вход по аккаунту


Ordering Phenomena and Phase Transitions in a Spin-Crossover CompoundЧUncovering the Nature of the Intermediate Phase of [Fe(2-pic)3]Cl2EtOH.

код для вставкиСкачать
Order and Disorder
Ordering Phenomena and Phase Transitions in
a Spin-Crossover Compound—Uncovering
the Nature of the Intermediate Phase of
Dmitry Chernyshov, Marc Hostettler, Karl W. Trnroos,
and Hans-Beat Brgi*
Many properties of solid materials are intimately related to
their crystal structures, and changes in properties are often
accompanied by phase transitions. Here we show that a
deeper understanding of the various phases of spin-crossover
compounds, their phase transitions and properties, calls for
multiple structure determinations as a function of external
variables; the usual “one phase–one crystal structure” protocol is insufficient.
The spin state of several first-row transition-metal complexes may be switched between high and low under the
influence of an external perturbation such as a change in
temperature or pressure or irradiation with light. The
characteristics of the bi- or multistable compounds, for
example, transition temperatures, hysteresis properties, thermochromism, photochromism, and photomagnetism as well
as thermodynamic and kinetic quantities, depend on the
interplay between the ligand-field strength at the transitionmetal ion and the interactions between the metal complex,
the counterions, and solvate molecules, as governed by the
crystal packing. The properties of spin-crossover complexes,
especially of iron(ii) complexes, have been widely studied
over the last several decades. Many publications stress the
potential—unproven so far—for applications of such compounds in nanotechnology, for example, in information
storage, as sensors, or as molecular switches.[1–4]
In octahedral d6 iron(ii) complexes the conversion is
between a low-spin (LS) and a high-spin (HS) state of
nearly equal energies (LS, S = 0, t2g6eg0 ; HS, S = 2, t2g4eg2). Spin
[*] Prof. H.-B. Brgi, Dr. D. Chernyshov, Dr. M. Hostettler,
Prof. K. W. Trnroos+
Laboratorium fr Kristallographie der Universit(t Bern
Freiestrasse 3, 3012 Bern (Switzerland)
Fax: (+ 41) 31-631-3996
[+] Permanent address:
Department of Chemistry, University of Bergen
Bergen (Norway)
[**] This work was supported by the Swiss National Science Foundation.
We thank the staff of the Swiss Norwegian Beam Lines (SNBL) at
the European Synchrotron Research Facility for their support, Dr. A.
Linden, University of Zrich, and Dr. T. Weber, ETHZ, for help with
initial experiments, and Prof. A. Hauser, University of Geneva, Prof.
G. M. Sheldrick, Universit(t Gttingen, Prof. P. TolAdano, UniversitA
de Picardie, and Prof. V. Dmitriev, SNBL, for advice and discussions.
The Universities of Bern and Bergen provided the means and
opportunity for a sabbatical leave for K.W.T. pic = picolylamine.
Supporting information for this article is available on the WWW
under or from the author.
Angew. Chem. Int. Ed. 2003, 42, 3825 –3830
DOI: 10.1002/anie.200351834
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
crossover is characterized by a transition curve which relates
the relative concentrations of the two spin states to temperature. There are two main types of such curves which show
either a continuous or discontinuous single step, or two such
steps. The three plateaus of a double-step transition are
associated with a HS phase, an intermediate phase (IP), and a
LS phase at elevated, intermediate, and low temperatures,
respectively. The IP is defined by the limiting temperatures
T1 > T2 and is usually found at approximately equal populations of the HS and LS states.
The ordering of spin states in some mononuclear complexes displaying two-step transitions has been investigated
by X-ray diffraction: an IP differing from the high-temperature phase has been described for [FeII{5-NO2-salN(1,4,7,10)}] which appears to show two, rather than one,
iron sites in the asymmetric unit (the ligand is six-coordinating and prepared by Schiff base condensation of 5-NO2salicylaldehyde with 1,4,7,10-tetraazadecane).[5] The crystal
structure of the HS phase of [FeII{HC(3,5-Me2pz)3}2](BF4)2
shows a single iron site at room temperature. Upon cooling,
the compound undergoes a transition to an IP with a
1:1 mixture of HS and LS states on each of two different
metal sites.[6] [FeII{4,4’-bis(1,2,4-triazole)}3](ClO4)2 has two
crystallographically independent iron sites at all temperatures. In the region of the plateau of the transition curve one
of them is occupied by HS complexes, the other by LS complexes.[7] An ordered IP has also been suggested for some
binuclear complexes showing a plateau in their transition
curves.[1, 8, 9] The theoretical approaches hitherto developed to
explain the magnetic behavior of compounds with two-step
transitions assume an Ising-type[10, 11] or Gorski–Bragg–Williams model that has been adapted to binary systems.[12] Both
models account for the ordered IP in terms of two different
iron lattices, one predominantly populated by HS molecules,
the other predominantly by LS molecules.
The compound [FeII(2-pic)3]Cl2·EtOH (1, pic = picolylamine) has been studied extensively during the last 20 years
and has become the prototypical example of a two-step
transition not associated with an ordered IP.[2, 3] Optical[13] and
magnetic data[14] as well as crystallographic unit-cell constants
in the range 108–290 K[15] show a plateau between T1 = 122 K
and T2 = 114 K (Figure 1 a). Complete crystal structures were
earlier determined at 298, 150, and 90 K[16] as well as at 227,
199, 171, 148, and 115 K.[17] All studies, including the one in
the plateau at 115 K, report a single iron site in the
asymmetric unit and no change of crystal symmetry. The
theoretical interpretation of this observation is based on an
Ising-type model with a single iron sublattice. Numerical
analysis of the model by Monte Carlo[18] and cluster variation
methods[19] reproduces the experimental transition curve
qualitatively, and suggests an intermediate state built from
ordered clusters of HS–LS pairs, whose correlation length has
been estimated to be on the order of next-nearest neighbor
distances, thereby excluding long-range ordering.[20]
We have reinvestigated 1 by X-ray diffraction with the aim
of characterizing these short-range correlations. Instead of
the expected diffuse scattering arising from short-range
order—and contrary to all diffraction evidence reported so
far—we have discovered additional Bragg reflections that
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. a) Comparison of spin-transition behavior from diffraction
(circles) and magnetization experiments (line). The HS fraction gHS(T)
was estimated from the observed average FeN bond lengths d(T) as
gHS(T) = 1[(dHSd(T))/(dHSdLS)], with dHS = 2.195 E and
dLS = 2.005 E. These values provide an overall agreement between
gHS(T) from the diffraction and magnetic data. Standard uncertainties
are approximately equal to the symbol sizes. b) Temperature dependence of the average FeN bond lengths hdi.[26] The bond lengths at the
two iron sites differ in the intermediate phase (the lines are guides to
the eye).
disclose an IP between two successive phase transitions at T1
~ 124(1) and T2 ~ 114(1) K.[21] At temperatures higher than T1
the asymmetric unit of the crystal structure comprises one
pair of chloride ions, one ethanol molecule, and a single
independent iron complex, the latter being predominantly,
but not entirely, in the HS state.[16, 17] The unit cell doubles in
size between T1 and T2 ; the resulting superstructure contains
four Cl ions, two EtOH molecules, and two different iron
complexes: one predominantly HS, the other predominantly
LS (Figure 1 b). The third phase appears at temperatures
below T2, its structure reverting to a single iron complex, but
now predominantly in the LS state.[16] Both phase transitions
are of the order-disorder type. The sequence of three phases
may be considered “re-entrant” in the sense that the unit cell
of the HS structure reappears in the LS structure, but differs
from that in the IP. Re-entrant behavior has so far not been
described for spin-crossover compounds. These observations
prompted us to characterize the transitions in 1 by means of
16 crystal structures obtained between 12 and 298 K.
At first glance the structure of the HS phase, crystals of
which are yellow, and that of the LS phase, crystals of which
are red (the same color schemes are used in the figures),
appear to be the same: the unit cell dimensions are similar,
the space groups the same, and all the iron atoms crystallographically equivalent. On closer inspection it becomes
evident that the unit-cell constants and the atomic coordinates change discontinuously with temperature between the
two phases. Figure 2 shows two typical examples: the monoclinic angle b and the x coordinates of the iron atoms. The
intermolecular vibrational frequencies, Debye temperatures,
and GrIneisen parameters are also different in the HS and
LS phases.[22, 23] If the transition between these phases were
not separated by an IP, the discontinuous changes in the
Angew. Chem. Int. Ed. 2003, 42, 3825 –3830
Figure 2. Monoclinic angle b (a) and x-coordinate of the Fe atoms (b)
as a function of temperature. The coordinate of Fe2 in the IP is given
as 1=2 x(Fe2) for easier comparison with x(Fe1). The standard uncertainties are approximately equal to the symbol sizes.
structural parameters would indicate a so-called isostructural
phase transition,[24] which is always of first order.[25]
The average FeN bond lengths of the two crystallographically independent iron complexes in the IP show that
one of the two independent iron atoms is predominantly HS,
the other predominantly LS (Figure 1 b).[26] However, their
positions coincide neither with those found in the HS nor
those in the LS phases, but are displaced significantly from
both (Figure 2 b). The spin ordering, which is characterized by
the difference in the HS populations at the two independent
sites, is zero at both transition temperatures and goes through
a maximum at T1/2 ~ 119 K, the temperature at which the
crystal as a whole contains equal concentrations of the two
spin states. On one site the HS population factor is about 0.85,
on the other it is approximately 0.13, which implies an order
parameter of 0.72.[27] The occurrence of a largely ordered IP
implies that long-range correlations of HS and LS complexes
are energetically more favorable than the short-range correlations postulated in the literature,[19, 20] even though the
ordering in the IP is not quite perfect.
It is worthwhile pointing out that distinguishing between
the different iron sites in the IP by means of optic, magnetic,
and MKssbauer data is very difficult, if not impossible. It is
trivial, however, from multitemperature structure analyses
(Figure 1). Furthermore, the characterization of the nature of
the phase transitions required the determination of crystal
structures over as large a range of temperatures as possible
(Figure 2).
In all three phases the crystal structures are built from
layers of hydrogen-bonded cationic complexes, chloride
anions, and alcohol molecules (Figure 3, left). The layers are
stacked through hydrophobic contacts between the hydrocarbon parts of the ligand and alcohol molecules (Figure 3,
right). The oxygen atom of the ethanol molecule was found to
be disordered over two sites between 298 and 100 K;
occupation of the major site varies from about 0.73 to 0.94
between 298 and 125 K and from about 0.85 to 1.0 below
115 K. The occupations of the two major sites by the alcohol
molecule in the IP are different: about 0.78 and 0.95.[28]
The two successive order-disorder transformations coincide with the two steps in the spin-transition curve (Figure 1)
and with maxima in the specific heat function cp(T),[29] with
the latter being typical for the latent heat and entropy changes
associated with first-order phase transitions. The structural
results presented here enable the HS/LS mixing entropy to be
calculated as a function of temperature from the spin
populations,[10, 12] which in turn are derived from the observed
average FeN bond lengths. The mixing entropy amounts to
Figure 3. Left: projection of one layer of the IP down the c-axis showing hydrogen bonds (dotted lines). Right: projection of the structure of the IP
down the b-axis showing the hydrophobic contacts at c/2 between layers of Fe complexes stacked along the c-axis. Only complexes between 0 and
b/2 are shown. Black: C and H, blue: N, red: O, green: Cl. The illustrations are based on the structure determined at 115 K and 50 % probability
ellipsoids. The two different iron, chlorine, and ethanol sites become indistinguishable in the HS and LS phases (a-glide operation in the left picture, translation (a + c)/2 in the right picture).
Angew. Chem. Int. Ed. 2003, 42, 3825 –3830
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
approximately 4 J K1 mol1 at most. The entropy corresponding to the disorder of the oxygen atoms of EtOH is of similar
magnitude to that of the spin-related one.[30] The entropies
determined from the residual disorder in the structures
together with those from spin multiplicities help to narrow
down other contributions to the total entropy, for example,
from phonons.
With the detailed experimental information reported
here, models describing spin-crossover behavior, especially
the occurrence of a plateau in the spin transition curve of 1
and related compounds may be assessed with respect to their
ability for predicting transition curves, structural changes, and
phase transitions as well as for interpreting the intermolecular—that is, supramolecular—interactions driving the transitions. All models include the notion of spatial correlations
between the states of different molecules, but differ in their
respective correlation lengths. Multitemperature diffraction
experiments are well-suited to probe such differences:
absence of correlation leads to a nearly homogenous background; short-range correlations extending over a few unit
cells produce modulated diffuse intensities; correlations at
the mesoscopic scale broaden the Bragg reflections whereas
long-range order produces sharp reflections.
Even after careful processing of the diffraction patterns of
freshly grown single crystals of 1, no diffuse scattering could
be discerned at any temperature, not even close to the phasetransition temperatures where disorder is most pronounced
(Figure 1). The lack of observable diffuse scattering and the
extra Bragg reflections in the IP suggest that the spin
transition does not induce short-range order of any importance, but rather the long-range order characteristic of the IP.
The Ising-type, HS/LS models derived from a Gorski–
Bragg–Williams or Slichter–Drickamer approach,[31, 32]
assume a single crystallographic iron site in a mean environment (mean field) and neglect correlation effects. These
models allow for continuous or discontinuous spin crossover
and are useful for describing the HS and LS parts of the
transition curve of 1 and for rationalizing the absence of
correlated diffuse scattering. They exclude the occurrence of
an intermediate phase and cannot explain the formation of
the plateau. Models abandoning the mean-field approximation interpret the plateau in the transition curve as an
“intermediate phase” built from correlated HS/LS pairs with
a short correlation length extending to next-nearest neighbor
distances.[18, 19] Since no modulated diffuse scattering has been
detected, these models do not describe the behavior of 1, but
may well be applicable to other, as yet unidentified, compounds.
More general mean-field models with appropriately
chosen numerical parameters account for disordered HS
and LS phases as well as for ordered IPs with two crystallographically independent iron sites.[10–12] They predict additional Bragg reflections in the IP, as observed in this study of
1. The maximum degree of order of 0.72 found in our
experiment is somewhat higher than those calculated from
two different two-sublattice models (0.65 from an Ising-type
model,[11] or ca. 0.62 from a Gorski–Bragg–Williams model
modified for binary mixtures[12]).
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The parameters of Ising-like models—namely the energy
difference and the interaction energy between HS and
LS molecules—appear simple and physically transparent.
However, interpretation of these parameters on a microscopic
level is very difficult, because they do not explicitly take into
account structural changes associated with the ordering of
spin states (atomic shifts, ordering of solvent, variation of size
and shape of unit cell), even though these changes do
contribute to both transition entropy and enthalpy.
A convenient alternative to analyze phase transitions is
Landau theory, which explicitly accounts for changes of
crystal symmetry and the coupling of ordering and spintransition processes. A Landau-type approach using a totally
symmetric quantity as an order parameter, for example, the
HS fraction gHS or associated scalar quantities such as metal–
ligand distances, has been developed for isostructural phase
transitions.[25] A discussion of the application of this approach
to two-step crossover systems will be given elsewhere.
In summary, we have shown that the structure of [Fe(2pic)3]Cl2·EtOH undergoes two first-order phase transitions on
cooling from a high-spin phase via an intermediate phase to a
low-spin phase. Comparisons of structural data over large
ranges of temperature clearly show that the structural
changes from the HS to the LS phase, both with a single
iron site, are discontinuous. The relationship between the
phases is isostructural. The IP reveals that, contrary to earlier
assertions, the intermediate plateau in the two-step transition
of this compound is the expression of a structure with two
different iron sites showing long-range order, very much like
that found for other compounds with two-step behavior. The
absence of discernible diffuse scattering suggests that shortrange correlations are very weak. The combination of the
isostructural HS and LS phases with the ordered IP implies
re-entrant phase-transition behavior, a relatively rare phenomenon.
The results presented here, probably the most detailed
and accurate structural description of an order-disorder phase
transition in a spin-crossover compound hitherto published,
will serve as a basis for developing and testing new models of
the coupling between spin-crossover processes and orderdisorder phase transitions. Such models will help to better
understand the many physical properties relevant for practical applications, for example, elastic, optical, and magnetic
properties, the latent heat and entropy changes associated
with the spin transition, as well as lattice and molecular
vibration. Thus, the completeness of the new data not only
corrects an oversight, but leads to new conclusions concerning
the phase diagram of [Fe(2-pic)3]Cl2·EtOH. Since the useful
properties of new materials often depend on changes in
crystal structure, phase transitions in particular, the approach
outlined here for a molecular spin-crossover compound
should be useful for other molecular materials as well.
Experimental Section
Magnetic susceptibility: Measurements between 2 and 298 K were
performed with a SQUID MPMS-XL5 (Quantum Design) instrument
on a freshly made[33] microcrystalline sample (13.3 mg) in fieldcooling mode (1003 Oe). The measured susceptibilities were cor-
Angew. Chem. Int. Ed. 2003, 42, 3825 –3830
rected for diamagnetic contributions,[14] and the paramagnetic susceptibilities cHT(T) and cLT(T) at high (230–300 K, HT) and low
temperatures (2–5 K, LT) described in terms of Curie and Weiss
constants (see Supporting Information). The high-spin fraction gHS(T)
follows from c(T) = gHS(T) cHT(T) + cLT(T) (5–230 K). The temperature dependence of gHS(T) agrees very well with that from our
diffraction data and well with published values,[14] except for a shift of
our T1 and T2 values to higher temperatures by about 1.5 K. This shift
is essentially attributed to the larger temperature range used for
modeling the cLT(T) function (2–300 versus 80–300 K in ref. [14]).
X-ray data: Single-crystal data were collected on the SNBL beam
line BM1A at the ESRF synchrotron in Grenoble (France) with a
MAR345 image-plate area detector at l = 0.7000(1) O, with f scans
of 1.5–2.08, and a readout pixel resolution of 150 mm. These measurements have been done on five different crystals: No. 1 at 12 K, no. 2 at
50 K, no. 3 at 100, 111, 113, 115, 117, 119, 121, 123, and 125 K, no. 4 at
130, and 200 K, and no. 5 at 165 K. Samples were cooled with N2 and
He cryostats (Oxford Cryostream series 600 and Oxford Diffraction
HeliJet open-flow, respectively). Intensities were integrated with
CrysAlis,[34] and reflections shadowed by the cryostats were removed
from the data by means of locally written software. Additional data at
143 and 298 K were collected on a sixth crystal with a SMART 1 K
CCD diffractometer by means of 0.38 w scans with MoKa radiation
(l = 0.71073 O) using SMART and integrated with SAINT.[35] Empirical absorption correction was made with SADABS[36] and the
structures refined with SHELXL97.[37] The standard uncertainties of
the cell constants reflect their precision and internal consistency for a
given crystal (Table 1). Their accuracy is about an order of magnitude
tional HS/LS unit cell with space group P21/c to that with B21/c is
[001] [010] [¯20̄1].
Further crystallographic data can be found in the Supporting
Information. CCDC-209849–209864 contain the supplementary crystallographic data for this paper. These data can be obtained free of
charge via (or from the
Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: (+ 44) 1223-336-033; or
Received: May 7, 2003 [Z51834]
Published online: July 28, 2003
Keywords: iron · low-temperature crystallography · N ligands ·
order–disorder phase transition · spin crossover
[1] E. KKnig, Struct. Bonding (Berlin) 1991, 76, 51.
[2] P. GItlich, A. Hauser, H. Spiering, Angew. Chem. 1994, 106,
2109; Angew. Chem. Int. Ed. Engl. 1994, 33, 2024.
[3] P. GItlich, Y. Garcia, H. Spiering, in Magnetism: Molecules to
Materials IV (Eds.: J. S. Miller, M. Drillon), Wiley-VCH, 2002.
[4] J. A. Real, A. B. Gaspar, V. Niel, M. C. Munoz, Coord. Chem.
Rev. 2003, 236, 121.
[5] D. Boinnard, A. Bousseksou, A. Dworkin, J. M. Savariault, F.
Varret, J.-P. Tuchagues, Inorg. Chem. 1994, 33, 271; 5-NO2-salN(1,4,7,10) = 1,10-bis(5-nitrosalicylaldehyde)-1,4,7,10-tetra-azadecane-O,O’,N,N’,N’’,N’’’.
[6] D. L. Reger, C. A. Little, V. G.
Young, Jr., M. Pink, Inorg. Chem.
2001, 40, 2870; HC(3,5-Me2pz)3 =
Table 1: Cell constants (a, b, c, b), cell volume (V), and measurement temperature (T) with crystal phase
tris(3,5-dimethylpyrazolyl)meof 1 (HS = high spin, LS = low spin, both with space group B21/c, IP = intermediate phase with space
group P21/c).
[7] Y. Garcia, O. Kahn, L. Rabardel,
T [K]
a [E]
b [E]
c [E]
b [8]
V [E ]
B. Chansou, L. Salmon, J.-P.
Tuchagues, Inorg. Chem. 1999, 38,
12 (LS)
50 (LS)
[8] V. Ksenofontov, H. Spiering, S.
100 (LS)
Reiman, Y. Garcia, A. B. Gaspar,
111 (LS)
N. Moliner, J. A. Real, P. GItlich,
113 (LS)
Chem. Phys. Lett. 2001, 348, 381.
115 (IP)
[9] V. Ksenofontov, A. B. Gaspar,
117 (IP)
J. A. Real, P. GItlich, J. Phys.
119 (IP)
Chem. B 2001, 105, 12 266.
121 (IP)
[10] H. Bolvin, Chem. Phys. 1996, 211,
123 (IP)
125 (HS)
[11] A. Bousseksou, J. Nasser, J.
130 (HS)
Linares, K. Boukheddaden, F.
143 (HS)
Varret, J. Phys. I 1992, 2, 1381.
165 (HS)
[12] A. B. Koudriavtsev, Chem. Phys.
200 (HS)
1999, 241, 109.
298 (HS)
[13] H. Romstedt, A. Hauser, H. Spiering, J. Phys. Chem. Solids 1998, 59,
lower, because of the difficulties in determining the crystal-to-image
[14] R. Jakobi, H. Spiering, P. GItlich, J. Phys. Chem. Solids 1992, 53,
plate distance accurately. The disordered oxygen atoms of the ethanol
[15] L. Wiehl, G. Kiel, C. P. KKhler, H. Spiering, P. GItlich, Inorg.
molecules were modeled with two positions. All non-hydrogen atoms
Chem. 1986, 25, 1565.
except for the minor disorder component of the alcohol oxygen atom
Mikami, M. Konno, Y. Saito, Acta Crystallogr. Sect. B 1980,
were refined anisotropically. Hydrogen atoms were input as riding
36, 275.
atoms and the isotropic displacement parameters of these fixed at
[17] B. A. Katz, C. E. Strouse, J. Am. Chem. Soc. 1979, 101, 6214.
1.2 Ueq(CH, CH2) and 1.5 Ueq(CH3, OH) of their respective parent
[18] T. Kohlhaas, H. Spiering, P. GItlich, Z. Phys. B 1997, 102, 455.
atoms. The hydrogen atom of the minor disorder component was not
[19] H. Romstedt, H. Spiering, P. GItlich, J. Phys. Chem. Solids 1998,
accounted for.
59, 1353.
To simplify the comparison between the HS/LS phases and the IP
[20] H. Spiering, T. Kohlhaas, H. Romstedt, A. Hauser, C. Brunthe nonconventional monoclinic space group description B21/c was
Yilmaz, J. Kusz, P. GItlich, Coord. Chem. Rev. 1999, 190–192,
chosen for the former. The transformation matrix from the convenAngew. Chem. Int. Ed. 2003, 42, 3825 –3830
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[21] These temperatures are redeterminations from a freshly prepared sample of 1. They agree to within less than 0.5 K with our
SQUID results (see Experimental Section).
[22] P. GItlich, H. KKppen, R. Link, H. G. SteinhUuser, J. Chem.
Phys. 1979, 70, 3977.
[23] E. Meissner, H. KKppen, C. P. KKhler, H. Spiering, P. GItlich,
Hyperfine Interact. 1987, 36, 1.
[24] The commonly used term “isostructural” phase transition does
not imply “equal structures”, but rather equal space groups and
equal Wyckoff positions. The unit-cell volumes and atomic
coordinates are generally different and their changes with
temperature discontinuous, as expected for a first-order phase
[25] P. TolVdano, V. Dmitriev, Reconstructive Phase Transitions in
Crystals and Quasicrystals, World Scientific, Singapore, 1996.
[26] Occupation of the Fe sites by HS and LS molecules implies
positional disorder amounting to about 0.2 O. As the resolution
of the diffraction experiments is only approximately 0.8 O the
disorder can not be resolved, but is modeled in terms of an
average between HS and LS positions weighted with the
respective population factors and a disorder contribution to
the atomic displacement parameters (ADP). Mean positions
imply that the experimentally determined FeN bond lengths
reported in Figure 1 b are to be interpreted as averages between
a long (HS) and a short (LS) value. Disorder contributions have
2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
been gauged from the differences in ADPs along internuclear
vectors, which tend to be largest where the mean distances
deviate most from the LS and HS values (K. Chandrasekhar,
H. B. BIrgi, Acta Crystallogr. Sect. B 1984, 40, 387).
The order parameter is defined as the difference between the
two populations.
Where comparable our results are close to those published
K. Kaji and M. Sorai, Thermochim. Acta 1985, 88, 185.
The calculation of entropies arising from spin disorder is given in
ref. [10] for oxygen disorder S/R = p ln p + (1p) ln (1p),
where p is the occupation factor of the major site.
C. P. Slichter, H. G. Drickamer, J. Chem. Phys. 1972, 56, 2142.
R. Boča, W. Linert, Monatsh. Chem. 2003, 134, 199.
M. Sorai, J. Ensling, P. GItlich, Chem. Phys. 1976, 18, 199.
CrysAlis Software System, Version 1.169, Oxford-diffraction
Ltd. Oxford, England, 2001.
SMART, Version 5.059; SAINT, Version 6.02a Bruker AXS
Inc., Madison, Wisconsin, USA, 2001.
G. M. Sheldrick, SADABS, Version 2.06, Empirical Absorption
Correction Program, University of GKttingen, GKttingen (Germany), 2002.
G. M. Sheldrick, SHELXL97, University of GKttingen, GKttingen (Germany), 1997.
Angew. Chem. Int. Ed. 2003, 42, 3825 –3830
Без категории
Размер файла
187 Кб
spina, nature, phenomena, ordering, intermediate, pic, crossover, transitional, compoundчuncovering, phase, cl2etoh
Пожаловаться на содержимое документа