Journal of Magnetism and Magnetic Materials 200 (1999) 44}56

Magnetocaloric e!ect and magnetic refrigeration

Vitalij K. Pecharsky, Karl A. Gschneidner Jr.*
Ames Laboratory and Department of Materials Science and Engineering, Iowa State University, 255 Spedding, Ames, IA 50011-3020, USA
Received 21 January 1999; received in revised form 9 April 1999
Abstract
The phenomenon of the magnetocaloric e!ect along with recent progress and the future needs in both the characteriza-
tion and exploration of new magnetic refrigerant materials with respect to their magnetocaloric properties are discussed.
Also the recent progress in magnetic refrigerator design is reviewed. 1999 Elsevier Science B.V. All rights reserved.
PACS: 65.40.#g; 65.50.#m; 75.30.Sg; 75.50.Cc
Keywords: Magnetocaloric e!ect; Magnetic refrigeration; Lanthanide materials; Ferromagnets; Paramagnets; Adiabatic demagneti-
zation
1. Magnetocaloric e4ect
1.1. Discovery and fundamentals
Magnetocaloric e!ect (MCE), or adiabatic tem-
perature change (

), which is detected as the

heating or the cooling of magnetic materials due to
a varying magnetic "eld, was originally discovered
in iron by Warburg [1]. The nature of the MCE
was explained and its practical use to reach ultra-
low temperatures in a process known as adiabatic
demagnetization was suggested independently by
Debye [2] and by Giauque [3].
The MCE is intrinsic to all magnetic materials
and is due to the coupling of the magnetic sub-
lattice with the magnetic "eld, which changes the
magnetic part of the entropy of a solid. Just as the
*Corresponding author. Tel.: #1-515-294-7931; fax: #1-
515-294-9579.
E-mail address: [email protected] (K.A. Gschneidner Jr.)
compression of a gas, the isothermal magnetizing of
a paramagnet or a soft ferromagnet reduces the
entropy and, in a reversible process, demagnetizing
(which is similar to the expansion of a gas) restores
the zero-"eld magnetic entropy of a system. The
thermodynamics of the MCE in a ferromagnet near
its magnetic ordering temperature (Curie temper-
ature,
'
) is illustrated schematically in Fig. 1. At
constant pressure the entropy of a magnetic solid,
S(, H), which is a function of both the magnetic
"eld strength (H) and the absolute temperature (),
is the combined total of the magnetic, S
`
, lattice,
S
'
, and electronic, S
'
contributions:
S(, H)"S
`
(, H)#S
'
()#S
'
(). (1)
It is shown for a ferromagnetic material in two
constant magnetic "elds (zero magnetic "eld, H
"
,
and a non-zero magnetic "eld, H

), together with
the corresponding magnetic and non-magnetic
terms. When the magnetic "eld is applied adiabati-
cally (i.e. when the total entropy of the system
0304-8853/99/$- see front matter 1999 Elsevier Science B.V. All rights reserved.
PII: S 0 3 0 4 - 8 8 5 3 ( 9 9 ) 0 0 3 9 7 - 2
Fig. 1. The S}T diagram illustrating the existence of the mag-
netocaloric e!ect. The solid lines represent the total entropy in
two di!erent magnetic "elds: H
"
"0 and H

'0. The horizon-

tal arrow shows

and the vertical arrow shows S

`
when
the magnetic "eld is changed from H
"
to H

. The dotted line

shows the combined lattice and electronic (non-magnetic) en-
tropy, and dashed lines show the magnetic entropy in the two
"elds. S
"
and
"
are zero "eld entropy and temperature, S

and

are entropy and temperature at the elevated magnetic

"eld H

.
remains constant during the magnetic "eld change)
in a reversible process, the magnetocaloric e!ect
(i.e. the adiabatic temperature rise,

"

!
"
) can be visualized as the isentropic di!er-
ence between the corresponding S()
&
functions as
shown in Fig. 1 by the horizontal arrow. The MCE
can be also expressed by means of the isothermal
magnetic entropy change (or simply magnetic en-
tropy change), S
`
"S

!S
"
, when the magnetic
"eld is applied isothermally. In the latter case it is
equal to the isothermal di!erence between the cor-
responding S()
&
functions as shown in Fig. 1 by
the vertical arrow. Therefore

and S
`
repres-
ent the two quantitative characteristics of the mag-
netocaloric e!ect, and it is obvious that both

and S
`
are functions of the initial temperature,

"
(i.e. the temperature before the magnetic "eld
was altered), and the magnetic "eld change,
H"H

!H
"
). It is easy to see (Fig. 1) that if
raising the magnetic "eld increases magnetic order
(i.e. reduces magnetic entropy, which is the case for
simple paramagnetic and ferromagnetic materials),
then

(, H) is positive and magnetic solid

heats up, while S
`
(, H) is negative. The signs of

(, !H) and S
`
(, !H) are corre-
spondingly reversed when the magnetic "eld is re-
duced.
The

and S
`
are correlated with the mag-
netization (M), the magnetic "eld strength, the heat
capacity at constant pressure (C), and the absolute
temperature by one of the fundamental Maxwell's
relations [4]

RS(, H)
RH
2
"

RM(, H)
R
&
, (2)
which for an isothermal}isobaric process after in-
tegration yields
S
`
(, H)"

&`
&

RM(, H)
R
&
dH. (3)
Eq. (3) indicates that the magnetic entropy change
is proportional to the derivative of magnetization
with respect to temperature at constant "eld and to
the magnetic "eld change. By combining Eq. (2)
with the corresponding dS equation it easy to
show [4] that the in"nitesimal adiabatic temper-
ature rise for the reversible adiabatic}isobaric pro-
cess is equal to
d"!

C(, H)
&

RM(, H)
R
&
dH. (4)
Hence, the adiabatic temperature rise is directly
proportional to the absolute temperature, to the
derivative of magnetization with respect to temper-
ature at constant "eld and to the magnetic "eld
change; and it is inversely proportional to the heat
capacity. After integrating Eq. (4) one gets the value
of the magnetocaloric e!ect as

(, H)
"!

&`
&

C(, H)
&

RM(, H)
R
&
dH. (5)
Eqs. (2)}(5) have a fundamental importance on
the understanding of the behavior of the MCE in
solids, and serve as a guide for the search of new
materials with a large magnetocaloric e!ect. First,
since the magnetization at constant "eld of para-
magnets and simple ferromagnets decreases with
V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56 45
increasing temperature [i.e. (RM/R)
&
(0], then
S
`
()
&
should be negative (Eqs. (2) and (3)),
while

()
&
should be positive (Eqs. (4) and
(5)), which agrees with Fig. 1. Second, in ferromag-
nets "(RM/R)
&
" is the largest at the
'
, and there-
fore, "S
`
()
&
" should peak at
'
(Eqs. (2) and (3)).
Third, although it is not straightforward from Eqs.
(4) and (5) because the heat capacity at constant
"eld is also anomalous near the
'
, it has been
shown [5] that

()
&
in ferromagnets peaks at
the Curie temperature when HP0. The behavior
of

()
&
should be similar to the behavior of
"S
`
()
&
", i.e. it will be gradually reduced both
below and above the
'
. Fourth, for the same
"S
`
()
&
", the

()
&
will be larger at a higher
absolute temperature, and also when the total heat
capacity of the solid is lower (Eq. (5)). The latter
point is critical in understanding the fact that para-
magnets display signi"cant

(, H) only at
temperatures close to absolute zero, where the lim-
ited value of "(RM/R)
&
" is easily o!set by the negli-
gible lattice heat capacity. Furthermore, at high
temperatures the measurable adiabatic heating (or
cooling) is expected only if the solid orders spon-
taneously, i.e. when value of "(RM/R)
&
" becomes
signi"cant.
1.2. Measurement of the magnetocaloric ewect
The magnetocaloric e!ect can be measured dir-
ectly or it can be calculated indirectly from the
measured magnetization or "eld dependence of the
heat capacity. Direct techniques always involve
measurements of the sample temperatures (
"
and

'
) in magnetic "elds H
"
and H
'
, where subscripts
0 and F designate initial and "nal magnetic "eld,
respectively. The

()
&
is then determined as
the di!erence between
'
and
"
for a given

"
and H"H
'
!H
"
.
Direct MCE measurements can be carried out
using contact (i.e. when the temperature sensor is in
direct thermal contact with the sample) and non-
contact techniques (i.e. when the sample temper-
ature is measured without the sensor being directly
connected to the sample) [6}14]. Since during the
direct MCE measurements a rapid change of the
magnetic "eld is required, the measurements can be
carried out on immobilized samples when the mag-
netic "eld change is provided either by charg-
ing/discharging the magnet, or by moving the
sample in and out of a uniform magnetic "eld
volume. Using immobilized samples and pulse
magnetic "elds direct MCE measurements in mag-
netic "elds from 1 to 40 T have been reported. The
use of electromagnets usually limits the magnetic
"eld strength to less than 2 T. Experimental appar-
ati, where the sample or the magnet are moved to
provide the varying magnetic "eld environment
usually employ superconducting or permanent
magnets, which limit the magnetic "eld range to
0.1}10 T. The accuracy of the direct experimental
techniques depends on the errors in thermometry,
errors in "eld setting, the quality of thermal insula-
tion of the sample (this becomes a critical source of
error when the MCE is large and thus disrupts the
adiabatic conditions), the quality of the compensa-
tion scheme to eliminate the e!ect of the changing
magnetic "eld on the temperature sensor reading.
Considering all these e!ects the accuracy is claimed
to be in the 5}10% range [6}14].
Unlike the direct MCE measurements, which
only yield the adiabatic temperature change, the
indirect experiments allow the calculation of both

()
&
and S
`
()
&
*heat capacity measure-
ments, or just S
`
()
&
* magnetization measure-
ments. Magnetization measured experimentally as
a function of temperature and magnetic "eld pro-
vides S
`
()
&
after numerical integration of
Eq. (3), and has been rightfully suggested as a useful
technique for the rapid screening of prospective
magnetic refrigerant materials [15]. The accuracy
of S
`
()
&
calculated from magnetization data
depends on the accuracy of the magnetic moment,
temperature, and magnetic "eld measurements. Be-
cause numerical integration is involved, and be-
cause the exact di!erentials (dM, d, and dH) are
substituted, respectively, by the measured M, ,
and H, the typical accuracy of S
`
()
&
from
magnetization measurements is reported to be in
the range of 3}10% [15]. The relative error may
become signi"cantly larger particularly for small
"S
`
()
&
" values.
The heat capacity measured as a function of
temperature in constant magnetic "elds, C()
&
,
provides the most complete characterization of mag-
netic materials with respect to their magnetocaloric
46 V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56
e!ect, since the entropy of magnetic solid can be
calculated from heat capacity as
S()
&"
"

2
"
C()
"

d#S
"
and
S()
&$"
"

2
"
C()
&

d#S
"&
, (6)
where S
"
and S
"&
are the zero temperature en-
tropies. In a condensed system these are the same
(i.e. S
"
"S
"&
) [16], and therefore, the calculation
of both

()
&
and S
`
()
&
is straightforward
[17], also see Fig. 1. The accuracy of the MCE
calculations using heat capacity data is critically
dependent on the accuracy of the heat capacity
measurements and data processing (Eq. (6)) because
both

()
&
and S
`
()
&
are small di!erences
between two large quantities (temperatures and
total entropies). The error in S
`
()
&
from heat
capacity is given [17] as
S
`
()
&
"S()
&"
#S()
&$"
, (7)
where S()
&"
and S()
&$"
are the errors in the
zero magnetic "eld entropy and non-zero magnetic
"eld entropy, respectively. The error in

()
&
is also proportional to the errors in the entropy, but
it is inversely proportional to the derivative of the
entropy with respect to temperature [17]

()
&
"S()
&"
/(dS()
&"
/d)
#S()
&$"
/(dS()
&$"
/d). (8)
It should be noted that Eqs. (7) and (8) yield the
absolute uncertainty in the calculated mag-
netocaloric e!ect, and therefore, the relative errors
rise signi"cantly for small MCE values (Figs. 2
and 3). Hence, assuming that the accuracy of the
heat capacity measurements is magnetic "eld inde-
pendent, the relative error in both

()
&
and
S
`
()
&
will be reduced for larger H.
1.3. Magnetocaloric ewect in paramagnets
As mentioned above, the MCE in paramagnets is
only expected to be measurable at temperatures
close to absolute zero, where the enhanced, but still
limited "(RM/R)
&
" is easily o!set by the negligible
Fig. 2. The S
`
in Gd for a magnetic "eld change from 0 to 5 T
calculated from the experimental heat capacity data measured in
0 and 5 T magnetic "elds (open circles, left-hand scale). The
dotted lines (left-hand scale) indicate the range of absolute errors
and the solid line shows the relative error (right-hand scale).
Fig. 3. The

in Gd for a magnetic "eld change from 0 to 5 T

calculated from the experimental heat capacity data measured in
0 and 5 T magnetic "elds (open circles, left-hand scale). The
dotted lines (left-hand scale) indicate the range of absolute errors
and the solid line shows the relative error (right-hand scale).
lattice heat capacity of a solid. Early research on
the magnetocaloric e!ect in paramagnets was car-
ried out because of the drive to reach ultra-low
temperature by adiabatic demagnetization cooling.
It is worth noting that the pioneering work of
Giauque and MacDougall, who studied low-tem-
perature magnetocaloric properties of paramagnetic
V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56 47
Gd
`
(SO
"
)
`
) 8H
`
O, showed that a temperature of
less than 1 K could be reached [3,18]. This work
and other important contributions to the low-tem-
perature behaviors of solids resulted in Giaugue
winning the 1949 Nobel prize in chemistry. Many
years later the low-temperature magnetocaloric ef-
fect was reported in a variety of paramagnetic salts
including ferric ammonium alum [19], chromic po-
tassium alum [20], and cerous magnesium nitrate
[21].
The low thermal conductivity of paramagnetic
salts is detrimental for adiabatic demagnetization
applications, and therefore, paramagnetic inter-
metallic compounds have attracted some attention
with respect to their magnetocaloric properties.
One of the most studied materials was PrNi
`
[22}32], which is still successfully used in nuclear
adiabatic demagnetization devices. In conjunction
with Cu as the low-temperature nuclear magnetic
demagnetization stage and PrNi
`
as the upper
stage a record was set for the lowest working tem-
perature (27 K) at which useful experiments could
be performed on materials, other than on the refrig-
erant itself [29]. Recent experimental and theoret-
ical data [32] show that at higher temperatures
(&3 K((&13 K) the magnetocaloric e!ect
in paramagnetic PrNi
`
is anomalous due to the
crystalline electric "eld e!ects, which cause the
paramagnetic system to cool upon magnetizing and
to warm upon demagnetizing.
A considerable e!ort has been devoted to the
studies of the MCE in paramagnetic garnets mainly
because of their high thermal conductivity, low
lattice heat capacity and very low ordering temper-
ature (generally below 1 K). With an ordering tem-
perature close to absolute zero it is possible to
obtain a large S
`
to maintain a signi"cant MCE
up to &20 K. Clark and Alben [33] report

approaching 6}10 K for H"11 T in ytterbium

and gadolinium iron garnets between 10 and 30 K.
Noticeable magnetocaloric e!ects were also re-
ported in neodymium gallium garnet (Nd
`
Ga
`
O
`
)
at 4.2 K [34], and in gadolinium gallium garnet
(Gd
`
Ga
`
O
`
) below 15 K[35,36]. A large enhance-
ment of the S
`
in magnetic nanocomposites based
on iron-substituted gadolinium gallium garnets
Gd
`
Ga
`V
Fe
V
O
`
(x)2.5) was recently observed
by Shull and co-workers [35,37,38].
1.4. Magnetocaloric ewect and order}disorder
magnetic phase transition
Spontaneous magnetic ordering of paramagnetic
solids upon lowering the temperature is a
cooperative phenomenon, which occurs at various
temperatures depending both on the nature of the
magnetic sublattice and on the strength of ex-
change interaction. When spontaneous magnetic
ordering occurs, the magnetic order parameter as
well as the bulk magnetization of a solid undergo
large changes in relatively narrow temperature in-
terval close to the respective Curie or NeH el temper-
ature contributing to a large "(RM/R)
&
" and thus
making it possible for a practical utilization of
the MCE (see Eqs. (2)}(5)). Even though it is not the
absolute magnetization, but the derivative of
the magnetization with respect to temperature,
which must be large to yield a large MCE, the 4f
metals (lanthanides) and their alloys have been
studied much more extensively than 3d metals
and their alloys simply because the available
theoretical magnetic entropy in the former is con-
siderably larger than in the latter. Most of the
research on the MCE has been associated with
either soft ferromagnetic materials ordering
from&4 to &77 Kfor applications such as helium
and hydrogen liquefaction, or materials ordering
near room temperature for applications such as
conventional air conditioning and refrigeration.
The scope and the size of this article do not allow
us to review all of the available data, and therefore
we will brie#y discuss only the most interesting
results.
1.4.1. Low temperatures: MCE in the range
&10}80 K
One of the most natural choices for low-temper-
ature magnetic refrigerant materials are the pure
lanthanide metals which order at low temperatures,
such as Pr, Nd, Er, and Tm. Of these four, the MCE
was studied for all but Pr (which does not order
magnetically) and the expectations for large MCE
values were not realized. Zimm et al. [39] report
that the MCE in Nd reached a

of &2.5 K at
"10 K for H"7 T. The presence of several
magnetic phase transitions, which occur in Er be-
tween &20 and &80 K brings about an almost
48 V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56
constant but quite small MCE (averaging between
4 and 5 K for a 7 T magnetic "eld change) in the
20}90 K temperature range [40]. Pure Tm orders
magnetically in a sinusoidally modulated ferromag-
netic structure at &56 K and becomes ferrimag-
netic at &32 K. These features of the Tm magnetic
structure restrict its MCE, which barely tops &3 K
at &56 K for H"7 T [41]. It has been also
observed that the MCE in Tm becomes negative
between &32 and &56 K in a weak magnetic "eld
of 1 T [41]. The main reason for these low

values is that most of magnetic phases in Nd, Er

and Tm are antiferromagnetic and ferrimagnetic,
and much of the entropy is involved in #ipping
spins to a ferromagnetic alignment.
The materials which display the largest MCE
between &10 and &80 K are intermetallic com-
pounds containing lanthanide metals. Among them
the best are RAl
`
, where R"Er, Ho, Dy and
Dy
"`
Ho
"`
[42}44], R"Dy
V
Er
V
, 0(x(1
[45,46], GdPd [47], and RNi
`
, where R"Gd
[48,49], Dy [49,50] and Ho [49,50]. The

for
selected materials is shown in Fig. 4. It is easy to
see, that the maximum MCE is signi"cantly re-
duced as temperature rises from &10 to &80 K,
which is associated with the rapid increase of the
lattice heat capacity in these alloys. The "eld de-
pendence of the MCE in this temperature range
varies from &1 to &2 K/T.
Fig. 4. The

in the best magnetic refrigerant intermetallic

materials in the temperature range from &10 to &80 K for
a magnetic "eld change from 0 to 7.5 T [42}50].
1.4.2. Intermediate temperatures: MCE in the range
&80}&250 K
This temperature range has not received nearly
as much attention as the other two ranges (10}80 K
and '250 K). First, there is the lack of the driving
force for magnetocaloric cooling to these temper-
atures. The second reason is the inherent minimum
of the /C ratio (where C is only the lattice and
electronic heat capacity) as shown for typical metal
(Cu) in Fig. 5 suggesting that the adiabatic heating
and cooling would be minimal in this temperature
range (Eqs. (4) and (5)). One of the best magnetic
refrigerant materials for the intermediate temper-
atures is pure Dy [11,51}54], with

reaching
&12 K at &180 K for H"7 T. Similarly, as
discussed earlier for the MCE in Tm, the complex-
ity of the magnetic structure of Dy brings about
a negative MCE for small magnetic "eld changes
(H(2 T). Recent studies of amorphous R
V
(

,
`
)
V
alloys, where R is a lanthanide metal,
and

and
`
are 3d transition metals [55}58]
indicate that these materials may be useful mag-
netic refrigerant materials between 100 and 200 K.
The "eld dependence of the MCE is just under
2 K/T for Dy, but rarely reaches more than 1 K/T
for all other studied magnetocaloric materials, in-
cluding the amorphous alloys. However, a recent
study [59,60] shows that the Gd
`
(Si
V
Ge
V
)
"
alloys
for 0.08)x )0.43 have extremely large S
`
and

values, 2}10 times larger than any of the above

mentioned materials (also see below, Section 1.5).
Fig. 5. The /C versus from lattice and electronic contribu-
tion in a typical metallic material.
V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56 49
1.4.3. Near room temperature MCE
The prototype material for the room temper-
ature range is the lanthanide metal Gd which or-
ders ferromagnetically at 294 K. Its magnetocaloric
e!ect has been studied independently by many
authors [5,8,14,61}66] and the

values at

'
are &6, 12, 16, and 20 K for H"2, 5, 7.5 and
10 T, respectively, reaching an impressive "eld de-
pendence of the MCE of &3 K/T in low magnetic
"elds which falls to &2 K/T in higher magnetic
"elds. A variety of Gd}R alloys, where R is another
lanthanide metal (such as Tb, Dy, Ho, or Er) and/or
Y were prepared in an attempt to improve the
MCE in Gd [67}74]. However, all alloying addi-
tions just shift the Curie temperature of Gd to
lower temperatures without any noticeable im-
provement in its MCE. The only known exception
is the probable enhancement of the low magnetic
"eld (H"1 T) MCE in nanocrystalline Gd}Y
alloys [75}77].
Most intermetallic compounds which order
magnetically near the room temperature but above
&290 K show a signi"cantly lower MCE than
that of Gd. A theoretical estimate [78] that the
MCE in Y
`
Fe
`
(
'
"&310 K) would be equal to
&50% of the MCE in Gd was recently veri"ed
experimentally [79]. Approximately, the same
magnitude of the MCE was also measured ex-
perimentally in Nd
`
Fe
`
with
'
"&324 K [79].
It has been suggested by [80] that there may be
an enhancement of the MCE in (Pr
`
Ce
"`
)Fe
`
,
but it has not been veri"ed experimentally. The
only intermetallic materials where the MCE
approaches that of Gd are Gd
`
Si
"
[60] with
the
'
"&335 K and germanium substituted
Gd
`
(Si
V
Ge
V
)
"
alloys, where 0.5(x)1 [59] with
&290(
'
)&335 K.
1.5. First-order magnetic phase transition and the
giant MCE
In an order}disorder magnetic phase transition
the existence of short-range order and spin #uctu-
ations well above the
'
often reduces the max-
imum possible "(RM/R)
&
" thus e!ectively reducing
the maximum MCE. A "rst-order phase trans-
formation theoretically should occur at constant
temperature, and therefore, "(RM/R)
&
" may be in"-
nitely large, consequently giving rise to a large,
i.e. giant magnetocaloric e!ect. One of the "rst
intermetallics, in which a giant, but also negative,
MCE was experimentally observed, is FeRh. It
undergoes an antiferromagnetic to ferromagnetic
"rst-order phase transition at &308 K, which is
accompanied by a MCE as large as !13 K for
H"2 T [81}83]. Unfortunately the giant MCE
in FeRh is irreversible and observed only upon
the application of the magnetic "eld to a virgin
sample.
Recently a series of Gd
`
(Si
V
Ge
V
)
"
alloys, where
0)x)0.5, was reported [59,84}87] to display
a S
`
at least two times larger than that of Gd near
room temperature and between 2 and 10 times
larger than the best magnetocaloric materials in the
low and intermediate (see above) temperature re-
gions. Besides the extremely large S
`
two addi-
tional features make these alloys unique, and likely
candidate magnetic refrigerant materials for the use
in highly e$cient magnetic refrigerators. The "rst is
the fact that their Curie temperature can be tuned
between &20 and &286 K by varying the Si to Ge
ratio and by introducing small alloying additions of
Ga to Gd
`
(Si
`
Ge
`
) [87]. This allows one to tune
the maximum giant magnetocaloric e!ect between
&20 and &305 K. The second is the fact, that
unlike FeRh, the giant magnetocaloric e!ect in the
Gd
`
(Si
V
Ge
V
)
"
alloys, where 0)x)0.5, is revers-
ible, i.e. it does not disappear after the "rst applica-
tion of the magnetic "eld. The di!erence in the
behavior of the MCE in FeRh and Gd
`
(Si
V
Ge
V
)
"
alloys is most likely associated with the di!erence
in the nature of the "rst order phase transition,
which in the former case is a magnetic order}order
transformation, while in the latter case it is a simul-
taneous magnetic and crystallographic phase
transition, i.e. it is a magnetic order}disorder, and
crystallographic order}order phase transformation
[86,88].
The third class of materials also displaying
a large magnetocaloric e!ect (comparable to Gd) is
based on di!erently substituted perovskite-like
LaMnO
`
materials with Y, Ca, Li and/or Na sub-
stituting for La, and Ti for Mn [89}92]. The
reversibility of the MCE as well as possible mecha-
nisms of its existence in these compounds has not
been reported.
50 V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56
2. Application of the magnetocaloric e4ect }
magnetic refrigeration
2.1. History of continuous magnetic refrigeration
The history of continuous magnetic refrigeration
can be traced to the work of Collins and Zimmer-
man [93], and Heer et al. [94], who built and tested
magnetic refrigerators (MR) operating between &1
and 0.73 K [93] and 0.2 K [94] by periodically
magnetizing and demagnetizing iron ammonium
alum. The latter apparatus extracted 123 erg/s
(12.3 J/s) from the cold reservoir at 0.2 K operat-
ing at 1/120 Hz frequency. It was not, however,
until Brown [95] reported a near room temper-
ature continuously operating MR, when it became
clear that magnetic refrigeration may be success-
fully utilized at signi"cantly higher temperatures
and achieve much larger temperature spans than
the maximumobserved MCE. Brown [95] was able
to attain a 473C no-load temperature di!erence
between the hot end (463C) and cold end (!13C) by
regenerating a column of #uid (80% water and
20% ethyl alcohol) using the lanthanide metal Gd
and a magnetic "eld change from 0 to 7 T. Follow-
ing the early work of Brown the concept of
active magnetic regenerator (AMR) refrigeration
was introduced by Steyert [96] and developed
by Barclay [97,98], and subsequently was brought
to life in various MR units operating at di!erent
temperatures [99}110]. Both the AMR concept
and the successful proof-of-principle room temper-
ature magnetic refrigerator are brie#y described
below.
2.2. Active magnetic regenerator cycle
A description of the AMR cycle was presented by
Zimm and DeGregoria [111]. In the AMR cycle,
a porous bed of a magnetic refrigerant material acts
as both the refrigerant (coolant) that produces re-
frigeration and the regenerator for the heat transfer
#uid. A schematic of the process is shown in Fig. 6.
Assume that the bed is at a steady state condition
with the hot heat exchanger at 753F (&243C) and
the cold heat exchanger at 403F (&53C). In Fig. 6a,
the initial temperature pro"le for the bed is in its
demagnetized state in zero magnetic "eld (dashed
line). When a magnetic "eld is applied to the refrig-
erant, each particle in the bed warms because of the
MCE to form the "nal magnetized bed temperature
pro"le (solid line). The amount each particle warms
is equal to

reduced by the e!ect of the heat

capacity of the heat transfer #uid in the pores
between the particles. Next, the 403F (&53C) #uid
#ows through the bed from the cold end to the hot
end (Fig. 6b). The bed is cooled by the #uid lower-
ing the temperature pro"le across the bed (the
dashed line to the solid line), and the #uid in turn is
warmed by the bed, emerging at a temperature
close to the temperature of the bed at the warm end.
This temperature is higher than 753F (&243C), so
heat is removed from the #uid at the hot heat sink
as the #uid #ows through the hot heat exchanger.
After the #uid #ow is stopped, the magnetic "eld is
removed, cooling the bed by the MCE (the dashed
line to the solid line in Fig. 6c). The refrigeration
cycle is completed by forcing the 753F #uid to #ow
from the hot to the cold end of the bed (Fig. 6d).
The #uid is cooled by the bed, emerging at a tem-
perature below 403F (&53C) and removes heat
from the cold sink as the #uid passes through the
cold heat exchanger.
The AMR cycle outlined above has several posit-
ive features useful for practical application in MR
devices. First, the temperature span of a single stage
can greatly exceed that of the MCE of the magnetic
refrigerant because the MCE of each individual
particle changes the entire temperature pro"le
across the bed. Second, because the bed acts as its
own regenerator, heat need not be transferred be-
tween two separate solid assemblies, but rather
between the solid particles in a single bed via the
action of a #uid. Third, the individual particles in
the bed do not encounter the entire temperature
span of the stage, and hence the bed may be made
into layers, each containing a magnetic material
with properties optimized for a particular temper-
ature range.
2.3. Recent achievements in near room temperature
refrigeration
The successful proof-of-principle MR, which was
recently described by Zimm et al. [112,113], oper-
ates near room temperature in a magnetic "eld
V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56 51
Fig. 6. The four steps of the active magnetic regenerator cycle: magnetizing (a), #ow from cold to hot (b), demagnetizing (c), and #ow
from hot to cold (d) [111].
between 1.5 to 5 T and uses approximately 3 kg of
Gd spheres packed in two magnetocaloric beds.
The heat transfer #uid is water. The beds are peri-
odically magnetized and demagnetized and the
#uid #ows are arranged as described above (AMR
cycle). A simple reciprocating machine, shown
schematically in Fig. 7, operates at low frequency
(&
'
Hz) and generates up to 600 W of cooling
power in a 5 T magnetic "eld (200 W of cooling
power in a 1.5 T "eld). Its e$ciency approaches
60% at 5 T (30% at 1.5 T) of Carnot with a coe$-
cient of performance (COP) approaching 15 at 5 T
(3 at 1.5 T). The e$ciency and COP are calculated
with the seal friction subtracted o!. The maximum
52 V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56
Fig. 7. A schematic sketch of the reciprocating proof-of-prin-
ciple magnetic refrigerator [112,113].
cooling power and COP are reported for a tem-
perature span equal to approximately
`
of the
maximum observed temperature span of 383C for
H"5 T.
3. Magnetocaloric e4ect and magnetic refrigeration
} future perspectives
The last quarter of the 20th century has seen
a continuous increase of interest in the mag-
netocaloric e!ect and magnetic refrigeration. The
number of publications, worldwide, has been rising
continuously during the last 35 years (Fig. 8) from
a total of 11 in 1965}1969 to a total of 127 in
1990}94 and an estimated 155 in 1995}1999 period.
This is about an 8% growth rate per year. Judging
Fig. 8. The number of publications concerning the mag-
netocaloric e!ect and magnetic refrigeration for the last 35 years
as cited in Chemical Abstracts. The "gure for 1995}1999 in-
cludes our projection for the second-half of 1998 and the entire
year for 1999.
by the number of publications, the research activ-
ities in the US are lagging behind the rest of the
world, although the major achievements [the dis-
covery of the giant magnetocaloric e!ect in
Gd
`
(Si
V
Ge
V
)
"
, and the design and the testing of
a reliable proof-of-principle, near room temper-
ature magnetic refrigerator] were accomplished re-
cently in the United States [59,84}87,112,113]. The
reduction in the US publication rate in the past "ve
years is probably associated with signi"cant reduc-
tion of the US defense, NIST and NASA research
expenditures, which occurred in late 1980s and
early 1990s. Research in Japan also has decreased
signi"cantly in the past 5 to 10 years, but this has
been more than o!set by recent increased research
on the magnetocaloric e!ect in China, Canada
and Europe. Furthermore, many of the new "nd-
ings are being published in letter-type journals
(particularly those which originate in China), which
also increases the number of non-US produced
reports.
We expect future developments in all aspects of
magnetic refrigeration to be rapidly forthcoming
because the traditional vapor compression techno-
logy is asymptotically approaching its technical
limitations for additional increases in energy e$-
ciency, and because of environmental concerns,
while magnetic refrigeration promises signi"cant
V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56 53
improvements in both. Magnetic refrigerant mate-
rials and magnetic refrigerator design represent the
two sides of the same scienti"c and technological
problem * the development of new magnetic re-
frigeration technology as an energy e$cient and
environmentally safe alternative to existing vapor
compression refrigeration. First, it is without
a doubt that new and better magnetic refrigerant
materials will be discovered as more research e!ort
will be concentrated on the search for and charac-
terization of new classes of metallic and non-metal-
lic magnetic materials for potential applications in
di!erent ranges of temperatures, from close to ab-
solute zero to well above room temperature. This
will also include basic studies on the fundamental
nature of magnetic ordering and how it a!ects the
magnetocaloric properties. Extensive basic re-
search is required, not only with regard to the
exploration and characterization of new magnetic
refrigerant materials, but also with respect to
methods of economical production and processing
them into porous media (including parallel plates,
spheres, jelly rolls, wires, etc.) suitable for making
magnetocaloric beds. Second, there is also no
doubt that new designs of magnetic refrigerators
using superconducting and permanent magnets will
be developed for a variety of large- and small-scale
applications for cooling from well above room tem-
perature to temperatures as low as 1 K. Extensive
basic and applied research is required not only
with respect to improvements in the thermodyn-
amic behavior and the mechanical con"guration of
the new magnetic refrigeration devices, but also
with respect to creating the varying magnetic "eld
environment to power-up the refrigerator to im-
prove the temperature lift across di!erently con-
"gured magnetocaloric beds, and in turn to
improve the overall magnetic refrigerator perfor-
mance.
Acknowledgements
The Ames Laboratory is operated for the US
Department of Energy (DOE) by Iowa State Univer-
sity under Contract No. W-7045-ENG-82. This work
was supported by the O$ce of Basic Energy Sciences,
Materials Sciences Division of the US DOE.
References
[1] E. Warburg, Ann. Phys. 13 (1881) 141.
[2] P. Debye, Ann. Phys. 81 (1926) 1154.
[3] W.F. Giauque, J. Amer. Chem. Soc. 49 (1927) 1864.
[4] A.H. Morrish, The Physical Principles of Magnetism,
Wiley, New York, 1965 (Chapter 3).
[5] A.M. Tishin, K.A. Gschneidner Jr., V.K. Pecharsky, Phys.
Rev. B 59 (1999) 503.
[6] V.M. Belova, V.I. Nikolaev, V.M. Stuchebnikov, Prib.
Tekh. Eksp. 1 (1974) 209.
[7] B.K. Ponomarev, Prib. Tekh. Eksp. 3 (1983) 153.
[8] G. Green, J. Chafe, J. Stevens, J. Humphrey, Adv. Cryog.
Eng. 35B (1990) 1165.
[9] H. Kato, K. Nara, M. Okaji, Cryogenics 31 (1991)
425.
[10] W. Otowski, C. Glorieux, R. Hofman, J. Thoen, Ther-
mochim. Acta 218 (1993) 123.
[11] B.R. Gopal, R. Chahine, M. Foeldeaki, T.K. Bose, Rev.
Sci. Instrum. 66 (1995) 232.
[12] R.Z. Levitin, V.V. Snegirev, A.V. Kopylov, A.S. Lagutin,
A. Gerber, J. Magn. Magn. Mater. 170 (1997) 223.
[13] B.R. Gopal, R. Chahine, T.K. Bose, Rev. Sci. Instr. 68
(1997) 1818.
[14] S.Yu. Dan'kov, A.M. Tishin, V.K. Pecharsky, K.A.
Gschneidner Jr., Rev. Sci. Instrum. 68 (1997) 2432.
[15] M. Foldeaki, R. Chahine, T.K. Bose, J. Appl. Phys. 77
(1995) 3528.
[16] M.W. Zemansky, Heat and Thermodynamics, 5th edi-
tion, McGraw-Hill, NY, 1968.
[17] V.K. Pecharsky, K.A. Gschneidner Jr., Adv. Cryog. Eng.
42A (1997) 423.
[18] W.F. Giauque, I.D.P. MacDougall, Phys. Rev. 43 (1933)
768.
[19] A.H. Cooke, Proc. Roy. Soc. A 62 (1949) 269.
[20] B. Bleaney, Proc. Roy. Soc. A 204 (1950) 203.
[21] A.H. Cooke, H.J. Du!us, W.P. Wolf, Philos. Mag. 44
(1953) 623.
[22] R.S. Craig, S.G. Sankar, N. Marzouk, V.U.S. Rao, W.E.
Wallace, E. Segal, J. Phys. Chem. Solids 33 (1972)
2267.
[23] K. Andres, P.H. Schmidt, S. Darack, AIP Conf. Proc. 24
(1975) 238.
[24] K. Andres, W.P. Sprenger, D.D. Oshero!, in: M. Krusius,
M. Vuorio (Eds.), Proceedigns of the 14th International
Conference on Low Temperature Physics, Vol. 4, North-
Holland, Amsterdam, Netherlands, 1975, p. 1.
[25] K. Andres, S. Darack, Physica 86-88B#C (1977) 1071.
[26] C. Buchal, K.J. Fischer, M. Kubota, R.M. Mueller,
F. Pobell, J. Phys. Lett. 39 (1978) L457.
[27] R. Hunik, E. Bongers, J.A. Konter, W.J. Huiskamp,
J. Phys. Colloq. 2 (1978) 1155.
[28] R.M. Mueller, C. Buchal, H.R. Folle, M. Kubota,
F. Pobell, Phys. Lett. A 75A (1980) 164.
[29] H. Ishimoto, N. Nishida, T. Furubayashi, M. Shinohara,
Y. Takano, Y. Miura, K. Ono, J. Low Temp. Phys. 55
(1984) 17.
54 V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56
[30] L. Skrbek, J. Stehno, J. Sebek, J. Tintera, J. Prusak,
S. Safrata, Cesk. Cas. Fyz. 36 (1986) 479.
[31] G.S. Burkhanov, O.D. Chistyakov, N.B. Kol'chugina,
Yu.M. Bun'kov, J. Stehno, L. Skrbek, J. Sebek,
Vysokochist. Veshchestva 1 (1990) 204.
[32] P.J. von Ranke, V.K. Pecharsky, K.A. Gschneidner, B.J.
Korte, Phys. Rev. B 58 (1998) 14436.
[33] A.E. Clark, R.S. Alben, J. Appl. Phys. 41 (1970) 1195.
[34] V. Nekvasil, V. Roskovec, F. Zounova, P. Novotny,
Czech. J. Phys. 24 (1974) 810.
[35] R.D. McMichael, J.J. Ritter, R.D. Shull, J. Appl. Phys. 73
(1993) 6946.
[36] R.Z. Levitin, V.V. Snegirev, A.V. Kopylov, A.S. Lagutin,
A. Gerber, J. Magn. Magn. Mater. 170 (1997) 223.
[37] R.D. Shull, R.D. McMichael, J.J. Ritter, Nanostruct. Ma-
ter. 2 (1993) 205.
[38] R.D. Shull, IEEE Trans. Magn. 29 (1993) 2614.
[39] C.B. Zimm, P.M. Ratzmann, J.A. Barclay, G.F. Green,
J.N. Chafe, Adv. Cryog. Eng. 36 (1990) 763.
[40] C.B. Zimm, P.L. Kral, J.A. Barclay, G.F. Green, W.G.
Patton, Proceedings of the 5th International Cryocooler
Conference, Wright Research and Development Center,
Wright Patterson Air Force Base, Ohio, 1988, p. 49.
[41] C.B. Zimm, J.A. Barclay, H.H. Harkness, G.F. Green,
W.G. Patton, Cryogenics 29 (1989) 937.
[42] T. Hashimoto, K. Matsumoto, T. Kurihara, T.
Numuzawa, A. Tomokiyo, H. Yayama, T. Goto, S. Todo,
M. Sahashi, Adv. Cryog. Eng. 32 (1986) 279.
[43] T. Hashimoto, T. Kuzuhara, K. Matsumoto, M. Sahashi,
K. Inomata, A. Tomokiyo, H. Yayama, Jpn. J. Appl.
Phys. 26 (1987) 1673.
[44] T. Hashimoto, T. Kuzuhara, M. Sahashi, K. Inomata,
A. Tomokiyo, A. Yayama, J. Appl. Phys. 62 (1987) 3873.
[45] H. Takeya, V.K. Pecharsky, K.A. Gschneidner Jr.,
J.O. Moorman, Appl. Phys. Lett. 64 (1994) 2739.
[46] K.A. Gschneidner Jr., H. Takeya, J.O. Moorman,
V.K. Pecharsky, S.K. Malik, C.B. Zimm, Adv. Cryogen.
Eng. 39 (1994) 1457.
[47] J.A. Barclay, W.C. Overton Jr., C.B. Zimm, in: U. Ekren,
A. Schmid, W. Weber, H. Wuhl (Eds.), LT-17 Contrib-
uted Papers, Elsevier Science, Amsterdam, 1984, p. 157.
[48] C.B. Zimm, E.M. Ludeman, M.C. Severson, T.A. Henn-
ing, Adv. Cryog. Eng. 37B (1992) 883.
[49] Tiezhong Ma, V.K. Pecharsky, K.A. Gschneidner Jr.,
unpublished.
[50] A. Tomokiyo, H. Yayama, H. Wakabayashi, T.
Kuzuhara, T. Hashimoto, M. Sahashi, K. Inomata, Adv.
Cryog. Eng. 32 (1986) 295.
[51] K.P. Belov, L.A. Chernikova, E.V. Talalaeva, R.Z. Levi-
tin, T.V. Kudryavtseva, S. Amadezzi, V.I. Ivanovski, Zh.
Eksp. Teor. Fiz. 58 (1970) 1923.
[52] B.K. Ponomarev, V.G. Tissen, S.P. Zapasskii, N.I.
Moreva, Fiz. Met. Metalloved. 44 (1977) 257.
[53] S.M. Benford, J. Appl. Phys. 50 (1979) 1868.
[54] S.A. Nikitin, A.S. Andreenko, A.M. Tishin, A.M. Ar-
kharov, A.A. Zherdev, Fiz. Met. Metalloved. 60 (1985)
689.
[55] X.Y. Liu, J.A. Barclay, B.R. Gopal, M. Foldeaki, R.
Chahine, T.K. Bose, P.J. Schurer, J.L. LaCombe, J. Appl.
Phys. 79 (1996) 1630.
[56] X.Y. Liu, J.A. Barclay, M. Foldeaki, B.R. Gopal, R.
Chahine, T.K. Bose, Adv. Cryog. Eng. 42A (1997) 431.
[57] M. Foldeaki, A. Giguere, B.R. Gopal, R. Chahine, T.K.
Bose, X.Y. Liu, J.A. Barclay, J. Magn. Magn. Mater. 174
(1997) 295.
[58] M. Foldeaki, R. Chahine, B.R. Gopal, T.K. Bose, X.Y.
Liu, J.A. Barclay, J. Appl. Phys. 83 (1998) 2727.
[59] V.K. Pecharsky, K.A. Gschneidner Jr., Appl. Phys. Lett.
70 (1997) 3299.
[60] V.K. Pecharsky, K.A. Gschneidner Jr., unpublished.
[61] S.A. Nikitin, E.V. Talalaeva, L.A. Chernikova, G.E.
Chuprikov, T.I. Ivanova, G.V. Kazakov, G.A. Yarkho,
Zh. Eksp. Teor. Fiz. 74 (1978) 205.
[62] T.I. Ivanova, R.Z. Levitin, S.A. Nikitin, E.V. Talalaeva,
L.A. Chernikova, Fiz. Met. Metalloved. 51 (1981) 893.
[63] S.M. Benford, G.V. Brown, J. Appl. Phys. 52 (1981)
2110.
[64] B.K. Ponomarev, J. Magn. Magn. Mater. 61 (1986) 129.
[65] C. Glorieux, J. Caerels, J. Thoen, J. Appl. Phys. 80 (1996)
3412.
[66] S.Yu. Dan'kov, A.M. Tishin, V.K. Pecharsky, K.A.
Gschneidner Jr., Phys. Rev. B 57 (1998) 3478.
[67] S.A. Nikitin, A.S. Andreenko, N.P. Arutyunyan, Fiz.
Tverd. Tela 20 (1978) 3685.
[68] S.A. Nikitin, A.S. Andreenko, A.K. Zvezdin, A.F. Pop-
kov, Izv. Akad. Nauk SSSR, Ser. Fiz. 44 (1980) 1343.
[69] E.V. Talalaeva, L.A. Chernikova, S.A. Nikitin, T.I.
Ivanova, A.S. Andreenko, Fiz. Met. Metalloved. 54 (1982)
495.
[70] S.A. Nikitin, A.S. Andreenko, A.M. Tishin, A.M. Ar-
kharov, A.A. Zherdev, Fiz. Met. Metalloved. 59 (1985)
327.
[71] S.A. Nikitin, A.M. Tishin, Pis'ma Zh. Tekh. Fiz. 14 (1988)
735.
[72] S.A. Nikitin, A.M. Tishin, S.V. Red'ko, Fiz. Met. Metal-
loved. 66 (1988) 86.
[73] G.S. Burkhanov, S.Yu. Dan'kov, S.A. Nikitin, A.M.
Tishin, O.D. Chistyakov, Vysokochist. Veshchestva 58
(1993) 3231.
[74] M. Foldeaki, W. Schnelle, E. Gmelin, P. Benard, B. Ko-
szegi, A. Giguere, R. Chahine, T.K. Bose, J. Appl. Phys. 82
(1997) 309.
[75] Y.Z. Shao, J. Zhang, J.K.L. Lai, C.H. Shek, J. Appl. Phys.
80 (1996) 76.
[76] Y.Z. Shao, J.K.L. Lai, C.H. Shek, J. Magn. Magn. Mater.
163 (1996) 103.
[77] Y.Z. Shao, Z. Xiong, J. Zhang, J. Zhang, Sci. China, Ser.
A 39 (1996) 748.
[78] H. Oesterreicher, F.T. Parker, J. Appl. Phys. 55 (1984)
4334.
[79] S.Yu. Dankov, V.V. Ivtchenko, A.M. Tishin, V.K.
Pecharsky, K.A. Gschneidner Jr., unpublished.
[80] S. Jin, L. Liu, Y. Wang, B. Chen, J. Appl. Phys. 70 (1991)
6275.
V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56 55
[81] S.A. Nikitin, G. Myalikgulev, A.M. Tishin, M.P. An-
naorazov, K.A. Asatryan, A.L. Tyurin, Phys. Lett. A 148
(1990) 363.
[82] M.P. Annaorazov, K.A. Asatryan, G. Myalikgulyev, S.A.
Nikitin, A.M. Tishin, A.L. Tyurin, Cryogenics 32 (10)
(1992) 867.
[83] M.P. Annaorazov, S.A. Nikitin, A.L. Tyurin, K.A. Asat-
ryan, A.Kh. Dovletov, J. Appl. Phys. 79 (1996) 1689.
[84] V.K. Pecharsky, K.A. Gschneidner Jr., Phys. Rev. Lett. 78
(1997) 4494.
[85] V.K. Pecharsky, K.A. Gschneidner Jr., J. Magn. Magn.
Mater. 167 (1997) L179.
[86] V.K. Pecharsky, K.A. Gschneidner Jr., J. Alloys Com-
pounds 260 (1997) 98.
[87] V.K. Pecharsky, K.A. Gschneidner Jr., Adv. Cryo. Eng.
43 (1998) 1729.
[88] L. Morellon, P.A. Algarabel, M.R. Ibarra, J. Blasco, B.
Garcia-Landa, Z. Arnold, F. Albertini, Phys. Rev. B 58
(1998) R14721.
[89] X.X. Zhang, J. Tajada, Y. Xin, G.F. Sunm, K.W. Wong,
X. Bohigas, Appl. Phys. Lett. 69 (1996) 3596.
[90] W. Zhong, W. Chen, W. Ding, N. Zhang, Y. Du, Q. Yan,
Solid State Commun. 106 (1998) 55.
[91] W. Chen, W. Zhong, D. Hou, W. Ding, Y. Du, Q. Yan,
Chinese Phys. Lett. 15 (1998) 134.
[92] X. Bohigas, J. Tejada, E. Del Barco, X.X. Zhang, M. Sales,
Appl. Phys. Lett. 73 (1998) 390.
[93] S.C. Collins, F.J. Zimmerman, Phys. Rev. 90 (1953)
991.
[94] C.V. Heer, C.B. Barnes, J.C. Daunt, Rev. Sci. Instr. 25
(1954) 1088.
[95] G.V. Brown, J. Appl. Phys. 47 (1975) 3673.
[96] W.A. Steyert, J. Appl. Phys. 49 (1978) 1216.
[97] J.A. Barclay, W.A. Steyert, U.S. Patent No. 4,332,135,
June 1, 1982.
[98] J.A. Barclay, U.S. Patent No. 4,408,463, October 11, 1983.
[99] L.D. Kirol, M. Dacus, Adv. Cryog. Eng. 33 (1988) 757.
[100] V.A. Al'tov, V.M. Brodyanskii, V.I. Karagusov, V.V.
Kurguzov, Yu.V. Sunyavskii, V.V. Sychev, Sov. Phys.
Dokl. 33 (1988) 759.
[101] G. Green, J. Shafe, Adv. Cryog. Eng. 35 (1990) 1165.
[102] A.J. DeGregoria, J.A. Barclay, P.J. Claybaker, S.R.
Jaeger, S.F. Kral, R.A. Pax, J.R. Rowe, C.B. Zimm, Adv.
Cryog. Eng. 35 (1990) 1125.
[103] A.J. DeGregoria, L.J. Feuling, J.F. Lattsch, J.R. Rowe,
J.R. Trueblood, A.A. Wang, Adv. Croyg. Eng. 37B (1992)
875.
[104] D. Janda, T. DeGregoria, J. Johnson, S. Jral, G. Kinrad,
Adv. Cryog. Eng. 37B (1992) 891.
[105] N.V. Filin, I.I. Mikhailov, A.L. Dovbish, P.L. Ronjin,
IEEE Trans. Magn. 28 (1992) 953.
[106] C.B. Zimm, J.W. Johnson, R.W. Murphy, Adv. Cryog.
Eng. 41 (1996) 1675.
[107] G.F. Nellis, J.L. Smith Jr., Adv. Cryog. Eng. 41 (1996)
1665.
[108] T. Numazawa, Teion Kogaku 32 (1997) 192.
[109] T. Ochi, H. Masatomi, Y. Hasegawa, R. Aoki, T. Ogushi,
K. Yabu-uchi, in: T. Haruyama, T. Mitsui, K. Yamafuji
(Eds.), Proceedings of the 16th International Cryo. Eng.
Conference/Intern. Cryo. Mater. Conference, Elsevier
Science, Oxford, 1997, p. 399.
[110] K. Ohira, S. Matsuo, H. Furumoto, in: T. Haruyama, T.
Mitsui, K. Yamafuji (Eds.), Proceedings of the 16th Inter-
national Cryo. Eng. Conference/Intern. Cryo. Mater.
Conference, Elsevier Science, Oxford, 1997, p. 403.
[111] C.B. Zimm, A.J. DeGregoria, in: H.S. Kwok, D.T. Show,
H.J. Naughton (Eds.), Proceedings of the 6th Interna-
tional Conference Superconduct. Applic., Bu!alo, NY,
September 15}17, AIP, New York, 1993, p. 471.
[112] C. Zimm, A. Jastrab, A. Sternberg, V. Pecharsky, K.
Gschneidner Jr., M. Osborne, I. Anderson, Adv. Cryog.
Eng. 43 (1998) 1759.
[113] K.A. Gschneidner, V.K. Pecharsky, C.B. Zimm, Mater.
Technol. 12 (1997) 145.
56 V.K. Pecharsky, K.A. Gschneidner Jr. / Journal of Magnetism and Magnetic Materials 200 (1999) 44}56