ARTICLE IN PRESS
M.G. Johnston, W.T.A. Harrison / Journal of Solid State Chemistry 177 (2004) 4680–4686
4681
stainless steel bomb. The bomb was heated to 140 1C for
4 days, followed by cooling to room temperature over a
few hours. The solid product, consisting of a mass of
brown rods and needles, was recovered from the
supernatant liquors by vacuum filtration and washing
with water and acetone. Powder XRD (see below)
indicated the presence of a small (visual estimate,
based on peak heights, of perhaps 2%) quantity of
MnIIMnI2IIO(SeO3)3. This phase [6] also crystallizes as
brown needles, making phase separation extremely
difficult.
PbFe2(SeO3)4 initially arose from a very surprising
reaction: 2 g MgSO4 ꢀ 7H2O (BDH Analar), 5 g SeO2
(Lancaster Synthesis), and 15 ml deionised H2O were
sealed in a 23 ml hydrothermal bomb and heated to
200 1C for two days, followed by cooling to room
temperature over a few hours. The solid product,
consisting of a tiny yield (o1 mg) of yellow–green,
faceted, chunks of PbFe2(SeO3)4 was recovered by
vacuum filtration and washing with water and acetone.
This reaction is reproducible. No lead or iron com-
pounds had been recently prepared using the hydro-
thermal bomb in question; thus, presumably, impurities
in the starting material(s) reacted to form the PbFe2
(SeO3)4 product. This is supported by trace-elemental
analysis results: the Lancaster SeO2 sample contained
4.24 (9) mg/kg Pb whereas the lead content of the
magnesium nitrate heptahydrate was negligible (below
detection limits). The iron probably arose from the
Analar MgSO4 ꢀ 7H2O [Fe content=9.37 (11) mg/kg]
whereas the Lancaster SeO2 contained negligible Fe (at
or below detection limits).
EDX indicated the presence of Pb, Fe, and Se in the
crystalline chunks, which was confirmed in the crystal-
structure study described below. A rational hydrother-
mal synthesis started from 0.662 g (2 mmol) Pb(NO3)2,
1.616 g (4 mmol) Fe(NO3)3 ꢀ 9H2O, 0.888 g (8 mmol)
SeO2 (BDH Analar with negligible Pb content as
determined by trace element analysis) and 15 ml H2O,
which were sealed in a 23-ml bomb and heated to 200 1C
for 2 days, followed by cooling to room temperature
over a few hours. Product recovery as above led to a
mixture of tiny (o0.02 mm) yellow cuboids of PbFe2
(SeO3)4 accompanied by brown whiskers and needles of
Fe2O(SeO3)4 [7] (identified on the basis of its unit cell
parameters), in about a 50:50 ratio (visual estimate).
(SeO3)6 showed two distinct weight losses (–39.8%
occurring between 350 and 700 1C and –14.4% between
800 and 1200 1C). The overall weight loss of 54.2%
correlates reasonably well with that calculated (55.4%)
for the decomposition scheme Ba2Mn3(SeO3)6 (s)-
Ba2Mn3O6 (s)+6 SeO2 (g), allowing for the presence of
a small quantity of Mn3O(SeO3)3. The identity of the
intermediate decomposition stage at 700 1C is not
known. There was insufficient amount of pure sample
of PbFe2(SeO3)4 to perform similar measurements. The
temperature ranges for the decomposition steps reported
here for Ba2Mn3(SeO3)6 are wide (up to 400 1C), but
similar sluggish decompositions have been seen for other
selenites [1].
The diffuse reflectance spectrum (reference standard
BaSO4) of Ba2Mn3(SeO3)6 was recorded on a Shimadzu
UV-3100 spectrometer at room temperature and con-
verted according to the Kubelka Munk method. It
shows a broad, relatively weak absorbance (ꢀ ꢁ 5)
centered at about 500 nm (20,000 cm–1), which probably
5
corresponds to a nominal Eg-5T2g transition [8] for
Mn3+ (i.e., promotion of an electron from a t2g-like
state to an eg-like state assuming approximate octahe-
dral symmetry). Any spin-forbidden d–d transitions
associated with Mn2+ are expected to be up to two
orders of magnitude weaker [8].
2.2. Structure determinations
The structures of Ba2Mn3(SeO3)6 (brown rod,
ꢂ0.04 Â 0.04 Â 0.37 mm3) and PbFe2(SeO3)4 (yellow–
green chunk, ꢂ0.082 Â 0.045 Â 0.045 mm3) were deter-
mined by standard X-ray methods using data collected
on a Bruker SMART CCD diffractometer (MoKa
(
radiation, l ¼ 0:71073 A; T=2572 1C). In each case, a
hemisphere of data was collected using o scans (0.31
slices) with the aid of the SMART and SAINT
programs [9], and empirical absorption corrections
(min./max. transmission coefficients=0.053/0.515 and
0.181/0.329 for Ba2Mn3(SeO3)6 and PbFe2(SeO3)4,
respectively) was made with SADABS [10] during data
reduction. The systematic absences for Ba2Mn3(SeO3)6
indicated space group P21/c (No. 14) which was
assumed for the remainder of the refinement. For
PbFe2(SeO3)4, the situation is somewhat less clear cut.
¯
Structure solution and refinement in space group P1
(No. 2) resulted (see below) in a generally satisfactory
answer, except for excessively large anisotropy of the
lead atom displacement parameters [Umax/Umin=7.7].
Thus, a model involving positional disorder of the Pb
species over two adjacent sites related by inversion
symmetry was developed [refined d(PbyPbi)=0.449
(3) A; symmetry code (i)=–x, Ày, –z; Umax/Umin=2.9].
A refinement was also performed in space group P1
(No. 1), but this also led to disorder over two adjacent,
nonsymmetry-related sites [d(PbyPb)=0.495 (5) A] and
2.1. Characterization
Powder XRD (Bruker D8 diffractometer, CuKa
(
radiation, l ¼ 1:5418 A; T=201 C) on a well-ground
sample of Ba2Mn3(SeO3)6 yielded a pattern in good
agreement with a simulation based on the single crystal
structure accompanied by a few weak lines correspond-
ing to Mn3O(SeO3)3 [6]. TGA (Mettler Toledo Star
system; ramp at 10 1C/min under argon) for Ba2Mn3