Magnetocaloric effects of ferromagnetic erbium mononitride

Article abstract of DOI:10.1016/j.jallcom.2005.04.061

The rare earth mononitride, ErN, has been synthesized by the carbothermic reduction and hot isostatic press methods. The magnetocaloric effect of ErN has been evaluated by calculating the magnetic entropy changes, ΔS, from the magnetization data sets and from the heat capacity measured at various temperatures and applied fields. The two results are in good agreement with each other. The ΔS value of ErN is the highest at 7.5 K and higher than that of ErNi₂ reported as the candidate material for the magnetic refrigerant of cryogenic technology. Heat capacity curve against temperature of ErN has a peak at 4.4 K at zero-field. The peak value of the heat capacity of ErN is 507 kJ K^-1 m^-3. This value is larger than those of Er₃Ni used commercially as a magnetic regenerator for the Gifford-McMahon refrigerator. The present results indicate that ErN is a promising material for the magnetic refrigerant and regenerator of cooling systems working above the boiling temperature of helium.

Full text of DOI:10.1016/j.jallcom.2005.04.061

Journal of Alloys and Compounds 408–412 (2006) 191–195
Magnetocaloric effects of ferromagnetic erbium mononitride
a,∗
a
a
a
Takashi Nakagawa , Takayuki Arakawa , Kengo Sako , Naoto Tomioka ,
a
b
b
c
Takao A. Yamamoto , Takafumi Kusunose , Koichi Niihara , Koji Kamiya ,
Takenori Numazawac
a
Graduate School of Engineering, Osaka University, Suita, Osaka 565-0871, Japan
b
Institute of Scientific and Industrial Research, Osaka University, Ibaraki 567-0047, Japan
c
Tsukuba Magnet Laboratory, National Institute for Materials Science, Tsukuba, Ibaraki 305-0003, Japan
Available online 13 June 2005
Abstract
The rare earth mononitride, ErN, has been synthesized by the carbothermic reduction and hot isostatic press methods. The magnetocaloric
effect of ErN has been evaluated by calculating the magnetic entropy changes, ꢀS, from the magnetization data sets and from the heat capacity
measured at various temperatures and applied fields. The two results are in good agreement with each other. The ꢀS value of ErN is the
highest at 7.5 K and higher than that of ErNi
capacity curve against temperature of ErN has a peak at 4.4 K at zero-field. The peak value of the heat capacity of ErN is 507 kJ K
This value is larger than those of Er Ni used commercially as a magnetic regenerator for the Gifford–McMahon refrigerator. The present
2
reported as the candidate material for the magnetic refrigerant of cryogenic technology. Heat
−
1
−3
m .
3
results indicate that ErN is a promising material for the magnetic refrigerant and regenerator of cooling systems working above the boiling
temperature of helium.
©
2005 Elsevier B.V. All rights reserved.
Keywords: Erbium; Nitride; Magnetic entropy; Heat capacity; Magnetic refrigerant; Magnetic regenerator
1
. Introduction
earths, ErAl2, HoNi2 or DyxHo1 − xAl2, which have been
proposed as candidates for the magnetic refrigerants for
the application [2–5]. For helium liquefaction, we need
materials with the larger magnetic entropy change at a
temperature slightly higher than the boiling point of helium.
ErN is expected to be an effective magnetic refrigerant for
helium liquefaction, because rare earth atom concentration in
mononitride is larger than that in metal [3] and the magnetic
transition temperature of ErN has been reported as 3.4–6 K
[6–8].
To realize the helium liquefaction, it is also necessary
to get materials with large heat capacity at temperatures
below 10 K for the magnetic regenerator, because the heat
capacity of helium gas has a peak at 8 K [9]. Since the
early 1990s, the development of rare-earth-based magnetic
regenerator such as Er3Ni, having high heat capacity at
temperatures below 10 K, has allowed improvement in the
cooling power of regenerative-cycle cryocoolers at very
low temperatures, permitting the achievement of working
The magnetic refrigeration system provides good energy
efficiency, especially at cryogenic temperatures [1]. It is,
therefore, prospective to apply this system to the helium or
hydrogen liquefaction process. To establish this refrigeration
technology, it is necessary to obtain the efficient and reliable
magnetic refrigerant that possesses a large magnetocaloric
effect (MCE) around and above the boiling points of helium
and hydrogen (4 and 20 K). The MCE is evaluated by the
magnetic entropy change ꢀS induced by demagnetization of
the refrigerant. The value of ꢀS is usually maximized around
the magnetic transition temperature, such as the Curie point.
We have found that HoN, DyN, TbN and GdN show larger
ꢀ
S around their Curie temperatures (18, 21, 44 and 60 K)
than those of the intermetallic compounds containing rare
∗
Corresponding author. Tel.: +81 6 6879 7887; fax: +81 6 6879 7887.
E-mail address: nakagawa@mit.eng.osaka-u.ac.jp (T. Nakagawa).
0
925-8388/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2005.04.061

1
92
T. Nakagawa et al. / Journal of Alloys and Compounds 408–412 (2006) 191–195
refrigeration temperatures in the liquid helium [10]. The
use of rare earth compounds having a heat capacity peak
at a low temperature has made it possible for today’s
Gifford–McMahon (GM) cryocoolers to achieve cooling
powers higher than 1 W at 4.2 K [11]. If materials with
larger heat capacity at temperature below 10 K than Er3Ni
are available, the cooling efficiency of GM cryocoolers is
improved. Heat capacity of a magnetic material usually
shows a peak at its transition temperature. Concentration of
Er per volume in ErN is higher than that in Er3Ni. Therefore,
ErN is also expected as one of the candidate materials
for the magnetic regenerator working around the boiling
temperature of helium, because the concentration of Er per
volume in ErN is higher than that in Er3Ni.
The MCE is also evaluated from heat capacity C(T). Absolute
entropy is expressed by:
ꢀ
T
CH (T)
SH (T) =
dT.
(3)
T
0
In this calculation, CH(T) below 2 K was assumed to be
Debye’s heat capacity, that is, CH(T) was in proportion to
3
T . Therefore, ꢀS is given by:
ꢀS(T) = S0(T) − SH (T)
(4)
The present magnetization data sets M(T, H), and specific
heat data sets C(T, H) were substituted into these equations,
and numerical calculations were carried out to obtain ꢀS(T,
H).
In this paper, we report on magnetic entropy change ꢀS
and heat capacity C of ErN.
3. Results and discussion
2
. Experimental
The XRD pattern of ErN synthesized by the HIP method
is shown in Fig. 1, which indicates that the sample is a single-
phase material of the mononitride with a NaCl-type structure.
As forthe sample prepared by carbothermic reduction, almost
the same profile was obtained. The lattice parameters deter-
mined by the XRD pattern obtained from the sample prepared
by carbothermic reduction is shown in Fig. 2. The lattice
parameters of other rare earth mononitrides (GdN, TbN,
DyN and HoN) are also plotted against atomic number in
Fig. 2 [2,3,6–8,12]. The plots show a fairly good linear rela-
tion, which indicates that these mononitrides share a crystal
structure with the lattice parameter that varies smoothly with
atomic number.
The sample of ErN was prepared by carbothermic reduc-
tion method performed in a nitrogen gas stream. It is quite
the same as that employed in our previous works [2,3]. The
sample of ErN were also synthesized by hot isostatic press
(
HIP) methods, because consolidate sample was needed
for heat capacity measurement. A rectangular-shaped
metal sheet of Er (99.9% purity, 5 mm × 5 mm × 1 mm for
magnetization measurement and 1 mm × 1 mm × 0.1 mm
for heat capacity measurement) was heated at 1873 K for 2 h
under 130 MPa nitrogen atmosphere (99.9999% purity) with
HIP equipment (O2-Dr.HIP; Kobelco Co. Ltd.). To examine
the phase occurring in the product, the XRD pattern was
measured with a diffractometer (RINT Ultima+; Rigaku
Corporation) using Cu K␣ radiation. The magnetization, M,
was measured with a superconducting quantum interference
device magnetometer (MPMS system, Quantum Design Inc.)
under different applied fields H, up to 5 T, and at different
temperatures T, from 100 to 5 K. The specific heat, C, was
measured with Heat Capacity System (Oxford Instruments
Inc.) under 0 and 5 T, and at different temperatures T, from
Fig. 3 shows magnetization M versus T curves of ErN syn-
thesized by carbothermic reduction at different field strengths
H. For the sample prepared by the HIP method, almost the
2
5 to 2 K.
The MCE was evaluated by calculating ꢀS induced by
isothermal demagnetization from H to 0, on the basis of the
equation:
ꢀ
ꢁ
ꢂ
0
∂
S
ꢀS =
dH.
(1)
∂
H
H
T
The integrand is replaced by another expression by applying
Maxwell’s relation, (∂S/∂H)T = (∂M/∂T)H, so that ꢀS is given
by:
ꢀ
ꢁ
ꢂ
0
∂
M
ꢀS =
dH
(2)
∂
T
H
H
Fig. 1. X-ray diffraction patterns of ErN prepared by HIP method.

T. Nakagawa et al. / Journal of Alloys and Compounds 408–412 (2006) 191–195
193
Fig. 4. Magnetic entropy change ꢀS vs. temperature curves at different
applied fields before demagnetization for ErN synthesized by HIP method.
ErN induced by the demagnetization from 5 to 0 T, together
with the plots of GdN, TbN, DyN and HoN [3]. ErN has
the largest ꢀS values among them. The peak temperature of
Fig. 2. Lattice parameters of ErN synthesized by the carbothermic reduction
method and those of other rare earth nitrides.
ꢀS of ErNi2, which is a candidate material for the magnetic
refrigerant, is 6.6 K and the ꢀS value at this temperature
−
1
−3
[13]. The maximum
same results have been obtained. These plots show clearly
order–disorder (ferromagnetic to paramagnetic) transition
with increasing temperature. The magnetic entropy change
has been reported as 253 kJ K
m
ꢀS value of magnetic materials tends to be larger as tran-
sition temperature decreases. In spite of this tendency, the
maximum ꢀS value of ErN is 1.4 times as large as that of
ErNi2. This result indicates that ErN is a promising magnetic
ꢀ
S of ErN synthesized by the HIP method was evaluated by
executing numerical calculation on the basis of Eq. (2), and
the results are shown in Fig. 4. The ꢀS versus temperature
curves are noticed to have peaks around 7.5 K, regardless of
applied field. This temperature well agrees with Curie tem-
perature TC of ErN estimated from Arrott plots as shown
in Fig. 5. In Fig. 6, the ꢀS versus temperature curve of
Fig. 5. Curie temperatures determined from the Arrott plots (denoted as AP
in the legend), and the temperature of the maximum ꢀS (denoted as ꢀS in
the legend) of ErN synthesized by the carbothermic reduction.
Fig. 3. Magnetization vs. temperature curves at different applied fields for
ErN synthesized by the carbothermic reduction.

1
94
T. Nakagawa et al. / Journal of Alloys and Compounds 408–412 (2006) 191–195
Fig. 6. Magnetic entropy change ꢀS induced by demagnetization from 5
to 0 T for ErN synthesized by the carbothermic reduction. For comparison,
plots for other rare earth nitrides are shown together.
Fig. 8. Entropy S vs. temperature curves calculated from heat capacity of
ErN at applied fields of 0 and 5 T.
refrigerant material for the cryogenic system at cryogenic
temperatures.
values of absolute entropy in Fig. 8 by use of Eq. (4). Fig. 9
shows ꢀS of ErN as a function of temperature. The two of the
three curves show the results estimated from magnetization
data sets of ErN synthesized by the carbothermic reduction
and the HIP methods, and the other curve shows the results
estimated from heat capacity measurements. The two curves
estimatedfrommagnetizationdatawellagreewitheachother.
The results calculated from heat capacity measurements are
also in moderate agreement with those obtained from mag-
netization measurements. These consistencies guarantee the
accuracy of the results.
The heat capacity of ErN synthesized by the HIP methods
is given in Fig. 7 as a function of temperature at applied fields
of 0 and 5 T. For comparison, those of He gas (1 MPa), Pb and
intermetallic compounds of Er3Ni and Gd0.5Er0.5Rh [9], used
commercially as magnetic regenerator for GM cryocoolers,
are also shown. This plot clearly indicates that the ErN has
a potential for a regenerator working near the helium boiling
temperature. Fig. 8 shows absolute entropy of ErN at applied
fields of 0 and 5 T calculated by the use of Eq. (3). We can also
calculate the magnetic entropy change ꢀS of ErN from the
Fig. 9. Magnetic entropy change ꢀS vs. temperature curves of ErN calcu-
lated from the data sets of magnetization and heat capacity measurements
when demagnetized from 5 to 0 T.
Fig. 7. Heat capacity vs. temperature curves at applied fields of 0 and 5 T
for ErN synthesized by the HIP method.

T. Nakagawa et al. / Journal of Alloys and Compounds 408–412 (2006) 191–195
195
4
. Summary
References
[
[
[
[
1] V.K. Pecharsky, K.A. Gshneidner Jr, J. Magn. Magn. Mater. 200
Magnetocaloric effect of ErN has been evaluated by
(1999) 44.
calculating the magnetic entropy changes, ꢀS, from the
magnetization data sets and heat capacity measured at
different temperatures and applied fields. The two sets of
results have well agreed with each other. The ꢀS value of
ErN is the highest at 7.5 K and higher than that of ErNi2
reported as one of candidate materials of magnetic refrigerant
for cryogenic refrigerant. The heat capacity curve against
temperature of ErN has a peak at 4.4 K at zero-field. The
2] T. Nakagawa, K. Sako, T. Arakawa, T.A. Yamamoto, J. Alloys
Compd. 364 (2004) 53.
3] T.A. Yamamoto, T. Nakagawa, K. Sako, T. Arakawa, H. Nitani, J.
Alloys Compd. 376 (2004) 17.
4] T. Tomokiyo, H. Yayama, H. Wakabayashi, T. Kuzuhara, T.
Hashimoto, M. Sahashi, K. Inomata, Adv. Cryog. Eng. 32 (1986)
295.
[
5] T. Hashimoto, K. Matsumoto, T. Kurihara, T. Numazawa, A.
Tomokiyo, H. Yayama, T. Goto, S. Todo, M. Sahashi, Adv. Cryog.
Eng. 32 (1986) 279.
−
1
−3
peak value of the heat capacity of ErN is 507 kJ K
m .
[
[
[
[
6] O. Vogt, K. Mattenerger, J. Alloys Compd. 223 (1995) 226.
7] G. Busch, P. Junod, O. Vogt, Phys. Lett. 6 (1963) 79.
8] A. Vendl, J. Nucl. Mater. 79 (1979) 246.
This value is larger than those of Er3Ni used commercially
as a magnetic regenerator for the Gifford–McMahon refrig-
erator. The present results indicate that ErN is a promising
material for a magnetic refrigerant and regenerator for
cooling systems working above the boiling temperature of
helium.
9] T. Hashimoto, R. Li, T. Kuriyama, M. Nagao, J. Cryog. Soc. Jpn.
31 (1996) 131 (in Japanese).
[10] M. Sahashi, Y. Tokai, T. Kuriyama, H. Nakagome, R. Li, M. Ogawa,
T. Hashimoto, J. Adv. Cryog. Eng. 35 (1990) 1175.
11] T. Satoh, A. Onishi, R. Li, H. Asami, Y. Kanazawa, Adv. Cryog.
[
[
[
Eng. 41 (1996) 1631.
12] D.X. Li, T. Haga, H. Shida, T. Suzuki, T.S. Kwon, G. Kido, J. Phys.:
Condens. Matter 9 (1997) 10777.
13] A. Tomokyo, M. Sahashi, K. Inomata, Adv. Cryog. Eng. 32 (1986)
Acknowledgement
The authors thank Emeritus Prof. T. Tabata (Osaka
Preference University) for his critical reading of the
manuscript.
1
95.
[14] H.R. Child, M.K. Wilkinson, J.W. Cable, W.C. Koehler, E.O. Wollan,
Phys. Rev. 131 (1963) 922.