O.L. Sologub et al. / Journal of Alloys and Compounds 337 (2002) 10–17
11
simultaneously measured with the following analytical
condition: 20 kV–20 nA for Ce La, Pt La and B Ka. The
can be explained by the fact that the surface state of the
sample and the standard can be slightly different (oxide
surface layer, conductive carbon layer), the B–Ka line
being easily absorbed by the surface layer and leads to an
overestimation of the boron content.
standards used for the calibration were CeAl for Ce, and
2
the pure elements for Pt and B. Typical beam sizes were
3
around 100 nm and the analyzed volume 1 mm . Back-
scattered electrons images (BSE) were recorded to evalu-
ate the homogeneity of the samples. The correction model
used for the quantification was the X-Phi model instead of
the PAP model because the X-Phi model is more accurate
for the quantification of light elements.
The laboratory X-ray powder diffraction data were
collected with a Bruker D8 diffractometer from a flat
rotating sample using Cu Ka radiation in the range 10–
4
. Structural investigations for the CePt B compound
3
4
.1. Ab initio structure solution from powder diffraction
Indexing the conventional X-ray powder diffraction data
of CePt B using the CRYSFIRE program package [10–14]
3
1
208 with a step width of 0.028 and a constant counting
suggested a tetragonal cell. A synchrotron powder pattern
was recorded to obtain better resolution. Use of the
synchrotron data (23 reflections, 2umax5408) increased the
figures of merit giving lattice parameters a54.0051(7),
c55.0760(8) A. Careful examination of the data and
application of a pattern decomposition method with cell
constraints (Le Bail method [15] by iterating the Rietveld
time of 15 s per step. Synchrotron data were collected at
KEK-PF BL20B [9] in an evacuated Debye–Scherrer
camera using a 0.3-mm diameter capillary and image
˚
plates spanning the range 5–858 at a wavelength of 0.75 A
˚
and effective step size of 0.018.
Magnetic measurements were performed in the tempera-
ture range 2–300 K using a MPMS SQUID (Quantum
Design) system.
XAS (X-ray absorption spectroscopy) was performed at
the French synchrotron radiation facility (LURE) in Orsay
using the X-ray beam of the DCI storage ring (working at
[16] decomposition formula) were applied; the structure
factors were extracted from the powder pattern with the
help of the PCW 2.3 program [17] and choosing the P4mm
space group [18].
The SHELXS-86 program [19] was used to locate the
heavy atom sites (Ce and Pt). After refining these heavy
atom coordinates using SHELXL-97 program package
1
.85 GeV and |320 mA) on the EXAFS D21 station. A
double Si 311 crystal was used as a monochromator. The
rejection of third-order harmonics was achieved with the
help of two parallel mirrors adjusted to cut off energies
higher than |9 keV. Experiments were carried out in the
[20], the boron site was located in the Fourier difference
synthesis. Using the extracted structure factors, the
SHELXL-97 refinement with complete structural model
gave R512%.
energy range 5640–5820 eV, which contains the L -edge
III
of Ce. Finely powdered samples were spread on an
adhesive Kapton tape and four such tapes were stacked
together for preparing a sample layer of thickness suffi-
cient enough to ensure a good signal. Moreover, this was
also helpful in eliminating, to a good extent, any sample-
free regions in the path of the radiation. X-ray absorption
spectra were measured at two fixed temperatures, 300 and
4.2. Rietveld refinement for CePt B samples from X-ray
powder diffraction
3
The structure of CePt B was refined from X-ray powder
3
data collected from two samples (a and b) using the atomic
parameters obtained above. Both samples contained small
amounts of CePt B and CePt B impurities with the most
1
0 K.
2
4
intense reflections occurring in the range 378,2u,418.
From the amplitude of the strongest impurity lines as
3
. Phase investigations for the CePt B compound
compared to the 100% intense line of CePt B, we can
3
3
estimate an impurity content of about 3–5%. Rietveld
refinement was performed using FULLPROF98 [21,22].
An experimentally determined Ka /Ka ratio of 0.5,
The BSE image reveals a biphasic material with very
fine elongated precipitates. The dimensions of the precipi-
tates were around 1 mm large and 10 mm long. For this
reason, the quantification is not easy and the beam volume
interaction must contain the matrix and the precipitate
leading to an inaccurate quantification of the precipitates.
Indeed, the composition of the matrix is homogeneous,
while the composition of the precipitates shows a tendency
to an increase of the platinum content. The average
composition of the matrix (after renormalization to one
cerium) is CePt2.97(9)B1.38(15). It can be noticed that the
boron content is higher than the nominal composition. This
1
2
cos u50.7998 for the monochromator polarization correc-
tion and a pseudo-Voigt profile shape function were used.
The background was refined with a polynomial function.
In the final cycles of the least-squares refinements, the
refined positional parameters and isotropic thermal param-
eters for the atoms were allowed to vary while the
occupancy parameter for the boron atom was held con-
stant. Further details on data collection and Rietveld
refinements are given in Table 1.
Experimental, calculated and difference X-ray diffrac-