18
T.A. Yamamoto et al. / Journal of Alloys and Compounds 376 (2004) 17–22
has no concern with material degradation due to crystal
structure change.
HoN
In this paper, we report on our finding of other RE
mono-nitrides, HoN and TbN, which exhibit ꢀS much
larger than those of the materials we reported previously.
2. Experimental
Mono-nitride samples of HoN and TbN were synthesized
by the carbothermic reduction method performed in nitro-
gen gas stream. It is quite the same method as that employed
in our previous work [13]. Powders of Ho2O3 or Tb4O7
of 99.99% purity and amorphous carbon were thoroughly
mixed with a mortar and pestle. The amorphous carbon was
charged as twice as the required stoichiometric amount. The
mixed powder was shaped into pellets and heated at 1773 K
in a reaction tube through which 99.9995% purity nitrogen
gas flowed during the reaction for 15 h. The reaction occur-
ring when Ho is nitrided is expressed as follows:
20
40
60
2
80
100
120
/ degree
Fig. 1. Lattice parameters of rare earth mono-nitrides.
The integrand is replaced by another expression by applying
Maxwell’s relation, (∂S/∂H)T = (∂S/∂H)H , so that ꢀS is
given by the following equation:
0 ꢁ
ꢂ
ꢀ
∂M
∂T
ꢀS =
dH
(3)
H
T
Ho2O3 + 6C + N2 → 2HoN + 3CO + 3C
(1)
The present magnetization data sets M(T, H) were substi-
tuted into this equation and numerical calculations were car-
ried out to obtain ꢀS(T, H).
To compare the present experiment with our previous results,
we also synthesized monolithic GdN and DyN again in the
same nitriding batch together with HoN and TbN. To exam-
ine the phase occurring in the product, the powder XRD pat-
tern was measured with a diffractometer (RINT Ultima+;
Rigaku Corporation) using Cu K␣ radiation. The magneti-
zation, M, was measured with a superconducting quantum
interference device magnetometer (MPMS system, QUAN-
TUM DESIGN, Inc.) under various applied fields H, up to
5 T, and at various temperatures T, from 100 to 5 K. The
measurements were carried out in a sequence; the applied
field was swept downward at a constant temperature, and
then the sample was cooled down to the next temperature at
for the measurement was shielded from air and prepared in
an argon-filled glove box connected to the reaction system.
No special treatment was performed to remove the ex-
cess carbon noticed in Eq. (1), so that an amount of free
carbon should have remained in the product. Carbonitride
did not seem to form in our samples because the present
magnetic data sets and lattice parameters reasonably agreed
with those of the mono-nitrides so far reported [18–21]. The
amorphous free carbon is magnetically invisible so that it
contributes only to the weight of each product. However, we
took the weight of carbon into account according to Eq. (1)
in calculating the specific magnetization. The weight of free
carbon was not more than a few weight percent.
3. Results and discussion
The XRD pattern of HoN is shown in Fig. 1, which in-
dicates that the sample is a single-phase material of the
mono-nitride in the NaCl type structure. The pattern of
TbN was of the same quality. The lattice parameters de-
termined for the RE mono-nitrides are plotted against the
atomic number of the involved RE element in Fig. 2, in
which plots for the GdxDy1−xN are shown together [13].
The plots show a fairly good linear relation, which indicates
Tb
Dy
Gd
Ho
0.500
0.495
0.490
0.485
3.2
3.3
3.4
3.5
This work
GdxDy1-xN [13]
Busch et al. [18]
Vogt et al. [19]
Vendl [20]
Li et al. [21]
The MCE was evaluated by calculating the magnetic en-
tropy change ꢀS induced by isothermal demagnetization
from H to 0, which is expressed by
64
65
66
Atomic number
67
A
Fig. 2. Powder XRD patterns of HoN and TbN synthesized by the
carbothermic reduction method. Data for the solid solutions of GdN and
DyN are also given against the average atomic number (open circles (᭺)
except for those at A = 64 and 66).
0 ꢁ
ꢂ
ꢀ
∂S
ꢀS =
dH
(2)
∂H
H
T