ARTICLE IN PRESS
Journal of Solid State Chemistry 181 (2008) 1472– 1479
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Thermal expansion of Cr2xFe22xMo3O12, Al2xFe22xMo3O12 and
Al2xCr22xMo3O12 solid solutions
M. Ari a, P.M. Jardim a, B.A. Marinkovic a,, F. Rizzo a, F.F. Ferreira b
a
b
Departamento de Ciência dos Materiais e Metalurgia, Pontifı́cia Universidade Católica de Rio de Janeiro—PUC-Rio, Rua Marquês de São Vicente 225, Rio de Janeiro, RJ, Brazil
Laboratório Nacional de Luz Sı́ncrotron (LNLS), CP 6192, CEP 13083-970, Campinas, SP, Brazil
a r t i c l e in f o
a b s t r a c t
Article history:
Received 23 November 2007
Received in revised form
27 February 2008
Accepted 15 March 2008
Available online 27 March 2008
The transition temperature from monoclinic to orthorhombic and the thermal expansion of the
orthorhombic phase were investigated for three systems of the family A2M3O12: Cr2xFe22xMo3O12,
Al2xFe22xMo3O12 and Al2xCr22xMo3O12. It was possible to obtain a single-phase solid solution in all
studied samples (x ¼ 0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1).
A linear relationship between the transition temperature and the fraction of A3+ cations (x) was
observed for each system. In all orthorhombic solid solutions studied here the observed thermal
expansion was anisotropic. These anisotropic thermal expansion properties of crystallographic axes a, b
and c result in a low positive or near-zero overall linear coefficient of thermal expansion (al ¼ aV/3). The
relationship between the size of A3+ cations in A2M3O12 and the coefficient of thermal expansion is
discussed. Near-zero thermal expansion of Cr2Mo3O12 is explained by the behavior of Cr–O and Mo–O
bond distances, Cr–Mo non-bond distances and Cr–O–Mo bond angles with increasing temperature,
estimated by Rietveld analysis of synchrotron X-ray powder diffraction data.
& 2008 Elsevier Inc. All rights reserved.
Keywords:
X-ray diffraction
Negative thermal expansion
Crystal structure
Molybdates
1. Introduction
There are several studies about molybdates and tungstates of
A2M3O12 family reporting the phenomenon of negative thermal
expansion (NTE) after phase transition from monoclinic to orthorhombic structure. Both structures consist of corner-sharing AO6
octahedra and MO4 tetrahedra. It appears that a transverse thermal
motion of oxygen in the A–O–M linkage within lower-density
orthorhombic structure produces NTE [1–4]. These structures
present large chemical flexibility, which means that A3+ cation can
be a transition metal or rare earth that accepts octahedral position,
while M6+ is W6+ or Mo6+ [5,6]. This feature gives many options for
cation substitution and recent literature reports several studies on
the thermal expansion properties of tungstates and molybdates of
A2M3O12 family and their solid solutions [7–11]. Therefore, the
extensive chemical flexibility makes this family rather promising for
tailoring of new materials (solid solutions) with controllable
coefficient of thermal expansion.
The A2M3O12 family presents anisotropic thermal expansion and
probably due to this reason there is currently a high discrepancy
between the measurements of thermal expansion coefficients based
on dilatometric and diffraction methods. Mary and Sleight [12]
explained this difference by the formation of microcracks during
Corresponding author. Fax: +51 21 35271248.
E-mail address: bojan@puc-rio.br (B.A. Marinkovic).
0022-4596/$ - see front matter & 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2008.03.015
specimen cooling and their elimination on heating, giving an extrinsic
NTE contribution to the bulk thermal expansion. For example, the
thermal expansion coefficient for Al2W3O12 was reported as
+2.2 106/1C based on X-ray diffraction (XRD) data [6], while
3.0 106/1C was measured by dilatometry [12]. Sivasubramanian
et al. [13] corroborated this explanation by showing that after cooling,
In2W3O12 bulk specimen had enlarged its volume to about 0.1%, what
was attributed to the formation of microcracks on cooling.
It seems from the literature that there is a correlation between
cationic radii (A3+) and thermal expansion coefficient of orthorhombic phases in A2M3O12 family. This correlation is manifested
in a way that the members of A2M3O12 family with larger A3+
cations demonstrate more negative overall linear thermal expansion coefficient (al ¼ aV/3) than the ones with smaller A3+ [14]. In
terms of axial thermal expansion, the phases with small A3+, such
as Al3+, generally have two crystallographic axes with small
negative or near-zero thermal expansion, while the third one
shows large positive expansion [15]. On the other hand, the
phases with large A3+, such as Y2W3O12 [15,16] and Y2Mo3O12
[17,18], possess NTE along all three crystallographic axes. A
classical rationalization of this correlation has been given by
Forster et al. [14] in terms of small distortions of polyhedra
(octahedra and tetrahedra) required for appearance of NTE in this
structure. In the case of small (A3+) cations, oxygen–oxygen
repulsion is enhanced resulting in more rigid polyhedra and
therefore these compounds usually do not show overall negative
linear thermal expansion. On the other hand, the oxygen–oxygen
ARTICLE IN PRESS
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
repulsion is diminished in octahedra with large A3+ cation
resulting in a possibility of forming more distorted polyhedra on
heating, allowing A2M3O12 compounds with large A3+ cation to
exhibit strong NTE in all crystallographic directions [14,15].
A relationship between phase transition temperature in
A2M3O12 family (from monoclinic P21/a to orthorhombic Pbcn)
and the cation (A3+) electronegativity was proposed by Evans et al.
[6]. It was suggested that as the electronegativity of the A3+ cation
increases, the effective negative charge of oxygen anions
decreases. Thus, in these cases the oxygen– oxygen repulsion
decreases, while oxygen– oxygen attractive forces cause the
transition to higher density monoclinic phase to occur at higher
temperatures than in the phases with less electronegative A3+
cations.
In order to shed more light on the effect of the cation
substitution on the thermal expansion coefficient and the temperature of phase transition from monoclinic to orthorhombic
phase in the A2M3O12 family, we studied through X-ray powder
diffraction (XRPD) and thermal methods the following solid
solutions: Cr2xFe22xMo3O12, Al2xCr22xMo3O12 and Al2xFe22x
Mo3O12 (x ¼ 0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1). As far as the authors
are aware of, this is the first time that the thermal expansion
properties of these solid solutions are reported. However, the
pure orthorhombic Al2Mo3O12, Cr2Mo3O12 and Fe2Mo3O12 were
already studied through dilatometry and differential scanning
calorimetry (DSC) by Tyagi et al. [19], being reported that all
three phases have NTE. In the present work, thermal expansion
coefficients are calculated for orthorhombic phases from highresolution XRPD data, using a synchrotron radiation facility
and it was demonstrated that these solid solutions, including
the three pure phases, have near-zero or low positive overall
linear thermal expansion, but not NTE as claimed by Tyagi et al.
[19]. This difference has been addressed to the influence of
the extrinsic contributions to thermal expansion recorded by
dilatometry.
The general relationship between the size of cation A3+ and
coefficient of thermal expansion of A2M3O12 phases with
orthorhombic structure has been also discussed. Near-zero
thermal expansion of orthorhombic Cr2M3O12 was discussed
based on the behavior of Cr–O and Mo–O bond distances, Cr–Mo
non-bond distances and Cr–O–Mo bond angles.
1473
Rh-coated ultra low expansion glass mirror placed before the
monochromator, which also provides filtering of high-energy
photons (third- and higher-order harmonics). A vertically focused
beam was used in the experiments in a spot of 1.0 103 m
(vertical) 2 103 m (horizontal) into the sample position. The
experiments were performed in the vertical scattering plane, i.e.,
perpendicular to the linear polarization of the incident photons.
Wavelength and the zero point were determined from 11 welldefined reflections of the SRM640c silicon standard. The diffracted
beam was analyzed with a Ge(111) crystal analyzer and detected
with a Na(Tl)I scintillation counter with a pulse-height discriminator in the counting chain. The incoming beam was also
monitored by a scintillation counter for normalization of the
decay of the primary beam.
The powder sample was measured in flat-plate geometry and
data were recorded at different temperatures for 0.5 s at each 2y
in steps of 0.0041 from 101 to 601.
A calibration curve for the furnace (see Supplementary
information) was obtained using a NIST Si sample and its cell
parameter variation with temperature, provided by Yim and Paff
[21]. This calibration curve was additionally verified and validated
following the phase transition monoclinic to orthorhombic for
different A2Mo3O12 compounds and comparing it with DSC results
(see Supplementary information).
A typical high-resolution XRPD pattern of the sample
Cr2Mo3O12 recorded at D10B-XPD beamline showing experimental, calculated and difference profile after Rietveld refinement, is
illustrated in Fig. 1. The unit-cell and crystal structure parameters
were refined from high-resolution diffraction patterns, through
the Le Bail and Rietveld methods, using the program TopasAcademic and considering the space group Pbcn for all orthorhombic phases. As the starting models for Rietveld refinements of
investigated solid solutions, orthorhombic Al2Mo3O12 and Fe2
Mo3O12 crystal structure descriptions from the ICSD files 80448
and 80449 were used. In the lack of a crystal structure description
for the orthorhombic Cr2Mo3O12, the crystal structure of
Fe2Mo3O12 (ICSD 80449) was used as the starting model for
Cr2Mo3O12. The refined crystal structures of orthorhombic
Cr2Mo3O12 at 420, 550, 650 and 740 1C have been deposited in
the Inorganic Crystal Structure Database (ICSD) receiving numbers
418845, 418846, 418847 and 418848. Details of the Rietveld
refinement are included in the Supplementary information.
2. Experimental
6000
Cr2Mo3O12 - 420°C
5000
4000
Intensity (a.u.)
Cr2xFe22xMo3O12, Al2xCr22xMo3O12 and Al2xFe22xMo3O12
solid solutions, (x ¼ 0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1) were produced
by solid-state reaction from a stoichiometric mixture of Cr2O3
(Vetec, 99.5%), Al2O3 (ALCOA, 99.5%), Fe2O3 (Vetec, 99.5%) and
MoO3 (Fluka, 99.9%) powders, respectively. The reactants were
preheated at 500 1C for 2 h, then weighed, activated mechanically
by ball milling for 10 h and pressed at 250 MPa. The pellets were
heated in alumina crucibles at 670 1C for 30 h and cooled down in
the furnace.
XRPD patterns at room temperature were collected using a
Siemens D-5000 diffractometer equipped with Cu sealed tube and
graphite monochromator installed after the sample. The stepscanning mode was 0.021/10 s. High-temperature XRPD (HTXRPD)
of the samples were collected at four different temperatures
always above the phase transition temperature from monoclinic
to orthorhombic structure at the XRPD (D10B-XPD) beamline [20]
of the Brazilian Synchrotron Light Laboratory (LNLS), placed after
a dipolar source. X-rays of wavelength 1.3777(3) Å were selected
by a double-bounce Si(111) monochromator, with water-refrigeration in the first crystal, while the second one is bent for sagittal
focusing. The beam is vertically focused or collimated by a bent
3000
2000
1000
0
10
15
20
25
30
35 40
2 Theta
45
50
55
Fig. 1. High-resolution XRPD pattern for Cr2Mo3O12 at 420 1C. Experimental profile
(dot line), calculated and difference profiles (full lines).
ARTICLE IN PRESS
1474
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
Thermogravimetric (TG) analysis of Al2Mo3O12, Cr2Mo3O12 and
Fe2Mo3O12 were performed in air from room temperature up to
800 1C using a SETARAM thermal analyzer. The heating rate was
10 1C/min. The DSC analyses of all samples were performed in a
Perkin-Elmer DSC-7 with a cycle including heating to 600 1C and
cooling down to room temperature under argon atmosphere. The
heating and cooling rates were 10 1C/min. The phase transition
temperatures were determined using the PYRES software.
A scanning electron microscope (SEM) Zeiss DSM 960 equipped
with an X-ray energy dispersive spectrometer was used in order to
verify the phase composition of the synthesized samples.
3. Results and discussion
3.1. Monoclinic solid solutions
All obtained samples for Cr2xFe22xMo3O12, Al2xCr22xMo3O12
and Al2xFe22xMo3O12 solid solutions, appeared to be singlephased, as confirmed by room temperature XRPD and SEM. In
accordance to that, it can be inferred that Fe3+ (0.645 Å), Cr3+
(0.615 Å) and Al3+ (0.535 Å) [22] can be totally substituted among
themselves in A2Mo3O12 structure due to the relatively small
differences in their cationic radii.
The room temperature XRPD patterns of Cr2xFe22xMo3O12,
Al2xCr22xMo3O12 and Al2xFe22xMo3O12 indicate that all solid
solutions and pure phases appear in the monoclinic structure
(P21/a). The variations of lattice parameters of Cr2xFe22xMo3O12,
Al2xCr22xMo3O12 and Al2xFe22xMo3O12 solid solutions as a
function of fraction (x) are shown in Fig. 2 and Supplementary
information. It is observed that crystallographic axes a, b, c and
volume V increase linearly with increasing of A3+ cation size and
its content in the compound, while the b angle varies inversely.
This dependence is in good agreement with the Vegard’s law. A
very similar trend was observed for Al2xScxW3O12 solid solution
in the orthorhombic structure [23].
3.2. Phase transition and hygroscopicity
The transition temperatures from monoclinic to orthorhombic
structure for Al2Mo3O12, Cr2Mo3O12 and Fe2Mo3O12 phases were
reported [19] to be in the regions of 200–250, 350–400 and
500–550 1C, respectively, and this is in accordance with our
findings (see Table 1 and Supplementary information). In the
present work, however, the transition temperatures for the
solid solutions: Cr2xFe22xMo3O12, Al2xCr22xMo3O12 and Al2x
Fe22xMo3O12 (Table 1 and Supplementary information) are
reported as well. The DSC curves on heating show an endothermic
peak for all specimens corresponding to the phase transition from
monoclinic to orthorhombic. A linear relationship between the
transition temperature and the fraction of A3+ cations (x) was
observed for each system.
The relationship between the phase transition temperature
and the electronegativity of A3+ cations, previously suggested by
Evans et al. [6], was confirmed for each of the solid solutions
studied here, considering electronegativity values for Fe3+, Cr3+
and Al3+ (1.83, 1.66 and 1.61 [24], respectively) and the transition
temperature of their solid solutions in A2Mo3O12 family.
Al2Mo3O12, Cr2Mo3O12 and Fe2Mo3O12 phases show a very low
water loss from room temperature to 800 1C of 0.15, 0.09 and
0.25 wt%, respectively. Comparing these with rare-earth molybdates [17,18,25], that have weight losses at level of 6–8 wt% over
the same temperature region, it could be inferred that Al, Cr and
Fe-molybdates are not hydrated. This was an expected feature,
based on the cation size differences between Al3+, Cr3+ and Fe3+ on
one side and rare earths on another, as well as on the fact that the
monoclinic structure in A2M3O12 family is denser than the
orthorhombic one. Namely, the differences in A3+ cation sizes
together with more dense monoclinic structure reflect over the
size of intracrystalline channels of A2Mo3O12 compounds. Therefore, smaller cations and denser structure form channels with
smaller entrance, blocking at this manner water accommodation
within the crystal structure.
3.3. Thermal expansion of orthorhombic solid solutions
The relationship between lattice parameters and temperature
for Cr2xFe22xMo3O12, Al2xCr22xMo3O12 and Al2xFe22xMo3O12
solid solutions follows a linear dependence as can be seen in
Fig. 3 and Supplementary information.
It is important to be observed (Fig. 4) that in all solid solutions
studied here the coefficient of thermal expansion (a) along a
direction is positive, varying between 6.25 and 7.96 106/1C. On
the other hand, the lattice parameter c shows low negative and
near-zero thermal expansion, having a coefficient of thermal
expansion ac between 2.47 and 0.16 106/1C. The thermal
expansion properties of lattice parameter b are even more
complex, being:
low negative to near-zero for Cr2xFe22xMo3O12 solid solution,
varying between 1.78 106 for Cr2Mo3O12 and 0.62
106/1C for Fe2Mo3O12;
Table 1
Phase transition temperatures of Cr2Fe22xMo3O12, Al2Cr22xMo3O12 and Al2
Fe22xMo3O12 solid solutions
Fig. 2. Variation of volume of Cr2Fe22xMo3O12 (m symbols), Al2Cr22xMo3O12
(K symbols) and Al2Fe22xMo3O12 (’ symbols) in monoclinic structure at room
temperature as a function of fraction (x). Standard uncertainties are included but
are too small to appear in the graphs.
Cr2xFe22xMo3O12
Al2xCr22xMo3O12
Al2xFe22xMo3O12
x
T (1C)
x
T (1C)
x
T (1C)
0
0.1
0.3
0.5
0.7
0.9
1
512.5
500.3
483.5
465.2
441.2
418.2
403
0
0.1
0.3
0.5
0.7
0.9
1
403.02
374.04
327.79
283.71
244.31
214.4
200.22
0
0.1
0.3
0.5
0.7
0.9
1
512.5
484.1
430.2
368.6
305.3
239.8
200.2
ARTICLE IN PRESS
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
1475
Fig. 3. Variation of volume of (a) Cr2Fe22xMo3O12, (b) Al2Cr22xMo3O12, and (c) Al2Fe22xMo3O12 solid solutions vs. temperature. Standard uncertainties are included but
are too small to appear in the graphs.
near-zero for Al2xCr22xMo3O12 solid solution, varying between
0.63 106 for Cr1.4Al0.6Mo3O12 and 0.71 106/1C for
Al1.4Cr0.6Mo3O12 and
low positive for Al2xFe22xMo3O12 being 2.01 106/1C for
AlFeMo3O12.
These thermal expansion properties of crystallographic axes a, b
and c result in a low positive or near-zero overall linear coefficient
of thermal expansion (al ¼ aV/3) for the studied Cr2xFe22xMo3O12,
Al2xCr22xMo3O12 and Al2xFe22xMo3O12 solid solutions (Fig. 4).
Standard deviations for calculated a values are very small as can
be seen from Fig. 4, which enables reliable comparison among
different solid solutions studied.
Overall linear coefficient of thermal expansion (al) for
Fe2Mo3O12 has been already reported by Cheng et al. [9] as
1.14 106/1C, while Yang et al. [26] reported for Cr2Mo3O12 an
al ¼ 1.61 106/1C, both al being calculated from conventional
XRD data. However, overall linear coefficient of thermal expansion
(al) for Fe2Mo3O12 and Cr2Mo3O12 calculated from our highresolution XRPD patterns in the present work were
al ¼ 1.72 106 and 0.67 106/1C, respectively. The observed
discrepancies can be attributed to the following contributions.
The first one can be errors in the temperature calibration of the
furnace. Secondly, it can be the difference in the resolution of
synchrotron and conventional XRD patterns. This can be proved
by simulation of high- and low-resolution XRD patterns of a same
A2Mo3O12 crystal structure, where the standard uncertainties of
the cell parameters calculated by Rietveld refinement are higher
for the low-resolution pattern and also different cell parameter
values were obtained for each case. The third contributions can be
the manner how coefficient of thermal expansion is calculated,
since there are different ways to perform this calculation. Here,
we used natural logarithmic plots (ln V versus T, for example) and
linear regression (with R values superior to 99%) to calculate
coefficient of thermal expansion form the inclination. Only in
some cases of a calculated for b-axis R-values were lower than
99%. On the other hand, Tyagi et al. [19] measured the Al2Mo3O12
compound by dilatometry and obtained a low negative linear
thermal expansion (al ¼ 2.83 106/1C), and a very strong
negative linear thermal expansion for Cr2Mo3O12 and Fe2Mo3O12
(al ¼ 9.39 106 and 14.82 106/1C, respectively). Actually, it
can be generalized for the A2M3O12 family (Fig. 5) that there are
large misfits between the linear coefficients of thermal expansion
(CTE) measured by diffraction and dilatometric methods. There
are probably several extrinsic sources causing discrepancy, but by
ARTICLE IN PRESS
1476
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
Fig. 4. Thermal expansion coefficients for (a) Cr2Fe22xMo3O12 solid solution as a function of Cr3+ content, (b) Al2Cr22xMo3O12 solid solution as a function of Al3+ content
and (c) Al2Fe22xMo3O12 solid solution as a function of Al3+ content.
Fig. 5. Overall linear thermal expansion coefficients for molybdates and tungstates
of A2M3O12 family, obtained by X-ray or neutron diffraction and dilatometry
[6–19,28,31,32,34–36]; results reported in this work (K). The full line is a guide for
the eyes.
far the most cited one is the formation of microcracks on cooling
due to the anisotropic thermal expansion of these phases. This
seems to be a rather reasonable explanation, considering that in
the cubic ZrW2O8 family, dilatometric and diffraction measurements of linear CTE are in good agreement [27]. Certainly, in some
cases, another source of discrepancy can be the use of not
sufficiently well sintered pellets for dilatometric measurements.
However, a more systematic study is necessary to fully understand this phenomenon, especially having in mind a potential
application of materials of the A2M3O12 family with tailored
intrinsic coefficient of thermal expansion, where extrinsic contribution to thermal expansion must be minimized.
In Fig. 5, all published overall linear CTE for orthorhombic
A2M3O12 family obtained through diffraction methods (closed
squares ’ for molybdates and closed triangles m for tungstates)
as a function of A3+ cation radii size, together with dilatometric
results, are plotted. The majority of the overall linear CTE, plotted
in Fig. 5, has been calculated for the temperature interval between
room temperature and 800 1C (see Supplementary information for
details). This plot confirms the general trend for overall linear CTE
for orthorhombic A2M3O12 family as dependent on A3+ cation radii
size. However, tailoring intrinsic thermal expansion properties for
ARTICLE IN PRESS
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
A2M3O12 phases and their solid solutions through substitutions of
cation A3+, using the rationalization that larger A3+ cations
provoke more NTE and vice-versa, is not always straightforward
(Fig. 5). Namely, our results indicate that Cr2Mo3O12 and
Fe2Mo3O12 do not exactly follow the established relationship
between overall linear CTE and the size of A3+ cation. This means
that Cr2Mo3O12 has lower overall linear CTE (al ¼ 0.67 106/1C)
than Fe2Mo3O12 (al ¼ 1.72 106/1C), although the size of Cr3+ in
octahedral coordination is smaller (0.615 Å) than the size of Fe3+
(0.645 Å) [22] in the same coordination. Similar discrepancy from
the general relationship was recently noticed for Tm2Mo3O12 and
Tm2W3O12 phases that show much lower NTE (al ¼ 4.03 106
and 3.95 106/1C) [28] than other rare-earth molybdates and
tungstates of the A2M3O12 family (see Fig. 5). As a matter of fact,
some of our results on cationic substitution in the Cr2xFe22x
Mo3O12, Al2xCr22xMo3O12 and Al2xFe22xMo3O12 solid solutions
reveal that overall linear CTE does not always change (Fig. 4)
according to A3+ size effect. As can be seen from Fig. 4a overall
linear coefficient of thermal expansion in Cr2xFe22xMo3O12
changes in a linear way with the increase of the fraction (x), of
larger A3+ cation, although in the opposite manner considering the
cation sizes of Cr3+ and Fe3+, as previously discussed. For the
Al2xCr22xMo3O12 system, a decrease of al with the increase of
fraction (x) of the larger A3+ is observed (Fig. 4b). However, the
Al2xFe22xMo3O12 system shows a more complex dependence
between overall linear coefficient of thermal expansion and the
fraction (x) (Fig. 4c). In this system, AlFeMo3O12 and Al1.4Fe0.6
Mo3O12 phases have more positive al than the pure phases
(Al2Mo3O12 and Fe2Mo3O12). Actually, in recent literature on
thermal expansion in A2M3O12 family there are also few examples
of such unexpected behavior of overall linear coefficient of
thermal expansion as a function of fraction (x) of larger A3+
cations [7,29]. Peng et al. [7] reported for the Y2xNdxW3O12
system that when the fraction (x) of larger cation Nd3+ is increased
from x ¼ 0.1 to 0.4, al became less negative, instead of becoming
more negative. Similar was observed for Y2xSmxW3O12 system
[29]. However, for some other systems such as: Nd2xErxW3O12
[8], Fe2xErxMo3O12 [9], Er2xCexW3O12 [10] and Er2xSmxW3O12
[11], the overall linear coefficient of thermal expansion become
more negative with increasing values of the fraction (x) of larger
cation, as can be expected from the general trend observed in
Fig. 5. Considering our results on thermal expansion of Cr2Mo3O12
and the reports from literature on Y2xNdxW3O12, Y2xSmxW3O12,
Tm2Mo3O12 and Tm2W3O12 [7,28,29], it occurs that in some
systems exists another factor, apart from A3+ cationic size, that can
influence their thermal expansion properties. This another factor
for A2M3O12 system can be suggested based on Chapman et al.
[30] study on the effect of M2+ substitution on the magnitude of
the NTE for a family of Prussian Blue analogs, M2+Pt4+(CN)6 for
M2+ ¼ Mn, Fe, Co, Ni, Cu, Zn, Cd. The authors found that NTE varies
widely with M2+ substitution and explained this variation in
thermal expansion being primarily due to the different bond
strengths of M2+-cyanide binding interaction, i.e., decreased
structural flexibility is associated with stronger M2+–N bonds
resulting in reduced NTE. It can be therefore suggested that the
strength of A3+–O bonds in A2M3O12 also plays a role in the NTE
behavior of these materials. This might explain unexpected
observation that Cr2Mo3O12 shows a lower overall linear CTE
than Fe2Mo3O12.
The changes of crystal structure details of Cr2Mo3O12 orthorhombic phase as a function of temperature (from 420 to 740 1C),
such as bonding distances Cr–O, Mo–O, non-bonding distance
Cr–Mo and angles Cr–O–Mo, were evaluated by the Rietveld
method. Due to the large crystallite sizes of Al2Mo3O12 and
Fe2Mo3O12 it was not possible to perform Rietveld refinements on
the diffraction patterns of these two phases. Table 2 resumes
1477
Table 2
Rietveld reliability factors for the refined diffraction patterns of Cr2Mo3O12
Temperature (1C)
RB (%)
Gof
Rwp (%)
Rexp (%)
420
550
650
740
3.17
3.81
3.69
4.15
1.28
1.30
1.30
1.31
21.25
21.37
21.26
21.38
16.52
16.43
16.36
16.32
Number of data points was 12,500, number of refined parameters was 42 and
number of considered diffraction lines was 225 for all four temperatures.
Rietveld reliability factors for the refined diffraction patterns,
while all details of Rietveld refinement are given in Supplementary information. In some previous detailed neutron diffraction
studies of A2M3O12 phases (Sc2W3O12 [31], Sc2Mo3O12 [32] and
Y2W3O12 [16]) it was generally established that W–O and Mo–O
apparent bond lengths undergo contraction during heating. The
same occurred for Mo–O in Cr2Mo3O12 phase [dav(Mo–O) ¼
1.72494.25 105 T Å, calculated by linear regression], as illustrated in Fig. 6a. This is actually an expected feature considering
that Mo6+–O bonds are extremely strong, having practically a
zero thermal expansion especially when Mo6+ occupies tetrahedral sites [33], and the possibility of existence of transverse
vibration modes of oxygen atoms, as it is the case in A2M3O12
family. In the case of Sc2W3O12, Sc2Mo3O12 and Y2W3O12 the
authors [16,31,32] also found that Y–O and Sc–O (in Sc2W3O12)
bond lengths do not change significantly with increasing
temperatures, while in the case of Sc2Mo3O12, the Sc–O bond
length even underwent an observable apparent contraction
[dav(Sc–O) ¼ 2.09732.7 105 T Å] [32]. This is an interesting
feature because A3+(VI)–O bonds are weaker than the M6+(IV)–O
ones and in a normal (positive) thermal expansion material these
bonds generally possess a linear coefficient of thermal expansion
(a) of around 6–8 106/1C [33]. The absence of this expansion for
A3+(VI)–O in Sc2W3O12, Sc2Mo3O12 and Y2W3O12 verified by
Rietveld refinements is explained by transverse vibration mode
of oxygen that compensates for the expected thermal expansion of
A3+(VI)–O bond lengths [16]. In Cr2Mo3O12, however, the Cr–O
bond length increases [dav(Cr–O) ¼ 1.9688+2.55 105 T Å] as
demonstrated in Fig. 6b. Nevertheless, it has to be considered
that our study was carried out using an X-ray and not a neutron
source, that would be preferable for the high accuracy of
determination of oxygen positions and thus of metal–oxygen
bond lengths. However, through smooth increase of isotropic
temperature factors of Cr3+, Mo6+ and O2 as a function of
temperature (see Supplementary information) it can be assessed
[16,31] that the structural parameters of Cr2Mo3O12 phase (for the
temperature range 420–740 1C) refined through Rietveld method
(see Supplementary information) formed a reliable structural
model with no significant systematic flaws. Another indication of
reliability of the structural model is based on good Rietveld
reliability factors (Table 2) (mean standard uncertainties for Cr–O
bond distances and Cr–Mo non-bond distances calculated by
Rietveld refinements, Fig. 6, are at the same level as observed by
Evans et al. [32] for Sc2Mo3O12 from neutron diffraction data).
Even if we take mean bond length of Cr–O at 740 1C (1.9903 Å) as a
somewhat higher than expected and remove this point from the
plot, it can be verified that the Cr–O mean bond length still
increases with temperature [dav(Cr–O) ¼ 1.9775+7.79 106 T Å]
(Fig. 6b). Therefore, it can be rationalized that transverse
vibrations of oxygen are not strong enough in orthorhombic
Cr2Mo3O12 phase (a near-zero thermal expansion material) to
compensate for Cr–O bond length increase, as it is the case for
Sc2W3O12, Sc2Mo3O12 and Y2W3O12 phases (NTE materials). As a
matter of fact, the mean non-bonding distance Cr–O–Mo is
ARTICLE IN PRESS
1478
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
Fig. 6. Bond distances, non-bond distances and bond angles as a function of temperature for Cr2Mo3O12 phase (a) mean apparent Mo–O bond distances, (b) mean apparent
Cr–O bond distances, (c) Cr–Mo non-bond distances and (d) Cr–O–Mo bond angles.
maintained almost unchanged with increasing temperatures
(Fig. 6c), which corroborates the conclusion that transverse
vibrations of oxygen are not strong enough in orthorhombic
Cr2Mo3O12 phase to provoke a negative overall coefficient of
thermal expansion. A lack of strong fluctuations in mean
Cr–O–Mo bond angles can be confirmed as well, from Fig. 6d
(standard uncertainties of the measured Cr–O–Mo angles are at
the level of 0.251, which is typical for Rietveld refinement results
[31]). However, if one closely inspects each of the Cr–O–Mo angles
it can be concluded that some of them increase and some others
decrease with temperature. Since the crystal structure of A2M3O12
family can be described as built from corner-shared polyhedra
(AO6 and MO4) forming bc planes, being connected themselves
along the a axis just through A–O3–M2 linkages, it can be
assumed that the nature of A–O3–M2 angle variation with
temperature influences thermal expansion along the crystallographic axis a. It occurs that in Cr2Mo3O12 this angle
Cr1–O3–Mo2 increases with temperature (Fig. 6d). Therefore, it
is reasonable that aa is positive (6.09 106/1C). From five other
angles, two of them, being Cr1–O2–Mo1 and Cr1–O5–Mo2,
contract on heating. These are the very same angles that also
contract in Sc2Mo3O12 phase [32]. Actually, according to Evans
and Mary [32] these two angles undergo abrupt increase on
displacive phase transition from monoclinic to orthorhombic
phase and therefore, these will be easily distort in orthorhombic
phase, probably being responsible for NTE along the axes b
(1.78 106/1C) and c (2.47 106/1C) in Cr2Mo3O12.
4. Conclusions
Single-phased solid solutions were obtained for Cr2xFe22x
Mo3O12, Al2xFe22xMo3O12 and Al2xCr22xMo3O12 (x ¼ 0, 0.1, 0.3,
0.5, 0.7, 0.9 and 1). The relationship between cell parameters and
the size of cation A3+ obeys Vegard’s law for all solid solutions. The
phase transition temperature increases with the A3+ cation
electronegativity and there is a linear relationship between the
transition temperature and the fraction of A3+ cations (x).
ARTICLE IN PRESS
M. Ari et al. / Journal of Solid State Chemistry 181 (2008) 1472–1479
The thermal expansion coefficients for the orthorhombic phase
and phase transition temperatures could be generally tailored by
choosing larger or smaller A3+ cations. Nevertheless, our results
showed that in some phases and solid solutions the A3+ cation size
is not the only factor which influences the thermal expansion
coefficient, as it is observed for Cr2Mo3O12 and Fe2Mo3O12.
Near-zero overall coefficient of thermal expansion of orthorhombic Cr2Mo3O12 is tentatively explained in terms of temperature
behavior of Mo–O and Cr–O bond lengths, Cr–O–Mo non-bond
lengths and Cr–O–Mo angles.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
Supplementary information
The atomic positions and isotropic thermal factors of orthorhombic Cr2Mo3O12 at 420, 550, 650 and 740 1C have been deposited
at the Fachinformationzentrum karlsruhe, Abt. PROKA, 76344
Eggenstein-Leopoldshafen, Germany, as supplementary material
ICSD numbers 418845–418848 and can be obtained by contacting
the FIZ quoting the article details or the corresponding ICSD number.
Acknowledgments
The authors are grateful to the Brazilian Synchrotron Light
Laboratory (LNLS) for the beam time and financial support under
the Projects D10B-XPD 5869/06 and D10B-XPD 5831/06.
M.S. Ari thanks CNPq and Capes for financial supports.
Appendix A. Supplementary information
Supplementary data associated with this article can be found
in the online version at doi:10.1016/j.jssc.2008.03.015
References
[1] J.S.O. Evans, T.A. Mary, A.W. Sleight, Phys. B: Condens. Matter 241–243 (1998)
311–316.
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
1479
A.W. Sleight, Annu. Rev. Mater. Sci. 28 (1998) 29–43.
A.W. Sleight, Inorg. Chem. 37 (1998) 2854–2860.
J.S.O. Evans, J. Chem. Soc. Dalton Trans. (1999) 3317–3326.
J.Z. Tao, A.W. Sleight, J. Solid State Chem. 173 (2003) 442–448.
J.S.O. Evans, T.A. Mary, A.W. Sleight, J. Solid State Chem. 133 (1997) 580–583.
J. Peng, M.M. Wu, H. Wang, Y.M. Hao, Z. Hu, Z.X. Yu, D.F. Chen, R. Kiyanagi,
J.S. Fieramosca, S. Short, J. Jorgensen, J. Alloys Compd. 453 (2008) 49–54.
M.M. Wu, J. Peng, Y.Z. Cheng, H. Wang, Z.X. Yu, D.F. Chen, Z.B. Hu, Solid State
Sci. 8 (2006) 665–670.
Y.Z. Cheng, M.M. Wu, J. Peng, X.L. Xiao, Z.X. Li, Z.B. Hu, R. Kiyanagi,
J.S. Fieramosca, S. Short, J. Jorgensen, Solid State Sci. 9 (2007) 693–698.
M.M. Wu, Y.Z. Cheng, J. Peng, X.L. Xiao, D.F. Chen, R. Kiyanagi, J.S. Fieramosca,
S. Short, J. Jorgensen, Z.B. Hu, Mater. Res. Bull. 42 (2007) 2090–2098.
M.M. Wu, J. Peng, Y.Z. Cheng, X.L. Xiao, Y.M. Hao, Z.B. Hu, Mater. Sci. Eng. B 137
(2007) 144–148.
T.A. Mary, A.W. Sleight, J. Mater. Res. 14 (3) (1999) 912–915.
V. Sivasubramanian, T.R. Ravindran, R. Nithya, A.K. Arora, J. Appl. Phys. 96 (1)
(2004) 387.
P.M. Forster, A. Yokochi, A.W. Sleight, J. Solid State Chem. 140 (1998) 157–158.
D.A. Woodcock, P. Lightfoot, C. Ritter, J. Solid State Chem. 149 (2000) 92–98.
P.M. Forster, A.W. Sleight, Int. J. Inorg. Mater. 1 (1999) 123–127.
B.A. Marinkovic, P.M. Jardim, R.R. de Avillez, F. Rizzo, Solid State Sci. 7 (2005)
1377–1383.
S. Sumithra, A.M. Umarji, Solid State Sci. 8 (2006) 1453–1458.
A.K. Tyagi, S.N. Achary, M.D. Mathews, J. Alloys Compd. 339 (2002) 207–210.
F.F. Ferreira, E. Granado, W. Carvalho Jr., S.W. Kycia, D. Bruno, R. Droppa Jr.,
J. Synchrotron Rad. 13 (2006) 46–53.
W.M. Yim, R.J. Paff, J. Appl. Phys. 45 (1974) 1456–1457.
R.D. Shannon, Acta Crystallogr. A 32 (1976) 751.
V. Sivasubramanian, T.R. Ravindran, S. Kalavathi, A.K. Arora, J. Electroceram.
17 (2006) 57–60.
A.L. Allred, J. Inorg. Nucl. Chem. 17 (1961) 2115–2221.
K. Nassau, H.J. Levinstein, G.M. Loiacono, J. Phys. Chem. Solids 26 (1965)
1805–1816.
Y.M. Yang, L.Ch. Li, M. Feng, Chin. J. Inorg. Chem. 23 (3) (2007) 382–386.
J.S.O. Evans, T.A. Mary, T. Vogt, M.A. Subramanian, A.W. Sleight, Chem. Mater.
8 (12) (1996) 2809–2823.
X.L. Xiao, Y.Z. Cheng, J. Peng, M.M. Wu, D.F. Chen, Z.B. Hu, R. Kiyanagi,
J.S. Fieramosca, S. Short, J. Jorgensen, Solid State Sci. 10 (2008) 321–325.
Z.X. Yu, J. Peng, H. Wang, M.M. Wu, Y.Z. Cheng, Z.B. Hu, D.F. Chen, Sci. China
Ser. E—Tech. Sci. 51 (2008) 25–32.
K.W. Chapman, P.J. Chupas, C.J. Kepert, J. Am. Chem. Soc. 128 (2006)
7009–7014.
J.S.O. Evans, T.A. Mary, A.W. Sleight, J. Solid State Chem. 137 (1998) 148–160.
J.S.O. Evans, T.A. Mary, Int. J. Inorg. Mater. 2 (2000) 143–151.
R.M. Hazen, C.T. Prewitt, Am. Mineral. 62 (1977) 309–315.
S. Sumithra, A.K. Tyagi, A.M. Umarji, Mater. Sci. Eng. B 116 (2005) 14–18.
N. Imanaka, M. Hiraiwa, G. Adachi, H. Dabkowska, A. Dabkowski, J. Cryst.
Growth 220 (2000) 176–179.
S. Sumithra, A.M. Umarji, Mater. Res. Bull. 40 (2005) 167–176.