ARTICLE IN PRESS
E.N. Dombrovski et al. / Journal of Solid State Chemistry 177 (2004) 312–318
314
2
˚
agreement with the cell parameters determined from
powder X-ray diffraction data. Indexes belonging to the
fluorite subcell and the hexagonal supercell are marked
with the subscript letters f and h, respectively. The
reciprocal lattice vectors of the fluorite subcell and the
hexagonal supercell are related by the matrix:
(ADP) (B19 A ). The refinement of the occupancy
factor of this position gave the value of 0.71(3).
However, in the Ba Ln F structure the F8 atom is
4
3 17
shifted from the ideal (0,0,1/2) position into a general
18f position with an occupancy factor of 1/6. The
refinement of the F8 atoms in 18f position resulted in
even slightly negative ADP. Some fluorine atoms also
had negative ADPs since the correct refinement of the
atomic displacement parameters of such light scatterers
as fluorine is hampered by the presence of Ba and Bi in
the structure and by strongcorrelations between the
ADPs of the anion positions, which were symmetrically
dependent in the fluorite subcell. This circumstance was
overcome by the refinement of all fluorine atoms with
the common ADP. ADPs for Ba and Bi cations were
refined independently in an isotropic approximation.
The final refinement led to good values of reliability
factors R =0.020 and R =0.036. The crystallographic
0
1
C
3
=2 ꢁ1=2 ꢁ1
B
@
ꢁ1=2
2
ꢁ1
3=2 A
2
2
Only –h + k + l=3n reflections were observed, which
%
%
corresponds to possible R3m; R3m; R3 and R3 space
groups. The sample was found to be very sensitive to the
intense electron beam. Under beam irradiation the
superstructure reflections gradually become weaker
and finally disappear, then the sample decomposes. To
avoid decomposition, ED patterns were taken with
widely spread weak electron beam. However, high-
resolution electron microscopy investigation of
Ba Bi F was not possible since the sample decomposes
I
P
parameters, reliability factors, atomic coordinates and
the most relevant interatomic distances for Ba Bi F
3 17
4
are listed in Tables 1–3. The experimental, calculated
and difference X-ray patterns are shown in Fig. 2.
Further details of the crystal structure investigation can
be obtained from the Fachinformationszentrum Karls-
ruhe, 76344 Eggenstein-Leopoldshafen, Germany (Fax:
(49) 7247-808-666; E-mail: crysdata@fiz.karlsruhe.de),
on quotingthe depository number CSD-391210, the
names of the authors and the chemical formula.
The Rietveld refinement from X-ray powder diffrac-
tion data does not allow one to differentiate between
two possible alternatives of arrangement of the F8
atoms in the structure: either these atoms are located at
3b position with an occupancy of 0.71 or at 18f position
with an occupancy of 1/6. The former results in an
overall amount of anions in the structure less than 17
and requires a partial replacement of fluorine by oxygen
to keep the electroneutrality. Oxygen may appear in the
substance because of partial pyrohydrolysis that cannot
be totally excluded. The resultingcomposition of the
phase then corresponds to Ba Bi F16.72O0.14. The latter
4
3 17
under a focused electron beam.
The Ba:Bi ratio was determined by EDX analysis
performed with BaL and BiM lines at 12 points on four
different crystallites. The determined Ba:Bi=
4
.02(5):2.98(5) ratio confirms the proposed Ba Bi F
4 3 17
composition. Thus, the almost monophase stochio-
metric sample of Ba Bi F was obtained when starting
BiF was taken with a small excess in relation to this
4
3 17
3
composition. Probably the excessive BiF was removed
3
from the reaction due to its volatility and possible
reduction.
The crystal structure of Ba Bi F was refined from
4
3 17
X-ray powder diffraction data. Careful analysis of the
raw X-ray powder profile revealed the presence of three
reflections of admixture phases with intensities of 0.4–
1
.2%. The admixture phases were identified as BaF2
˚
d=3.57 A, I/I =0.8%) and metallic Bi (d=3.279 A,
˚
(
I/I =1.2%; d=2.273 A, I/I =0.4%). Their appearance
o
˚
o
o
can be explained by a partial reduction of BiF3 to
metallic Bi, which often leads to a difference between the
initial bulk composition and the final stoichiometry.
Since the admixture reflections have very low intensities
and do not overlap with the reflections of the main
phase, their impact to the diffraction profile was
subtracted usingprofile decomposition prior to the
refinement.
4
3
was proved by means of monocrystals Ba Ln F
3 17
(Ln=Y, Yb) studies [20].
4
Table 1
Selected parameters from Rietveld refinement of X-ray powder data
for Ba Bi
4
3 17
F
The cation and anion positions for the initial structure
model were taken from the Ba Ln F (Ln=Y, Yb) [20]
structure, which has similar cell parameters: Ba Y F —
Space group
˚
a (A)
˚
c (A)
R3%
11.2300(2)
20.7766(5)
4
3 17
4
3 17
Z
Cell volume (A )
3
Calculated density (g/cm )
6
2269.16(8)
˚
˚
a=11.075(1) A,
c=20.372(2) A;
Ba Yb F —a=
4
3
3
17
˚
˚
1.000(1) A, c=20.262(1) A. The R3 space group was
˚
%
1
chosen by analogy with the Ba Ln F compounds. At a
6.583
9p2yp102; 0.01
2y range, step (deg)
Number of reflections
Refinable parameters
4
3 17
281
31
first step the F8 atom was placed at (0,0,1/2) position
3b) with full occupancy, but this resulted in an
abnormally high atomic displacement parameter
(
R
I
, R
P
, RwP
0.020; 0.036; 0.047