Facile synthesis of interstitial metal nitrides with the filled β -manganese structure

Article abstract of DOI:10.1016/S0022-4596(02)00171-8

A new family of nitrides, Ni_2-xM′_xMo₃N (M′=Co or Pd; 0≤x≤1.5), has been prepared pure by nitridation of commercially available crystalline metal oxides under reducing conditions (10% H₂ in N₂). The simple synthesis employs standard solid-state techniques and does not require the preparation of reactive precursors. Substitution of Ni by Co or Pd leads to a linear increase of the unit cell volume with composition. The temperature, composition, and magnetic-field dependence of the molar magnetisation suggest that the introduction of Co, but not Pd, increases the degree of electron localisation in Ni_2-xM′_xMo₃N. The same synthetic method has also lead to the formation, in mixtures, of the new phases π-Co₂Mo₃N and Pd₂N.

Full text of DOI:10.1016/S0022-4596(02)00171-8

Journal of Solid State Chemistry 172 (2003) 138–147
Facile synthesis of interstitial metal nitrides with the filled
b-manganese structure
Timothy J. Prior and Peter D. Battle
ꢀ
Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QR, UK
Received 17 September 2002; received in revised form 5 November 2002; accepted 18 November 2002
Abstract
A new family of nitrides, Ni2ꢁx
0
0
M
x
Mo
3
N (M =Co or Pd; 0pxp1:5), has been prepared pure by nitridation of commercially
2 2
available crystalline metal oxides under reducing conditions (10% H in N ). The simple synthesis employs standard solid-state
techniques and does not require the preparation of reactive precursors. Substitution of Ni by Co or Pd leads to a linear increase of
the unit cell volume with composition. The temperature, composition, and magnetic-field dependence of the molar magnetisation
0
suggest that the introduction of Co, but not Pd, increases the degree of electron localisation in Ni2ꢁx
method has also lead to the formation, in mixtures, of the new phases p-Co
r 2003 Elsevier Science (USA). All rights reserved.
M
x
Mo
3
N. The same synthetic
2 3 2
Mo N and Pd N.
Keywords: Interstitial nitride; Beta-manganese structure; Nitride synthesis; Nitride magnetism
1
. Introduction
ternary nitrides has been attempted by a number of
routes. The majority of known ternary nitrides have
been synthesised from binary nitrides and metals or by
the ammoniation of ternary oxides. Other routes that
have been utilised include ammoniation of metal
sulphides [13] or amorphous mixed metal precursors.
[14] Frequently the preparation of these precursors
requires air-sensitive [14] or complex handling techni-
ques (e.g., freeze-drying [15]) while the purity of
products is often doubtful when ammonia is employed;
binary metal nitrides, [14] alloys, [16] or pure metals [17]
being common impurities. It is notable that there is no
simple, generally applicable method for the formation of
pure ternary metal nitrides.
Metal nitrides have attracted considerable attention
due to a number of interesting properties which they
display, for example hardness, [1,2] electrical [3,4] or
thermal [5] conductivity, superconductivity, [6] magnet-
ism, [7] and catalytic activity. [8,9] However, compared
with the enormous range of ternary and higher metal
oxides which is known, there is a dearth of mixed metal
nitrides. There are a number of reasons for this, not least
the refractory (and hence unreactive) nature of many
binary nitrides and the extreme strength of the bond in
dinitrogen. Elder et al. [10] have provided a means of
predicting which ternary nitrides might be accessible.
This is valuable when the appropriate thermodynamic
data are available, but in general it is impossible to
predict a priori whether a particular nitride will be
stable.
The incorporation of non-metals in interstitial sites
within metals and alloys leads to materials with
increased hardness. This is believed to be due to
strengthening of the metal–metal bonding. [2] One
structure type adopted by a number of interstitial mixed
metal compounds (borides, carbides, nitrides) which has
been receiving interest is the filled b-Mn structure
(sometimes denoted p-phase), M2T3X; which is exem-
plified by Al Mo C. [18,19] M atoms lie on the 8c
The synthesis of binary nitrides is normally accom-
plished by one of two methods: reaction of the metal
with nitrogen, often using a liquid metal flux, [11,12] or
reaction of the metal oxide with ammonia. Synthesis of
2
3
ꢀ
positions of a cubic unit cell with space group symmetry
P4132 forming a single (10,3)-a network. [20] The space
Corresponding author. Fax: +44-1865-272690.
E-mail address: peter.battle@chem.ox.ac.uk (P.D. Battle).
0022-4596/03/$ - see front matter r 2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0022-4596(02)00171-8

T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
139
within this network is filled by corner-sharing T X
6
2.1. Synthesis of p -phases
Ni Mo N: (1) A stoichiometric mixture of NiO
and MoO3 was intimately ground and fired under
flowing synthesis gas, for 72 h at 7751C and 24 h at
10001C, with intermittent regrinding. (2) A stoichio-
octahedra. (Fig. 1) The structure is entirely analogous to
that of b-Mn, differing only by the addition of the
interstitial non-metal atoms.
Our attempts to synthesise nickel molybdenum oxides
2
3
IV
under conditions suitable for the formation of Mo
(
1
standard ceramic synthesis techniques under flowing
0% hydrogen in nitrogen) did not yield the desired
metric mixture of NiO and MoO was intimately ground
2
and fired under flowing synthesis gas, for 48 h at 8001C
and 24 h at 9001C and 10001C, with intermittent
regrinding.
product; instead a mixture containing the b-Mn inter-
stitial nitride Ni Mo N was formed. Further experi-
mentation showed that pure Ni Mo N can be
2
3
Cobalt-doped phases: Stoichiometric mixtures of
NiO, Co O , and MoO were intimately ground and
2
3
deliberately prepared in a simple procedure in which
intimately ground crystalline metal oxides are heated
under flowing synthesis gas (10% H in N ) and that the
3
4
3
fired under flowing synthesis gas for 48 h at 7001C,
7501C, 10001C, and 11001C, with intermittent regrind-
ing.
2
2
same method can be used to prepare the p-phase solid
solutions Ni2 Co Mo N and Ni Pd Mo N. During
Palladium-doped phases: Stoichiometric mixtures of
NiO, PdO, and MoO were intimately ground and fired
ꢁx
x
3
2ꢁx
the course of these syntheses we were able to show that
x
3
3
our protocol also simplifies the preparation of the Z-
under flowing synthesis gas, for 48 h at 700, 750, and
twice at 11001C, with intermittent regrinding.
phases M3Mo N (M=Fe, Co) [16] and leads to the
formation of a new phase, tentatively identified as Pd N.
3
2
We describe below the synthesis and structural char-
acterisation of these phases and a preliminary survey of
their magnetic properties.
2.2. Synthesis of Z-phases
Stoichiometric mixtures of MoO and either Co O or
4
3
3
Fe O were intimately ground and fired under flowing
2
3
2
. Experimental
synthesis gas, for 48 h at 700, and 24 h each at 7501C,
10001C, and 11001C, with intermittent regrinding.
Reagents purities: Cobalt oxide (Johnson Matthey,
9.9985%), Iron oxide (Alfa Aesar, 99.998%), Nickel
oxide (Johnson Matthey, 99.998%), Molybdenum
dioxide (Alfa Aesar, 99%), Molybdenum trioxide (Alfa
Aesar, 99.9995%), Palladium oxide (Johnson Matthey,
In all cases samples were cooled to room temperature
9
under flowing gas after switching off the furnace. No
study of the effects of variation of the cooling rate
was carried out. No special handling techniques were
applied to samples following their removal from the
furnace.
9
9.9%). Synthesis gas: 10% H in N (Air Products).
2 2
2 3
Fig. 1. Filled b-Mn structure of Ni Mo N. (a) Single (10,3)-a network of nickel atoms. (b) Nickel network (black spheres) filled with corner-sharing
Mo N octahedra (grey).
6

140
T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
2
.3. Characterisation
magnetisation was measured as a function of field
(
ꢁ50pH=kOep50) after cooling to the measuring
The progress of synthesis was charted by X-ray powder
diffraction using a Philips diffractometer operating with
CuKa and Ka radiation. Relatively high-resolution data
to be used in quantitative analyses were collected using a
Siemens D5000 diffractometer nominally operating with
temperature in a field of 50 kOe. Attempts were made
to fit the measured magnetic susceptibility (defined as
w ¼ M=H) to a Curie–Weiss law, modified to include a
temperature-independent term. However, this function
proved inadequate in many cases, and the discussion of
the magnetic properties given below is largely qualita-
tive.
1
2
CuKa radiation over the angular range 10p2y=1p125
1
with a step size of 0.021. However, imperfect mono-
chromation introduced a contribution of approximately
2
% Ka which was allowed for in the data analysis.
2
Rietveld refinement [21] of the crystal structures was
carried out using the GSAS [22] suite of programs.
Backgrounds were fitted using a shifted Chebyshev
polynomial of the first kind. Peaks shapes were modelled
using a pseudo-Voigt function.
3. Synthesis and structural chemistry
3.1. Ni Mo N
2
3
Rietveld analysis of X-ray powder diffraction data
(Fig. 2) showed that Ni Mo N crystallises with the filled
TGA was performed using a Rheometric Scientific
STA1500 instrument. Samples were loaded into plati-
num or alumina pans and heated under a flow
2
3
b-Mn structure in the isomorphous pair of space groups
P4 32 and P4 32: Crystal data are summarised in Table
1. It contains a net of nickel atoms, the holes in which
1
3
ꢁ1
(
30 ml min ) of pure oxygen. Metal analysis was
effected by ICP emission spectroscopy using a Thermo
Jarell Ash Atomscan 16 spectrometer. Nitrogen analysis
was performed using a Carlo Erber 1106 CHN analyser.
Magnetic measurements were carried out using a
Quantum Design MPMS 5000 SQUID magnetometer.
The sample magnetisation (M) was measured as a
function of temperature (5pT=Kp350) during warm
up after cooling to 5 K in either zero field (ZFC) or in
the measuring field of 100 Oe (FC). In selected cases, the
are filled by Mo N octahedra. Within the net, the
6
distance
between
neighbouring
nickel
atoms
(2.4694(6) A, 3-coordinate nickel) is similar to that for
˚
˚
the pure metal (2.492(1) A, 12-coordinate). The current
synthesis of Ni Mo N affirms the structural description
2
3
of Herle et al. [17] although in contrast to their
preparation which involved ammoniation of amorphous
precursors prepared under air sensitive conditions, our
X-ray diffraction patterns showed no evidence for the
Fig. 2. Observed ( Á ), calculated (-), and difference X-ray powder diffraction profiles for Ni
2 3
Mo N at room temperature; tick marks indicate
positions of allowed reflections from the Ka1 diffraction. The final structural refinement converged with Rwp ¼ 0:1218:

T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
141
presence of molybdenum nitride (Mo N) in the final
2
product.
of formation of the compound. Following the initial
heating at 7001C, X-ray powder diffraction patterns
The X-ray powder diffraction pattern of the TGA
oxidation product showed the presence of only NiMoO4
and MoO . The increase in mass upon complete
showed the presence of MoO , NiO, and the desired
2
product. This suggests that prior to the formation of the
final product, a stepwise reduction of molybdenum
trioxide is occurring. Synthesis employing a stoichio-
metric mixture of MoO and NiO under H /N also lead
3
oxidation of Ni Mo N to NiMoO and MoO is in
2
3
4
3
good agreement with the calculated value: Calc. 38.6%,
Found 37(1)%. Satisfactory chemical analysis data were
obtained for Ni Mo N: Calc. % Ni 27.99, Mo 68.65, N
2
2
2
to the formation of Ni Mo N, although this is less
2
3
convenient as it is rather difficult to obtain the starting
material MoO highly pure.
2
3
3
Ni Mo N0.96).
.34, Found % Ni 27.94, Mo 68.65, N 3.20 (gives
2
2
3
Ex situ X-ray powder diffraction data charting the
course of reaction gave an insight into the mechanism
3.2. Ni2ꢁxCo
Mo
N (x ¼ 0:5; 0.75, 1, and 1.5)
x
3
Compounds in the solid solution obtained by
substitution of Ni by Co may be synthesised using the
same protocol. The initial cobalt source (Co O ) is more
Table 1
Atomic positions for filled b-manganese nitrides Ni
Co Mo
2
Mo
3
N and
2
3
N
3
4
2
difficult to reduce than the nickel oxide, therefore a
number of firings at 11001C were necessary.
˚
Atom Wyckoff
position
x
y
z
Uiso/A
Rietveld analyses of X-ray patterns (Fig. 3) showed
that these compounds are isostructural with the parent
phase Ni Mo N, displaying a smooth increase in cubic
unit cell parameter with increasing cobalt content,
consistent with the larger size of cobalt. (Fig. 4) It is
not possible to determine by X-ray diffraction whether
there is any ordering of the nickel and cobalt within the
structure and these two elements were, therefore,
assumed to be disordered over the 8c sites. There is no
˚
Ni
Ni
2
Mo
3
N Space group P4132; a0 ¼ 6:63582ð3Þ A, Rwp ¼ 0:1218
8c
0.0670(1) 0.0670(1)
0.0670(1)
0.20153(6) 0.45153(6) 0.0086(7)
0.0108(9)
2
3
Mo
N
12d
4a
1/8
3/8
3/8
3/8
˚
0.012(3)
Co
Co
Mo
N
2
Mo
3
N Space group P4132; a0 ¼ 6:6653ð1Þ A, Rwp ¼ 0:0425
8c
0.0671(3) 0.0671(3)
0.0671(3)
0.4524(2)
3/8
0
12d
4a
1/8
3/8
0.2024(2)
3/8
0.06(3)
0
Fig. 3. Observed ( Á ), calculated (-), and difference X-ray powder diffraction profiles for NiCoMo
3
N at room temperature; tick marks indicate
positions of allowed reflections from the Ka1 diffraction. The final structural refinement converged with Rwp ¼ 0:051:

142
T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
evidence however to suggest the incorporation of the
first row transition metals into the Mo N octahedra.
the more cobalt-rich phase is preferred: reduction-
nitridation of a 1:1 stoichiometric mixture of cobalt
and molybdenum oxides yielded a pure sample of
Co Mo N. Before the final firing small amounts of
6
Satisfactory TGA oxidation analyses were obtained for
all compositions within the series, indicating full
nitrogen occupancy and ideal metal stoichiometry.
These are detailed in Table 2. Chemical analysis of the
nitrogen content for one representative composition,
3
3
Co O , MoO and Co Mo N were present, but these
3
4
2
2
3
reacted to give the desired product.
Co Mo N again crystallises with the filled b-Mn
2
3
NiCoMo N, was consistent with the expected analysis:
3
structure. (Table 1) However, X-ray powder diffraction
studies of this reaction product (Fig. 5) showed the
Calc. % N 3.34, Found % N 3.37. For Ni0.5Co1.5Mo N,
3
peaks corresponding molybdenum metal (2.7(1)%
by weight) and to the known phase Co Mo N [16]
3
3
(2.1(2)% by weight) were present in the X-ray diffrac-
tion pattern.
Table 2
Ratio of initial and final mass of Ni2ꢁxM
complete oxidation under flowing oxygen
Mo N (M=Co, Pd) upon
x 3
3
.3. Co Mo N
2 3
Ni2ꢁx
M Mo
x
3
N
x
It proved possible to synthesise this previously
0.5
0.75
1
1.5
2
unknown phase, albeit not in a pure form, by reduction
under nitriding conditions. Previous attempts to synthe-
sise this phase by plasma nitridation of oxides or via air-
sensitive precursors have yielded only the Z-phase
Co Mo N. [16] Indeed our studies also suggest that
M=Co
M=Pd
Calculated
Observed
1.31
1.33
1.31
1.33
1.31
1.33
1.31
1.31
1.31
1.35
Calculated
Observed
1.37
1.35
1.36
1.33
1.35
1.28
1.33
1.31
1.31
1.28
3
3
Fig. 4. Variation of unit cell parameter with composition for Ni2ꢁx
x 3
M Mo N (M=Co, Pd). Linear fits have been made through the data points.
Error bars are contained within the points.

T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
143
Fig. 5. Observed ( Á ), calculated (-), and difference X-ray powder diffraction profiles for Co
2 3
Mo N at room temperature; the top tick marks indicate
positions of allowed reflections from the Ka1 diffraction for Co
final structural refinement converged with Rwp ¼ 0:0425:
2
Mo
3
3 3
N, the middle marks correspond to Co Mo N, and the lowest are for Mo. The
presence of small amounts of molybdenum metal
4.8(3)% by weight) and Co Mo N (7.7(3)% by weight).
3.4. Fe Mo N
3
3
(
3
3
Upon prolonged heating the favoured products were
Co Mo N and molybdenum. We emphasise that no
impurities were observed in Ni2 Co Mo N over the
composition range 0pxp1:
This phase has been prepared previously in a pure
form by a method involving ammoniation of amor-
phous oxide precursors. [16] The X-ray diffraction
pattern (Fig. 6) of the product prepared by our simpler
3
3
ꢁx
x
3

144
T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
Fig. 6. Observed ( Á ), calculated (-), and difference X-ray powder diffraction profiles for Fe
3 3
Mo N at room temperature; tick marks indicate
positions of allowed reflections from the Ka1 diffraction. The final structural refinement converged with Rwp ¼ 0:0441:
route is in excellent agreement with that reported
˚
included in the data analysis. (Table 3) This phase has
not been reported previously.
previously (a ¼ 11:0802ð2Þ A).
0
3
.5. Ni2 Pd Mo N (x ¼ 0:5; 0.75, 1, 1.5, and 2)
4. Magnetism
ꢁx
x
3
Palladium may also be introduced into the network of
Measurements of magnetisation as a function of
temperature for Ni Mo N showed that the compound is
transition metal atoms. Synthesis was accomplished in
the same fashion. Rietveld refinement against X-ray
diffraction data (Fig. 7) shows that all of the compounds
in this series crystallise with the filled b-Mn structure
with no apparent ordering of Ni and Pd. Satisfactory
TGA oxidation analyses were obtained for all composi-
tions within the series. (Table 2) As the amount of
palladium included is increased there is a smooth
increase in the lattice parameter, as shown in Fig. 4.
No impurity phases were detected for 0pxp1:5; but the
X-ray powder diffraction pattern of the known
Pd Mo N [23] did contain weak peaks which could
2
3
weakly Pauli paramagnetic. Consistent magnetic beha-
viour was displayed across a number of samples, in
contrast to the variation in behaviour previously
reported. [17] We ascribe this to the absence of
impurities in our samples. The low magnetic suscept-
ibility (Fig. 8a) is consistent with the Pauli paramagnet-
ism expected of a delocalised electron system, although
the Curie-like component and the small hysteresis
observed at low temperatures show that this cannot be
a complete description of this compound. The replace-
ment of 25% Ni by Co causes a marked increase in both
the susceptibility (Fig. 8b) and the level of hysteresis; the
data plotted in Fig. 9 show a remanent magnetisation of
2
3
not be indexed using the cubic unit cell of the main
phase. Attempts to fit the impurity peaks using a
mixture of possible impurities, including Mo N and Pd,
ꢁ
4
2.72 ꢂ 10 mB per formula unit at 5 K. The suscept-
2
were unsatisfactory. Indexing of these peaks suggested
the presence of a phase with a small orthorhombic unit
cell, similar to that of Co N. A significant improvement
ibility (Fig. 8c) of the 50%—substituted composition
NiCoMo N shows a relatively rapid increase at B80 K,
3
ꢁ4
with a remanent magnetisation of 2.66 ꢂ 10 mB per
2
was made to the fit when 2.29(9)% by weight of a phase
with the Co N structure and composition Pd N was
formula unit at 5 K. However, the susceptibility
(Fig. 8d) of Ni0.5Co1.5Mo N, although comparable in
2
2
3

T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
145
Fig. 7. Observed ( Á ), calculated (-), and difference X-ray powder diffraction profiles for Pd
2
Mo
N and the lower set correspond to Pd
3
N at room temperature; upper tick marks indicate
N. The final structural refinement
positions of allowed reflections from the Ka1 diffraction for Pd
converged with Rwp ¼ 0:1375:
2
Mo
3
2
magnitude to that of NiCoMo N, shows less hysteresis.
3
In contrast, the introduction of Pd causes only a small
change in susceptibility and relatively little hysteresis
transition series. Energy bands formed from the 3d
orbitals of the former will be comparable in width to
those involving 3d Ni orbitals and narrower than those
which have a large contribution from the 4d orbitals of
Pd. The disorder on the transition metal sublattice in the
(Fig. 8e). The two chosen substituents, Co and Pd, come
respectively from the first ð3dÞ and second ð4dÞ

146
T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
Table 3
Atomic positions for Pd
N with the Co N structure
2 2
Atom
Wyckoff position
x
y
z
N
Pd
2a
4g
0
0
0
0.351(4)
0
0.242(5)
Space group
Pmnn (no. 58)
˚
a/A
2.808(1)
4.863(2)
4.4869(9)
˚
b/A
˚
c/A
3
Fig. 9. Magnetisation of Ni1.5Co0.5Mo N as a function of applied field
at 5 K.
causes partial electron localisation and therefore an
increase in the magnetic susceptibility, whereas the 4d
levels of Pd are sufficiently broad to overcome this
effect. However, the replacement of Ni by Co (but not
Pd) changes the number of electrons in the system, and
this factor will also have to be taken into account in
producing a more comprehensive account of the
magnetic behaviour.
5. Conclusion
The principal purpose of this publication is to
demonstrate that the solid-state synthesis of ternary,
low-valent nitrides under reducing conditions can be a
simple and clean procedure. We have prepared samples
of the p-phase Ni Mo N with a higher level of purity
2
3
than was obtained previously when more complex
synthetic procedures were used. We have also used the
method to prepare for the first time pure samples of the
0
0
solid solutions Ni M Mo N (M =Co, 0pxp1;
2
ꢁx
x
3
0
M =Pd, 0pxp1:5) and impure samples of Ni Co
0
.5
1.5
Mo N, p-Co Mo N and Pd N. We have also been able
3
2
3
2
to prepare pure the known Z-phases, Co Mo N and
3
3
Fig. 8. Variation of molar magnetic susceptibility with temperature
for: (a) Ni Mo N, (b) Ni1.5Co0.5Mo N, (c) NiCoMo N, (d) Ni0.5
Co1.5Mo N and (e) Ni1.5Pd0.5Mo N.
Fe Mo N. We hope that in the future the method will
lead to the synthesis of many new nitrides with
interesting properties.
3 3
2
3
3
3
3
3
substituted compositions will favour localisation of the
d electrons, as will the relatively narrow energy width of
the 3d bands. Our observations can thus be explained by
assuming that the disorder introduced by Co-doping
Acknowledgments
We thank EPSRC for financial support.

T.J. Prior, P.D. Battle / Journal of Solid State Chemistry 172 (2003) 138–147
147
References
[13] F. Tessier, R. Marchand, Y. Laurent, J. Eur. Ceram. Soc. 17
1997) 1825–1829.
14] K.S. Weil, P.N. Kumta, J. Grins, J. Solid State Chem. 146 (1999)
2–35.
15] S. Alconchel, F. Sapina, D. Beltran, A. Beltran, J. Mater. Chem.
(1999) 749–755.
16] S.K. Jackson, R.C. Layland, H.-C. zur Loye, J. Alloy. Compd.
91 (1999) 94–101.
(
[
[
[
[
[
[
[
[
[
[
[
1] L.E. Toth, in: Transition Metal Carbides and Nitrides, Refractory
Materials, Vol. 7, Academic Press, New York, 1971.
2
2] S.T. Oyama, The Chemistry of Transition Metal Carbides and
Nitrides, Blackie Academic and Professional, London, 1996.
3] S. Yoshida, Proceedings of the 32nd IEEE Conference on
Electrical Components 1982, pp. 530–535.
9
2
17] P.S. Herle, M.S. Hegde, K. Sooryanarayana, T.N.G. Row,
G.N. Subbanna, Inorg. Chem. 37 (1998) 4128–4130.
18] W. Jeitschko, H. Nowotny, F. Benesovsky, Monatsh. Chem.
4] R.M. Fix, R.G. Gordon, D.M. Hoffman, Chem. Mater. 2 (1990)
235–241.
[
[
5] K. Watari, J. Ceram. Soc. Japan 109 (2001) S7–S16.
6] F.H. Horn, W.T. Zieglar, J. Am. Chem. Soc. 69 (1947)
9
5 (1964) 1212–1218.
19] W. Jeitschko, H. Nowotny, F. Benesovsky, Monatsh. Chem.
4 (1963) 247–251.
2762–2769.
9
[7] J.M.D. Coey, H. Sun, J. Magn. Magn. Mater. 87 (1990) L251–L254.
[8] C.J.H. Jacobsen, Chem. Commun. (2000) 1057–1058.
[9] P.W. Lednor, Catal. Today 15 (1992) 177–337 (Complete issue).
20] A.F. Wells, Three-dimensional Nets and Polyhedra, Wiley, New
York, 1977.
[
[
21] H.M. Rietveld, J. Appl. Cryst. 2 (1969) 65–71.
22] A.C. Larson, R.B. von-Dreele, General Structure Analysis System
[
10] S.H. Elder, F.J. DiSalvo, L. Torpor, A. Navrotsky, Chem. Mater.
(1993) 1545–1553.
11] F.J. DiSalvo, Mater. Sci. Forum 325–6 (2000) 3–10.
5
(
GSAS), Los Alamos National Laboratories, 1990.
23] A. El-Himri, F. Sapina, R. Ibanez, A. Beltran, J. Mater. Chem.
1 (2001) 2311–2314.
[
[
[
12] O. Reckeweg, J.C. Molstad, F.J. DiSalvo, J. Alloy. Compd. 315
1
(2001) 134–142.