer
p
ˆ
2qD,dimer for intramolecular
oscillations (since the integrals over all Debye frequencies
have to be normalized by the number of molecules, the
normalization has to be done by N for bridged molecules and
by 2N for monomer molecules), respectively. The experimen-
tal ratio of the two Debye temperatures is 0.77, which lies
between the calculated values (intermolecular: 0.71, intra-
molecular: 0.79). This result confirms the assignment of
monomeric and dimeric molecules to the related signals in the
spectra.
Figure 5. Mössbauer spectra of sample B L ´ FeCl3 a ± c) after heating in air
to 1308C, at different temperatures; d) after storage for a few days in air at
RT.
The spectra of samples B and C (L ´ FeCl3; Figure 4b ± d)
show broad signals IIIB and IIIC, which are supposed to be
equivalent. Because of the weakness of IIIB, we shall discuss
only IIIC. Variation of temperature between 4 and 200 K does
not change the relative area of signal IIIC, which is constant
at all temperatures and amounts to 46 Æ 2% of the total area
of all the lines added together. The fitted linewidth GIIIC and
the small quadrupole splitting DIIIC are, however, highly
correlated, which causes large uncertainties in both the
parameter values (Table 2). A significant reduction of signal
IIIC was obtained on heating this sample in an argon
Thermal transformations: An irreversible transformation
above 1308C was observed by DSC measurements and
polarizing microscopy for samples B and C (L ´ FeCl3) under
Ar. Similar behavior was found for sample A (L ´ FeCl3 ´
6H2O). For a more detailed description of these transforma-
tions, we performed Mössbauer spectroscopy on all the
96
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0401-0096 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 1

Liquid Crystalline Iron Complexes
93±99
samples after thermal treatment in air. In principle, the results
are comparable for all three samples. However, we explain the
results on the basis of the spectra of sample B (L ´ FeCl3).
Figure 5a ± c shows a temperature scan (40 ± 250 K) of sam-
ple B. The spectra were recorded immediately after heating to
1308C in air and subsequent cooling to the indicated temper-
ature. The spectrum in Figure 5d was taken after storage of
the heated sample for a few days at RT in air or pure oxygen,
which led to equivalent results.
The observed differences between spectra 5a ± c and those
obtained in the initial state prove that there has been a
fundamental change of the iron environment: after
heating, the initial Fe(iii) compound has changed and shows
at least two quadrupole doublets typical for Fe(ii)
1
1
(D ˆ 2.90 Æ 0.01 mms , d ˆ 1.04 Æ 0.03 mms
and D ˆ
1
1
2.40 Æ 0.01 mms , d ˆ 1.00 Æ 0.06 mms ). Their intensities
become negligible for higher temperatures (Figure 5c) due to
their small f-factor. The quadrupole splittings decrease
strongly with increasing temperature; this indicates that the
iron(ii) centers are in the high-spin state.
The values D and d of doublet IV in Figure 5d come close to
those of the doublets I and II of the spectra of the same
sample before thermal treatment. However, dIV ˆ 0.37 Æ
1
1
0.03 mms and DIV ˆ 1.30 Æ 0.05 mms do not agree per-
fectly with either doublet I or with doublet II. The relatively
high quadrupole splitting and the extremely small f-factor
indicate a dimeric bridged structure again, but one which
contains two similar iron environments that cannot be
distinguished. This behavior implies that heating the samples
for the first time leads to an irreversible change of the bridged
iron centers from an asymmetric species with two different,
well-defined iron environments to a modified bridged struc-
ture with two identical iron sites. Repeated heating of this
reoxidized sample leads to the Fe(ii)-dimeric configuration
again, independent of the atmosphere during the heating
process. This reversible Fe(ii) ± Fe(iii) cycle has been run
several times without any significant changes to either of the
two different kinds of spectra (a ± c and d).
The oxidation and reduction process involving binuclear
molecules suggests the eventual appearance of a mixed-valent
species containing a Fe(ii) ± Fe(iii) bridge. The dynamics of
charge fluctuations should show up in relaxational broadening
and eventual averaging of the hyperfine interactions for fast
(>GHz) fluctuations that could give more insight into the
chemical process. However, our data gives no indication of the
existence of such a species for the whole Fe(ii) ± Fe(iii) cycle.
Figure 6. Mössbauer relaxation spectra at different temperatures observed
for sample A L ´ FeCl3 ´ 6H2O before heating to 1308C, and for all samples
after the thermal transformation in air.
signal IIIA transforms to the magnetic sextet, and the relative
amount of the sextet increases with decreasing temperature.
This indicates a wide distribution of relaxation frequencies.
This is typical for a wide distribution of particle volumes and/
or of magnetic anisotropies. The magnetic splitting reaches its
saturation below about 9 K and corresponds to an effective
magnetic hyperfine field of 50 T, which is typical for high-spin
Fe(iii). The angle between the main axis of the electric field
gradient and the direction of the effective magnetic field can
be derived by the relative line intensities of the six-line
pattern and the distances between the lines 1 ± 2 and 5 ± 6.[33]
The Mössbauer spectra reveal this angle to be 688, which is
clearly distinct from the magic angle of 558 for a random
encirclement.
Even in this temperature range, the lines are still broad-
ened, which indicates a considerable inhomogeneous static
distribution of the magnetic hyperfine interaction.
Low-temperature behavior: Some Mössbauer spectra of
sample A obtained below 30 K are shown in Figure 6. A
defined magnetic hyperfine pattern with six lines develops
from the doublet IIIA. We have attributed these to the
monomeric species. Similar spectra are obtained at low
temperatures for samples B and C after heating in air above
1308C.
No clear onset of the magnetic splitting can be defined, and
the resonance lines are severely broadened, which points to a
relaxational phenomenon with a fluctuation rate of the local
magnetic field at the iron nucleus slowing down below
the nuclear Larmor precession frequency. The doublet
Similar relaxational spectra are usual for small particles of
different iron oxides or hydroxides whose magnetizations
become blocked at low temperatures. Such particles could
have precipitated in our samples as decomposition products.
There are, however, arguments against this interpretation.
Firstly, the recoil-free fraction of the observed signal is very
low: for oxide precipitates one would expect a strong
absorption signal to persist above room temperature, which
is not the case here. Secondly, the relative amount of
signal IIIA with respect to the total spectral area increases
upon slower cooling and decreases for fast quenching. These
Chem. Eur. J. 1998, 4, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0401-0097 $ 17.50+.25/0
97

F. J. Litterst, G. Llattermann et al.
FULL PAPER
reversible changes are not expected for a decomposition
product; rather, they reflect the kinetics of the monomer
formation.
Slow paramagnetic relaxation of isolated Fe3‡ as the origin
of the relaxation spectra can be excluded, since the splitting of
Kramerꢁs doublets for an S-state ion is very small compared to
the thermal energies. We thus propose that the magnetic
hyperfine patterns arise from magnetic ordering between the
monomeric species. This means that the monomeric mole-
cules form regular structures that enable the iron centers to
couple magnetically through exchange interactions and to
establish a static magnetic order at low temperatures
(<40 K). This magnetic order becomes possible only for the
expected regular stacking of the monomeric molecules, which
brings the magnetic ions into close proximity. Both the
dynamic and the static broadenings observed result from
short-range order connected with structural disorder, but also
with the low-dimensional character of the magnetic coupling.
A thorough investigation of the magnetic properties and their
dependence on the molecular structure is in progress.
Scheme 1. Reaction scheme of iron(iii) complexes with the 3,4-bis(decyl-
oxy)benzyl-substituted 1,4,7-triazacyclononane ligand L deduced from
Mössbauer spectroscopy.
for 30 min (annealing in air for 2 days gives the same results)
increased the oxygen content and decreased the chlorine
content in the sample, as shown by elemental analysis, which
demonstrates the suggested replacement of chloro ligands.
From these facts we propose a possible reaction scheme in
terms of molecular structures (Scheme 2). A pentacoordinat-
ed iron(iii) center (belonging here to the asymmetric dimer)
has also been postulated in binuclear haemerythrin iron
complexes.[35] The existence of m-hydroxo or m-oxo bridges has
been reported in complexes with a 1,4,7-triazacyclononane
derivative core.[36±38,41] Additionally, responsibility for the
formation of oxo bridges in a liquid crystalline, ferrocene-
containing Schiff base has been ascribed to atmospheric
moisture.[42] With respect to the redox process, a similar
mechanism involving oxygen and H2O was established for the
interconversion of various forms of haemerythrin.[42]
With respect to liquid crystallinity, the investigations
presented here lead to the conclusion that the low-temper-
ature mesophase, M1, is related to the asymmetric dimer. The
high-temperature mesophase, M2, formed after the thermal
treatment at 1308C must then result either from the monomer
or the dimeric compound formed by the hydrolysis process.
Analogous or similar complexes of iron(iii)chloride with
[9]aneN3 derivatives have always been reported as mono-
meric.[36±38] Bridged binuclear complexes have only been
obtained by subsequent specific reactions.[36±38,41] To our
knowledge, a thermally induced redox equilibrium has never
been described. Only an air oxidation of a binuclear iron(ii)
Discussion and Conclusions
Mössbauer spectroscopic measurements demonstrate that the
peculiar thermal behavior of the liquid crystalline iron(iii)
complexes with the 1,4,7-tris[3,4-bis(decyloxy)benzyl]-1,4,7-
triazacyclononane ligand, L, is due to the existence of various
monomeric and dimeric species as well as reversible redox
reactions. The results are summarized in Scheme 1.
The presence of water or oxygen is of crucial importance for
the formation of the monomeric iron(iii) species as well as for
the redox reaction. Under inert gas in anhydrous conditions,
samples B and C (L ´ FeCl3) do not form the monomeric
iron(III) species (see Figure 5d). In the presence of water, the
monomeric species is formed directly in sample A (L ´ FeCl3 ´
6H2O; see Figure 5a). In samples B and C it is formed after
thermal treatment or even if kept in air (Figure 5d).
On the other hand, the reduction of Fe(iii) to Fe(ii) occurs
on heating A, B, or C above 1308C; for A (L ´ FeCl3 ´ 6H2O)
the reduction occurs even under argon.
Therefore, the reduction process seems to
be connected with the presence of waterÐ
air humidity in the case of B and C, crystal
water in the case of A. It is possible that
the process depends on the formation of
aquo or hydroxo ligands (doublet IV),
which would have replaced chloro ligands.
A further hydrolysis product could then be
the monomeric iron(iii) species. Another
indication for the role of water in a
hydrolysis and reduction process is the
observation that the thermal transforma-
tion around 1308C apparently occurred
much faster for L ´ FeCl3 ´ 6H2O, which
allows the detection of the transformation
in DSC measurements (see Figure 3).
Finally, annealing the water-free sample
L ´ FeCl3 (B, C) in water vapor at 1308C
Scheme 2. Possible reaction scheme for hydrolysis of iron(iii) complexes with ligand L.
98
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998 0947-6539/98/0401-0098 $ 17.50+.50/0 Chem. Eur. J. 1998, 4, No. 1

Liquid Crystalline Iron Complexes
93±99
[18] V. G. Bekeshev, V. Y. Rochev, E. F. Makarov, Mol. Cryst. Liq. Cryst.
1990, 192, 131.
[19] V. G. Bekeshev, V. Y. Rochev, E. F. Makarov, Hyperfine Interact. 1991,
67, 661.
[20] Y. Galyametdinov, G. Ivanova, K. Griesar, A. Prosvirin, I. Ovchinikov,
W. Haase, Adv. Mater. 1992, 4, 739.
[21] W. Haase, K. Griesar, M. F. Iskander, Y. Galyametdinov, Mol. Cryst.
Liq. Cryst. 1993, 3, 115.
[22] D. Huang, J. Yang, L. Zhang, Y. Liu, S. Xiang, Hecheng Huaxue 1994,
2, 1.
[23] N. E. Domracheva, Y. G. Galyametdinov, R. A. Manapov, A. V.
Prosvirin, I. V. Ovchinnikov, W. Haase, K. Griesar, Fiz. Tverd. Tela
(St. Petersburg) 1994, 36, 2154.
[24] Y. G. Galyametdinov, O. N. Kadkin, A. V. Prosvirin, Izv. Akad. Nauk.
Ser. Khim. 1994, 941.
[25] M. Marcos, J. L. Serrano, P. J. Alonso, J. I. Martinez, Adv. Mater. 1995,
7, 173.
complex with a substituted [9]aneN3 ligand to a stable iron(iii)
product has been observed.[36] On the other hand, redox
processes involving metal complexes of 1,4,7-substituted
[9]aneN3 ligands are described for binuclear complexes with
manganese,[43] which can be used in detergent additives as
very efficient catalysts for low-temperature bleaching proc-
esses.[44,45]
It is obviously significant that the described behavior of the
presented iron complexes is closely related to the presence of
ligand L. At the present, we do not know whether it is only the
steric situation of the bulky substituents, the influence of long
alkyl side chains on transition temperatures, or the liquid
crystalline state, with its additional degree of order, which is
responsible for the described bridging and redox reactions of
this example of metallomesogen.
[26] K. Hanabusa, J. Higashi, T. Koyama, H. Shirai, N. Hojo, A. Kurose,
Makromol. Chem. 1989, 190, 1.
[27] P. Singh, M. D. Rausch, R. W. Lenz, Polymer Bull. 1989, 22, 247.
[28] P. J. Alonso, J. I. Martinez, L. Oriol, M. Pinol, J. L. Serrano, Adv.
Mater. 1994, 6, 663.
[29] G. Staufer, G. Lattermann, Makromol. Chem. 1991, 192, 2421.
[30] W. Stieler, M. Hillberg, F. J. Litterst, C. Böttger, J. Hesse, Nucl.
Instrum. Methods Phys. Res. Sect. B 1995, 235.
Acknowledgments: U.S., S.S., and G.L. gratefully acknowledge the
financial support of the Deutsche Forschungsgemeinschaft (DFGLa662/
1 ± 2).
Received: April 29, 1997 [F682]
[31] W. Wagener, M. Hillberg, R. Feyerherm, W. Stieler, F. J. Litterst, T.
Pöhlmann, O. Nuyken, J. Phys.: Condens. Matter 1994, 6, L391.
[32] M. Wittenberg, J. H. Wendorff, Marburg, personal communication.
[33] N. N. Greenwood, T. C. Gibb, Mössbauer Spectroscopy, Chapman &
Hall Ltd., London, 1971.
[34] K. S. Murray, Coord. Chem. Rev. 1974, 12, 1.
[35] R. G. Wilkins, Chem. Soc. Rev. 1992, 171.
[36] J. R. Hartmann, R. L. Rardin, P. Chaudhuri, K. Pohl, K. Wieghardt, B.
Nuber, J. Weiss, G. C. Papaefthymiou, R. B. Frankel, S. J. Lippard, J.
Am. Chem. Soc. 1987, 109, 7387.
[37] R. Hotzelmann, K. Wieghardt, U. Flörke, H.-J. Haupt, Angew. Chem.
1990, 102, 720; Angew. Chem. Int. Ed. Engl. 1990, 129, 645.
[38] C. Flassbeck, K. Wieghardt, Z. Anorg. Allg. Chem. 1992, 608, 60.
[39] J. G. Dash, D. P. Johnson, W. M. Visscher, Phys. Rev. 1968, 168, 1087.
[40] R. M. Housley, F. Hess, Phys. Rev. 1966, 146, 517.
[41] K. Wieghardt, S. Drüecke, P. Chaudhuri, Z. Naturforsch. B 1989, 44,
1093.
[42] Y. G. Galyametdinov, O. N. Kadkin, A. V. Prosvirin, Russ. Chem. Bull.
1994, 43, 887.
[43] U. Bossek, T. Weyermüller, K. Wieghardt, B. Nuber, J. Weiss, J. Am.
Chem. Soc. 1990, 112, 6387.
[1] A.-M. Giroud-Godquin, P. M. Maitlis, Angew. Chem. 1991, 103, 370,
Angew. Chem. Int.Ed. Engl. 1991, 30, 375.
[2] P. Espinet, M. A. Esteruelas, L. A. Oro, J. L. Serrano, E. Sola, Coord.
Chem. Rev. 1992, 117, 215.
[3] S. A. Hudson, P. M. Maitlis, Chem. Rev. 1992, 93, 861.
[4] L. Oriol, J. L. Serrano, Adv. Mater. 1995, 7, 348.
[5] Metallomesogens (Ed.: J. L. Serrano), VCH, Weinheim, 1996.
[6] G. Lattermann, S. Schmidt, R. Kleppinger, J. H. Wendorff, Adv. Mater.
1992, 4, 30.
[7] S. Schmidt, G. Lattermann, R. Kleppinger, J. H. Wendorff, Liq. Cryst.
1994, 16, 693.
[8] H. Zheng, T. M. Swager, J. Am. Chem. Soc. 1994, 116, 761.
[9] D. W. Bruce, Adv. Mater. 1994, 6, 699.
[10] D. W. Bruce, X.-H. Liu, J. Chem. Soc. Chem. Commun. 1994, 729.
[11] S. Morrone, G. Harrison, D. W. Bruce, Adv. Mater. 1995, 7, 665.
[12] D. W. Bruce, X.-H. Liu, Liq. Cryst. 1995, 18, 165.
[13] T. M. Swager, H. Zeng, Mol. Cryst. Liq. Cryst. 1995, 260, 301.
[14] U. Stebani, G. Lattermann, M. Wittenberg, J. H. Wendorff, Angew.
Chem. Int. Ed. Engl. 1996, 35, 1858.
[44] A. E. Comyns, Nature, 1994, 369, 609.
[15] Y. G. Galyametdinov, G. I. Ivanova, I. V. Ovchinikov, Izv. Akad.
Nauk. SSSR Ser. Khim. 1989, 1931.
[45] R. Hage, J. E. Iburg, J. Kerschner, J. H. Koek, E. L. M. Lempers, R. J.
Martens, U. S. Racherla, S. W. Russel, T. Swarthoff, M. R. P. van Vliet,
J. B. Warnaar, L. van der Wolf, B. Krijnen, Nature, 1994, 329, 637.
Ã
[16] L. Ziminski, J. Malthete, J. Chem. Soc. Chem. Commun. 1990, 1495.
Ã
[17] P. Jacq, J. Malthete, Liq. Cryst. 1996, 21, 291.
Chem. Eur. J. 1998, 4, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0401-0099 $ 17.50+.25/0
99