B Lasse 1975
B Lasse 1975
B Lasse 1975
Abstraet--I.R. and Raman spectra are reported for the following three systems: Ba2CaMo~ xTe~O~,
Ba2_~Sr~MgWO~and Ba2Ca~_~Mg~WO~.In the first series the internal vibrations of the M6+O~octahedra do not
influence each other. The intensity of v~ (MoO~)is five times that of v~ (TeO~). In the second system there is a
gradual, but no linear variation with the composition parameter x. In the third system selection rules are lifted and
new bands appear. Where possible these changes in the spectra are related to the structure of the compositions.
1~47
1348 G. BLASSE
Ba2CaMoO6
vl Ba2CaMoO6
.o_
V5
V4
v3 v2
r"
J\ 8 0
I
600
i i
400
v1(Te) .._i-~m~. cm-1
Fig. 2. I.R. spectrum of Ba2CaMoO6 in the Mo-O stretching
i o Ba2CaM°olTeo.906v5 frequencyregion.
3o/
maxima as a function of x and the observed bands are
v2(Te) broader for the solid solutions than for the end members
i i ,
860 660 400 cm-' ?
Fig. 1. Raman spectra of Ba=CaMo06and Ba2CaMoo1Teo.906.
2C
T~,). The spectrum shows, however, that the threefold
degeneracy has been removed completely. Its Raman
spectrum (Fig. I) shows also that the deviation from Oh B.10
symmetry must be considerable, since components of the
v3 and v4 modes (both Ttu and, therefore, forbidden in the
Raman spectrum) are observed in the Raman spectrum.
This is not the case for the corresponding tungstate and 8 4 0 ~ x B.O0
tellurate [1]. Note that this assignment differs slightly from
Ref. [1] and agrees with Ref. [4].
The marked distortion of the MoO6 octahedron in
Ba2CaMoO6 may be due to the fact that the Mo6+ ion is
somewhat too small to fit an octahedral oxygen ion 82(2 790
coordination. This in turn may be related to the
non-existence of a number of molybdates with ordered
perovskite structure, whereas the corresponding tung-
states and tellurates are known[8, 9]. 80C ' 780
O 1 2
(b) The system Ba2-xSrxMgWO6 ----~D-- X
This series also forms a solid solution series for all Fig. 3. Lattice parameter a, position of v, frequency and half
values of x. There is a slight deviation from Vegard's rule width of m, ~v~I2,in the systemBa2-xSrxMgWOt.
Vibrational spectra of solid solutionseries 1349
(especially in the Raman spectra). The data for v~ in the shows a more extensive splitting than the corresponding
Raman spectra are given in Fig. 3 as an illustrative band for Ba2MgWO6 indicating a larger deviation from
example. We note that the curve of Av~/2vs x is strongly cubic site symmetry for the W6÷ ion in SrMgWO6. We
asymmetric due to the large difference in width of the v, further observed that v2 appears weakly in the Raman
peak in the Raman spectra of Ba2MgWO6and Sr2MgWO6. spectrum of SrzMgWO6(at 590 cm-~), whereas it is absent
This may indicate a larger deviation from complete in the Ba compound. This value agrees with those
crystallographic order in Ba2MgWO6 than in observed by Lentz[4] for other tungstates and may
Sr2MgWO6[10]. The increase of Av~/2for mixed occupa- indicate that the previous assignment of the extremely
tion of the A sites substantiates the arguments used in weak 680 cm 1 band in the Raman spectrum of
Ref.[10] to propose a deviation from complete order in Ba2MgWO6[1] is incorrect. The different intensity of v2
these compounds. indicates that the bonding in the WO6 octahedra of the Sr
It is further noteworthy that the position of the internal and the Ba compound is slightly different.
WO6 vibrational bands is stronger influenced by replacing
Ba in Ba2MgWO6by Sr than by replacing Sr in Sr2MgWO6 (c) The system Ba2Ca~-,Mg, WO6
by Ba (see Fig. 3). The experimental results suggest that This series shows a broad miscibility gap from x is
the non-linearity of the v~ position vs x is connected with about 0.3 to about 0.8. This is probably due to the
the non-linearity of the lattice parameter vs x (Fig. 3). A difference between the ionic radii of Mg~+ and Ca2÷. We
possible explanation may be the following. The Ba note, however, that this broad miscibility gap forms an
compound shows a smaller deviation from cubic sym- exception to the rule of thumb that the (linear) lattice
metry than the Sr compound (see also below). This parameters of isomorphous compounds must differ less
suggests that in the Sr compound there is a larger amount than 5 per cent to have complete miscibility[12]. In the
of cooperative buckling of the corner-shared octahedra present case this difference is only 3.5 per cent. Note that
which enlarges the unit cell[8, 11]. The deviation from Ba2 xSrx MgWO6 where this value is 3 per cent does not
Vegard's rule in the Sr-rich region of the system can then show a miscibility gap. Of all systems described in this
be ascribed to the occurrence of buckling. Introduction of paper the single-phase compositions Ba2Cal-xMg~WO6
Sr in the Ba compound will cause in first approximation a show the strongest influence of a mixed occupation of one
shortening of bond lengths, resulting in an increasing of the crystallographic sites on their spectra.
value of a.o. the v~frequency. At higher Sr concentrations Figure 5 shows Raman spectra of Mg-rich compositions
there will also occur buckling of the octahedra as certainly and Fig. 6 those of Ca-rich compositions. In Fig. 7 the i.r.
is present in the Sr compound. This may influence the v3 band of the end members Ba2MgWO6and Ba2CaWO6is
internal WO6 frequencies less than bond length shorten- compared
ing, resulting in a curve of the type observed experimen- First we consider the magnesium-rich single-phase
tally. region. Inspection of Fig. 5 learns that the replacement of
Finally we compare the spectra of the Sr and the Ba end
members of this series. In Fig. 4 the i.r.-active v3 has been
drawn for both compounds. The v3 band for Sr2MgWO6
ffl
g
5 yk_
%
V1 V 3
i
BOO 700
i i
600
i
500
- - - ~ m l , . ca-1
~
700
i
600 500
.• Ba2MgWO6
Fig. 4. I.R. spectra of Sr2MgWO6 and Ba2MgWO~in the W-O a6o ' 660 4()O :m -1
stretchingfrequency region. Fig. 5. Raman spectra of Ba2MgWO6and Ba2Mgo.sCao.2WO6.
1350 G. BLASSE
Ba2CaWO6
I
_L A WO6 group with a mixed surroundings does not
posses inversion symmetry as the regular WO6 group in
Ba2MgWO6. Consequently the exclusion principle is not
strictly valid. We therefore assign the weak Raman peaks
800 6C)0 ' 4()0 cm 4
in the region 700-750 cm -~ to components of the v~ mode
Fig. 6. Raman spectra of Ba2CaWO6and Ba2Cao.7~Mg,o 25WO6.
of WO6 groups with mixed surroundings and the weak i.r.
peak around 810 cm -~ to the v~ mode of these groups.
The changes in the Ca-rich single-phase region are less
pronounced. This is shown in Fig. 6 for the Raman
8a2MgWO6 BaaCaWO6
spectra. The v~ peak shifts only slightly to higher
wavenumbers upon introduction of Mg in Ba2CaWOr. At
the same time a new band appears at about 700 cm -1 (see
Fig. 6). In the i.r. spectrum the v3 band shows a slight
broadening. No new bands occur. The Raman band at 700
cm -1 is assigned to v3 due to the absence of inversion
t-
.0
~ V3 symmetry for WO6 octahedra with mixed surroundings.
In Fig. 7 the v3 bands of Ba2MgWO6 and Ba:CaWO6 are
compared. It is clear that the splitting in the case of the
calcium compound is more pronounced than for the
magnesium compound. The former compound will there-
fore show a larger deviation from cubic site symmetry for
\ tungsten. The result of the replacement of Ca in
700
I 600
I I
500 700
I I
600 500
influence of Mg upon the vibrational spectra of
Ba2CaWO6 in contrast with the influence of Ca upon the
Cl'~"1 spectra of Ba2MgWO6 where bond distances, not bond
angles are influenced. This argument runs parallel with
Fig. 7. I.R. spectra of Ba2MgWO6and Ba~CaWO6in the W-O considerations used above for the system
stretchingfrequency region. Ba2-x Srx MgWOr.
8. See, e . g . J . B . Goodenough and J. M. Longo, Landolt 10. J. H. G. Bode and A. B. van Oosterhout, to be published.
Bi~rnstein (Edited by K. H. Hellwege), Group III, Vol. 4a, p. 11. H. D. Megaw, Trans. Faraday Soc. A42, 224 (1946).
275. Springer Verlag (Berlin) (1970). 12. See e.g.H. Remy, Lehrbuch derAnorganische Chimie, Band I,
9. G. J. Dirksen, unpublished results from this laboratory. p. 335. Akad. Verlagsgesellschaft, Leipzig (1960).