Reducing ultra-low thermal expansion of β-Zr2O(PO 4)2 by substitutions?

Article abstract of DOI:10.1016/j.materresbull.2010.08.005

The ultra-low thermal expansion of β-Zr₂O(PO ₄)₂ could still be reduced by chemical modifications, according to previous works on uranium and thorium IV homologues that evidenced a contracting effect of big cations. However, the present study shows that the only isovalent cations likely to substitute for Zr^IV or P^V are either harmful, or unstable, chemically or thermally. Furthermore, aliovalent modifications seem to be forbidden by the crystal structure of the material, leaving probably very few perspectives of improvement. Vanadium however improves significantly the crystallinity of the material and lowers the α-β transition temperature.

Full text of DOI:10.1016/j.materresbull.2010.08.005

Materials Research Bulletin 45 (2010) 1996–2000
Contents lists available at ScienceDirect
Materials Research Bulletin
journal homepage: www.elsevier.com/locate/matresbu
Reducing ultra-low thermal expansion of
b-Zr O(PO ) by substitutions?
2 4 2
C e´ line Barreteau, Damien Bregiroux, Guillaume Laurent, Gilles Wallez *
UPMC Univ. Paris 06, CNRS-UMR 7574, ENSCP-ParisTech, Laboratoire de Chimie de la Mati e` re Condens e´ e de Paris, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France
A R T I C L E I N F O
A B S T R A C T
Article history:
2 4 2
The ultra-low thermal expansion of b-Zr O(PO ) could still be reduced by chemical modifications,
Received 21 July 2010
Accepted 16 August 2010
Available online 26 August 2010
according to previous works on uranium and thorium IV homologues that evidenced a contracting effect
of big cations. However, the present study shows that the only isovalent cations likely to substitute for
IV
V
Zr or P are either harmful, or unstable, chemically or thermally. Furthermore, aliovalent modifications
seem to be forbidden by the crystal structure of the material, leaving probably very few perspectives of
improvement. Vanadium however improves significantly the crystallinity of the material and lowers the
Keywords:
A. Ceramics
a–b transition temperature.
A. Structural Materials
B. Sol–gel chemistry
C. X-ray diffraction
D. Thermal expansion
ß 2010 Elsevier Ltd. All rights reserved.
1
. Introduction
Oxides with low or negative thermal expansion (L/NTE) make a
The NZP family (from rhombohedral archetype NaZr
2 4 3
(PO ) )
(
Fig. 1b) gathers countless formulations through substitution of any
+
size-compatible mono- or multivalent cation for Na in the so-called
4+
small but remarkable class of materials. The formers, endowed
with a high resistance to thermal shocks, have been used since the
beginning of the XXth century for various high temperature
applications, while the latters can be used to elaborate zero-
expansion composites and constitute an original field of investi-
gation for the academic. Research and design in this domain have
long been ruled by empirical considerations and, as pointed by Roy
et al. in 1989, progresses remained slow due to a limited
understanding of the real causes of thermal contraction [1].
However, at the end of the following decade, thanks to the
improvements of diffraction techniques, the pioneer works of
Sleight led to a phenomenological classification in which the
thermal evolutions of the polyhedra networks appeared as key
parameters to understand the contraction mechanisms [2].
At this time, the anhydrous zirconium phosphates had been
extensively studied by ceramists who evidenced remarkable LTE
materials [3,4]. A high chemical and thermal stability, low toxicity,
moderate cost and easiness of elaboration strongly contributed to
the interest for these materials.
M1 sixfold coordinated site. Iso- and aliovalent substitutions for Zr
5
+
in octahedral site and P in tetrahedral site are also possible. The M2
sites, vacant in the archetype at room temperature, can be occupied
+
by supplementary low-valence cations like Na , thus still extending
the spectrum of chemical compositions. Driven by inter-cations
Coulombicrepulsions, the strongthermalexpansion along the c-axis
is balanced through a ring-like deformation mode by a contraction
along the perpendicular axes [7,8].
The structure and low expansion mechanism of
2 4 2
a-Zr O(PO )
(Fig. 1c) have been characterized only recently [9,10], but this
metastable monoclinic form sustains an irreversible martensitic
transition to
b between 1120 and 1220 8C, thus forbidding any
ꢁ6
ꢁ1
application despite a low expansion coefficient (+2.6 ꢀ 10
K )
ꢁ
6
ꢁ1
[4,10]. However, the hafnium analogue (+2.9 ꢀ 10
undergo the transition before decomposition, around 1500 8C.
Orthorhombic -Zr O(PO -ZP, Fig. 1d) features ZrO
7
K
) does not
a–b
b
2
4
)
2
(b
pentagonal bipyramids that form chains by edges-sharing along
the a-axis [11]. The non-phosphate oxygen builds a linear unit
with two zirconium atoms. The ultra-low expansion of
b
-ZP
ꢁ1
K ) has been discovered in 1985 [12] and explained
(+1.5 ꢀ 10ꢁ
6
ZrP
2
O
7
(cubic) (Fig. 1a) shows a 3 ꢀ 3 ꢀ 3-fold superstructure
that collapses at high temperature. Above the transition point, this
diphosphate [5] and its divanadate counterpart [6] show
respectively LTE and NTE that result from the thermal rocking
recently by a dual mechanism involving a reduction of the cell
cavities under the effect of inter-cations repulsions and
a
transverse rocking of the oxygen atoms in twofold coordination
[13]. Also thermally and chemically stable, it is undoubtedly the
most useful among zirconium phosphates regarding applications
6 4
of the corner-connected ZrO and P/VO polyhedra.
(
refractories, composites with SiC fibers and dimensionally stable
components up to 1550 8C). It has been commercialized since the
late 1990s, mainly under porous form. Paradoxically, -ZP remains
far less studied by the academic community than the NZP’s and
b
*
Corresponding author. Tel.: +33 1 53 73 79 46; fax: +33 1 46 34 74 89.
E-mail address: gilles.wallez@upmc.fr (G. Wallez).
0025-5408/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2010.08.005

[
(
F
i
g
.
_
1
)
T
D
$
F
I
G
]
C. Barreteau et al. / Materials Research Bulletin 45 (2010) 1996–2000
1997
Fig. 1. From top left (a) to bottom right (d): Negative thermal expansion mechanisms in the zirconium phosphates: blue arrows show polyhedra or oxygen rocking, red arrows
ring-like deformations or atomic movements resulting from inter-cations repulsions, black arrows show the global expansion/contraction. Artwork made with Balls & Sticks
[
17]. (For interpretation of the references to color in text, the reader is referred to the web version of the article.)
diphosphate. To this date, the only known isotypes are U
2
O(PO
4
)
2
2. Experimental
IV
IV
˚ ˚ ˚
and Th O(PO ) [14] (U : 1.09 A, Th: 1.14 A (est.), Zr : 0.92 A in
2 4 2
sevenfold coordination [15]) which exhibit a remarkable negative
2.1. Syntheses
expansion (resp. ꢁ1.2 and ꢁ1.5 10ꢁ6
K
ꢁ1
[16]). For safety concerns,
the main interest of these compounds is to evidence a clear
relationship between structural and thermal properties: the bigger
the cation, the stronger the rocking, the lower the expansion. In
these compounds, the steric effect of the cation consists in giving
its oxygen neighbors enough space to rock independently of the
other ones, even if the latters are blocked by a higher coordinence
4 2 4
Wet route syntheses are carried out from NH H PO (Sigma–
Aldrich, 98.5%), ZrOCl
2
ꢂ8H
2
O (Sigma–Aldrich, 99.5%) and ade-
quate soluble salts of cations for substitutions weighted in
proportions corresponding to the expected final formulas, then
dissolved in distilled water and mixed to obtain gels. We observed
that remaining chlorides replace the phosphates as counteranions
(as it is the case in
b
-ZP). According to this mechanism, and at
2
and provoke their loss and formation of ZrO during annealing. So,
variance with the rigid units model that prevails for global
polyhedra rocking, substitutions of cations with higher radii for
the gels have to be washed and separated by centrifugation
several times until elimination of Cl anions from the supernatant
ꢁ
IV
V
Zr or even P could be an attractive way to bring the thermal
expansion of -ZP closer to zero. Whereas the literature is
(AgNO
at various temperatures depending on the composition. For dry
route syntheses, ZrO(NO O (Prolabo, 99%) is preferred to
ꢂ2H
ZrO for its higher reactivity. Specific informations are reported in
Table 1.
3
test). The precursors are then dried at 300 8C and annealed
b
surprisingly scarce on this subject, the present work reports a wide
spectrum of experiments and proposes conclusions based upon
structural criteria about the possibility to improve this material.
3
)
2
2
2
Table 1
Summary of the synthesis experiments for substituted forms of
b-ZP. WR and DR refer to wet and dry routes.
Expected substitution
Specific reagents
Heat treatment
Result
IV
IV
a
˚
3
III
Ce ! Zr , WR
CeO
id.
2
99.9% in HNO
3
1 M
1200–1500 8C
b
b
a
-ZP (683.87(7) A ) + Ce PO
4
4
air or O
1200–1500 8C
2
IV
IV
˚
3
III
5
% Ce ! Zr , WR
-ZP (683.92(8) A ) + Ce PO
air or O
1100 8C
1200 8C
1400 8C
1400 8C
2
V
V
a
As ! P , WR or DR
As
NH
NbCl
Nb
Mg(NO
WO
2
O
5
ꢂxH
VO
99% in EtOH
2
O 99.99%
-Form Zr O(AsO
2
4
)
2
, see Table 2
V
V
b
1
5
7
5
5
5
5
5
5
5
5
–12% V ! P , WR
4
3
99% in HNO
3
0.1 M
-ZP
See Section 3.2
V
V
a
˚
3
% Nb ! P , WR
5
b
b
b
-ZP (683.80(9) A ) + Nb
2
Zr
6
O
17
17
V
V
a
-ZP (683.94(7) A ) + Nb
˚
3
.3% Nb ! P , DR
2
O
5
99.8% , 5 mass% +
b
2
Zr
6
O
II
IV
a
-ZP (684.04(8) A ) + unidentified phase
˚
3
% Mg ! Zr
3
)
2
ꢂ6H
2
O 99%
1200–1400 8C
VI
V
a
% W ! P , WR
H
2
4
99% in NH
3
2 M
III
IV
d
˚
3
% Tb ! Zr
Tb
4
O
7
in HNO
3
1 M
1400 8C
b
b
b
-ZP (684.05(3) A ) + TbPO
4
+ unidentified phase
˚
-ZP (684.33(3) A ) + ScPO + unidentified phase
4
VI
V
a
% W ! P , WR
H
2
WO
99.9% in HNO
4
99% in NH
3
2 M
III
IV
a
3
% Sc ! Zr
Sc
2
O
3
3
1 M
1400 8C
VI
V
% W ! P , WR
IV
V
c
-ZP (684.04(8) A ) + unidentified phase
˚
3
% Si ! P
SiC
8
O
4
H
20 (TEOS), 98%
1250 8C
1450 8C
VI
V
a
% W ! P , WR
H
2
WO
4
99% in NH
3
2 M
id. + ZrO
2
a
b
c
Sigma–Aldrich.
Merck.
Fluka.
d
Laboratory preparation.

1
998
C. Barreteau et al. / Materials Research Bulletin 45 (2010) 1996–2000
Table 2
Cell parameters ( A˚ ,8) and thermal expansion coefficients 20–750 8C (8C ) for
ꢁ1
a-Zr
2
O(PO
4
)
2
[9,10] and
c/
a-Zr
2
O(AsO
4
)
2
(this work).
a/a
a
b/a
b
a
c
b
l V
V/a = (a )/3
a
-Zr
-Zr
2
O(PO
4
)
2
10.2726(6)
6.5957(3)
ꢁ1.1
10.0665(5)
7.0
95.433(3)
95.383(2)
679.00(6)
2.6
1
.7
10.3547(5)
.3
a
2
O(AsO
4
)
2
6.8345(3)
ꢁ1.2
10.3568(4)
7.6
729.71(6)
2.5
1
2
.2. Analytical methods
appear around 12 or 16%, likewise, as shown on Fig. 2, the increase
of the cell volume tend to reduce with V rate. Even within the
solubility limits, the color of the samples evolves gradually from
pale yellow to amber with the vanadium rate.
The products are identified by X-ray diffraction on a Panalytical
X’Pert Pro apparatus. The diffractometer is equipped with an
Anton-Paar furnace for high temperature X-ray diffraction. The cell
parameters are refined by Rietveld analysis [18]. A particular
V
As a rule of thumb, considering the nearly equal radii of V and
V
As on the one hand, and the similarity of the densities of
a
-ZP and
attention is paid to the cell parameters and volume of the
b
-ZP
b
-ZP (Dr = 0.7%) on the other hand, the
b
-Zr
2
O((P,V)O
4
)
2
cell
phase in order to observe even faint chemical variations of the
volume should evolve roughly like for the
a
-Zr O((P,As)O
2
4
)
2
solid
31
51
3
˚
material. P and V solid state nuclear magnetic resonance (NMR)
spectra were recorded on a Bruker Avance III 300 spectrometer
solution (Table 2), at about 0.5 A /% (Fig. 2, red line). Actually, the
experimental slope is lower than expected, thus calling for a
1
with a magnetic field of 7 T corresponding to 300.29 MHz for H,
thorough analysis of the substitution mechanism.
31
51
31
1
21.56 MHz for
P and 78.93 MHz for
V. A Bruker triple
The P NMR spectrum of pure b-ZP shows a strong single peak
resonance Magic Angle Spinning (MAS) probe was used with
a
[
F
g
_
)
D
F
G
t ꢁ23.1 ppm (Fig. 3) that corresponds to the unique P site. This
31
51
spinning frequencies of either 14 or 11.5 kHz. P and V spectra
were respectively referenced to H PO 85% and VOCl
3
4
3
.
3
. Results and discussion
.1. Isovalent substitution for Zr
Besides actinides, the only tetravalent cations bigger than Zr
IV
3
IV
IV
IV
˚
˚
are Pb (0.95 A) and Ce (1.06 A). The former has to be dismissed
because of its volatility which makes it unsuitable for high
IV
temperature applications. The attempts to elaborate Ce
2 4 2
O(PO )
III
under air by dry and wet routes only yield monazite Ce PO
known to be a highly stable mineral. A Zr/Ce = 95/5 mol ratio
precursor gives likewise a mixture of CePO and unsubstituted b-
4
,
4
ZP. Similar experiments made under oxygen flow give the same
results. In a previous work, we evidenced the thermal decomposi-
IV
tion of the first known cerium IV phosphate (Ce
4 4 4 2 7
(PO ) (P O ))
III
III
4 3 3
into a mixture of Ce PO and Ce (PO ) around 850 8C [19], also
accounting for the extreme difficulty to accommodate Ce in an
anhydrous phosphate environment [20]. So, isovalent substitu-
IV
IV
tions for Zr seem to be limited to the previously known actinide-
based isotypes. Besides, stable tetravalent cations bigger than Zr
IV
simply do not exist.
Fig. 2. Black plot: experimental variation of the b-ZP cell volume vs. vanadium rate.
V
3.2. Isovalent substitution for P
Red line corresponds to the expected variation. (For interpretation of the references
to color in text, the reader is referred to the web version of the article.)
[F(ig._3)TD$FIG]
V
Likewise, the substitution of a bigger pentavalent cation (As
V
V
V
˚
˚
˚
˚
(
0.475 A), V (0.495 A) or Nb (0.62 A)) for P (0.31 A) in the
tetrahedral site could conceivably reduce the thermal expansion,
like for V-substituted ZrP
2
O
7
[6]. We obtain inedite Zr
2
O(AsO
4
)
2
by
both wet and dry routes and find it to be isotype with
a
-ZP
(monoclinic I2/m). This allows supposing that a continuous solid
solution exists between the phosphate and the arsenate. Cell
parameters and thermal expansion coefficients for both com-
pounds are reported in Table 2: the arsenate is found to have also a
low expansion, but not significantly lower than that of the
phosphate.
2 4 2
a-Zr O(AsO ) has been heated in search for a b form,
but decomposition occurs around 1300 8C, before a transition
could be observed. This suggests that a bigger pentavalent cation in
the tetrahedral site stabilizes the
a
form, whereas a compared
-HP (Hf homologue) evidences an inverse
steric effect of the tetravalent cation [9].
IV
study of
a-ZP and a
V
V
Although very similar in size to As , V is found to substitute
V
31
Fig. 3. P NMR spectra of pure and V-substituted b-ZP.
only in small amounts for P : ZrO
2
and Zr(P,V)
2
O
7
diffraction peaks

[
(
F
i
g
.
_
4
)
T
D
$
F
I
G
]
C. Barreteau et al. / Materials Research
[
i
.
5
T
$
I
]
Bulletin 45 (2010) 1996–2000
1999
51
Fig. 4. V NMR spectra of V-substituted
b-ZP. *Indicate spinning side bands.
value is in good agreement with the 25 ppm reported by Clayden
for a P(–O–Zr) environment [21]. The peak shifts progressively
with the increasing V rate, while a shoulder grows up around
22 ppm, probably due to moderate deformation of the P site
caused by the substitution of the neighboring P’s by a V (shortest
4
Fig. 5. Amorphous,
a
,
b
and coexistence domains of Zr
2
O((P,V)O
4
)
2
solid solutions.
ꢁ
Relative amounts of both phases are inferred from the scale factors (Rietveld
analysis, = 50:50 mixture used as standard).
a:b
˚
d = 3.95 A). Both peaks cease to evolve between 12 and 16 V%,
accounting for the already observed saturation. Even at high V%,
neither an increase of the spinning side bands or a strong decrease
Improved diffusion mechanisms, either through a vanadium-
51
rich liquid phase (corresponding to the amorphous peaks of
V
31
of the relaxation delay is observed on P spectra, indicating the
absence of detectable paramagnetic species in the samples,
NMR) or involving structural vacancies can be responsible for these
beneficent effects that could allow the elaboration of well-
V
meaning that Vanadium is only as V .
crystallized
The effect of Nb was mentioned in 1987 by Yamai et al., who
observed a slight increase of the cell volume for a -ZP sintered at
1400 8C in addition with 5 mass% Nb [23]. Considering that the
b-ZP refractories at moderate temperature.
51
V
On V NMR (Fig. 4), the central peak is strongly shifted
compared to V poly-oxo-vanadates, but such a value is not
unusual for V in a zirconium oxide environment [22]. Its splitting
ꢁ813 and ꢁ834 ppm) can be ascribed to the second-order nuclear
quadrupole interaction caused by the strong asymmetry inducing
electric field gradient in the host tetrahedron (in pure -ZP, the
x
O
y
b
2 5
O
V
V
IV
(
ionic radius of Nb is intermediate between those of P and Zr , this
would account for a partial substitution of Nb for P . However, this
experiment, made again in the same conditions, gives us a sintered
V
V
b
edge shared with Zr results in a low O–P–O angle of 96.88). On both
sides of the main peak, spinning side bands are observed at
multiples of the spinning frequency. One can notice that at 16 V% a
broad peak appears around ꢁ620 ppm attributable to a secondary
amorphous vanadate phase. The spinning frequency was changed
to better see this peak.
composite (89% compact) of unmodified
Despite Nb sometimes occurs in fourfold coordination, it could be
b
-ZP and Nb
2 6 17
Zr O .
V
too bulky to take place into the
b-ZP structure.
3.3. Aliovalent substitutions
V
IV
V
Whereas the partial insertion of V appears undoubtful on the
Simultaneous substitutions for Zr and P by cations of
ifferent oxidation states are also conceivable, provided that the
basis of XRD and NMR, it seems however poorly stable in this
environment: the prolonged annealing (1200 8C, 10 h) of an 8 V%
dried sample results in a 1.5% weight loss, to be compared to its
d
[(Fig._6)TD$FIG]
2 5
nominal 2.3 mass% assay in V O (the weight of a pure b-ZP sample
heatedinthesameconditionsremainsunchanged).Inthesametime,
weak diffractionpeaksofanunknownphaseappear, accountingfor a
decomposition, probably due to the evaporation of vanadium oxide.
It seems probable that this loss also occurs during the synthesis, thus
explaining the volumic anomaly observed, but it is difficult to
measure precisely the real V rate in the
b-ZP phase of these samples
and to conclude about possible vacancies left by the P/V
2 5
O
loss and
affecting the cell volume. Nonetheless, no significant variation of the
ꢃ 1.6 ꢀ 10ꢁ
6
K ) in the
ꢁ1
coefficient of thermal expansion (1.5 ꢃ
a
solubility domain (0, 2, 6, 8, 10 and 12 V%) was measured by high
temperature XRD and Rietveld analysis.
On the opposite, vanadium has remarkable structural effects on
the material. As shown in Fig. 5, the crystallization (DTA
measurements) and the
a–b transition (XRD measurements)
occur at lower temperatures and, concerning the latter, within a
reduced thermal range when the V rate increases. Even in faint
amounts, vanadium also enhances the crystallinity of the
which diffraction peaks are narrower (Fig. 6). At last, whereas pure
-ZP needs to be ground several times and annealed up to 1400 8C
to eliminate residual ZrO and ZrP , samples with 2 V% treated at
000 8C appear single-phase.
b form,
Fig. 6. Zoom on the XRD powder patterns of 0 V%
V% -ZP (heated 1000 8C, black). Note the broadening of the 0 V% peaks
HWHM = 0.128) compared to those of the 2% sample (0.108), and the presence of
residual ZrP and ZrO in the former. (For interpretation of the references to color
in text, the reader is referred to the web version of the article.)
b-ZP (heated 1300 8C, red) and
2
b
b
(
2
2 7
O
2
O
7
2
1

2
000
C. Barreteau et al. / Materials Research Bulletin 45 (2010) 1996–2000
global charge of the cell remains unchanged. Various schemes have
been considered, involving size-compatible bivalent and trivalent
probably forbidden by the structural peculiarities of the
material, thus leaving no clear way to increase its intrinsic
resistance to thermal shocks. To our opinion, adjusting extrinsic
microstructural parameters such as porosity, secondary phases
IV
cations in substitution for Zr (penta- and hexavalent cations are
VI
V
all smaller), while W has been chosen to replace P for its thermal
and chemical stability as well as its compatibility with a
tetrahedral environment. As shown in Table 1, each of these
potential dopants eventually gives an unwanted compound in
or microcracks is a more promising way to improve
b-ZP,
whereas the bulk itself should be considered as stuck to the
archetype formulation. Considering its technological impor-
tance, such attempts as ours to modify the composition might
have been carried out previously and have given similar results,
therefore remained unreported. So, we hope this in-depth study
addition to pure
b-ZP, meaning a failure of the substitution.
Kim et al. recently reported the elaboration of single-phase
b
-
III
IV
ZP doped with 1–10% Tb in substitution for Zr and presented the
diffraction pattern of the 3% Tb sample [24]. The authors gave no
informations about the cell parameters, but a careful examination
on
b-ZP will give ceramists a clearer insight on its crystal–
chemical properties and help them saving time and efforts in
their works on this valuable material.
evidences the diffraction peaks of xenotime-like TbPO
4
at
2
u
= 19.68 and 25.68, like for our sample, leading to suppose that
the substitution was not really effective. The light emission peaks
they reported result undoubtedly of f-level transitions of Tb , but
Acknowledgments
III
keeping in mind their low sensitivity to the environment, one must
This work was funded by the French program MATINEX (CNRS-
CEA). The authors are very grateful to El e´ onore Noussan and
Laetitia Corde, students of Chimie-ParisTech for their precious help
during their training period and to Dr. Christel Gervais-Stary
(LCMCP) for her comments on the NMR spectra.
4
remark that they also match with those of TbPO .
Another aliovalent substitution has been thought up for the
IV
tetrahedral site only, to be partly occupied by equal amounts of Si
VI
V
and W , both bigger than P . This attempt is also unsuccessful.
The systematic failures of aliovalent substitutions can be
understood on structural bases. First, one should consider that the
Appendix A. Supplementary data
ZrO
7
polyhedra are paired by vertices through Zr–O–Zr linkages
(
Fig. 1d): to our knowledge, such oxo units occur only for high-
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.materresbull.2010.08.005.
III
valence cations (ꢄIV); otherwise (i.e. for a M O
distribution of bond strengths would be exceedingly assymmetric
1 v.u. for the oxo bond and only 2 v.u. to be shared between the 6
7
polyhedron) the
(
References
other ones). In these conditions, no efficient aliovalent substitution
can be made for Zr . This calls for a comparison with the NZP
structure, in which Zr occupies an independent octahedron and
IV
[1] R. Roy, D.K. Agrawal, H.A. McKinstry, Annu. Rev. Mater. Sci. 19 (1989) 59–81.
IV
[2] A.W. Sleight, Inorg. Chem. 37 (1998) 2854–2860.
can be replaced by any size-compatible cation with valency III to V,
thus allowing an easy control of the thermal expansion. Another
[3] D.E. Harrison, H.A. McKinstry, F.A. Hummel, J. Am. Ceram. Soc. 37 (1954) 277–280.
[
[
4] J. Alamo, R. Roy, J. Am. Ceram. Soc. 67 (1984) 80–82.
5] N. Khosrovani, V. Korthuis, A.W. Sleight, T. Vogt, Inorg. Chem. 35 (1996) 485–489.
structural peculiarity of
b
-ZP should be put forward for
[6] V. Korthuis, N. Khosrovani, A.W. Sleight, N. Roberts, R. Dupree, W.W. Warren,
Chem. Mater. 7 (1995) 412–417.
[7] J.P. Boilot, J.P. Salani e´ , G. Desplanches, D. Le Potier, Mater. Res. Bull. 14 (1979)
IV
understanding the failure of the simultaneous substitution of Si
VI
V
7 4
and W for P : the edge-sharing between ZrO and PO . Such a
1469–1477.
feature is rather uncommon between cations with so high charges,
and increased Coulombic interactions, like those caused by W for
example, could make it impossible. Indeed, to our knowledge, a
[8] D.K. Agrawal, C.Y. Huang, H.A. McKinstry, Int. J. Thermophys. 12 (1991) 697–710.
VI
[9] G. Wallez, J.P. Souron, M. Quarton, Solid State Sci. 8 (2006) 1061–1066.
10] G. Wallez, D. Bregiroux, M. Quarton, J. Solid State Chem. 181 (2008) 1413–1418.
11] W. Gebert, E. Tillmanns, Acta Crystallogr. B31 (1975) 1768–1770.
12] I. Yamai, T. Ota, J. Am. Ceram. Soc. 68 (1985) 273–278.
[
[
[
IV
4
WO tetrahedron never shares an edge with a M -based
VI
+
V
polyhedron. At variance once again, W can replace (Na + P )
in NZP to give orthorhombic Zr (PO (WO ), a somewhat different
structure in which the tungstate tetrahedra are corner-connected
25,26]. It is also nowadays used in commercial ultra-low
expansion materials.
[13] G. Wallez, S. Launay, J.P. Souron, M. Quarton, E. Suard, Chem. Mater. 15 (2003)
793–3797.
3
2
4
)
2
4
[
[
14] N. Dacheux, N. Clavier, G. Wallez, M. Quarton, Solid State Sci. 9 (2007) 619–627.
15] R.D. Shannon, Acta Crystallogr. A32 (1976) 751–767.
[
[16] G. Wallez, S. Launay, M. Quarton, N. Dacheux, J.L. Soubeyroux, J. Solid State Chem.
77 (2004) 3575–3580.
1
[
[
17] S.J. Kang, T.C. Ozawa, Balls & Sticks, V.1.51, 2004.
18] J. Rodriguez-Carvajal, FULLPROF.2k: Rietveld, Profile Matching and Integrated
Intensity Refinement of X-ray and Neutron Data, V 1.9c, Laboratoire L e´ on
Brillouin, CEA, Saclay, France, 2001.
4
. Conclusion
[
[
19] M. Nazaraly, G. Wallez, C. Chan e´ ac, E. Tronc, M. Quarton, J.P. Jolivet, J. Phys. Chem.
Solids 67 (2006) 1075–1078.
20] D. Bregiroux, O. Terra, F. Audubert, N. Dacheux, V. Serin, R. Podor, D. Bernache-
Assollant, Inorg. Chem. 46 (2007) 10372–10382.
[21] N.J. Clayden, J. Chem. Soc. Dalton Trans. 8 (1987) 1877–1881.
22] J.R. Sohn, S.G. Cho, Y.I. Pae, S. Hayashi, J. Catal. 159 (1996) 170–177.
23] I. Yamai, T. Ota, J. Am. Ceram. Soc. 70 (1987) 585–590.
Despite the substitution of bigger cations in
appear as an appealing way to reduce its thermal expansion,
b-ZP could
IV
IV
very few isovalent species proved to be compatible: Th and U
V
V
for Zr; V for P in low amounts because of its volatility, and As
[
[
that only yields of
a-form. Either unsafe, volatile or chemically
unstable, the isovalent cations do not give useful materials, but
vanadium improves the crystallinity and reduces the transition
point significantly. This could be an interesting way to elaborate
this material at a lower cost. Aliovalent substitutions are very
[24] S.W. Kim, T. Masui, H. Matsushita, N. Imanaka, Chem. Lett. 38 (2009) 1100–
101.
1
[
[
25] J.S.O. Evans, T.A. Mary, A.W. Sleight, J. Solid State Chem. 120 (1995) 101–104.
26] T. Isobe, T. Umezome, Y. Kameshima, A. Nakajima, K. Okada, Mater. Res. Bull. 44
(2009) 2045–2049.