J. Huot et al. / Journal of Alloys and Compounds 280 (1998) 306–309
307
Fe. In such composition, a complete formation of
Mg2FeH6 will leave some unreacted Mg and Fe.
milling, there is evidence of formation of an orthorhombic
metastable g-MgH2 [9]. The origin and condition of
appearance of this metastable phase by ball milling will be
discussed elsewhere. The 10 h of milling diffraction
pattern showed the first appearance of Mg2FeH6 phase
with the MgH2 and Fe phases still present. After 30 h of
milling, the Mg2FeH6 phase is more abundant and the
MgH2 phase has disappeared. Moreover, the Mg peaks are
now present. This is an indication that all hydrogen
initially in the MgH2 phase has been used to synthesized
Mg2FeH6, leaving some unreacted magnesium. This is
confirmed by the fact that further milling up to 60 h does
not significantly change the phase distribution.
A Rietveld refinement was performed on the 60 h
sample. By simple inspection of the powder diffraction
pattern, it is very hard to identify the presence of MgO.
However, the residue of a Rietveld refinement using only
the phases Mg2FeH6, Mg and Fe clearly shows the
presence of diffraction peaks attributed to MgO. This is an
example of the power of Rietveld refinement for minor
phases identification [10].
2. Experimental details
Pure magnesium hydride (98%) was provided by Th.
Goldschmidt. It was mixed with iron powder (Atomet 1001
from Quebec Fer/Titane, 99.5%) inside an argon filled
glove box. An oxygen content of 0.12 wt.% in the iron
powder was measured with a LECO Nitrogen/Oxygen
detector, model TC-136. The milling was carried out with
a Spex mill 8000 and a vial and balls of stainless steel. The
ball to powder weight ratio was 10:1. Milling was carried
out for up to 60 h. At regular intervals, small amount of
powder was taken for analysis. The X-ray powder diffrac-
tion were performed on a Philips X’pert system with Cu
K(radiation. The X-ray diffraction pattern was analyzed by
the Rietveld method using RIETAN software [8]. Lattice
parameters, phase abundance, crystallite size and strain
were extracted from the Rietveld refinement analyses. The
thermal behavior was studied by pressurized differential
scanning calorimeter (PDSC) using a TA 2190 calorimeter
at heating rate of 20 K min21 under 10 bar of hydrogen.
The pressure–composition isotherms and kinetics of hy-
drogen absorption/desorption were measured on a special-
ly designed gas titration apparatus.
To judge the quality of Rietveld refinement, the most
meaningful indicator is the ‘‘R-weighted pattern’’ Rwp
,
which measures the weighted difference between the
calculated and measured intensities. The ‘‘R-expected’’
value Re is an estimation of the minimum value of Rwp
.
The ‘‘R-Bragg’’ RB measures the difference between
calculated and ‘‘experimental’’ intensities of the Bragg
reflections. The mathematical description of these quan-
tities can be found in [11]. A reliable indicator of the
goodness of the fit is the ratio Rwp /Re5S. An S value less
than 1.3 is usually considered satisfactory [11].
3. Results and discussion
A very good fit was obtained by fitting the phases
Mg2FeH6, Mg, Fe, and MgO. After refinement, S51.20
and the Bragg factors for Mg2FeH6, Mg, Fe and MgO
were respectively 2.70%, 2.79%, 1.34% and 3.06%. For
each phase, the phase abundance, lattice parameter, crys-
tallite size and strain determined from Rietveld refinement
is shown in Table 1. The strain value is almost the same
for each phase. Similarly, the crystallite size is of the same
order of magnitude for all phases. Despite the fact that
broad peaks are not ideal for lattice parameter determi-
nation, the value for Mg2FeH6 is surprisingly good, as
shown in Table 2. The value found in this work is in good
3.1. Crystallographic parameters
In Fig. 1 we present the X-ray powder diffraction
patterns as a function of milling time. After only 1 h of
Table 1
Crystallographic parameters deduced from Rietveld refinement of the
X-ray diffraction pattern of the 60 h milled sample
Phase
Weight
%
Lattice parameter
nm
Crystallite size
(nm)
Strain
(%)
Mg2FeH6
Fe
Mg
52(1)
31(1)
10(2)
a50.6444(3)
a50.2874(1)
a50.322(2)
c50.520(4)
a50.423(1)
27.0(6)
19.0(2)
9(1)
1.62(1)
1.13(1)
1.7(1)
Fig. 1. X-ray powder diffraction pattern of 2MgH21Fe mixture as a
function of milling time m, MgH2; j, g MgH2; d, Fe; ♦, Mg2FeH6; .,
Mg; 3, MgO.
MgO
7(1)
12(4)
1 (1)