Tegus Thesis

Novel Materials for Magnetic
Refrigeration

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor
aan de Universiteit van Amsterdam
op gezag van de Rector Magnificus
prof. mr. P.F. van der Heijden
ten overstaan van een door het college voor promoties
ingestelde commissie, in het openbaar te verdedigen
in de Aula der Universiteit
op donderdag 23 oktober 2003 te 11.00 uur

door
Tegusi
geboren te Chifeng (P.R. China)

AMSTERDAM 2003

Promotiecommissie

Promotores
Prof. dr. F.R. de Boer
Prof. dr. K.H.J. Buschow
Co-promotor
Dr. E. Brck

Overige leden
Prof. dr. W.C. Sinke
Dr. A.M. Tishin
Dr. M. ter Brake
Prof. dr. J.J.M. Franse
Prof. dr. M.S. Golden
Dr. A. de Visser

Faculteit der Natuurwetenschappen, Wiskunde en Informatica

The work described in this thesis was supported by the Dutch
Technology Foundation (STW), the applied science division of Netherlands
Organization for Scientific Research (NWO) and the technology program of
the ministry of Economic Affairs, carried out in the Materials Physics group
at the

Van der Waals-Zeeman Instituut,
Universiteit van Amsterdam
Valckenierstraat 65-67, 1018 XE Amsterdam, the Netherlands

Where a limited number of copies of this thesis is available
ISBN: 90 5776 107 6

Printed in the Netherlands by PrinterPartners Ipskamp B.V.,
P.O. Box 333, 7500 AH Enschede

Cover illustration: Ton Riemersma

To my parents, my wife Dagula and my son Yiliqi

Contents

1 Introduction.. 1
1.1 General introduction 1
1.2 The magnetocaloric effect 2
1.3 Magnetic refrigeration... 3
1.4 Outline of this thesis . 5
References .. 5

2 Theoretical aspects .... 7
2.1 Gibbs free energy .... 8
2.2 Magnetic entropy .... 9
2.3 The Bean-Rodbell model ... 11
References .. 14

3 Experimental ... 15
3.1 Sample preparation .. 15
3.1.1 Arc melting and ball milling 15
3.1.2 Crystal growth .. 17
3.2 Sample characterization 18
3.3 Magnetic measurements 19
3.4 Specific- heat measurements . 19
3.5 Electrical-resistivity measurements ... 20
3.6 Determination of the magnetocaloric effect ... 21
3.6.1 Determination of the magnetocaloric effect from
magnetization measurements 21
3.6.2 Determination of the magnetocaloric effect from

ii
specific-heat measurements .. 22
References . 24

4 Magnetic phase transitions and magnetocaloric effect in
Gd-based compounds 25
4.1 Introduction 25
4.2 GdRu
2
Ge
2
.. 26
4.2.1 Introduction 26
4.2.2 Experimental . 27
4.2.3 Results and discussion ... 27
4.2.4 Conclusions ... 33
4.3 Single-crystalline Gd
5
Si
1.7
Ge
2.3
33
4.3.1 Introduction .. 33
4.3.2 Cyrstal growth and characterization . 34
4.3.3 Magnetic properties .. 36
4.3.4 Specific heat . 41
4.3.5 Magnetocaloric effect .. 43
4.3.6 Discussion and conclusions .. 45
References .. 48

5 Magnetocaloric effect in hexagonal MnFeP
1-x
As
x

compounds 51
5.1 Introduction .. 51
5.2 Sample preparation and characterization . 53
5.3 Structural properties . 54
5.4 Magnetic properties .. 56
5.5 Specific heat and dc susceptibility 60

iii
5.6 Magnetocaloric effect ... 62
5.7 Electrical resistivity and magnetoresistivity . 68
5.8 A model description of the first-order magnetic
phase transition .. 71
5.9 Discussion and conclusions 77
References.... 82

6 Effects of Mn/Fe ratio on the magnetocaloric properties of
hexagonal MnFe(P,As) compounds 85
6.1 Introduction 85
6.2 Experimental . 87
6.3 Results and discussion .. 87
6.3.1 Structural and magnetic properties 87
6.3.2 Magnetocaloric properties 94
6.3.3 Electrical resistivity ..... 102
6.4 Conclusions .. 104
References ..... 105

Summary ...... 107
Samenvatting .. 110
Publications ....... 113
Acknowledgments ......... 117

1

Chapter 1

Introduction

1.1 General introduction

Modern society relies very much on readily available cooling. Next to the food
storage and transport, air-conditioning in buildings and cars gains more
importance, and in the near future it is envisaged that superconducting electronics
may be operated at liquid-nitrogen temperatures. These developments call for
energy-efficient and versatile refrigeration technology.
The vapor-compression refrigerators have become ubiquitous in a large
number of cooling applications. However, the use of chlorofluorocarbons (CFCs)
and hydrochlorofluorocarbons (HCFCs) as working fluids has raised serious
environmental concerns, primarily for the role in the destruction of the ozone layer
[1, 2] and the global warming. Replacement by fluid hydrofluorocarbons (HFCs),
which contain no chlorine and therefore have no ozone depletion potential, is not
without problems because the HFCs are greenhouse gases [3] with higher global
warming potential than CO
2
. In addition, the efficiency of the vapor-compression
refrigeration systems is not expected to be significantly improved in the future.
Thus, due to slow improvement of the efficiency and serious concern for the
environment, alternative technologies that use either inert gases or no fluid at all
become attractive solutions to the environment problems.
2 Chapter 1

Magnetic refrigeration has been in use in scientific applications for a long
time for cooling below 1 K. But there are no commercial applications at
temperatures around room temperature due to the fact that the magnetocaloric
effect (MCE) is relatively weak in most ferromagnetic materials at these
temperatures. Only gadolinium exhibits a considerable MCE, about 2 K/T, at room
temperature. Recently, this technique has been demonstrated [4] as a promising
alternative for the conventional gas-compression/expansion technique generally in
use today. But the major problem in magnetic refrigeration is still to find working
materials with a large MCE in different temperature regions.
In 1997, Pecharsky and Gschneidner [5] have reported the discovery of the
so-called giant MCE in the Gd
5
(Si
x
Ge
1-x
)
4
system. Subsequently, a large number of
materials were reported as candidate materials for magnetic cooling [6-9].
Although considerable success has been achieved in developing magnetic
refrigerants, the search for novel working materials is still an important task, in
particular, in order to develop suitable materials for room-temperature applications
in lower fields, which can be generated by permanent magnets [10, 11].
The motivation of our research project, namely to explore new materials for
magnetic cooling, was two-fold. From the application point of view, we have
focused on finding potential refrigerants, specifically among Mn- or Fe-based
compounds, in order to establish the appropriateness for room-temperature
magnetic-cooling application. From the fundamental point of view, our motivation
was to gain a deeper insight into the fundamental relations between the MCE and
magnetic phase transitions, compositions, and the thermomagnetic properties of
solid magnetic materials. This insight may serve as a guide in the search for new
materials suitable for application.

1.2 The magnetocaloric effect

The MCE is defined as the thermal response of a magnetic material to an applied
magnetic field and is apparent as a change in its temperature. It was discovered by
Introduction 3

Warburg [12] in 1881 and is intrinsic to all magnetic materials. In the case of a
ferromagnetic material as depicted in Fig. 1.1, the material heats up when it is
magnetized and cools down when it is removed out of the magnetic field.
The magnitude of the MCE of a magnetic material is characterized by the
adiabatic temperature change
ad
T , or by the isothermal magnetic-entropy change
m
S due to a varying magnetic field. The nature of the MCE in a solid is the result
of the entropy variation due to the coupling of the magnetic spin system with the
magnetic field [13]. For the various aspects of the MCE and magnetic refrigeration
we refer to Kuzman and Tishin [14], Gschneidner and Pecharsky [15], and Tishin
[16].

1.3 Magnetic refrigeration

Magnetic refrigeration is a method of cooling based on the MCE. The heating and
cooling caused by a changing magnetic field are similar to the heating and cooling
of a gaseous medium in response to compression and expansion. A schematic
representation of a magnetic-refrigeration cycle is depicted in Fig. 1.1.
When a ferromagnetic material containing atoms that carry magnetic
moments is placed in an external magnetic field, the field forces the magnetic
moments to align, reducing the magnetic entropy. Since the total entropy is
constant under adiabatic conditions, the reduced part of the magnetic entropy is
transferred from spin subsystem to lattice subsystem via spin and lattice coupling.
This causes an increase of the lattice entropy, which makes the atoms vibrate more
rapidly, and results in an increase of temperature of the material. Conversely, when
the material is taken out of the magnetic field, the moments randomize again and
remove entropy from the lattice, creating a cooling effect.
In 1926, Debye [17] and Giauque [18] have independently proposed the
principle of adiabatic magnetic cooling, which utilizes the MCE of paramagnetic
salts, as a means of reaching temperatures below the boiling point of liquid helium.

4 Chapter 1

Figure 1.1: Schematic representation of a magnetic-refrigeration cycle in
which heat is transported from the heat load to its surroundings. Initially
randomly oriented magnetic moments are aligned by a magnetic field,
resulting in heating of the material. This heat is removed from the material to
its surroundings by a heat-transfer medium. On removing the field, the
magnetic moments randomize, which leads to cooling of the magnetic
material to below the ambient temperature. Depending on the operating
temperature, the heat-transfer medium may be water (with antifreeze) or air,
and, for very low temperatures, helium.

In 1933, Giauque and MacDougall [19] have put this idea into practice and have
experimentally demonstrated the use of the MCE to achieve temperatures below 1
K. From then on, the MCE has been successfully utilized to achieve ultra-low
temperatures by employing a process known as adiabatic demagnetization. In 1976,
Brown [20] has reported a prototype of a room-temperature magnetic refrigerator
and demonstrated that magnetic refrigeration can be realized in the room-
temperature region. In 2001, Astronautics Corporation of America [4] has realized
the worlds first successful room-temperature magnetic refrigerator, in which
Introduction 5

permanent magnets were used to generate the magnetic field. This achievement
moves the magnetic refrigerator a step closer to commercial applications.

1.4 Outline of this thesis

The work presented in this thesis is a study of the MCE and related physical
properties of several systems of intermetallic compounds.
The theoretical aspects of the MCE and the Bean-Rodbell model that
describes magnetic phase transitions observed in MnFeP
1-x
As
x
compounds are
presented in Chapter 2. In Chapter 3, a short review is given of the experimental
techniques and set-ups that have been employed for the sample preparation, the
characterization and the investigation of the physical properties of the materials
studied in this thesis.
Chapter 4 is designated to the MCE and related physical properties of the
Gd-based compounds GdRu
2
Ge
2
and Gd
5
Si
1.7
Ge
2.3
. The isothermal magnetic-
entropy change of GdRu
2
Ge
2
was determined by means of both magnetization and
specific-heat measurements, which are in good agreement. We have also grown a
single crystal of Gd
5
Si
1.7
Ge
2.3
and have studied the thermomagnetic properties and
the MCE of this single crystal.
The highlight of the work is the discovery of the giant MCE in transition-
metal-based compounds of the type MnFeP
1-x
As
x
. A systematic study of the MCE
and related physical properties of the MnFeP
1-x
As
x
compounds is presented in
Chapter 5. The magnetocaloric properties of MnFe(P,As)-based compounds can be
improved by varying the Mn/Fe ratio. This is reported in Chapter 6. The last part is
the summary of the present thesis work.

6 Chapter 1

References

[1] F. Drake, M. Purvis and J. Hunt, Public Understand. Sci. 10 (2001) 187.
[2] Montreal Protocol on Substances that Deplete the Ozone Layer, United
Nations (UN), New York, NY, USA, 1987.
[3] Kyoto Protocol to the United Nations Framework Convention on Climate
Change, United Nations (UN), New York, NY, USA, 1997.
[4] http://www.external.ameslab.gov/News/release/01magneticrefrig.htm.
[5] V.K. Pecharsky and K.A. Gschneidner, Jr., Phys. Rev. Lett. 78 (1997) 4494.
[6] F.X. Hu, B.G. Shen and J.R. Sun, Appl. Phys. Lett. 76 (2000) 3460.
[7] H. Wada and Y. Tanabe, Appl. Phys. Lett. 79 (2001) 3302.
[8] O. Tegus, E. Brck, K.H.J. Buschow and F.R. de Boer, Nature 415 (2002)
150.
[9] A. Fujita, S. Fujieda, Y. Hasegawa and K. Fukamichi, Phys. Rev. B 67
(2003) 104416.
[10] J.M.D. Coey, J. Magn. Magn. Mater. 248 (2002) 441.
[11] S.J. Lee, J.M. Kenkel, V.K. Pecharsky and D.C. Jiles, J. Appl. Phys. 91
(2002) 8543.
[12] E. Warburg, Ann. Phys. Chem. 13 (1881) 141.
[13] V.K. Pecharsky, K.A. Gschneidner, Jr., A.O. Pecharsky and A.M. Tishin,
Phys. Rev. B 64 (2001) 144406.
[14] M.D. Kuzmin and A.M. Tishin, Cryogenics 32 (1992) 545.
M.D. Kuzmin and A.M. Tishin, Cryogenics 33 (1993) 868.
[15] K.A. Gschneidner, Jr., and V.K. Pecharsky, Ann. Rev. Mater. Sci. 30 (2000)
387.
[16] A.M. Tishin, Magnetocaloric effect in the vicinity of phase transitions, in
Handbook of Magnetic Materials, Vol. 12, Edited by K.H.J. Buschow
(North Holland, Amsterdam, 1999) pp. 395-524.
[17] P. Debye, Ann. Physik 81 (1926) 1154.
[18] W.F. Giauque, J. Amer. Chem. Soc. 49 (1927) 1864.
[19] W.F. Giauque and D.P. MacDougall, Phys. Rev. 43 (1933) 768.
[20] G.V. Brown, J. Appl. Phys. 47 (1976) 3673.
7

Chapter 2

Theoretical aspects

The magnetocaloric effect (MCE) of a magnetic material is associated with the
magnetic-entropy change of the material. The theoretical aspects of the MCE have
been discussed in Refs. 1 and 2. According to thermodynamics, the MCE is
proportional to M/T at constant field and inversely proportional to the field
dependence of the specific heat c
p
(T,B). In the temperature region of a magnetic
phase transition, the magnetization changes rapidly and, therefore, a large MCE is
expected in this region [3,4]. However, the critical behavior of the physical
quantities in the phase-transition region is so complicated that there is no unified
theory. The theoretical description of MCE is still far from complete. Therefore,
the adiabatic temperature change T
ad
of a given material can only be determined
by using experimental methods.
The understanding of magnetic phase transitions and the evaluation of the
entropy change associated with the magnetic phase transitions, therefore, form an
important part of this thesis. In this chapter, we will first introduce the theoretical
background of the MCE. Then, we will give outlines of the Bean-Rodbell model
[5] that we will use to describe the first-order magnetic phase transition in the
hexagonal MnFe(P,As)-type compounds.

8 Chapter 2

2.1 Gibbs free energy

The thermodynamic properties of a system are fully determined by the Gibbs free
energy or free enthalpy of the system. The system we consider here consists of a
magnetic material in a magnetic field B at a temperature T under a pressure p. The
Gibbs free energy G of the system is given by

MB pV TS U G + = , (2.1)

where U is the internal energy of the system, S the entropy of the system, and M
the magnetization of the magnetic material. The volume V, magnetization M, and
entropy S of the material are given by the first derivatives of the Gibbs free energy
as follows

. ) , , (
) , , (
) , , (
,
,
,
p B
p T
B T
T
G
p B T S
B
G
p B T M
p
G
p B T V
=
(2.2)

The specific heat of the material is given by the second derivative of the
Gibbs free energy with respect to temperature

. ) , (
2
2
p
p
T
G
T B T c

= (2.3)

Theoretical aspects of the MCE 9

By definition, if the first derivative of the Gibbs free energy is discontinuous
at the phase transition, then the phase transition is of first order. Therefore, the
volume, magnetization, and entropy of the magnetic material are discontinuous at a
first-order phase transition. If the first derivative of the Gibbs free energy is
continuous at the phase transition but the second derivative is discontinuous, then
the phase transition is of second order.

2.2 Magnetic entropy

The total entropy of a magnetic material in which the magnetism is due to localized
magnetic moments, as for instance in lanthanide-based materials, is presented by

), , , ( ) , , ( ) , , ( ) , , ( p B T S p B T S p B T S p B T S
m e l
+ + = (2.4)

where S
l
represents the entropy of the lattice subsystem, S
e
the entropy of
conduction-electron subsystem and S
m
the magnetic entropy, i.e. the entropy of the
subsystem of the magnetic moments. In magnetic solids exhibiting itinerant-
electron magnetism, separation of these three contributions to the total entropy is,
in general, not straightforward because the 3d electrons give rise to the itinerant-
electron magnetism but also participate in the conduction. Separation of the lattice
entropy is possible only if electron-phonon interaction is not taken into account.
Since the entropy is a state function, the full differential of the total entropy
of a closed system is given by

.
,
,
,
dB
B
S
dp
p
S
dT
T
S
dS
p T
B T
B p

=
(2.5)

Among these three contributions, the magnetic entropy is strongly field
dependent, and the electronic and lattice entropies are much less field dependent.
Therefore, for an isobaric-isothermal (dp = 0; dT = 0) process, the differential of
the total entropy can be represented by
10 Chapter 2

.
,
dB
B
S
dS
p T
m
= (2.6)

For a field change from the initial field B
i
to the finalfield B
f
, integration of
Eq. (2.6) yields for the total entropy change

) , ( ) , ( ) , ( ) , ( B T S B T S B T S B T S
m i f
= = , (2.7)

where B = B
f
- B
i
. This means that the isothermal-isobaric total entropy change
of a magnetic material in response to a field change B is also presented by the
isothermal-isobaric magnetic-entropy change.
The magnetic-entropy change is related to the bulk magnetization, the
magnetic field and the temperature through the Maxwell relation

.
) , ( ) , (
, , p B p T
m
T
B T M
B
B T S
(2.8)

Integration yields

dB
T
B T M
B T S
p B
B
B
m
f
i ,
) , (
) , (

= . (2.9)

On the other hand, according to the second law of thermodynamics

.
) , (
,
T
B T c
dT
dS p
p B
=
(2.10)

Integration yields


.
) , (
) , (
'
0
'
'
0
dT
T
B T c
S B T S
T
p
+ = (2.11)

In the absence of configurational entropy, the entropy will be zero at T = 0 K, so
that the value of S
0
is usually chosen to be zero. Therefore, the entropy change in
response to a field change B is given by

'
0
'
' '
) , ( ) , (
) , ( dT
T
B T c B T c
B T S
T
i p f p

= , (2.12)

where ) , (
'
f p
B T c and ) , (
'
i p
B T c represent the specific heat at constant pressure p
in the magnetic field B
f
and B
i
, respectively.

2.3 The Bean-Rodbell model

Bean and Rodbell [5] have proposed a phenomenological model that describes the
first-order phase transition in MnAs. Blois and Rodbell have used this model to
explain the first-order magnetic phase transition observed for MnAs [6]. Zach et al.
[7] have used this model in a semiquantitative analysis of the magnetic phase
transition in the MnFeP
1-x
As
x
series of compounds. In this section, we will
introduce the Bean-Rodbell model.
This model correlates strong magnetoelastic effects with the occurrence of a
first-order phase transition. The central assumption in the model of Bean and
Rodbell is that the exchange interaction (or Curie temperature) is strongly
dependent on the interatomic spacing. In this model, the dependence of Curie
temperature on the volume is represented by

], / ) ( 1 [
0 0 0
V V V T T
C
+ = (2.13)

12 Chapter 2

where T
C
is the Curie temperature, whereas T
0
would be the Curie temperature if
the lattice were not compressible, and V
0
would be the volume in the absence of
exchange interaction. The coefficient may be positive or negative.
In the Bean-Rodbell model, the critical behavior of the magnetic system is
analyzed on the basis of the Gibbs free energy consisting of the following
contributions

,
press entropy elastic Zeeman exch
G G G G G G + + + + = (2.14)

where G
exch
, G
Zeeman
, G
elastic
, G
entropy
, and G
press
represent the exchange interaction,
the Zeeman energy, the elastic energy, the entropy term, and the pressure term,
respectively. Within the molecular-field approximation, for arbitrary spin j, Eq.
(2.14) is given by [6]

pV S S T
V
V V
K
B T Nk
j
j
G
l j C B
+ +
+
= ) (
) (
2
1
1 2
3
2
0
0
0
2
, (2.15)

where N is the number of magnetic atoms per kilogram, k
B
the Boltzmann constant,
0
the saturation magnetization per kilogram at 0 K, the relative magnetization
(M/
0
), K the compressibility, S
j
the entropy of the spin subsystem, and S
l
the
entropy of the lattice subsystem. Inserting Eq. (2.13) into Eq. (2.15) and
minimizing the expression for G with respect to volume, we obtain the equilibrium
volume for arbitrary j

.
) 1 ( 2
3
2
0
0 0
0
pK T KT Nk
j
j
V V
V V
B
+
+
=
(2.16)

This result shows that the magnetization depends on the volume change. The
term T, in which is the lattice thermal expansion coefficient, is from the thermal
expansion.
Inserting Eqs. (2.13) and (2.16) into Eq. (2.15), we obtain

,
) 0 ( ) (
)) ( 1 )(
1
(
2
3
1 8
9 ) 0 ( ) (
0 0
0 2
4 2
0
2
0 0
+
=
B
j j
B
B
B
Nk
S S
T
T
T Nk
B
T pK
j
j
K T Nk
j
j
V T Nk
G G
(2.17)

The implicit dependence of the magnetization on temperature is obtained by
minimizing Eq. (2.17) with respect to . In the case of absence of external pressure
(p = 0), we obtain

,
) (
1
2
/
0
0 0
3
0

+ +
=
j
B
B j j j
S
Nk
T
T Nk B b a
T
T
(2.18)

where

.
] 1 ) 1 2 [( 2
)] 1 ( 4 [ 5
,
)] 1 ( 2 [
1 ) 1 2 (
5
9
,
1
3
2
0
0
4
2
4
4
KT Nk
V j
j j
j
j
b
j
j
a
B j
j
j
+
+
=
+
+
=
+
=
(2.19)

Here,
j
is an important parameter, involving the parameters K and that are
related to the volume change. In the molecular-field approximation, the spin
entropy, S
j
, is a function of the relative magnetization and can be expressed as a
series in even powers of [5]. For the compounds MnFeP
1-x
As
x
(0.25 < x < 0.65),
magnetization measurements at 5 K show that the saturation magnetization is about
4
B
/f.u., from which we conclude that the angular momentum j equals 2 (assuming
that g = 2). In this case, Eq. (2.18) becomes

14 Chapter 2

.
2 606 . 0 867 . 0 2
/ 867 . 0 2
0
5 3
0 0
3
2
0

T
T Nk B
T
T
B
+ +
+ +
= (2.20)

This equation may express the temperature and field dependence of the
magnetization of MnFeP
1-x
As
x
in the vicinity of the phase transition. In Chapter 5,
we will fit our experimental results based on Eqs. (2.17) and (2.20), and will give a
model description of the first-order magnetic phase transition in MnFeP
1-x
As
x

compounds.

References

[3] K.A. Gschneidner, Jr., and V.K. Pecharsky, Mater. Sci. Eng. A287 (2000)
301.
[4] O. Tegus, E. Brck, L. Zhang, Dagula, K.H.J. Buschow and F.R. de Boer,
Physica B 319 (2002) 174.
[5] C.P. Bean and D.S. Rodbell, Phys. Rev. 126 (1962) 104.
[6] R.W. de Blois and D.S. Rodbel, Phys. Rev. 130 (1963) 1347.
[7] R. Zach, M. Guilot and J. Tobola, J. Appl. Phys. 83 (1998) 7237.
15

Chapter 3

Experimental

3.1 Sample preparation

3.1.1 Arc melting and ball milling

Generally, intermetallic compounds are prepared by melting. In this way, also the
intermetallic compounds investigated in this present thesis were prepared by arc
melt ing appropriate amounts of the constituent elements, typically about 5 g, of at
least 99.9 % purity in a water-cooled copper crucible pre-evacuated to better than
210
-6
mbar and refilled with high-purity Ar gas. In order to obtain homogeneous
samples, arc melting was repeated several times. In order to eliminate the stress
and to obtain a homogeneous single-phase sample with large grains, the ingots
were annealed at appropriate temperatures for several days, depending on the series
of alloys.
The vapor pressures of P and As are too high to prepare the intermetallic
compounds containing P and/or As by means of arc melting. Therefore, instead of
arc melting, the ball-milling technique was used to prepare fine metallic powders
of these compounds. This technique is also suitable to accomplish a wide range of
chemical reactions. Milling devices include vibratory and planetary mills; products
include amorphous and nanocrystalline materials, and solid solutions. During
milling, solid-state reactions are initiated through repeated deformation and
fracture of the powder particles. In this thesis work, a solid-state reaction method
16 Chapter 3

was used for the preparation of the samples discussed in Chapters 5 and 6. First,
the mixture of starting materials with appropriate amounts was ball milled, then the
powder was sealed in a molybdenum crucible under argon atmosphere and placed
in a quartz ampoule for a couple of hours at a temperature at which the reaction
could take place. Subsequent annealing was performed at a temperature below the
reaction temperature.

Figure 3.1: Schematic representation of the high-energy vibratory mill
used in the present work.

The vibratory ball mill used in the present study is presented in Fig. 3.1. The
device consists of a stainless-steel vial with a hardened-steel bottom, the central
part of which consists of a tungsten-carbide disk. Inside the vial, a single hardened-
steel ball with a diameter of 6 cm is kept in motion by a water-cooled vibrating
frame. The amount of milled sample varied from about 5 to 10 g in this thesis
work. The device is evacuated during the milling down to a pressure of 10
-6
mbar
in order to prevent reactions with the gas atmosphere.
Experimental 17

3.1.2 Crystal growth

A single crystal of Gd
5
Si
1.7
Ge
2.3
was grown with the traveling-floating-zone
method in an adapted NEC double-ellipsoidal-type image furnace. A schematic
picture of the image furnace is shown in Fig. 3.2.

Figure 3.2: A schematic picture of the NEC SC-N35HD image furnace
(taken from [1]).

The furnace consists of two mirrors, which are plated with gold for enhanced
reflectivity and corrosion resistance. The heat sources are two halogen lamps. The
filaments of the two halogen lamps are positioned in the focus of each of the
mirrors, and are projected on the common focal point of the two mirrors. In this
way, the input power is concentrated on the molten zone between the feed and the
18 Chapter 3

seed. The temperature of the molten zone is controlled by controlling the dc-
voltage of the two lamps.
Feeds were prepared by arc-melting the pure starting materials into a button,
which was then cast into a cylindrical rod of 4 to 5 mm diameter. A quartz tube
served as growth chamber. Before the growth, the chamber was evacuated to a
pressure of 10
-6
mbar, and then filled with about 900 mbar Ar gas. During the
growth, the Ar atmosphere was continuously purified with Ti - Zr getter. The feed
and seed were counter-rotated with speeds of 20 rpm. The pulling speed of the
shafts was 3 mm/h for the Gd
5
Si
1.7
Ge
2.3
crystal. After the growth, the sample was
slowly cooled down to room temperature. In this way, a single crystal of
Gd
5
Si
1.7
Ge
2.3
was obtained with a diameter of 4 mm.

3.2 Sample characterization

Powder x-ray diffraction (XRD) patterns were taken at room temperature by means
of a Philips diffractometer with Cu K

radiation. In this way, the main phase as
well as the impurity phases can be detected, when the latter are present in amounts
of at least 5 vol. %. The crystal structure and the lattice parameters (with an
accuracy of 0.5 %) of the crystalline materials were analyzed by means of a
refinement procedure using Philips Xpert software. Laue x-ray back-scattering
diffraction was used to examine the single-crystallinity and to determine the
crystallographic directions of the single crystal. A Laue photo gives information on
the crystal quality of the surface over an area of about 1 mm
2
. By taking several
pictures at different positions information on single-crystallinity is obtained. The
crystallographic directions of the samples were determined by using the software
Orient Express [2].
Electron-probe microanalysis (EPMA) was used to check the homogeneity
and the stoichiometry, and also the single-crystallinity of the samples. The
measurements were performed in the JEOL JXA-8621 equipment [3] at the FOM-
ALMOS facility at the Kamerlingh Onnes Laboratory, University of Leiden.
Experimental 19

3.3 Magnetic measurements

A Quantum Design MPMS2-type SQUID magnetometer [4] was employed to
investigate the temperature and magnetic-field dependence of the magnetization.
This SQUID magnetometer is capable of measuring magnetization values in the
range of 10
-12
to 10
3
Am
2
with an accuracy of 0.1 % in the temperature range from
1.7 to 400 K and in the field range from 5 T to 5 T. The samples used for
magnetization measurements in the SQUID magnetometer are single crystals,
polycrystalline bulk pieces and powders.
Measurements of the magnetization loop in fields higher than 5 T and of the
magnetoresistance were performed in an Oxford Instruments MagLab system [5] in
the temperature range from 1.7 to 400 K and in the field range from 9 to 9 T. The
sensitivity of the system for magnetization measurements is 10
-6
Am
2
.

3.4 Specific-heat measurements

Specific-heat measurements on GdRu
2
Ge
2
at low-temperatures and in high-
magnetic field were performed in the Amsterdam 17 T specific-heat measurement
set-up [6]. In this set-up, the magnetic-field dependence of the specific heat can be
measured at temperatures between 300 mK and 90 K, by using the well known, and
reliable, semi-adiabatic method. Electrical heat pulses with a duration of 15 to 30 s
are applied to a sample holder, which is made of gold-plated cold-rolled silver. The
temperature before and after the heat pulse is monitored by a so-called combination
thermometer [7], which exhibits a very limited field dependence. The
3
He cryostat
is a closed system, working with a room-temperature gas-storage vessel, and a
cryopump for cooling the system down to 300 mK.
Specific-heat measurements on Gd
5
Si
1.7
Ge
2.3
were performed in the
temperature range from 4.2 to 300 K in a home-built set-up [8]. The accuracy of
the measurement is better than 1 % in the whole temperature range. The sample,
with a flat surface and a mass of about 100 mg, was fixed on a sapphire plate by N-
20 Chapter 3

type apiezon for good thermal contact. The temperature of the sample is monitored
by a Cernox resistance-temperature sensor over the whole range of temperatures. A
carbon-glass thermometer and a PID temperature controller are used to control the
environment temperature. In this way, semi-adiabatic measurements can be
performed. The magnetic-field dependence of the specific heat can be obtained by
positioning the system into a superconducting magnet system.
Because of its high ordering temperature, the specific-heat measurements of
MnFeP
0.45
As
0.55
(I) were performed in a PPMS system [9] at the Kamerlingh Onnes
Laboratory, University of Leiden. In this system, the specific heat can be measured
in the temperature range from 1.7 to 400 K.

3.5 Electrical-resistivity measurements

The electrical resistance was measured by means of the four-point method using an
Oxford Instruments MagLab system [5]. The system at the Van der Waals-Zeeman
Institute is equipped with a 9 T magnet and is capable of measurements between
1.7 K and 400 K. The ac-frequency range is from 1 Hz to 10 kHz. The maximum
input current (both ac and dc) is 250 mA. The sensitivity of the voltage
measurements is about 2 nV.
The electrical resistivity is obtained from the electrical resistance by

l
A
R = , (3.1)

where R is the electrical resistance, A the cross section of the sample
perpendicular to the current direction, and l the distance between the voltage
contacts. The magnetoresistance is obtained from the electrical-resistance
measurements in an applied magnetic field by

Experimental 21

) , 0 (
) , 0 ( ) , (
) , (
T B
T B T B
T B
=
=
=

, (3.2)

where B is the magnetic field and T is the absolute temperature.

3.6 Determination of the magnetocaloric effect

There are several ways to determine the MCE in a magnetic material
experimentally. Clark and Callen [10], Kuhrt et al. [11], and Ponomarev [12] have
directly measured the temperature of the sample with a thermocouple during the
application or removal of a magnetic field. For the principal scheme of these direct-
measurement methods and set-ups we refer to Dankov et al. [13]. The accuracy of
the direct measurements depends on the errors in thermometry, the errors in field
setting rates, the quality of thermal isolation of the samples, and the quality of the
compensation circuitry to eliminate the effect of the changing magnetic field on the
temperature sensors. As Pecharsky and Gschneidner [14] have pointed out, the
accuracy is claimed to be in the 5 to 10 % range. Larger errors will occur if one of
the above mentioned issues affecting the accuracy are not resolved properly. Other
techniques for determining the MCE are indirect. Indirect methods that are often
used include the ones based on magnetization measurements and on specific-heat
measurements in a constant magnetic field.

3.6.1 Determination of the magnetocaloric effect from magnetization
measurements

For magnetization measurements made at discrete temperatures, the integral in Eq.
(2.9) can be numerically evaluated by

,
) , ( ) , (
) , (
'
'
i
i i i
i i i i i i
m
B
T T
B T M B T M
B T S
=

(3.3)
22 Chapter 3

where ) , (
'
i i
B T M and ) , (
i i
B T M represent the values of the magnetization at a
magnetic field B
i
at the temperatures
'
i
T and T
i
, respectively. T is the mean value of
'
i
T and T
i
.
i
B is the step of field increase, and B = B
f
- B
0
. In the experiments that
we have conducted, the field is varied from B
0
= 0 to a field B = B
f
. The
accumulation of experimental errors in the determination of ) , ( B T S
m
has been
analyzed by Pecharsky and Gschneidner [15], and the validity of using this method,
even for a first-order magnetic phase transition, is discussed in Refs. 16 and 17.
Although a magnetization measurement by means of a SQUID magnetometer is the
most accurate method to determine the bulk magnetization of a magnetic material,
the accumulated errors in the determination of magnetic-entropy change
) , ( B T S
m
can be as high as 20 to 30 %, mainly because of the poor thermal
contact between the sample and the thermocouple. Nevertheless, this method is
often used to quickly establish the potential magnetocaloric properties of a
magnetic material.

3.6.2 Determination of the magnetocaloric effect from specific-heat
measurements

A specific-heat measurement is the most accurate method of determining heat
effects in a material. The total entropy change of a magnetic material can be
derived from the specific heat by using Eq. (2.12). According to Eq. (2.7), this
entropy change is equal to the magnetic-entropy change for an isobaric-isothermal
process. This means that we can also obtain the magnetic-entropy change from the
field dependence of the specific-heat measurements by using Eq. (2.12).
The determination of the absolute value of the adiabatic temperature change
in different magnetic materials is a rather complicated task. By combining Eqs.
(2.6), (2.8), and (2.10), the infinitesimal adiabatic temperature change for the
adiabatic-isobaric process is found to be
Experimental 23

.
) , (
) , (
,
dB
T
M
B T c
T
B T dT
p B p

= (3.4)

By integration of Eq. (3.4), the adiabatic temperature change for a field change
from B
i
to B
f
is given by

.
) , (
) , (
,
dB
T
M
B T c
T
B T T
f
i
B
B
p B p
ad

= (3.5)

Analytical integration of Eq. (3.5) is actually impossible since both the
magnetization and the specific heat are material dependent and generally unknown
functions of temperature and magnetic field in the vicinity of the phase transition.
Above the Debye temperature, the lattice specific heat of solids approaches the
Dulong-Petit limit of 3R. Therefore, at higher temperatures, if the specific heat can
be considered to be only weakly dependent on temperature, and the variation of T/
c
p
(T,B) with temperature is slow compared with the variation of the magnetization
with temperature, then, Eq. (3.5) can be simplified to

). , (
) , (
) , ( B T S
B T c
T
B T T
m
p
ad
= (3.6)

Obviously, MCE is large when ( )
p B
T M
,
/ is large and c
p
(T,B) is small at the same
temperature. Since ( )
p B
T M
,
/ peaks around the magnetic ordering temperature, a
large MCE is expected in the vicinity of the temperature of the magnetic phase
transition. The determination of the MCE from magnetization, specific heat, or the
combined magnetization and specific-heat data can be used to characterize the
magnetocaloric properties of magnetic refrigerant materials. Magnetization data
provides the magnetic-entropy change ) , ( B T S
m
. Specific heat at constant field
provides both magnetic-entropy change ) , ( B T S
m
and adiabatic temperature
24 Chapter 3

change ) , ( B T T
ad
. However, an analysis of experimental errors in the MCE as
derived from magnetization measurements and specific-heat measurements has
been shown that the accumulation of experimental errors may be as high as 20 % to
30 % near room temperature [15].

References

[1] N.P. Duong, Correlation between magnetic interactions and magnetic
structures in some antiferromagnetic rare-earth intermetallic compounds,
Ph.D. Thesis, University of Amsterdam (2002).
[2] OrientExpress, Version 2.03, ILL, Cyberstar S.A.
[3] http://www.cameca.fr/html/epma_technique.html.
[4] http://www.qdusa.com.
[5] http://www.science.uva.nl/research/mmm/eindex.html.
[6] http://www.science.uva.nl/research/mmm/17Tindex.html.
[7] J.C.P. Klaasse, Rev. Sci. Instrum. 68 (1997) 89.
[8] N.H. Kim-Ngan, Magnetic phase-transitions in NdMn
2
and related
compounds, Ph.D. Thesis, University of Amsterdam, 1993.
[9] http://www.physics.leidenuniv.nl/sections/cm/msm/welcome.htm.
[10] A.E. Clark and E. Callen, Phys. Rev. Lett. 23 (1969) 307.
[11] C. Kuhrt, T. Schitty and K. Brner, Phys. Stat. Sol. (a) 91 (1985) 105.
[12] B.K. Ponomarev, J. Magn. Magn. Mater. 61 (1986) 129.
[13] S.Yu. Dankov, A.M. Tishin, V.K. Pecharsky and K.A. Gschneidner, Jr.,
Rev. Sci. Instr. 68 (1997) 2432. S.Yu. Dankov, A.M. Tishin, V.K.
Pecharsky and K.A. Gschneidner, Jr., Phys. Rev. B 57 (1998) 3478.
387.
[15] V.K. Pecharsky and K.A. Gschneidner, Jr., J. Appl. Phys. 86 (1999) 565.
[16] K.A. Gschneidner, Jr., V.K. Pecharsky, E. Brck, H.G.M. Duijn and E.M.
Levin, Phys. Rev. Lett. 85 (2000) 4190.
[17] J.R. Sun, F.X. Hu and B.G. Shen, Phys. Rev. Lett. 86 (2000) 4191.

25

Chapter 4

Magnetic phase transitions and
magnetocaloric effect in Gd-based
compounds

4.1 Introduction

Magnetic cooling takes advantage of the entropy difference between the
magnetized and the demagnetized state of the working material. The entropy
change depends on both the properties of the material and the strength of the
applied magnetic field. The magnetocaloric effect (MCE) in heavy rare-earth
metals and their compounds has been studied intensively for some decades because
of the fact that they possess the largest magnetic moments and, therefore, the
largest available magnetic entropy. Among them, Gd
3+
has a
8
S
7/2
ground state and
the highest effective exchange coupling around room temperature. Thus, for
magnetic-cooling purposes, Gd metal and its compounds appear to be superior to
others in the sub-room-temperature range.
The discovery of the giant MCE in the ferromagnetic (FM) material
Gd
5
Si
2
Ge
2
[1] and the high magnetic fields that can be generated by high-energy-
product permanent magnets [2] have revived interest in magnetic cooling as a
technology competitive with vapor-cycle refrigeration. In the meantime, the basic
26 Chapter 4

understanding of the MCE and its relationship with the magnetic phase transitions,
and the methods of determination of MCE in the investigated materials have been
improved. In the second section of this chapter, we present a study of the magnetic
phase transition and the MCE in the compound GdRu
2
Ge
2
by means of both
magnetization and specific-heat measurements. In the third section, we present the
magnetic phase transitions and the dependence of the MCE in single-crystalline
Gd
5
Si
1.7
Ge
2.3
on the crystallographic directions.

4.2 GdRu
2
Ge
2

4.2.1 Introduction

Recently, large MCEs have been observed in materials that exhibit a magnetic
field-induced first-order phase transition. Pecharsky and Gschneidner [1] have
discovered the so-called giant MCE in Gd
5
Ge
2
Si
2
, originating from a first-order
structural and magnetic transition at T
C
= 276 K. Wada et al. [3] have observed a
large MCE in DyMn
2
Ge
2
, originating from two successive first-order phase
transitions at 36 and 40 K. The ternary rare-earth compounds of the type GdT
2
X
2
(T = transition metal; X = Si, Ge) have been studied intensively because of the
large variety of structural and physical properties shown by these phases [4, 5].
Duong [6] has studied the magnetic properties of GdT
2
Ge
2
(T = 3d, 4d) and found
that there are several types of magnetic phase transitions in these compounds. For
instance, GdRu
2
Ge
2
displays a field-induced magnetic phase transition at low field
strengths. In this section, we report on the magnetic phase transitions and the
determination of the MCE in this compound by means of specific-heat
measurements and magnetization measurements.

4.2.2 Experimental

A polycrystalline sample of GdRu
2
Ge
2
was prepared by repeatedly arc-melting
appropriate amounts of the starting materials with a purity of 99.9 wt.% and
Magnetic phase transitions and MCE in Gd-based compounds 27

subsequent annealing at 1073 K for two weeks in a 200 mbar Ar atmosphere. The
annealed sample was examined by x-ray diffraction (XRD) and found to be mainly
single phase with the tetragonal ThCr
2
Si
2
-type structure. The composition of the
sample was examined by electron-probe micro-analysis (EPMA) and found to be
mainly the stoichiometric composition GdRu
2
Ge
2
with a small amounts of second
phase of Gd oxide [6].
The temperature and the field dependence of the magnetization of the sample
was measured in a SQUID magnetometer and an Oxford Instruments MagLab
system in the temperature range from 5 to 300 K.
Specific-heat measurements were performed by means of a semi-adiabatic
heat-pulse method in the temperature interval from 300 mK to 70 K in fields of 0,
2, 4, 6, and 15 T. The magnetic-entropy changes in GdRu
2
Ge
2
were determined
from magnetization data by using Eq. (3.3) and from the specific-heat data by
using Eq. (2.12). Finally, the adiabatic temperature change T
ad
was obtained by
using Eq. (3.6).

4.2.3 Results and discussion

The temperature dependence of the specific heat of GdRu
2
Ge
2
, measured on a
polycrystalline sample, is shown in Fig. 4.1 in the representation c
p
/T vs T. Two
distinct peaks at T
1
= 29 K and T
2
= 33 K are observed in the zero-field specific-
heat curve, indicating two separate phase transitions exist in this compound. The
temperature dependence of the magnetization measured in a field of 0.1 T is shown
in Fig. 4.2. This result agrees with the result of the specific-heat measurements,

28 Chapter 4

0 10 20 30 40 50 60 70
0.0
0.2
0.4
0.6
0.8
1.0
1.2

0 T
2 T
4 T
6 T
15 T
GdRu
2
Ge
2
c
p
/
T

(
J
/
m
o
l
K
2
)
T (K)
T
1
T
2

Figure 4.1: Temperature dependence of the specific heat, plotted as c
p
/T vs
T, of GdRu
2
Ge
2
in zero field and in fields of 2, 4, 6, and 15 T. These results
have been taken from Ref. [6].

0 20 40 60 80 100
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
T
2
= 33 K T
1
= 29 K GdRu
2
Ge
2

M

(
A
m
2
/
k
g
)
T (K)

Figure 4.2: Temperature dependence of the magnetization of GdRu
2
Ge
2

determined on a sample cooled to 5 K in zero field. The measurement was
made on heating in a field of 0.1 T.


although the two separate transitions are less clear in the M(T) curve. The peak in
the c
p
/T curve at lower temperature T
1
still persists in a field of 2 T, but has become
markedly broadened and has disappeared in the curves measured in higher fields.
The peak at higher temperature has become invisible in the 2 T field curve due to
broadening. At about 34 K, the specific heat in a field of 2 T exceeds the value in
zero field. These results indicate that the observation of T
1
and T
2
strongly depends
on the strength of the applied field.

0
10
20
30
40
B = 0 T
GdRu
2
Ge
2

0 10 20 30 40 50 60
0
10
20
30
40
B = 2 T
4 T
6 T
15 T

S

(
J
/
k
g
K
)

T (K)

Figure 4.3: Total entropy of GdRu
2
Ge
2
as a function of temperature and
magnetic field, derived from specific-heat measurements.

The temperature dependence of the total entropy S in different fields was
obtained by integrating c
p
/T with respect to T by using Eq. (2.11). The results are
depicted in Fig. 4.3. The zero-field S(T) curve shows a small change at the
transitions. In the non-zero-field S(T) curves, no clear changes are found at the
transitions. The field dependence of the magnetization of GdRu
2
Ge
2
at 5 K, 15 K
and 31 K, measured with increasing field and subsequent decreasing field, is shown
30 Chapter 4

in Fig. 4.4. The M(B) curves measured at 5 K and 15 K show a rapid increase in
magnetization at 1 T and 0.8 T, respectively. The saturation moment at 9 T is about
7.4
B
/Gd-atom, which is close to the value for the free Gd
3+
ion. These results
show that the field-induced magnetic transition is of the antiferromagnetic to FM
type and that the transition is magnetically reversible. This transition is closely

0 2 4 6 8 10
0
1
2
3
4
5
6
7
8
increasing field
decreasing field
31 K
15 K
5 K

M

(
/
f
.
u
.
)
0
H (T)
GdRu
2
Ge
2

Figure 4.4: Field dependence of the magnetization of GdRu
2
Ge
2
at 5, 15
and 31 K, measured with increasing fie ld and subsequently with decreasing
field.

associated with the fact that the PM Curie temperature is positive (
p
= 37.2 K)
[6]. It means that the overall interaction between the Gd moments in this compound
is FM and that the antiferromagnetic ground state is rather instable. The broadening
of the transition in the c
p
/T(T) curves measured at H 0 can be ascribed to this
field-induced magnetic transition.
Figure 4.5 shows a set of magnetic isotherms of GdRu
2
Ge
2
, measured in the


0 1 2 3 4 5
0
10
20
30
40
50
60
70
80
T

=

5

K
75 K
5 K
GdRu
2
Ge
2

M

(
A
m
2
/
k
g
)
0
H (T)

Figure 4.5: Magnetic isotherms of GdRu
2
Ge
2
between 5 and 75 K,
measured with increasing magnetic field.

0 20 40 60 80 100
-1
0
1
2
3
4
5
6
GdRu
2
Ge
2
0-2 T from magnetization
0-4 T from magnetization
0-2 T from specific heat
0-4 T from specific heat

T (K)
-

S

(
J
/
k
g
K
)

Figure 4.6: Comparison of the magnetic-entropy changes of GdRu
2
Ge
2

derived from the magnetization and from the specific heat.

32 Chapter 4

temperature range from 5 to 75 K and fields up to 5 T, with increasing temperature
steps of 5 K. From these magnetization data, the isothermal magnetic-entropy
change
m
S has been derived by using Eq. (3.3). The results are shown in Fig. 4.6
together with the results obtained from the specific-heat measurements by using
Eq. (2.12).
m
S is seen to peak around 33 K, which is close to the magnetic-
ordering temperature. The negative values of -
m
S obtained below 20 K for a field
change from 0 to 2 T are due to the fact that the material is in the antiferromagnetic
state, in which the external field reduces the magnetic order rather than enhances
it. The maximal values of -
m
S are 1.7 J/kgK and 5.0 J/kgK in 2 T and 4 T,
respectively. The entropy changes associated with the two successive transitions
are only a small fraction of the maximum available magnetic entropy of Gd,
Rln(2J+1) = 110 J/kgK (J = 7/2 for Gd
3+
). The magnetic-entropy changes
determined by means of the two types of measurements agree quite well. This
confirms that magnetic measurements form a reliable method to determine the
isothermal magnetic-entropy change of magnetic materials.

0 10 20 30 40 50
-2
0
2
4
6

0 - 2 T
0 - 4 T
0 - 6 T
GdRu
2
Ge
2

T
a
d

(
K
)
T (K)

Figure 4.7: Adiabatic temperature change
ad
T in GdRu
2
Ge
2
between 5
and 50 K for magnetic field changes from 0 to 2, 0 to 4, and 0 to 6 T.

The temperature dependence of the adiabatic temperature change
ad
T in
the temperature range from 5 to 50 K upon field changes ranging from 0 to 2, 0 to
4 and 0 to 6 T has been derived from the specific-heat measurements by using Eq.
(3.6). The results are shown in Fig. 4.7. The maximum values of
ad
T are
approximately 1.5, 3.5 and 4.5 K for field change from 0 to 2, 0 to 4 and 0 to 6 T,
respectively. The profiles of
ad
T are similar to those of
m
S , although there is
some broadening of the peaks.

4.2.4 Conclusions

GdRu
2
Ge
2
orders antiferromagnetically below T
2
= 33 K. The antiferromagnetic
ground state is rather unstable. At 5 K, a field-induced transition occurs at a field of
1 T. This transition is closely associated with the overall interaction between the
Gd moments as indicated by the positive paramagnetic Curie temperature. We
found that the maximum value of the adiabatic temperature change
ad
T
is about
4.5 K in 6 T, which is a moderate MCE for a rare-earth compound in the
temperature range below 40 K. The results confirm that specific-heat and
magnetization measurements can both be employed to assess the MCE of a
magnetic material.

4.3 Single-crystalline Gd
5
Si
1.7
Ge
2.3

4.3.1 Introduction

The discovery of the giant MCE in Gd
5
Si
2
Ge
2
[1] has led to a revival of the
research dealing with magnetic refrigeration. This compound belongs to the
pseudo-binary system Gd
5
(Si
x
Ge
1-x
)
4
, in which the magnetic properties change
from antiferromagnetic to ferromagnetic (FM) upon increasing the Si content x.
The composition range 0.24 x 0.5 is of special interest since a giant MCE, giant
magnetoresistance [7], and colossal magnetostriction [8] are observed in this
34 Chapter 4

composition range. All these unusual features are related to a first-order magnetic
phase transition accompanied by a structural transition from the low-temperature
orthorhombic FM state to the high-temperature monoclinic paramagnetic (PM)
state. This magneto-structural transition can be induced by temperature and/or by
magnetic field.
A better understanding of the nature of the first-order phase transition in the
Gd
5
(Si
x
Ge
1-x
)
4
system is of fundamental and practical importance. Especially the
relation between the structural and magnetic phase transitions from the low-
temperature orthorhombic FM state to the high-temperature monoclinic PM state
and the giant MCE in the Gd
5
(Si
x
Ge
1-x
)
4
alloys is intriguing. Choe et al. [9] have
studied the formation and breaking of the covalent bonds between Si(Ge) and
Ge(Si) atoms in Gd
5
Si
2
Ge
2
, and have pointed out that the structural transition
occurs by a shear mechanism in which the (Si,Ge)-(Si,Ge) dimers increase their
distance by 0.859(3) which leads to twinning. The structural transition changes
the electronic structure and provides on micro-structural level an explanation of the
change in magnetic behavior with temperature in this system. Both the magnetic
and the crystal structure are easily affected by temperature and/or magnetic field,
indicating a strong coupling between the magnetism and the lattice. The changes in
the magnetic and crystallographic parameters of such a system may lead to unusual
phenomena, such as unusual magnetic behavior and the spontaneous generation of
an electrical voltage in Gd
5
(Si
x
Ge
1-x
)
4
during the transition [10].
In order to obtain more insight into the mechanism of this unusual physical
behavior and the MCE in the Gd
5
(Si
x
Ge
1-x
)
4
system, we have grown a single crystal
of Gd
5
Si
1.7
Ge
2.3
and studied the magnetic and magnetocaloric properties and their
relationship with the structural and magnetic phase transitions.

4.3.2 Crystal growth and characterization

A single crystal of Gd
5
Si
1.7
Ge
2.3
was grown by means of the traveling-floating-zone
method in an adapted NEC double-ellipsoid image furnace. The starting materials

were 4N Gd (from Ames Lab., USA), 6N Si and 6N Ge. The crystal was grown
under an Ar atmosphere of 900 mbar with a speed of 3 mm/h. The feed and seed
were counter-rotated at 22 and 31 rpm, respectively. The characteristic region
between 2 = 20 and 40

of a powder XRD pattern of Gd
5
Si
1.7
Ge
2.3
is shown in
Fig. 4.8. Si powder was added as internal standard. The diagram was indexed
within the monoclinic structure (space group P112
1
/a) with the unit-cell parameters
a = 7.585 , b = 14.800 , c = 7.777 , = 93.29

. The unit cell contains four
formula units and has a volume = 871.6
3
. The molar volume V
m
equals 1.312
x 10
-4
m
3
/mol. These crystallographic data are in good agreement with literature
values [11,12].

26 28 30 32 34 36 38 40
Si
Cu K
Gd
5
Si
1.7
Ge
2.3

I
n
t
e
n
s
i
t
y

(
a
r
b
.

u
n
i
t
)
2 (deg.)

Figure 4.8: XRD pattern of the Gd
5
Si
1.7
Ge
2.3
collected at room temperature.
The open circles represent the observed data and the lines represent the
calculated XRD pattern. The vertical bars indicate the calculated positions
of the Bragg reflections for Cu K
a 1
radiation. The difference between the
experimental and calculated intensities is shown at the bottom as a solid
line.

36 Chapter 4

Figure 4.9: EPMA micrograph of a Gd
5
Si
1.7
Ge
2.3
single crystal.

The as-grown single crystal was checked as regards composition,
homogeneity and single-crystallinity by means of EPMA and x-ray Laue
backscattering. The EPMA micrograph as presented in Fig. 4.9 shows that the
crystal is homogeneous. A slight gradient of the Si content has been detected along
the growth direction of the crystal. The average composition of the crystal is 56.3
at. % Gd, 18.4 at. % Si and 25.3 at. % Ge, which corresponds to the actual formula
Gd
5.06
Si
1.66
Ge
2.28
.

4.3.3 Magnetic properties

Figure 4.10 shows the temperature dependence of the magnetization of single-
crystalline Gd
5
Si
1.7
Ge
2.3
, measured in a magnetic field of 50 mT and 5 T,
respectively, applied along the three principal crystallographic axes. The magnetic
ordering is observed as a pronounced change in magnetization, which is clearly
different from a second-order FM phase transition, indicating that the transition is
of first order. The Curie temperature T
C
, determined as the temperature


0 50 100 150 200 250 300 350 400
0
10
20
30
40
0
1
2
3
4
5
5 T
B //a-axis
//b-axis
//c-axis

M

(
B
/

f
.
u
.
)
T (K)

M

(
B
/

f
.
u
.
)
50 mT
B // a-axis
// b-axis
// c-axis
Gd
5
Si
1.7
Ge
2.3

Figure 4.10: Temperature dependence of the magnetization of Gd
5
Si
1.7
Ge
2.3

in a field of 50 mT (top panel) and 5 T (bottom panel), respectively,
measured with the field direction along the three principal axes a ([100]), b
([010]), and c ([001]).

corresponding to the extreme of dM/dT, is 240.2 K. In the FM state, the M(T)
curves measured for Gd
5
Si
1.7
Ge
2.3
in a field of 50 mT display anisotropic behavior.
This anisotropy in the M(T) curves of Gd
5
Si
1.7
Ge
2.3
disappears in a field of 5 T as
indicated in the bottom panel of Fig. 4.10. This is, therefore, possibly due to
domain-wall displacement and/or rotation of the magnetization of the domains,
while one cannot exclude that it is related to a spin reorientation. In the
paramagnetic state, the temperature dependence of the inverse dc magnetic
susceptibility of Gd
5
Si
1.7
Ge
2.3
obeys the Curie-Weiss law in the higher-temperature
region with an effective moment
eff
of 8.2, 8.4, and 8.6
B
/Gd-atom for the a, b,
and c direction, respectively. These values are slightly larger than the theoretical
38 Chapter 4

value of 7.94
B
for a free Gd
3+
ion and the experimental value of 7.98
B
for Gd
metal. The PM Curie temperature
p
equals 204 K.
In order to determine the change of the Curie temperature with applied
magnetic field and the thermal hysteresis, we have performed measurements of the
temperature dependence of the magnetization with increasing and decreasing
temperature in fields of 1, 2, 3, 4 and 5 T. The results are shown in Fig. 4.11. It is
clear that the Curie temperature increases with increasing field and that there is a
large thermal hysteresis.

0
10
20
30
0
10
20
30
0
10
20
30
180 200 220 240 260 280 300
0
10
20
30
0
10
20
30
B //a-axis
Gd
5
Si
1.7
Ge
2.3
1 T

M

(
B
/
f
.
u
.
)
2 T

3 T

T (K)
5 T
4 T

Figure 4.11: Temperature dependence of the magnetization of
Gd
5
Si
1.5
Ge
2.3
, measured with increasing and decreasing temperature in a
field of 1, 2, 3, 4, and 5 T, respectively.

Figure 4.12 shows the magnetic isotherms of single-crystalline Gd
5
Si
1.7
Ge
2.3

measured at 5 K and at several temperatures around T
C
, measured with the
magnetic-field direction along the three principal crystallographic axes [100], [010]


0 1 2 3 4 5 6
0
15
30
0
15
30
0
15
30

260 K
257.5 K 250 K 240 K
5 K
B // c - axis
0
H(T)
240 K

5 K
230 K
252.5K
257.5 K
260 K
B // a- axis
M

(
B
/
f
.
u
.
)
B
c1
B
c2
B
c3
B
c4

250 K
257.5 K
240 K
260 K
265 K
230 K
5 K
B // b- axis

Figure 4.12: Magnetic isotherms of Gd
5
Si
1.7
Ge
2.3
along the three principal
axes at 5 K and at several temperatures, which are in the vicinity of the
Curie temperature, measured with increasing () and decreasing () field.

and [001] with increasing and decreasing magnetic fields. The spontaneous
magnetization at 5 K is 7.08, 7.18, and 7.28
B
/Gd-atom along the a, b, and c axis,
respectively. These values are slightly larger than the value of 7.0
B
/Gd-atom for
the free Gd
3+
ion. Below T
C
, the magnetization curves show FM behavior. Above
T
C
, the magnetization increases linearly in low field strengths, which is
characteristic for simple PM behavior. Above a lower field B
c1
, the magnetization
40 Chapter 4

increases sharply and saturates at a higher field B
c2
, indicating that the applied
magnetic field gives rise to a field-induced PM to FM phase transition. With
decreasing field, a reverse magnetic phase transition FM to PM starts at the field
B
c3
and ends at the field B
c4
. The field hysteresis observed in the field dependence
of the magnetization, which is about 1 T, also indicates that the transition is of first
order.

236 240 244 248 252 256 260 264
0
1
2
3
4
5
data from
M-T, B//[100]
M-T, [100]
M-B, [001]
M-B, [001]
M-B, [010]
M-B, [010]
M-B, [100]
M-B, [100]
Gd
5
Ge
2.3
Si
1.7
FM
PM

B

(
T
)
T (K)
Figure 4.13: Magnetic phase diagram of Gd
5
Si
1.7
Ge
2.3
.() and () indicate
the data from the temperature dependence of the magnetization. (, ), (,
) and (, ) indicate the data from the field dependence of the
magnetization, measured along the [001], [010] and [100] direction,
respectively. The solid lines are guides to the eye.

The magnetic phase diagram of Gd
5
Si
1.7
Ge
2.3
constructed from the
temperature dependence of the magnetization and the field dependence of the
magnetization, measured with the field direction along three principal
crystallographic axes. The results are shown in Fig. 4.13. The critical-field values
were taken as the mean value of B
c1
and B
c2
for increasing field and, B
c3
and B
c4
for
decreasing field, respectively. Both sets of experimental data are in good

agreement with each other and there are no clear differences along the three
directions. T
C
almost linearly increases with increasing field at a rate of 4.4 K/T.
Extrapolation of the temperature dependence of the critical-field lines to zero field
shows that the zero-field Curie temperature T
C
of Gd
5
Si
1.7
Ge
2.3
is 240.1 K for a
measurement with increasing temperature and 235.3 K for decreasing temperature,
respectively. It indicates that a thermal hysteresis of about 5 K exists between the
increasing- and decreasing-temperature measurements. This is in good agreement
with the result observed in thermal-expansion measurements [13].

4.3.4 Specific heat

Usually, the specific heat of a material at constant pressure behaves anomalously
near the magnetic phase transition and hence measurements of the specific heat can
be a useful tool for studying the nature of a given magnetic phase transition. Figure
4.14 shows the temperature dependence of the specific heat of Gd
5
Si
1.7
Ge
2.3
,
measured with increasing temperature in zero field.

0 50 100 150 200 250 300
0
1
2
3
4
5
6
7
8
zero field
i ncreasi ng T
Gd
5
Si
1. 7
Ge
2.3
D
= 237 K
= 32. 3 mJ/ mol K
2
T
C
= 239 K

c
p
/
T

(
J
/
m
o
l
K
2
)
T ( K)

Figure 4.14: Temperature dependence of the specific heat of Gd
5
Si
1.7
Ge
2.3
,
in the representation c
p
/T vs T, measured with increasing temperature in
zero field. The solid line is a Debye fit with
D

= 237 K.
42 Chapter 4

190 200 210 220 230 240 250 260 270
0
1
2
3
4
5
6
7
8
T '
C
= 246 K
T
C
= 239 K
0 T
2 T
Gd
5
Si
1.7
Ge
2.3

c
p
/
T

(
J
/
m
o
l
K
2
)
T (K)

Figure 4.15: Temperature dependence of the specific heat of Gd
5
Si
1.7
Ge
2.3
,
in the representation c
p
/T vs T, measured with increasing temperature in
zero field and in 2 T.

The peak position corresponds to a T
C
of about 239 K, which is slightly smaller
than the value of T
C
obtained from magnetic measurement. The sharpness and the
large amplitude of the peak suggest that the phase transition in Gd
5
Si
1.7
Ge
2.3
is of
first order. A fit of the low-temperature data to the formula c
p
/T = + T
2
yields
an electronic contribution to the specific heat = 32.3 mJ/molK
2
and a Debye
temperature
D
= 237 K.
Figure 4.15 shows the specific heat of Gd
5
Si
1.7
Ge
2.3
measured across the
phase-transition region in zero field and in 2 T with increasing temperature. The
magnetic field suppresses the magnetic part of the specific heat and shifts the peak
position to a higher temperature
'
C
T , which is about 246 K at 2 T. This value is
slightly smaller than the value of T
C
(in 2 T) obtained from the magnetic
measurement.

4.3.5 Magnetocaloric effect

The magnetic-entropy changes of Gd
5
Si
1.7
Ge
2.3
, which have been derived from the
magnetic isotherms measured with increasing temperature and increasing field
along the three principal axes by using Eq. (3.3), are displayed in Fig. 4.16.

200 220 240 260 280 300 320
0
10
20
30
40
50
// c-axis 0 - 1 T
0 - 2 T
0 - 3 T
0 - 4 T
0 - 5 T

T (K)
0
10
20
30
40
50
// a-axis

0 - 1 T
0 - 2 T
0 - 3 T
0 - 4 T
0 - 5 T
-

S
m

(
J
/
k
g
K
)
0
10
20
30
40
50
// b-axis
B
0 - 1 T
0 - 2 T
0 - 3 T
0 - 4 T
0 - 5 T

Figure 4.16. Temperature dependence of the magnetic-entropy changes of
Gd
5
Si
1.7
Ge
2.3
along the three principal crystallographic axes, for field
changes from 0 to 1, 0 to 2, 0 to 3, 0 to 4 and 0 to 5, derived from the
magnetization data.
44 Chapter 4

0 50 100 150 200 250
0
50
100
150
200
250
300
350
400
Gd
5
Si
1.7
Ge
2.3

S

(
J
/
k
g
K
)
T (K)

Figure 4.17: Temperature dependence of the total entropy of Gd
5
Si
1.7
Ge
2.3
determined from the zero-field specific heat. The inset shows an enlarged
view of the entropy at the transition.

The shape of -
m
S consists of a spike and a plateau part. The spike is probably
related to an irreversible magnetization process and most likely associated with the
fact that the magnetic transition occurs simultaneously with the crystal-structure
change, and it boosts the value of -
m
S to higher values. With increasing field, the
plateau part saturates and extends to higher temperatures.
The entropy evolution as a function of temperature,
= TdT T c T S
p
/ ) ( ) (
,
of the system can directly be obtained from the specific-heat data. The S(T) curve
in zero field is depicted in Fig. 4.17. The entropy change associated with the
transition is about S = 11 J/kgK and the latent heat L = T
C
S = 2.63 kJ/kg.

230 235 240 245 250 255
330
335
340
345
350
355
360
365
370

S

(
J
/
k
g
K
)
T (K)

4.3.6 Discussion and conclusions

First, we discuss the magnetic phase transition and the magnetic interactions in
Gd
5
Si
1.7
Ge
2.3
. On the basis of the observed discontinuous behavior of the
magnetization and the entropy at the transition, and the constructed magnetic phase
diagram of Gd
5
Si
1.7
Ge
2.3
, we conclude that the phase transition observed in this
material is a first-order phase transition. From the specific-heat measurements, we
have determined the latent heat involved in this phase transition is about 2.63
kJ/kgK. According to the study reported in Ref. 8, the origin of the transition is a
simultaneous structural and magnetic phase transition. The crystal structure adopts
the orthorhombic structure in the FM state and changes into the monoclinic
structure in the PM state. The major crystallographic structure change occurs due
to the breaking of covalent-like Si-Si, Si-Ge and Ge-Ge bonds at the transition
from the FM state to the PM state [9,14].
The large effective magnetic moment, the abrupt change in the magnetization
at the transition, and the anisotropy observed in Gd
5
Si
1.7
Ge
2.3
cannot well be
explained in the framework of the indirect RKKY 4f-4f exchange interaction.
Therefore, there may exist other exchange interactions that play an important role
in this compound. One possible exchange interaction in this compound may be the
indirect exchange between the 4f-electron spins via polarization of the 5d-electron
spins [15]. The experimental observation of saturation magnetic moments at 5 K
that are slightly larger than the Gd free-ion moment and the somewhat enhanced
effective magnetic moment support the occurrence of polarized 5d-electron spins in
this compound. Another possible interaction is a Gd-Si(Ge)-Gd superexchange
interaction in the low-temperature FM phase propagating through the interlayer
covalent-like bonds [15]. The fact that the long-range FM order is abruptly
destroyed at the transition where the material becomes PM is because of the
breaking Si(Ge)-Ge(Si) bond between the slabs occurs at the structural
transformation which leads to the disappearance of the superexchange interaction.
46 Chapter 4

However, the determination of these interactions is complicated because they are
strongly dependent on composition and temperature [9].
Secondly, we discuss the MCE observed in this compound. The MCE in this
phase-transition region is extremely large. Not only the magnitude of the MCE is
large, but also the full width at the half maximum of the MCE with respect to the
field change is large. The maximum value of -
m
S for a field change from 0 to 5
T is 44.6 J/kgK at the spike and around 30 J/kgK at the plateau part. These values
are consistent with the results reported earlier on polycrystalline material [16]. If
we take 30 J/kgK as maximum value of the magnetic-entropy change -
m
S (max)
then the full width at the half maximum (
FWHM
T = T
2
T
1
) [16] is about 19 K for
a field change from 0 to 5 T. A recent study [17] has shown that the plateau part
perfectly matches the S values given by the Clausius-Clapeyron equation
S=MdB
c
/dT (where M is the jump of the magnetization at the magneto-
structural transition, and dB
c
/dT is the rate of critical-field change with
temperature). From the linear relation between B
c
and T, we have obtained dB
c
/dT
= 0.23 T/K. If we take the value of M at the T
C
as M , then we obtain S as 32
J/kgK for a field change from 0 to 1 T and 35 J/kgK for a field change from 0 to 5
T. These values roughly match with the values of S on the plateau part shown in
Fig. 4.16. However, we should mention that the value of dB
c
/dT can be easily
determined from the phase diagram (see Fig. 4.13), but the determination of the
jump of the magnetization at the transition is complicated because the first-order
transition does not occur infinitely fast. One may determine M in different ways
from the M(T) curve or from the M(B) curve and may obtain different values.
Finally, we discuss the magnetic anisotropy in this compound. The magnetic
anisotropy of the Gd-based compounds is usually negligible due to the spherical
symmetry of the 4f orbitals of Gd
3+
ions and the isotropic nature of the RKKY
interaction. There are, however, some indications for anisotropic behavior in this
material. First, the temperature dependence of the magnetization in low field

exhibits an anisotropic behavior. As we discussed this anisotropic behavior may
belong to the domain-wall displacement and/or rotation of the magnetization of the
domains. However, the anisotropy may arise from anisotropic exchange
interactions. Duijn [18] has proposed a possible mechanism for the occurrence of
anisotropy in the Gd
5
(Si
1- x
Ge
x
)
4
compounds. Our magnetic measurements show
that the magnetic anisotropy is negligibly small and/or has only a minor effect on
the magnetization along three principal crystallographic axes as well as the MCE in
the compound Gd
5
Si
1.7
Ge
2.3
, as indicated by the similar magnetic behavior and
magnitude of -
m
S obtained along the three principal axes. It should be noted that
these magnetic parameters are determined for the monoclinic structure, while the
FM phase adopts the orthorhombic structure. The other anisotropic behavior in this
compound that should be mentioned is found in the thermal-expansion
measurements along the principal crystallographic axes that have been performed
on the same single crystal [13]. They show that the magnetic and structural phase
transition occur at one and the same temperature. The thermal expansion shows a
pronounced anisotropy between the bc-plane and the a-axis. The resulting steps in
L L / for the b- and c-axis attain negative values of 2.0 x 10
-3
and 2.1 x 10
-3
,
upon heating, respectively, while for the a-axis the step is positive and much larger,
6.8 x 10
-3
. The volume change V V / at T
C
is positive and amounts to 2.7 x 10
-3
.
By combining the specific-heat and the thermal-expansion data and by making use
of the Clausius-Clapeyron relation, we extract a hydrostatic pressure dependence of
T
C
equal to dT
C
/dp = 3.2 0.2 K/kbar. This value is in good agreement with the
value dT
C
/dp = 3.46 K/kbar extracted from thermal-expansion measurements under
hydrostatic pressure for a Gd
5
Si
1.8
Ge
2.2
sample

[8]. The result points out that the
pressure effect is strongly anisotropic. Uniaxial pressure along the a-axis enhances
T
C
, while uniaxial pressure in the bc-plane suppresses T
C
. This provides important
information how to chemically substitute the system in order to further enhance T
C
.
In conclusion, we have studied the magnetic and magnetocaloric properties
of a single crystal of Gd
5
Si
1.7
Ge
2.3
. The bulk-property measurements show that the
48 Chapter 4

unusual magnetic properties and the giant MCE in this compound is associated
with a simultaneous magnetic and structural transition. This phase transition is of
first order. The magnetic properties observed in this compound suggest that not
only the indirect RKKY exchange interaction, but also the indirect 4f-electron
spins coupling via polarization of 5d-electron spins and the Gd-Si(Ge)-Gd
superexchange interaction may play important roles in governing the magnetic
properties of Gd
5
Si
1.7
Ge
2.3
. The MCE in this compound system is large. There are
some indications of the magnetic anisotropy in this compound. But the magnetic
anisotropy has a negligible effect on the MCE in this compound. However, an
accurate determination of the magnetic structure, the magnetic interactions and the
microstructure of the Gd
5
Si
1.7
Ge
2.3
compound are still required for a full
understanding of the unusual magnetic behavior observed in this interesting alloy
system.

References

[3] H. Wada, Y. Tanabe, K. Hagiwara and M. Shiga, J. Magn. Magn. Mater. 218
(2000) 203.
[4] A. Szytula and J. Leciejewicz, Magnetic properties of ternary intermetallic
compounds of the RT
2
X
2
type, Ch. 83 of Handbook on the physics and
chemistry of rare earths, Vol. 12, K.A. Gschneidner, Jr., and Eyring Eds.
(North-Holland, Amsterdam, 1998).
[5] A. Szytula, Magnetic properties of ternary intermetallic compounds, Ch. 2 of
Handbook of Magnetic Materials, Vol. 6, K.H.J. Buschow Ed. (North-
Holland, Amsterdam, 1991).
[6] N.P. Duong, Correlation between magnetic interactions and magnetic
structures in some antiferromagnetic rare earth intermetallic compounds,
Ph.D. Thesis, University of Amsterdam (2002).
[7] L. Morollon, J. Stankiewicz, B. Garcia-Landa, P.A. Algarabel and M.R.
Ibarra, Appl. Phys. Lett. 73 (1998) 3462.

[8] L. Morollon, P.A. Algarabel, M.R. Ibarra, J. Blasco, B. Garcia-Landa, Z.
Arnold and F. Albertini, Phys. Rev. B 58 (1998) R14721.
[9] W. Choe, V.K. Pecharsky, A.O. Pecharsky, K.A. Gschneidner, Jr., V.G.
Young, Jr., and G.J. Miller, Phys. Rev. Lett. 84 (2000) 4617.
[10] E.M. Levin, V.K. Pecharsky and K.A. Gschneidner, Jr., Phys. Rev. B 63
(2001) 174110.
[11] V.K. Pecharsky and K.A. Gschneidner, Jr., J. Alloys Compds. 260 (1997) 98.
[12] V.K. Pecharsky, A.O. Pecharsky and K.A. Gschneidner, Jr., J. Alloys
Compds. 344 (2002) 362.
[13] M. Nazih, A. de Visser, L. Zhang, O. Tegus and E. Brck, Solid State
Commun. 126 (2003) 255.
(1999) 7993.
(2000) R14625.
387.
[17] F. Casanova, X. Batle, A. Labarta, J. Marcos, L. Manosa and A. Planes,
Phys. Rev. B 66 (2002) 10040 (R).
[18] H.G.M. Duijn, Magnetotransport and magnetocaloric effects in intermetallic
compounds, Ph.D. Thesis, University of Amsterdam (2000).

50 Chapter 4

51

Chapter 5

Magnetocaloric effect in hexagonal
MnFeP
1-x
As
x
compounds

5.1 Introduction

Since the discovery of the giant magnetocaloric effect (MCE) in Gd
5
Ge
2
Si
2
at
Ames Laboratory [1], the research on magnetic refrigerant materials has been
strongly intensified worldwide. Currently, most research groups study the MCE in
rare-earth-based materials because the large moments of the rare-earth atoms imply
the possibility of large MCE. However, especially for the important applications
around room temperature only a very limited number of rare-earth compounds
(usually the ordering temperature of the rare earth compounds is below room
temperature) are suitable because the MCE is optimal around the magnetic-
ordering temperature. The largest MCE known so far in rare-earth materials near
room temperature is observed in Gd metal. The maximum MCE in Gd occurs at the
temperature where it orders ferromagnetically (294 K). When the magnetic field
changes from 0 to 1.5 T, the MCE is about 4 K, and it is 11 K when the magnetic
field changes from 0 to 5 T [2]. It is unclear whether the giant-MCE material
Gd
5
Ge
2
Si
2
, which was reported to exceed the reversible MCE in any known
magnetic material by at least a factor of two, will be suitable for practical
application. Because of the low Curie temperature (about 276 K) and the relatively
large thermal hysteresis, this may not be the case.
52 Chapter 5

On the other hand, a number of transition-metal-based materials, such as
FeRh [3, 4], MnP
1-x
As
x
[5], Mn
2
Sb [6], MnAs
1-x
Sb
x
[7], and La
1-x
Ca
x
MnO
3
[8, 9]
have been investigated with respect to their MCE. In general, the MCE in
transition-metal-based materials is lower than in rare-earth alloys in the same
temperature range. Interestingly, FeRh exhibits an unusual and irreversible, but
giant, MCE as large as 13 K at 307 K for a field change from 0 to 1.95 T. This
effect is related to a first-order metamagnetic phase transition. The lanthanum-
manganese perovskite oxides La
1-x
Ca
x
MnO
3
, known as colossal-magnetoresistance
(CMR) material, also show considerable magnetic-entropy changes. In these CMR
materials, only the Mn ions have magnetic moments. However, Pecharsky et al.
[10] and Sun et al. [11] have pointed out that the CMR materials do not seem
promising candidates for magnetic refrigeration as previously claimed in many
reports, because of their relatively small adiabatic temperature change.
The results on the above-mentioned Mn-based materials have in common
that their MCE can be rather large. This motivated us to study the MCE in the
vicinity of the first-order metamagnetic phase transition in other Mn-based
materials. These materials have various phase transitions and frequently order
around room temperature, and Mn ions can have relatively large magnetic
moments compared to other transition metals. However, the magnetic moments of
Mn are generally about two times smaller than those of the heavy rare-earth
elements. Enhancement of the MCE associated with magnetic-moment alignment
may be achieved through the induction of a first-order phase transition.
Among Mn-based compounds, the hexagonal MnFeP
1-x
As
x
compounds that
are stable for 0.15 < x < 0.66 exhibit peculiar magnetic properties [12-14]
associated with a first-order metamagnetic transition. Our recent studies [15,16]
have shown that MnFeP
1-x
As
x
compounds possess a large magnetic-entropy change
with the same magnitude as Gd
5
Ge
2
Si
2
. This result is of significant importance,
because it not only makes these compounds attractive candidates for working
materials in magnetic refrigeration but also indicates significant progress in the
MCE in MnFeP
1-x
As
x
compounds 53

search for new magnetic refrigerant materials. In this chapter, we report on a
detailed study of the magnetic and magnetocaloric properties of the hexagonal
MnFeP
1-x
As
x
compounds.

5.2 Sample preparation and characterization

Polycrystalline samples of MnFeP
1-x
As
x
compounds with nominal compositions x
= 0.25, 0.35, 0.45, 0.50, 0.53, 0.55 and 0.65 were synthesized by a solid-state
reaction. The starting materials used in our sample preparation are the binary
compounds Fe
2
P (purity 99.5 %, Alfa Aesar) and FeAs
2
(purity 99.5 %, Alfa
Aesar), pure Mn chips (purity 99.99 %) and red-P powder (purity 99.99 %). In
order to obtain homogeneous samples, appropriate proportions of the starting
materials were ball milled in a high-energy vibratory ball-mill before the solid-state
reaction.
Figure 5.1 presents the x-ray diffraction (XRD) patterns of MnFeP
0.5
As
0.5

after various periods of milling. With increasing milling time, the characteristic

30 35 40 45 50 55 60
M
n
M
n
M
n
M
n
F
e
2
A
s
F
e
2
A
s F
e
2
A
s
F
e
2
A
s
F
e
2
A
s
F
e
2
A
s
MnFeP
0.5
As
0.5
1 h
10 h
20 h
40 h
90 h

I
n
t
e
n
s
i
t
y

(
a
.
u
.
)
2

Figure 5.1: XRD patterns of MnFeP
0.5
As
0.5
after various periods of milling.
54 Chapter 5

peaks of the starting materials become broadened. After 90 hours milling, only the
broad profile of Mn is visible. For complete homogenization, all the samples are
milled up to 200 hours.
The solid-state reaction was performed in a molybdenum crucible. First, the
obtained mixture was sealed in the crucible in a 100 - 200 mbar Ar atmosphere.
Then, this crucible was heated at 1273 K for 100 hours, followed by a
homogenization process at 923 K for 120 hours. Finally, the crucible was slowly
cooled down to ambient conditions.
The powder XRD patterns of the samples show that the MnFeP
1-x
As
x

compounds with x = 0.25, 0.35, 0.45, 0.50, 0.53, 0.55 and 0.65 crystallize in the
hexagonal Fe
2
P-type of structure (space group m P 2 6 ) with a small amount of
MnO as a second phase. From the broadening of the XRD patterns, the mean grain
size is estimated to be about 100 nm. The homogeneity and stoichiometry of the
samples with x = 0.45, 0.53, and 0.55 were checked by means of electron-probe
microanalysis (EPMA). Also the MnFeP
0.45
As
0.55
sample contains an extra phase
with Mn, which is probably MnO as detected by XRD. The MnFeP
0.47
As
0.53
sample
also contains an extra phase with Mn. The actual composition of the main phase of
MnFeP
0.47
As
0.53
is 30.1 at. % Mn, 34.5 at. % Fe, 16.0 at. % P and 18.8 at. % As,
which corresponds to the formula Mn
0.93
Fe
1.04
P
0.48
As
0.56
. The composition of the
sample is very sensitive to the starting materials and the preparation process. For
the composition x = 0.55, we used two samples for the measurements presented in
this chapter. Because their properties are slightly different, one is indicated as
MnFeP
0.45
As
0.55
(I), and the other as MnFeP
0.45
As
0.55
(II).

5.3 Structural properties

The MnFeP
1-x
As
x
compounds crystallize in three different types of structures: the
orthorhombic Co
2
P type (Pnma, No. 62) for low As contents (0 x 0.15), the
hexagonal Fe
2
P type ( m P 2 6 , No. 189) for intermediate As contents (0.15 < x
MCE in MnFeP
1-x
As
x
compounds 55

0.66), and the tetragonal Fe
2
As type (P4/nmm, No. 129) for the highest As contents
(0.66 < x 1) [14]. In the present thesis, only the isostructural hexagonal series of
compounds is considered in which the large MCEs are observed.
A schematic drawing of the structure of the hexagonal MnFeP
1-x
As
x

compounds is shown in Fig. 5.2. There are two different metal sites: Fe(3f) at the
tetragonal position (x
1
, 0, 0) with local symmetry m2m and Fe(3g) at the pyramidal
position (x
2
, 0, 1/2) with the same local symmetry. There are also two different
non-metal sites: one two-fold position P(2c) at (1/3, 2/3, 0) and one single position
P(1b) at (0, 0, 1/2) [14]. Substitution of Mn for Fe in this structure leads to
preferential Mn occupation of the 3g sites and substitution of As for P leads to a
random distribution of As over the 1b and 2c sites. Therefore, the MnFeP
1-x
As
x

compounds magnetically consist of two basal planes alternating along the
hexagonal c-axis, one containing the Mn atoms and the other one containing the Fe
atoms. The shortest Mn-Mn distance within the Mn-layer is d
Mn-Mn
=
a
2
2 2
3 3 1 x x +
[17], in which a represents the lattice parameter. It should be

Figure 5.2: Schematic representation of the Fe
2
P-type of structure. The
volume shown contains three unit cells.
56 Chapter 5

Table 5.1: Lattice parameters determined by means of XRD at room
temperature, and Curie temperatures of MnFeP
1-x
As
x
compounds.

Nominal, x 0.25 0.35 0.45 0.50 0.53 0.55(I) 0.65
a () 6.0392 6.0677 6.1080 6.1290 6.1628 6.1739 6.2120
c () 3.4870 3.4874 3.4900 3.4805 3.4946 3.4511 3.4633
c/a 0.5774 0.5748 0.5714 0.5679 0.5670 0.5590 0.5575
T
C
(K) 168 213 240 282 290 300(I)
307(II)
332

noted that, because 1/2 < x
2
< 2/3, only the a axis and not the b axis is a two-fold
axis. The structural parameters x
1
and x
2
depend on the constituents of the
compound under consideration, and can be obtained from intensity fits to XRD or
neutron-diffraction patterns [14].
The lattice parameters of MnFeP
1-x
As
x
determined by means of XRD at room
temperature, and the Curie temperature T
C
are listed in Table 5.1. The lattice
parameter c remains constant in the paramagnetic (PM) state and displays a
pronounced decrease in the ferromagnetic (FM) state, while the parameter a
increases markedly with increasing x both in the PM and the FM state. The c/a
ratio decreases with increasing x.

5.4 Magnetic properties

The temperature dependence of the magnetization of the MnFeP
1-x
As
x
compounds,
measured in a field of 50 mT, is shown in Fig. 5.3. The compounds with x > 0.25
are FM and exhibit a sharp magnetic phase transition. Only the compound with x =
0.25 behaves differently, the magnetization having a much smaller value and an
anomaly in the M(T) curve around 40 K.
Figure 5.4 shows the composition dependence of the Curie temperature (left
axis) and the spontaneous magnetization (right axis) of MnFeP
1-x
As
x
at 5 K. The

MCE in MnFeP
1-x
As
x
compounds 57

0 50 100 150 200 250 300 350 400
0
5
10
15
20
25
x
0.25
0.35
0.45
0.5
0.55(I)
0.65
MnFeP
1-x
As
x

50 mT

M

(
A
m
2
/
k
g
)
T (K)

MnFeP
1-x
As
x
compounds in a field of 50 mT.

0.2 0.3 0.4 0.5 0.6 0.7
160
200
240
280
320
0
1
2
3
4
T
C

(
K
)
X

M
s

(
B
/
f
.
u
.
)

Figure 5.4: Composition dependence of the Curie temperature (left axis)
and the spontaneous magnetization (right axis) of MnFeP
1-x
As
x
compounds
at 5 K.
58 Chapter 5

Curie temperature, which is listed in Table 5.2, was determined as the temperature
where the first derivative of the magnetization with respect to temperature has an
extreme value. Upon substitution of As for P, the Curie temperature increases
linearly from 168 K for x = 0.25 to 332 K for x = 0.65, indicating that the Curie
temperature is very sensitive to the variation of the P/As ratio. The spontaneous
magnetization at 5 K is about 4
B
per formula unit, but it slightly decreases with
increasing x.

270 280 290 300 310 320 330 340 350 360
0
20
40
60
80
100
120
MnFeP
0.45
As
0.55
(II)
1 T
2 T
3 T
4 T
5 T

M

(
A
m
2
/
k
g
)
T (K)

MnFeP
0.45
As
0.55
(II), measured in constant fields of 1, 2, 3, 4, and 5 T with
temperature increasing and decreasing in steps of 1 K.

For MnFeP
0.45
As
0.55
(II), we have measured the M(T) curves in fields of 1, 2,
3, 4, and 5 T with increasing and decreasing temperature. The results are shown in
Fig. 5.5. Based on these measurements, we have constructed a magnetic phase
diagram, which is shown in Fig. 5.6. The result shows that the Curie temperature of
the sample increases linearly with applied field at a rate of dB dT
C
/ = 3.3 K/T.

MCE in MnFeP
1-x
As
x
compounds 59

304 308 312 316 320 324 328
0
1
2
3
4
5
Decreasing B
Increasing T
Decreasing T
Increasing B

MnFeP
0.45
As
0.55
(II)
PM phase
FM phase
B

(
T
)
T (K)

Figure 5.6: Magnetic phase diagram of MnFeP
0.45
As
0.55
(II). The arrows
indicate the phases in the history-dependent region.

0 1 2 3 4 5
0
20
40
60
80
100
Increasing B
Decreasing B
3
1
4

K
3
1
2

K
3
0
8

K
3
0
4

K
3
0
0

K
MnFeP
0.45
As
0.55
(I)
M

(
A
m
2
/
k
g
)
0
H (T)

Figure 5.7: Magnetic-field dependence of the magnetization of
MnFeP
0.45
As
0.55
(I), measured with increasing and decreasing field in the
vicinity of the phase transition.

60 Chapter 5

Extrapolation of the temperature dependence of the critical fields to zero field
shows that the PM to FM transition occurs at 302.8 K on cooling, and that the
inverse transition occurs at 306.6 K on heating. This indicates a thermal hysteresis
of 3.8 K. As we have seen, the transition becomes smoother in higher fields, and
the thermal hystersis decreases slightly with increasing field.
Figure 5.7 shows the magnetic isotherms of MnFeP
0.45
As
0.55
(I) in the vicinity
of its Curie temperature measured with increasing and decreasing field. The
magnetization processes show that there exists a field-induced magnetic phase
transition from the PM to the FM state. At low fields, the phase transition exhibits
a stepwise discontinuity in the magnetization, but at higher fields the transition
becomes smoother. The hysteresis is limited to a small field range of about 0.5 T
and does not extend to zero field.

5.5 Specific heat and dc susceptibility

The specific heat of MnFeP
0.45
As
0.55
(I) was measured in zero field with temperature
decreasing from 390 to 250 K, utilizing the adiabatic heat-pulse relaxation method
in the PPMS described in Chapter 3. The result is given in Fig. 5.8. The
temperature corresponding to the peak is 294 K. This is slightly smaller than the
value of 296 K, which is the Curie temperature determined from the magnetic
measurement with decreasing temperature in a field of 50 mT. The dotted line in
Fig. 5.8 represents the high-temperature limit of the molar lattice specific heat 9 R
(R is the universal gas constant).
The evolution of the entropy of MnFeP
0.45
As
0.55
(I) in the temperature range
from 250 to 390 K is depicted in the inset of Fig. 5.8. The entropy exhibits a
discontinuous change at the transition and this entropy change S associated with
the transition is about 5.2 J/molK (31.4 J/kgK) The latent heat involved in the
transition is determined as L = T
C
S = 1.53 kJ/mol (9.2 kJ/kg).

MCE in MnFeP
1-x
As
x
compounds 61

260 280 300 320 340 360 380 400
0
50
100
150
200
250
MnFeP
0. 45
As
0. 55
(I)

c
p

(
J
/
m
o
l
K
)
T ( K)

Figure 5.8: Temperature dependence of the specific heat of
MnFeP
0.45
As
0.55
(I) measured in zero field with decreasing temperature.
The dotted line indicates the value of 9R, which is the high-temperature
limit of the lattice specific heat. The inset shows the temperature
dependence of the total entropy difference S(T) - S(250) of
MnFeP
0.45
As
0.55
(I), derived from the specific-heat data.

For the MnFeP
1-x
As
x
compounds, anomalous behavior is found in the
magnetic susceptibility just above the Curie temperature. As representative
examples, the temperature dependence of the reciprocal susceptibility of the
MnFeP
1-x
As
x
compounds with x = 0.35, 0.45 and 0.55(II), measured in a field of
1T, is shown in Fig. 5.9. It is seen that, at higher temperatures, the compounds
exhibit Curie-Weiss behavior. Near the Curie temperature, the reciprocal
susceptibility abruptly drops to zero. A further analysis of the temperature
dependence of the PM susceptibility in terms of the Bean-Rodbell model, that we

250 275 300 325 350 375
0
10
20
30
40
zero field
MnFeP
0.45
As
0.55
(I)

S

(
T
)

-

S

(
2
5
0
)

(
J
/
m
o
l
K
)
T (K)
62 Chapter 5

0 50 100 150 200 250 300 350 400
0.0
0.1
0.2
0.3
0.4
0.5
0.6
x = 0.35
x = 0.55(II)
x = 0.45
MnFeP
1-x
As
x

1
/

(
T
k
g
/
A
m
2
)
T (K)

Figure 5.9: Temperature dependence of the reciprocal susceptibility of
MnFeP
0.55
As
0.45
.

have introduced in Chapter 2, will be presented in Section 5.8 in conjunction with
the analysis of the temperature dependence of the magnetization of MnFeP
1-x
As
x

compounds.

5.6 Magnetocaloric effect

In order to determine the magnetic-entropy changes in the MnFeP
1-x
As
x
system, we
have carried out measurements of the field dependence of the magnetization at
different temperatures across the Curie temperature of each sample. A
representative measurement result is presented in Fig. 5.10.
The magnetic-entropy changes
m
S have been derived from the
magnetization data on the basis of Eq. (3.3). The results for field changes from 0 to
2 and 0 to 5 T are shown in Fig. 5.11. It is found that the MnFeP
1-x
As
x
compounds
exhibit large magnetic-entropy changes, for instance, the maximum values of the

MCE in MnFeP
1-x
As
x
compounds 63

0 1 2 3 4 5
0
20
40
60
80
100
120 274--342 K T = 4 K MnFeP
0.45
As
0.55
(I)
M

(
A
m
2
/
k
g
)
0
H (T)

Figure 5.10: Magnetic isotherms of MnFeP
0.45
As
0.55
(I) in the vicinity of the
Curie temperature, measured with increasing temperature and field.

150 175 200 225 250 275 300 325 350 375
0
5
10
15
20
25
30
35

0- 2 T
0- 5 T
x=0.35
x=0.5
x=0.25
x=0.65
x=0.55(I)
x=0.45
MnFeP
1-x
As
x

-

S
m

(
J
/
k
g
K
)
T (K)

Figure 5.11: Magnetic-entropy changes of MnFeP
1-x
As
x
for field changes
from 0 to 2 and 0 to 5 T.

64 Chapter 5

magnetic-entropy changes are 20 J/kg K for x = 0.45 for a field change from 0 to 2
T and 33 J/kg K for x = 0.35 for a field change of 0 to 5 T. The large magnetic-
entropy changes of the MnFeP
1-x
As
x
compounds are concentrated in a certain
temperature interval. This interval increases with increasing field change. For
comparison, the magnetic-entropy changes of the metal Gd and the giant MCE
compound Gd
5
Ge
2
Si
2
are shown in Fig. 5.12 together with those of
MnFeP
0.45
As
0.55
(I). It is evident that the magnetic-entropy changes of
MnFeP
0.45
As
0.55
(I) are larger than of Gd and comparable with the ones of
Gd
5
Ge
2
Si
2
.

250 260 270 280 290 300 310 320 330 340 350
0
2
4
6
8
10
12
14
16
18
20
Gd
5
Ge
2
Si
2
5T
2T
5T
2T
5T
2T
MnFeP
0.45
As
0.55
(I)
Gd

-

S
m

(
J
/
k
g
K
)
T (K)

Figure 5.12: Magnetic-entropy changes of MnFeP
0.45
As
0.55
(I)

for field
changes from 0 to 2 and 0 to 5 T, derived from the magnetization data,
compared with those for Gd and Gd
5
Ge
2
Si
2
(after [1]).

In order to further examine the MCE in the MnFeP
1-x
As
x
compounds, also a direct
measurement was made by Tishins group at Moscow State University. The
temperature change of MnFeP
0.45
As
0.55
(II) was measured under adiabatic conditions
MCE in MnFeP
1-x
As
x
compounds 65

with a continuous registration of the temperature change upon fast increase of the
applied magnetic field from 0 to 1.45 T. The rate of the field change was 0.4 T/s.
The accuracy of these measurements was about 5-20 % depending on the
temperature interval. More details of these direct measurements of the MCE are
given in Ref. 18. The results are shown in Fig. 5.13. The largest value of the MCE
is about 4 K for a field change from 0 to 1.45 T. This value is about same as that
for Gd.

300 305 310 315
0
1
2
3
4
MnFeP
0.45
As
0.55
(II)

0 - 1.45 T

T
a
d

(
K
)
T (K)

Figure 5.13: Temperature dependence of the adiabatic temperature change
ad
T of MnFeP
0.45
As
0.55
(II) for a field change from 0 to 1.45 T. This
measurement was performed by Tishins group at Moscow State
University.

The field dependence of
ad
T measured at different temperatures in the
vicinity of the Curie temperature is shown in Fig. 5.14. It can be seen that
ad
T increases with increasing field. The MCE is as high as 3 K for a field change
from 0 to 1 T, and it is 2.2 K when field changes from 0 to 0.8 T at 306 K.
66 Chapter 5

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0

MnFeP
0.45
As
0.55
(II)
T = 304.7 K
T = 305.5 K
T = 306.0 K
T = 307.4 K
T
a
d

(
K
)
()

ad
T of MnFeP
0.45
As
0.55
(II) as
a function of applied-magnetic-field change, obtained by direct
measurements in the vicinity of T
C
. The measurement was performed by
Tishins group at Moscow State University.

The refrigerant capacity or cooling power is one of the most important
parameters for magnetic refrigeration. It is defined as [19]

, ) , (
2
1
=
T
T
m
dT B T S q (5.4)

where T
1
and T
2
are the temperatures of the cold reservoir and the hot sink,
respectively. It indicates how much heat can be transferred between the cold and
hot parts in a single ideal refrigeration cycle. Figure 5.15 shows the dependence of
the refrigerant capacity (left axis) of MnFeP
0.45
As
0.55
(I) on the field change in the
corresponding full-width-at-half-maximum temperature interval,
FWHM
T (right
axis).
MCE in MnFeP
1-x
As
x
compounds 67

0 1 2 3 4 5
0
50
100
150
200
250
300
350
0
5
10
15
20
25
30
35
MnFeP
0.45
As
0.55
(I)
q

(
J
/
k
g
)
B (T)

T
F
W
H
M

(
K
)

Figure 5.15: Field-change dependence of the refrigerant capacity and the
full-width-at-half-maximum temperature interval of MnFeP
0.45
As
0.55
(I).

150 175 200 225 250 275 300 325 350
0
400
500
600
700
Temperature span from T
c
-25 to T
c
+25 K
MnFeP
1-x
As
x
x = 0.35
x = 0.65
x = 0.45
x = 0.55(I)
x = 0.5
Gd
Gd
5
Ge
2
Si
2
Field Change 0 - 5 T

q

(
J
/
k
g
)
Temperature (K)
O

Figure 5.16: Refrigerant capacities of MnFeP
1-x
As
x
compounds and the
materials Gd and Gd
5
Ge
2
Si
2
, for a field change of 5 T and a temperature
span from T
C
25 to T
C
+ 25 K.
68 Chapter 5

The refrigerant capacity almost linearly increases with the field change up to
5 T. In order to compare the refrigerant capacities of our samples with those of Gd
and Gd
5
Ge
2
Si
2
, we used a field change of 5 T and a temperature range from T
C
- 25
to T
C
+ 25 K for the calculations. The results are shown in Fig. 5.16. It is clear that
the refrigerant capacity of the MnFeP
1-x
As
x
compounds is larger than that of Gd
metal, which is used as magnetic refrigerant in the prototypes of magnetic
refrigerators

5.7 Electrical resistivity and magnetoresistivity

Bearing in mind the use of MnFe(P,As) materials in magnetic refrigerators, next to
the magnetocaloric properties also the electrical and heat conductivity are of
utmost importance. There is hardly any information on the electrical-transport
properties of these materials. The electrical resistance can also be useful for a more
detailed investigation of the magnetic phase transition because it is very sensitive
to changes in the interactions between magnetic ions. The availability of electrical-
resistance data would make it possible to compare the critical magnetic fields
derived from magnetic and electrical measurements and to understand the role of
the electron-phonon and electron-magnon interactions. In order to study this, we
have selected one of the samples, MnFeP
0.55
As
0.45
, for measurements of the
electrical resistance and magnetoresistance. Figure 5.17 shows the temperature
dependence of the electrical resistivity of MnFeP
0.55
As
0.45
, normalized to its room
temperature value measured during cooling of the sample. It can be seen that there
is an anomaly at T
cr
= 231 K. Below T
cr
, the resistance increases with increasing
temperature and has metallic character, but above T
cr
it decreases dramatically in a
narrow temperature range and then recovers the metal-like temperature
dependence. The total contribution from both electron-phonon scattering and
electron-magnon scattering in the PM phase is smaller than in the FM phase which
is in contrast with normal FM metallic materials.
MCE in MnFeP
1-x
As
x
compounds 69

0 50 100 150 200 250 300
0.00
0.25
0.50
0.75
1.00
1.25
1.50
zero field
MnFeP
0.55
As
0.45

(
T
)
/
(
T

=

2
9
5

K
)
T (K)

Figure 5.17: Temperature dependence of the electrical resistivity of
MnFeP
0.55
As
0.45
, normalized to the room temperature value, measured with
decreasing temperature.

The isothermal magnetic-field dependence of the hysteresis loops of the
magnetoresistance, /
0
= (R(B,T) R(0,T))/R(0,T), of MnFeP
0.55
As
0.45
in the
temperature interval from 243 to 265 K is shown in Fig. 5.18. An increase of the
magnetic field leads to an increase of the electrical resistance, beginning at a
critical field B
cr1
and ending at B
cr2
. Hence, between 243 and 265 K the sample is in
the PM phase in zero-field, but the application of a magnetic field exceeding B
cr1

brings it into the FM regime. The field-induced PM-FM transition ends at B
cr2
.
During the decrease of the magnetic field, the reversible FM-PM transition begins
at B
cr3
and ends at B
cr4
.

70 Chapter 5

-10 -8 -6 -4 -2 0 2 4 6 8 10
0
20
40
60
80
0
20
40
60
80
0
20
40
60
80
0
20
40
60
80
0
20
40
60
80

243.4 K

0
H (T)

248.1 K

254.2 K

B
cr4
B
cr3
B
cr2
B
cr1
258.8 K

264.5 K

0
(
%
)

Figure 5.18: Isothermal magnetic-field dependence of the
magnetoresistance of MnFeP
0.55
As
0.45
in the temperature range from 243 K
to 265 K, measured with increasing and decreasing magnetic field.

MCE in MnFeP
1-x
As
x
compounds 71

240 245 250 255 260 265
0
1
2
3
4
5
6
7
8
Decreasing T
Increasing T
Decreasing B
Increasing B
MnFeP
0.55
As
0.45
FM phase
PM phase
B

(
T
)
T (K)

Figure 5.19: Magnetic phase diagram of MnFeP
0.55
As
0.45
.
The solid lines are guides to the eye.

The magnetic phase diagram of MnFeP
0.55
As
0.45
is given in Fig. 5.19. The
critical fields displayed in this figure were determined as midpoints of the
transition curves in increasing and decreasing magnetic field in Fig. 5.18. The
behavior of the electrical resistance in a magnetic field reflects the presence of
magnetic field hysteresis for the complete PM-FM transitions and indicates that
also the field-induced transition is of first order.

5.8 A model description of the first-order magnetic phase
transition

The occurrence of a first-order magnetic phase transition in MnFeP
1-x
As
x

compounds is a very striking and intriguing feature. In this section, we will present
a description of the magnetic properties of the MnFeP
1-x
As
x
compounds in terms of
the Bean-Rodbell model [20] that we have introduced in Chapter 2. First, we will
analyze the temperature dependence of the PM susceptibility of MnFeP
1-x
As
x

72 Chapter 5

compounds presented in Section 5.5. After this, we will present the results of an
analysis of the temperature dependence of the magnetization of the MnFeP
1-x
As
x

compounds.
Taking into account that the Curie temperature strongly depends on the
interatomic spacing, Blois and Rodbell have proposed the following expressions
for the PM susceptibility [21].

,
) 1 (
*
0
*
0
T T
C
T T T
C
T T
C
C

=
+
=
(5.1)

where C is the Curie constant, a is the coefficient of linear thermal expansion, and
is the slope of the dependence of T
C
on the volume

0
*
1 T
C
C

= and
0
0 *
0
1 T
T
T

= . (5.2)

The Curie constant is given by

, 3 / ) 1 (
2 2
B B
k j j g N C + = (5.3)

where N is the number of magnetic ions per formula unit,
B
is the Bohr magneton
and, g is the gyromagnetic ratio (about 2). The average total angular momentum
number j for a formula unit is estimated to be two from the saturation moment.
Figure 5.20 shows the Curie-Weiss fit of the temperature dependence of the
reciprocal PM susceptibility of MnFeP
1-x
As
x
compounds with x = 0.35, 0.45 and
0.55 by using Eq. 5.1.

MCE in MnFeP
1-x
As
x
compounds 73

0 50 100 150 200 250 300 350 400
0.0
0.1
0.2
0.3
0.4
0.5
0.6
x = 0.35
x = 0.55(II)
x = 0.45
MnFeP
1-x
As
x

1
/

(
T
k
g
/
A
m
2
)
T (K)

Figure 5.20: The Curie-Weiss fit of the temperature dependence of the
reciprocal PM susceptibility of MnFeP
1-x
As
x
compounds with x = 0.35,
0.45 and 0.55.

Table 5.2: The parameters C
*
,
*
0
T , T
0
and T
0
for MnFeP
1-x
As
x

compounds with x = 0.35, 0.45 and 0.55.

x C*(K
B
/Tf.u.)
*
0
T (K)
T
0
(K)
0
T
0.35 10.43 101 53 0.48
0.45 8.56 181 114 0.37
0.55(II) 6.15 299 263 0.12

From the fitting, we have determined the apparent Curie constant C
*
, the
apparent Curie temperature
*
0
T . The parameters T
0
and T
0
are calculated by
using Eq. 5.2 and 5.3. The results are listed in Table 5.2. We will use these
parameters in the following analysis of the temperature dependence of the
1-x
As
x
compounds.
74 Chapter 5

As a representative example of the analysis of the temperature dependence of
the magnetization, we have selected MnFeP
0.45
As
0.55
(II) for a detailed analysis.
Figure 5.21 shows the relative magnetization (which is M/
0
) of
MnFeP
0.45
As
0.55
(II) as a function of temperature for different values of the
parameter
2
(for simplification, the subscript of is neglected in the following
discussion) obtained by evaluating Eq. (2.20). The parameters that yield the best
match with the experimentally observed temperature dependence of the
0.45
As
0.55
(II), measured in a field of 1 T, are = 1.75, T
0
=
263 K, and T
0
= 0.12. In order to understand the role of the parameter , we
have calculated the (T) curves in zero field, on the basis of the obtained
parameters and for different values. This is shown in Fig. 5.22. It can be seen

200 225 250 275 300 325 350
0.0
0.2
0.4
0.6
0.8
1.0
j = 2
B = 1 T
T
0
= 263 K
T
0
= 0.12
(a) = 0
(b) = 1
(c) = 1.75
(d) = 2
(d)
(c)
(b)
(a)
Experimental data
MnFeP
0.45
As
0.55
(II)

T(K)

Figure 5.21: Relative magnetization in 1 T vs temperature for j = 2, T
0

= 263 K and different values of also shown is the experimentally
determined relative magnetization for MnFeP
0.45
As
0.55
(II).
MCE in MnFeP
1-x
As
x
compounds 75

160 180 200 220 240 260 280 300 320
0.0
0.2
0.4
0.6
0.8
1.0
T
2 T
1
0
= 0.12
T
0
= 263 K
MnFeP
0.45
As
0.55
(II)
j = 2
B = 0 T
1.75 1.5
1 0.5 = 0

T (K)

Figure 5.22: Relative magnetization vs temperature in zero field for
different values calculated for the case of MnFeP
0.45
As
0.55
(II).

that = 1 separates the first-order and the second-order transition. The curves
with < 1 correspond to a continuous change in the magnetization. In this case,
the temperature T
1
is the Curie temperature being also the paramagnetic Curie
temperature. The transition is of second order. If > 1, then a discontinuous
change (indicated by dashed vertical lines) occurs in the magnetization, showing
the transition is of first order. The temperature T
2
is the limiting temperature of the
FM state. Up to this temperature, with increasing temperature the system is found
in the FM phase.
We may illustrate some of the features of the first-order transition from the
evolution of the Gibbs-free-energy isotherms. Figure 5.23 shows the evolution of
the Gibbs free energy of MnFeP
0.45
As
0.55
(II) in the vicinity of the Curie
temperature, which is about 307 K (see Fig. 5.6), in zero field. Just above the Curie

76 Chapter 5

-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
-6
-4
-2
0
2
4
6
315 K
312
310
308
306
304
302
300

(
G
-
G
0
)
/
N
k
B
MnFeP
0.45
As
0.55
B = 0 T
j = 2
= 1.75
T
0
= 0.12
T
0
= 263 k

Figure 5.23: Reduced Gibbs-free-energy (G()-G(0))/Nk
B
isotherms vs
in the vicinity of the transition as calculated on the basis of Eq. (2.17) for j
= 2; = 1.75; T
0
= 263 K; T
0
= 0.12 and B = 0 T.

temperature, the Gibbs free energy has two shallow local minima besides the
absolute minimum at = 0, indicating the existence of a metamagnetic state.
These free-energy minima that are separated by an energy barrier determine the
metastable state of the magnetization, and are strongly dependent on temperature.
The minima disappear when the temperature increases. When the temperature
decreases from high temperature where the system is in the stable PM phase down
to T
1
, the PM phase becomes metastable. So T
1
is the limiting temperature of the
PM state when temperature decreases and, therefore, corresponds to the PM Curie
temperature. The FM state and the PM state coexist in the temperature interval T
1
<
T < T
2
. Therefore, there exists a temperature T
C
, which is the Curie temperature of
the first-order phase transition, at which the stability of the two phases is equal, in
MCE in MnFeP
1-x
As
x
compounds 77

this interval. From Fig. 5.23, we may estimate this T
C
to be around 308 K, which is
in good agreement with the experimental value of 307 K (see Fig. 5.6).

Table 5.3: Composition dependence of the parameter for
MnFeP
1-x
As
x
compounds.

x 0.35 0.45 0.55 0.65
2.43 1.90 1.75 1.47

We carried out the same fitting procedure for the MnFeP
1-x
As
x
compounds
and obtained the composition dependence of the parameter that is listed in Table
5.3. All values of are larger than one, indicating that these compounds exhibit a
first-order phase transition. This is consistent with the experimental observations.
Zach et al. [22] have carried out a similar analysis and obtained similar parameters
= 1.62 and T
0
= 250 K for MnFeP
0.5
As
0.5
. We conclude that the first-order phase
transition behavior, such as the discontinuous change of the magnetization, the
thermal hyteresis and the magnetic field-induced transition can be quite well
understood on the basis of the Bean-Rodbell model.

5.9 Discussion and conclusions

We will now present a more general discussion of the magnetic properties of the
MnFeP
1-x
As
x
compounds based on the results described above. At low
temperatures, the investigated MnFeP
1-x
As
x
compounds have a FM ground state.
With increasing temperature, the compounds undergo a phase transition from the
FM state to the PM state. In the FM state, the moment of the MnFeP
1-x
As
x

compounds is about 4
B
/f.u. at 5 K, which is in good agreement with the results of
neutron-diffraction measurements on the MnFeP
1-x
As
x
compounds with x = 0.3 and
0.5 [23]. These magnetic moments originate from the spin moments of the 3d
78 Chapter 5

electrons of the Mn and Fe atoms. Because of the itinerant character of the 3d
electrons, the magnetic moments are governed by the Fe 3d-band and Mn 3d-band,
and the main magnetic interactions in this system are direct exchange interactions.
As we see in Fig. 5.4, the Curie temperature of the MnFeP
1-x
As
x
compounds,
which is a measure of the magnetic interactions, increases linearly with increasing
As contents, indicating that the FM interactions become stronger. The reason is that
the substitution of As with larger covalent radius (1.18 ) for P (1.10 ) leads to
an expansion of the lattice in the a-b plane. This expansion probably results in a
weakening of the magnetic interactions between the Mn- moments and between the
Fe-moments, which are claimed to be antiferromagnetic [23]. Thus, variation of the
composition as well as the presence of impurities and vacancies have a strong
effect on the magnetic interactions. Different manners of sample preparation
probably result in small differences in the stoichiometry, which may be the reason
that the magnetic ordering temperatures observed for our samples are different
from the values reported by Bacmann et al. [14]. The sharp transition and the big
difference between the paramagnetic Curie temperature and the Curie temperature
indicate that the exchange interaction in this system is strongly temperature
dependent.
Based on the occurrence of a discontinuous change in the magnetization at
the transition temperature, on the thermal hysteresis in the temperature dependence
of the magnetization, and on the very sharp peak of the specific heat at the
transition temperature, we have established that the magnetic phase transition is of
first order. Moreover, as illustrated in Fig. 5.7, the transition can be induced by the
application of an external magnetic field. The stepwise change of the magnetization
at the transition and the field hysteresis observed in the field dependence of the
magnetization present evidence that also the field-induced transition is of first
order.
The first-order transition observed in the MnFeP
1-x
As
x
system has some
similarities with the transition observed in the Gd
5
(Ge
x
Si
1- x
)
4
system. In both
MCE in MnFeP
1-x
As
x
compounds 79

systems, the observed PM-FM transitions are rather sharp and occur with a certain
thermal hysteresis. The PM-FM transition can be induced by an applied field. In
the Gd
5
(Ge
x
Si
1- x
)
4
system, the transition is accompanied by a simultaneous
structural transition as discussed in Section 3 of Chapter 4, but in the MnFeP
1-x
As
x

system there is no crystal-symmetry change at the magnetic phase transition.
However, the mechanism of the first-order transition in these compounds is not
clear, yet. Bean and Rodbell [20] have proposed that a first-order phase transition
could be driven by a strong dependence of the exchange interactions on the
interatomic distances, in which the lattice distortion due to the magnetoelastic
effect plays an important role. An applied magnetic field leads to a magnetoelastic
effect that results in an enhanced exchange interaction and thus in an effectively
enhanced applied field. Band-structure calculations have shown that the
magnetoelastic phase transitions observed in this system can be associated with the
distances between the magnetic atoms as well as with changes in the density of
states (DOS) near the Fermi level, mainly due to the DOS of the Fe 3d electrons
[14,24]. According to these calculations, the strong magnetic interactions between
the Fe-layer and the Mn-layer and between the Mn-moments lead to a contraction
of the lattice parameter c and result in a first-order type magnetic phase transition.
On the basis of the Bean-Rodbell model, we have proposed a model
description of the first-order transition observed in the MnFeP
1-x
As
x
system. The
transition observed in this system is associated with a double minimum in the
Gibbs free energy as a function of magnetization. At a certain temperature, the
applied field shifts the energy minimum of the FM state to lower values than that
of the PM state above T
C
, resulting in the metamagnetic transition. The reasonably
good degree of agreement between the Bean-Rodbell model and the experimental
observations, and the equally reasonable values of the parameters obtained from
the fittings lend credence to the applicability of the model in its cardinal features to
the hexagonal MnFeP
1-x
As
x
compounds.
80 Chapter 5

Now, we discuss the MCE observed in the hexagonal MnFeP
1-x
As
x
compounds. The entropy changes associated with the first-order magnetic phase
transition have been derived from the magnetization data by using Eq. (3.3). As we
have seen in Fig. 5.11, the maximum values of the isothermal entropy change are
quite large, for instance, as high as 20 J/kgK for x = 0.45 for a field change from 0
to 2 T. As we have mentioned in Chapter 3, the magnetic-entropy change derived
from the magnetization data does not guarantee a high accuracy for determining the
MCE. But the results obtained from the magnetization data provide a reasonable
estimate of the MCE in a material and of the possible origin of the MCE. The large
magnetic-entropy changes in the MnFeP
1-x
As
x
system should be attributed to the
comparatively high 3d moments and, principally, to the rapid change of the
magnetization at the transition.

240 260 280 300 320 340 360
0
20
40
60
80
100
120
140
Gd
MnFeP
0.45
As
0.55
(I)

M

(
A
m
2
/
k
g
)
T (K)

MnFeP
0.45
As
0.55
(I) and Gd, measured with increasing temperature in a
field of 1 T.

MCE in MnFeP
1-x
As
x
compounds 81

To illustrate the reason why this system has a large MCE, we have made a
comparison of the temperature dependence of the magnetization of Gd metal and
MnFeP
0.45
As
0.55
(I). As we see in Fig. 5.24, in Gd metal the moment fully develops
only at low temperatures and strongly decreases with increasing temperature due to
RKKY interaction. The transition observed in Gd metal, which is a second-order
phase transition, gives rise to a relatively small value of
B
T M ) / ( . The strong
and direct exchange interactions between the 3d moments in transition-metal
compounds lead to perfect long-range magnetic order below the ordering
temperature and a sharper transition at T
C
. As we have mentioned, the sharp
transition in the MnFeP
1-x
As
x
compounds originates from the strong magnetoelastic
coupling, which leads to a modification of the distances between the magnetic ions,
and involves competing intra- and inter-atomic interactions.
The direct measurements of the MCE in MnFeP
0.45
As
0.55
have confirmed the
large MCE in this system. The results obtained from the direct measurements show
that the MnFeP
1-x
As
x
compounds exhibit a large MCE in low magnetic field, such
as 0.8 T. This result is very important for practical applications because lower
fields like 0.8 T are much easier to generate by permanent magnets than higher
fields like 2 T.
Finally, we discuss the temperature and field dependence of the electrical
resistance of MnFeP
0.55
As
0.45
. The results presented in Section 5.7 indicate that the
PM-FM phase transition can be induced both by temperature and magnetic field.
The former type of transition takes place from a high-resistance FM state at low
temperature to a low-resistance PM state at high temperature. The latter type of
transition leads to a positive magnetoresistance peak above T
C
. The magnetic phase
diagram based on the electrical resistance data shows that the FM-PM transition
has a field hysteresis of about 1 T. It is interesting to note that the transition at T
cr
is
accompanied by a change of the c/a ratio [23], which may lead to a change in the
Fermi surface topology and affect the electron-phonon scattering.
82 Chapter 5

In conclusion, we have prepared the hexagonal MnFeP
1-x
As
x
compounds by
means of ball milling and a solid-state reaction method. In the MnFeP
1-x
As
x
system,
the magnetic and structural properties are strongly related to a first-order magnetic
phase transition. The first-order phase transition in this compound system can
reasonably well be described by the Bean-Rodbell model. The MCE associated
with this first-order transition is large. Besides the large MCE, two additional
features make these materials excellent candidates for magnetic refrigerants in
room-temperature applications. The first is the fact that their Curie temperature can
be tuned between 168 K and 332 K by varying the P/As ratio between 1.5 and
about 0.5. This in turn allows one to tune the maximum MCE in this temperature
range, without losing the large MCE. The second is the fact that, unlike FeRh, the
giant MCE in the MnFeP
1-x
As
x
compounds is reversible.

References

[2] C.B. Zimm, A. Jastrab, A. Sternberg, V.K. Pecharsky, K.A. Gschneidner, Jr.,
M. Osborne and I. Anderson, Adv. Cryog. Eng. 43 (1998) 1759.
[3] S.A. Nikitin, G. Myalikgulev, A.M. Tishin, M.P. Annaorazov, K.A. Asatryan
and A.L. Tyurin, Phys. Lett. A 148 (1990) 363.
[4] M.P. Annaorazov, S.A. Nikitin, A.L Tyurin, K.A. Asatryan and A.K.
Dovletov, J. Appl. Phys. 79 (1996) 1689.
[5] C.H. Kuhrt, T.H. Schittny and K. Brner, Phys. Stat. Sol. A 91 (1985) 105.
[6] R.B. Flippen and F.J. Darnell, J. Appl. Phys. 34 (1963) 1094.
[8] X.X. Zhang, J. Tejada, Y. Xin, G.F. Sun, K.W. Wong and X. Bohigas, Appl.
Phys. Lett. 69 (1996) 3596.
[9] Z.B. Guo, Y.W. Du, J.S. Zhu, H. Huang, W.P. Ding and D. Feng, Phys. Rev.
Lett. 78 (1997) 1142.
[10] V.K. Pecharsky, and K.A. Gschneidner, Jr., J. Appl. Phys. 90 (2001) 4614.
[11] Young Sun, M.B. Salamon and S.H. Chun, J. Appl. Phys. 92 (2002) 3235.
[12] L. Pytlik and A. Zieba, J. Magn. Magn. Mater. 51 (1985) 199.
MCE in MnFeP
1-x
As
x
compounds 83

[13] R. Zach, M. Guillot and R. Fruchart, J. Magn. Magn. Mater. 89 (1990) 221.
[14] M. Bacmann, J.L. Soubeyrousx, R. Barrett, O. Fruchart, R. Zach, S. Niziol,
and R. Fruchart, J. Magn. Magn. Mater. 134 (1994) 59.
[15] O. Tegus, E. Brck, K.H.J. Buschow and F.R. de Boer, Nature 415 (2002)
150.
[16] O. Tegus, E. Brck, L. Zhang, Dagula, K.H.J. Buschow and F.R. de Boer,
Physica B 319 (2002) 174.
[17] E. Brck, Hybridization in cerium and uranium intermetallic compounds,
Ph.D. Thesis, University of Amsterdam, 1991.
[18] A.M. Tishin, Magnetocaloric effect in the vicinity of phase transitions, in
Handbook of Magnetic Materials, Vol. 12, pp. 398-518, Edited by K.H.J.
Buschow, Elsevier Science Publ. Amsterdam 1999.
[19] K.A. Gschneidner, Jr., V.K. Pecharsky, A.O. Pecharsky and C.B. Zimm,
Mater. Sci. Forum 315-317 (1999) 69.
[20] C.P. Bean and D.S. Rodbell, Phys. Rev. 126 (1962) 104.
[21] R.W. Blois and D.S. Rodbell, Phys. Rev. 130 (1963) 1347.
[22] R. Zach, M. Guillot and J. Tobota, J. Appl. Phys. 83 (1998) 7237.
[23] O. Beckman and L. Lundgren, Compounds of transition elements with non-
metals, in Handbook of Magnetic Materials, Vol. 6, pp 186-276, Edited by
K.H.J. Buschow, Elsevier Science Publ. Amsterdam 1991.
[24] R. Zach, M. Bacmann, D. Fruchart, P. Wolfers, R. Fruchart, M. Guillot, S.
Kaprzyk, S. Niziol and J. Tobola, J. Alloys Compds. 262-263 (1997) 508.

84 Chapter 5

85

Chapter 6

Effects of Mn/Fe ratio on
the magnetocaloric properties of
hexagonal MnFe(P,As) compounds

6.1 Introduction

The three most important factors for the realization of domestic magnetic
refrigeration are a low-cost magnetic-field source, an active magnetic refrigerant,
and a proper design of the thermodynamic refrigeration cycle. The best choice of
the field source for domestic applications would be a permanent magnet. However,
the field generated by permanent magnets is typically below 1 T, but it is also
possible to generate a field of 2 T with high-energy-product magnets [1,2].
Usually, the magnetocaloric effect (MCE) of a magnetic material is very small for
such a low-field change. This calls for new materials that possess a large MCE.
Recently developed new materials exhibit a very large MCE [3-5]. For example, as
we have reported in Chapter 5, the compound MnFeP
0.45
As
0.55
exhibits an adiabatic
temperature change of about 3 K at a field change of 1 T around 30C. These
achievements are very promising for developing domestic magnetic refrigerators as
an alternative for the conventional gas compression/expansion refrigerators in use
today. The large MCE observed in the new materials reported in Refs. [3]-[5] is
related to a first-order phase transition and is strongly material dependent.
Therefore, a better understanding of the magnetocaloric properties of these
86 Chapter 6

materials is essential for developing new materials. Recent developments in new
designs of prototypes of magnetic refrigerators [6,7] have brought the magnetic-
refrigeration technology a step closer toward room-temperature applications. Due
to an enhanced lattice entropy above 20 K in relevant materials, the Ericsson cycle
is most suited for the high-temperature applications.
As reported in Chapter 5, the new materials MnFeP
1-x
As
x
that consist of Mn,
Fe, P, and As with a ratio of Mn:Fe:(P+As) = 1:1:1, have advantages over existing
magnetic coolants. It exhibits a large MCE, which is larger than that of Gd metal
and its operating temperature can be easily tuned from 168 to about 332 K by
adjusting the P/As ratio between 1.5 and 0.5 without losing the large MCE [8]. The
main effect of varying the composition of the non-magnetic P and As is a variation
of the lattice parameters and a change of the magnetic-ordering temperature. In this
case, as may be expected, the size of the magnetic moments is hardly affected.
Only beyond 65 at. % P, the moment is reduced. The MCE increases with
decreasing magnetic-ordering temperature. The large entropy change is associated
with a field-induced first-order phase transition. The magnetoelastic effect plays an
important role in the phase transition, and may lead to an enhancement of the
exchange interactions and thus a change of the molecular field. This may enhance
the effective magnetic field.
We have studied the effects of substitution of some elements on the MCE of
MnFe(P,As)-based compounds [9-12]. We have found that the variation of the
composition of the magnetic elements Mn and Fe directly changes the magnetic
moments and simultaneously induces a change in the magnetic interactions. We
may expect an enhanced MCE with increasing Mn content since Mn has a larger
magnetic moment than Fe. We have also been aware of the fact that thermal
hysteresis or field hysteresis is one of the main obstacles for applications. Although
this hysteresis is intrinsic to the first-order phase transition, it may be reduced by
precisely adjusting the magnetic interactions. In this chapter, we report on our
Effects of Mn/Fe ratio on the magnetocaloric properties of MnFe(P,As) 87

studies that were motivated to improve and to optimize the MCE observed in the
MnFe(P,As)-based compounds by varying the Mn/Fe ratio.

6.2 Experimental

Polycrystalline samples of Mn
2-x
Fe
x
P
0.5
As
0.5
with x = 1.2, 1.1, 1.0, 0.9, 0.8, 0.7, 0.6,
0.5, 0.4, and Mn
1.1
Fe
0.9
P
0.47
As
0.53
were prepared by ball milling, followed by a
solid-state reaction as reported in Chapter 5. The starting materials are the binary
compounds Fe
2
P (99.5 % pure, Alfa Aesar), FeAs
2
(99.5 % pure, Alfa Aesar), Mn
chips (99.99 % pure, Alfa Aesar), and red-P powder (99.5 % pure). X-ray
diffraction (XRD) was used for the characterization of the phases and for the
determination of the unit-cell parameters.
The temperature and field dependence of the magnetization of the samples
were measured with a Quantum Design SQUID magnetometer in the temperature
range from 5 to 400 K and in magnetic fields from 0 to 5 T. The resistance
measurements were carried out in an Oxford Instruments MagLab system using a
four-point method. The sample used in the measurements had dimensions 2 x 2 x
10 mm
3
. The measurements were performed with 10 mA ac current with a
frequency of 63 Hz, and the direction of the ac current was perpendicular to the
field direction.
Direct measurements of the MCE were carried out at Moscow State
University. The rate of the field change was up to 0.4 T/s. The accuracy of the
measurements was about 5-20 %.

6.3 Results and discussion

6.3.1 Structural and magnetic properties

Figure 6.1 shows the XRD patterns of the Mn
2-x
Fe
x
P
0.5
As
0.5
compounds. With
increasing Mn content, we observe a small shift of the peak positions to lower

88 Chapter 6

20 30 40 50 60 70 80
0
3
2
1
2
2
1
3
1
1
2
1
0
3
0
1
2
0
0
2
1
1
1
1
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
x = 1.2

I
n
t
e
n
s
i
t
y

(
a
r
b
.

u
n
i
t
)
2 (deg)

Figure 6.1: XRD patterns of Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x = 1.2,
1.1, 1.0, 0.9, 0.8, 0.7, 0.6, 0.5 and 0.4, measured at room temperature.
The indices of the main reflections of the Fe
2
P type are also given.

angle. This indicates a small increase of the unit-cell volume as may be expected
from the slightly larger atomic volume of Mn compared to that of Fe. Refinement
of the structure results in the lattice parameters as summarized in Table 6.1, where,
we see that with increasing Mn content, mainly the lattice parameter a increases

while the parameter c remains unchanged. This effect is quite similar to that caused
by variation of the P (and As) content.

Table 6.1: Lattice parameters of Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x =
1.2, 1.1, 1.0, 0.9, 0.8, 0.7, 0.6, 0.5, and 0.4 at room temperature obtained
through refinement of XRD data.

Nominal, x a() c() c/a v(
3
)
1.2 6.1126 3.4770 0.569 112.5
1.1 6.1224 3.4743 0.567 112.8
1.0 6.1243 3.4765 0.568 112.9
0.9 6.1344 3.4744 0.566 113.2
0.8 6.1470 3.4980 0.569 114.5
0.7 6.1372 3.4820 0.567 113.6
0.6 6.1507 3.4810 0.566 114.0
0.5 6.1610 3.4970 0.568 115.0
0.4 6.1621 3.4982 0.568 115.0

Figure 6.2 shows the magnetic-field dependence of the magnetization of the
Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x = 1.2, 1.1, 1.0, 0.9, 0.8, 0.7, 0.6, 0.5 and 0.4 at
5 K. The Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x = 1.2, 1.1, 1.0, 0.9, 0.8 and 0.7 are
ferromagnetic (FM). The observed magnetic moment varies between 3.8 and 4.2
B
/f.u.. The largest moments observed are 4.2
B
/f.u. for Mn
1.2
Fe
0.8
P
0.5
As
0.5
and 4.1
B
/f.u. for Mn
1.1
Fe
0.9
P
0.5
As
0.5
, probably due to the higher moment of Mn compared
to that of Fe. In the case of excess of Fe, we would expect a reduction of the
magnetic moment, but the moment remains almost unchanged. This indicates that
Fe has a higher moment at the 3g sites than at the 3f sites. These results are in
agreement with neutron-diffraction results [13] and band-structure calculations
[14]. Because both Mn
2
P and Mn
2
As are antiferromagnetically ordered [15], we
expect that for the Mn-rich compounds beyond some amount of Mn substitution
the FM ground state will be destroyed. This is seemingly the case for the
compounds in which 40 - 60 % of the Fe is replaced by Mn, which are clearly not
FM.
90 Chapter 6

0 1 2 3 4 5 0 1 2 3 4 5
0
2
4
0
2
4
0 1 2 3 4 5
0
2
4

x = 0.9

x = 0.8

x = 0.7

x = 1.1
0
H (T)

x = 1.2
0
H (T)

0
H(T)

x = 1.0
x = 0.4

M

(
B
/
f
.
u
.
)
x = 0.5

M

(
B
/
f
.
u
.
)
x = 0.6

M

(
B
/
f
.
u
.
)

Figure 6.2: Field dependence of the magnetization of Mn
2-x
Fe
x
P
0.5
As
0.5

compounds with x = 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1 and 1.2, measured at
5 K.

In Fig. 6.3, we have plotted the temperature dependence of the magnetization
of Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x = 0.4, 0.5, 0.6 and 0.7, 0.8, 0.9, 1.0, 1.1 and
1.2, measured in an applied field of 50 mT. In the Mn
2-x
Fe
x
P
0.5
As
0.5
compounds
with x =0.4, 0.5 and 0.6, obviously no FM order occurs. Instead, we observe
complex magnetic behavior below the magnetic-ordering temperature, and
additionally we observe some differences between the field-cooled and zero-field-
cooled measurements, which may indicate a complicated spin structure in these
compounds at low temperatures. As a magnetocaloric material, these materials are
of less interest. The Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x = 0.7, 0.8, 0.9, 1.0, 1.1
and 1.2 are FM. The sample with x = 0.9 has exactly the same critical temperature


0 50 100 150 200 250 300 350 400
0
10
20
30
x = 0.7
x = 0.8
x = 0.9
x = 1.0
x = 1.1
x = 1.2
Mn
2-x
Fe
x
P
0.5
As
0.5
B = 50 m T

M

(
A
m
2
/
k
g
)
T (K)
0.0
0.1
0.2
0.3 x = 0.4
x = 0.5
x = 0.6

Figure 6.3: Temperature dependence of the magnetization of the
Mn
2-x
Fe
x
P
0.5
As
0.5
compounds with x = 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1
and 1.2, measured with increasing temperature in a field of 50 mT. For x =
0.4, 0.5 and 0.6, also the measurements with decreasing temperature are
included.

T
C
= 282 K as MnFeP
0.5
As
0.5
. Further increase of the Mn content leads to the
expected reduction of the Curie temperature as can be seen for the samples with x
= 0.8 and 0.7. In these samples, the Curie temperature is strongly reduced to 240
and 203 K, respectively. The samples with x = 1.1 and 1.2 exhibit an increase of
the Curie temperature to 319 and 322 K, respectively. No thermal hysteresis is
observed in the samples with x = 1.1 and 1.2. These results suggest that the
magnetic interactions and the magnetic phase transition do not only depend on the

92 Chapter 6

250 260 270 280 290 300 310 320 330 340 350
0
20
40
60
80
100
120
Mn
1.1
Fe
0.9
P
0.47
As
0.53
1 T
2 T
3 T
4 T
5 T

M

(
A
m
2
/
k
g
)
T (K)

Mn
1.1
Fe
0.9
P
0.47
As
0.53
, measured with increasing temperature in fields of 5, 4,
3, 2, and 1 T. The inset shows the field dependence of the Curie
temperature.

distances between the magnetic atoms but that they are also related to the
electronic structure of the magnetic atoms, presumably to the density of the 3d-
electron states near the Fermi level.
In order to adjust the ordering temperature of the material around room
temperature (assuming 293 K), we have prepared an additional sample with
composition Mn
1.1
Fe
0.9
P
0.47
As
0.53
. Figure 6.4 displays the temperature dependence
of the magnetization of Mn
1.1
Fe
0.9
P
0.47
As
0.53
, measured in fields of 5, 4, 3, 2, and 1
T with increasing temperature from 250 K to 350 K and, after this, decreasing the
temperature to 250 K. The Curie temperature increases almost linearly with
increasing field at a rate of 4.2 K/T. This rate is much larger than for

0 1 2 3 4 5
285
290
295
300
305
310
dT/dB = 4.2 K

T

(
K
)
B (T)

0 1 2 3 4 5
0
20
40
60
80
100
120
3
0
4
K
3
0
0

K
2
9
6
K
2
9
2
K
2
8
8

K
2
8
4

k
2
8
0

K
Mn
1.1
Fe
0.9
P
0.47
As
0.53

M

(
A
m
2
/
k
g
)
B
H (T)

Figure 6.5: Magnetic isotherms of Mn
1.1
Fe
0.9
P
0.47
As
0.53
in the vicinity of
the Curie temperature.

MnFeP
0.45
As
0.55
(3.3 K/T), and almost same as for Gd
5
Si
1.7
Ge
2.3
(4.4 K/T). The
extrapolated zero-field Curie temperature is about 286 K. The Curie temperature
has been determined as the temperature corresponding to the minimum point of the
first derivative of the M(T) curve.
Figure 6.5 shows the magnetic isotherms of Mn
1.1
Fe
0.9
P
0.47
As
0.53
, measured
with increasing field and subsequent decreasing field with steps of 50 mT. The
results show that, at temperatures above T
C
, a first-order transition from the PM
state to the FM state can be induced by application of a magnetic field. The critical
field needed to induce the transition increases with increasing temperature, whereas
the field hysteresis decreases with increasing temperature. The transition is much
smoother than the one observed in MnFeP
0.45
As
0.55
. There is no remanence when
the field decreases to zero, indicating that the transition can be reproduced upon
cycling through zero field.

94 Chapter 6

6.3.2 Magnetocaloric properties

The isothermal
m
S of the samples has been derived from the magnetic isotherms
by using Eq. (3.3). As a representative example, Fig. 6.6 shows the magnetic
isotherms of Mn
1.1
Fe
0.9
P
0.5
As
0.5
in the vicinity of the Curie temperature (T
C
= 282
K). The magnetic isotherms display three different types of behavior: below 282
K, they are FM, above 298 K, they are paramagnetic (PM), and between these two
temperatures, there exists a field-induced magnetic phase transition from the PM
state to the FM state. The critical field of the phase transition increases with
temperature at a rate about 0.23 T/K. From the magnetic isotherms of each of the
samples, we have derived the isothermal magnetic-entropy changes in the samples.
To avoid unnecessary errors, all magnetization measurements have been
performed in the same manner with an increasing temperature step of 4 K and

0 1 2 3 4 5
0
20
40
60
80
100
120
T

=

4

K
310 K
270 K
Mn
1.1
Fe
0.9
P
0.5
As
0.5

M

(
A
m
2
/
k
g
)
0
H (T)

1.1
Fe
0.9
P
0.5
As
0.5
, measured with
increasing field and increasing temperature between 270 and 310 K with
temperature steps of 4 K.

180 200 220 240 260 280 300 320 340
0
5
10
15
20
25
30
Mn
2-x
Fe
x
P
0.5
As
0.5
x = 1.2 x = 1.1
x = 1.0
x = 0.9
x = 0.8
x = 0.7
B
0 - 1 T
0 - 2 T

-

S
m

(
J
/
k
g
K
)
T (K)

Figure 6.7: Isothermal magnetic-entropy change in Mn
2-x
Fe
x
P
0.5
As
0.5

compounds as a function of temperature for field changes from 0 to 1 and 0
to 2 T.

increasing field steps of 0.1 T up to 1 T and steps of 0.2 T between 1 and 5 T. The
results are shown in Fig. 6.7. The maximum magnetic-entropy changes are
observed in the compound Mn
1.1
Fe
0.9
P
0.5
As
0.5
, being about 17 and 25 J/kgK for
field changes from 0 to 1 and 2 T, respectively. The Fe-richer samples with x = 1.1
and x = 1.2 exhibit an increase of T
C
and a simultaneous reduction of the
magnetic-entropy change. Moreover, the peak of the magnetic-entropy change
becomes much broader than for other compositions. For the other compositions,
both the T
C
and the magnetic-entropy change decrease with increasing Mn
contents.
The magnetic isotherms of Mn
1.1
Fe
0.9
P
0.47
As
0.53
were measured in the vicinity
of the ordering temperature with the magnetic field increasing from 0 to 3 T. The
result is shown in Fig. 6.8. The isothermal magnetic-entropy change of

96 Chapter 6

0.0 0.5 1.0 1.5 2.0 2.5 3.0
0
20
40
60
80
100
120
272 -- 312 K; T = 4 K
Mn
1.1
Fe
0.9
P
0.47
As
0.53

M

(
A
m
2
/
k
g
)
0
H (T)

1.1
Fe
0.9
P
0.47
As
0.53
, measured with
increasing field in the temperature range from 272 to 312 K with
temperature steps of 4 K and field steps of 0.1 T.

270 280 290 300 310 320 330
0
5
10
15
20
25
0 - 1 T
0 - 2 T
0 - 3 T

Mn
1.1
Fe
0.9
P
0.47
As
0.53

-
S
m

(
J
/
k
g
K
)
T (K)

Figure 6.9: Temperature dependence of the isothermal magnetic-entropy
change in Mn
1.1
Fe
0.9
P
0.47
As
0.53
for field changes from 0 to 1, 0 to 2, and 0 to
3 T.

Mn
1.1
Fe
0.9
P
0.47
As
0.53
was derived from the magnetization data by using Eq. (3.3). In
this case, we used temperature steps of 4 K and field steps of 0.1 T.
The magnetic-entropy changes of Mn
1.1
Fe
0.9
P
0.47
As
0.53
resulting from field
changes from 0 to 1, 0 to 2, and 0 to 3 T are shown in Fig. 6.9. The maximum
values of the magnetic-entropy changes for field changes of 1, 2, and 3 T are found
to be about 12, 21, and 23 J/kgK, respectively. The corresponding values for
FWHM
T are 5, 6, and 9 K, respectively. The refrigerant capacity, in the temperature
range from 291 to 297 K for a field change of 2 T, is about 94 J/kg. It should be
mentioned, however, that the refrigerant capacity, which is essentially the integral
under the curves in Fig. 6.9, is not much altered. For magnetic-refrigeration cycles
in low magnetic fields, it may be of importance to absorb a large amount of heat
over a narrow temperature interval.

286 288 290 292 294 296 298
0
1
2
3
4
B
0 - 1.45 T
Mn
1.1
Fe
0.9
P
0.47
As
0.53

T
a
d

(
K
)
T (K)

ad
T of Mn
1.1
Fe
0.9
P
0.47
As
0.53
as
a function of temperature, determined by direct measurement for a field
change from 0 to 1.45 T. This measurement was performed by Tishins
group at Moscow State University.

98 Chapter 6

The isothermal magnetic-entropy change is a fundamental parameter but not the
only one for a magnetic refrigerant. The adiabatic temperature change appears to
be more important for cooling applications. In order to assess the suitability of the
suggested compounds, we have collaborated with Tishins group at Moscow State
University and have performed direct measurements of the adiabatic temperature
change of Mn
1.1
Fe
0.9
P
0.47
As
0.53
. Figure 6.10 shows the results for a field change
from 0 to 1.45 T. The largest achieved value of
ad
T due to this field change is 4.2
K.

0.4 0.6 0.8 1.0 1.2 1.4 1.6
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
291.1 K
292.2 K
293.4 K
295.4 K
Mn
1.1
Fe
0.9
P
0.47
As
0.53

T
a
d

(
K
)
B (T)

ad
T of Mn
1.1
Fe
0.9
P
0.47
As
0.53
in
the vicinity of T
C
as a function of applied magnetic field obtained by means
of direct measurements. This measurement was performed by Tishins
group at Moscow State University.

Figure 6.11 shows the field dependence of
ad
T of Mn
1.1
Fe
0.9
P
0.47
As
0.53
at
four different temperatures in the transition region. As we see, the MCE increases
with increasing field, and is not saturated up to the maximal applied field of 1.45 T.
This means that the values of the adiabatic temperature change will further increase
with increasing magnetic field. Around 292 K, the adiabatic temperature change

reaches 3 K for a field change from 0 to 1 T. This is larger than the value of 2.07
K/T for Gd and the same as the value of 3 K/T reported for Gd
5
Si
2
Ge
2
[16].
In order to compare the magnetocaloric properties of the present compounds
with those of other systems, we present relevant data for these materials in Table
6.2. It can be seen that the largest magnetic-entropy change is observed in MnAs,
which is about 30 J/kgK for a field change from 0 to 2 T. But the corresponding
ad
T is only 4.7 K. The largest MCE, as high as about 13 K for a field change of 2
T, has been reported for the FeRh system. This is almost three times larger than

Table 6.2: Magnetic-ordering temperature T
C
, isothermal magnetic-entropy
change S
m
, and adiabatic temperature change
ad
T of MnFe(P,As)-
based compounds, compared with other materials.

Material

T
C
(K) - S
m
(J/kgK)
0-2 T
T
ad
(K)

Ref.

MnFeP
1-x
As
x
x = 0.35
0.45
0.50
0.55

213
240
282
300(I)

13
20
16.5
15(I)

3.9 (0-1.45T)*
present
work

Mn
2-x
Fe
x
P
0.5
As
0.5
x = 0.90
0.80
0.70

282
236
205

25.5
16.1
10.8

present
work
Mn
1.1
Fe
0.9
P
0.47
As
0.53
290 21 4.2 (0-1.45T)* present
work
Gd 294 5 5.7 (0-2 T) [17]
Fe
49
Rh
51
316 22 12.9 (0-2 T) [18]
Fe
49
Rh
51
313 12 8.4 (0-2 T) [19]
MnAs 318 31 4.7 (0-2 T) [20]
Gd
5
Si
2
Ge
2
278 14 7.3 (0-2 T) [3]
Gd
5
Si
1.97
Ge
2.03
262 - 2-3 (0-1.4 T)* [21]
La(Fe
0.9
Si
0.1
)
13
La(Fe
0.88
Si
0.12
)
13
H
0.5
La(Fe
0.89
Si
0.11
)
13
H
1.3

188
233
291
25
20
24
4 (0-1.4 T)*
6 (0-2 T)
6.9 (0-2 T)
[22]
[5]
[5]

100 Chapter 6

that of Gd metal in the vicinity of the magnetic phase transition. Some direct
measurements (denoted by * in Table 6.2) performed at Moscow State University
show that the MnFe(P,As)-based compounds have larger a MCE than Gd metal in
the same temperature range.
As has been pointed out in Ref. 23, the adiabatic temperature change mainly
depends on the change of the magnetic-ordering temperature upon application of a
magnetic field. We have observed dT
C
/dB of 3.3 K/T for hexagonal MnFeP
1-x
As
x

and 4.2 K/T for Mn
1.1
Fe
0.9
P
0.47
As
0.53
.
A first-order transition is always associated with thermal or field
hysteresis. In the hysteresis region, the magnetic state of the material is
uncertain, depending on the history. Hysteresis will result in a reduced
efficiency of the refrigeration cycle, which will be especially important if
one works in low magnetic fields. For the samples with equal amounts of Fe

280 285 290 295 300 305 310
0
20
40
60
80
100
Mn
1.1
Fe
0.9
P
0.47
As
0. 53

B = 1 T
M

(
A
m
2
/
k
g
)
T (K)

Mn
1.1
Fe
0.9
P
0.47
As
0.53
, measured in a field of 1 T with increasing and, after
this, decreasing temperature.

and Mn, the thermal hysteresis is between 3 and 4 K, and the field hysteresis
about 0.5 T. In Fig. 6.12, we show the temperature dependence of the
magnetization of a Mn
1.1
Fe
0.9
P
0.47
As
0.53
sample, measured in a constant
magnetic field of 1 T with increasing and, after this, decreasing temperature.
The observed thermal hysteresis is less than 2 K and, as we see in Fig. 6.5, the field
hysteresis is only about 0.2 T. This reduction of the thermal hysteresis makes it
feasible to employ refrigeration cycles even in fields below 2 T. The lower the
field, however, the more stages of magnetocaloric cycles one may have to employ
to achieve the desired temperature change.
The possibility to reduce the thermal hysteresis of the first-order phase
transition is the most important result of this study. The driving force for the first-
order character of the magnetic transition is clearly a magnetoelastic interaction
between the Mn- and Fe-containing planes in the hexagonal MnFe(P,As)
compounds. In contrast to compounds like Gd
5
Si
2
Ge
2
, MnAs or FeRh, which also
show a first-order magnetoelastic transition, the transition in MnFe(P,As) is not
associated with a change in crystal symmetry but merely with a change in c/a ratio.
This difference forms the basis of the rather low hysteresis observed. If the crystal
symmetry is altered in the magnetic phase transition, the magnetic state is locked
in, resulting in the generally observed large hysteresis. In the case of a change in
c/a ratio, only the exchange interaction between the sublattices varies. Addition of
Mn on the Fe sublattice may introduce some competing interactions. On
approaching the critical distance between the sublattices, already small thermal
fluctuations lead to a release of the locking. This idea is also corroborated by the
fact that substitution of 10 % of the Fe by Mn, which does not alter the lattice
parameter c and thus the distance between the sublattices, does not affect T
C
.

102 Chapter 6

6.3.3 Electrical resistivity

Figure 6.13 shows the temperature dependence of the electrical resistivity of
Mn
1.1
Fe
0.9
P
0.47
As
0.53
measured in zero field with decreasing temperature in the
temperature interval from 5 to 310 K. With decreasing temperature, an anomaly
occurs in the electrical resistivity around 290 K. This temperature corresponds to
the magnetic phase transition. Above this temperature, the electrical resistance of
the sample decreases rapidly in a narrow temperature range and then becomes less
temperature dependent. A similar behavior of the electrical resistivity has been
reported for the isostructural compound MnRhP [24]. At 300 K, the electrical
resistivity of Mn
1.1
Fe
0.9
P
0.47
As
0.53
is about 1800 cm. This value is between the
resistivity, which is about 400 cm, of Fe
2
P [25,26] and the resistivity, which is
about 3000 cm of Mn
2
P [27] at room temperature, and also of the same order as
that of Gd
5
Si
2
Ge
2
, which is about 1100-2800 cm [28].

0 50 100 150 200 250 300
0
500
1000
1500
2000

290 K
Zero Field
Mn
1.1
Fe
0.9
P
0.47
As
0.53
c
m
)
T(K)

Figure 6.13: Temperature dependence of the electrical resistivity of
Mn
1.1
Fe
0.9
P
0.47
As
0.53
, measured with decreasing temperature in zero field.


The magnetic-field dependence of the isothermal electrical resistance of
Mn
1.1
Fe
0.9
P
0.47
As
0.53
, measured at four different temperatures, is shown in Fig. 6.14.
At 275 K, which is in the FM region, the electrical resistance of the sample
decreases with increasing magnetic field, indicating that the magnetic contribution
to the resistance is suppressed by the applied field. At temperatures of 295, 303,
and 310 K, which are in the PM region, we observe a field-induced phase transition

0 1 2 3 4 5
0.044
0.046
0.048
0.050
0.052
310 K
303 K
295 K
275 K

Mn
1.1
Fe
0.9
P
0.47
As
0.53
R

(
)
B(T)

Figure 6.14: Field dependence of the isothermal resistance of
Mn
1.1
Fe
0.9
P
0.5
As
0.5
, measured with increasing and, after this, decreasing
field.

with a field hysteresis of less than 0.3 T in the electrical-resistance curves. The
resistance has a step of about 8 % at the transition. These features also indicate that
the transition is of first order. The critical field increases with a rate of about 0.22
T/K for increasing temperature.

104 Chapter 6

6.4 Conclusions

Variation of the Mn/Fe ratio has the following effects on the magnetic and the
magnetocaloric properties of the hexagonal Mn
2-x
Fe
x
P
0.5
As
0.5
compounds:
1. The magnetic moment increases by 2 % and 5 % when going from
MnFeP
0.5
As
0.5
to Mn
1.1
Fe
0.9
P
0.5
As
0.5
and Mn
1.2
Fe
0.8
P
0.5
As
0.5
, respectively.
The maximum magnetic moment found in Mn
1.2
Fe
0.8
P
0.5
As
0.5
is about 4.2
B
/f.u..
2. The Curie temperature, which is also the working point of a magnetic
refrigerant, increases from 203 K to 330 K in the composition range 0.7
x 1.2.
3. The magnetic-entropy change is maximal in the compound
Mn
1.1
Fe
0.9
P
0.47
As
0.53
. The maximum values are 21 and 23 J/kgK for field
changes of 0 to 1 and 0 to 2 T, respectively. A direct measurement of the
MCE shows that the adiabatic temperature change is 4.2 K for a field
increase from 0 to 1.45 T, and as high as 3 K for a field increase from 0 to
1 T around 292 K.
4. The thermal hysteresis is smaller than 2 K and the field hysteresis is about
0.3 T in the Mn
2- x
Fe
x
P
0.5
As
0.5
system, both values being smaller than those
observed in the MnFeP
1-x
As
x
system.
5. The electrical resistivity of the compounds is about 1800 cm around
room temperature. The steplike change and the field hysteresis in the
electrical resistance of the compound Mn
1.1
Fe
0.9
P
0.47
As
0.53
also confirm that
the transition is of first order.
Summarising, we have observed an enhanced MCE and a reduced thermal
hysteresis in the compounds with 10 % excess of Mn: Mn
1.1
Fe
0.9
P
0.47
As
0.53
and
Mn
1.1
Fe
0.9
P
0.5
As
0.5
. These excellent features emphasize the large potential of the
present compounds for room-temperature magnetic-refrigeration applications.


References

[2] S.J. Lee, J.M. Kenkel, V.K. Pecharsky and D.C. Jiles, J. Appl. Phys. 91
(2002) 8543.
[3] V.K. Pecharsky and K.A. Gechneidner, Jr., Phys. Rev. Lett, 78 (1997) 4494.
[4] O.Tegus, E. Brck, K.H.J. Buschow and F.R. de Boer, Nature 415 (2002)
150.
[5] A. Fujita, S. Fujieda, Y. Hasegawa and K. Fukamichi, Phys. Rev. B 67
(2003) 104416.
[6] http://www.external.ameslab.gov/news/relase/01magneticrefrig.htm.
[7] http://www.jarn.co.jp/News/2003_Q2/30520_Chubu_Magnet_cool.htm.
[8] O.Tegus, E. Brck, L. Zhang, W. Dagula, K.H.J. Buschow and F.R. de Boer,
Physica B 319 (2002) 174.
[9] E. Brck, O. Tegus, X.W. Li, F.R. de Boer and K.H. J. Buschow, Physica B
327 (2003) 431.
[10] O. Tegus, L. Zhang, W. Dagula, E. Brck, K.H.J. Buschow and F.R. de
Boer, J. Appl. Phys. 93 (2003) 7655.
[11] X.W. Li, O. Tegus, L. Zhang, W. Dagula, E. Brck, K.H.J. Buschow and
F.R. de Boer, submitted to IEEE Trans. Magn..
[12] O. Tegus, E. Brck, X.W. Li, L. Zhang, W. Dagula, F. R. de Boer and K.H.J.
Buschow, submitted to J. Magn. Magn. Mater..
[13] M. Bacmann, J.L. Soubeyroux, R. Barrett, R. Zach, S. Niziol and R.
Fruchart, J. Magn. Magn. Mater. 134 (1994) 59.
[14] J. Tobola, M. Bacmann, D. Fruchart, S. Kaprzyk, A.Koumina, S. Niziol, J.L.
Soubeyroux, P. Wolfers and R. Zach, J. Magn. Magn. Mater. 157-158 (1996)
708.
[15] O. Beckman and L. Lundgren, in Handbook of Magnetic Materials, Edited
by K.H.J. Buschow (North Holland, Amsterdam, 1991), Vol. 6, pp. 181-287.
[16] K.A. Gschneidner, Jr. and V.K. Pecharsky, Ann. Rev. Mater. Sci. 30 (2000)
387.
[17] S.Yu. Dankov, A.M. Tishin, V.K. Pecharsky and K.A. Gschneidner, Jr.,
Phys. Rev. B 57 (1998) 3478.
[18] M.P. Annaorazov, K.A. Asatryan, G. Myalikgulyev, S.A. Nikitin, A.M.
Tishin and A.L. Tyurin, Cryogenics 32 (1992) 867.
106 Chapter 6

[19] M.P. Annaorazov, S.A. Nikitin, A.L. Tyurin, K.A. Asatryan and A.Kh.
Dovletov, J. Appl. Phys. 79 (1996) 1689.
[21] A.S. Chernyshov, D.A. Filippov, M.I. Ilyn, R.Z. Levitin, A.O. Pecharsky,
V.K. Pecharsky, K.A. Gschneidner, Jr., V.V. Snegirev and A.M. Tishin,
Phys. Met. Metall. 93 (2002) S19.
[22] F.X. Hu, Max Ilyn, A.M. Tishin, J.R. Sun, G.J. Wang, Y.F. Chen, F. Wang,
Z.H. Cheng and B.G. Shen, J. Appl. Phys. 93 (2003) 5503.
[23] V.K. Pecharsky, K.A. Gschneidner, Jr., A.O. Pecharsky and A.M. Tishin,
Phys. Rev. B 64 (2001) 144406.
[24] I. Kanomata, T. Narita, N. Suzuki, K. Sato, T. Harada, N. Ogawara, K.
Koyama and M. Motokawa, J. Alloys Compds. 334 (2002) 68.
[25] S. Chiba, J. Phys. Soc. Japan, 15 (1960) 581.
[26] D. Bellavance, J. Mikkelson and A. Wold, J. Solid State Chem. 2 (1970)
285.
[27] I.G. Fakidov and V.P. Krasovskii, Phys. Met. Metall. 7 (1959) 159.
(1999-I) 7993.

107

Summary

This thesis presents a study of the magnetocaloric effect (MCE) and related
physical properties of several intermetallic compounds: GdRu
2
Ge
2
, Gd
5
Ge
2.3
Si
1.7
,
and MnFe(P,As). Of particularly interest is the discovery of a very large MCE in
MnFe(P,As) compounds around room temperature and low fields that can be
generated by permanent magnets. This is a promising step in the direction of
refrigeration technology by means of the magnetocaloric effect.
After a brief overview on the conventional refrigeration techniques, the
significance of developing magnetic refrigeration and the motivation for this study
are pointed out in Chapter 1. The adiabatic temperature change and the isothermal
magnetic-entropy change are the two characteristic parameters for evaluating the
magnetocaloric properties of a magnetic material. The theoretical aspects of the
MCE and of magnetic phase transition are presented in Chapter 2. Since a large
MCE may be expected in the vicinity of a magnetic phase transition, our study is
focused on the MCE associated with such transitions, in particular first-order phase
transitions
For the determination of the MCE, we have used both magnetic and
specific-heat measurements, that are described in Chapter 3. Additionally, we also
present direct measurements, which were performed by Tishins group at Moscow
State University, of the MCE in some of the MnFe(P,As) compounds. A reliable
characterization technique for the MCE and a better understanding of the MCE and
related physical properties of the newly found compounds Gd
5
(Ge,Si)
4
are,
therefore, crucial in fundamental as well as technological investigations for new
magnetic refrigerants. Therefore, in Chapter 4, the characterization techniques and
methods to determine the MCE were tested by studying the MCE in GdRu
2
Ge
2
.
108 Summary

The results of this study show that the isothermal magnetic- entropy change
derived from magnetization measurements is in good agreement with the values
obtained by means of specific-heat measurements in magnetic field, thus
confirming that it is reliable to investigate the MCE in new materials by means of
simple magnetization measurements. In Chapter 4, we also present the results of a
study of the MCE and of related physical properties of a single crystal of
Gd
5
Ge
2.3
Si
1.7
. This compound exhibits a large magnetic-entropy change, at least
three times larger than that of Gd, in the vicinity of its first-order simultaneous
structural and magnetic phase transition.
We have found that MnFe(P,As) compounds are very promising novel
magnetic coolants with a large MCE in the room-temperature region. In Chapter 5,
we present a study of the MCE and related physical properties, such as
magnetization and electrical resistivity, of the hexagonal MnFe
1-x
As
x
compounds.
These materials that contain elements that easily evaporate, such as P and As, have
been prepared by ball milling and subsequent solid-state reaction. The hexagonal
MnFeP
1-x
As
x
compounds exhibit a MCE that is as large as the one found in
Gd
5
(Ge,Si)
4
. This MCE is also associated with a first-order magnetic phase
transition from the low-temperature ferromagnetic (FM) phase to the high-
temperature paramagnetic (PM) phase. An applied magnetic field induces a
transition from PM state to FM state. The first-order phase transition in the
MnFe(P,As) compounds studied is adequately described by the Bean-Rodbell
model, indicating the magnetoelastic effect plays an important role in the phase
transition. Besides their large MCE, two additional features make the MnFe(P,As)
compounds promising candidate materials for magnetic refrigerants in room-
temperature applications. The first one is the fact that their Curie temperature can
be tuned between 168 K and 332 K by varying the P/As ratio between 1.5 and 0.5,
which allows one to tune the maximum MCE in this temperature range without
losing the large MCE. The second important feature is that, unlike for instance in
FeRh, the large MCE in the MnFe(P,As) compounds is reversible, and that the
Summary, Samenvatting 109

thermal hysteresis observed in the compounds is smaller than that observed in
Gd
5
(Ge,Si)
4
system. It is also commercially attractive that the price of the
MnFe(P,As) is much cheaper than Gd.
In Chapter 6, we present further studies on the magnetocaloric properties of
the MnFe(P,As)-based compounds in which the Mn/Fe ratio is varied. The
experimental results show that a small addition of Mn enhances the MCE and
reduces the thermal hysteresis. The best magnetocaloric properties are achieved for
Mn
1.1
Fe
0.9
P
0.47
As
0.53
, which exhibits a maximum adiabatic temperature change of 3
K for a field change from 0 to 1 T at 292 K. The thermal hysteresis is less than 2 K.

110 Summary

Samenvatting

Magnetisch koelen is een mogelijk alternatief voor het welbekende gascompressie-
en verdampingsproces. Een warmtepomp is namelijk ook te realiseren met een
magnetische vaste stof die een groot magnetocalorisch effect vertoont. In de
koelcyclus van zon pomp warmt de magnetische stof op als een magneetveld
wordt aangelegd en koelt af als het veld wordt uitgeschakeld. Doordat zulke
pompen zeer compact kunnen worden gebouwd, zijn ze ideaal voor allerlei huis-,
tuin- en keukentoepassingen, zoals de koelkast, een aircosysteem, de koeling van
een computer of ook een warmtepomp voor verwarming, waarvoor nu nog de
gascompressor de meest gebruikte vorm is. Andere voordelen van magnetisch
koelen zijn: zuiniger, geen gebruik van broeikasgassen en geluidsarm. Een
magnetische koeler is echter vrij duur, omdat het materiaal dat thans als actief
koelmedium wordt gebruikt (gadolinium) relatief zeldzaam is en op den duur
oplost in het water dat als warmtemedium wordt gebruikt
Onderzoek naar geavanceerde magnetische materialen die een groot
koeleffect in lage magneetvelden vertonen is daarom van groot belang. Dit
is het onderwerp van het in dit proefschrift beschreven onderzoek. Uit
theoretische overwegingen, die in hoofdstuk 2 zijn uiteengezet, volgt dat
een groot magnetocalorisch effect verwacht kan worden bij een scherpe
magnetische faseovergang. Daarom is vooral onderzoek gedaan aan
materialen die een bijzondere magnetische faseovergang vertonen. Om het
magnetocalorisch effect te bepalen, zijn magnetische en calorische metingen
verricht. Aan enkele preparaten is in samenwerking met de groep van
Professor Tishin van de Staatsuniversiteit van Moskou de
Summary, Samenvatting 111

temperatuurverandering gemeten tengevolge van een verandering van het
aangelegde magnetisch veld.
Omdat er in de literatuur discussie is ontstaan of magnetische metingen altijd
een goede beschrijving van het magnetocalorisch effect leveren, is in hoofdstuk 4
een nauwkeurige vergelijking gemaakt tussen de resultaten van calorische
metingen en van magnetische metingen. Deze metingen leveren binnen de
foutmarges eenzelfde magnetocalorisch effect voor GdRu
2
Ge
2
een materiaal dat
een zeer scherpe veldgenduceerde overgang vertoont. Hiermee is aangetoond dat
voor de bepaling van het magnetocalorisch effect van een materiaal volstaan kan
worden met de veel eenvoudigere magnetisatiemetingen. Metingen aan een
nkristal van Gd
5
Ge
2.3
Si
1.7
tonen een magnetocalor isch effect dat drie keer groter
is dan in zuiver Gd.
Een eerste grote doorbraak in magnetische koeling is de ontwikkeling van
een nieuw materiaal namelijk een legering van mangaan, ijzer, fosfor en arseen
MnFe(P,As), beschreven in hoofdstuk 5. Na het bepalen van allerlei eigenschappen
van deze legering konden wij vaststellen dat er sprake is van een geschikt
materiaal. Wij hebben aangetoond, dat het mogelijk is met dit materiaal te koelen
met lage magneetvelden gegenereerd m.b.v. permanente magneten. Door het
gebruik van het magneetveld van permanente magneten kan een warmtepomp zeer
energie-efficint gemaakt worden omdat alleen voor het warmtetransport arbeid
moet worden verricht. Een bijzonder interessante eigenschap van de nieuwe
legering, naast de veel lagere materiaalkosten, is dat de temperatuur waarbij deze
legering het beste koelt, het zogeheten werkpunt, kan worden ingesteld door een
kleine verandering van de samenstelling. Hiermee kan het werkpunt tussen 80 en
+70C worden gevarieerd. Een warmtepomp die over dit temperatuurinterval gaat
koelen zal zijn gebaseerd op een soort cascade die in meerdere stappen werkt.
Belangrijk voor de levensduur van een magnetocalorische koeler is ook het feit dat
dit materiaal niet oplost in water. Zelfs in zoutzuur lost het niet op, daarom ook is
arseen in deze legering volkomen ongevaarlijk.
112 Summary

Verdere verbetering van het magnetocalorisch effect in MnFe(P,As) is
gevonden bij verhoging van het aandeel van Mn in deze legering. Dit is beschreven
in hoofdstuk 6. De grootste adiabatische temperatuurverandering wordt gevonden
voor een legering waarin 10 % van het ijzer is vervangen door mangaan.
Ten slotte, de techniek van magnetisch koelen is bijzonder geschikt voor
kleinschalige systemen. Voor een gewone koelkast is niet meer dan 0.1 liter (een
kopje koffie) magnetisch materiaal nodig. Met magnetisch koelen kan de industrie
niet alleen koelsystemen ontwikkelen voor koelkasten, maar ook voor aircos en
voor computers. Ook voor de levensmiddelenindustrie kan het interessant zijn
koelwagens van deze koeltechniek te voorzien. Het spaart immers ruimte in
vergelijking met de huidige koelsystemen. Grote installaties als koelhuizen of
kunstijsbanen zullen hoogstwaarschijnlijk voortborduren op de geijkte
koelmethoden, omdat daar ruimtebesparing en geluidsoverlast veel minder
belangrijk zijn.

Tegus Thesis

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Tegus Thesis

Hochgeladen von

Copyright:

Verfügbare Formate

Novel Materials for Magnetic

Das könnte Ihnen auch gefallen