2782
Jinbo Yang et al
Fe1−xNix · C2O4 · 2H2O (0 6 x 6 0.6) through reaction between appropriate amounts of
aqueous solutions of oxalic acid and sulphates of iron and nickel. The precipitate was then
filtered, washed and dried. The x-ray diffraction patterns of oxalates indicated a single
phase of isostructure. Next the dry powder was put into a quartz tube furnace in a gaseous
flow of a mixture of ammonia and hydrogen to be nitrogenated with an optimized flow ratio
◦
H2/NH3 = 1:3 and temperature T = 500–600 C. Thus iron–nickel nitrides were obtained.
The XRD patterns for different compositions of (Fe1−xNix)4N compounds showed that all
0
compounds are of γ -Fe4N type single phase [6]. M o¨ ssbauer spectra were recorded using
a FH1913 conventional constant-acceleration spectrometer at room temperature (RT) and
7
7 K. A 10 mCi 57Co(Rh) source was used. The spectrometer was calibrated at room
temperature with respect to the standard α-Fe. Magnetic measurements were carried using
a vibrating sample magnetometer at 77 K and RT.
3. Results and discussion
3.1. M o¨ ssbauer spectra
The M o¨ ssbauer spectra of (Fe1−xNix)4N at T = 77 K are shown in figure 1. Since the unit
0
cell of γ -Fe4N has two non-equivalent sites, i.e. the cubic corner site and face centre site
with a ratio of 1:3, the M o¨ ssbauer spectra were unfolded into two components corresponding
0
to two iron sites [7]. Because the easy magnetization direction of γ -Fe4N is parallel to
[
100], the iron atoms at face centre sites were split into two groups with a ratio of 2:1.
0
The M o¨ ssbauer spectra of γ -Fe4N were fitted to three sextets with relative line intensity
c
f
f
Fe :Fe :Fe = 1:2:1 [8]. The component with largest hyperfine field and zero quadrupole
II
I
splitting was assigned to the Fe atoms at corner sites, and other components were assigned
f
f
to Fe and Fe .
I
II
Due to the substitution of iron by nickel in Fe4N, the neighbour configurations of iron
atoms are changed, so a good fit with only three subspectra cannot be obtained. A best fit
of the spectra of (Fe1−xNix)4N was achieved using five subspectra, corresponding to four
groups of six lines for face centre iron and one group of six lines for corner iron. The
−
1
widths of Lorentz lines were all in the range 0.25–0.36 mm s , and the intensity ratios
within each subspectrum (I2/I1 and I3/I1) were kept the same in all fitted subspectra.
c
One can see from figure 1 that the outer spectrum is assigned to Fe and the other
f
f
groups of six-lines are assigned to FeI and Fe . The intensities of outer peaks decrease
II
with increasing nickel content until nearly disappearing at x = 0.5, when corner iron atoms
are completely replaced by nickel atoms.
The hyperfine fields for different samples of (Fe1−xNix)4N compounds are listed in
table 1. The average hyperfine fields Bhf for iron decrease monotonically with nickel
substitution. This is similar to Bhf of the high-nickel-concentration Fe–Ni bulk alloys,
c
which is always the case in the nickel-substituted alloys. It is interesting that Bhf for Fe
increases from x = 0.0 to x = 0.6 in (Fe1−xNix)4N at 77 K.
The iron magnetic moments µFe can be estimated from hyperfine fields by applying a
conversion factor A which is the same for all sites [9],
Bhf = AµFe
(1)
where Bhf is the average hyperfine fields of iron.
In order to calculate iron magnetic moments in (Fe1−xNix)4N (x
6= 0), the correct
conversion factor is to be decided. According to experimental results, the average magnetic
moment per iron atom in Fe4N is 2.21 µB at 0 K and the average hyperfine field is −26.93 T
Yang, Jinbo
